(cid:9)
`
`BINDING PROPERTIES OF IMMUNOGLOBULIN COMBINING
`SITES SPECIFIC FOR TERMINAL OR
`NONTERMINAL ANTIGENIC DETERMINANTS IN DEXTRAN*
`
`By JOHN CISAR,f ELVIN A . KABAT, MARIANNE M . DORNER,§ AND JERRY LIAO
`
`(From the Departments of Microbiology, Human Genetics and Development, and Neurology, College
`of Physicians and Surgeons, Columbia University, and the Neurological Institute, Presbyterian
`Hospital, New York 10032)
`
`Since the discovery by Landsteiner (1, 2) that low molecular weight haptens
`would react with antihapten antibody to inhibit precipitation by azoprotein
`antigens, and of Marrack and Smith (3) that binding of haptens to antibody
`could be demonstrated by equilibrium dialysis, such interactions have been used
`for identifying antigenic determinants and for establishing specificities and sizes
`of antibody combining sites (cf. 4, 5) .
`Specificities for carbohydrate antigens often appear to be directed toward the structure
`and linkage of terminal nonreducing sugars . This was first recognized by Goebel, Avery,
`and Babers (6), who found that the structure of the terminal nonreducing sugar was of
`predominant importance in determining cross-reactions with rabbit antisera to various
`disaccharides conjugated to horse serum globulin . Later, Karush (7) showed that
`lactoside-specific rabbit antibodies, produced by immunization with a lactosyl-azoprotein
`conjugate, had the major portion of their binding energies directed against the terminal
`,d-linked galactose . With glycoproteins and polysaccharides there are many examples of
`terminal antigenic determinants . These include the A, B, H, and Lewis specificities of
`human blood group substances (8, 9), terminal glucuronic acid residues which are
`involved in the specificity of Type II pneumoccous polysaccharide (10-13 cf. 4, 5, 14), and
`terminal fl-N-acetylglucosamine for Group A (15) and a-N-acetylgalactosamine for Group
`C (16) specificities in streptococcal polysaccharides (cf. 17) . Certain determinants in
`teichoic acids (17-21 cf. 22) and in the somatic antigens of Salmonella (23, 24) and
`Shigella (25) are also of this type .
`Antigenic determinants which do not require a terminal end group have also been
`recognized . Polysaccharides such as the type III and VIII pneumococcal polysaccharides
`which were shown to be linear (26) and thus to have one nonreducing end per milecule had
`long been known to precipitate with homologous antisera and also to cross-react (cf. 4) .
`The determinant of the Type III polysaccharide (S III) was shown to consist of three
`repeating units of the disaccharide, cellobiuronic acid (27) . Moreover, Heidelberger and
`
`Science Foundation GB 35243-X-1, X-2 and
`*Aided by a grant from the National
`BMS-72-02219-A02, and in part by a General Research Support Grant from the United States Public
`Health Service to Columbia University .
`f Present address: School of Dentistry, University of Colorado Medical Center, Denver, Colorado
`80220.
`§ Present address: Medizinische Universitaets-Klinik Bergheimer Strasse 58, 6900 Heidelberg,
`Germany.
`THE JOURNAL OF EXPERIMENTAL MEDICINE - VOLUME 142, 1975
`
`435
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1033 - Page 1
`
`

`

`(cid:9)
`
`436
`
`BINDING PROPERTIES OF IMMUNOGLOBULIN COMBINING SITES
`
`Rebers (28) attributed the cross-reactivity between anti-S IV sera and S II to (1 --a
`3) -linked L-rhamnose units which occur at intervals along linear sugar chains . Nontermi-
`nal determinants are important in Shigella (25) and Salmonella (24), examples being deter-
`minants 3 and 15 of group E Salmonella which involve repeating [a D Gal(1- 6)D Man-
`(1-4)L Rham(1--,3) ] sequences (29) . In addition, certain specificities for teichoic acids
`(30 cf . 22) are of this type . It has been suggested that steric factors play a role in the proper
`exposure of certain terminal (31-33) and nonterminal (34) sugar determinants .
`Dextrans are branched polymers of a-linked D glucopyranosyl units and their relatively
`simple structures have made them especially useful in immunochemical studies of
`homopolysaccharide antigens . Specificities involving all -" 6)- and all - 2)-, all - 3)-,
`all - 4)-linked D glucopyranosyl units have been described (35-43) . The reactions of all
`6) -specific antidextrans have been thought to occur at the nonreducing ends of glucose
`chains with the terminal all - 6)-linked D glucopyranosyl residue being immunodomi-
`nant . The evidence for this has come from the structure of these polysaccharides which
`have several nonreducing ends but only one reducing end, and from inhibition of precipi-
`tation studies with isomaltose oligosaccharides from which it appeared that a majority
`of the binding energy was contributed by the terminal nonreducing glucose (44, 45, cf.
`4, 5) . In addition, rabbits which had been immunized with bovine serum albumin (BSA)'
`conjugates of isomaltotrionic acid (IM3-CONH-BSA) (46), and isomaltohexaonic acid
`(IM6-CONH-BSA) (47), produced antibodies whose quantitative precipitin reactions
`with dextrans were similar to those of all - 6)-specific antidextrans . However, the best
`inhibitor for certain rabbit sera to IM3-CONH-BSA was larger than the trisaccharide
`which suggested as one possibility that specificities for nonterminal antigenic determi-
`nants such as (-0-(1 - 6)-D Glc-a-0-) sequences might exist (46) . Moreover, Richter (48)
`has shown that rabbit antibodies to a dextran-protein conjugate made with a dextran frag-
`ment of mol wt 4,400 could precipitate a synthetic linear dextran (49, 50) which implied
`that these antibodies were capable of reacting with nonterminal glucosyl residues . Con-
`canavalin A, which is thought to react at terminal nonreducing ends of dextran chains,
`does not precipitate with synthetic linear dextran (51) . These findings raise two questions
`of fundamental importance : (a) Are antidextrans and antibodies against isomaltosyl oligo-
`saccharides coupled to protein, mixtures of antibodies with specificities directed toward
`terminal nonreducing as well as nonterminal oligosaccharide sequences? (b) Can anti-
`bodies against terminal determinants cross-react at nonterminal locations along the dex-
`tran chain?
`The existence of several homogeneous BALB/c mouse myeloma proteins with specifici-
`ties for dextrans (52-55), fructosans (54-56), and galactans (57-59) provides a very useful
`approach to these questions . Three dextran-reactive myeloma globulins, W3129, W3434
`and QUPC 52, have been shown (55) to have all -+ 6) specificities . Proteins W3129 and
`W3434 differed slightly in specificity (55) and idiotype (60), but both showed maximum
`complementarity to isomaltopentaose (IM5) while protein QUPC 52 showed maximum
`binding for isomaltohexaose (IM6) and was not related idiotypically to W3129 .
`We have now determined the binding energies for the interactions of proteins
`W3129 and QUPC 52 with each member of the isomaltose series and with methyl
`aDglucoside . These results indicate the protein W3129 has a nonreducing
`terminal specificity for ce(1 - 6) chains of dextran while protein QUPC 52 has
`nonterminal specificity . This was confirmed by precipitin reactions with a
`synthetic linear dextran which failed to precipitate protein W3129 but precipi-
`tated QUPC 52 . The correlation between the binding properties of these
`myeloma proteins and their precipitin reactions with the synthetic dextran (49,
`50) provides further evidence that this dextran reacts immunochemically as a
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1033 - Page 2
`
`

`

`(cid:9)
`(cid:9)
`(cid:9)
`
`J. CISAR, E. A. KABAT, M. M. DORNER, AND J. LIAO
`
`437
`completely linear molecule. In addition, quantitative precipitin studies with
`linear and branched dextrans favor the concept that some human antidextran
`molecules also react only at terminal nonreducing ends while others can react at
`nonterminal locations along linear dextran chains . Moreover, rabbit antisera
`against IM3 or IM6 coupled to BSA' contain fractions of molecules which
`precipitate linear dextran and thus must be capable of nonterminal reactions
`with dextran chains . These findings provide a fundamental insight into the
`immunochemistry of dextran and probably other polysaccharide antigens .
`
`Materials and Methods
`Human and Rabbit Antisera and Mouse Myeloma Proteins .
`Human antidextrans are listed in
`Table I along with the dextrans employed for immunization . Serum 1 D., .,, was a pool obtained by
`plasmapheresis of subject 1 seven times over a period of 2 mo, and was taken 18 1/2 years after
`immunization with dextran B1255. Rabbit antisera against IM3-CONH-BSA (46) and IM6-CONH-
`BSA (47) were those studied previously . BALB/c mouse IgA myeloma proteins W3129, QUPC 52,
`W3082, and UPC 61 have been described (55) and were from serum or ascites fluid generously
`provided by Doctors Melvin Cohn and Martin Weigert (The Salk Institute, San Diego, Calif.), and
`Dr . Michael Potter of NIH.
`Purification of Human Antidextrans and Mouse Myeloma Proteins .
`Purification of antidextrans
`and myeloma proteins was by a standard batch-wise immunoadsorption procedure (63-65) using
`Sephadex G75 (Pharmacia Fine Chemicals, Inc ., Piscataway, N. J.) for human antidextrans and
`proteins W3129 and QUPC 52 and levan gel (66, 68, 69) for human antilevans and proteins W3082 and
`UPC 61 . After adsorption of protein, the insoluble gels were washed with 0.01 M phosphate-buffered
`saline, pH 7.2, 0 .02% sodium azide (PBS) until supernates gave negligible absorption at 280 nm .
`Hapten elution was at 37°C for 1 h and proteins were freed of hapten by at least two passes through
`columns of Bio-Gel P-10 (Bio-Rad Laboratories, Richmond, Calif.) (68, 69) and concentrated by
`ultrafiltration with collodion bags (Schleicher & Schuell, Inc., Keene, N. H.) or above a Diaflow
`UM-10 membrane (Amicon Corporation, Lexington, Mass .) .
`Antidextrans from 3,726 ml of 1 D,,-,o which contained 15 .1 mg antidextran nitrogen (N) were
`adsorbed on to 922 mg Sephadex G75 and eluted with IM3 and IM6 in a batch-wise modification of
`the procedure described by Harisdangkul and Kabat (68) . The packed Sephadex G75 was eluted first
`with 270 mg of IM3 in 15 ml PBS. This eluate (1 D_ IM3El 1) was concentrated and the
`ultrafiltrate containing most of the IM3 was used for the second IM3 elution and after this, the
`process was repeated a third time . The Sephadex was then washed with PBS and the washings were
`combined with the second and third IM3 eluates to give 1 DS,-, o IM3E12-3 . Then 202 mg IM3 in 15 ml
`PBS were added to the packed Sephadex and this eluate plus the next two were combined to give 1
`D6,_BO IM3E1 4-6. The washed Sephadex was then extracted six times as above starting with 90 mg
`IM6 in 15 ml PBS and these were pooled to give 1 D,,_,p IM6E1 1-6. The total recovery from all
`elutions was 10 .2 mg antidextran N (68%) of which 1 D5,_  IM3E1 1, El 2-3, El 4-6, and IM6E1 1-6
`accounted for 52, 21, 8, and 19% respectively . After removal of antidextrans, human antilevans were
`purified from 1 D5,_  sera by elution from levan gel with 0.1 M acetate buffer, pH 3 .7 (66, 68).
`Dextran- and fructosan-specific mouse myeloma proteins were eluted with methyl anglucoside
`and sucrose respectively, which had previously been dialyzed to remove polysaccharides (70) and
`recrystallized from ethanol. With protein W3129, 45 .5 mg of myeloma N was adsorbed to 4 g of
`Sephadex G75, and eluted three times with about 200 mg methyl a n glucoside per mg of myeloma N;
`the final recovery was 36 mg W3129 N (79%) . With protein QUPC 52, 92 mg of myeloma N was
`adsorbed to 6.3 g of Sephadex, washed, and eluted 10 times with methyl anglucoside at about
`420 mg methyl anglucoside per mg of myeloma N; the final recovery was only 26 mg QUPC 52 N
`(28%). The G75 was then eluted repeatedly in the cold (30 min per elution) with 0.1 M glycine HCI
`buffers ranging from pH 3.0 to pH 1.8 and eluates were neutralized immediately after collection . The
`
`this paper: BSA, bovine serum albumin; IM2, isomaltose ; IM3,
``Abbreviations used in
`isomaltotriose ; IM4, isomaltotetraose; IM5, isomaltopentaose ; IM6, isomaltohexaose; IM7, isomal-
`toheptaose ; PBS, phosphate-buffered saline .
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1033 - Page 3
`
`

`

`(cid:9)
`(cid:9)
`(cid:9)
`
`438
`
`BINDING PROPERTIES OF IMMUNOGLOBULIN COMBINING SITES
`
`TABLE I
`Human Antidextrans
`
`Linkages
`
`all -~ 6)
`
`a(1 - 3)-
`like
`
`a(1- 4)-
`like
`
`Reference
`
`86
`
`88
`
`88
`
`86
`
`96
`
`0
`
`14
`
`10
`
`4
`
`4
`
`0
`
`4, 5, 36, 37, 45, 61, 63, 65,
`66, 67, 68
`4, 5, 36, 44, 45, 61, 63, 64,
`65, 66, 67
`4, 5, 35, 36, 37, 44, 45, 61,
`66,67
`4,37
`
`4, 5, 45, 62, 64
`
`62
`
`Antidextran
`
`Immunizing
`dextran
`
`ID,+_e0
`
`201),o
`
`30D,
`
`116D,
`
`176D,
`
`219D,
`
`B1255
`(Native)
`OP 155
`(Clinical)
`OP 163
`(Clinical)
`S-5-A-1 .0
`(Clinical)
`NRC Fr . 4
`(Clinical)
`APC-54
`(Clinical)
`
`final yields were : 2.4 mg N at pH 3.0 (one elution) 11 .8 mg N at pH at pH 2.7 to 2.4 (four elutions),
`13 .7 mg N at pH 2.0 (three elutions), and 4.6 mg N at pH 1.8 (three elutions) giving a total recovery
`in the low pH elutions of 32 .5 mg QUPC 52 N (35%) . All eluates were between 90 and 95% reactive
`with Sephadex G75. With proteins W3082 and UPC 61, 1.2 mg myeloma N was adsorbed onto 65 mg
`levan gel and eluted twice with 300 mg sucrose per mg myeloma N which removed virtually all ad-
`sorbed protein.
`Dextrans .
`Native dextran B512 and clinical dextran N-150N have been described (61 cf . 4) . Prof.
`Conrad Schuerch, Chemistry Department, State University College of Forestry, Syracuse, N. Y.,
`kindly provided dextran D3 which by chemical methods appears to be a completely linear polymer of
`a (I - 6)-linked D glucopyranosyl units having a mol wt by viscosity of 36,500 (49, 50). Enzymatic
`degradation experiments indicate that synthetic dextran contains 1-2% of structural flaws (71) .
`Qligosaccharides . The isomaltose oligosaccharides, methyl anglucopyranoside, the ,B
`(2
`1)-linked series of fructose oligosaccharides and sucrose were those described (55) . Methyl
`a-isomaltoside and methyl a-isomaltotrioside (72) were provided by Dr . Allene Jeanes and have been
`studied previously (4, 65) .
`Tritium-labeled isomaltoheptitol Q1H]IM7-OH) was prepared by overnight reduction at 4°C of 90
`umol IM7 with 25 gmol [8111 sodium borohydride (383 mCi/nmol), and the reaction was stopped by
`adding small amounts of 25% acetic acid . The partially reduced and tritiated IM7 was then
`completely reduced with a 20-fold molar excess of NaBH,. After 2 days at 4°C, excess NaBH, was
`destroyed and the [IHIIM7-OH was purified by ethanol elution from a charcoal-celite column,
`followed by preparative paper chromatography with a propanol-ethanol-water (6 :1 :3) solvent system
`and final chromatography on a Bio-Gel P-2 column (minus 400 mesh, 70 x 1 .9 cm) (73) .
`Immunochemical Methods .
`Quantitative precipitin and inhibition assays were done by a
`microtechnique (4) and total N in the washed precipitates was measured by the ninhydrin method
`(74) . The separation of IgA monomer and polymer fractions of protein QUPC 52 and quantitative
`immunoadsorption experiments with dextran-reactive proteins using Sephadex G75 were like those
`described for protein W3129 (55) .
`Isoelectric Focusing.
`Analytical isoelectric focusing in a thin layer of 5% polyacrylamide gel (3%
`cross-linked) containing 2% carrier ampholytes (ampholine pH 3.5-10) was performed with an LKB
`2117 Multiphor unit according to the directions of the manufacturer (LKB-Producter AB, Sweden).
`Gels were fixed overnight in 12% trichloroacetic acid then rinsed in several changes of water to remove
`carrier ampholytes before staining with 0.02% Coomassie Blue in methanol-acetic acid-water
`(45:9 :46) and destaining with a solution of methanol (25%) and acetic acid (10%) . Preparative
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1033 - Page 4
`
`

`

`(cid:9)
`(cid:9)
`(cid:9)
`
`439
`J. CISAR, E. A. KABAT, M. M. DORNER, AND J. LIAO
`isoelectric focusing of 1 D-,o IM3El 2-3 (1.6 mg N) was done in a 110 ml LKB column at 4°C using
`3% carrier ampholytes (pH 3.5-10) in a sucrose gradient as described by the manufacturer .
`Electrofocusing was for 3 days with a final potential of 700 V.
`Equilibrium Dialysis .
`Equilibrium dialysis and displacement experiments were carried out in
`lucite microdialysis chambers (75) at 25°C f 0.1 ° and PBS was the buffer in all experiments. Dialysis
`tubing was boiled in 2 M Na,CO, and rinsed thoroughly before use. Control experiments (4)
`established that there was no detectable bindingofthe ligand [9H ]IM7-OH to the dialysis membrane,
`to normal human -y-globulin (4.1 mg/ml) or to the levan-reactive IgA mouse myeloma protein Y5476
`(14 mg/ml) (55) . Experiments were done with equal volumes (50 or 75 u1) on each side of the
`membrane and dialysis cells were allowed to reach equilibrium over a 72-h period of continuous
`mixing ; however, displacement experiments with protein QUPC 52 were given an extra day because
`of the high concentrations of competitors employed . Control experiments indicated that equilibrium
`was probably attained in less than 24 h. Samples (20 ,l) for radioactivity measurements were taken
`from the hapten side only (76) and counted in duplicate or triplicate . The [9H]IM7-OH gave about 2.9
`x 10° cpm/,mol in Bray's solution (77) and 3.5 x 10° cpm/,umol in Insta-Gel (Packard Instrument
`Co., Downers Grove, Ill.). In displacement experiments, nonlabeled competitors did not cause any
`detectable quenching ofcounts except for the highest amounts of methyl a D glucoside employed with
`protein QUPC 52 and this occurred only in Insta-Gel and was corrected for. The free concentration of
`competitor was calculated by the method of Nisonoff and Pressman (78).
`Fluorescence Titrations .
`Fluorescence titrations were performed with square quartz semimicro
`cuvettes (5 mm light path, 0.6 ml capacity) and an Aminco-Bowman spectrophotofluorometer
`equipped with a cell jacket thermostated at 25°C (68) . Haptens were added in 5-20,ul volumes to 250
`,ul portions of protein solutions having optical densities of about 0.18 (1 cm light path) at 280 nm.
`None of the haptens absorbed in the 280 nm region and the excitation wave length was set between
`280 and 285 nm depending on the protein and emission was measured between 345 and 350 nm.
`Values for Q... were determined at the end of each titration by averaging the results from two
`additions of a concentrated ligand solution, each resulting in a greater than 97% saturation of sites.
`All titrations were done against a reference cell containing the solution being titrated and in this way
`it was possible to detect and correct for instrument drift. It was also necessary to plan experiments so
`that each addition of hapten resulted in at least a 3% quenching of fluorescence after correction for
`dilution .
`
`Results
`Purified antidextrans from 1
`Isoelectric Focusing of Human Antibodies.
`D5,-60 were 90-95% reactive with Sephadex G75 and analytical isoelectric
`focusing revealed a restriction in heterogeneity as compared with whole -y-
`globulin from the same individual . The patterns for antidextrans eluted with IM3
`(Fig. 1 A,IM3El 1, IM3E1 2-3, and IM3E1 4-6) were similar with three major
`bands near pH 7.0 and several less prominent bands between pH 6.0 and pH 8 .5 .
`The antidextrans eluted with IM6 (Fig. 1 A, IM6E1 1-6) were qualitatively like
`the IM3 eluates, but the three major bands near pH 7 .0 were less obvious while
`those between pH 7.0 and pH 8 .0 accounted for a greater proportion of the total
`protein . Purified antilevan from 1 D,60 (Fig. 1 A, L) and antibodies to blood
`group A substance from 1 D33-a, (Fig. 1 A, A) also were restricted in heterogeneity
`and visibly different from each other and from the 1 D54-60 antidextrans . The
`anti-A antibodies had relatively low isoelectric points . A pH 3.5 to pH 10 gradient
`was used for the preparative separation of 1 Ds,-6o IM3E1 2-3 into seven fractions
`(Fig. 1 B and C) which were studied later.
`Equilibrium Dialysis of Mouse Myeloma Proteins and Human Antidextrans.
`Purified BALB/c mouse myeloma proteins W3129 and QUPC 52 were approxi-
`mately 95% reactive with Sephadex G75, had two binding sites for [3H)IM7-OH
`per IgA monomer, and gave linear Scatchard plots indicating homogeneous bind-
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1033 - Page 5
`
`

`

`(cid:9)
`(cid:9)
`
`440
`
`BINDING PROPERTIES OF IMMUNOGLOBULIN COMBINING SITES
`
`A. Analytical isoelectric focusing of 7-globulin (,yG) and purified antibodies from
`1 .
`FIG.
`subject 1. Antidextrans from 1 D5 ,_ eo were those eluted from Sephadex G75 by sequential
`extraction with IM3 (IM3 eluate 1, IM3 eluate 2-3, and IM3 eluate 4-6) then IM6 (IM6 eluate
`1-6) . The antilevan (L) from 1 D,,_ 8, and anti-blood group A antibodies (A) from 1 D- yG
`fraction of Ga1NAc eluate (79) are shown. Approximately 5.2 Ag N of all samples were applied
`to the gel . (B) Preparative isoelectric focusing of 10 mg 1 D6,-eo IM3 eluate 2-3 showing how
`fractions I to VII were collected. (C) Analytical isoelectric focusing of 1 D,,-,, IM3 eluate 2-3
`and fractions I to VII from B. Approximately 3.3 kg N of all samples were applied to the gel.
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1033 - Page 6
`
`

`

`
`
`(cid:9)(cid:9)
`
`J. CISAR, E. A. KABAT, M. M. DORNER, AND J. LIAO
`441
`ing (Fig. 2 A), as did the heterogeneity indices of 1.0 from Sips plots (not shown)
`for protein QUPC 52 and 0.9 for W3129. Both proteins were mixtures of mono-
`meric and polymeric IgA, and the homogeneous binding shows that polymeriza-
`tion of IgA monomers does not alter the binding constants toward monovalent
`haptens. The binding constant of protein W3129 (K11 = 1 .0 x 105 M- ') was about
`10-fold greater than that of protein QUPC 52 (K11 = 8.4 x 103 M- ') .
`As expected from their isoelectric focusing patterns (Fig. 1 A), binding of
`[3H]IM7-OH by human antidextrans 1 D5,-1111 IMM 1 and 1 De,-s11 IMM 1-6 was
`heterogeneous as revealed by nonlinear Scatchard plots (Fig. 2 B) . From Sips
`
`Scatchard plots of equilibrium dialysis data at 25°C with ['Hlisomalthoheptitol . (A)
`FIG. 2.
`Protein W3129 (") at 10 .5 mg/ml, and protein QUPC 52 (O) at 20 .6 mg/ml. (B) Purified
`human antidextrans, 1 D_ IM3 eluate 1 (O) at 6.68 mg/ml, and 1 D,,_,o IM6 eluates 1-6 (")
`at 2.56 mg/ml. Calculations were done with a mol wt of 150,000 for myeloma proteins and
`human antidextrans .
`
`analyses (not shown) heterogeneity indexes of 0 .6 were calculated for both the
`IM3 and IM6 antibody . Association constants for 1 D5,-1111 IMM 1 (K11 =1 .1 x 105
`M- ') and 1 Ds,-611 IMM 1-6 (K11 =1 .8 x 105 M- ') differed slightly and were
`similar to the value obtained for myeloma protein W3129.
`Equilibrium Dialysis Displacement Studies with Myeloma Proteins W3129
`Association constants for isomaltose oligosaccharides and
`and QUPC 52.
`methyl aDglucoside were determined by measuring the abilities of unlabeled
`oligosaccharides to displace [3H]IM7-OH from the binding sites of proteins
`W3129 and QUPC 52. For both proteins, displacement curves with IM6 and IM7
`were the same as self displacement curves calculated (78) for [3H]IM7-OH and,
`thus, the association constants for IM6 and IM7 with proteins W3129 and QUPC
`52 are 1 .0 x 105 M- ' and 8.4 x 103 M- ' respectively. With each protein, a
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1033 - Page 7
`
`

`

`442
`
`BINDING PROPERTIES OF IMMUNOGLOBULIN COMBINING SITES
`
`displacement curve was calculated for a hypothetical competitor with K=1 M-1
`and thus a AF' = 0 cal/mol (8F0 = -2.303 RT log K), and this serves as a use-
`ful reference as the distance between it and the displacement curve of a given
`competitor is directly proportional to the AF' for the competitor .
`With protein W3129 (Fig . 3 A and Table II) over half of the binding energy for
`IM6 and IM7 is directed against methyl aDglucoside or IM2 which are similar.
`These are much less active than IM3 or IM4 which are equal and are bound
`almost as firmly as longer oligosaccharides. Isomaltopentaose is most efficient
`at displacing [3H ]IM7-OH' and even somewhat more active than IM6 or IM7.
`
`ISMALTOHEPTAOSE
`o ISOMALTOHEXAOSE
`ISOMALTOPENTAOSE
`a SOMA LTOTETRAOSE
`ISOMALTOTRIOSE
`A ISOMALTOSE
`¢ METHYL X D GLUCOSIDE
`
`IM 5
`IM 4
`o f
`` IM 3
`
`A
`
`A
`A
`
`X
`A d a
`Me DylX'
`
`Ms'
`IM 5
`1M 4
`1M 3
`MeelQgic
`
`e.4 X103
`5.9 X 103
`2.e X 103
`z0X 102
`'"1" 6 X 100
`
`i
`
`I
`
`i
`
`I
`
`~
`
`A
`
`1
`
`5
`m
`b 4
`_o
`
`3
`
`F
`
`Q
`
`F
`
`UOV
`
`0
`
`-40
`
`0
`
`+I .0
`
`-L0
`-3.0
`LOG FREE CONCENTRATION OF COMPETITOR (M)
`FIG. 3 . Equilibrium dialysis displacement at 25°C of ['Hlisomaltoheptitol by unlabeled
`isomaltose oligosaccharides and methyl aDglucoside. (A) Protein W3129 (10.5 mg/ml) with a
`concentration of ['H]isomaltoheptitol giving r = 0.62 in the absence of competitor . (B) Protein
`QUPC 52 (20.6 mg/ml) with a concentration of ['H]isomaltoheptitol giving r = 0.39 in the
`absence of competitor. The dashed displacement curve corresponds to a hypothetical
`competitor with Ke = 1 M- `. The dashed horizontal line signified the concentration of bound
`['H]isomaltoheptitol in the absence of competitor .
`
`Thus, while maximum complementarity is for IM5, 91% of the total binding
`energy is directed toward IM3 and 56% is for methyl aDglucoside.
`The pattern of displacement reactions for protein QUPC 52 (Fig. 3 B and Table
`II) were strikingly different than those of myeloma W3129. Most significant were
`the findings that displacement by methyl aDglucoside and IM2, which could not
`be studied at higher concentrations than those shown in Fig. 3 B, were close to
`the hypothetical displacement curve having K=1 M-1, and these competitors
`were bound at QUPC 52 sites with only about 5% of the total binding energy for
`IM6 or IM7 (Table II) . By contrast, IM3 was much more active and 72% of the
`total binding energy was directed against the trisaccharide. There were increases
`
`
`(cid:9)(cid:9)
`
`
`(cid:9)(cid:9)(cid:9)
`
`(cid:9)(cid:9)
`
`
`
`
`
`
`
`(cid:9)(cid:9)(cid:9)(cid:9)(cid:9)(cid:9)(cid:9)(cid:9)
`
`
`
`
`
`
`
`
`
`(cid:9)(cid:9)(cid:9)(cid:9)(cid:9)
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1033 - Page 8
`
`

`

`
`
`
`
`(cid:9)(cid:9)(cid:9)
`
`.r
`
`_
`
`O F
`" F
`
`e
`
`. O
`.-~ p,
`
`F
`
`aC
`
`CO
`
`'o
`U A o .n
`w w
`
`.2
`
`Q~
`
`~
`
`M
`
`W
`
`~ ~ ~ S
`u G
`
`r,
`
`G0
`
`_
`
`F
`O
`
`y0
`
`d V
`
`S OMB
`
`rM- u°-~
`
`O O O O
`~f] O ~0 O
`
`I
`
`I
`
`I
`
`I
`
`O O O_ O
`X X X X
`~=. o
`O
`
`c~S
`
`CD omc-
`oao'a
`
`_ F
`
`o '~ O
`a+
`O
`U FA
`
`OD
`G
`
`~
`
`A
`
`w
`4 v
`
`cN m
`
`O
`
`(V
`O
`7 v
`
`FOM . Or
`a."
`
`v
`
`;,
`
`E
`F
`0 0 `d w
`o F
`A
`
`U A
`
`CD A
`
`'y' C a
`
`O O O
`X X X
`o, o 0
`
`N
`O
`~ X
`ao
`
`O
`
`O N
`cc) ~o
`
`m wM 0C
`
`TL
`
`LV
`
`T
`N
`T
`c
`
`W is
`
`d U
`
`0 0 o
`
`S
`c- rvi r~
``
`I
`I
`I
`I
`
`O O O O
`X X X X
`M O O N
`c-
`N .+ 1~ M
`
`dO
`
`N
`y
`a~ .O O . Oy
`- -O o
`
`O
`-Q t
`w u m o
`m
`tD CD CD T
`T
`T
`T
`0
`
`.
`
`id
`U
`m
`eon
`0
`
`F
`O
`
`eo
`F
`
`m
`
`wv
`
`FC
`
`N
`
`a
`
`Y
`
`Q
`
`Ca
`
`ro
`
`Ca
`
`.
`
`e
`v
`oa ~
`
`w
`I
`
`_ 0
`x
`
`~ c
`I
`I
`I
`
`0 0 0
`x x x
`.~ M N
`
`N O ~'
`Q) CZ) O
`
`VO' S ~
`CJ ~ N d'
`
`O
`~/
`
`000
`~
`
`Q L
`
`F
`
`raq
`A i0
`.~
`OL
`
`C
`
`G O
`7 O+ G
`U-D 8~
`w A
`
`I
`
`I
`
`I
`
`I
`++
`- O O O O O
`X
`X X X X
`d; o c? ~ e;
`co v .. M
`o
`
`N El L
`'U `~
`vp
`41 .
`
`UP.
`
`O
`C~'
`
`ti
`T
`is
`
`s S~m~"'~ n
`
`s0.
`a
`
`R a
`
`ro
`a
`
`.
`
`d
`
`U
`
`wD
`
`T
`
`TOFW wO
`
`0a C
`
`7 WN
`
`S T a~0 t VO' N
`
`O O O O O O
`O <Y M CD t . d'
`C~J M M O (p o
`CD M M
`I
`1
`I
`
`O O O O O O
`X X X X X X
`cD u7 u; ., O c0,
`M N I oo I
`
`~0 (b M M ~
`
`S
`
`0 0 0 0 o 0
`o~fJ
`~
`~ N N N
`N [~ u] M M .y
`I
`i
`I
`I
`I
`I
`
`0 0 0 0 0 0
`
`F
`O
`
`7
`.L
`
`w G
`A
`
`Gr. m
`d V
`
`en
`F
`TS
`
`F A
`
`C F
`C '~
`
`O ~
`O
`
`d U
`
`V
`w 'o
`w-
`"" O
`U
`p., A
`.C
`G .E
`
`o
`
`.
`
`F R
`
`a
`CV p
`o Y
`3 ;a
`F F
`
`SJ
`a
`
`dO
`
`O
`
`14,
`Q.
`
`"
`
`m
`C
`Tf
`
`d'
`OJ L~ a0 W
`M o~NN
`u~ a' c,
`I
`I
`I
`I
`O O O O
`O O
`X X X X X X
`d' O CA O O (O
`
`-
`
`g g a 6,
`
`O O S S O COD
`.~ c~
`ci
`N t` CO CD M M
`I
`+
`- O O O O O O
`xxxxxx
`c- ~ ~n ~r o 0
`.~
`.. co co co ao
`
`T W S S
`
`pO
`
`6° S
`
`w O cC F
`c'
`ov
`o a
`F
`U A N
`Cs . m
`d O
`
`o
`U A O A
`
`W is
`d U
`
`"
`
`EM
`bi
`
`_.
`
`o
`A U
`,I7
`c en
`C (y
`'sy xn
``-'
`
`a
`
`C
`
`O
`
`M
`
`0 d
`
`.
`
`d 0 C C
`
`X X X X X X
`O O N to fD O
`M '+ N N N m.
`
`s 88
`
`O O
`
`c= c~
`II
`
`O O
`
`X X
`to o
`C+J M
`
`O
`
`c+S
`I
`
`O
`
`x
`M
`O
`
`_ g
`O 'O
`
`F A
`
`- F
`O
`
`U a
`
`w m
`u
`4
`
`a
`
`row W
`
`O ' m
`M o W
`m F'" c
`C " O
`o
`Q U w
`{ZI a +-~
`
`d
`000 m
`O v
`N
`
`EOw
`
`N
`F
`V
`m
`
`7 E
`
`w
`
`F c6 C
`o
`'_
`° ~
`.G
`UA
`w Cd
`O
`Q
`
`O 7
`
`U A O
`
`, F
`
`W
`d V
`
`F.
`d
`o
`a.
`
`y Q
`" U
`
`F 00
`~r~
`m
`
`w
`TS Q
`EM
`00
`~A_
`
`o 0 0 0 oom
`Soo
`"~ "~ CD CD GV N
`t- n CO CO d' V'
`I
`I
`I
`I
`I
`I
`
`0 0 0 0 0 0
`
`xxxxxx
`rt- Lq
`
`G To
`
`000 00
`LnLn °O
`cD P CO CD C er
`I
`I
`I
`I
`I
`I
`O O O O O O
`
`-o
`
`~
`
`p b
`
`C- ~fJ e M N aA.,
`
`:U
`
`W
`o
`
`=
`
`o y
`F O
`F
`cp.
`F.
`
`a0 0U U
`D 00
`O
`
`T
`
`T
`
`oO
`TO
`
`C G
`v U
`F
`W F
`p O V
`V U y
`F
`
`Q O
`
`7b
`
`.No_.
`~csc~o o
`U
`wWWW
`M Pr.
`Co CV N
`
`wNw
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1033 - Page 9
`
`

`

`(cid:9)
`
`
`
`
`
`
`
`(cid:9)(cid:9)(cid:9)(cid:9)
`
`444
`
`BINDING PROPERTIES OF IMMUNOGLOBULIN COMBINING SITES
`
`in binding energy between IM3 and IM4 and from IM4 to IM5 . IM5 seemed to be
`slightly less active than IM6 or IM7 but this difference, although known to exist
`from inhibition studies (55) is so small that it probably is within experimental
`error in the present data. Thus, maximum complementarity is for the hexasac-
`charide, and while 72% of the total binding energy is for IM3, almost none of this
`can function in the binding of IM2 or methyl a D glucoside.
`Fluorescence Studies with Mouse Myeloma Proteins and Human Antidex-
`trans. The binding of isomaltose oligosaccharides and methyl a D glucoside by
`protein W3129 was associated with a quenching of the protein's fluorescence and
`Q,,,. values from several titrations varied from 16 to 20% with IM3 through IM7
`and from 20 to 23% with IM2 and methyl a D glucoside. The binding data (Fig. 4
`A) from fluorescence titrations are in good agreement with results from
`equilibrium dialysis displacement and from quantitative precipitin inhibition
`assays (Table U) . However, one difference was that IM5, IM6, and IM7 gave
`similar association constants in fluorescence titrations while in equilibrium
`dialysis displacement IM6 and IM7 were less active than IM5 . Both methods
`show that maximum binding occurs with IM5 and gave similar Ko values for its
`binding by protein W3129. Titrations with monomer and polymer fractions (not
`shown) gave the same results as with unfractionated W3129 and supported the
`previous conclusion that polymerization of IgA subunits does not effect binding
`of a monovalent hapten . Compounds other than isomaltose oligosaccharides also
`
`PROTEIN W.3129
`
`- PROTEINS W3082 and UPC61
`
`I
`-4 .0
`
`I
`-3,0
`LOG C (M)
`(A) Sips plots of data from fluorescence titrations at 25°C of dextran-specific
`FIG. 4 .
`myeloma protein W3129 with isomaltose oligosaccharides and methyl anglucoside (symbols
`as in Fig. 3) and of (B) fructosan-specific myeloma proteins W3082 and UPC 61 with IB (2 --.
`1)-linked fructose oligosaccharides (2FlG and 3F1G) and with sucrose. All proteins were at
`20 fag N/ml, and the excitation and emission wave lengths with W3129 were 280 and 345 nm
`respectively, and 285 and 345 nm respectively with W3082 and UPC 61 .
`
`1
`-2 .0
`
`1
`-I .0
`
`I
`0
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1033 - Page 10
`
`

`

`(cid:9)
`
`J. CISAR, E. A. KABAT, M. M. DORNER, AND J. LIAO
`
`445
`
`quenched the fluorescence of protein W3129, but with these it was often not
`possible to reach Q... . By using QMRX obtained from adding IM5, association
`constants were estimated for the a (1 -" 2), a (1 -" 3), and a (1-+ 4) -linked glucose
`disaccharides, kojibiose, nigerose, and maltose, respectively, as being between
`102 M-1 and 103 M- ', and were lower than those of IM2 and methyl aDglucoside.
`Methyl #Dglucoside, methyl aDglucoside, methyl aDmannoside, trehalose (aD-
`glucopyranosyl aDglucopyranoside) and glucose all had K,, values less than 5 x
`10' M- '. These findings emphasize the specificity which the W3129 site has for
`isomaltosyl structures. The binding of isomaltose oligosaccharides by protein
`QUPC 52 was not associated with a change in fluorescence .
`The fructosan-specific myeloma proteins W3082 and UPC 61 (55) were
`indistinguishable and had Qmax values from 6 to 9% with the tetrasaccharide
`MG, 5 to 7% with the trisaccharide 2F1G, and 10 to 13% with sucrose. Both
`proteins had association constants of 3.6 x 105 M- ' with 3FIG and 2F1G and 6.3
`x 102 M- ' with sucrose (Fig. 4 B and Table II) . Thus, 50% of the total binding
`energy was directed against the nonreducing, ,Q-linked, fructosyl unit of sucrose
`which represents the immunodominant group. The oligosaccharides 2FIG and
`3FIG present two and three fructosyl units respectively in # (2 , 1) linkage plus
`an additional ,B linkage and were similar, although earlier studies (55) showed
`that 3FIG was slightly more active than 2F1G (Table II) . In the fluorescence
`data this small difference would be expected to be within experimental error.
`Purified human antidextrans were then