Downloaded from
`Downloaded from
`
`http://www.jimmunol.org/
`http://www.jimmunol.org/
`
` by guest on November 16, 2015
` by guest on November 16, 2015
`
`This information is current as
`of November 16, 2015.
`
`Binding Motifs of Copolymer 1 to Multiple
`Sclerosis- and Rheumatoid
`Arthritis-Associated HLA-DR Molecules
`Masha Fridkis-Hareli, John M. Neveu, Renee A. Robinson,
`William S. Lane, Laurent Gauthier, Kai W. Wucherpfennig,
`Michael Sela and Jack L. Strominger
`
`1999; 162:4697-4704; ;
`J Immunol(cid:160)
`http://www.jimmunol.org/content/162/8/4697
`
`References
`
`cites 46 articlesThis article
`
`, 22 of which you can access for free at:
`http://www.jimmunol.org/content/162/8/4697.full#ref-list-1
`
`Subscriptions
`
`Information about subscribing to is online at: The Journal of Immunology
`
`http://jimmunol.org/subscriptions
`
`
`
`Permissions
`
`Submit copyright permission requests at:
`http://www.aai.org/ji/copyright.html
`
`Email Alerts
`
`Receive free email-alerts when new articles cite this article. Sign up at:
`http://jimmunol.org/cgi/alerts/etoc
`
` is published twice each month by
`The Journal of Immunology
`The American Association of Immunologists, Inc.,
`9650 Rockville Pike, Bethesda, MD 20814-3994.
`Copyright © 1999 by The American Association of
`Immunologists All rights reserved.
`Print ISSN: 0022-1767 Online ISSN: 1550-6606.
`
`Page 1 of 9
`
`YEDA EXHIBIT NO. 2036
`MYLAN PHARM. v YEDA
`IPR2015-00644
`
`

`
`Downloaded from
`
`http://www.jimmunol.org/
`
` by guest on November 16, 2015
`
`Binding Motifs of Copolymer 1 to Multiple Sclerosis- and
`Rheumatoid Arthritis-Associated HLA-DR Molecules1
`
`Masha Fridkis-Hareli,* John M. Neveu,† Renee A. Robinson,† William S. Lane,†
`Laurent Gauthier,‡ Kai W. Wucherpfennig,‡ Michael Sela,§ and Jack L. Strominger,2*‡
`
`Copolymer 1 (Cop 1, poly (Y, E, A, K)) is a random synthetic amino acid copolymer effective in the treatment of relapsing forms
`of multiple sclerosis (MS). Cop 1 binds promiscuously, with high affinity and in a peptide-specific manner to purified MS-
`associated HLA-DR2 (DRB1*1501) and rheumatoid arthritis-associated HLA-DR1 (DRB1*0101) or HLA-DR4 (DRB1*0401)
`molecules. In the present work at least 95% of added Cop 1 could be bound to recombinant “empty” HLA-DR1 and -DR4, and
`80% could be bound to HLA-DR2 proteins. Amino acid composition, HPLC profiles, and sequencing patterns of Cop 1 eluted by
`acid extraction from HLA-DR molecules were similar to those of the unseparated Cop 1. Protruding N-terminal ends of Cop 1
`bound to HLA-DR1, -DR2, or -DR4 molecules were then treated with aminopeptidase I, followed by elution, HPLC, and pool
`sequencing. In contrast to untreated or unbound Cop 1, this material exhibited distinct motifs at some positions with increases in
`levels of E at the first and second cycles, of K at the second and third cycles, and of Y (presumably at P1 of the bound peptide)
`at the third to fifth cycles, regardless of the HLA-DR molecule employed. No preference was seen at the following cycles that were
`mainly A. These first pooled HLA-DR binding epitopes provide clues to the components of Cop 1 that are biologically active in
`suppressing MS and possibly rheumatoid arthritis. The Journal of Immunology, 1999, 162: 4697– 4704.
`
`C opolymer 1 (Cop 13, poly (Y, E, A, and K)) is a synthetic
`
`amino acid copolymer effective both in suppression of
`experimental allergic encephalomyelitis (EAE) (1–12)
`and in the treatment of relapsing forms of multiple sclerosis (MS)
`(13, 14). The mechanisms proposed for the activity of Cop 1 in-
`volve binding to class II MHC molecules on APCs (9), leading to
`induction of Ag-specific suppressor cells (4, 6) and/or competition
`with myelin Ags for activation of specific effector T cells (7, 8).
`Indeed, the binding of Cop 1 to purified human HLA-DR mole-
`cules within the peptide binding groove has been reported (15).
`Cop 1 inhibited the binding of HA 306 –318 peptide, a high-affinity
`epitope of influenza virus, to both HLA-DR1 (DRB1*0101) and
`-DR4 (DRB1*0401) molecules, and of myelin basic protein
`(MBP) 84 –102, a human immunodominant epitope of MBP, to
`HLA-DR2 (DRB1*1501) molecules (15). Moreover, Cop 1 has
`been recently found to compete with collagen type II (CII) 261–
`273, a candidate autoantigen in rheumatoid arthritis (RA), for
`binding to RA-associated HLA-DR1 (DRB1*0101) and -DR4
`
`*Department of Molecular and Cellular Biology, and †Microchemistry Facility, Har-
`vard University, Cambridge, MA 02138; ‡Dana Farber Cancer Institute, Harvard
`Medical School, Boston, MA 02115; and §Department of Immunology, The Weiz-
`mann Institute of Science, Rehovot, Israel
`
`Received for publication July 14, 1998. Accepted for publication January 19, 1999.
`
`The costs of publication of this article were defrayed in part by the payment of page
`charges. This article must therefore be hereby marked advertisement in accordance
`with 18 U.S.C. Section 1734 solely to indicate this fact.
`1 This work was supported by grants from the National Institutes of Health (R35-CA
`47554 and N01-AI 45198) and Teva Pharmaceutical Industries, Ltd., Israel. M.F.-H.
`is the recipient of a National Multiple Sclerosis Society advanced postdoctoral fel-
`lowship. K.W.W. is a Harry Weaver Neuroscience Scholar of the National Multiple
`Sclerosis Society.
`2 Address correspondence and reprint requests to Dr. Jack L. Strominger, Department
`of Molecular and Cellular Biology, Harvard University, 7 Divinity Avenue, Cam-
`bridge, MA 02138. E-mail address: jlstrom@fas.harvard.edu
`3 Abbreviations used in this paper: Cop 1, copolymer 1; CII, collagen type II; HA,
`influenza virus hemagglutinin; MBP, myelin basic protein; MS, multiple sclerosis;
`RA, rheumatoid arthritis; EAE, experimental allergic encephalomyelitis; RP, reverse
`phase.
`
`(DRB1*0401) molecules, and to inhibit CII-reactive T cell clones
`(16). The characterization of the active component(s) of the mix-
`ture of random polypeptides has thus been of particular importance
`in view of the therapeutic applications of Cop 1 in MS and pos-
`sibly RA patients.
`Because Cop 1 is a mixture of random polypeptides, it may
`contain different sequences that bind to different HLA proteins;
`in this case only a fraction out of the whole mixture would be
`an “active component.” Alternatively, the whole mixture may
`be competent, i.e., all polypeptides binding to any HLA-DR
`molecule. In view of the crystallographic analysis of several
`HLA-DR molecules complexed with the antigenic peptides
`(17–19), as well as the binding motifs of natural MHC-associ-
`ated ligands that were elucidated by acid-extraction and se-
`quencing (20 –25), it has been intriguing to find out whether all
`four amino acids that compose Cop 1 are involved in its binding
`in the groove of HLA-DR molecules.
`The study here was undertaken to attempt to identify these ac-
`tive components present in the bound Cop 1 and to determine their
`binding motifs. To isolate the bound fraction of Cop 1 with no
`interference from endogenous peptides, recombinant “empty”
`HLA-DR1, -DR2, and -DR4 molecules produced in insect cells
`were employed. Because the average length of the Cop 1 polypep-
`tides used is 75– 80 amino acids, the epitopes lying in the groove
`of HLA-DR molecules are likely to be found internally in the
`polypeptide chains with protruding ends, making direct analysis of
`the bound amino acid sequences by microchemical methods very
`complicated. To access these regions, N-terminal peptidase treat-
`ment of the protruding ends of Cop 1 polypeptides was employed.
`This approach proved to be useful in trimming of N-terminal ends
`of peptides that protrude out of class II MHC proteins, while pro-
`tecting epitopes bound to the groove from proteolysis (26, 27).
`Various characteristics relating to the fraction bound and its mo-
`tifs, including amino acid composition, HPLC profiles, and pool
`sequencing, together with the immunological recognition of these
`fractions, are presented here.
`
`Copyright © 1999 by The American Association of Immunologists
`
`0022-1767/99/$02.00
`
`Page 2 of 9
`
`YEDA EXHIBIT NO. 2036
`MYLAN PHARM. v YEDA
`IPR2015-00644
`
`

`
`Downloaded from
`
`http://www.jimmunol.org/
`
` by guest on November 16, 2015
`
`4698
`
`BINDING MOTIFS OF COPOLYMER 1 TO HLA-DR1, -DR2, AND -DR4 MOLECULES
`
`Materials and Methods
`Protein expression and purification
`
`Soluble HLA-DR1, -DR2, and -DR4 molecules were expressed in Dro-
`sophila S2 cells as described (18, 28, 29). Cells were grown in roller bottles
`in ExCell 401 medium (JRH Biosciences, Lenexa, KS) supplemented with
`0 –5% fetal bovine serum (Sigma Chemicals, St. Louis, MO) at 26°C. Cells
`were harvested 4 –5 days after induction by 1 mM CuSO4. Immunoaffinity
`purification of HLA-DR1, -DR2, and -DR4 molecules was performed as
`reported previously (18). Briefly, supernatant from harvested cells was se-
`quentially passed through protein A, protein G, and protein A-LB3.1 col-
`umns, followed by elution of the bound HLA-DR with 50 mM 3-[cyclo-
`hexylamino]-1-propanesulfonic acid (CAPS) (pH 11.5) and neutralized
`with 200 mM phosphate (pH 6.0). Proteins were concentrated on a Cen-
`triprep 10 membrane (Amicon, Beverly, MA).
`
`Antigens and Antibodies
`
`Cop 1 is a synthetic random copolymer prepared by polymerization of the
`N-carboxyanhydrides of L-tyrosine, g-benzyl-L-glutamate, L-alanine, and
`e,N-trifluoroacetyl-L-lysine (1) (the end product is a mixture of acetate salts
`of random polypeptides). Cop 1, batch 52596, in the molar ratio of 1 Y:1.5
`E:4.3 A:3.3 K, with an average m.w. of 8150, was obtained from Teva
`Pharmaceutical Industries (Petach Tikva, Israel). Rabbit anti-Cop 1 poly-
`clonal Abs (IgG fraction, biotin-labeled) were also from Teva Pharmaceu-
`tical Industries.
`
`Treatment of HLA-DR-Cop 1 complexes with aminopeptidase I
`
`Cop 1 (1 mM) was incubated with recombinant water-soluble “empty”
`HLA-DR1, -DR2, or -DR4 molecules (100 mM) in PBS for 40 h at 37°C.
`Aminopeptidase I (2 units) (Sigma Chemicals) was added for the last 18 h
`of incubation. Samples were then spin-concentrated to a final volume of
`;100 ml using Centricon 10 ultrafiltration devices (Beverly, MA). Bound
`Cop 1 was eluted from HLA-DR by addition of acetic acid (10%) and
`incubated at 70°C for 15 min, followed by ultrafiltration and vacuum con-
`centration in a SpeedVac (Savant), as described previously (12).
`
`HPLC separation and microsequencing
`
`After elution as above, typically 5–10% of the Cop 1 mixtures were frac-
`tionated by microbore HPLC using a Zorbax C18 1.0-mm reverse-phase
`(RP) column (Saratoga, CA) on a Hewlett-Packard 1090 HPLC with 1040
`diode array detector. At a flow rate of 54 ml/min, Cop 1 was eluted with a
`gradient of 0.055% trifluoroacetic acid in acetonitrile (0% at 0 –10 min,
`33% at 73 min, and 60% at 105 min). Strategies for peak selection, RP
`separation, and Edman microsequencing have been described previously
`(22, 30). Pooled fractions were submitted to automated Edman degradation
`on a Hewlett-Packard G1005A protein sequencer using the manufacturer’s
`Routine 3.5 analytical method.
`
`PAGE
`
`SDS-PAGE was conducted with the NOVEX (San Diego, CA) minicell
`electrophoresis system. Separation gel was 10% in acrylamide and stacking
`gel was 5%. HLA-DR1-Cop 1 complexes were run under nonreducing
`conditions for 1 h at 200 V, stained with Coomassie Brilliant blue, fixed for
`3 h in 10% methanol/10% acetic acid, and dried on Cellophane paper
`(Bio-Rad, Richmond, CA) at 25°C.
`
`Ab binding assay
`
`The cross-reactivity between Cop 1 and its fractions was detected by direct
`ELISA assay using biotinylated anti-Cop 1 polyclonal Abs. Cop 1 or frac-
`tions were diluted to 0.4 and 2.0 mg/ml and plated in duplicate on a 96-well
`microtiter immunoassay plates (PRO-BIND, Falcon, Lincoln Park, NJ)
`(100 ml per well). All incubations were for 1 h at37°C and washes were
`three times with Tris-buffered saline (TBS)/0.05% Tween-20 (TBS 5 137
`mM sodium chloride, 25 mM Tris (pH 8.0), 2.7 mM potassium chloride).
`The wells were then blocked with TBS/3% BSA, followed by addition of
`biotinylated anti-Cop 1 Abs (1:5000, 100 ml per well). Ab-ligand com-
`plexes were detected using streptavidin-conjugated alkaline phosphatase
`(1:3000, Bio-Rad) and p-nitrophenyl phosphate in triethanolamine buffer
`(Bio-Rad). The absorbance at 410 nm was monitored by a microplate
`reader (Dynatech MR4000).
`
`Table I. Amino acid composition of Cop 1 bound to and eluted from
`recombinant HLA-DR1, -DR1 and -DR4 molecules
`
`Cop 1a
`
`Untreated
`Eluted from DR1
`Eluted from DR2
`Eluted from DR4
`
`Amountb
`
`22.3
`21.5 (96)
`17.8 (80)
`21.7 (97)
`
`Y
`
`0.8
`0.8
`1.0
`0.8
`
`E
`
`1.7
`1.3
`2.0
`1.1
`
`A
`
`4.3
`4.3
`4.3
`4.3
`
`K
`
`3.2
`4.3c
`4.0c
`5.7c
`
`a Cop 1 (8150) either incubated with recombinant HLA-DR1, -DR2, and -DR4
`molecules and eluted by acid extraction, or untreated, was subjected to amino acid
`analysis as described in Materials and Methods. Values represent molar ratios of
`amino acids and were calculated with reference to A, whose yields are not reduced by
`the hydrolysis process.
`b Nanomole copolymer in the sample, determined by amino acid analysis after
`acid extraction. Percent binding is shown in parentheses.
`c Higher levels of K may be due to a contamination with isophenylthiourea. This
`interpretation is compatible with the observation that no such increase was observed
`in the more precise pool sequencing data in Fig. 2.
`
`Results
`Isolation of the bound fraction of Cop 1
`Quantitation and amino acid analysis of Cop 1 bound to HLA-
`DR1, -DR2, and -DR4 molecules. Cop 1 was incubated with wa-
`ter-soluble HLA-DR1, -DR2, or -DR4 molecules at the molar ratio
`of 1:1 for 40 h at 37°C. These recombinant “empty” HLA-DR
`molecules are usually stably assembled in the presence of exog-
`enously added peptide Ag. However, Cop 1 can substitute for pep-
`tides in promoting stabilization and with no interference from en-
`dogenous peptides (15). Unbound Cop 1 was separated from
`bound Cop 1 by Centricon ultrafiltration. Bound Cop 1 was then
`extracted from the HLA-DR by acid treatment (22) and subjected
`to amino acid analysis. At least 95% of added Cop 1 was bound to
`isolated HLA-DR1 and -DR4, and 80% was bound to HLA-DR2
`proteins (Table I). Cop 1 eluted from HLA-DR1, -DR2, and -DR4
`molecules showed ratios of Y:E:A:K similar to unseparated Cop 1
`(Table I). These results indicate that the bound fraction of Cop 1
`reflects the amino acid composition of the whole mixture and sug-
`gest that there is little or no preferential binding of Cop 1 compo-
`nents to different HLA-DR proteins. These data were supported by
`a different set of experiments in which Cop 1 was incubated with
`an excess of HLA-DR1, -DR2, and -DR4 molecules that had been
`purified from human homozygous EBV-transformed B cell lines,
`and then passed through a size-exclusion column. Nearly all of the
`Cop 1 was found in the fractions corresponding to the high m.w.
`complexes with each of the HLA-DR molecules (data not shown),
`with ,10% in each case at the lower m.w. position of Cop 1, also
`indicating that most of the copolymer binds to these molecules.
`HPLC separation. To further characterize the bound fraction of
`Cop 1 by means of hydrophobicity and size, samples were sepa-
`rated on RP-HPLC (Fig. 1) using an acetonitrile gradient, as de-
`scribed in Materials and Methods. In contrast to typical profiles of
`single peptides or peptide pools (22, 31–34), untreated Cop 1
`showed a very broad peak with several smaller peaks, which
`spread between ;40- and 75-min elution time (Fig. 1A). This elu-
`tion profile is characteristic of a mixture of random polypeptides
`and resembles HPLC separations of other batches of Cop 1 (un-
`published data). Similar profiles were obtained when Cop 1 was
`eluted from HLA-DR1 (Fig. 1B), -DR2 (Fig. 1C), or -DR4 (Fig.
`1D) molecules, suggesting that the bound fraction is similar to the
`whole Cop 1 mixture in its chemical properties.
`Pool sequencing. To analyze the sequence of Cop 1 bound to
`HLA-DR1, -DR2, and -DR4 molecules, HPLC fractions within the
`area described in the previous section, were pooled and sequenced.
`In all cases, the four amino acids of Cop 1 showed random pat-
`terns, with significantly higher levels of A over E, Y, and K (Fig.
`
`Page 3 of 9
`
`YEDA EXHIBIT NO. 2036
`MYLAN PHARM. v YEDA
`IPR2015-00644
`
`

`
`The Journal of Immunology
`
`4699
`
`Downloaded from
`
`http://www.jimmunol.org/
`
` by guest on November 16, 2015
`
`FIGURE 1. Separation of untreated Cop 1 (A), Cop 1 eluted from HLA-DR1 (B), -DR2 (C), and -DR4 (D) molecules on RP-HPLC. From 5% to 10%
`of the protein mixtures were fractionated by microbore HPLC using a Zorbax C18 1.0-mm RP column on a Hewlett-Packard 1090 HPLC with 1040 diode
`array detector. At a flow rate of 54 ml/min, Cop 1 was eluted with a gradient of 0.055% trifluoroacetic acid in acetonitrile (0% at 0 –10 min, 33% at 73
`min, and 60% at 105 min). Upper solid line, absorbance at 205 nm; lower solid lines, absorbance at 277 and 292 nm.
`
`2, A–D), which corresponds to the initially higher molar ratio of A
`in Cop 1 (1). There was no sequence specificity or preferential
`positioning of any amino acid of Cop 1, indicating that the bound
`fraction is also random, similar to the entire Cop 1. The yields of
`other amino acids resulting from endogenous peptides of HLA-DR
`molecules were minor (data not shown). It should be noted that
`these data represent sequencing of the first 20–25 amino acids
`from the N termini of bound Cop 1 polypeptides, which most
`likely protrude from the groove of HLA-DR molecules, and there-
`fore do not reflect the actual binding motif(s).
`Recognition of bound Cop 1 by anti-Cop 1 Abs. Anti-Cop 1 poly-
`clonal Abs were used to determine whether fractions of Cop 1
`eluted from different HLA-DR molecules contain epitopes similar
`to unseparated Cop 1. Binding assays were conducted as described
`in Materials and Methods. The results (Fig. 3) showed that all the
`fractions were similarly recognized by anti-Cop 1 Abs, suggesting
`that these fractions share similar epitopes with each other and with
`Cop 1.
`
`Characterization of binding motifs of Cop 1
`Treatment of Cop 1 bound to HLA-DR1, -DR2, or -DR4 mole-
`cules with aminopeptidase I. To determine the actual binding mo-
`tifs, Cop 1 was incubated with HLA-DR molecules at the molar
`ratio of 10 Cop 1:1 HLA-DR in PBS for 40 h at 37°C. Amino-
`peptidase I, a metalloprotein isolated from Streptomyces griseus
`(35), was added for the last 18 h of incubation, to remove N-
`terminal ends of Cop 1 polypeptides, protruding from the groove
`
`of the HLA-DR molecules, as well as to digest the unbound Cop
`1. The resulting Cop 1-HLA-DR complexes were analyzed by
`SDS-PAGE. As shown in Fig. 4, Cop 1-DR1 complexes were re-
`sistant to SDS-induced dissociation, forming higher m.w. com-
`plexes with HLA-DR1 ab heterodimers that were observed as nu-
`merous bands on the polyacrylamide gel above the molecular mass
`protein standard of 50 kDa (lane 5), and as described previously
`(15). In the presence of aminopeptidase I (a 33-kDa protein ap-
`pearing as a thin band below the molecular mass protein standard
`of 35 kDa, lanes 2, 4, and 6), all the unbound Cop 1 (a smear in
`the lower part of the gel, lanes 1 and 5) was completely digested
`(lanes 2 and 6), whereas Cop 1-DR1 complexes were protected
`(lane 6). It should be noted that upon incubation of aminopeptidase
`I with HLA-DR1 alone a complex was formed, represented by a
`band below the 50-kDa molecular mass protein standard (lane 4)
`probably caused by binding of some aminopeptidase I-derived di-
`gestion products to HLA-DR1. Similar treatment with aminopep-
`tidase I was applied to Cop 1 bound to HLA-DR2 and -DR4 mol-
`ecules. Bound Cop 1 remaining after aminopeptidase I treatment
`was eluted from HLA-DR by acid extraction, as described in Ma-
`terials and Methods.
`HPLC separation. After elution, Cop 1 digestion products were
`separated on RP-HPLC (Fig. 5) using an acetonitrile gradient, as
`described in Materials and Methods. In contrast to a very broad
`peak corresponding to untreated Cop 1 (Figs. 2A and 5A), proteo-
`lytic Cop 1 fragments eluted from HLA-DR1 (Fig. 5C), -DR2 (Fig.
`5E) or -DR4 (Fig. 5G) showed profiles similar to peptide pools
`
`Page 4 of 9
`
`YEDA EXHIBIT NO. 2036
`MYLAN PHARM. v YEDA
`IPR2015-00644
`
`

`
`4700
`
`BINDING MOTIFS OF COPOLYMER 1 TO HLA-DR1, -DR2, AND -DR4 MOLECULES
`
`FIGURE 2. Pool sequencing of untreated
`Cop 1 (A), and Cop 1 eluted from HLA-DR1
`(B), -DR2 (C), and -DR4 (D) molecules. HPLC
`fractions were pooled, concentrated, and sub-
`mitted to automated Edman degradation on a
`Hewlett-Packard G1005A protein sequencer us-
`ing the manufacturer’s Routine 3.5.
`
`Downloaded from
`
`http://www.jimmunol.org/
`
` by guest on November 16, 2015
`
`and second cycles, of K at the second and third cycles, and of Y
`(presumably at P1 of the bound peptide) at the third to fifth cycle,
`for peptides bound to HLA-DR1 (Fig. 6C), -DR2 (Fig. 6E), or
`-DR4 (Fig. 6G). This spread of residues over two or three positions
`in the pool sequencing data is probably caused by the ragged N
`termini of the Cop 1 components after aminopeptidase treatment,
`similarly to other treated naturally processed class II MHC ligands
`
`isolated from purified human HLA-DR molecules (31–33). On the
`other hand, almost no peptides were eluted from “empty” HLA-
`DR1 (Fig. 5D), -DR2 (Fig. 5F), or -DR4 (Fig. 5H) molecules, or
`from a total Cop 1 digestion with no HLA-DR added (Fig. 5B).
`Peaks for further analysis were selected in the region where the
`untreated Cop 1 was eluted (Figs. 2A and 5A), between ;40 and
`75 min elution time. For each HLA-DR molecule, only peaks cor-
`responding to Cop 1 peptides, which did not overlap those eluted
`from that same HLA-DR molecule with no Cop 1, were pooled for
`sequencing.
`Pool sequencing. To analyze the sequence of the Cop 1 bound to
`HLA-DR1, -DR2, and -DR4 molecules, Cop 1 fractions were
`pooled and sequenced. In contrast to random patterns of the un-
`treated Cop 1, showing no sequence specificity or preferential po-
`sitioning of any of the four amino acids that comprise Cop 1 (Figs.
`3A and 6A), significantly higher levels of E were found at the first
`
`FIGURE 3. Detection of Cop 1 bound to HLA-DR1, -DR2, and -DR4
`molecules by anti-Cop 1 Abs. Bound fractions were diluted and plated in
`duplicates on a 96-well microtiter plates, followed by blocking with
`TBS/3% BSA and addition of biotinylated anti-Cop 1 polyclonal Abs. For
`other details see Materials and Methods. Background levels were ,10% of
`the binding.
`
`FIGURE 4. SDS-PAGE of soluble Cop 1-DR1 complexes treated with
`aminopeptidase I. Recombinant “empty” HLA-DR1 molecules (100 mM
`per sample) were incubated with unlabeled Cop 1 (8, 150) (1 mM) in PBS
`for 40 h at 37°C. Aminopeptidase I (2 units) was added for the last 18 h of
`incubation. Cop 1 alone (lane 1), or treated with aminopeptidase I (lane 2);
`HLA-DR1 alone (lane 3), or treated with aminopeptidase I (lane 4); and
`HLA-DR1-Cop 1 complexes alone (lane 5), or treated with aminopeptidase
`I (lane 6). Separation gel was 10% in acrylamide and stacking gel was 5%.
`HLA-DR1-Cop 1 complexes were run under nonreducing conditions for
`1 h at 200 V and stained with Coomassie Brilliant blue.
`
`Page 5 of 9
`
`YEDA EXHIBIT NO. 2036
`MYLAN PHARM. v YEDA
`IPR2015-00644
`
`

`
`Downloaded from
`
`http://www.jimmunol.org/
`
` by guest on November 16, 2015
`
`The Journal of Immunology
`
`4701
`
`FIGURE 5. Separation of untreated Cop 1
`(A); Cop 1 digested with aminopeptidase I
`(B); Cop 1 digested and eluted from recom-
`binant HLA-DR1 (C), -DR2 (E), and -DR4
`(G); or endogenous peptides digested and
`eluted from HLA-DR1 (D), -DR2 (F), and
`-DR4 (H) molecules on RP-HPLC. Other de-
`tails are as in the legend to Fig. 1.
`
`(25). Position 3 was assumed to correspond to P1 because in the
`structure of the HA 306–318/HLA-DR1 complex P-2 is at the flush
`end of the groove and P1 is amino acid 3, i.e., Y308, in a deep
`pocket (17). For HLA-DR2, both Y and A levels were increased at
`cycle 3 (Fig. 6E). However, HLA-DR2 (DRB1*1501) has a P1
`pocket lined by b86 Val, which could not accommodate Y but can
`accommodate A (18, 32, 33). Although Y can be accommodated at
`the P4 pocket, no enrichment of Y at this position (cycles 6, 7) was
`observed (Fig. 6E). No sequence specificity or preferential posi-
`tioning was observed for anchor positions following P1 (P4, P6, or
`P9 of HLA-DR1 or -DR4; P4, P7, of DR2b molecules) (Fig. 6, C,
`E, and G). In all the samples the levels of A were higher than E,
`Y, and K, which corresponds to the initially higher molar ratio of
`A in Cop 1 (1). The yields of Y, E, A, and K (Fig. 6, D, F, and H),
`as well as of other amino acids (data not shown), resulting from
`endogenous peptides of HLA-DR molecules or from a complete
`Cop 1 digestion by aminopeptidase I (Fig. 6B), were minor.
`
`Discussion
`Previous studies of Cop 1 activity in EAE and MS focused on the
`effects of the whole mixture on immune cells (1–14). Different
`preparations of Cop 1 containing random polypeptides varying in
`size exhibited consistent characteristics as reflected in suppression
`
`of EAE (1–5), T cell recognition (6 – 8, 10), binding to class II
`MHC molecules (9 –12, 15), or clinical efficacy (13, 14, 36). Thus,
`in view of the accumulated information on antigenic peptide bind-
`ing, especially in autoimmune diseases (32–34), characterization
`of Cop 1 components has been of particular interest. The present
`study addressed the question whether different determinants of
`Cop 1 might be selected for various MHC interactions, or whether
`the same type of determinant(s) are universally involved. Recom-
`binant “empty” HLA-DR molecules were employed to isolate and
`characterize the bound fraction of Cop 1 with no interference from
`endogenous peptides. These HLA-DR molecules have been used
`previously for binding, alignment, and crystallographic studies of
`various exogenously added Ags (17, 18, 28, 29, 31, 37, 38). Sev-
`eral parameters indicated that similar Cop 1 fractions bound to the
`different HLA-DR proteins. 1) Amino acid analyses of Cop 1
`eluted from HLA-DR1, -DR2, and -DR4 molecules revealed sim-
`ilar amino acid ratios of Y:E:A:K in all fractions and intact Cop 1.
`Moreover, the sequencing patterns of the N-termini of the intact
`bound polypeptides showed a random distribution. 2) Chromato-
`graphic profiles on RP-HPLC reflected complexity and similarity
`of the bound peptide pools. To access the m.w. distribution, Cop 1
`was also analyzed by matrix-assisted laser-desorption mass spec-
`trometry. However, because of the random nature of Cop 1 or that
`
`Page 6 of 9
`
`YEDA EXHIBIT NO. 2036
`MYLAN PHARM. v YEDA
`IPR2015-00644
`
`

`
`Downloaded from
`
`http://www.jimmunol.org/
`
` by guest on November 16, 2015
`
`4702
`
`BINDING MOTIFS OF COPOLYMER 1 TO HLA-DR1, -DR2, AND -DR4 MOLECULES
`
`FIGURE 6. Pool sequencing of un-
`treated Cop 1 (A); Cop 1 digested with
`aminopeptidase I (B); Cop 1 digested,
`and eluted from recombinant HLA-
`DR1 (C), -DR2 (E), and -DR4 (G); or
`endogenous peptides digested and
`eluted from HLA-DR1 (D), -DR2 (F),
`and -DR4 (H) molecules. For details
`see legend to Fig. 2.
`
`material eluted from HLA-DR1, -DR2, or -DR4 molecules, at-
`tempts to determine individual mass values were unsuccessful. No
`major peaks were evident; rather, a wide range of masses was seen
`with the highest point of the curve corresponding to the average
`m.w. of Cop 1 (data not shown). 3) Epitopes on the bound fractions
`and intact Cop 1, as recognized by anti-Cop 1 polyclonal Abs,
`were also similar. These findings together suggest that despite its
`random and heterogenous nature, most (all) components of the
`Cop 1 mixture are fully potent in terms of binding capacity for all
`three allotypes examined. The average molecular mass of the Cop
`1 employed was 8500 Da (i.e., 75– 80 amino acids in length). Ty-
`rosine, a major anchor residue for many DR allotypes, is present at
`0.8 of every 10 residues (Table I). In the case of HLA-DR1 and
`-DR4, which have a P1 pocket that is lined by residue b86Gly, this
`tyrosine residue would serve as a primary anchor (17, 18). In the
`case of DR2, which has b86Val, the tyrosne residue is presumably
`too large for the P1 pocket, but can be accommodated in the P4
`
`pocket; in MBP 85–99, Phe92 is accommodated in the P4 pocket
`(33). The high fraction of Cop 1 that binds suggests that nearly
`every polypeptide within this mixture with an average size of
`75– 80 amino acids includes a 13-amino acid stretch that is ac-
`commodated within the groove.
`The N-terminal sequence analysis of the first 20 amino acid
`residues of bound Cop 1 most likely represents material that pro-
`trude from the grooves of the HLA-DR molecules. Determination
`of the actual peptide sequence(s) from Cop 1 found within the
`binding grooves of HLA-DR proteins, has been an important goal.
`To directly access these amino acids, digestion of the protruding
`ends of Cop 1 polypeptides with an N-terminal peptidase was em-
`ployed. This approach was proven to be useful in trimming of
`N-terminal ends of peptides that protrude out of class II MHC
`proteins, while protecting epitopes bound to the groove from pro-
`teolysis (26, 27).
`
`Page 7 of 9
`
`YEDA EXHIBIT NO. 2036
`MYLAN PHARM. v YEDA
`IPR2015-00644
`
`

`
`Downloaded from
`
`http://www.jimmunol.org/
`
` by guest on November 16, 2015
`
`The Journal of Immunology
`
`4703
`
`Table II. Binding motifs of Cop 1 for HLA-DR1, -DR2 and -DR4
`molecules
`
`Amino Acid Residues (relative positions)a
`
`HLA-DR
`
`DRB1*0101
`DRB1*0401
`DRB1*1501
`
`-2
`
`E
`E
`E
`
`-1
`
`K
`K
`K
`
`1
`
`Y
`Y
`Y, A
`
`4
`
`A
`A
`A
`
`6
`
`A
`A
`A
`
`7
`
`A
`A
`A
`
`9
`
`A
`A
`A
`
`a This interpretation is derived from the pooled sequencing data of Fig. 6.
`
`Detailed peptide motifs for class II MHC molecules were elu-
`cidated by pool sequencing of natural MHC-associated ligands
`(25). A similar approach was employed here for Cop 1, a mixture
`of random polypeptides of various lengths. Because of the nature
`of its synthesis, isolation of individual components has been tech-
`nically impossible (1). Regardless of the HLA-DR molecule, Y
`was found at the first anchor position (the third residue in the
`sequence analysis), followed by A in the subsequent pockets (Fig.
`6). These data are in line with the peptide-binding motifs of HLA-
`DR1 (DRB1*0101) (17, 21, 22, 39, 40) or -DR4 (DRB1*0401)
`(18, 22, 41– 43) molecules. For HLA-DR2 (DRB1*1501), how-
`ever, no aromatic residue was found in the first P1 anchor, whereas
`the second (P4) could have Y (19, 32, 33). Anomalously, however,
`in the Cop 1 bound to HLA-DR2 Y as well as A was also appar-
`ently enriched at P1, but not at P4. The efficacy of Cop 1 might be
`improved by substituting V for Y in the copolymer, since V89 is
`found in MBP 85-99 in the P1 pocket. Substitution of F for Y
`might be even better, since it would fit tightly into the P1 pocket
`as well as fitting into the P4 pocket. The increase in E at P-2 and
`K at P-1 is not fully understood. The enrichment of K near the N
`termini of naturally processed peptides bound to HLA-DR1 was
`observed previously and mistakenly interpreted as an anchor res-
`idue (21). These residues may contribute to the stable interactions
`of Cop 1 with the HLA-DR molecules and the TCR, similarly to
`residue K at P-1 of HA 306 –318 peptide complexed with HLA-
`DR1 (17), or Y at P-1 and possibly Q at P-2 of CII 1168 –1180
`complexed with HLA-DR4 (18). Also, removal of P at P-2 and K
`at P-1 from HA 306 –318 peptide resulted in a lower binding to
`some HLA-DR1 and -DR4 alleles (44). Peptide flanking residues
`that lie outside the MHC anchor positions at the C-terminal end
`were shown previously to influence immunogenicity (45); no com-
`parable information is available at the N-terminal end. However,
`P-2 and P-1 in the crystal structures of the HA 306 –318/HLA-DR1
`and CII 1168 –1180/HLA-DR4 complexes point upwards and most
`likely contact the TCR (17, 18). Moreover, peptide analogues of
`HA 306 –318 with amino acid substitutions at position 307K (P-1)
`anergized HA 306 –318-specific DR1-restricted human T cell
`clones (46). Binding motifs of Cop 1 are summarized in Table II.
`These results all suggest that Cop 1 contains promiscuous class
`II MHC binding motifs. Once bound to the groove of HLA-DR
`molecules, Cop 1 may act as either a blocking peptide or as an
`antagonist or partial agonist, resulting in suppression of autoim-
`mune T cell responses (e.g., by induction of T suppressor cells) or
`anergy, or both. Further studies will indicate whether any of these
`sequences are potentially useful for mapping the T cell epitopes,
`and possibly in the treatment of MS and RA in humans.
`
`Acknowledgments
`We thank Drs. Olaf Rotzschke and Kirsten Falk for fruitful discussions and
`Anastasia Haykov and Michal Mandelboim for expert technical assistance.
`
`References
`1. Teitelbaum, D., A. Meshorer, T. Hirshfeld, R. Arnon, and M. Sela. 1971. Sup-
`pression of experimental allergic encephalomyelitis by a synthetic polypeptide.
`Eur. J. Immunol. 1:242.
`2. Teitelbaum, D., C. Webb, A. Meshorer, R. Arnon, and M. Sela. 1973. Suppres-
`sion by severa