Structure and Molecular Interactions of a Unique Antitumor Antibody Specific for N-Glycolyl GM3

Abstract

N-Glycolyl GM3 ganglioside is an attractive target antigen for cancer immunotherapy, because this epitope is a molecular marker of certain tumor cells and not expressed in normal human tissues. The murine monoclonal antibody 14F7 specifically recognizes N-glycolyl GM3 and shows no cross-reactivity with the abundant N-acetyl GM3 ganglioside, a close structural homologue of N-glycolyl GM3. Here, we report the crystal structure of the 14F7 Fab fragment at 2.5 Å resolution and its molecular model with the saccharide moiety of N-glycolyl GM3, NeuGcα3Galβ4Glcβ. Fab 14F7 contains a very long CDR H3 loop, which divides the antigen-binding site of this antibody into two subsites. In the docking model, the saccharide ligand is bound to one of these subsites, formed solely by heavy chain residues. The discriminative feature of N-glycolyl GM3 versus N-acetyl GM3, its hydroxymethyl group, is positioned in a hydrophilic cavity, forming hydrogen bonds with the carboxyl group of Asp H52, the indole NH of Trp H33 and the hydroxyl group of Tyr H50. For the hydrophobic methyl group of N-acetyl GM3, this environment would not be favorable, explaining why the antibody specifically recognizes N-glycolyl GM3, but not N-acetyl GM3. Mutation of Asp H52 to hydrophobic residues of similar size completely abolished binding. Our model of the antibodycarbohydrate complex is consistent with binding data for several tested glycolipids as well as for a variety of 14F7 mutants with replaced VL domains. N-Glycolyl GM3 ganglioside is an attractive target antigen for cancer immunotherapy, because this epitope is a molecular marker of certain tumor cells and not expressed in normal human tissues. The murine monoclonal antibody 14F7 specifically recognizes N-glycolyl GM3 and shows no cross-reactivity with the abundant N-acetyl GM3 ganglioside, a close structural homologue of N-glycolyl GM3. Here, we report the crystal structure of the 14F7 Fab fragment at 2.5 Å resolution and its molecular model with the saccharide moiety of N-glycolyl GM3, NeuGcα3Galβ4Glcβ. Fab 14F7 contains a very long CDR H3 loop, which divides the antigen-binding site of this antibody into two subsites. In the docking model, the saccharide ligand is bound to one of these subsites, formed solely by heavy chain residues. The discriminative feature of N-glycolyl GM3 versus N-acetyl GM3, its hydroxymethyl group, is positioned in a hydrophilic cavity, forming hydrogen bonds with the carboxyl group of Asp H52, the indole NH of Trp H33 and the hydroxyl group of Tyr H50. For the hydrophobic methyl group of N-acetyl GM3, this environment would not be favorable, explaining why the antibody specifically recognizes N-glycolyl GM3, but not N-acetyl GM3. Mutation of Asp H52 to hydrophobic residues of similar size completely abolished binding. Our model of the antibodycarbohydrate complex is consistent with binding data for several tested glycolipids as well as for a variety of 14F7 mutants with replaced VL domains. The interest in antibodies that specifically recognize tumor markers has revived during the last few years, boosted by the approval, in 1997, of the first monoclonal antibody for cancer treatment by the United States Food and Drug Administration (FDA). Today, a total of six antibodies have been approved by the FDA for the therapy of several types of malignancies, and dozens of different antibodies with potential therapeutic effects in cancer are in clinical trials. One attractive target antigen for cancer immunotherapy is the N-glycolyl GM3 ganglioside (1Malykh Y.N. Schauer R. Shaw L. Biochimie (Paris). 2001; 83: 623-634Crossref PubMed Scopus (239) Google Scholar), a membrane-associated glycosphingolipid with a terminal N-glycolylated sialic acid residue (NeuGc). 1The abbreviations used are: NeuGc5′-N-glycolylneuraminidateFabantigen-binding fragments of immunoglobulinsCDRcomplementarity determining regionCLconstant light chainCHconstant heavy chainFvFab variable domainsGalgalactoseGlcglucosemAbmonoclonal antibodyNeuAc5′-N-acetylneuraminidatePDBProtein Data BankR.m.s.d.root mean square differencesr.s.c.c.real space correlation coefficientVLvariable light chainVHvariable heavy chainPEGpolyethylene glycolELISAenzyme-linked immunosorbent assayPBSphosphate-buffered saline.1The abbreviations used are: NeuGc5′-N-glycolylneuraminidateFabantigen-binding fragments of immunoglobulinsCDRcomplementarity determining regionCLconstant light chainCHconstant heavy chainFvFab variable domainsGalgalactoseGlcglucosemAbmonoclonal antibodyNeuAc5′-N-acetylneuraminidatePDBProtein Data BankR.m.s.d.root mean square differencesr.s.c.c.real space correlation coefficientVLvariable light chainVHvariable heavy chainPEGpolyethylene glycolELISAenzyme-linked immunosorbent assayPBSphosphate-buffered saline. This tumor marker is expressed in certain tumor cells, such as melanoma and breast tumors (2Müthing J. Stener H. Peter-Katalinic J. Marx U. Bethke V. Neuman V. Lehmann J. J. Biochem. 1994; 116: 64-73Crossref PubMed Scopus (38) Google Scholar, 3Marquina G. Waki H. Fernández L.E. Kon K. Carr A. Valiente O. Pérez R. Ando S. Cancer Res. 1996; 56: 5165-5171PubMed Google Scholar), but is otherwise absent from normal human tissues (4Varki A. Biochimie. (Paris). 2001; 83: 615-622Crossref PubMed Scopus (87) Google Scholar), which opens up the possibility of using antibodies specific toward these molecules both for diagnosis and immunotherapy. 5′-N-glycolylneuraminidate antigen-binding fragments of immunoglobulins complementarity determining region constant light chain constant heavy chain Fab variable domains galactose glucose monoclonal antibody 5′-N-acetylneuraminidate Protein Data Bank root mean square differences real space correlation coefficient variable light chain variable heavy chain polyethylene glycol enzyme-linked immunosorbent assay phosphate-buffered saline. 5′-N-glycolylneuraminidate antigen-binding fragments of immunoglobulins complementarity determining region constant light chain constant heavy chain Fab variable domains galactose glucose monoclonal antibody 5′-N-acetylneuraminidate Protein Data Bank root mean square differences real space correlation coefficient variable light chain variable heavy chain polyethylene glycol enzyme-linked immunosorbent assay phosphate-buffered saline. However, targeting N-glycolyl GM3 gangliosides is no easy task, as this glycosphingolipid shares common structural features with many other gangliosides expressed on the cell surface. In particular, N-glycolyl GM3 is highly similar to N-acetyl GM3, which is present in most human tissues. In fact, what should be targeted corresponds to an extremely subtle chemical modification: the difference between the N-glycolyl group in NeuGc and the N-acetyl group of NeuAc consists of the addition of a single oxygen atom to the N-acetyl moiety (CH2OH instead of CH3 in the context of a trisaccharide) (see Fig. 1). A murine monoclonal antibody (mAb), termed 14F7, that has this singular property has recently been described (5Carr A. Mullet A. Mazorra Z. Vázquez A.M. Alfonso M. Mesa C. Rengifo E. Pérez R. Fernández L.E. Hybridoma. 2000; 19: 241-247Crossref PubMed Scopus (100) Google Scholar, 6Carr A. Mesa C. del Carmen Arango M. Vázquez A.M. Fernández L.E. Hybrid. Hybridomics. 2002; 21: 463-468Crossref PubMed Scopus (36) Google Scholar). It exhibits high specificity toward N-glycolyl GM3 and strongly recognizes human melanoma and breast cancer tissues. 14F7 is quite unique, not only because of its binding specificity and rather strong binding affinity (in the nanomolar range) 2G. Rojas, A. Talavera, Y. Munoz, E. Rengifo, U. Krengel, J. Ångström, J. Gavilondo, and E. Moreno, submitted manuscript.2G. Rojas, A. Talavera, Y. Munoz, E. Rengifo, U. Krengel, J. Ångström, J. Gavilondo, and E. Moreno, submitted manuscript. for N-glycolyl GM3, but also because of several other properties. In particular, 14F7 is an IgG antibody, belonging to the IgG1 subclass, whereas most other anti-ganglioside antibodies are IgMs, as a result of only a primary immune response to carbohydrate antigens (7Portoukalian J. Clin. Rev. Allergy Immunol. 2000; 19: 73-78Crossref PubMed Scopus (8) Google Scholar). Furthermore, in experiments carried out in mice, the 14F7 monoclonal antibody showed a remarkable inhibition of the growth of solid tumors (6Carr A. Mesa C. del Carmen Arango M. Vázquez A.M. Fernández L.E. Hybrid. Hybridomics. 2002; 21: 463-468Crossref PubMed Scopus (36) Google Scholar). Recently, a pilot clinical trial with radiolabeled 14F7 has been completed, with very encouraging results showing that this antibody is able to specifically target breast tumors and their metastases in vivo. 3A. Casaco, personal communication.3A. Casaco, personal communication. To date, only very few structures of carbohydrate-binding antibodies have been reported (8Cygler M. Rose D.R. Bundle D.R. Science. 1991; 253: 442-445Crossref PubMed Scopus (247) Google Scholar, 9Cygler M. Wu S. Zdanov A. Bundle D.R. Rose D.R. Biochem. Soc. Trans. 1993; 21: 437-441Crossref PubMed Scopus (32) Google Scholar, 10Rose D.R. Przybylska M. To R.J. Kayden C.S. Oomen R.P. Vorberg E. Young N.M. Bundle D.R. Prot. Sci. 1993; 2: 1106-1113Crossref PubMed Scopus (56) Google Scholar, 11Bundle D.R. Baumann H. Brisson J.-R. Gagné S.M. Zdanov A. Cygler M. Biochemistry. 1994; 33: 5183-5192Crossref PubMed Scopus (103) Google Scholar, 12Zdanov A. Li Y. Bundle D.R. Deng S.-J. MacKenzie C.R. Narang S.A. Young N.M. Cygler M. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 6423-6427Crossref PubMed Scopus (146) Google Scholar, 13Jeffrey P.D. Bajorath J. Chang C.Y.Y. Yelton D. Hellström I. Hellström K.E. Sheriff S. Nat. Struct. Biol. 1995; 2: 466-471Crossref PubMed Scopus (92) Google Scholar, 14Pichla S.L. Murali R. Burnett R.M. J. Struct. Biol. 1997; 119: 6-16Crossref PubMed Scopus (37) Google Scholar, 15Villeneuve S. Souchon H. Riottot M.-M. Mazié J.-C. Lei P.-S. Glaudemans C.P.J. Ková[caron]c P. Fournier J.-M. Alzari P.M. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 8433-8438Crossref PubMed Scopus (105) Google Scholar, 16Vyas N.K. Vyas M.N. Chervenak M.C. Johnson M.A. Pinto B.M. Bundle D.R. Quiocho F.A. Biochemistry. 2002; 41: 13575-13586Crossref PubMed Scopus (88) Google Scholar, 17Calarese D.A. Scanlan C.N. Zwick M.B. Deechongkit S. Mimura Y. Kunert R. Zhu P. Wormald M.R. Stanfield R.L. Roux K.H. Kelly J.W. Rudd P.M. Dwek R.A. Katinger H. Burton D.R. Wilson I.A. Science. 2003; 300: 2065-2071Crossref PubMed Scopus (673) Google Scholar). This is remarkable in the view of the more than 250 entries containing Fab structures that have been deposited in the Protein Data Bank (18Berman H.M. Westbrook J. Feng Z. Gilliland G. Bhat T.N. Weissig H. Shindyalov I.N. Bourne P.E. Nucleic Acids Res. 2000; 28: 235-242Crossref PubMed Scopus (27319) Google Scholar). One of the possible reasons for this unbalance could be that carbohydrates in general are not very immunogenic, producing mostly low affinity antibody responses of the IgM isotype, as mentioned above. 14F7 is thus an exception to the rule, thanks to a successful immunization procedure (5Carr A. Mullet A. Mazorra Z. Vázquez A.M. Alfonso M. Mesa C. Rengifo E. Pérez R. Fernández L.E. Hybridoma. 2000; 19: 241-247Crossref PubMed Scopus (100) Google Scholar). Here, we report the crystal structure of the unliganded Fab fragment of 14F7 at 2.5 Å resolution. This structure provided the starting point for computer docking studies with the saccharide moiety of N- glycolyl GM3, NeuGcα3Galβ4Glcβ. The molecular model a for of the binding site to tumor Fab 14F7 from using and the of the heavy chain on a A in with at for with and on a to the and the the The Fab For the Fab on a and the by on during with a by and a of at a of The and to using a experiments from a of Fab 14F7 but the could be to a few provided that the by using the at The rather with the of from to at to with the addition of as The crystal from to and from many tested for and of The space group with cell a to a of 2.5 J. Biol. 33: PubMed Scopus Google with one used for data of in formed by a of the in and the containing and at Data at at the in on a of the in a in strongly we to data at different on the a long To and a crystal used at data to 2.5 Å resolution and with J. 21: Scopus Google Scholar, J. 21: Scopus Google Scholar, J. 1993; Scopus Google Scholar), in for the data are in and in to the resolution for the data Wilson of of of Protein chain chain molecules from in bonds from in for chain chain from R.A. R. A. 1991; Scopus Google in a The structure by with the J. A. 1994; Scopus Google Scholar), using the Fab as the The and the model to with domains chain constant domains and heavy chain constant by P.D. R.J. Proc. Natl. Sci. U. S. A. 1997; PubMed Scopus Google Scholar). of used for the PubMed Scopus Google Scholar). and with P.D. P. J. M. R.J. D. PubMed Scopus Google with molecular M. A. 1991; PubMed Scopus Google in several molecules at of and than and and at the molecules are in the results are in I. The structure from In a C. J. PubMed Scopus Google by R.A. J. 1993; Google Scholar), of the residues are in the most of the and are in the residues are in one is in a a is positioned in a in region and is in a similar of the in many reported Fab The real space correlation coefficient for this residue a is The for the constant domains have not been but from the data G. Wu Nucleic Acids Res. 2001; PubMed Scopus Google The for first as the that be a of the data that Trp is by a residue at The real space correlation for these two residues are and Here, we the for the variable which for this antibody in the heavy chain to and to of Protein and crystal structure of the 14F7 Fab fragment as the starting for and docking The and constant domains and molecules thus only the Fab variable domains for and using the States S. M. J. Scopus Google Scholar). and hydrogen to the an carried out in in to the of to during of a a similar with the to the to of A constant and a Å used for The terminal moiety of N-glycolyl GM3 used in the docking procedure using the on in the for the The carbohydrate docking of the not in are as described in A. M.-M. Bourne Y. C. Pérez S. J. 1991; PubMed Scopus Google Scholar, G. Schauer R. J. Biochemistry. PubMed Scopus Google Scholar, I. Biochemistry. 1993; PubMed Scopus Google in the using to the space of the binding using the structure as starting of the in the complementarity determining and several other residues to It should be that not to the but also the The structure from to during and at this during a with a The as in used to bonds to The as for that at the in a using an of of the a of the I. Science. PubMed Scopus Google Scholar, D.A. J. 1996; PubMed Scopus Google Moreno, to be This the of a of for binding of a single by using a and of the binding site E. K. 2002; PubMed Scopus Google and the used for for The a of from the and into using a of as only for of the the one from for and studies that Asp H52 a in the binding of In to the of this residue for the molecular of the saccharide two targeting the 14F7 residue Asp H52 for One used to Asp H52 to the other one for to The in the single chain antibody fragment in the from 14F7, which the region and exhibits binding as the 14F7 monoclonal The for using a single chain with a different VL is that we could not an fragment with the of 14F7, to in from used as for chain The first using and a single that to the of the the antibody at at and at The thus in the of using and used to a of the with a single that to the of the the antibody used as antibody fragment containing the with and and into the and in in containing and of from using with an using both to the and the antibody of antibody fragments from the with on a and from the with polyethylene glycol and 2.5 J. G. J. Biol. 1991; PubMed Scopus Google Scholar). of antibody fragments by enzyme-linked immunosorbent assay with N-glycolyl GM3 in with and with in during at antibody fragments in and on the for at the with an to in to the The and the and hydrogen in The with The with a tested for the binding of N-glycolyl GM3. The as a for N-glycolyl GM3 and using from and Fig. using from of Fab crystal structure of the 14F7 Fab fragment at 2.5 Å resolution and to of and with The is very well for a 2.5 Å structure for is to the data a few in the structure are well In particular, the residues and from the light chain and the heavy could not be in the For the heavy chain is but not for this to be Fab structures deposited in the data this to for their and as and by a the crystal structure of Fab 14F7 is highly similar to other Fab with one The complementarity determining region CDR H3 of 14F7 with at is the both in the data and in the with acid residues (see This long from the of the the binding site into two that we to as the VL and of the CDR H3 of 14F7 to other carbohydrate-binding antibodies present in the Protein Data Bank The is The of the CDR H3 are between H3 in a In the crystal structure of 14F7, the long CDR H3 is well by and has rather low for this region is with for the that this of the structure is quite However, a of the from The and two of which in with the from and and one of forming a hydrogen to a residue in the constant region of the and the hydrophobic of the of these are with molecules in the crystal by that are not long by with other residues are in general very It is thus that the of the CDR H3 is an by and that this loop, which is only by to molecules in the is more at its the molecular we that the of CDR H3 is quite and that no of its are to be binding of the In this is the of at the of CDR H3 and The of these are with other in the structure and very with the with to the and of CDR both the VL and are rather the of CDR H3 is toward the A is by the of this loop, by the of and with several from CDR and and one single residue from the light chain of the 14F7 to in complex with its carbohydrate we to a model of the complex of the antibody with its N-glycolyl GM3. The of the binding site strongly that the binding site for N-glycolyl GM3 is in the in the described above. from the crystal structure of the Fab fragment of 14F7 and using molecular we an of which to different binding site docking with the terminal of N-glycolyl GM3, in a model of the complex with the crystal structure of unliganded 14F7, no in the computer from docking the of the of the residues. The in CDR H3 with root mean square differences of The residue that is most by the docking is Tyr are for Tyr and Trp (in of differences with the crystal Trp H33 is Asp H52 exhibits a of its The for the ligand is of the total of This well with the for similar molecules reported in the such as as no NeuGc is present in the which from to of structures are In the model, the hydroxyl group of the N-glycolyl moiety of the sialic acid is in a highly hydrophilic by the carboxyl group of Asp H52 and the indole NH of Trp the of the N-glycolyl hydroxymethyl group the of Tyr forming a hydrophobic (see Fig. and Tyr which is positioned more also a hydrogen to the N-glycolyl The oxygen of the N-glycolyl moiety is to hydrogen to This residue a hydrogen to of the hydrogen between and the 14F7 Fab chain H33 indole H52 chain Å in a the group of the sialic acid residue and the chain NH of as well as hydroxyl of the as in Fig. and The residue also in hydrogen the oxygen to and its oxygen belonging to to The are close to for the G. Schauer R. J. Biochemistry. PubMed Scopus Google and in with the for the antibody, which also E. Vázquez A.M. I. Fernández L.E. J. Pérez R. PubMed Scopus (32) Google Scholar), with a specificity are for the glucose moiety to the using In the crystal a bound is positioned between and the Tyr hydroxyl group, forming hydrogen bonds to both residues. on the complex with this into between and the carboxyl of the sialic acid of the and Tyr would be to the in the model of the However, is possible that the bound to the is bound to 14F7, more that this is not the for the The sialic acid is and would be more strongly by a rather than by an with a between and the carboxyl group of as by docking model, is the most molecular for we would to point out that in the model, ligand solely with heavy chain a single residue of VL is in with the carbohydrate Trp which is to the binding is at a of Å from the atom of the sialic acid The of to the 14F7 monoclonal antibody in between the N-glycolyl group of NeuGc and the N-acetyl group of NeuAc present in most other gangliosides a difference to the addition of one oxygen atom to the N-acetyl moiety in the context of a Our results that is the hydrophilic environment the hydroxymethyl group of NeuGc that the with Asp H52, Trp and Tyr for of the the computer docking model be as of the structure of the complex in the as would be possible for a crystal several results the features of the structural we that in the 14F7 model, ligand solely with heavy chain In this is to the results from a for other In this the 14F7 heavy chain variable region with a of light chain variable from both and and than one producing antibody fragments able to specifically recognize the antigen and of for The VL complementarity determining of these both with other and also with the CDR in Fig. these the the binding by 14F7, that the VL region is not for of to computer model, residues in binding the N-glycolyl group of GM3 are Asp H52, Trp and Tyr H50. Asp H52 as a for we two point mutants of this by should be mentioned that the two an at to rather than single The are to and thus the mutants and residue is not positioned in the binding to docking and with its chain into the we that this antigen binding both mutants out to no binding to of Data for on the of gangliosides sialic a have recently been tested for binding to the 14F7 antibody, in a strong solely of (5Carr A. Mullet A. Mazorra Z. Vázquez A.M. Alfonso M. Mesa C. Rengifo E. Pérez R. Fernández L.E. Hybridoma. 2000; 19: 241-247Crossref PubMed Scopus (100) Google Scholar). the for an N-glycolyl group out binding of ganglioside an N-acetyl group, the sialic acid is bound at a such as tested which the binding of the antibody E. Vázquez A.M. I. Fernández L.E. J. Pérez R. PubMed Scopus (32) Google Scholar). The binding of to 14F7 be from the complex by between the terminal residue and and However, more the of a terminal NeuGc residue such as and In the of the of an to results in an of the sialic acid the of these two Furthermore, different are most at that an more binding epitope as a to the of an on the sialic be that the terminal NeuGc of would have to be binding to 14F7 to However, the structures for of the possible low of the E. Vázquez A.M. I. Fernández L.E. J. Pérez R. PubMed Scopus (32) Google that the residues in with the in different thus for the of to of 14F7 with most the antigen binding site is positioned in a at the between the light and heavy chain variable and residues from both VL and to ligand binding. This is in to to are only with heavy chain acid residues. For Fab 14F7, is no at the the long CDR H3 from this and divides the potential binding site into two subsites. It as no that a of 14F7 with other Fab which in complex with their carbohydrate antigens (see that the saccharide in are in a in a with the long CDR H3 of A exception is the structure of the Fab fragment D.A. Scanlan C.N. Zwick M.B. Deechongkit S. Mimura Y. Kunert R. Zhu P. Wormald M.R. Stanfield R.L. Roux K.H. Kelly J.W. Rudd P.M. Dwek R.A. Katinger H. Burton D.R. Wilson I.A. Science. 2003; 300: 2065-2071Crossref PubMed Scopus (673) Google Scholar), a antibody a with this Fab fragment that its bound antigen at a site in that corresponds to the of However, should also be that the antibody has an to the of the which to the two antibodies in In the present we have the crystal structure of Fab 14F7, a antibody specific for N-glycolyl GM3. the structure as a starting studies with the saccharide moiety of the NeuGcα3Galβ4Glcβ. The molecular model of the carbohydrate complex the binding of 14F7 and is consistent with glycosphingolipid binding data for a of 14F7 The model the for of the antibody to its