Portal de Periódicos da CAPES

The use of antibody fragments for crystallization and structure determinations

1995; Elsevier BV; Volume: 3; Issue: 12 Linguagem: Inglês

10.1016/s0969-2126(01)00266-0

1878-4186

Ladislau C. Kovari, Cory Momany, Michael G. Rossmann,

Monoclonal and Polyclonal Antibodies Research

Structural studies of some of the most interesting biological macromolecules are frequently aborted because of an inability to form suitably diffracting crystals. Although advances have been made in crystallizing soluble and membrane proteins [1Blundell T.L. Johnson L.N. Protein Crystallography. Academic Press, London1976Google Scholar, 2McPherson A. Preparation and Analysis of Protein Crystals. John Wiley Co, New York1982Google Scholar, 3Ducruix A. Giege R. Crystallization of Nucleic Acids and Proteins: A Practical Approach. Oxford University Press, New York1992Google Scholar, 4Michel H. Crystallization of Membrane Proteins. CRC Press, Orlando1990Google Scholar, 5McPherson A. Malkin A.J. Kuznetsov Y.G. The science of macromolecular crystallization.Structure. 1995; 3: 759-768Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar, 6McPherson A. Crystallization of proteins from polyethylene glycol.J. Biol. Chem. 1976; 251: 6300-6303Abstract Full Text PDF PubMed Google Scholar], crystallization is frequently inhibited by heterogeneity, insolubility, molecular flexibility or a polydisperse character in solution. Viral capsid subunits and membrane proteins are, for example, members of this intransigent group. Current crystallization tools include commercial crystallization kits [7Cudney R. Patel S. Weisgraber K. Newhouse Y. McPherson A. Screening and optimization strategies for macromolecular crystal growth.Acta Cryst. D. 1994; 50: 414-423Crossref PubMed Google Scholar, 8McPherson A. Two approaches to the rapid screening of crystallization conditions.J. Cryst. Growth. 1992; 122: 161-167Crossref Scopus (53) Google Scholar, 9Stura E.A. Satterthwait A.C. Calvo J.C. Kaslow D.C. Wilson I.A. Reverse screening.Acta Cryst. D. 1994; 50: 448-455Crossref PubMed Google Scholar, 10D'Arcy A. Crystallizing proteins — a rational approach?.Acta Cryst. D. 1994; 50: 469-471Crossref PubMed Google Scholar, 11Shaw Stewart P.D. Khimasia M. Predispensed gradient matrices − a new rapid method of finding crystallization conditions.Acta Cryst. D. 1994; 50: 441-442Crossref PubMed Google Scholar], sparse matrix sampling [[12]Jancarik J. Kim S.-H. Sparse matrix sampling: a screening method for crystallization of proteins.J. Appl. Cryst. 1991; 24: 409-411Crossref Scopus (2062) Google Scholar], crystallization robots [[13]Ward K.B. Perozzo M.A. Zuk W.M. Automating crystallization experiments.in: Ducruix A. & Giege R Crystallization of Nucleic Acids and Proteins: A Practical Approach. Oxford University Press, New York1992: 291-309Google Scholar], surveys of crystallization data bases [[14]Gilliland G.L. Tung M. Blakeslee D.M. Ladner J.E. Biological macromolecule crystallization database, version 3.0: new features, data and the NASA archive for protein crystal growth data.Acta Cryst. D. 1994; 50: 408-413Crossref PubMed Google Scholar] and seeding techniques [[15]Stura E.A. Wilson I. Seeding techniques.in: Ducruix A. & Giege R Crystallization of Nucleic Acids and Proteins: A Practical Approach. Oxford University Press, New York1992: 99-125Google Scholar]. A useful addition to this growing repertoire is co-crystallization with antibody fragments such as Fabs or Fvs. Fab fragments have reasonable solubility and bind specifically to selected antigens with equilibrium constants in the range of 105 to 108 M−1 [[16]Roitt I. Brostoff J. Male D. Immunology. 3rd. Mosby Press, London1993Google Scholar]. An Fab–antigen complex is therefore likely to have suitable properties for forming crystals (Table 1; Corrigendum). Thus, they can effectively transform aggregated material into a soluble, monodisperse sample suitable for crystallization. Laver [[17]Laver W.G. Crystallization of antibody–protein complexes in protein and nucleic acid crystallization.in: Carter C.W Series Methods: A Companion to Methods in Enzymology. Academic Press, San Diego1990: 70-74Google Scholar] observed that some proteins can be crystallized only when complexed with a cognate Fab fragment. This approach can sometimes be applied more readily than molecular engineering of point mutants or truncations to obtain a modified protein with improved crystallization properties. Furthermore, antibodies are frequently available from related biochemical studies.Table 1Antibody–protein antigen complexes studied crystallographically.ProteinAntibody fragmentReferencesLysozyme∗D1.3[29]Fischmann T.O. Poljak R.J. et al.Crystallographic refinement of the three-dimensional structure of the Fab D1.3–lysozyme complex at 2.5 å resolution.J. Biol. Chem. 1991; 266 (91302305): 12915-12920Abstract Full Text PDF PubMed Google ScholarLysozyme∗HyHEL-5[30]Sheriff S. Davies D.R. et al.Three-dimensional structure of an antibody–antigen complex.Proc. Natl. Acad. Sci. USA. 1987; 84 (88068538): 8075-8079Crossref PubMed Scopus (597) Google ScholarLysozyme∗HyHEL-10[31]Padlan E.A. Davies D.R. et al.Structure of an antibody–antigen complex: crystal structure of the HyHEL-10 Fab–Lysozyme complex.Proc. Natl. Acad. Sci. USA. 1989; 86 (89345580): 5938-5942Crossref PubMed Scopus (464) Google ScholarLysozyme∗F9.13.7[32]Lescar J. Souchon H. Alzari P. Crystal structures of pheasant and guinea fowl egg-white lysozymes.Protein Sci. 1994; 3 (94339839): 788-798Crossref PubMed Scopus (19) Google ScholarNeuraminidase∗NC41[33]Tulip W.R. Varghese J.N. Laver W.G. Webster R.G. Colman P.M. Refined crystal structure of the influenza virus N9 neuraminidase–NC41 Fab complex.J. Mol. Biol. 1992; 227 (92395656): 122-148Crossref PubMed Scopus (207) Google ScholarHuman rhinovirus 14∗F17-IA[34]Liu H. Baker T.S. et al.Structure determination of an Fab fragment that neutralizes human rhinovirus 14 and analysis of the Fab–virus complex.J. Mol. Biol. 1994; 240 (94300590): 127-137Crossref PubMed Scopus (48) Google ScholarHIV-1 reverse transcriptase†Fab28[21]Jacobo-Molina A. Arnold E. et al.Crystal structure of human immunodeficiency virus type 1 reverse transcriptase complexed with double-stranded DNA at 3.0 å resolution shows bent DNA.Proc. Natl. Acad. Sci. USA. 1993; 90 (93317673): 6320-6324Crossref PubMed Scopus (1100) Google ScholarHIV-1 capsid protein p24†Fab25.3[19]Nermut M.V. Thomas D. et al.Fullerene-like organization of HIV gag-protein shell in virus-like particles produced by recombinant Baculovirus.Virology. 1994; 198 (94082461): 288-296Crossref PubMed Scopus (116) Google ScholarCytochrome c oxidase†Fv7E2[23]Iwata S. Ostermeier C. Ludwig B. Michel H. Structure at 2.8 å resolution of cytochrome c oxidase from Paracoccus denitrificans.Nature. 1995; 376 (95379947): 660-669Crossref PubMed Scopus (1939) Google Scholar∗In these complexes the Fab was used to study antigen–antibody interactions. †In these complexes the Fab was used as a crystallization tool to aid in the structure determination of the protein antigen. Corrigendum Open table in a new tab 1Corrigendum: 15 January 1996Table 1 of this Ways & Means article should have included the antibody-protein complex between CD4 and FabW3/25 as an example in which a Fab was used to aid crystallization [[35]Davis S.J. Brady R.L. Barclay A.N. Harlos K. Dodson G.G. Williams A.F. Crystallization of a soluble form of the rat T-cell surface glycoprotein CD4 complexed with Fab from the W3/25 monoclonal antibody.J. Mol. Biol. 1990; 213 (90250775): 7-10Crossref PubMed Scopus (23) Google Scholar]. ∗In these complexes the Fab was used to study antigen–antibody interactions. †In these complexes the Fab was used as a crystallization tool to aid in the structure determination of the protein antigen. Corrigendum Table 1 of this Ways & Means article should have included the antibody-protein complex between CD4 and FabW3/25 as an example in which a Fab was used to aid crystallization [[35]Davis S.J. Brady R.L. Barclay A.N. Harlos K. Dodson G.G. Williams A.F. Crystallization of a soluble form of the rat T-cell surface glycoprotein CD4 complexed with Fab from the W3/25 monoclonal antibody.J. Mol. Biol. 1990; 213 (90250775): 7-10Crossref PubMed Scopus (23) Google Scholar]. An early use of an antibody specifically for crystallizing an otherwise insoluble protein was for the crystallization of HIV capsid protein p24 [[18]Prongay A.J. McClure J. et al.Preparation and crystallization of a human immunodeficiency virus p24-Fab complex.Proc. Natl. Acad. Sci. USA. 1990; 87 (91088635): 9980-9984Crossref PubMed Scopus (38) Google Scholar]. Dimers of recombinant human immunodeficiency virus (HIV) capsid protein p24 associate into large oligomers with a complex distribution of sizes [[19]Nermut M.V. Thomas D. et al.Fullerene-like organization of HIV gag-protein shell in virus-like particles produced by recombinant Baculovirus.Virology. 1994; 198 (94082461): 288-296Crossref PubMed Scopus (116) Google Scholar]. Although this oligomerization is appropriate for the protein as it forms the ∼1000 å long core of the HIV virion, this property interfered with attempts to crystallize the protein. Detergents, such as β-octyl-glucoside, decreased aggregation but alone were inadequate for obtaining crystals. Only after screening crystallization conditions with several different Fab fragments that recognize different p24 epitopes were crystals obtained that were suitable for X-ray diffraction analysis. A subsequent example of the use of a Fab fragment to crystallize a protein was that of HIV reverse transcriptase (RT) [20Jacobo-Molina A. Arnold E. et al.Crystals of a ternary complex of human immunodeficiency virus type 1 reverse transcriptase with a monoclonal antibody Fab fragment and double-stranded DNA diffract X-rays to 3.5 å resolution.Proc. Natl. Acad. Sci. USA. 1991; 88 (92073393): 10895-10899Crossref PubMed Scopus (66) Google Scholar, 21Jacobo-Molina A. Arnold E. et al.Crystal structure of human immunodeficiency virus type 1 reverse transcriptase complexed with double-stranded DNA at 3.0 å resolution shows bent DNA.Proc. Natl. Acad. Sci. USA. 1993; 90 (93317673): 6320-6324Crossref PubMed Scopus (1100) Google Scholar], which was crystallized as a ternary complex of RT–Fab–DNA. In this case, the Fab was thought to act as a ‘molecular clamp’ that immobilized a region of the enzyme. Although other structures of RT were obtained without the use of a Fab, its inclusion in RT–Fab–DNA complex resulted in a particularly rapid structure determination [[21]Jacobo-Molina A. Arnold E. et al.Crystal structure of human immunodeficiency virus type 1 reverse transcriptase complexed with double-stranded DNA at 3.0 å resolution shows bent DNA.Proc. Natl. Acad. Sci. USA. 1993; 90 (93317673): 6320-6324Crossref PubMed Scopus (1100) Google Scholar]. Use of Fv instead of Fab fragments is likely to have some advantages as there will be no flexible elbow to inhibit crystallization. The most recent example of an antibody–protein complex utilized an Fv fragment to crystallize cytochrome c oxidase, a membrane protein, from Paracoccus denitrificans [22Ostermeier C. Iwata S. Ludwig B. Michel H. Fv fragment-mediated crystallization of the membrane protein bacterial cytochrome c oxidase.Nat. Struct. Biol. 1995; 2: 842-846Crossref PubMed Scopus (158) Google Scholar, 23Iwata S. Ostermeier C. Ludwig B. Michel H. Structure at 2.8 å resolution of cytochrome c oxidase from Paracoccus denitrificans.Nature. 1995; 376 (95379947): 660-669Crossref PubMed Scopus (1939) Google Scholar]. The Fv fragment recognizes the periplasmic domain of the oxidase, but does not substitute for the membrane portion. Instead, the polar surface of the oxidase is increased by complex formation, thereby permitting formation of the major lattice packing contacts. Practical aspects of crystallizing antigen–antibody complexes involve selection of the antibody, preparation of a homogenous Fab species and the preparation of a complex with proper stoichiometry. The ideal antibody should not alter the native antigen conformation or interfere with its biological activity. For proteins that have a measurable enzymatic activity, conformation can be inferred by assaying the activity in the presence of the Fab. In the case of the HIV RT [[20]Jacobo-Molina A. Arnold E. et al.Crystals of a ternary complex of human immunodeficiency virus type 1 reverse transcriptase with a monoclonal antibody Fab fragment and double-stranded DNA diffract X-rays to 3.5 å resolution.Proc. Natl. Acad. Sci. USA. 1991; 88 (92073393): 10895-10899Crossref PubMed Scopus (66) Google Scholar], the monoclonal antibody was selected that had the highest avidity to the RT heterodimer. In contrast, the HIV p24 crystallization trials were set up with Fabs recognizing different epitopes. Bead-immobilized papain is particularly convenient for Fab preparation as the protease can be removed by centrifugation or filtration. The progress of digestion can be monitored by examining aliquots with reducing and non-reducing SDS-PAGE to find the conditions that maximize Fab production. Following the papain treatment, the Fc portion must be removed by, for instance, anion-exchange chromatography with a NaCl gradient. The Fab fragment used to crystallize HIV p24 could be further resolved into nine fractions (Figure 1) by ion-exchange chromatography using a pH gradient. Eight of these nine fractions were able to form a complex with HIV p24 in solution, shown by native, or isoelectric focusing (IEF), gel analysis, although only one of these complexes gave useful crystals. Unfortunately, the single Fab species that crystallized represented only about 15%, by weight, of the total Fab isolated from cell culture. Native gel electrophoresis can assay whether the purified Fab fragment has retained its antigen-binding capacity (Figure 2). IEF gels are also useful for determining the optimal stoichiometry, by examining mixtures with different ratios of Fab to antigen. If the isoelectric points of the Fab and antigen are different, then it should be possible to resolve the Fab, the complex and the antigen. The best stoichiometry should be that ratio with the least intense bands for the individual components. It is necessary to have determined the amino-acid sequence of the light and heavy chain variable domains before interpreting the electron density of the antigen–antibody complex. The resultant cDNA sequencing represents an additional investment of time. Sequence determination [[24]Larrick J.W. Borrebaeck A.K. et al.Rapid cloning of rearranged immunoglobulin genes from human hybridoma cells using mixed primers and the polymerase chain reaction.Biochem. Biophys. Res. Comm. 1989; 160 (89273586): 1250-1256Crossref PubMed Scopus (144) Google Scholar] for the Fab used for crystallization of HIV p24 (derived from a mouse IgG1 with a k light chain) was accomplished with the ‘Ig-Prime’ kit purchased from Novagen Inc. (597 Science Drive, Madison, WI 53711). The amino-acid sequence of the Fv (murine IgG1, k) fragment that interacts with the integral membrane protein cytochrome c oxidase shows an unusual cysteine residue in the complimentarity-determining region (CDR)–H2 at position H50. Michel and co-workers presumed that this cysteine, which is not involved in disulfide bond formation, might be the cause of low recovery of the Fv fragment from the Escherichia coli periplasmic space. After mutating this residue to a serine, the amount of Fv that could be purified increased approximately 20-fold [[25]Ostermeier C. Essen L.-O. Michel H. Crystals of an antibody Fv fragment against an integral membrane protein diffracting to 1.28 å resolution.Proteins. 1995; 21 (95232124): 74-77Crossref PubMed Scopus (23) Google Scholar] while the binding characteristics to cytochrome c oxidase remained unaltered. The development of recombinant DNA technology for immunodiagnostic and immunotherapeutic applications is likely to lead to simpler methods for producing Fv fragments. Phage display technology [26Barbas C.F. Kang A.S. Lerner R.A. Benkovic S.J. Assembly of combinatorial antibody libraries on phage surfaces: The gene III site.Proc. Natl. Acad. Sci. USA. 1991; 88 (91376069): 7978-7982Crossref PubMed Scopus (990) Google Scholar, 27Burton D.R. Barbas C.F. Human antibodies from combinatorial libraries.in: Clark M Protein Engineering of Antibody Molecules for Prophylactic and Therapeutic Applications in Man. Academic Titles, Nottingham, U.K1993: 65-82Google Scholar] coupled with a large combinatorial immunoglobulin library, can also rapidly produce a large number of monoclonal antibodies against a given antigen [[28]Hoogenboom H.R. Marks J.D. Griffiths A.D. Winter G. Building antibodies from their genes.in: Moller G Immunological Reviews 130. Munksgard, Copenhagen1992: 41-68Google Scholar]. The presence of a Fab or Fv fragment in the crystal can aid in the crystallographic structure determination itself, as the fragment can be used for molecular replacement or as a recipient of heavy atom labels. Unfortunately, the molecular replacement phases, based on the cognate Fab, in the structure determination of HIV p24, were insufficient to solve the structure. They had to be augmented by multiple isomorphous replacement phases, single wavelength anomalous dispersion of a lead derivative and averaging over the two non-crystallographically related complexes in the asymmetric unit of the crystal. It may also be useful to crystallize the Fabfragment on its own and use the resultant structure as the molecular replacement model, although the elbow angle is likely to differ in different environments. In the structure determination of HIV-p24, the Fab fragment also served to identify the p24 electron density associated with the residues known to be the epitope for Fab binding. The crystal lattice contacts were primarily between the Fv or Fab fragments in the structure determinations of the cytochrome c oxidase and HIV p24, respectively. Although this was an advantage in the former, it was a problem in the latter where the occupancy of the antigen within the Fab crystal lattice was incomplete and variable. However, particularly for small, single domain, monomeric proteins, the formation of a crystal lattice based only on contacts between the larger and more soluble Fab fragments is likely to be an advantage. We thank Andrew Prongay for his earlier contributions to the crystallization of the HIV p24 Fab complex, Jan McClure (Bristol Myers-Squibb Pharmaceutical Corporation, Seattle, WA) for providing the anti-p24 monoclonals, Lorna Ehrlich and Carol Carter (SUNY, Stony Brook, NY) for the recombinant HIV p24, and Tianwei Lin for discussions on antibody sequencing. We are grateful for an MRC and an NIH grant for MGR and for a research scholar award to LCK from the American Foundation for AIDS Research. Ladislau C Kovari, Cory Momany and Michael G Rossmann, Department of Biological Sciences, Purdue University, West Lafayette, IN 47907-1392, USA.