1H nuclear magnetic resonance studies of hen lysozyme-N-acetylglucosamine oligosaccharide complexes in solution. Application of chemical shifts for the comparison of conformational changes in solution and in the crystal.

Two-dimensional 1H nuclear magnetic resonance spectroscopy has been used to examine the complexes formed in solution between hen egg-white lysozyme and N-acetylglucosamine (GlcNAc) oligosaccharides. Changes in chemical shift have been measured for resonances of the majority of residues of lysozyme on binding the monomer, dimer and trimer of GlcNAc. The three inhibitors induce very similar changes in chemical shift, and these increase slightly with the length of the oligosaccharide. The largest changes are confined principally to the vicinity of site C in the active site cleft of the enzyme. These changes in chemical shift have been compared with differences in the ring current chemical shifts calculated from the crystal structures of unbound and GlcNAc3 bound lysozyme. This comparison suggests that the major conformational changes of residues in the vicinity of site C of the enzyme, that are caused by the binding of GlcNAc3, observed in the diffraction studies are at least consistent with the changes that occur in solution. Small changes in chemical shift are observed in the enzyme in regions remote from the active site, which indicate that the effects of inhibitor binding are felt throughout the enzyme. These changes in chemical shift correlate to a lesser extent than those near site C with the changes in chemical shift predicted from changes in conformation observed in the crystal structures. The results illustrate that chemical shifts are useful in assessing the significance of small conformational changes in proteins, although the usefulness of this approach will be limited by the resolution of the crystallographic structures, as well as the uncertainties in the origins of the chemical shift. Although conformational changes in site C account for many of the changes in the NMR spectrum of lysozyme, evidence is, however, presented for multiple binding sites for the GlcNAc oligosaccharides in solution, perhaps involving partial occupancy of site D.

Refinement of an enzyme complex with inhibitor bound at partial occupancy. Hen egg-white lysozyme and tri-N-acetylchitotriose at 1.75 A resolution.

The structure of the tri-N-acetylchitotriose inhibitor complex of hen egg-white lysozyme has been refined at 1.75 A resolution, using data collected from a complex crystal with ligand bound at less than full occupancy. To determine the exact value of the inhibitor occupancy, a model comprising unliganded and sugar-bound protein molecules was generated and refined against the 1.75 A data, using a modified version of the Hendrickson & Konnert least-squares procedure. The crystallographic R-factor for the model was found to fall to a minimum at 55% bound sugar. Conventional refinement assuming unit occupancy was found to yield incorrect thermal and positional parameters. Application of the same refinement procedures to an earlier 2.0 A data set, collected independently on different complex crystals by Blake et al. gave less consistent results than the 1.75 A refinement. From an analysis of the high resolution structure a detailed picture of the protein-carbohydrate interactions in the non-productive complex has emerged, together with the conformation and mobility changes that accompany ligand binding. The specificity of interaction between the protein and inhibitor, bound in subsites A to C of the active site, is seen to be generated primarily by an extensive network of hydrogen bonds, both to the protein itself and to bound solvent molecules. The latter also play an important role in maintaining the structural integrity of the active site cleft in the apo-protein.

Molecular dynamics simulations of native and substrate-bound lysozyme. A study of the average structures and atomic fluctuations.

Molecular dynamics simulations of hen egg-white lysozyme in the free and substrate-bound states are reported and the nature of the average structures and atomic fluctuations are analyzed. Crystallographic water molecules of structural importance, as determined by hydrogen-bonding, were included in the simulations. Comparisons are made between the dynamics and the X-ray results for the atomic positions, the main-chain and side-chain dihedral angles, and the hydrogen-bonding geometry. Improvements over earlier simulations in the potential energy function and methodology resulted in stable trajectories with the C alpha co-ordinates within 1.5 A of the starting X-ray structure. Structural features analyzed in the simulations agreed well with the X-ray results except for some surface residues. The Asx chi 2 dihedral distribution and the geometry of hydrogen bonding at reverse turns show differences; possible causes are discussed. The relation between the magnitudes and time-scales of the residue fluctuations and secondary structural features, such as helices beta-sheets and coiled loops, is examined. Significant differences in the residue mobilities between the simulations of the free and substrate-bound states were found in a region of the enzyme that is in direct contact with the substrate and in a region that is distant from the active-site cleft. The dynamic behavior of the structural water molecules is analyzed by examining the correlation between the fluctuations of the water oxygens and the lysozyme heavy-atoms to which they are hydrogen-bonded.

Antibody-combining sites: prediction and design.

For maximum value, a predicted model of an antibody-combining site should have an accuracy approaching that of an X-ray structure (1.6-2.7 A). In addition, the method by which the combining site is modelled should make no demands on the user of a sort that require arbitrary or subjective decisions to be made during the process. We have made substantial progress towards this objective and some recent results are reviewed. In addition, we describe how the modelling protocols developed can aid in the design of novel features within the antibody-combining site. The particular design example reported here suggests an approach for the introduction of metal-binding sites to create metallo-antibodies. This type of modification may be useful in the design of immunobiosensors, the induction of catalytic activity or simply as an alternative to metal chelates in the preparation of antibodies for imaging.

Molecular modeling of antibody combining sites.

Each of the six CDRs of Gloop2 is shown with the modeled structure in. Overall, the results obtained using the combined algorithm are similar in accuracy to those achieved using the canonical method of Chothia et al. However, the canonical method is limited to those loops where the key residues identified by Chothia are present. With the number of antibody structures currently available, it is not possible to classify CDR-H3 into canonical ensembles. Additionally, a small percentage of examples in the remaining CDRs do not match the current canonical classifications and the protein engineer may well wish to mutate the key residues, precluding the use of Chothia's method for modeling the resulting conformation. Thus the best approach appears to be to use Chothia's method (at least to model the backbone conformation) when the loop to be modeled is represented in the database of canonical structures. Any other loops, either unrepresented among the known canonicals (including CDR-H3), or where mutations have been made to the key residues, may then be modeled by the combined algorithm presented here.

Generation of an antibody with enhanced affinity and specificity for its antigen by protein engineering.

A detailed description of the interactions between an antibody and its epitope is necessary to allow an understanding of the way in which antibodies bind to antigenic surfaces presented by foreign molecules. Ideally this should be done by analysis of crystal structures of antibody-antigen complexes, but so far only two of these are available. An alternative strategy combines molecular modelling with site-directed mutagenesis (SDM) and using this we have generated a preliminary model of the complex between Gloop2, an antibody raised against a peptide containing the 'loop' determinant of hen egg-white lysozyme (HEL) which also binds the native protein, and its epitope on the protein surface. The main predictions from our model were; (1) that the surface of interaction between the antibody and the antigen is large (20 A X 15 A) and involves all the complementarity-determining regions (CDRs), (2) that electrostatic interactions were important in the formation of the complex, and (3) that conformational changes in either the loop or in the CDRs may occur during the formation of the complex. Here we report SDM studies which test some of these predictions; removal of two charged residues at the periphery of the combining site increases the affinity of the antibody for its antigen over 8-fold and decreases its ability to cross-react with closely-related antigens. This result is at variance with our original prediction but can be accommodated within our newly refined model; the role of electrostatics in antigen-antibody interactions is now questionable.

Modeling antibody hypervariable loops: a combined algorithm.

To be of any value, a predicted model of an antibody combining site should have an accuracy approaching that of antibody structures determined by x-ray crystallography (1.6-2.7 A). A number of modeling protocols have been proposed, which fall into two main categories--those that adopt a knowledge-based approach and those that attempt to construct the hypervariable loop regions of the antibody ab initio. Here we present a combined algorithm requiring no arbitrary decisions on the part of the user, which has been successfully applied to the modeling of the individual loops in two systems: the anti-lysozyme antibody HyHel-5, the crystal structure of which is as a complex with lysozyme [Sheriff, S., Silverton, E. W., Padlan, E. A., Cohen, G. H., Smith-Gill, S. J., Finzel, B. C. & Davies, D. R. (1987) Proc. Natl. Acad. Sci. USA 84, 8075-8079], and the free antigen binding fragment (Fab) of the anti-lysozyme peptide antibody, Gloop2. This protocol may be used with a high degree of confidence to model single-loop replacements, insertions, deletions, and side-chain replacements. In addition, it may be used in conjunction with other modeling protocols as a method by which to model particular loops whose conformations are predicted poorly by these methods.

Engineering antibody affinity and specificity.

A combination of ab initio calculations, "knowledge-based prediction", molecular graphics and site-directed mutagenesis has enabled us to probe the molecular details of antibody:antigen recognition and binding and to alter the affinity and specificity of an antibody for its antigen. The significance of electrostatic hydrogen bonding, hydrophilic/hydrophobic patch matching and van der Waals interactions as well as CDR:CDR interactions are discussed in relation to the results of site-directed mutagenesis experiments on the anti-lysozyme antibody Gloop2. The ability to generate reconstructed antibodies, chimeric antibodies, catalytic antibodies and the use of modelled antibodies for the design of drugs is discussed.

The prediction and characterization of metal binding sites in proteins.

The rational engineering of novel functions into proteins can only be attempted when the underlying structural scaffold on which the new function is displayed and the structure of the target protein are both well understood. To introduce functions mediated by metals it is therefore necessary to identify the principal liganding residues for the chosen metal, the required architecture of the metal-ligand complex and sites within the target protein that could accommodate such sites. Here we present a method that applies structural information from the protein data bank to the ab initio design and characterization of novel metal binding sites. The prediction method has been tested on 28 metalloprotein structures from the Brookhaven Protein Data Bank. It successfully identified > 90% of the metal binding sites. In addition, we have used the method to design and characterize zinc binding sites in two antibody structures. Metal binding studies on one of these putative metalloantibodies showed metal binding, confirming the predictive power of the method.

Antigen mobility in the combining site of an anti-peptide antibody.

The interaction between a high-affinity antibody, raised against a peptide incorporating the loop region of hen egg lysozyme (residues 57-84), and a peptide antigen corresponding to this sequence, has been probed by proton NMR. The two-dimensional correlated spectroscopy spectrum of the antibody-antigen complex shows sharp, well-resolved resonances from at least half of the bound peptide residues, indicating that the peptide retains considerable mobility when bound to the antibody. The strongly immobilized residues (which include Arg-61, Trp-62, Trp-63, and Ile-78) do not correspond to a contiguous region in the sequence of the peptide. Examination of the crystal structure of the protein shows that these residues, although remote in sequence, are grouped together in the protein structure, forming a hydrophobic projection on the surface of the molecule. The antibody binds hen egg lysozyme with only a 10-fold lower affinity than the peptide antigen. We propose that the peptide could bind to the antibody in a conformation that brings these groups together in a manner related to that found in the native protein, accounting for the high crossreactivity.

NMR structure of human erythropoietin and a comparison with its receptor bound conformation.

The solution structure of human erythropoietin (EPO) has been determined by nuclear magnetic resonance spectroscopy and the overall topology of the protein is revealed as a novel combination of features taken from both the long-chain and short-chain families of hematopoietic growth factors. Using the structure and data from mutagenesis studies we have elucidated the key physiochemical properties defining each of the two receptor binding sites on the EPO protein. A comparison of the NMR structure of the free EPO ligand to the receptor bound form, determined by X-ray crystallography, reveals conformational changes that may accompany receptor binding.

Efficiency of signalling through cytokine receptors depends critically on receptor orientation.

Human erythropoietin is a haematopoietic cytokine required for the differentiation and proliferation of precursor cells into red blood cells. It activates cells by binding and orientating two cell-surface erythropoietin receptors (EPORs) which trigger an intracellular phosphorylation cascade. The half-maximal response in a cellular proliferation assay is evoked at an erythropoietin concentration of 10 pM, 10(-2) of its Kd value for erythropoietin-EPOR binding site 1 (Kd approximately equal to nM), and 10(-5) of the Kd for erythropoietin-EPOR binding site 2 (Kd approximately equal to 1 microM). Overall half-maximal binding (IC50) of cell-surface receptors is produced with approximately 0.18 nM erythropoietin, indicating that only approximately 6% of the receptors would be bound in the presence of 10 pM erythropoietin. Other effective erythropoietin-mimetic ligands that dimerize receptors can evoke the same cellular responses but much less efficiently, requiring concentrations close to their Kd values (approximately 0.1 microM). The crystal structure of erythropoietin complexed to the extracellular ligand-binding domains of the erythropoietin receptor, determined at 1.9 A from two crystal forms, shows that erythropoietin imposes a unique 120 degrees angular relationship and orientation that is responsible for optimal signalling through intracellular kinase pathways.

Engineering a soluble extracellular erythropoietin receptor (EPObp) in Pichia pastoris to eliminate microheterogeneity, and its complex with erythropoietin.

The extracellular ligand-binding domain (EPObp) of the human EPO receptor (EPOR) was expressed both in CHO (Chinese Hamster Ovary) cells and in Pichia pastoris. The CHO and yeast expressed receptors showed identical affinity for EPO binding. Expression levels in P. pastoris were significantly higher, favoring its use as an expression and scale-up production system. Incubation of EPO with a fourfold molar excess of receptor at high protein concentrations yielded stable EPO-EPObp complexes. Quantification of EPO and EPObp in the complex yielded a molar ratio of one EPO molecule to two receptor molecules. Residues that are responsible for EPOR glycosylation and isomerization in Pichia were identified and eliminated by site-specific mutagenesis. A thiol modification was identified and a method was developed to remove the modified species from EPObp. EPObp was complexed with erythropoietin (EPO) and purified. The complex crystallized in two crystal forms that diffracted to 2.8 and 1.9 A respectively. (Form 1 and form 2 crystals were independently obtained at AxyS Pharmaceuticals, Inc. and Amgen, Inc. respectively.) Both contained one complex per asymmetric unit with a stoichiometry of two EPObps to one EPO.