High affinity RNase S-peptide variants obtained by phage display have a novel "hot-spot" of binding energy.

Using phage display mutagenesis, high affinity variants of RNase S-peptide were produced that bind to RNase S-protein over 100-fold more tightly than the wild type S-peptide. The S-peptide: S-protein interface was further characterized using "biased" phage display libraries, where each targeted residue was constrained to be either polar or nonpolar. The use of these tailored libraries placed constraints on the type of interactions present during affinity maturation process and allowed more amino acids to be randomized simultaneously. These results, in conjunction with kinetic association and dissociation constants determined by surface plasmon resonance (SPR), highlight the role of a single mutation (A5W) in increasing S-peptide binding affinity. High affinity S-peptide variants were only identified when tryptophan was present in the phage display library at position 5, suggesting that this residue is a "hot-spot" of binding energy in the high affinity variants. Analysis of SPR data in the presence of denaturant suggests that the increased affinity is a result of increased hydrophobic interactions in the transition state rather than a stabilization of helical structure.

The structure and activity of a monomeric interferon-gamma:alpha-chain receptor signaling complex.

BACKGROUND: Interferon-gamma (IFN-gamma) is a homodimeric cytokine that exerts its various activities by inducing the aggregation of two different receptors. The alpha chain receptor (IFN-gammaRalpha) is a high affinity receptor that binds to IFN-gamma in a symmetric bivalent manner to form a stable, intermediate 1:2 complex. This intermediate forms a binding template for the subsequent binding of two copies of the second receptor, beta chain receptor (IFN-gammaRbeta), producing the active 1:2:2 signaling complex. RESULTS: A single chain monovalent variant of IFN-gamma (scIFN-gamma) was constructed and complexed to one copy of the extracellular domain (ECD) of IFN-gammaRalpha. The structure of this 1:1 complex was determined and the hormone-receptor interface shown to be characterized by a number of hydrophilic interactions mediated by several highly ordered water networks. The scIFN-gamma interface consists of segments from each of the monomer chains of the homodimer. The principal hydrophobic contact of the receptor involves a tripeptide segment of the receptor having an unusual and high energy conformation. Despite containing only one binding site for IFN-gammaRalpha, the monovalent scIFN-gamma molecule has significant activity in antiviral biological assays. CONCLUSIONS: ScIFN-gamma binds the ECD of IFN-gammaRalpha through a highly hydrated interface with an important set of hormone-receptor contacts mediated through structured waters. Although the interface is highly hydrated, it supports tight binding and has a considerable degree of specificity. The biological activity of scIFN-gamma confirms that the scIFN-gamma:IFN-gammaRalpha complex represents a productive intermediate and that it can effectively recruit the other required component, IFN-gammaRbeta, to signal based on the 1:1:1 complex.

Ternary complex between placental lactogen and the extracellular domain of the prolactin receptor.

The structure of the ternary complex between ovine placental lactogen (oPL) and the extracellular domain (ECD) of the rat prolactin receptor (rPRLR) reveals that two rPRLR ECDs bind to opposite sides of oPL with pseudo two-fold symmetry. The two oPL receptor binding sites differ significantly in their topography and electrostatic character. These binding interfaces also involve different hydrogen bonding and hydrophobic packing patterns compared to the structurally related human growth hormone (hGH)-receptor complexes. Additionally, the receptor-receptor interactions are different from those of the hGH-receptor complex. The conformational adaptability of prolactin and growth hormone receptors is evidenced by the changes in local conformations of the receptor binding loops and more global changes induced by shifts in the angular relationships between the N- and C-terminal domains, which allow the receptor to bind to the two topographically distinct sites of oPL.

Biosynthetic phage display: a novel protein engineering tool combining chemical and genetic diversity.

BACKGROUND: Molecular diversity in nature is developed through a combination of genetic and chemical elements. We have developed a method that permits selective manipulation of both these elements in one protein engineering tool. It combines the ability to introduce non-natural amino acids into a protein using native chemical ligation with exhaustive targeted mutagenesis of the protein via phage-display mutagenesis. RESULTS: A fully functional biosynthetic version of the protease inhibitor eglin c was constructed. The amino-terminal fragment (residues 8-40) was chemically synthesized with a non-natural amino acid at position 25. The remaining carboxy-terminal fragment was expressed as a 30-residue peptide extension of gIIIp or gVIIIp on filamentous phage in a phage-display mutagenesis format. Native chemical ligation was used to couple the two fragments and produced a protein that refolded to its active form. To facilitate the packing of the introduced non-natural amino acid, residues 52 and 54 in the carboxy-terminal fragment were fully randomized by phage-display mutagenesis. Although the majority of the observed solutions for residues 52 and 54 were hydrophobic - complementing the stereochemistry of the introduced non-natural amino acid - a significant number of residues (unexpected because of stereochemical and charge criteria) were observed in these positions. CONCLUSIONS: Peptide synthesis and phage-display mutagenesis can be combined to produce a very powerful protein engineering tool. The physical properties of the environment surrounding the introduced non-natural residue can be selected for by evaluating all possible combinations of amino acid types at a targeted set of sites using phage-display mutagenesis.

Deciphering the role of the electrostatic interactions involving Gly70 in eglin C by total chemical protein synthesis.

Eglin c from the leech Hirudo medicinalis is a potent protein inhibitor of many serine proteinases including chymotrypsin and subtilisins. Unlike most small protein inhibitors whose solvent-exposed enzyme-binding loop is stabilized primarily by disulfide bridges flanking the reactive-site peptide bond, eglin c possesses an enzyme-binding loop supported predominantly by extensive electrostatic/H-bonding interactions involving three Arg residues (Arg48, Arg51, and Arg53) projecting from the scaffold of the inhibitor. As an adjacent residue, the C-terminal Gly70 participates in these interactions via its alpha-carboxyl group interacting with the side chain of Arg51 and the main chain of Arg48. In addition, the amide NH group of Gly70 donates an H-bond to the carbonyl C=O groups of Arg48 and Arg51. To understand the structural and functional relevance of the electrostatic/H-bonding network, we chemically synthesized wild-type eglin c and three analogues in which Gly70 was either deleted or replaced by glycine amide (NH(2)CH(2)CONH(2)) or by alpha-hydroxylacetamide (HOCH(2)CONH(2)). NMR analysis indicated that the core structure of eglin c was maintained in the analogues, but that the binding loop was significantly perturbed. It was found that deletion or replacement of Gly70 destabilized eglin c by an average of 2.7 kcal/mol or 20 degrees C in melting temperature. As a result, these inhibitors become substrates for their target enzymes. Binding assays on these analogues with a catalytically incompetent subtilisin BPN' mutant indicated that loss or weakening of the interactions involving the carboxylate of Gly70 caused a decrease in binding by approximately 2 orders of magnitude. Notably, for all four synthetic inhibitors, the relative free energy changes (DeltaDeltaG) associated with protein destabilization are strongly correlated (slope = 0.94, r(2) = 0. 9996) with the DeltaDeltaG values derived from a decreased binding to the enzyme.

The 2.0 A structure of bovine interferon-gamma; assessment of the structural differences between species.

The structure of bovine interferon-gamma (IFN-gamma) was determined by multiple isomorphous replacement at 2.0 A resolution. Bovine IFN-gamma crystallizes in two related crystal forms. Crystal form 1 diffracts to 2.9 A resolution and is reproducible and stable to derivatization. Crystal form 2 diffracts to 2.0 A resolution, but shows significant non-isomorphism from crystal to crystal. The previously determined structures of several different species of INF-gamma were either at too low a resolution [human, 1hig; Ealick et al. (1991), Science, 252, 698-702] or were too inaccurate [bovine, 1rfb; Samudzi & Rubin (1993), Acta Cryst. D49(6), 505-512; rabbit, 2rig; Samudzi et al. (1991), J. Biol. Chem. 266(32), 21791-21797] for the structure to be solved by molecular replacement. The structure was solved in crystal form 1 using two derivatives produced by chemically modifying two free cysteine residues that were introduced by site-directed mutagenesis (Ser30Cys, Asn59Cys). After model building and refinement, the final R value was 21.8% (R(free) = 30.9%) for all data in the resolution range 8.0-2.9 A. The crystal form 1 structure was then used as a molecular-replacement model for crystal form 2 data collected from a flash-cooled crystal. Subsequent model building and refinement, using all data in the resolution range 15.0-2.0 A, gave an R value of 19.7% and an R(free) of 27.5%. Pairwise comparison of C(alpha) positions of bovine IFN-gamma (BOV) and the previously determined 1rfb and 2rig structures indicated some significant differences in the models (r.m.s.d. values for BOV to 1rfb, 4.3 A; BOV to 2rig, 4.0 A). An assessment of the quality of the structures was made using the 3D-1D algorithm [Eisenberg et al. (1992), Faraday Discuss. 93, 25-34]. The resulting statistical scoring indicated that BOV was consistent with expected criteria for a 2.0 A structure, whereas both 1rfb and 2rig fell below acceptable criteria.

Probing intermolecular backbone H-bonding in serine proteinase-protein inhibitor complexes.

BACKGROUND: Intermolecular backbone H-bonding (N-H.O=C) is a common occurrence at the interface of protein-protein complexes. For instance, the amide NH groups of most residues in the binding loop of eglin c, a potent serine proteinase inhibitor from the leech Hirudo medicinalis, are H-bonded to the carbonyl groups of residues in the target enzyme molecules such as chymotrypsin, elastase and subtilisins. We sought to understand the energetic significance of these highly conserved backbone-backbone H-bonds in the enzyme-inhibitor complexes. RESULTS: We synthesized an array of backbone-engineered ester analogs of eglin c using native chemical ligation to yield five inhibitor proteins each containing a single backbone ester bond from P3 to P2' (i.e. -CONH-to -COO-). The structure at the ligation site (P6-P5) is essentially unaltered as shown by a high-resolution analysis of the subtilisin-BPN'-eglin c complex. The free-energy changes (DeltaDeltaGNH-->O) associated with the binding of ester analogs at P3, P1 and P2' with bovine alpha-chymotrypsin, subtilisin Carlsberg and porcine pancreatic elastase range from 0-4.5 kcal/mol. Most markedly, the NH-->O substitution at P2 not only stabilizes the inhibitor but also enhances binding to the enzymes by as much as 500-fold. CONCLUSIONS: Backbone H-bond contributions are context dependent in the enzyme-eglin c complexes. The interplay of rigidity and adaptability of the binding loop of eglin c seems to play a prominent role in defining the binding action.

Crystallization and preliminary X-ray analysis of a 1:1 complex between a designed monomeric interferon-gamma and its soluble receptor.

A variant of human interferon-gamma (IFN-gamma) has been created in which the two chains of the homodimeric cytokine were linked N- to C-terminus by an eight residue polypeptide linker. The sequence of this linker was derived from a loop in bira bifunctional protein, and was determined from a structural database search. This "single-chain" variant was used to create an IFN-gamma molecule that binds only a single copy of the alpha-chain receptor, rather than the 2 alpha-chain receptor: 1 IFN-gamma binding stoichiometry observed for the native hormone. Crystals have been grown of a 1:1 complex between this single-chain molecule and the extracellular domain of its alpha-chain receptor. These crystals diffract beyond 2.0 A, significantly better than the 2.9 A observed for the native 2:1 complex. Density calculations suggest these crystals contain two complexes in the asymmetric unit; a self-rotation function confirms this conclusion.

Crystallization of ovine placental lactogen in a 1:2 complex with the extracellular domain of the rat prolactin receptor.

Growth hormone and prolactin control somato-lactogenic biology. While high-resolution crystal structures have been determined for receptor complexes of human growth hormone, no such information exists for prolactin. A stable 1:2 complex was formed between ovine placental lactogen, a close prolactin homologue, and two copies of the extracellular portion of the rat prolactin receptor. Using synchrotron radiation, native data have been collected to 2.3 A. Crystals contain one complex per asymmetric unit. The crystal structure of this complex will shed light on the structural reasons for cross-reactivity and specificity among the endocrine hormones, placental lactogen, prolactin and growth hormone.

Crystal structures of bovine chymotrypsin and trypsin complexed to the inhibitor domain of Alzheimer's amyloid beta-protein precursor (APPI) and basic pancreatic trypsin inhibitor (BPTI): engineering of inhibitors with altered specificities.

The crystal structures of the inhibitor domain of Alzheimer's amyloid beta-protein precursor (APPI) complexed to bovine chymotrypsin (C-APPI) and trypsin (T-APPI) and basic pancreatic trypsin inhibitor (BPTI) bound to chymotrypsin (C-BPTI) have been solved and analyzed at 2.1 A, 1.8 A, and 2.6 A resolution, respectively. APPI and BPTI belong to the Kunitz family of inhibitors, which is characterized by a distinctive tertiary fold with three conserved disulfide bonds. At the specificity-determining site of these inhibitors (P1), residue 15(I)4 is an arginine in APPI and a lysine in BPTI, residue types that are counter to the chymotryptic hydrophobic specificity. In the chymotrypsin complexes, the Arg and Lys P1 side chains of the inhibitors adopt conformations that bend away from the bottom of the binding pocket to interact productively with elements of the binding pocket other than those observed for specificity-matched P1 side chains. The stereochemistry of the nucleophilic hydroxyl of Ser 195 in chymotrypsin relative to the scissile P1 bond of the inhibitors is identical to that observed for these groups in the trypsin-APPI complex, where Arg 15(I) is an optimal side chain for tryptic specificity. To further evaluate the diversity of sequences that can be accommodated by one of these inhibitors, APPI, we used phage display to randomly mutate residues 11, 13, 15, 17, and 19, which are major binding determinants. Inhibitors variants were selected that bound to either trypsin or chymotrypsin. As expected, trypsin specificity was principally directed by having a basic side chain at P1 (position 15); however, the P1 residues that were selected for chymotrypsin binding were His and Asn, rather than the expected large hydrophobic types. This can be rationalized by modeling these hydrophilic side chains to have similar H-bonding interactions to those observed in the structures of the described complexes. The specificity, or lack thereof, for the other individual subsites is discussed in the context of the "allowed" residues determined from a phage display mutagenesis selection experiment.

Hydroxyl and water molecule orientations in trypsin: comparison to molecular dynamic structures.

A comparison is presented of experimentally observed hydroxyl and water hydrogens in trypsin determined from neutron density maps with the results of a 140ps molecular dynamics (MD) simulation. Experimental determination of hydrogen and deuterium atom positions in molecules as large as proteins is a unique capability of neutron diffraction. The comparison addresses the degree to which a standard force-field approach can adequately describe the local electrostatic and van der Waals forces that determine the orientations of these hydrogens. The molecular dynamics simulation, based on the all-atom AMBER force-field, allowed free rotation of all hydroxyl groups and movement of water molecules making up a bath surrounding the protein. The neutron densities, derived from 2.1A D2O-H2O difference Fourier maps, provide a database of 27 well-ordered hydroxyl hydrogens. Virtually all of the simulated hydroxyl orientations are within a standard deviation of the experimentally-observed positions, including several examples in which both the simulation and the neutron density indicate that a hydroxyl group is shifted from a 'standard' rotamer. For the most highly ordered water molecules, the hydrogen distributions calculated from the trajectory were in good agreement with neutron density; simulated water molecules that displayed multiple hydrogen bonding networks had correspondingly broadened neutron density profiles. This comparison was facilitated by development of a method to construct a pseudo 2A density map based on the hydrogen atom distributions from the simulation. The degree of internal water molecules is shown to result primarily from the electrostatic environment surrounding that water molecule as opposed to the cavity size available to the molecule. A method is presented for comparing the discrete observations sampled in a dynamics trajectory with the time-averaged data obtained from X-ray or neutron diffraction studies. This method is particularly useful for statically-disordered water molecules, in which the average location assigned from a trajectory may represent a site of relatively low occupancy.

A comparison of neutron diffraction and molecular dynamics structures: hydroxyl group and water molecule orientations in trypsin.

A comparison is presented of experimentally observed hydroxyl and water hydrogen atoms in trypsin determined from neutron density maps with the results of a 140 ps molecular dynamics simulation. Experimental determination of hydrogen and deuterium atom positions in molecules as large as proteins is a unique capability of neutron diffraction. The comparison addresses the degree to which a standard force-field approach can adequately describe the local electrostatic and van der Waals forces that determine the orientations of these hydrogen atoms. The molecular dynamics simulation, based on the all-atom AMBER force-field, allowed free rotation of all hydroxyl groups and movement of water molecules making up a bath surrounding the protein. The neutron densities, derived from 2.1 A 2H2O-H2O difference Fourier maps, provide a database of 27 well-ordered hydroxyl hydrogen atoms. Virtually all of the simulated hydroxyl orientations are within a standard deviation of the experimentally observed positions, including several examples in which both the simulation and the neutron density indicate that a hydroxyl group is shifted from a "standard" rotamer. For the most highly ordered water molecules, the hydrogen distributions calculated from the trajectory were in good agreement with neutron density; simulated water molecules that displayed multiple hydrogen-bonding networks had correspondingly broadened neutron density profiles. This comparison was facilitated by development of a method to construct a pseudo 2 A density map based on the hydrogen atom distributions from the simulation. This method is particularly useful for statically disordered water molecules, in which the average location assigned from a trajectory may represent a site of relatively low occupancy. The degree of disorder of internal water molecules is shown to result primarily from the electrostatic environment surrounding that water molecule as opposed to the cavity size available to the molecule.

Structure of the growth hormone-receptor complex and mechanism of receptor signaling.

The structure of the growth hormone-receptor complex discussed here is the first such system to be studied at the level of atomic detail and provides unique information that elucidates the mechanism of signal transduction of an important receptor family. The growth hormone receptor is a single-pass receptor, with an extracellular protein domain, a transmembrane domain and an intracellular protein domain. Structural data, obtained by crystallography, indicate that there are actually two growth hormone receptors that encapsulate the bound hormone. Although the topology of the hormone is asymmetric, the receptors can use their same sequence of residues to bind to different structural motifs by changing conformation. This mechanism of aggregation controls signal transduction. It may be possible to use this information in the design of radiolabeled ligands for molecular nuclear medicine studies involving the concentration or occupancy of growth receptors.

The production and X-ray structure determination of perdeuterated Staphylococcal nuclease.

Staphylococcal Nuclease (SNase) has been chosen as a model protein system to evaluate the improvement in neutron diffraction data quality using fully perdeuterated protein. Large quantities of the protein were expressed in Escherichia coli grown in medium containing deuterated amino acids and deuterated water (D2O) and then purified. The mean perdeuteration level of the non-exchangable sites in the protein was found to be 96% by electrospray ionization mass spectrometry. The perdeuterated enzyme was crystallized and its X-ray structure determined. Crystals of perdeuterated SNase have been grown to 1.5 mm3. Crystallization conditions, space group and cell parameters were found to be the same for both native and perdeuterated forms of the protein. Comparison of these two forms of SNase revealed no significant structural differences between them at the atomic resolution of 1.9 A. Data collection using crystals of the perdeuterated protein is scheduled at the Brookhaven High Flux Beam Reactor.

The X-ray structure of a growth hormone-prolactin receptor complex.

The human pituitary hormones, growth hormone (hGH) and prolactin (hPRL), regulate a large variety of physiological processes, among which are growth and differentiation of muscle, bone and cartilage cells, and lactation. These activities are initiated by hormone-receptor binding. The hGH and hPRL receptors (hGHR and hPRLR, respectively) are single-pass transmembrane receptors from class 1 of the haematopoietic receptor superfamily. This classification is based on sequence similarity in their extracellular domains, notably a highly conserved pentapeptide, the so-called 'WSXWS box', the function of which is controversial. All ligands in class 1 activate their respective receptors by clustering mechanisms. In the case of hGH, activation involves receptor homodimerization in a sequential process: the active ternary complex containing one ligand and two receptor molecules is formed by association of a receptor molecule to an intermediate 1:1 complex. hPRL does not bind to the hGH receptor, but hGH binds to both the hGHR and hPRLR, and mutagenesis studies have shown that the receptor-binding sites on hGH overlap. We present here the crystal structure of the 1:1 complex of hGH bound to the extracellular domain of the hPRLR. Comparisons with the hGH-hGHR complex reveal how hGH can bind to the two distinctly different receptor binding surfaces.

Comparison of the intermediate complexes of human growth hormone bound to the human growth hormone and prolactin receptors.

The crystal structures of complexes of human growth hormone (hGH) with the growth hormone and prolactin receptors (hGHR and hPRLR, respectively), together with the mutational data available for these systems, suggest that an extraordinary combination of conformational adaptability, together with finely tuned specificity, governs the molecular recognition processes operative in these systems. On the one hand, in the active 1:2 ligand-receptor complexes, 2 copies of the same receptor use the identical set of binding determinants to recognize topographically different surfaces on the hormone. On the other hand, comparing the 1:1 hGH-hGHR and hGH-hPRLR complexes, 2 distinct receptors use this same set of binding determinants to interact with the identical binding site on the ligand, even though few residues among the binding determinants are conserved. The structural evidence demonstrates that this versatility is accomplished by local conformational flexibility of the binding loops, allowing adaptation to different binding environments, together with rigid-body movements of the receptor domains, necessary for the creation of specific interactions with the same binding site.

The crystal structure of affinity-matured human growth hormone at 2 A resolution.

A variant of human growth hormone (hGH), in which 15 mutations were introduced with phage display mutagenesis to improve receptor binding affinity by 400-fold, yielded two related crystal forms diffracting to high resolution. The structure of this variant was determined in both crystal forms, one at 2.0 A resolution and one at 2.4 A resolution, using molecular replacement with wild-type hGH taken from the receptor complex structure as a search model. Crystallographic refinement of the 2 A structure gave an R-value R-value of 18.5% for data in the resolution range 8 to 2 A. The final model consists of residues 1 to 128 and 155 to 191, with three side-chains modeled in alternative conformations, together with 77 water molecules. Comparison of the structure with wild-type hGH shows that most of the secondary structural elements are unchanged. The exception is the first turn of the third helix in the four-helix bundle core, which is unraveled in the present variant. Analysis of the two related packing environments suggests that this change is caused by crystal packing forces. A large change in the orientation of a short segment of helix found in the connection between the first two core helices is interpreted as evidence for rigid-body variability of this helical segment. Analysis of the mutations in light of the structure of the wild-type hGH/receptor complex shows that six of the mutations are buried in the hormone, whereas the remaining nine involve residues that interact with the receptor in the complex.

X-ray structures of the antigen-binding domains from three variants of humanized anti-p185HER2 antibody 4D5 and comparison with molecular modeling.

The X-ray structures of 1 Fv and 2 Fab humanized anti-p185HER2 antibody fragments (IgG1-kappa) have been determined at a resolution between 2.7 A and 2.2 A. The antibodies are three different versions of a human antibody framework onto which the antigen recognition loops from a murine antibody (4D5) have been grafted. The sequences of the three versions differ in the framework region at positions L55, H78 and H102. The version 8 Fv fragment crystallizes in space group P2(1) with cell parameters a = 37.6 A, b = 63.4 A, c = 90.2 A, beta = 98.2 degrees, with two molecules per asymmetric unit, and has been refined against data 10.0 A-2.2 A to an R-factor of 18.3%. Versions 4 and 7 Fabs crystallize in space group P1 with cell parameters a = 39.2 A, b = 80.2 A, c = 86.1 A, alpha = 113.1 degrees, beta = 92.7 A, gamma = 102.6 A and two molecules per asymmetric unit. Version 4 has been refined against data 10.0 A-2.5 A resolution to an R-factor of 17.9%. Version 7 has been refined against data 10 A-2.7 A to an R-factor of 17.1%. The X-ray structures have been used to assess the accuracy of structural predictions made via molecular modeling, and they confirm the structural role of certain framework residues and the conformations of five of six complementarity determining regions (CDRS). The average deviation of the model from the X-ray structures is within the range observed among the X-ray structures for 81% of the C alpha atoms. Of the hydrogen bonds common to the X-ray structures, 94% of the main-chain-main-chain and 79% of the main-chain-side-chain ones were predicted by the model. The side-chain conformation was predicted correctly for 79% of the buried residues. The third CDR in the heavy chain is variable, differing by up to 8 A between molecules within an asymmetric unit. The structural relationship between variable domains of light and heavy chains is not significantly altered by the absence of constant domains in the Fv molecule. The antigen-binding potential of an unusual light chain sequence has been confirmed. The arginine at position 66 interacts with the first light chain CDR, but in a fashion somewhat different than predicted. A substitution of a leucine for an alanine side-chain directed between the beta-sheets has only relatively small and local effects.(ABSTRACT TRUNCATED AT 400 WORDS)