Mutational and secondary structural analysis of the basolateral sorting signal of the polymeric immunoglobulin receptor.

The 17-juxtamembrane cytoplasmic residues of the polymeric immunoglobulin receptor contain an autonomous basolateral targeting signal that does not mediate rapid endocytosis (Casanova, J. E., G. Apodaca, and K. E. Mostov. Cell. 66:65-75). Alanine-scanning mutagenesis identifies three residues in this region, His656, Arg657, and Val660, that are most essential for basolateral sorting and two residues, Arg655 and Tyr668, that play a lesser role in this process. Progressive truncations suggested that Ser664 and Ile665 might also play a role in basolateral sorting. However, mutation of these residues to Ala or internal deletions of these residues did not affect basolateral sorting, indicating that these residues are probably not required for basolateral sorting. Two-dimensional NMR spectroscopy of a peptide corresponding to the 17-mer signal indicates that the sequence Arg658-Asn-Val-Asp661 has a propensity to adopt a beta-turn in solution. Residues COOH-terminal to the beta-turn (Arg662 to Arg669) seem to take up a nascent helix structure in solution. Substitution of Val660 with Ala destabilizes the turn, while mutation of Arg657 to Ala does not appear to affect the turn structure. Neither mutation detectably altered the stability of the nascent helix in the COOH-terminal portion of the peptide.

Structure-based strategies for drug design and discovery.

Most drugs have been discovered in random screens or by exploiting information about macromolecular receptors. One source of this information is in the structures of critical proteins and nucleic acids. The structure-based approach to design couples this information with specialized computer programs to propose novel enzyme inhibitors and other therapeutic agents. Iterated design cycles have produced compounds now in clinical trials. The combination of molecular structure determination and computation is emerging as an important tool for drug development. These ideas will be applied to acquired immunodeficiency syndrome (AIDS) and bacterial drug resistance.

Hydration of macromolecules.

Frozen protein and nucleic acid solutions at -35 degrees C show relatively narrow (100 milligauss) proton nuclear magnetic resonance signals which are assignable to water that is sufficiently mobile to reduce the dipolar broadening normally associated with solids. Hydration was found to be 0.3 to 0.5 gram of water per gram of protein. Nucleic acids are three to five times as hydrated as proteins. Conformational changes in solution produce detectable changes in linewidth or amount of "bound" water, or both. The very fact that the water signals can be observed by high resolution nuclear magnetic resonance suggests that it is not "ice-like" in any literal sense, although it is clearly less mobile than liquid water at the same temperature. A simple model is described which considers both surface hydration and trapped water.

Real-time color graphics in studies of molecular interactions.

Studies of the structures and interactions of large biological molecules require both coordinate data and three-dimensional visualization. Orthodox molecular models often bear a tenuous relationship to the coordinate data. In contrast, computer graphics requires that the display directly and accurately represent the data, and storage of modified configurations and recovery of original structures are simple. Software has been developed that allows real-time display of color line and surface displays of several interacting molecules, while quantitatively monitoring the stereochemistry.

Mutants of bovine pancreatic trypsin inhibitor lacking cysteines 14 and 38 can fold properly.

It is a generally accepted principle of biology that a protein's primary sequence is the main determinant of its tertiary structure. However, the mechanism by which a protein proceeds from an unfolded, disordered state to a folded, relatively well-ordered, native conformation is obscure. Studies have been initiated to examine the "genetics" of protein folding, with mutants of bovine pancreatic trypsin inhibitor (BPTI) being used to explore the nature of the specific intramolecular interactions that direct this process. Previous work with BPTI chemically modified at cysteines 14 and 38 indicated that transient disulfide bond formation by these residues contributed to efficient folding at 25 degrees C. In the present work, mutants of BPTI in which these cysteines were replaced by alanines or threonines were made and the mutant proteins were produced by a heterologous Escherichia coli expression system. At 25 degrees C in vitro, the refolding behavior of these mutants was characterized by a pronounced lag. However, when expressed at 37 degrees C in E. coli, or when refolded at 37 degrees or 52 degrees C in vitro, the mutant proteins folded readily into the native conformation, albeit at a rate somewhat slower than that exhibited by wild-type BPTI. These results indicate that, at physiological temperatures, BPTI lacking cysteines 14 and 38 can refold quantitatively.

Structure-activity studies of interleukin-2.

The critical role of interleukin-2 (IL-2) in immune response heightens the need to know its structure in order to understand its activity. New computer-assisted predictive methods for the assignment of secondary structure together with a method to predict the tertiary structure of a protein from data on its primary sequence and secondary structure were applied to IL-2. This method generated four topological families of structures, of which the most plausible is a right-handed fourfold alpha-helical bundle. Members of this family were shown to be compatible with existing structural data on disulfide bridges and monoclonal antibody binding for IL-2. Experimental estimates of secondary structure from circular dichroism and site-directed mutagenesis data support the model. A region likely to be important in IL-2 binding to its receptor was identified as residues Leu36, Met38, Leu40, Phe42, Phe44, and Met46.

Structure-based discovery of inhibitors of thymidylate synthase.

A molecular docking computer program (DOCK) was used to screen the Fine Chemical Directory, a database of commercially available compounds, for molecules that are complementary to thymidylate synthase (TS), a chemotherapeutic target. Besides retrieving the substrate and several known inhibitors, DOCK proposed putative inhibitors previously unknown to bind to the enzyme. Three of these compounds inhibited Lactobacillus casei TS at submillimolar concentrations. One of these inhibitors, sulisobenzone, crystallized with TS in two configurations that differed from the DOCK-favored geometry: a counterion was bound in the substrate site, which resulted in a 6 to 9 angstrom displacement of the inhibitor. The structure of the complexes suggested another binding region in the active site that could be exploited. This region was probed with molecules sterically similar to sulisobenzone, which led to the identification of a family of phenolphthalein analogs that inhibit TS in the 1 to 30 micromolar range. These inhibitors do not resemble the substrates of the enzyme. A crystal structure of phenolphthalein with TS shows that it binds in the target site in a configuration that resembles the one suggested by DOCK.

Water proton magnetic resonance studies of normal and sickle erythrocytes. Temperature and volume dependence.

The temperature and cell volume dependence of the NMR water proton line-width, spin-lattice, and spin-spin relaxation times have been studied for normal and sickle erythrocytes as well as hemoglobin A and hemoglobin S solutions. Upon deoxygenation, the spin-spin relaxation time (T2) decreases by a factor of 2 for sickle cells and hemoglobin S solutions but remains relatively constant for normal cells and hemoglobin A solutions. The spin-lattice relaxation time (T1) shows no significant change upon deoxygenation for normal or sickle packed red cells. Studies of the change in the NMR linewidth, T1 and T2 as the cell hydration is changed indicate that these parameters are affected only slightly by a 10-20% cell dehydration. This result suggests that the reported 10% cell dehydration observed with sickling is not important in the altered NMR properties. Low temperature studies of the linewidth and T1 for oxy and deoxy hemoglobin A and hemoglobin S solutions suggest that the "bound" water possesses similar properties for all four species. The low temperature linewidth ranges from about 250 Hz at -15 degrees C to 500 Hz at -36 degrees C and analysis of the NMR curves yield hydration values near 0.4 g water/g hemoglobin for all four species. The low temperature T1 data go through a minimum at -35 degrees C for measurements at 44.4 MHz and -50 degrees C for measurements at 17.1 MHz and are similar for oxy and deoxy hemoglobin A and hemoglobin S. These similarities in the low temperature NMR data for oxy and deoxy hemoglobin A and hemoglobin S suggest a hydrophobically driven sickling mechanism. The room temperature and low temperature relaxation time data for normal and sickle cells are interpreted in terms of a three-state model for intracellular water. In the context of this model the relaxation time data imply that type III, or irrotationally bound water, is altered during the sickling process.

Protein docking and complementarity.

Predicting the structures of protein-protein complexes is a difficult problem owing to the topographical and thermodynamic complexity of these structures. Past efforts in this area have focussed on fitting the interacting proteins together using rigid body searches, usually with the conformations of the proteins as they occur in crystal structure complexes. Here we present work which uses a rigid body docking method to generate the structures of three known protein complexes, using both the bound and unbound conformations of the interacting molecules. In all cases we can regenerate the geometry of the crystal complexes to high accuracy. We also are able to find geometries that do not resemble the crystal structure but nevertheless are surprisingly reasonable both mechanistically and by some simple physical criteria. In contrast to previous work in this area, we find that simple methods for evaluating the complementarity at the protein-protein interface cannot distinguish between the configurations that resemble the crystal structure complex and those that do not. Methods that could not distinguish between such similar and dissimilar configurations include surface area burial, solvation free energy, packing and mechanism-based filtering. Evaluations of the total interaction energy and the electrostatic interaction energy of the complexes were somewhat better. Of the techniques that we tried, energy minimization distinguished most clearly between the "true" and "false" positives, though even here the energy differences were surprisingly small. We found the lowest total interaction energy from amongst all of the putative complexes generated by docking was always within 5 A root-mean-square of the crystallographic structure. There were, however, several putative complexes that were very dissimilar to the crystallographic structure but had energies that were close to that of the low energy structure. The magnitude of the error in energy calculations has not been established in macromolecular systems, and thus the reliability of the small differences in energy remains to be determined. The ability of this docking method to regenerate the crystallographic configurations of the interacting proteins using their unbound conformations suggests that it will be a useful tool in predicting the structures of unsolved complexes.

Molecular docking to ensembles of protein structures.

Until recently, applications of molecular docking assumed that the macromolecular receptor exists in a single, rigid conformation. However, structural studies involving different ligands bound to the same target biomolecule frequently reveal modest but significant conformational changes in the target. In this paper, two related methods for molecular docking are described that utilize information on conformational variability from ensembles of experimental receptor structures. One method combines the information into an "energy-weighted average" of the interaction energy between a ligand and each receptor structure. The other method performs the averaging on a structural level, producing a "geometry-weighted average" of the inter-molecular force field score used in DOCK 3.5. Both methods have been applied in docking small molecules to ensembles of crystal and solution structures, and we show that experimentally determined binding orientations and computed energies of known ligands can be reproduced accurately. The use of composite grids, when conformationally different protein structures are available, yields an improvement in computational speed for database searches in proportion to the number of structures.

Protein-ligand dynamics. A 96 picosecond simulation of a myoglobin-xenon complex.

A 96 picosecond dynamics trajectory of myoglobin with five xenon-probe ligands in internal cavities is examined to study the effect of protein motions on ligand motion and internal cavity fluctuations. Average structural and energetic properties indicate that the simulation is well behaved. The average protein volume is similar to the volume of the X-ray model and the main-chain atom root-mean-square deviation between the X-ray model and the average dynamical structure is 1.25 A. The protein volume oscillates 3 to 4% around the volume of the X-ray structure. These fluctuations lead to changes in the internal free volume and in the size, shape and location of atom-sized cavity features. Transient cavities produced in the simulation have a crucial role in the movement of two of the ligands. One of the ligands escapes to the protein surface, whilst a second ligand travels through the protein interior. Complex gating processes involving several protein residues are responsible for producing the necessary pores through which the ligand passes between transient cavities or packing defects.

Solution structure of the GTPase activating domain of alpha s.

We have used heteronuclear three-dimensional NMR spectroscopy to determine the solution structure of a 141 residue protein containing the GTPase activating domain from the alpha chain of the heterotrimeric G protein Gs. The domain contains six alpha-helices and is stable and structured in solution despite having been excised from the intact Gs protein. The N-terminal ten and C-terminal 11 residues of the protein are unstructured in solution while the core is well determined by the 2483 distance and torsion restraints derived from the NMR spectra. The final ensemble of 14 structures, generated with a hybrid distance geometry/simulated annealing protocol, have an average to-the-mean backbone root-mean-square deviation of 0.39 A for the core residues 89 to 201. The majority of the structure is remarkably similar to that observed for the cognate domains in crystal structures of the homologous proteins alpha t and alpha i1. However, the orientations of the second helix and the subsequent interhelical loops differ markedly among the three proteins. This structural divergence, along with functional studies of chimeric proteins, suggests that this region of the domain interacts with either the downstream effector adenylyl cyclase or with some other intermediary protein.

Studies of synthetic helical peptides using circular dichroism and nuclear magnetic resonance.

We have designed a set of 17-residue synthetic peptides to be monomeric helices in aqueous solution. Circular dichrosim experiments indicate the presence of helical structure in aqueous solution at low temperature and low pH. The two-dimensional nuclear magnetic resonance results for one of the peptides show a segment of ten residues which clearly meets all of the criteria for the existence of helical structure at both 5 degrees C and 15 degrees C. The first four residues of the peptide are in a largely extended conformation. Calculations suggest that residues 5 through 14 are significantly helical at 5 degrees C. When the temperature is increased, circular dichroism spectra indicate that the helical content decreases. At 15 degrees C, the 3JN alpha coupling constants increase in the helical region, indicating an increase in motion or conformational averaging in the helical segment. None of the peptides has pH titration behavior consistent with salt bridge stabilization of helical conformation. Our data lend themselves to interpretation with the helix dipole model and specific side-chain interactions. When the N and C termini charges are removed the helical content of the peptides increases. The amount of helicity increases as the pH is lowered, due to the ionization of His16. Much of the helical stabilization appears to be due to a specific side-chain interaction between His16 and Tyr12.

Computational studies of the interaction of myoglobin and xenon.

Computational studies are used to investigate the energies of xenon binding to myoglobin and to describe pathways through the protein interior for a metmyoglobin-xenon complex. Empirical energy calculations indicate a favorable enthalpic contribution of 0.6 to 4.2 kcal/mol to xenon binding for four experimentally determined xenon sites. These calculated enthalpies help to explain the different xenon occupancies observed experimentally. A fifth site, modeled in place of the iron co-ordinated water molecule in the distal cavity, is also predicted to bind xenon. The largest contribution to the binding energy is from van der Waals' interactions with smaller contributions from polarization and protein strain terms. Ligand trajectory calculations as well as a new geometric algorithm define a connecting network of channel-like pathways through the static protein structure. One or two pathways appear to lead most easily from each major internal cavity to the protein surface. The importance of these channels in protein dynamics and their implications as routes for ligand motion are discussed.

Predicting the structure of protein complexes: a step in the right direction.

In a blind test of protein-docking algorithms, six groups used different methods to predict the structure of a protein complex. All six predicted structures were close enough to the experimental complex to be useful; nevertheless, several important details of the experimental complex were missed or only partially predicted.

Structure-based design and combinatorial chemistry yield low nanomolar inhibitors of cathepsin D.

BACKGROUND: The identification of potent small molecule ligands to receptors and enzymes is one of the major goals of chemical and biological research. Two powerful new tools that can be used in these efforts are combinatorial chemistry and structure-based design. Here we address how to join these methods in a design protocol that produces libraries of compounds that are directed against specific macromolecular targets. The aspartyl class of proteases, which is involved in numerous biological processes, was chosen to demonstrate this effective procedure. RESULTS: Using cathepsin D, a prototypical aspartyl protease, a number of low nanomolar inhibitors were rapidly identified. Although cathepsin D is implicated in a number of therapeutically relevant processes, potent nonpeptide inhibitors have not been reported previously. The libraries, synthesized on solid support, displayed nonpeptide functionality about the (hydroxyethyl)amine isostere. The (hydroxyethyl)amine isostere, which targets the aspartyl protease class, is a stable mimetic of the tetrahedral intermediate of amide hydrolysis. Structure-based design, using the crystal structure of cathepsin D complexed with the peptide-based natural product pepstatin, was used to select the building blocks for the library synthesis. The library yielded a 'hit rate' of 6-7% at 1 microM inhibitor concentrations, with the most potent compound having a Ki value of 73 nM. More potent, nonpeptide inhibitors (Ki = 9-15 nM) of cathepsin D were rapidly identified by synthesizing and screening a small second generation library. CONCLUSIONS: The success of these studies clearly demonstrates the power of coupling the complementary methods of combinatorial chemistry and structure-based design. We anticipate that the general approaches described here will be successful for other members of the aspartyl protease class and for many other enzyme classes.

Modeled structure of the 75-kDa neurotrophin receptor.

Motifs in ligand-binding domains of the neurotrophin (NTR) and lymphotoxin (TNFR-I) receptors define a family of receptors that mediates programmed cell death. We have explored relationships of architecture and function in this family through a molecular model of NTR, also called p75NGFR or LANR. Modeling by homology took advantage of four modular subdomains in the crystal structure of TNFR-I that also occur in NTR. Hypothetical complexes between the model and a ligand structure (for nerve growth factor, NGF) were then examined using docking software. NTR appears to bind in the dimer interface of NGF, making two sets of contacts. NTR subdomains III and IV provide the ligand-contact surfaces, in contrast to TNFR, in which subdomains II and III contact TNF-beta. NTR subdomain II appears to have been evolutionarily modified, potentially contributing to an interface between receptor subunits. These and other specific predictions of the model will require experimental confirmation.

Macromolecular docking of a three-body system: the recognition of human growth hormone by its receptor.

Human growth hormone (hGH) binds to its receptor (hGHr) in a three-body interaction: one molecule of the hormone and two identical monomers of the receptor form a trimer. Curiously, the hormone-receptor interactions in the trimer are not equivalent and the formation of the complex occurs in a specific kinetic order (Cunningham BC, Ultsch M, De Vos AM, Mulkerrin MG, Clauser KR, Wells JA, 1991, Science 254:821-825). In this paper, we model the recognition of hGH to the hGHr using shape complementarity of the three-dimensional structures and macromolecular docking to explore possible binding modes between the receptor and hormone. The method, reported previously (Hendrix DK, Kuntz ID, 1998, Pacific symposium on biocomputing 1998, pp 1234-1244), is based upon matching complementary-shaped strategic sites on the molecular surface. We modify the procedure to examine three-body systems. We find that the order of binding seen experimentally is also essential to our model. We explore the use of mutational data available for hGH to guide our model. In addition to docking hGH to the hGHr, we further test our methodology by successfully reproducing 16 macromolecular complexes from X-ray crystal structures, including enzyme-inhibitor, antibody-antigen, protein dimer, and protein-DNA complexes.

Comparison of solution structures of mutant bovine pancreatic trypsin inhibitor proteins using two-dimensional nuclear magnetic resonance.

Structural perturbations due to a series of mutations at the 30-51 disulfide bond of bovine pancreatic trypsin inhibitor have been explored using NMR. The mutants replaced cysteines at positions 30 and 51 by alanine at position 51 and alanine, threonine, or valine at position 30. Chemical shift changes occur in residues proximate to the site of mutation. NOE assignments were made using an automated procedure, NASIGN, which used information from the wild-type crystal structure. Intensity information was utilized by a distance geometry algorithm, VEMBED, to generate a series of structures for each protein. Statistical analyses of these structures indicated larger averaged structural perturbations than would be expected from crystallographic and other information. Constrained molecular dynamics refinement using AMBER at 900 K was useful in eliminating structural movements that were not a necessary consequence of the NMR data. In most cases, statistically significant movements are shown to be those greater than approximately 1 A. Such movements do not appear to occur between wild type and A30A51, a result confirmed by crystallography (Eigenbrot, C., Randal, M., & Kossiakoff, A.A., 1990, Protein Eng. 3, 591-598). Structural alterations in the T30A51 or V30A51 mutant proteins near the limits of detection occur in the beta-loop (residues 25-28) or C-terminal alpha-helix, respectively.

Molecular similarity based on DOCK-generated fingerprints.

An alternative method for defining molecular similarity is presented. By using the docking program DOCK and a reference panel of protein binding sites, fingerprints for a set of molecules have been generated, based on calculated interaction energies. These binding patterns allowed us to calculate matrices of similarity coefficients which subsequently were used for nearest-neighbor searches within the database. Our results indicate that the method is suitable for finding significant similarities of compounds of the same biological activity. Although the overall performance of a traditional 2D similarity method is better in the test systems investigated, our 3D approach can be regarded as complementary since it is able to detect similarities independent of the covalent structure of the compounds. Thus it should be a useful 3D database-searching tool for rational lead discovery.