Crystallographic structures of the ligand-binding domains of the androgen receptor and its T877A mutant complexed with the natural agonist dihydrotestosterone.

The structures of the ligand-binding domains (LBD) of the wild-type androgen receptor (AR) and the T877A mutant corresponding to that in LNCaP cells, both bound to dihydrotestosterone, have been refined at 2.0 A resolution. In contrast to the homodimer seen in the retinoid-X receptor and estrogen receptor LBD structures, the AR LBD is monomeric, possibly because of the extended C terminus of AR, which lies in a groove at the dimerization interface. Binding of the natural ligand dihydrotestosterone by the mutant LBD involves interactions with the same residues as in the wild-type receptor, with the exception of the side chain of threonine 877, which is an alanine residue in the mutant. This structural difference in the binding pocket can explain the ability of the mutant AR found in LNCaP cells (T877A) to accommodate progesterone and other ligands that the wild-type receptor cannot.

Hydrophobicity profiles for protein sequence analysis.

Hydrophobic interactions are a major force in protein folding and numerous hydropathy scales have been developed to quantify the relative hydrophobicity of the amino acids. Hydropathy profiles can be used to examine the surface features of proteins in order to generate hypotheses that can be confirmed experimentally. This unit describes the application of hydrophobicity plots to typical problems and provides suggested uses for a few selected scales.

Protein secondary structure prediction.

This unit describes procedures developed for predicting protein structure from the amino acid sequence. The first of the four sections is an overview and brief history of structure prediction schemes. The second section describes four distinct prediction schemes, with emphasis on their differences. In the third part each prediction scheme is used to evaluate three proteins that have different folding patterns. The final section is a comparison of the prediction results and suggestions for secondary structure prediction.

High-throughput and virtual screening: core lead discovery technologies move towards integration.

In addition to high-throughput screening (HTS), the main lead discovery technology employed by most pharmaceutical companies today is virtual screening (VS). Although the two techniques have somewhat different philosophical origins, they contain many synergies that can potentially enhance the lead discovery process. Here, we describe many of the latest developments in VS technology with particular emphasis on their potential impact on HTS in, for example, focussed screening and data mining. In addition, we highlight key issues that need to be addressed before the potential of such efforts can be fully realized.

Simulated annealing with restrained molecular dynamics using a flexible restraint potential: theory and evaluation with simulated NMR constraints.

A new functional representation of NMR-derived distance constraints, the flexible restraint potential, has been implemented in the program CONGEN (Bruccoleri RE, Karplus M, 1987, Biopolymers 26:137-168) for molecular structure generation. In addition, flat-bottomed restraint potentials for representing dihedral angle and vicinal scalar coupling constraints have been introduced into CONGEN. An effective simulated annealing (SA) protocol that combines both weight annealing and temperature annealing is described. Calculations have been performed using ideal simulated NMR constraints, in order to evaluate the use of restrained molecular dynamics (MD) with these target functions as implemented in CONGEN. In this benchmark study, internuclear distance, dihedral angle, and vicinal coupling constant constraints were calculated from the energy-minimized X-ray crystal structure of the 46-amino acid polypeptide crambin (ICRN). Three-dimensional structures of crambin that satisfy these simulated NMR constraints were generated using restrained MD and SA. Polypeptide structures with extended backbone and side-chain conformations were used as starting conformations. Dynamical annealing calculations using extended starting conformations and assignments of initial velocities taken randomly from a Maxwellian distribution were found to adequately sample the conformational space consistent with the constraints. These calculations also show that loosened internuclear constraints can allow molecules to overcome local minima in the search for a global minimum with respect to both the NMR-derived constraints and conformational energy. This protocol and the modified version of the CONGEN program described here are shown to be reliable and robust, and are applicable generally for protein structure determination by dynamical simulated annealing using NMR data.

Mutational analysis of the endothelin type A receptor (ETA): interactions and model of selective ETA antagonist BMS-182874 with putative ETA receptor binding cavity.

Endothelin (ET) receptor antagonism is a potential therapeutic intervention in the treatment of vascular diseases. To elucidate the mechanism of antagonist-ET receptor complex formation, the interactions of four chemically distinct antagonists were investigated using a combination of genetic and biochemical approaches. By site-specific mutagenesis we previously demonstrated that Tyr129 in the second transmembrane domain was critical for high-affinity, subtype-selective binding to the A subtype of ET (ETA) receptors [Krystek et al. (1994) J. Biol. Chem. 269, 12383-12386]. Affinities of the constrained cyclic pentapeptide BQ-123, the pyrimidinylbenzenesulfonamide bosentan, the indancarboxlic acid SB 209670, and the naphthalenesulfonamide BMS-182874 were decreased 20-1000-fold in Tyr129Ala, Tyr129Ser, and Tyr129His ETA receptor mutants. Substitution of Tyr129 with Phe or Trp did not alter the high-affinity binding of BQ-123, bosentan, or SB 209670. BMS-182874 binding affinity was decreased 10-fold in Tyr129Phe and Tyr129trp ET receptors. These data indicate a role of aromatic interactions in the binding of these antagonists to ETA receptors an, in the case of BMS-182874, also suggested a hydrogen bond with the tyrosine hydroxyl. This hypothesis was supported by structure-activity data with analogs of BMS-182874 that varied the C-5 dimethylamino substituent on the naphthalene ring. Mutation of Asp126 and Asp133 also altered binding of BMS-182874 and C-5 analogs. In all cases, naphthalenesulfonamide binding was more severely affected by mutation of Asp133 than by mutation of Asp126. Phosphoinositide hydrolysis and extracellular acidification rate studies demonstrated the importance of Tyr129 to ETA-mediated signal transduction. On the basis of these data, two plausible models of the docked conformation of BMS-182874 in the ETA receptor are proposed as a starting point for further delineation of interactions that underlie antagonist-ETA receptor complex formation.

Aspartate mutation distinguishes ETA but not ETB receptor subtype-selective ligand binding while abolishing phospholipase C activation in both receptors.

The endothelin receptors, ETA and ETB, are G protein-coupled receptors (GPCR) that show distinctively different binding profiles for the endothelin peptides and other ligands. We recently reported that Tyr129 in the second transmembrane region (TM2) of the ETA receptor was critical for subtype-specific ligand binding [Krystek, S.R. et al. (1994) J. Biol. Chem. 269, 12383-12386]. Receptor models indicated that aspartic acids located one helical turn above (Asp133) and below (Asp126) Tyr129 in ETA had their side chains directed toward the putative binding cavity. Similarly in ETB, Asp147 and Asp154 are located one turn below and above His150, the residue that corresponds to Tyr129. Asp126 in ETA and Asp147 in ETB correspond to the highly conserved aspartate present in TM2 of many GPCR that has frequently been shown to be crucial for agonist efficacy. Mutagenesis of Asp126 of the human ETA receptor to alanine resulted in an unaltered affinity for ET-1, a 160-fold increase in ET-3 affinity and a decrease in affinity for the ETA selective naphthalenesulfonamide, BMS-182874. ET-1 activation of phospholipase C was abolished. In addition, despite the gain in binding affinity, ET-3 failed to activate phospholipase C, suggesting that Asp126 is required for signal transduction. Mutagenesis of Asp133 to alanine indicated that it was critical only for the binding of BMS-182874. In the ETB receptor, mutation of His150 to alanine or tyrosine indicated that it plays a minor role in ETB subtype-selective ligand binding; mutation of the aspartates in TM2 of ETB did not alter ligand binding. As in the Asp126 Ala ETA variant, ET-1 and ET-3 failed to increase intracellular levels of inositol phosphates in the Asp147Ala ETB mutant. Taken together, these data support the hypothesis that Asp126 and Asp133 flanking Tyr129 in TM2 of the ETA receptor play a role in defining ETA subtype-selective ligand binding but Asp147 and Asp154 that flank the His150 in TM2 of the ETB receptor do not. Furthermore, these data indicate that Asp126 in ETA and Asp147 in ETB are important for transmembrane signaling via phospholipase C.

Three-dimensional quantitative structure-activity relationships of sulfonamide endothelin inhibitors.

A three-dimensional quantitative structure-activity relationship (QSAR) using steric and electrostatic fields (comparative molecular field analysis, CoMFA) applied to 36 aryl sulfonamides assayed for endothelin receptor subtype-A (ETA) antagonism provided high cross-validation correlations (0.7) and showed promising predictive ability. The results were validated through trials using scrambled activities as well as trials using scrambled orientation of molecules. CoMFA was used to discriminate between alternate hypothetical biologically active conformations. CoMFA was also used to discriminate between two different molecular superpositions representing possible positioning within the receptor binding site. The preferred superposition supports hypotheses that suggest Tyr129 in the ETA receptor as a key residue for antagonist binding. Significant CoMFA results were obtained when crudely optimized geometries and simple charge schemes were used. The results improved on refinement, most substantially with refinement of the atomic charges.

A comparison of X-ray and NMR structures for human endothelin-1.

Direct comparisons between the recently solved X-ray and NMR structures of human endothelin-1 with respect to secondary structure, RMS deviations, surface accessibilities, and side-chain conformers indicate important differences in conformation, especially in the C-terminus, but also in the central loop region, that are important for defining the specificity of binding. These differences are larger than seen for other X-ray and NMR structures that have been compared. Comparisons between the X-ray structure and the NMR NOE constraints highlight the regions of flexibility and environment-induced diversity in the endothelin structures.

Mutation of peptide binding site in transmembrane region of a G protein-coupled receptor accounts for endothelin receptor subtype selectivity.

The molecular basis for endothelin (ET) isopeptide selectivity between ETA and ETB receptors was studied by examining ligand binding to several site-specific mutants of the human ETA receptor. Based on a computer-built three-dimensional model of the ETA receptor, five non-conserved amino acids, clustered around the putative ligand binding site, were targeted for mutation to alanine. Expression of the wild-type and mutant ETA receptors in COS-7 cells revealed that the binding profile of one of the ETA mutants, Tyr129-->Ala, was characteristic of the ETB receptor. In the Tyr129-->Ala ETA receptor mutant the affinity of two ETB-selective agonists, endothelin-3 and sarafotoxin S6c, was increased 10-200-fold, whereas that for two ETA-selective antagonists, BQ-123 and BMS-182874, was reduced 350-2,000-fold. Thus, mutation of a single amino acid in the second transmembrane region of the wild-type ETA receptor results in subtype conversion. In addition, these data represent the first example of peptide interactions with a transmembrane region of a G protein-coupled receptor and indicate that Tyr129, located in the second transmembrane region of the ETA receptor, is a critical component for determination of endothelin receptor subtype-selective ligand binding.

Identification of a subunit contact site of the alpha-subunit of follitropin.

Using synthetic peptides, we previously identified portions of the lutropin alpha-subunit that are in contact with the beta-subunit. In order to elucidate structure/function differences of the glycoprotein hormones, a similar study was conducted for follitropin. Peptides corresponding to the follitropin alpha-subunit were synthesized using standard solid-phase procedures. Purified peptides were incubated in the presence of alpha- and beta-subunits of follitropin, and subunit recombination was monitored using gel permeation chromatography and reverse-phase high-pressure liquid chromatography. Peptide alpha 33-58, corresponding to a highly conserved portion of alpha-subunit, completely inhibited subunit recombination for 24 h, and allowed only partial (61%-71%) recombination after 48 h. Peptide alpha 51-65 or alpha 61-78 inhibited subunit recombination partially at 24 h, but almost full (greater than 80%) recombination was observed by 48 h. Peptides corresponding to the rest of the alpha-subunit, alpha 1-15, alpha 11-27, alpha 22-39, and alpha 73-92, did not inhibit recombination of the alpha- and beta-subunits. The data suggest that alpha-subunits have similar residues in contact with regions of the beta-subunits of both lutropin and follitropin, specifically involving residues from continuous regions of the alpha-subunit (residues 45-75). The data suggest that this region contains multiple sites of contact with the beta-subunit.

Solution conformation of a cyclic pentapeptide endothelin antagonist. Comparison of structures obtained from constrained dynamics and conformational search.

The structure of a cyclic pentapeptide, cyclo-(D-Trp-D-Asp-L-Pro-D-Val-L-Leu), that has high selectively for the endothelin ETA receptor has been determined by NMR spectroscopy using constrained molecular dynamics and conformational search procedures. Structures obtained using two methods of refinement, namely (i) constrained molecular dynamics; and (ii) systematic searches of conformational space for optimal satisfaction of distance constraints, were compared to those obtained from systematic searches of conformational space without NMR data. The two different procedures of refinement produce similar conformations that are consistent with the NMR distance constraints. Conformational searches for optimal energy without any NMR distance constraints produced several low-energy structures, two of which have essentially the same backbone as those structures derived from distance-constrained procedures and one of these even reproduces several side-chain positions well. The pentapeptide backbone consists of a linked gamma- and beta-turn conformation, with the leucine and tryptophan as corner residues of the type II beta-turn. The side chains are highly ordered both in aqueous solvent and in dimethyl sulfoxide. In aqueous media the leucine side chain is directed towards the indole ring, presumably to reduce the non-polar surface exposure, producing unusual upfield shifts for the methyls (and particularly H gamma). This structural feature was reproduced in one of the structures obtained from conformational searches performed without NMR data. Exhaustive conformational searches appear to provide an alternative method for structure generation for cyclic peptides.

Conformational isomerism of endothelin in acidic aqueous media: a quantitative NOESY analysis.

The conformational features of endothelin-1 (ET-1) in mixed water/ethylene glycol media have been studied by two-dimensional 1H NMR experiments throughout the pH range 3.2-7.2. At pH less than 5 all backbone NH signals can be observed, and NOESY experiments provided a large set of dipolar cross-peaks. Cross-peak intensities from each experiment (different mixing times and H2O versus D2O) were converted to distance constraints using a novel algorithm (program DISCON) for removing spin diffusion effects and thus obtain cross-rates rather than cross-peak intensities. A set of 168 nonstereospecific distance bounds (average experimental precision, +/- 0.38 A) was used in dynamics simulated annealing refinements. Two consensus structural features were found--a reverse turn at Ser5----Asp8 and an alpha-helical stretch from Lys9 to Cys15; however, after constraint-free minimization, structures generated using XPLOR-1.5, CONGEN, and DISCOVER all violated at least 32% of the bounds by more than 0.2 A, which we ascribe to conformational isomerism. When the constraints were modified to reflect subsequent experimental data and to eliminate constraints that could not be obeyed by any single conformer structure, the relaxed structures still violated at least 15% of this more limited and looser set of constraints. Therefore, a modified procedure for constrained dynamics refinement (using XPLOR-2.1), which allows for conformational isomerism outside of the central helical core region, was developed. This "conformer search procedure" produced structures which fell into five tightly defined conformational clusters. The two most populated clusters correspond to a rotation of the 8,9-amide unit. The conformer which we propose as the major contributor at pH 3.2-5.8 was defined to a backbone rmsd of 0.51 A over residues 1----15. An alternative description of the motional averaging in segments of the endothelin structure as extensive randomization rather than rapid interconversion between a small number of discreet conformers was ruled out by an analysis of NH shift-temperature gradients and exchange rates. This analysis suggests that small delta delta/delta T values need not correlate with H-bonding for conformational mixtures. In ET-1 the greatest motional averaging occurs from Ser2 through Ser5 (not in the C-terminus) and may be so extensive as to approximate a flexible random coil population as high as 30%. The C-terminus shows less rapid and less extensive conformational averaging, but no definitive structures for individual conformers could be derived in the absence of stereospecific constraints. The pharmacological implications of the consensus structural features are discussed.

Stabilities of leucine zipper dimers estimated by an empirical free energy method.

The leucine zipper motif is a characteristic amino acid sequence found in dimeric DNA-binding proteins. Computer-generated models for leucine zippers were constructed as alpha-helical coiled dimers with leucine repeated every seventh residue. An empirical Gibbs free energy, delta G, function which incorporates hydrophobic force, electrostatic interactions, and conformational entropy loss as the major intermolecular interactions was used to estimate the delta G of dimer formation in fos, jun, and GCN4 zipper sequences. The calculations showed that complexes known to form stable homo- or heterodimers have favorable (negative) delta G, while other less stable complexes have unfavorable (positive) delta G. Leucines in position d of the coiled coil contribute large hydrophobic stabilization energies while residues in the a position contribute less to dimer stability. Hydrophobic contributions show little sequence specificity, however, and do not contribute significantly to homo/heterodimer preference. Charged residues in the e and g positions, on the other hand, determine homo/heterodimer specificity. In GCN4 homodimers, residues GLU el, Glu b2, Lys g2, and Lys e4 greatly contribute to dimer stability. The preferential stability of fos-jun heterodimer over the jun-jun and fos-fos homodimers is primarily due to the side chains Asp b1, Glu g1, Asp b2, Glu e2, Glu g2, Glu g3, and Lys a5 of the fos helix, and Arg c1, Lys g1, Lys b2, Lys e2, Arg e4, and Glu g4 of the jun helix.

Conformation of endothelin in aqueous ethylene glycol determined by 1H-NMR and molecular dynamics simulations.

The solution conformation of a 21-residue vasoconstrictor peptide endothelin-1 (ET-1) in water-ethylene glycol has been determined by two-dimensional 1H-NMR spectroscopy and constrained molecular dynamics simulations. The N-terminus (residues 1-4) appears to undergo conformational averaging and no single structure consistent with the NMR constraints could be found for this region. Residues 5-8 form a turn, and residues 9-16 exist in a helical conformation. A flexible 'hinge' between residues 8-9 allows various orientations of the turn relative to the helix. Another 'hinge' at residue 17 connects the extended C-terminus to the bicyclic core region (residues 1-15). Residues important for binding and biological activity form a contiguous surface on one side of the helix, with the two disulfides extending from the other side of the helix.

Identification of subunit contact sites on the alpha-subunit of lutropin.

Peptides corresponding to the entire sequence of the alpha-subunit of the human glycoprotein hormones were synthesized by using standard solid-phase procedures. Purified peptides were incubated in the presence of alpha- and beta-subunits of bovine lutropin, and subunit recombination was monitored by difference spectroscopy, reverse-phase high-pressure liquid chromatography, and gel filtration chromatography. Although the binding of alpha-peptides to either subunit could not be detected by these techniques, it was possible to demonstrate that some peptides could inhibit the recombination of alpha- and beta-subunits. Specifically, alpha-peptide 33-58 allowed only 0-11% of subunit recombination in 24 h (38-56% after 48 h), while alpha-peptide 51-65 allowed 10-60% of subunits to recombine in 24 h (65-94% in 48 h). Peptides 1-15, 11-27, 22-39, 61-78, and 73-92 of the alpha-subunit could not inhibit subunit recombination at any time or at any concentration tested. The data suggest that at least a portion of the alpha-subunit contact site has been identified, and results are discussed in terms of protein structure assessment tools.

Endothelin-like pulmonary vasoconstrictor peptide release by alpha-thrombin.

The endothelial cells lining the vessel wall can modulate vasomotor tone by releasing vasoactive factors, such as endothelial-derived constricting factors. We observed that alpha-thrombin, but not catalytically inactivated alpha-thrombin, mediated the release of two pulmonary vasoconstrictor peptides into the venous effluent of guinea pig lungs. These peptides elicited a slow-onset, long-lasting pulmonary vasoconstriction similar to the effect of endothelin, an endothelial-derived 21-amino acid vasoconstrictor peptide previously isolated from cells in culture. One of the isolated peptides coelutes with endothelin upon reverse-phase HPLC with an acetonitrile gradient and has a molecular weight comparable to endothelin as determined by gel-permeation HPLC. The other vasoconstrictor peptide elutes earlier than endothelin on reverse-phase HPLC and exhibits a lower molecular weight. The studies show the release of endothelin-like pulmonary vasoconstrictor peptides in the intact lung by alpha-thrombin, a central regulatory enzyme in hemostasis.

A proposed structural model of endothelin.

We have assessed the hydropathic nature, secondary structure, and flexibility of endothelin and report a testable model for the structure of this 21-amino acid vasoconstrictor peptide. The proposed model consists of a rigid (disulfide bonded), hydrophilic amino-terminal half of the molecule with two turns, and an extended hydrophobic sheet structure comprising the C-terminal half of the molecule. The extended structure may be stabilized by intermolecular hydrogen bonding, leading to dimers or higher order aggregates, and is hydrophobic enough to partition into an organic solvent such as ether, as long as the molecule can keep its hydrophilic amino terminus in an aqueous phase. To test this model, partitioning measurements were made on bioactive synthetic endothelin. It was found to partition at the interface of an ether/water system in a concentration- and surface area-dependent manner. A further indication as to the amphipathic nature of the molecule is that microcrystals of synthetic endothelin tend to grow at an organic-aqueous interface in a two-phase system.

Cleavage of a model peptide at its glycine residue by alkaline mercuric oxycyanide.

Treatment of peptides with excess HgO in the presence of alkaline cyanide leads to cleavage of the peptides at glycine residues. The reaction appears to involve both C- and N-mercuration with subsequent release of 2 mol mercury per mol of glycine. An intermediate glyoxylic acid residue in Schiff base linkage is postulated. Treatment of the heptapeptide Phe-Ala-Lys-Gly-Leu-Asp-Val with alkaline HgO and KCN for 6 h at 25 degrees resulted in greater than 90% cleavage, and the resultant reaction products were separated by reverse phase chromatography and identified by amino acid analysis. N-terminal products were approximately equimolar Phe-Ala-Lys, Phe-Ala-Lys-Gly, and Phe-Ala-Lys-amide. C-terminal products were predominantly Leu-Asp-Val (63%), plus Gly-Leu-Asp-Val (9%), and oxalyl-Leu-Asp-Val (8%). This method may be useful for cleavage of peptides or proteins containing glycine residues.

Inhibition of iodine-125-labeled human follitropin binding to testicular receptor by epidermal growth factor and synthetic peptides.

Two tetrapeptide sequence homologies between mouse epidermal growth factor precursor (mEGFP) and human follitropin (FSH) were revealed by a computer program that identifies identical residues among polypeptide sequences. The two tetrapeptides, Lys-Thr-Cys-Thr (KTCT) and Thr-Arg-Asp-Leu (TRDL), are present in the hormone-specific beta subunit of FSH from all species studied. These tetrapeptides are not present in the alpha subunit, which is common to all pituitary glycoprotein hormones. Both tetrapeptides are also found in mEGFP, and one tetrapeptide, TRDL, is located within the 53-residue form of mEGF purified from mouse submaxillary glands. Computer-generated hydropathy profiles predicted that both tetrapeptides are located in hydrophilic portions of the FSH beta subunit and that TRDL is in a hydrophilic portion of commercially available mEGF. Therefore, the tetrapeptides might be accessible to receptor binding sites for FSH. We report that mEGF inhibits binding of 125I-labeled human FSH to receptors in testis by 50% (I50) at a concentration of 1.8 X 10(-5) M. No binding inhibition was observed by GnRH or arginine-vasopressin at 10(-4) M, neither of which contain the tetrapeptide sequences. FSH beta subunit, which contains both tetrapeptides, also inhibited binding (I50 = 9 X 10(-8) M) of 125I-labeled human FSH to testis receptor. Thus, it appears that FSH beta subunit and mEGF are capable of inhibiting binding of FSH to testicular FSH receptors, presumably through interactions that include the homologous tetrapeptides. This presumption was supported by the observation that the synthetic tetrapeptides (KTCT or TRDL) were also active in inhibiting binding of 125I-labeled human FSH to testis receptor.