The SHAPES strategy: an NMR-based approach for lead generation in drug discovery.

BACKGROUND: Recently, it has been shown that nuclear magnetic resonance (NMR) may be used to identify ligands that bind to low molecular weight protein drug targets. Recognizing the utility of NMR as a very sensitive method for detecting binding, we have focused on developing alternative approaches that are applicable to larger molecular weight drug targets and do not require isotopic labeling. RESULTS: A new method for lead generation (SHAPES) is described that uses NMR to detect binding of a limited but diverse library of small molecules to a potential drug target. The compound scaffolds are derived from shapes most commonly found in known therapeutic agents. NMR detection of low (microM-mM) affinity binding is achieved using either differential line broadening or transferred NOE (nuclear Overhauser effect) NMR techniques. CONCLUSIONS: The SHAPES method for lead generation by NMR is useful for identifying potential lead classes of drugs early in a drug design program, and is easily integrated with other discovery tools such as virtual screening, high-throughput screening and combinatorial chemistry.     

In vivo incorporation of unnatural amino acids to probe structure, dynamics, and ligand binding in a large protein by nuclear magnetic resonance spectroscopy

In vivo incorporation of isotopically labeled unnatural amino acids into large proteins drastically reduces the complexity of nuclear magnetic resonance (NMR) spectra. Incorporation is accomplished by coexpressing an orthogonal tRNA/aminoacyl-tRNA synthetase pair specific for the unnatural amino acid added to the media and the protein of interest with a TAG amber codon at the desired incorporation site. To demonstrate the utility of this approach for NMR studies, 2-amino-3-(4-(trifluoromethoxy)phenyl)propanoic acid (OCF 3Phe), (13)C/(15)N-labeled p-methoxyphenylalanine (OMePhe), and (15)N-labeled o-nitrobenzyl-tyrosine (oNBTyr) were incorporated individually into 11 positions around the active site of the 33 kDa thioesterase domain of human fatty acid synthase (FAS-TE). In the process, a novel tRNA synthetase was evolved for OCF 3Phe. Incorporation efficiencies and FAS-TE yields were improved by including an inducible copy of the respective aminoacyl-tRNA synthetase gene on each incorporation plasmid. Using only between 8 and 25 mg of unnatural amino acid, typically 2 mg of FAS-TE, sufficient for one 0.1 mM NMR sample, were produced from 50 mL of Escherichia coli culture grown in rich media. Singly labeled protein samples were then used to study the binding of a tool compound. Chemical shift changes in (1)H-(15)N HSQC, (1)H-(13)C HSQC, and (19)F NMR spectra of the different single site mutants consistently identified the binding site and the effect of ligand binding on conformational exchange of some of the residues. OMePhe or OCF 3Phe mutants of an active site tyrosine inhibited binding; incorporating (15)N-Tyr at this site through UV-cleavage of the nitrobenzyl-photocage from oNBTyr re-established binding. These data suggest not only robust methods for using unnatural amino acids to study large proteins by NMR but also establish a new avenue for the site-specific labeling of proteins at individual residues without altering the protein sequence, a feat that can currently not be accomplished with any other method.     

Affinity Polymers Tailored for the Protein A Binding Site of Immunoglobulin G Proteins

Rational design in combination with a screening process was used to develop affinity polymers for a specific binding site on the surface of immunoglobulin G (IgG) proteins. The concept starts with the identification of critical amino acid residues on the protein interface and their topological arrangement. Appropriate binding monomers were subsequently synthesized. Together with a sugar monomer (2–5 equiv) for water solubility and a dansyl monomer (0.5 equiv) as a fluorescent label, they were subjected in aqueous solution to linear radical copolymerization in various compositions (e.g., azobisisobutyronitrile (AIBN), homogeneous water/DMF mixtures). After ultrafiltration and lyophilization, colorless dry water‐soluble powders were obtained. NMR spectroscopic and gel permeation chromatography (GPC) characterization indicated molecular weights between 30 and 500 kD and confirmed retention of monomer composition as well as the absence of monomers. In a competitive enzyme‐linked immunosorbent assay (ELISA) screen of the polymer libraries (20–50 members), few copolymers qualified as strong and selective binders for the protein A binding site on the Fc fragment of the antibody. Their monomer composition precisely reflected the critical amino acids found at the interface. The simple combination of a charged and a nonpolar binding monomer sufficed for selective submicromolar IgG recognition by the synthetic polymer. Affinities were confirmed by fluorescence titrations; they increased with decreasing salt load but remained largely unaltered at lowered pH. Other proteins, including those of similar size and isoelectric point (pI), were bound 10–1000 times less tightly. This example indicates that interaction domains in other proteins may also be targeted by synthetic polymers if their comonomer composition reflects the nature and arrangement of amino acid residues on the protein surface.     

SOS-NMR: a saturation transfer NMR-based method for determining the structures of protein-ligand complexes

An NMR-based alternative to traditional X-ray crystallography and NMR methods for structure-based drug design is described that enables the structure determination of ligands complexed to virtually any biomolecular target regardless of size, composition, or oligomeric state. The method utilizes saturation transfer difference (STD) NMR spectroscopy performed on a ligand complexed to a series of target samples that have been deuterated everywhere except for specific amino acid types. In this way, the amino acid composition of the ligand-binding site can be defined, and, given the three-dimensional structure of the protein target, the three-dimensional structure of the protein-ligand complex can be determined. Unlike earlier NMR methods for solving the structures of protein-ligand complexes, no protein resonance assignments are necessary. Thus, the approach has broad potential applications--especially in cases where X-ray crystallography and traditional NMR methods have failed to produce structural data. The method is called SOS-NMR for structural information using Overhauser effects and selective labeling and is validated on two protein-ligand complexes: FKBP complexed to 2-(3'-pyridyl)-benzimidazole and MurA complexed to uridine diphosphate N-acetylglucosamine.     

Site-selective screening by NMR spectroscopy with labeled amino acid pairs

A new method for site-selective screening by NMR is presented. The core of the new method is the dual amino acid sequence specific labeling technique. Amino acid X is labeled with (13)C and amino acid Y is labeled with (15)N. Provided only one XY pair occurs in the amino acid sequence, only one signal in the 1D carbonyl (13)C spectrum will display a splitting due to the (1)J(C'N) coupling. Using this labeling strategy it is possible to screen selectively for binding to a selected epitope without the need for sequence specific assignments. An HNCO spectrum (1D or 2D) can be used either directly as a screening experiment or indirectly to identify what signals to monitor in a 2D (1)H-(15)N correlation spectrum. Chemical shift perturbations upon addition of a potential ligand are easily detected even for large proteins due to the reduced spectral complexity resulting from the use of a selectively labeled sample. The new technique is demonstrated on the human adipocyte fatty acid binding protein FABP-4. Due to the reduced spectral complexity, the method should be applicable to larger proteins than are conventional methods.     

Simplification of protein NOESY spectra using bioorganic precursor synthesis and NMR spectral editing

A novel method is proposed for the analysis of protein NOEs in solution. In this approach, chemically synthesized precursor compounds for the amino acids valine, leucine, and isoleucine are used for amino acid specific labeling of these hydrophobic residues. The methodology is based on a novel synthetic route to 12C,1H,2H Val, Leu, and Ile side chains selectively labeled with 13CH3 only at the terminal methyl group. In an otherwise 12C,1H labeled protein, discrimination between protons bound to 12C and 13C (or 15N) can be achieved using standard isotope-editing NMR pulse schemes. This strategy significantly relieves problems with spectral overlap through selective observation of interresidue methyl NOEs and will thus be a powerful extension of existing biomolecular NMR methodology.     

Recommended criteria for the mass spectrometric identification of target peptides and proteins (<8 kDa) in sports drug testing

Mass spectrometry has become an invaluable tool for the identification of prohibited peptide hormones and proteins in doping control analysis. Regulatory authorities have established criteria for identifying banned drugs in doping control specimens, but these criteria do not address the specific issues for high molecular weight protein drugs such as molecular weight determination of multiply charged molecules, analysis of chemically or enzymatically derived degradation products, identification of amino acid sequence tags, etc. Technical considerations such as sample preparation methods (e.g. immunoaffinity purification), resulting analytes (e.g. intact compounds vs. chemically or enzymatically derived peptides), ionization modes, analyzer resolution, and the information provided by respective techniques are discussed in light of sports drug testing requirements using typical application examples.     

Cover Picture

The cover picture shows an array of several hundred synthetically produced variants of the 44 amino acids comprising the hyAP-WW protein domain. The array was produced by a stepwise SPOT synthesis on a cellulose membrane. At each synthesis site (spot) a WW domain is bound to the membrane through a C-terminal ester bond. The secondary structure of the WW domain is shown in green as a ribbon. The domains of the single spots differ only by a single amino acid. All the domains were tested simultaneously for their ability to bind to a peptide motif (red) common to many proteins. The binding was evident when visualized by chemoluminescence, with the Spots having various intensities. The systematic analysis undertaken here enabled molecular biology to be carried out that would have otherwise have required great effort, or not been done at all. This chemical technique also allows the construction of many non-genetically coded building blocks. Through the use of modern synthetic techniques for the coupling of peptides this method should also allow the synthesis of proteins. The combination of molecular biological and chemical methods opens up opportunities for the preparation of protein chips for diagnostics and drug discovery. More about this can be found in the article by Schneider-Mergener et al. on p. 897ff.     

Semi-Synthesis and Analysis of Chemically Modified Zif268 Zinc-Finger Domains

Total synthesis of proteins can be challenging despite assembling techniques, such as native chemical ligation (NCL) and expressed protein ligation (EPL). Especially, the combination of recombinant protein expression and chemically addressable solid-phase peptide synthesis (SPPS) is well suited for the redesign of native protein structures. Incorporation of analytical probes and artificial amino acids into full-length natural protein domains, such as the sequence-specific DNA binding zinc-finger motifs, are of interest combining selective DNA recognition and artificial function. The semi-synthesis of the natural 90 amino acid long sequence of the zinc-finger domain of Zif268 is described including various chemically modified constructs. Our approach offers the possibility to exchange any amino acid within the third zinc finger. The realized modifications of the natural sequence include point mutations, attachment of a fluorophore, and the exchange of amino acids at different positions in the zinc finger by artificial amino acids to create additional metal binding sites. The individual constructs were analyzed by circular dichroism (CD) spectroscopy with respect to the integrity of the zinc-finger fold and DNA binding.     

PI by NMR: Probing CH–π Interactions in Protein–Ligand Complexes by NMR Spectroscopy

While CH–π interactions with target proteins are crucial determinants for the affinity of arguably every drug molecule, no method exists to directly measure the strength of individual CH–π interactions in drug–protein complexes. Herein, we present a fast and reliable methodology called PI (π interactions) by NMR, which can differentiate the strength of protein–ligand CH–π interactions in solution. By combining selective amino‐acid side‐chain labeling with ¹H‐¹³C NMR, we are able to identify specific protein protons of side‐chains engaged in CH–π interactions with aromatic ring systems of a ligand, based solely on ¹H chemical‐shift values of the interacting protein aromatic ring protons. The information encoded in the chemical shifts induced by such interactions serves as a proxy for the strength of each individual CH–π interaction. PI by NMR changes the paradigm by which chemists can optimize the potency of drug candidates: direct determination of individual π interactions rather than averaged measures of all interactions.     

Site‐Specific Protection and Dual Labeling of Human Epidermal Growth Factor (hEGF) for Targeting,Imaging, and Cargo Delivery

Well‐defined human epidermal growth factor (hEGF) constructs featuring selectively addressable labels are urgently needed to address outstanding questions regarding hEGF biology. A protein‐engineering approach was developed to provide access to hEGF constructs that carry two cysteine‐based site‐specific orthogonal labeling sites in multi‐milligram quantities. Also, a site‐selective (de)protection and labeling approach was devised, which allows selective modification of these hEGF constructs. The hEGF, featuring three native disulfide bonds, was expressed featuring additional sulfhydryl groups, in the form of cysteine residues, as orthogonal ligation sites at both the N and C termini. Temporary protection of the N‐terminal cysteine unit, in the form of a thiazolidine ring, avoids interference with protein folding and enables sequential labeling in conjunction with the cysteine residue at the C terminus. Based on thus‐generated hEGF constructs, sequential and site‐specific labeling with a variety of molecular probes could be demonstrated, thus leading to a biological fully functional hEGF with specifically incorporated fluorophores or protein cargo and native cellular targeting and uptake profiles. Thus, this novel strategy provides a robust entry to high‐yielding access of hEGF and rapid and easy site‐specific and multifunctional dual labeling of this growth factor.     

NMR and Molecular Genetics as Tools for Investigating Protein Interactions in Membrane Systems

In our laboratory, we have applied the tools of nuclear magnetic resonance (NMR) spectroscopy and molecular genetics to investigate the structural and dynamic properties of membrane-associated proteins and their interactions with membrane components. There are two general classes of membrane proteins, i.e., intrinsic and peripheral ones. For the intrinsic membrane proteins, we have chosen the membranebound D-lactate dehydrogenase (D-LDH) of Escherichia coli as a model to study protein-lipid interactions in membranes. D-LDH is a respiratory enzyme of molecularweight 65, 000 containing flavin adenine dinucleotide (FAD) as a cofactor. The activity of purified D-LDH is enhanced up to 100-fold by lipids and detergents. The gene for D-LDH has been sequenced, and production of the enzyme amplified up to 300-times normal levels. We have biosynthetically incorporated 5-fluorotryptophan (5F-Trp) into D-LDH and studied the five Trp residues by ¹⁹F-NMR spectroscopy. In order to gain additional information using ¹⁹F-NMR, site-specific, oligonucleotide-directed mutagenesis has been used to insert a sixth Trp into D-LDH at various positions throughout the 571-amino acid chain. These mutant D-LDHs are being characterized biochemically and through NMR. For peripheral membrane proteins, we have chosen two periplasmic binding proteins, histidine-binding protein J (J protein) of Salmonella Typhimurium and glutamine-binding protein (GlnBP) of E. coli as models to investigate the structure-function relationship in periplasmic binding protein-mediated active transport systems. These two proteins both have molecular weights of approximately 25, 000. By using mutant J proteins and GlnBPs and site-specific, oligonucleotide-directed mutagenesis techniques, we have assigned several resonances to specific amino acid residues. We are investigating the relationship between ligand-induced conformational changes in these two proteins and their roles in the active transport of ligand across the cell membrane. We have found that a combination of isotopic labeling, biochemistry, molecular biology, and NMR is a very useful approach to investigate various interactions of membrane-associated protein systems.     

Binding of low molecular weight inhibitors promotes large conformational changes in the dengue virus NS2B-NS3 protease: fold analysis by pseudocontact shifts

The two-component dengue virus NS2B-NS3 protease (DEN NS2B-NS3pro) is an established drug target, but inhibitor design is hampered by the lack of a crystal structure of the protease in its fully active form. In solution and without inhibitors, the functionally important C-terminal segment of the NS2B cofactor is dissociated from DEN NS3pro ("open state"), necessitating a large structural change to produce the "closed state" thought to underpin activity. We analyzed the fold of DEN NS2B-NS3pro in solution with and without bound inhibitor by nuclear magnetic resonance (NMR) spectroscopy. Multiple paramagnetic lanthanide tags were attached to different sites to generate pseudocontact shifts (PCS). In the face of severe spectral overlap and broadening of many signals by conformational exchange, methods for assignment of (15)N-HSQC cross-peaks included selective mutation, combinatorial isotope labeling, and comparison of experimental PCSs and PCSs back-calculated for a structural model of the closed conformation built by using the structure of the related West Nile virus (WNV) protease as a template. The PCSs show that, in the presence of a positively charged low-molecular weight inhibitor, the enzyme assumes a closed state that is very similar to the closed state previously observed for the WNV protease. Therefore, a model of the protease built on the closed conformation of the WNV protease is a better template for rational drug design than available crystal structures, at least for positively charged inhibitors. To assess the open state, we created a binding site for a Gd(3+) complex and measured paramagnetic relaxation enhancements. The results show that the specific open conformation displayed in the crystal of DEN NS2B-NS3pro is barely populated in solution. The techniques used open an avenue to the fold analysis of proteins that yield poor NMR spectra, as PCSs from multiple sites in combination with model building generate powerful information even from incompletely assigned (15)N-HSQC spectra.     

A combinatorial selective labeling method for the assignment of backbone amide NMR resonances

A combinatorial selective labeling (CSL) method is presented for the assignment of backbone amide NMR resonances, which has a particular application in the identification of protein-ligand interaction sites. The method builds on the dual amino acid selective labeling technique. In the CSL method a number of different samples are produced, each with a different pattern of labeled amino acids. By analyzing peak intensities in HSQC and 2D HNCO spectra of these samples, a large number of combinations of amino acid pairs can be simultaneously assigned. We demonstrate the method on the 27 kDa protein GFP. The samples can be produced rapidly and cost-effectively in a commercially available in vitro translation system. The method greatly simplifies the process of backbone assignment and would be very straightforward to automate.     

Binding site elucidation of hydantoin-based antagonists of LFA-1 using multidisciplinary technologies: evidence for the allosteric inhibition of a protein--protein interaction

The binding site on the lymphocyte function-associated antigen-1 (LFA-1) of a class of hydantoin-based antagonists of leukocyte cell adhesion has been identified. This site resides in the inserted-domain (I-domain) of the CD11a chain at a location that is distal to residues known to be required for interactions with the intercellular adhesion molecules. This finding supports the hypothesis that the molecules are antagonizing cell adhesion via an allosteric modification of LFA-1. The binding site was identified using an integrated immunochemical, chemical, and molecular modeling approach. Antibodies that map to epitopes on the I-domain were blocked from binding to the purified protein by the hydantoins, indicating that the hydantoin-binding site resides on the I-domain. Photoaffinity labeling of the I-domain followed by LC/MS and LC/MS/MS analysis of the enzymatic digest identified proline 281 as the primary amino acid residue covalently attached to the photoprobe. Distance constraints derived from this study coupled with known SAR considerations allowed for the construction of a molecular model of the I-domain/inhibitor complex. The atomic details of the protein/antagonist interaction were accurately predicted by this model, as subsequently confirmed by the X-ray crystal structure of the complex.     

RAMPED-UP NMR: multiplexed NMR-based screening for drug discovery

The crucial step in drug discovery is the identification of a lead compound from a vast chemical library by any number of screening techniques. NMR-based screening has the advantage of directly detecting binding of a compound to the target. The spectra resulting from these screens can also be very complex and difficult to analyze, making this an inefficient process. We present here a method, RAMPED-UP NMR, (Rapid Analysis and Multiplexing of Experimentally Discriminated Uniquely Labeled Proteins using NMR) which generates simple spectra which are easy to interpret and allows several proteins to be screened simultaneously. In this method, the proteins to be screened are uniquely labeled with one amino acid type. There are several benefits derived from this unique labeling strategy: the spectra are greatly simplified, resonances that are most likely to be affected by binding are the only ones observed, and peaks that yield little or no information upon binding are eliminated, allowing the analysis of multiple proteins easily and simultaneously. We demonstrate the ability of three different proteins to be analyzed simultaneously for binding to two different ligands. This method will have significant impact in the use of NMR spectroscopy for both the lead generation and lead optimization phases of drug discovery by its ability to increase screening throughput and the ability to examine selectivity. To the best of our knowledge, this is the first time in any format that multiple proteins can be screened in one tube.     

Ligand-induced conformational heterogeneity of cytochrome P450 CYP119 identified by 2D NMR spectroscopy with the unnatural amino acid (13)C-p-methoxyphenylalanine

Conformational dynamics are thought to play an important role in ligand binding and catalysis by cytochrome P450 enzymes, but few techniques exist to examine them in molecular detail. Using a unique isotopic labeling strategy, we have site specifically inserted a (13)C-labeled unnatural amino acid residue, (13)C-p-methoxyphenylalanine (MeOF), into two different locations in the substrate binding region of the thermophilic cytochrome P450 enzyme CYP119. Surprisingly, in both cases the resonance signal from the ligand-free protein is represented by a doublet in the (1)H,(13)C-HSQC spectrum. Upon binding of 4-phenylimidazole, the signals from the initial resonances are reduced in favor of a single new resonance, in the case of the F162MeOF mutant, or two new resonances, in the case of the F153MeOF mutant. This represents the first direct physical evidence for the ligand-dependent existence of multiple P450 conformers simultaneously in solution. This general approach may be used to further illuminate the role that conformational dynamics plays in the complex enzymatic phenomena exhibited by P450 enzymes.     

A Formylglycine-Peptide for the Site-Directed Identification of Phosphotyrosine-Mimetic Fragments**

Discovery of protein-binding fragments for precisely defined binding sites is an unmet challenge to date. Herein, formylglycine is investigated as a molecular probe for the sensitive detection of fragments binding to a spatially defined protein site . Formylglycine peptide 3 was derived from a phosphotyrosine-containing peptide substrate of protein tyrosine phosphatase PTP1B by replacing the phosphorylated amino acid with the reactive electrophile. Fragment ligation with formylglycine occurred in situ in aqueous physiological buffer. Structures and kinetics were validated by NMR spectroscopy. Screening and hit validation revealed fluorinated and non-fluorinated hit fragments being able to replace the native phosphotyrosine residue. The formylglycine probe identified low-affinity fragments with high spatial resolution as substantiated by molecular modelling. The best fragment hit, 4-amino-phenyl-acetic acid, was converted into a cellularly active, nanomolar inhibitor of the protein tyrosine phosphatase SHP2.     

Chaperone-like activity of alpha-crystallin and other small heat shock proteins

Small heat shock proteins (sHsps) are a large family of proteins with monomeric molecular weight of 12-43 kDa, present within the prokaryotic and eukariotic cell as large oligomeric complexes, ranging in size from 200-800 kDa. Unlike the high molecular weight Hsps, which are involved in protein folding in vivo, under normal conditions, sHsps play an important role in protecting organism from stress. SHsps share an evolutionarily conserved sequence of 80-100 amino acids, located in the C-terminal region, and called "alpha-crystallin domain"; its role in subunits interactions has been recently underlined by site-directed spin labeling studies and by fluorescence resonance energy transfer data. The N-terminal region, preceding the alpha-crystallin domain, is variable in length and amino acid sequence, contributing to structural diversity between different sHsps and having a role in multimerization. The alpha-Crystallin domain is followed by C-terminal extension, a polar structure, involved in protein solubility, which share no sequence homology. Expression of sHsps is induced in response to various kinds of stress including heat shock, oxidative stress, osmostress, or ischemia, but some sHsps are expressed constitutively under physiological conditions. In vitro, sHsps selectively bind and stabilize proteins and prevent their aggregation at elevated temperatures in an ATP-independent way and protect enzymes against heat-induced inactivation. Our own studies focused on the chaperone-like activity of alpha-crystallin, the major protein component of vertebrate lens, using another system than heat-induced aggregation. Our data demonstrated that alpha-crystallin specifically protects enzymes against inactivation by different posttranslational modifications such as glycation, carbamylation and aldehyde binding, and also reactivates GuHCl-denatured enzymes. Complex formation between alpha-crystallin and the denatured enzymes, was suggested as a mechanism of protection.     

The binding site for an inhibitor of squalene:hopene cyclase determined using photoaffinity labeling and molecular modeling.

BACKGROUND: The squalene:hopene cyclases (SHCs) are bacterial enzymes that convert squalene into hopanoids, a function analogous to the action of oxidosqualene cyclases (OSCs) in eukaryotic steroid and triterpenoid biosynthesis. We have identified the binding site for a selective, potent, photoactivatable inhibitor of an SHC. RESULTS: SHC from Alicyclobacillus acidocaldarius was specifically labeled by [3H]Ro48-8071, a benzophenone-containing hypocholesteremic drug. Edman degradation of a peptide fragment of covalently modified SHC confirmed that Ala44 was specifically modified. Molecular modeling, using X-ray-derived protein coordinates and a single point constraint for the inhibitor, suggested several geometries by which Ro48-8071 could occupy the active site. CONCLUSIONS: A covalent complex of a potent inhibitor with a squalene cyclase has been characterized. The amino acid modification and molecular modeling suggest that Ro48-8071 binds at the junction between the central cavity and substrate entry channel, therefore inhibiting access of the substrate to the active site.