Scaffold hopping via ANCHOR.QUERY: ß-lactams as potent p53-MDM2 antagonists†.

Using the pharmacophore-based virtual screening platform ANCHOR.QUERY, we morphed our recently described Ugi-4CR scaffold towards a ß-lactam scaffold with potent p53-MDM2 antagonizing activities. 2D-HSQC and FP measurements confirm potent MDM2 binding. Molecular modeling studies are used to understand the observed SAR in the ß-lactam series.

Solution structure of PCP, a prototype for the peptidyl carrier domains of modular peptide synthetases.

BACKGROUND: Nonribosomal peptide synthetases (NRPSs) are large modular enzymes responsible for the synthesis of a variety of microbial bioactive peptides. They consist of modules that each recognise and incorporate one specific amino acid into the peptide product. A module comprises several domains, which carry out the individual reaction steps. After activation by the adenylation domain, the amino acid substrate is covalently tethered to a 4'-phosphopantetheinyl cofactor of a peptidyl carrier domain (PCP) that passes the substrate to the reaction centres of the consecutive domains. RESULTS: The solution structure of PCP, a distinct peptidyl carrier protein derived from the equivalent domain of an NRPS, was solved using NMR techniques. PCP is a distorted four-helix bundle with an extended loop between the first two helices. Its overall fold resembles the topology of acyl carrier proteins (ACPs) from Escherichia coli fatty acid synthase and actinorhodin polyketide synthase from Streptomyces coelicolor; however, the surface polarity and the length and relative alignment of the helices are different. The conserved serine, which is the cofactor-binding site, has the same location as in the ACPs and is situated within a stretch of seven flexible residues. CONCLUSIONS: The structure of PCP reflects its character as a protein domain. The fold is well defined between residues 8 and 82 and the structural core of the PCP domain can now be defined as a region spanning 37 amino acids in both directions from the conserved serine. The flexibility of the post-translationally modified site might have implications for interactions with the cooperating proteins or NRPS domains.

Structure and multiple conformations of the kunitz-type domain from human type VI collagen alpha3(VI) chain in solution.

BACKGROUND: The Kunitz-type inhibitor motif is found at the C terminus of the human collagen alpha3(VI) chain. This 76-residue module (domain C5) was prepared in recombinant form and showed high stability against proteases; however, it lacked any inhibitory activity against trypsin, thrombin, kallikrein and several other proteases. We have undertaken the determination of the three-dimensional (3D) structure of domain C5 in solution, by nuclear magnetic resonance (NMR), in order to establish the structural basis for the properties of this protein. RESULTS: The 7 N-terminal and 12 C-terminal residues of domain C5 are disordered in the solution structure. The 55-residue core, which shows high homology to bovine pancreatic trypsin inhibitor, retains the characteristic fold of all members of the Kunitz-type inhibitor family. 24 residues of this main structural body show more than one resonance, symptomatic of multiple conformations slowly exchanging on the NMR time scale. In addition, significant proton chemical exchange line broadening is observed for residues in the vicinity of the disulfide bridge between residues 20 and 44: this indicates interconversion, on the micro- to millisecond time scale, between multiple conformations. CONCLUSION: The NMR study demonstrates that domain C5 is a highly dynamic molecule at temperatures studied (between 10 and 30 degrees C). Indeed, some 44% of the main body structure of C5 showed multiple conformations. The existence of multiple conformations was not necessarily expected in view of the conformational constraints imposed by the 3D structure of proteins as rigid as C5; it should therefore be considered in the interpretation of its structural and dynamical properties. The accessibility of the inhibitory binding loop (Gly18 [P4] to Leu25 [P4']) should be relatively unaffected by this conformational exchange and thus would not explain the unusual specificity of C5. Most serine proteinase inhibitors that, like C5, have an arginine at the P1 position inhibit trypsin; the lack of trypsin inhibition of C5 must therefore arise from the amino-acid side-chain composition of the adjoining positions in the binding loop.

Structure of human cyclin-dependent kinase inhibitor p19INK4d: comparison to known ankyrin-repeat-containing structures and implications for the dysfunction of tumor suppressor p16INK4a.

BACKGROUND: The four members of the INK4 gene family (p16(INK4a), p15(INK4b), p18(INK4c) and p19(INK4d)) inhibit the closely related cyclin-dependent kinases CDK4 and CDK6 as part of the regulation of the G1-->S transition in the cell-division cycle. Loss of INK4 gene product function, particularly that of p16(INK4a), is found in 10-60% of human tumors, suggesting that broadly applicable anticancer therapies might be based on restoration of p16(INK4a) CDK inhibitory function. Although much less frequent, defects of p19(INK4d) have also been associated with human cancer (osteosarcomas). The protein structures of some INK4 family members, determined by nuclear magnetic resonance (NMR) spectroscopy and X-ray techniques, have begun to clarify the functional role of p16(INK4a) and the dysfunction introduced by the mutations associated with human tumors. RESULTS: The crystal structure of human p19(INK4d) has been determined at 1.8 A resolution using multiple isomorphous replacement methods. The fold of p19(INK4d) produces an oblong molecule comprising five approximately 32-residue ankyrin-like repeats. The architecture of the protein demonstrates the high structural similarity within the INK4 family. Comparisons to other ankyrin-repeat-containing proteins (GABPbeta, 53BP2 and myotrophin) show similar structures with comparable hydrogen-bonding patterns and hydrophobic interactions. Such comparisons highlight the splayed beta-loop geometry that is specific to INK4 inhibitors. This geometry is the result of a modified ankyrin structure in the second repeat. CONCLUSIONS: Among the INK4 inhibitors, the highest amino acid sequence conservation is found in the helical stacks; this conservation creates a conserved beta-loop geometry specific to INK4 inhibitors. Therefore, in addition to models which predict that the conserved helix alpha6 is responsible for CDK inhibition, a binding mode whereby the loops of INK4 proteins bind to the CDKs should also be considered. A similar loop-based interaction is seen in the complex formed between the ankyrin-repeat-containing protein GABPbeta and_GABPalpha. This mode of binding would be consistent with the observation that p16(INK4a) is sensitive to deleterious mutations found throughout this tumor suppressor protein; these mutations probably destabilize the three-dimensional structure.

Relaxation matrix refinement of the solution structure of squash trypsin inhibitor.

The structure of the small squash trypsin inhibitor CMTI-I is refined by directly minimizing the difference between the observed two-dimensional nuclear Overhauser enhancement (NOE) intensities and those calculated by the full relaxation matrix approach. To achieve this, a term proportional to this difference was added to the potential energy function of the molecular dynamics program X-PLOR. Derivatives with respect to atomic co-ordinates are calculated analytically. Spin diffusion effects are thus accounted for fully during the refinement. Initial structures for the refinement were those determined recently by solution nuclear magnetic resonance using the isolated two-spin approximation to derive distance range estimates. The fits to the nuclear magnetic resonance data improve significantly with only small shifts in the refined structures during a few cycles of conjugate gradient minimization. However, larger changes (approximately 1 A) in the conformation occur during simulated annealing, which is accompanied by a further reduction of the difference between experimental and calculated two-dimensional NOE intensities. The refined structures are closer to the X-ray structure of the inhibitor complexed with trypsin than the initial structures. The root-mean-square difference for backbone atoms between the initial structures and the X-ray structure is 0.96 A, and that between the refined structures and the X-ray structure 0.61 A.

Determination of the complete three-dimensional structure of the trypsin inhibitor from squash seeds in aqueous solution by nuclear magnetic resonance and a combination of distance geometry and dynamical simulated annealing.

The complete three-dimensional structure of the trypsin inhibitor from seeds of the squash Cucurbita maxima in aqueous solution was determined on the basis of 324 interproton distance constraints, 80 non-nuclear Overhauser effect distances, and 22 hydrogen-bonding constraints, supplemented by 27 phi backbone angle constraints derived from nuclear magnetic resonance measurements. The nuclear magnetic resonance input data were converted to the distance constraints in a semiquantitative manner after a sequence specific assignment of 1H spectra was obtained using two-dimensional nuclear magnetic resonance techniques. Stereospecific assignments were obtained for 17 of the 48 prochiral centers of the squash trypsin inhibitor using the floating chirality assignment introduced at the dynamical simulated annealing stage of the calculations. A total of 34 structures calculated by a hybrid distance geometry-dynamical simulated annealing method exhibit well-defined positions for both backbone and side-chain atoms. The average atomic root-mean-square difference between the individual structures and the minimized mean structure is 0.35(+/- 0.08) A for the backbone atoms and 0.89(+/- 0.17) A for all heavy atoms. The precision of the structure determination is discussed and correlated to the experimental input data.

Nuclear magnetic resonance solution and X-ray structures of squash trypsin inhibitor exhibit the same conformation of the proteinase binding loop.

A comparison of the solution nuclear magnetic resonance (n.m.r.) structures of squash trypsin inhibitor from seeds of the squash Cucurbita maxima with the X-ray structure of a trypsin complex of the inhibitor shows that the n.m.r. and X-ray structures are similar in terms of the global folding and secondary structure. The average atomic root-mean-square difference between the 36 n.m.r. structures on the one hand and the X-ray structure is 0.96 A for the backbone atoms and 1.95 A for all heavy atoms. The n.m.r. and X-ray structures exhibit extremely similar conformations of the primary proteinase binding loop. Despite the overall similarity, there are small differences between the mean computed structure and the X-ray structure. The n.m.r. structures have slightly different positions of the segments from residues 16 to 18, and 24 and 25. The n.m.r. results show that the inclusion of stereospecific assignments and precise distance constraints results in a significant improvement in the definition of the n.m.r. structure, making possible a detailed analysis of the local conformations in the protein.

Homonuclear three-dimensional NOE-NOE nuclear magnetic resonance/spectra for structure determination of proteins in solution.

The solution structures of two proteins (CMTI-I, a trypsin inhibitor from Cucurbita maxima, and hisactophilin, an actin binding protein of 118 amino acids) have been determined based on the NOE data derived solely from the homonuclear 3D NOE-NOE magnetic resonance spectroscopy. Two different approaches for extraction of the structural information from the 3D NOE-NOE experiment were tested. One approach was based on the transformation of the 3D intensities into distance constraints. In the second, and more robust approach, the 3D NOE intensities were used directly in structure calculations, without the need to transform them into distance constraints. A new 2D potential function representing the 3D NOE-NOE intensity was developed and used in the simulated annealing protocol. For CMTI-I, a comparison between structures determined with the 3D NOE-NOE method and various 2D NOE approaches was carried out. The 3D data set allowed better definition of the structures than was previously possible with the 2D NOE procedures that used the isolated two-spin approximation to derive distance information.

Backbone dynamics of the CDK inhibitor p19(INK4d) studied by 15N NMR relaxation experiments at two field strengths.

The four members of the INK4 gene family, p16(INK4a), p15(INK4b), p18(INK4c) and p19(INK4d), are known to bind to and inhibit the closely related cyclin-dependent kinases CDK4 and CDK6 as part of the regulation of the G1/S transition in the cell division cycle. Loss of INK4 gene product function, and particularly that of p16(INK4a), is found in human cancer. 15N NMR relaxation rates of p19(INK4d) were analyzed using the reduced spectral density mapping method. Most of the backbone of p19(INK4d) exists in a well-defined structure of limited conformational flexibility on the nanosecond to picosecond time-scales. Introducing appropriate scaling to account for the effects of anisotropy, a considerable amount of exchange broadening was found for several residues throughout the sequence, especially residues in the second ankyrin repeat and in the beginnings and ends of loops connecting ankyrin repeats. A possible mode of binding between p19(INK4d) and CDK4 and CDK6 could therefore involve the loop segments of p19(INK4d). The average overall correlation time taumeff was determined to be 13.6 ns, reflecting the tendency of p19(INK4d) to aggregate.

Structure of severin domain 2 in solution.

The three-dimensional structure of domain 2 of severin in aqueous solution was determined by nuclear magnetic resonance spectroscopy. Severin is a Ca(2+)-activated actin-binding protein that servers F-actin, nucleates actin assembly, and caps the fast-growing ends of actin filaments. The 114-residue domain consists of a central five-stranded beta-sheet, sandwiched between a parallel four-turn alpha-helix and, on the other face, a roughly perpendicular two-turn alpha-helix. There are two distinct binding sites for Ca2+ located near the N and C termini of the long helix. Conserved residues of the gelsolin-severin family contribute to the apolar core of domain 2 of severin, so that the overall fold of the protein is similar to those of segment 1 of gelsolin and profilins. Together with biochemical experiments, this structure helps to explain how severin interacts with actin.

Ca2+-loaded spherulin 3a from Physarum polycephalum adopts the prototype gamma-crystallin fold in aqueous solution.

Spherulin 3a is the most abundantly expressed cytosolic protein in spherulating plasmodia of the slime mold Physarum polycephalum. High yields of unlabeled, uniformly 15N and uniformly 13C/15N-labeled recombinant spherulin 3a from Escherichia coli could be produced by a simple protocol described here. The three-dimensional solution structure of Ca2+-loaded spherulin 3a was determined by homo- and heteronuclear NMR spectroscopy. The structure of monomeric spherulin 3a consists of two pleated beta-sheets plus a short alpha-helix arranged into the gamma-crystallin fold. The beta-sheets comprise two intertwined Greek-key motifs. An additional N-terminal beta-strand is unique to spherulin 3a. Complexation of calcium ions greatly enhances overall conformational stability of the protein. The average atomic root-mean-square deviations (r.m.s.d.) for heavy atoms in beta-strands were 0.34(+/-0.16) A for the backbone atoms and 0.73(+/-0.40) A for all atoms. The corresponding r.m.s.d. values for heavy atoms in the whole protein were 0.62(+/-0.42) A for the backbone atoms and 0.99(+/-0.65) A for all atoms. We show the structural relationship between spherulin 3a, a myxomycete dormancy protein, and crystallins from the vertebrate eye lens. Since spherulin 3a has a structure corresponding to one domain of bovine gammaB(II)-crystallin, it represents a hypothetical ancestral gamma-crystallin precursor structure.

Structure of the nidogen binding LE module of the laminin gamma1 chain in solution.

The structure of the single LE module between residues 791 and 848 of the laminin gamma1 chain, which contains the high affinity binding site for nidogen, has been probed using NMR methods. The module folds into an autonomous domain which has a stable and unique three-dimensional (3D) structure in solution. The 3D structure was determined on the basis of 362 interproton distance constraints derived from nuclear Overhauser enhancement measurements and 39 phi angles, supplemented by 5 psi and 22 chi1 angles. The main features of the NMR structures are two-stranded antiparallel beta-sheets which are separated by loops and cross-connected by four disulfide bridges. The N-terminal segment which contains the first three disulfide bridges is similar to epidermal growth factor. The C-terminal segment has an S-like backbone profile with a crossover at the last disulfide bridge and comprises two three-residue long beta-strands that form an antiparallel beta-sheet. The LE module possesses an exposed nidogen binding loop that projects away from the main body of the protein. The side-chains of three amino acids which are crucial for binding (Asp, Asn, Val) are all exposed at the domain surface. An inactivating Asn-Ser mutation in this region showed the same 3D structure indicating that these three residues, and possibly an additional Tyr in an adjacent loop, provide direct contacts in the interaction with nidogen.

The domains of protein S from Myxococcus xanthus: structure, stability and interactions.

Protein S from Myxococcus xanthus is a member of the beta gamma-crystallin superfamily. Its N and C-terminal domains (NPS and CPS, respectively) show a high degree of structural similarity and possess the capacity to bind two calcium ions per domain. For NPS, their positions were determined by X-ray diffraction at 1.8 A resolution, making use of molecular replacement with the NMR structure as search model. The overall topology of NPS is found to be practically the same as in complete protein S. In natural protein S, the domains fold independently, with a significant increase in stability and cooperativity of folding in the presence of Ca2+. The recombinant isolated domains are stable monomers which do not show any tendency to combine to "nicked" full-length protein S. In order to investigate the stability and folding of natural protein S and its isolated domains, spectroscopic techniques were applied, measuring the reversible urea and temperature-induced unfolding transitions at varying pH. The increment of Ca2+ to the free energy of stabilization amounts to -10 and -5 kJ/mol for NPS and CPS, respectively. For both NPS and CPS, in the absence and in the presence of 3 mM CaCl2, the two-state model is valid. Comparing DeltaGU-->N for CPS (-21 kJ/mol at pH 7, liganded with Ca2+) with its increment in the intact two-domain protein, the stability of the isolated domain turns out to be decreased in a pH-dependent manner. In contrast, the stability of Ca2+-loaded NPS (DeltaGU-->N=-31 kJ/mol, pH 7) is nearly unchanged down to pH 2 where Ca2+ is released (DeltaGU-->N=-26 kJ/mol, pH 2). In intact protein S, the N-terminal domain is destabilized relative to NPS. Evidently, apart from Ca2+ binding, well-defined domain interactions contribute significantly to the overall stability of intact protein S.

Molecular architecture of the rod domain of the Dictyostelium gelation factor (ABP120).

The Dictyostelium discoideum gelation factor is a two-chain actin-cross-linking protein that, in addition to an N-terminal actin-binding domain, has a rod domain constructed from six tandem repeats of a 100-residue motif that has an immunoglobulin fold. To define the architecture of the rod domain of gelation factor, we have expressed in E. coli a series of constructs corresponding to different numbers of gelation factor rod repeats and have characterised them by chemical crosslinking, ultracentrifugation, column chromatography, matrix-assisted laser desorption ionisation (MALDI) mass spectrometry and NMR spectroscopy. Fragments corresponding to repeats 1-6 and 5-6 dimerise, whereas repeats 1-5 and single repeats 3 and 4 are monomeric. Repeat 6 interacts weakly and was present as monomer and dimer when analysed by analytical ultracentrifugation. Proteolytic digestion of rod5-6 resulted in the generation of two polypeptides that roughly corresponded to rod5 and part of rod6. None of these polypeptides formed dimers after chemical crosslinking. Stable dimerisation therefore appears to require repeats 5 and 6. Based on these data a model of gelation factor architecture is presented. We suggest an arrangement of the chains where only the carboxy-terminal repeats interact as was observed for filamin/ABP280, the mammalian homologue of gelation factor.

NMR characterization of structure, backbone dynamics, and glutathione binding of the human macrophage migration inhibitory factor (MIF).

Human macrophage migration inhibitory factor is a 114 amino acid protein that belongs to the family of immunologic cytokines. Assignments of 1H, 15N, and 13C resonances have enabled the determination of the secondary structure of the protein, which consists of two alpha-helices (residues 18-31 and 89-72) and a central four-stranded beta-sheet. In the beta-sheet, two parallel beta-sheets are connected in an antiparallel sense. From the total of three cysteines present in the primary structure of MIF, none was found to form disulfide bridges. 1H-15N heteronuclear T1, T2, and steady-state NOE measurements indicate that the backbone of MIF exists in a rigid structure of limited conformational flexibility (on the nanosecond to picosecond time scale). Several residues located in the loop regions and at the N termini of two helices exhibit internal motions on the 1-3 ns time scale. The capacity to bind glutathione was investigated by titration of a uniform 15N-labeled sample and led us to conclude that MIF has, at best, very low affinity for glutathione.

Improved strategies for the determination of protein structures from NMR data: the solution structure of acyl carrier protein.

The hybrid method that combines the early stages of a distance geometry program with simulated annealing in the presence of NMR constraints was optimized to obtain structures fully consistent with the observed NMR data. This was achieved by using more restrictive bounds of the NOE constraints than those usually used in the literature and by grouping the NOEs into classes dependent on the quality of the experimental NOE data. The 'floating' stereospecific assignment introduced at the simulated annealing stage of the calculations further improved the definition of the local conformation. An improved sampling and convergence property of the hybrid method was obtained by means of fitting the substructure obtained from the distance geometry program to different conformations. Compared to the standard hybrid methods, this procedure gave superior structures for a 77 amino acid protein, acyl carrier protein from Escherichia coli.

1H NMR assignments of sidechain conformations in proteins using a high-dimensional potential in the simulated annealing calculations.

A high-dimensional potential representing distance constraints for stereospecifically assignable diastereotopic proton or methyl pairs was incorporated into the dynamical simulated annealing protocol to calculate structure with stereospecifically determined sidechain conformations. The protocol is tested on nuclear magnetic resonance cross-relaxation data of a trypsin inhibitor from squash seeds, CMTI-I, and compared with two other methods of stereospecific assignment, the floating chirality and coupling constant methods. There is good agreement between the three methods in predicting the same stereospecific assignments. Because the high-dimensional potential uses more relaxed absolute distance constraints and also takes into account the relative distance constraint patterns, it avoids possible overinterpretation of the NOE data.

Secondary structure of human granulocyte colony-stimulating factor derived from NMR spectroscopy.

Recombinant 15N-, 13C-labeled human granulocyte colony-stimulating factor (rh-metG-CSF) has been studied by 2D and 3D NMR using uniformly labeled protein as well as residue-specific 15N-labeled samples. Assignment of the 1H, 15N backbone, and 60% 1H sidechain resonances has enabled the determination of the secondary structure of the protein. The secondary structure is dominated by alpha-helical regions with four stretches of helices between residues 11-41, 71-95, 102-124 and 144-170.

Comparison of the amide proton exchange behavior of the rapidly formed folding intermediate and the native state of an antibody scFv fragment.

We have investigated the stability of backbone amide protons of the intermediate and the native state of the scFv fragment of an antibody. Stopped flow experiments analyzed by MS and NMR detected the formation of an exchange protected intermediate within the deadtime of the stopped flow apparatus (17 ms). H/D exchange rates of the native protein identified a number of very stable backbone amide protons in the V(L) and the V(H) domains. In the V(L) domain, this slowly exchanging core of the scFv fragment is similar to the folding core of the intermediate, while the V(H) domain possesses a great number of very stable amide protons which are not stabilized to a significant degree in the folding intermediate of the scFv fragment.

Characterization of the linker peptide of the single-chain Fv fragment of an antibody by NMR spectroscopy.

A comparison of the single-chain Fv fragment of the antibody McPC603 (scFv) with its corresponding unlinked Fv fragment has been carried out with 15N-edited NMR spectroscopy. The two Fv fragments adopt the same structure, indicating that the linker does not perturb the folding of the domains. This also directly demonstrates that folding in vivo (Fv fragment) and in vitro (scFv fragment) leads to the same structure. The main differences in the spectra of the uniformly 15N-labeled scFv and Fv fragments are due to signals of Gly and Ser from the linker peptide of the scFv fragment. The linker peptide has been mapped with NMR spectra of 15N-glycine- and 15N-glycine/15N-serine-labeled scFv fragments. The 15N T2 relaxation data indicate that the linker peptide is more flexible than the rest of the molecule.