Miguel Llinás

The CCPN data model for NMR spectroscopy: Development of a software pipeline

To address data management and data exchange problems in the nuclear magnetic resonance (NMR) community, the Collaborative Computing Project for the NMR community (CCPN) created a “Data Model” that describes all the different types of information needed in an NMR structural study, from molecular structure and NMR parameters to coordinates. This paper describes the development of a set of software applications that use the Data Model and its associated libraries, thus validating the approach. These applications are freely available and provide a pipeline for high‐throughput analysis of NMR data. Three programs work directly with the Data Model: CcpNmr Analysis, an entirely new analysis and interactive display program, the CcpNmr FormatConverter, which allows transfer of data from programs commonly used in NMR to and from the Data Model, and the CLOUDS software for automated structure calculation and assignment (Carnegie Mellon University), which was rewritten to interact directly with the Data Model. The ARIA 2.0 software for structure calculation (Institut Pasteur) and the QUEEN program for validation of restraints (University of Nijmegen) were extended to provide conversion of their data to the Data Model. During these developments the Data Model has been thoroughly tested and used, demonstrating that applications can successfully exchange data via the Data Model. The software architecture developed by CCPN is now ready for new developments, such as integration with additional software applications and extensions of the Data Model into other areas of research. Proteins 2005.

Characterization of kringle domains of angiostatin as antagonists of endothelial cell migration, an important process in angiogenesis.

Angiogenesis is a complex process that involves endothelial cell proliferation, migration, basement membrane degradation, and neovessel organization. Angiostatin, consisting of four homologous triple‐disulfide bridged kringle domains, has previously been shown to exhibit profound inhibition of endothelial cell proliferation in vitro and angiogenesis in vivo. It was also demonstrated that angiostatin could suppress the growth of a variety of tumors via the blocking of angiogenesis. The primary aim of our study was to characterize the kringle domains of angiostatin for their inhibitory activities of endothelial cell migration in order to elucidate their contributions to the anti‐angiogenic function of angiostatin. In this report, we demonstrate for the first time that the kringles of angiostatin play different roles in inhibiting endothelial cell migration, a crucial process in angiogenesis. Kringle 4, which has only marginal anti‐proliferative activity, is among the most potent fragments in inhibiting endothelial cell migration (IC50 of approximately 500 nM). In contrast, kringle 1–3, which is equivalent to angiostatin in inhibiting endothelial cell proliferation, manifests only a modest anti‐migratory effect. The combination of kringle 1–3 and kringle 4 results in an anti‐migratory activity comparable to that of angiostatin. When kringle 1 is removed from kringle 1–3, the resulting kringle 2–3 becomes more potent than kringle 1–3. This implies that kringle 1, although virtually ineffective in inhibiting endothelial cell migration, may influence the conformation of kringle 1–3 to alter its anti‐migratory activity. We also show that disruption of the kringle structure by reducing/alkylating agents markedly attenuates the anti‐migratory activity of angiostatin, demonstrating the significance of kringle conformation in maintaining the anti‐angiogenic activity of angiostatin. Our data suggest that different kringle domains may contribute to the overall anti‐angiogenic function of angiostatin by their distinct anti‐migratory activities.—Ji, W. R., Castellino, F. J., Chang, Y., DeFord, M. E., Gray, H., Villarreal, X., Kondri, M. E., Marti, D. N., Llinás, M., Schaller, J., Kramer, R. A., and Trail, P. A. Characterization of kringle domains of angiostatin as antagonists of endothelial cell migration, an important process in angiogenesis. FASEB J. 12, 1731–1738 (1998)

Refined solution structure and ligand-binding properties of PDC-109 domain b. A collagen-binding type II domain.

We have determined, via 1H-n.m.r., the solution conformation of the collagen-binding b-domain of the bovine seminal fluid protein PDC-109 (PDC-109/b). The structure determination is based on 341 interproton distance estimates and 42 dihedral angle estimates: a set of 24 initial structures were computed; 12 using the variable target function program DIANA, and 12 using the metric matrix program DISGEO. These structures were optimized by restrained energy minimization and dynamic simulated annealing using the CHARMM and X-PLOR programs. The average pairwise root-mean-square difference (r.m.s.d) between the optimized DIANA (DISGEO) structures is 0.71 A (0.82 A) for the backbone atoms, and 1.73 A (2.03 A) for all atoms. Both sets of structures exhibit the same global fold, secondary structure and placement of most non-polar side-chains. Two central antiparallel beta-sheets, which lie roughly perpendicular to each other, and two irregular loops support a large, partially exposed, hydrophobic surface that defines a putative binding site. A test of a hybrid relaxation matrix-based distance refinement protocol (MIDGE program) was performed using a normalized 250 millisecond NOESY spectrum. The resulting distances were input to the molecular mechanics/dynamics procedures mentioned above in order to optimize the DIANA structures. Our results indicate that relaxation matrix refinement of distances is most useful when used conservatively for identifying underestimated distance constraints. 1H-n.m.r. monitored ligand titration experiments revealed definite, albeit weak, binding interactions for phenethylamine and leucine analogs (Ka less than or equal to 25 M-1). Residues perturbed by ligand binding include Tyr7, Trp26, Tyr33, Asp34 and Trp39. These results suggest that PDC-109/b may recognize specific leucine and/or isoleucine-containing sequences within collagen.

CLOUDS, a protocol for deriving a molecular proton density via NMR

We demonstrate the feasibility of computing realistic spatial proton distributions for proteins in solution from experimental NMR nuclear Overhauser effect data only and with minimal assignments. The method, CLOUDS, relies on precise and abundant interproton distance restraints calculated via a relaxation matrix analysis of sets of experimental nuclear Overhauser effect spectroscopy crosspeaks. The MIDGE protocol was adapted for this purpose. A gas of unassigned, unconnected H atoms is condensed into a structured proton distribution (cloud) via a molecular dynamics simulated-annealing scheme in which the internuclear distances and van der Waals repulsive terms are the only active restraints. Proton densities are generated by combining a large number of such clouds, each computed from a different trajectory. After filtering by reference to the cloud closest to the mean, a minimal dispersion proton density (foc) is identified. The latter affords a quasi-continuous hydrogen-only probability distribution that conveys immediate information on the protein surface topology (grooves, protrusions, potential binding site cavities, etc.), directly related to the molecular structure. Feasibility of the method was tested on NMR data measured on two globular protein domains of low regular secondary structure content, the col 2 domain of human matrix metalloproteinase-2 and the kringle 2 domain of human plasminogen, of 60 and 83 amino acid residues, respectively.

Domain Organization of the 39-kDa Receptor-associated Protein

The 39-kDa receptor-associated protein (RAP) is an endoplasmic reticulum resident protein that binds to the low density lipoprotein receptor-related protein (LRP) as well as certain members of the low density lipoprotein receptor superfamily and antagonizes ligand binding. In order to identify important functional regions of RAP, studies were performed to define the domain organization and domain boundaries of this molecule. Differential scanning calorimetry (DSC) experiments revealed that the process of thermal denaturation of RAP is highly reversible and occurs in a broad temperature range with two well resolved heat absorption peaks. A good fit of the endotherm was obtained with four two-state transitions suggesting these many cooperative domains in the molecule. A number of recombinant fragments of RAP were expressed in bacteria, and their domain composition and stability were characterized by DSC, circular dichroism, and fluorescence spectroscopy. The results confirmed that RAP is composed of four independently folded domains, D1, D2, D3, and D4, that encompass residues 1–92, 93–163, 164–216, and 217–323, respectively. The first and the fourth domains preserved their structure and stability when isolated, whereas the compact structure of the fragment corresponding to D2 seems to be altered when isolated from the parent molecule. Isolated D3 was partially degraded during isolation from bacterial lysates. The isolated D4 was capable of binding with high affinity to LRP whereas neither D1 nor D2 bound. At the same time a fragment containing both D1 and D2 exhibited high affinity binding to LRP. These facts combined with the thermodynamic analysis of the melting process of the fragments containing D1 and D2 indicate that these two domains interact with each other and that the proper folding of the second domain into a native-like active conformation requires presence of the first domain.

Proton magnetic resonance study of lysine-binding to the Kringle 4 domain of human plasminogen: the structure of the binding site

The binding of L-Lys, D-Lys and epsilon-aminocaproic acid (epsilon ACA) to the kringle 4 domain of human plasminogen has been investigated via one and two-dimensional 1H-nuclear magnetic resonance spectroscopy at 300 and 600 MHz. Ligand-kringle association constants (Ka) were determined assuming single site binding. At 295 K, pH 7.2, D-Lys binds to kringle 4 much more weakly (Ka = 1.2 mM-1) than does L-Lys (Ka = 24.4 mM-1). L-Lys binding to kringle 4 causes the appearance of ring current-shifted high-field resonances within the -1 approximately less than delta approximately less than 0 parts per million range. The ligand origin of these signals has been confirmed by examining the spectra of kringle 4 titrated with deuterated L-Lys. A systematic analysis of ligand-induced shifts on the aromatic resonances of kringle 4 has been carried out on the basis of 300 MHz two-dimensional chemical shift correlated (COSY) and double quantum correlated spectroscopies. Significant differences in the effect of L-Lys and D-Lys binding to kringle 4 have been observed in the aromatic COSY spectrum. In particular, the His31 H4 and Trp72 H2 singlets and the Phe64 multiplets appear to be the most sensitive to the particular enantiomers, indicating that these residues are in proximity to the ligand C alpha center. In contrast, the rest of the indole spectrum of Trp72 and the aromatic resonances of Trp62 and Tyr74, which are affected by ligand presence, are insensitive to the optical nature of the ligand isomer. These results, together with two-dimensional proton Overhauser studies and ligand-kringle saturation transfer experiments reported previously, enabled us to generate a model of the kringle 4 ligand-binding site from the crystallographic co-ordinates of the prothrombin kringle 1. The latter, although lacking recognizable lysine-binding capability, is otherwise structurally homologous to the plasminogen kringles.

The second type II module from human matrix metalloproteinase 2: Structure, function and dynamics

Abstract Background: Matrix metalloproteinase 2 (MMP-2, gelatinase A, 72 kDa type IV collagenase) has an important role in extracellular matrix degradation during cell migration and tissue remodeling. It is involved in development, inflammation, wound healing, tumor invasion, metastasis and other physiological and pathological processes. The enzyme cleaves several types of collagen, elastin, fibronectin and laminin. Binding to collagen is mediated by three repeats homologous to fibronectin type II modules, which are inserted in the catalytic domain in proximity to the active site. Results: We have determined the NMR solution structure of the second type II module from human MMP-2 (col-2). The module exhibits a typical type II fold with two short double-stranded antiparallel β sheets and three large loops packed around a cluster of conserved aromatic residues. Backbone amide dynamics, derived from 15 N relaxation experiments, correlate well with solvent accessibility and intramolecular hydrogen bonding. A synthetic peptide with the collagen consensus sequence, (Pro-Pro-Gly) 6 , is shown to interact with the module. Conclusions: Spectral perturbations induced by (Pro-Pro-Gly) 6 binding reveal the region involved in the interaction of col-2 with collagen. The binding surface comprises exposed aromatic residues Phe21, Tyr38, Trp40, Tyr47, Tyr53 and Phe55, and the neighboring Gly33–Gly37 segment.

Peptide Ligands for the Fibronectin Type II Modules of Matrix Metalloproteinase 2 (MMP-2)

The interaction of matrix metalloproteinase 2 (MMP-2) with gelatin is mediated by three repeats homologous to fibronectin type II (FN2) modules, which are inserted in the catalytic domain in proximity of the active site. We screened a random 15-mer phage display library to identify peptides that interact with the FN2 modules of MMP-2. Interestingly, the selected peptides are not gelatin-like and do not share a common, obvious sequence motif. However, they contain a high proportion of aromatic residues. The interactions of two peptides, WHWRH0RIPLQLAAGR and THSHQWRHHQFPAPT, with constructs comprising the in-tandem first and second and second and third FN2 modules of MMP-2 (Col-12 and Col-23, respectively) were characterized by NMR. Both peptides interact with Col-12 and Col-23 with apparent association constants in the mm −1 range. Peptide binding results in perturbation of signals from residues located in the gelatin-binding pocket and flexible parts of the molecule. Although the former finding suggests that the gelatin-binding site is involved in the contact, the interpretation of the latter is less straightforward and may well reflect both the direct and indirect effects of the interaction.

Protein structure elucidation from NMR proton densities

The NMR-generated foc proton density affords a template to which the molecule has to be fitted to derive the structure. Here we present a computational protocol that achieves this goal. HN atoms are readily recognizable from 1H/2H exchange or 1H/15N heteronuclear single quantum correlation (HSQC) experiments. The primary structure is threaded through the unassigned foc by leapfrogging along peptidyl amide HNs and the connected Hαs. Via a Bayesian approach, the probabilities of the sequential connectivity hypotheses are inferred from likelihoods of HN/HN, HN/Hα, and Hα/Hα interatomic distances as well as 1H NMR chemical shifts, both derived from public databases. Once the polypeptide sequence is identified, directionality becomes established, and the foc N and C termini are recognized. After a similar procedure, side chain H atoms are found, including discriminated cis/trans proline loci. The folded structure then is derived via a direct molecular dynamics embedding into mirror image-related representations of the foc and selected according to a lowest energy criterion. The method was applied to foc densities calculated for two protein domains, col 2 and kringle 2. The obtained structures are within 1.0–1.5 Å (backbone heavy atoms) and 1.5–2.0 Å (all heavy atoms) rms deviations from reported x-ray and/or NMR structures.

Model-independent refinement of interproton distances generated from 1H NMR overhauser intensities

A recursive method to refine interproton distances compatible with two-dimensional nuclear Overhauser effect (NOESY) experiments has been tested. Convergence does not depend on the initial estimate of the parameters. Hence, no approximate initial structure of the molecule is required: the iterative process can be started from the experimentally measured NOESY cross-peak volumes, supplemented with arbitrary cross-peak and autopeak values to obtain an initial NOESY matrix. The relaxation matrix is calculated from the NOESY matrix, and its diagonal elements (ϱi) are adjusted at each iteration until the difference between theoretical and experimental cross peaks is a minimum. The improvement comes from using interproton distances calculated from the off-diagonal (σij) elements to generate ϱi values. The method was applied to alumichrome, a rigid cyclohexapeptide of virtually identical solution and crystallographic structures. The experimental data consisted of the integrated volumes of NOESY cross peaks at 500 MHz. Convergence was tested by resorting to different initial conditions, one of them being a NOESY matrix in which the experimentally unobserved off diagonal elements were set equal to zero and the diagonal elements to 0.5. The iterations rapidly converge, in all cases, to a set of distances whose root-mean-squares deviation (rmsd) from the crystallographic distances is < 0.05 A. The acronym MIDGE (model-independent distance generation) for the procedure is proposed.