Protein folding/dynamics and function; NMR spectroscopy; computational methods for solving the fold using sparse, unassigned data; ligand-protein interaction; stochastic ligand docking; plasminogen and related modular constructs
Many extracellular proteins are built in mosaic-like fashion, namely, as assembled via a covalently linked array of modular units. Each unit, or domain, is characterized by its own fold and thermal stability profile. In the case of plasminogen, the proteinase is composed of an N-terminal "preactivation" domain, five kingle domains, and a trypsinogen-like protease zymogen that, upon activation, catalyzes proteolysis of major proteins, such as polymerized fibrin in blood clots or components of the extracellular matrix in tissues. Kringles afford the units responsible for plasminogen binding to the various potential substrates. With the aim of defining the supra-fold of the mosaic construct, we follow two approaches: (a) the NMR spectroscopic characterization (1-3) and (b) computational ab-initio modeling of single domain and multidomain constructs in water (in collaboration with the (Prof. Kim's group) (4). A main facet of these studies is to identify intramolecular interactions resulting from the supra-fold before and after the activation cleavage, and how effectors, including anti-fibrinolytic drugs, can modulate the supra-fold.
We have developed CLOUDS, an approach for "directly" solving the protein conformation from unassigned NMR data (5). The method is based on optimally matching the nuclear spin spatial distribution to a fold that minimizes the "energy" of the covalent framework vis-à-vis the empirical NMR constraints, usually spin-spin dipolar (Overhauser) connectivities. CLOUDS involves a suite of programs that "intelligently" discriminate and group signals in terms of Bayesian probabilistic criteria. The approach has been extended to the case of sparse constraints (6), as is the case when dealing with large proteins with specific isotopic labels (such as protons against a bleached deuterated background) that simplify the NMR spectra but results in sparse data. Optimal matching of signals to chemical groups is done via bipartite graph matching criterion.
By exploiting NMR chemical shift perturbation, it is often feasible to localize binding sites for specific ligands on the exposed protein surface(1-3). These sites can be structurally "refined" via docking algorithms that incorporate the experimental NMR constraints, which act as a filter. However, in some cases of biological relevance, as when the ligand is weakly interacting or no such "unique" localized protein site exists, such as areas of distributed electrostatic charges, chemical shift perturbation can provide a fuzzy, distributed picture of the interaction area. Such is the case, e.g., with heparin-like sulfated saccharides interacting with some kringle domains. In such cases, while computational docking can still localize potential binding site(s) Prima facie this would appear unsatisfactory in terms of defining the specificity of the interaction. The latter, however, can be exploited "stochasticastically" to sample and define patches on the protein surface that potentially could interact with polymeric heparin in blood or in the extracellular matrix (8). This approach has been exploited successfully on the kringle 3 domain of plasminogen and has been extended to other non-plasminogen kringles that exhibit heparin affinity as detected, e.g., via affinity chromatography against heparin-conjugated resins.
|Years||Position or Degree|
|2004||Visiting Professor, Unité de Bioinformatique Structurale, Institut Pasteur, Paris, France|
|2003–present||Member, Cambridge Collaborative Computation NMR (CCPN), UK|
|2003||Visiting Professor, Department of Biochemistry, Cambridge University, UK|
|1992||Visiting Professor, Department of Organic Chemistry, University of Barcelona, Spain|
|1990||Visiting Professor, Macromolecular Chemistry Unit, Catalonia Polytechnic University, Barcelona, Spain|
|1989||Visiting Professor, Department of Chemistry, University of Utrecht, The Netherlands|
|1988–present||Professor, Carnegie Mellon University|
|1976–1988||Associate Professor, Carnegie Mellon University|
|1974–1976||Assistant, Institut für Molekularbiologie und Biophysik, Swiss Federal Institute of Technology, Switzerland|
|1973||Lecturer, University of California, Berkeley|
|1971–1974||Postdoctoral Research Associate, University of California, Berkeley|
|1971||Ph.D. Biophysics, University of California, Berkeley|
|1963||Licentiate in Physics, Cordoba National University, Argentina|
|2007||Member, National Academy of Sciences, Argentina|
|2003||Lincei Academy Lecturer, Bressanone, Italy|
|2001||Honorary Member, Sociedad Argentina de Investigaciones en Quimica Organica (SAIQO)|
(1) D.N. Marti, C.-K. Hu, S.S.A. An, P. von Haller, J. Schaller & M. Llinás (1997). “Ligand Preference of Kringle 2 and Homologous Domains of Human Plasminogen: Canvassing Weak, Intermediate and High Affinity Binding Sites by 1H-NMR” Biochemistry 36, 11591-11604.
(2) M.D. Battistel, A. Grishaev, S.S.A. An, F.J. Castellino & M. Llinás (2009). "Solution Structure and Functional Characterization of Human Plasminogen Kringle 5". Biochemistry 48, 10208-10219.
(3) M.T. Christen, P. Frank, J. Schaller & M. Llinás (2010). "Human Plasminogen Kringle 3: Solution Structure, Functional Insights, Phylogenetic Landscape". Biochemistry 49, 7131-7150.
(4) H.J. Kim, M.Y. Choi, H.J. Kim & M. Llinás (2010). "Conformational Dynamics and Ligand Binding in the Multi-Domain Protein PDC109". PLoS ONE 5(2): e9180. doi:10.1371/journal.pone.0009180
(5) A. Grishaev & M. Llinás (2005). “Protein Structure Elucidation from Minimal NMR Data: the CLOUDS Approach” Meth. Enzymology 394, 261-295.
(6) G.A. Bermejo & M. Llinás (2008). "Deuterated Protein Folds Obtained Directly from Unassigned Nuclear Overhauser Effect Data". J. Am. Chem. Soc. 130, 3797-3805.
(7) G.A. Bermejo & M. Llinás (2010). "Structure-Oriented Methods for Protein NMR Data Analysis". Progr. NMR Spectroscopy 56, 311-328.(8) M.T. Christen (2010), Doctoral Dissertation, Carnegie Mellon University.