joint application of nmr and scattering
play

Joint Application of NMR and Scattering Haydyn Mertens PhD Methods - PowerPoint PPT Presentation

Joint Application of NMR and Scattering Haydyn Mertens PhD Methods combining SAXS with NMR Filtering/scoring NMR conformers Direct refinement of structure Refinement of domain/subunit positions Restraints: SAXS & NMR Orientation


  1. Joint Application of NMR and Scattering Haydyn Mertens PhD

  2. Methods combining SAXS with NMR Filtering/scoring NMR conformers Direct refinement of structure Refinement of domain/subunit positions

  3. Restraints: SAXS & NMR Orientation Shapes/envelopes RDC s PCS Interface Dimensions (Rg, Dmax) NOEs Chemical Shifts PRE

  4. Restraints: SAXS & NMR Distances: NOEs Chemical Shifts PRE Orientation: RDCs PCS SAXS provides shape/distance info

  5. Model Filtering

  6. Model filtering by SAXS Structures determined X-ray crystallography NMR Homology modeling Docking algorithms Score these models using SAXS data best fit to SAXS curves cluster models

  7. Model filtering by SAXS pyDockSAXS FTDock pyDock SAXS (CRYSOL) ~ 40% improvement in selection of correct solution Pons et al., J Mol Biol. 2010 Oct 22;403(2):217-30

  8. Docking Steps: pyDockSAXS FTDock (http://www.sbg.bio.ic.ac.uk/docking/ftdock.html) generates large pool of complexes Surface complementarity Electrostatics filter pyDOCK (http://www.cllgenome.es/servlet/pydock/) scores/ranks FTDock complexes Binding energy electrostatics desolvation vdw CRYSOL (http://www.embl-hamburg.de/biosaxs/atsas-online/) scores/ranks complexes (fit to SAXS)

  9. Docking Performance: pyDockSAXS Identification of near native solutions 43% 29% 23% 21% Rank (top N predictions for each benchmark member) Pons et al., J Mol Biol. 2010 Oct 22;403(2):217-30 Combined pyDOCK + CRYSOL is best!

  10. Docking Performance: pyDockSAXS Anisometry has an impact on selection Pons et al., J Mol Biol. 2010 Oct 22;403(2):217-30 Spherical complexes hard to select by SAXS

  11. Model filtering by SAXS pyDockSAXS combined scoring function: pyDockSAXS = E pyDOCK + w c X CRYSOL SAXS can aid in the selection of near-native docking models Pons et al., J Mol Biol. 2010 Oct 22;403(2):217-30

  12. Building a model with RDCs & SAXS

  13. Building a model with RDCs & SAXS Basic idea: Use known domain structures Orient domains with RDCs Reduce possible solutions with SAXS domain-distance ( Rg ) scoring/filtering

  14. Example: Calmodulin Calmodulin (CaM) Ca 2+ binding protein (4 EF-hand motifs) Binds multiple targets (multiple functions) Structure is flexible Solved in both extended & compact forms +ligand -ligand Extended form Compact form

  15. Example: Calmodulin-TFP (trifluoperazine) CaM compacts upon TFP binding decreased Dmax observed by SAXS Quaternary structure built from RDCs, domain structures and SAXS SAXS: Rg = 2.0 nm -> Rg = 1.8 nm +TFP Mattinen et al., Biophys J. (2002) 83:1177-1183

  16. Building CaM-TFP model RDCs determine PAS (Azz, Axx, Ayy) D NH from single alignment medium Orientations using free CaM domain structures Mattinen et al., Biophys J. (2002) 83:1177-1183

  17. Building CaM-TFP model 4 degenerate relative orientations SAXS constrains distance between domains ( Rg ) Reduced to a single solution SAXS 1LIN Solutions scored using CRYSOL Mattinen et al., Biophys J. (2002) 83:1177-1183

  18. SASREF can do this too! (in principle)

  19. Using RDCs in SASREF Collect RDCs Define domain orientations PALES (http://www.mpibpc.mpg.de/groups/zweckstetter/_links/software_pales.htm) Xplor-NIH python tools (http://nmr.cit.nih.gov/xplor-nih/) Use pre-oriented PDBs as input rigid bodies Allow only translations in SASREF

  20. Direct Refinement

  21. Direct Refinement Requires accurate and fast calculation of scattering intensity from structure What people currently use: Spherical harmonics (CRYSOL) Zernike potentials (SASTBX) Debye (Xplor-NIH, MODELLER)

  22. Quick reminder: intensity calc Debye: Form factors Distances Need to evaluate all distances Computationally expensive scales quadratically with number of atoms

  23. Quick reminder: intensity calc Improve speed: globbic approximation group atoms (eg. GASBOR)

  24. Quick reminder: intensity calc Multipole expansion (spherical harmonics) Very fast Computation time is linear with size

  25. Quick reminder: intensity calc Direct refinement with NMR & SAXS slow back-calculation of I(s) globbic approx can help reduce number of points bin experimental data less points for calc I(s) Excluded volume treatment Hydration layer treatment

  26. Xplor-NIH (direct refinement)

  27. Xplor-NIH example: E. coli Enzyme-I : HPr complex Active sugar-phosphate transfer in bacteria Phosporylation of E1 and transfer to HPr Significant differences in crystal and NMR structures E1 crystallographic dimers Crystal NMR Schwieters et al., JACS (2010) 132:13026-13045

  28. Xplor-NIH example: Strategy: backbone n-h RDCs show E1 nter subdomain orientations unchanged (free vs complex) thus assuming E1 cter dimerisation domain structure unchanged from E1 crystal structures: orientation E1 dimer from E1 nter RDCs Combine with SAXS/WAXS data Determine free E1 Introduce HPr and determine E1:HPr Schwieters et al., JACS (2010) 132:13026-13045

  29. Xplor-NIH example: Backbone n-h RDCs show E1 nter subdomain orientations unchanged (free vs complex) Schwieters et al., JACS (2010) 132:13026-13045

  30. Xplor-NIH: protocol E1 dimer E1-HPr dimer Schwieters et al., JACS (2010) 132:13026-13045

  31. Comparison to existing structures: Schwieters et al., JACS (2010) 132:13026-13045

  32. Xplor-NIH example: Using SANS to discriminate between 2 possible E1-HPr clusters? 2 H- E1 / 1 H- HPr in 40% D2O (see E1 only) Cluster-1 Cluster-2 Schwieters et al., JACS (2010) 132:13026-13045

  33. Model Mechanism: E1-HPr Schwieters et al., JACS (2010) 132:13026-13045 Relaxation to I through X-ray crystal intermediate (phosphorylated intermediate) Free E1 Relaxation to V through intermediate NMR/SAXS

  34. CNS (direct refinement)

  35. CNS: xafs.f module Direct refinement (RDCs & SAXS) Power series expansion of SAXS curve (FAST) Rg region ---> inter-domain distance Higher angles ---> domain positions Grid-search to account for solvation Gabel et al., JBNMR (2008) 41:199-208

  36. CNS example SAXS + RDCs reduces possible orientations reduces inter-domain translations Gabel et al., JBNMR (2008) 41:199-208

  37. DADIMODO (search algorithm)

  38. DADIMODO Genetic algorithm based Optimise multi-domain structures Adjusts regions of non-defined structure Start from crystal/nmr structure or homology model Objective function (SAXS & RDCs)

  39. DADIMODO Mareuil et al., Eur biophys J (2007) 37:95-104

  40. DADIMODO example gamma-S crystallin vertebrate eye lens component Target Domain Models Conformers (25) Mareuil et al., Eur biophys J (2007) 37:95-104

  41. DADIMODO example gamma-S crystallin Mareuil et al., Eur biophys J (2007) 37:95-104

  42. Some considerations Method of intensity calculation spherical harmonics FAST, accurate (to ~ 5 nm -1 ) Debye SLOW, accurate (to 5-10 nm -1 ) Treatment of solvent Envelope Explicit layer of water Solvent treatment paramount for WAXS

  43. Summary NMR and SAXS/SANS very complementary Scoring Direct refinement Rigid body refinement and docking But if data is rubbish --> model is rubbish! check overall parameters!!!

Recommend


More recommend