1. Visualisation/prediction 3D structures

2. Recognition ability is the basis of biological function 3D struture is key for recognition

3. Objectives  Visualize / understand 3D structures and their interactions Derive structure-function relationships  Predict 3D structure

4. http://www.pdb.org

6. Total entries

7. Protein folds

10. Structure prediction

11. aim  Structure prediction tries to build models of 3D structures of proteins that could be useful for understanding structure-function relationships.

13. The protein folding problem  The information for 3D structures is coded in the protein sequence  Proteins fold in their native structure in seconds  Native structures are both thermodynamically stables and kinetically available

14. AVVTW...GTTWVR ab-initio prediction  Prediction from sequence using first principles

15. Ab-initio prediction  “In theory”, we should be able to build native structures from first principles using sequence information and molecular dynamics simulations: “Ab-initio prediction of structure” Simulaciones de 1 s de “folding” de una proteína modelo (Duan-Kollman: Science, 277, 1793, 1998). Simulaciones de folding reversible de péptidos (20- 200 ns) (Daura et al., Angew. Chem., 38, 236, 1999). Simulaciones distribuidas de folding de Villin (36- residues) (Zagrovic et al., JMB, 323, 927, 2002).

16. ... the bad news...  It is not possible to span simulations to the “seconds” range  Simulations are limited to small systems and fast folding/unfolding events in known structures steered dynamics biased molecular dynamics  Simplified systems

17. Some protein from E.coli predicted at 7.6 Å (CASP3, H.Scheraga) Results from ab-initio  Average error 5 Å - 10 Å  Function cannot be predicted  Long simulations

18. comparative modelling  The most efficient way to predict protein structure is to compare with known 3D structures

19. Basic concept  In a given protein 3D structure is a more conserved characteristic than sequence Some aminoacids are “equivalent” to each other Evolutionary pressure allows only aminoacids substitutions that keep 3D structure largely unaltered  Two proteins of “similar” sequences must have the “same” 3D structure

20. Possible scenarios 1. Homology can be recognized using sequence comparison tools or protein family databases (blast, clustal, pfam,...). Structural and functional predictions are feasible 2. Homology exist but cannot be recognized easily (psi- blast, threading) Low resolution fold predictions are possible. No functional information. 3. No homology 1D predictions. Sequence motifs. Limited functional prediction. Ab-initio prediction

21. fold prediction

22. 3D struc. prediction

23. 1D prediction  Prediction is based on averaging aminoacid properties AGGCFHIKLAAGIHLLVILVVKLGFSTRDEEASS Average over a window

25. 1D prediction. Properties  Secondary structure propensitites  Hydrophobicity  Accesibility ...

26. Propensities Chou-Fasman Biochemistry 17, 4277 1978   turn

27. Some programs (www.expasy.org) BCM PSSP - Baylor College of Medicine Prof - Cascaded Multiple Classifiers for Secondary Structure Prediction GOR I (Garnier et al, 1978) [At PBIL or at SBDS] GOR II (Gibrat et al, 1987) GOR IV (Garnier et al, 1996) HNN - Hierarchical Neural Network method (Guermeur, 1997) Jpred - A consensus method for protein secondary structure prediction at University of Dundee nnPredict - University of California at San Francisco (UCSF) PredictProtein - PHDsec, PHDacc, PHDhtm, PHDtopology, PHDthreader, MaxHom, EvalSec from Columbia University PSA - BioMolecular Engineering Research Center (BMERC) / Boston PSIpred - Various protein structure prediction methods at Brunel University SOPM (Geourjon and Deléage, 1994) SOPMA (Geourjon and Deléage, 1995) AGADIR - An algorithm to predict the helical content of peptides

28. 1D Prediction  Original methods: 1 sequence and uniform parameters (25-30%)  Original improvements: Parameters specific from protein classes  Present methods use sequence profiles obtained from multiple alignments and neural networks to extract parameters (70-75%, 98% for transmembrane helix)

31. PredictProtein (PHD) 1. Building of a multiple alignment using Swissprot, prosite, and domain databases 2. 1D prediction from the generated profile using neural networks 3. Fold recognition 4. Confidence evaluation

32. PredictProtein Available information  Signal peptides SignalP  O-glycosilation NetOglyc  Chloroplast import signal CloroP  Consensus secondary struc. JPRED  Transmembrane TMHMM, TOPPRED  SwissModel

33. Methods for remote homology  Homology can be recognized using PSI-Blast  Fold prediction is possible using threading methods  Acurate 3D prediction is not possible: No structure-function relationship can be inferred from models

34. Threading  Unknown sequence is “folded” in a number of known structures  Scoring functions evaluate the fitting between sequence and structure according to statistical functions and sequence comparison

35. ATTWV....PRKSCT.......... 10.55.2>.......... SELECTED HIT

36. ATTWV....PRKSCT Sequence HHHHH....CCBBBB Pred. Sec. Struc. eeebb....eeebeb Pred. accesibility.......... Sequence GGTV....ATTW........... ATTVL....FFRK Obs SS BBBB....CCHH........... HHHB.....CBCB Obs Acc. EEBE.....BBEB........... BBEBB....EBBE

37. Technical aspects  Alignment: Dynamic programming (Needleman & Wunsch, 1970)  Scoring Function: w seq.P seq + w str. (P SS + P AC ) P seq : Dayhoff matrix, P SS y P AC : probability model on pred. SS and AC P seq : Dayhoff matrix, P SS y P AC : probability model on pred. SS and AC

38. Threading accurancy

40. 3D-PSSM Steps  Building of 1D/superfamily profile  Building of 3D/superfamily profile  Determine/predict secondary structure and accesibility  Best score from 1. Structure vs. query PSSM 2. Query vs. 1D-PSSM structures 3. Query vs. 3D-PSSM structures

42. Comparative modelling  Good for homology >30%  Accurancy is very high for homology > 60%

43. Remainder  The model must be USEFUL Only the “interesting” regions of the protein need to be modelled

44. Expected accurancy  Strongly dependent on the quality of the sequence alignment  Strongly dependent on the identity with “template” structures. Very good structures if identity > 60-70%.  Quality of the model is better in the backbone than side chains  Quality of the model is better in conserved regions

46. Steps 1. Alignment of template structures 2. Alignment of unknown sequence against template alignment 3. Build structure of conserved regions (SCR) 4. Build of unconserved regions (“loops” usually)

47. Optimization 1. Optimize side chain conformation 1. Energy minimization restricted to standard conformers and VdW energy 2. Optimize everything Global energy minimization with restrains Molecular dynamics

49. Quality test  No energy differences between a correct or wrong model  The structure must by “chemically correct” to use it in quantitative predictions

50. Prediction software SwissModel (automatic)  http://www.expasy.org/swissmod/ SwissModel Repository  http://swissmodel.expasy.org/repository/ 3D-JIGSAW (M.Stenberg)  http://www.bmm.icnet.uk/servers/3djigsaw/ Modeller (A.Sali)  http://salilab.org/modeller/modeller.html MODBASE (A. Sali)  http://alto.compbio.ucsf.edu/modbase- cgi/index.cgi

53. Resultspdbv

54. Final test  The model must justify experimental data (i.e. differences between unknown sequence and templates) and be useful to understand function.