1. Methods in 3D Structure Determination
David Wishart
June 2005
Methods in 3D Structure Determination
David Wishart
University of Alberta
Edmonton, AB
Lecture 3.1
(c) CGDN 2005

2. 3D Structure Review of Polypeptide Structure
Protein Structure Determination by X-ray
Protein Structure Determination by NMR
The Protein Data Bank
Lecture 3.1

3. Much Ado About Structure
Structure Function
Structure Mechanism
Structure Origins/Evolution
Structure-based Drug Design
Solving the Protein Folding Problem
Lecture 3.1

4. Protein Structures Are Complex
Lactate Dehydrogenase: Mixed a / b
Immunoglobulin Fold: b
Hemoglobin B Chain: a
Lecture 3.1

5. Proteins are Complex Average residue contains 8 “heavy” atoms
Average protein contains 300 amino acids
Average structure contains 2400 atoms
First structure (sperm whale myoglobin) took ~5 years with a team of ~15 key punch operators working around the clock to solve
Most structures still take 1 year to solve
Lecture 3.1

6. Solving Protein Structures
Only 2 kinds of techniques allow one to get atomic resolution pictures of macromolecules
X-ray Crystallography (first applied in Kendrew & Perutz)
NMR Spectroscopy (first applied in Ernst & Wuthrich)
Lecture 3.1

7. X-ray Crystallography
Lecture 3.1

8. X-ray Crystallography
Crystallization
Diffraction Apparatus
Diffraction Principles
Conversion of Diffraction Data to Electron Density
Resolution
Chain Tracing
Lecture 3.1

9. Crystallization
Protein Crystal
Lecture 3.1

10. Crystallization
Lecture 3.1

11. Diffraction Apparatus
Lecture 3.1

12. Diffraction Apparatus
Lecture 3.1

13. A Bigger Diffraction Apparatus
Synchrotron Light Source
Lecture 3.1

14. Diffraction Principles
nl = 2dsinq
Lecture 3.1

15. Diffraction Principles
Corresponding
Diffraction Pattern
A string of atoms
Lecture 3.1

16. Protein Crystal Diffraction
Diffraction Pattern
Lecture 3.1

17. Diffraction Apparatus
Lecture 3.1

18. Converting Diffraction Data to Electron Density
F T
Lecture 3.1

19. Fourier Transformation
i(xyz)(hkl)
F(x,y,z) = f(hkl)e d(hkl)
Converts from units of inverse space to cartesian coordinates
Lecture 3.1

20. Diffracting a Cat Diffraction data with phase information Real
Lecture 3.1

21. Reconstructing a Cat FT
Easy FT
Hard
Lecture 3.1

22. The Phase Problem
Diffraction data only records intensity, not phase information (half the information is missing)
To reconstruct the image properly you need to have the phases (even approx.)
Guess the phases (molecular replacement)
Search phase space (direct methods)
Bootstrap phases (isomorphous replacement)
Uses differing wavelengths (anomolous disp.)
Lecture 3.1

23. MAD & X-ray Crystallography
MAD (Multiwavelength Anomalous Dispersion
Requires synchrotron beam lines
Requires protein with multiple scattering centres (selenomethionine labeled)
Allows rapid phasing
Proteins can now be “solved” in just 1-2 days
Lecture 3.1

24. Resolution
1.2 Å
2 Å
3 Å
Lecture 3.1

25. Chain Tracing
Electron Chain Final
Density Trace Model
Lecture 3.1

26. Refinement
Lecture 3.1

27. Refinement R R = S(|Fo-Fc|)/S(Fo) Fc = calculated structure factor
# iterations R
R = S(|Fo-Fc|)/S(Fo)
Fc = calculated structure factor
Fo = observed structure factor
Lecture 3.1

28. The Final Result http://www-structure.llnl.gov/Xray/101index.html
ORIGX TRX 147
ORIGX TRX 148
SCALE TRX 149
SCALE TRX 150
SCALE TRX 151
ATOM N SER A TRX 152
ATOM CA SER A TRX 153
ATOM C SER A TRX 154
ATOM O SER A TRX 155
ATOM CB SER A TRX 156
ATOM OG SER A TRX 157
ATOM N ASP A TRX 158
ATOM CA ASP A TRX 159
ATOM C ASP A TRX 160
ATOM O ASP A TRX 161
ATOM CB ASP A TRX 162
Lecture 3.1

29. Bioinformatics & X-ray
Energy minimization and molecular dynamics (for structure refinement)
Structure quality assessment (threading, comparative geometry)
Structure comparison and classification (VAST, SCOP, CATH, CE)
Automated chain tracing (threading)
Domain, secondary structure, and supersecondary structure assignment
Lecture 3.1

30. NMR Spectroscopy
Radio Wave
Transceiver
Lecture 3.1

31. Principles of NMR
Measures nuclear magnetism or changes in nuclear magnetism in a molecule
NMR spectroscopy measures the absorption of light (radio waves) due to changes in nuclear spin orientation
NMR only occurs when a sample is in a strong magnetic field
Different nuclei absorb at different energies (frequencies)
Lecture 3.1

32. Principles of NMR
Lecture 3.1

33. Principles of NMR N N hn S S
Low Energy High Energy
Lecture 3.1

34. The Bell Analogy
CH3 NH CH
Lecture 3.1

35. FT NMR
Free Induction Decay FT
NMR spectrum
Lecture 3.1

36. Fourier Transformation
iwt
F(w) = f(t)e dt
Converts from units of time to units of frequency
Lecture 3.1

37. 1H NMR Spectra Exhibit...
Chemical Shifts (peaks at different frequencies or ppm values)
Splitting Patterns (from spin coupling)
Different Peak Intensities (# 1H)
Lecture 3.1

38. Protein NMR Spectrum
Lecture 3.1

39. 2D Gels & 2D NMR
SDS
PAGE
Lecture 3.1

40. Multidimensional NMR 1D 2D 3D MW ~ 500 MW ~ 10,000 MW ~ 30,000
Lecture 3.1

41. The NMR Process Obtain protein sequence Collect TOCSY & NOESY data
Use chemical shift tables and known sequence to assign TOCSY spectrum
Use TOCSY to assign NOESY spectrum
Obtain inter and intra-residue distance information from NOESY data
Feed data to computer to solve structure
Lecture 3.1

42. Multidimensional NMR
NOESY
TOCSY
Lecture 3.1

43. Assigning Chemical Shifts
Lecture 3.1

44. Measuring NOEs
Lecture 3.1

45. NMR Spectroscopy Chemical Shift Assignments NOE Intensities
J-Couplings
Distance
Geometry
Simulated
Annealing
Lecture 3.1

46. The Final Result ORIGX2 0.000000 1.000000 0.000000 0.00000 2TRX 147
SCALE TRX 149
SCALE TRX 150
SCALE TRX 151
ATOM N SER A TRX 152
ATOM CA SER A TRX 153
ATOM C SER A TRX 154
ATOM O SER A TRX 155
ATOM CB SER A TRX 156
ATOM OG SER A TRX 157
ATOM N ASP A TRX 158
ATOM CA ASP A TRX 159
ATOM C ASP A TRX 160
ATOM O ASP A TRX 161
ATOM CB ASP A TRX 162
Lecture 3.1

47. Bioinformatics & NMR
Energy minimization and molecular dynamics (for structure refinement)
Structure quality assessment (threading, comparative geometry)
Structure comparison and classification (VAST, SCOP, CATH, CE)
Automated assignment, fuzzy logic, chemical shift prediction
Lecture 3.1

48. X-ray Versus NMR X-ray NMR Producing enough protein for trials
Crystallization time and effort
Crystal quality, stability and size control
Finding isomorphous derivatives
Chain tracing & checking
Producing enough labeled protein for collection
Sample “conditioning”
Size of protein
Assignment process is slow and error prone
Measuring NOE’s is slow and error prone
Lecture 3.1

49. The PDB PDB - Protein Data Bank
Established in 1971 at Brookhaven National Lab (7 structures)
Primary archive for macromolecular structures (proteins, nucleic acids, carbohydrates)
Moved from BNL to RCSB (Research Collaboratory for Structural Bioinformatics) in 1998
Lecture 3.1

50.
The PDB
Lecture 3.1

51. The PDB
Contains coordinate data (primarily) from X-ray, NMR and modelling
Contains files in 2 formats
PDB format
mmCIF (macromolecular Crystallographic Information File Format)
Contains >21,000 entries
Currently growing exponentially
Lecture 3.1

52. PDB Growth
Lecture 3.1

53. Sequences vs. Structures
200000
180000
160000
140000
120000
Sequences
100000
Structures
80000
60000
40000
20000
Lecture 3.1

54. PDB Composition (2003)
Lecture 3.1

55. PDB Data Entry
Lecture 3.1

56. PDB File Format HEADER ELECTRON TRANSPORT 19-MAR-90 2TRX 2TRXA 1
COMPND THIOREDOXIN TRXA 2
SOURCE (ESCHERICHIA $COLI) TRX 4
AUTHOR S.K.KATTI,D.M.LE*MASTER,H.EKLUND TRX 5
REVDAT JAN-93 2TRXA HEADER COMPND TRXA 3
REVDAT OCT-91 2TRX TRX 6
JRNL AUTH S.K.KATTI,D.M.LE*MASTER,H.EKLUND TRX 7
JRNL TITL CRYSTAL STRUCTURE OF THIOREDOXIN FROM ESCHERICHIA 2TRX 8
JRNL TITL 2 $COLI AT 1.68 ANGSTROMS RESOLUTION TRX 9
JRNL REF J.MOL.BIOL V TRX 10
JRNL REFN ASTM JMOBAK UK ISSN TRX 11
REMARK 1
HEADER - PDB accession, date, function
CMPND - name of molecule or compound
SOURCE - origin or source of molecule (species)
REVDAT - revision dates
JRNL - primary reference (journal) describing structure
REMARK - a comment made by depositor
Lecture 3.1

57. PDB File Format REMARK - a comment made by depositor
REMARK 6 CORRECTION. CORRECT CLASSIFICATION ON HEADER RECORD AND TRXA 5
REMARK 6 REMOVE E.C. CODE. 15-JAN TRXA 6
SEQRES 1 A SER ASP LYS ILE ILE HIS LEU THR ASP ASP SER PHE ASP 2TRX 74
SEQRES 2 A THR ASP VAL LEU LYS ALA ASP GLY ALA ILE LEU VAL ASP 2TRX 75
SEQRES 3 A PHE TRP ALA GLU TRP CYS GLY PRO CYS LYS MET ILE ALA 2TRX 76
SEQRES 4 A PRO ILE LEU ASP GLU ILE ALA ASP GLU TYR GLN GLY LYS 2TRX 77
SEQRES 5 A LEU THR VAL ALA LYS LEU ASN ILE ASP GLN ASN PRO GLY 2TRX 78
SEQRES 6 A THR ALA PRO LYS TYR GLY ILE ARG GLY ILE PRO THR LEU 2TRX 79
SEQRES 7 A LEU LEU PHE LYS ASN GLY GLU VAL ALA ALA THR LYS VAL 2TRX 80
SEQRES 8 A GLY ALA LEU SER LYS GLY GLN LEU LYS GLU PHE LEU ASP 2TRX 81
SEQRES 9 A ALA ASN LEU ALA TRX 82
HET CU COPPER ++ ION TRX 100
HET CU COPPER ++ ION TRX 101
HET MPD METHYL-2,4-PENTANEDIOL TRX 102
HET MPD METHYL-2,4-PENTANEDIOL TRX 103
REMARK - a comment made by depositor
SEQRES - sequence of protein in 3 letter code
HET - names of heteroatoms
Lecture 3.1

58. PDB File Format FORMUL - chemical formula of heteroatoms
FORMUL CU 2(CU1 ++) TRX 110
FORMUL 4 MPD 8(C6 H14 O2) TRX 111
FORMUL 5 HOH *140(H2 O1) TRX 112
HELIX 1 A1A SER A LEU A DISORDERED IN MOLECULE B TRX 113
HELIX 2 A2A CYS A TYR A BENT BY 30 DEGREES AT RES TRX 114
HELIX 3 A3A ASN A ASN A TRX 115
HELIX A THR A TYR A DISTORTED H-BONDING C-TERMINS 2TRX 116
HELIX 5 A4A SER A LEU A TRX 117
SHEET 1 B1A 5 LYS A 3 THR A TRX 123
SHEET 2 B1A 5 LEU A 53 ASN A O VAL A N ILE A TRX 124
SHEET 3 B1A 5 GLY A 21 TRP A N TRP A O LEU A TRX 125
SHEET 4 B1A 5 PRO A 76 LYS A O THR A N PHE A TRX 126
SHEET 5 B1A 5 VAL A 86 GLY A N GLY A O LYS A TRX 127
SSBOND 1 CYS A CYS A TRX 143
FORMUL - chemical formula of heteroatoms
HELIX - location of helices as identified by depositor
SHEET location of beta sheets as identified by depositor
SSBOND - location and exisitence of disulfide bond
Lecture 3.1

59. PDB File Format
ORIGX TRX 146
ORIGX TRX 147
ORIGX TRX 148
SCALE TRX 149
SCALE TRX 150
SCALE TRX 151
ATOM N SER A TRX 152
ATOM CA SER A TRX 153
ATOM C SER A TRX 154
ATOM O SER A TRX 155
ATOM CB SER A TRX 156
ATOM OG SER A TRX 157
ATOM N ASP A TRX 158
ATOM CA ASP A TRX 159
ATOM C ASP A TRX 160
ATOM O ASP A TRX 161
ORIGXn - scaling factors to transform from orthogonal coords.
SCALEn - scaling factors to transform to fractional cryst. Coords.
ATOM - atomic coordinates of molecule
Lecture 3.1

60. PDB File Format Residue Name Atom # X coord (Å) Z coord (Å) B-factor
Atom Name
Residue #
Y coord (Å)
Occupancy
ATOM N SER A TRX 152
ATOM CA SER A TRX 153
ATOM C SER A TRX 154
ATOM O SER A TRX 155
ATOM CB SER A TRX 156
ATOM OG SER A TRX 157
ATOM N ASP A TRX 158
ATOM CA ASP A TRX 159
ATOM C ASP A TRX 160
ATOM O ASP A TRX 161
Lecture 3.1

61. PDB File Format Spacing is critical (Fortran compatible)
Often inconsistent (30+ years old)
Watch for unusual residues (ACE, SME)
Some files have 1 structure (X-ray), others have 2 structures (chain A and B in unit cell), others have >20 (NMR)
Some have missing atoms, others have hydrogens, others don’t
Lecture 3.1

62. Structure File Conversion
Alchemy (t) AMBER PREP (prep) Ball and Stick (bs)
Biosym .CAR (car) Boogie (boog) Cacao Cartesian (caccrt)
Cambridge CADPAC (cadpac) CHARMm (charmm) Chem3D Cartesian 1 (c3d1)
Chem3D Cartesian 2 (c3d2) CSD CSSR (cssr) CSD FDAT (fdat)
CSD GSTAT (c) Feature (feat) Free Form Fractional (f)
GAMESS Output (famout) Gaussian Z-Matrix (g) Gaussian Output (gauout)
Hyperchem (hin) MDL Isis (isis) Mac Molecule(macmol)
Macromodel (k) Micro World (micro) MM2 Input (mi)
MM2 Ouput (mo) MM3 (mm3) MMADS (mmads)
MDL MOLfile (mdl) MOLIN (molen) Mopac Cartesian (ac)
Mopac Internal (ai) Mopac Output (ao) PC Model (pc)
PDB (p) Quanta (quanta) ShelX (shelx)
Spartan (spar) Spartan Semi-Empirical (semi) Spartan Mol. Mechanics (spmm)
Sybyl Mol (mol) Sybyl Mol2 (mol2) Conjure (con)
Maccs 2d (maccs2) Maccs 3d (maccs3) UniChem XYZ (unixyz)
XYZ (x) XED (xed)
Lecture 3.1

63. Lecture 3.1

64. Lecture 3.1

65. Lecture 3.1

66. Lecture 3.1

67. Lecture 3.1

68. Conclusions
X-ray and NMR spectroscopy are both capable of determining 3D structures in a high throughput manner
X-ray is more powerful than NMR for coordinate determination, NMR is more powerful for understanding “processes” and motions
The Protein Databank (PDB) is the primary resource for protein structure information
Lecture 3.1