Macromolecular Structure Database group

1
Dimitris Dimitropoulos Chemistry the PDB MSDchem
2
The chemical database
3
MSDchem ligand dictionary

• Complete, clean, up to date collection of all
  the chemical species and small molecules in the
  PDB
• A ligand in MSDchem is a complete, distinct
  stereo isomer of a chemical compound
• Atoms and element types
• Bonds and bond orders
• Stereo configuration of atoms and bonds in cases
  of stereo-isomers (R/S E/Z)
• Atom names and coordinates are not fundamental
  properties

4
Role in the MSD database

• An integral component in the core of MSD
  database
• Relational reference from entities where a
  molecule or atom name is used in the PDB (protein
  residues and atoms)
• It is not possible for an ATOM line
• HETATM 4342 C2 PLA 86 14.227 11.195 -8.256
  1.00 67.95 C
• to be loaded if the PLA ligand is not defined
  or it does not include a C2 atom.

5
Chemistry and PDB

• Eliminate chemical inconsistencies from new PDB
  entries
• Structure and derived properties of a ligand
  apply automatically to residues and bound
  molecules that reference it
• The basic structure is carefully determined
  during curation, and a rich set of derived
  attributes is calculated for each ligand
• Graph isomorphism is being applied to check the
  consistency of the PDB, taking stereo-configuratio
  n into account
• Old legacy PDB entries are chemically
  corrected when loaded in the MSD database
• In thousands of cases errors are identified and
  corrected, involving most of them times
  inconsistent naming or different
  stereo-configuration
• Exchanged in cooperation with RCSB and the wwPDB

6
More than just the PDB codes

• All ligands are modelled as separate
  inter-related ligands and the appropriate one is
  referenced
• No distinction is made in the PDB between ribo-
  and deoxyribonucleotides (all are identified
  with the same residue name i.e., A, C, G, T, U,
  I)
• Modified nucleic acids are given as A etc
  regardless of modification
• No distinction between different topological
  variants (12 different variants can be found for
  HIS in PDB)

7
Derived information

• External scientific software (CACTVS, VEGA,
  CORINA, ACD-labs, CCP4, OELIB) together with in
  house development has been used to derive
• Stereochemistry (R/S E/Z)

DCM C4' S C3' R C1' S
DCF C4' R C3' S C1' R

• Smiles and detailed gifs
• Systematic IUPAC names

THIOALANINE (ALT) CC(N)C(O)S -
CC_at_H(N)C(O)S (2S)-2-aminopropanethioic O-acid
8
Derived information

• Fingerprints
• A bit string in hexadecimal form that indicates
  the presence or not of segments from predefined
  lists
• Useful for fast search and classification
• Different libraries of predefined lists can be
  set
• Currently calculated for the CACTVS library (500
  segments)
• Molecule Segments

BitString 1 0 1 0 1 0
Fingerprint 2A
9
(No Transcript)
10
Search options

• By ligand code
• By ligand name or synonym
• By formula or formula range
• By non stereo substructure
• By non stereo superstructure
• By exact stereo or non stereo structure
• By fingerprint similarity

11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
Click on EAA
Results of is superstructure of
15
EAA details
3-chloro-phenol
16
Results Viewers
17
PDB residue KWT
ltchemCompgt ltcodegtKWTlt/codegt ltnamegt(1S,6BR,9AS,11
R,11BR)-9A,11B-DIMETHYL-1-(METHYLOXY)METHYL-3,6,
9-TRIOXO-1,6,6B,7,8,9,9A,10,11,11B-DECAHYDRO-3H-FU
RO4,3,2-DEINDENO4,5-H2BENZOPYRAN-11-YL
ACETATElt/namegt ltnAtomsAllgt55lt/nAtomsAllgt
ltnAtomsNhgt31lt/nAtomsNhgt ltoverallChargegt0lt/overal
lChargegt ltstereoSmilesgtCOCC_at_H1OC(O)c2coc3C(O)
C4C(C_at__at_H(CC_at__at_5(C)C_at_H4CCC5O)OC(C)O)C_at_1(C
)c23lt/stereoSmilesgt ltsystematicNamegt(1S,6bR,9aS,11
R,11bR)-1-(methoxymethyl)-9a,11b-dimethyl-3,6,9-tr
ioxo-1,6,6b,7,8,9,9a,10,11,11b-decahydro-3H-furo4
,3,2-deindeno4,5-hisochromen-11-yl
acetatelt/systematicNamegt
18
Future targets

• Identify and model protein inhibitors as ligands
• Pre-classify functional groups for ligands and
  ligand atoms based on substructure fragments.
• Optimise and boost the performance of
  substructure searches
• Enhance visualisation and integration with other
  MSD tools