Transcript Macromolecular Structures

Macromolecular Structures:
A User’s Perspective
Mike Word, Ph.D.
GlaxoSmithKline & Duke University Biochemistry
November, 2003
Rational drug design
what we want to be doing
4cox
Illustration by David Goodsell
Structure  Function
1aos
(urea cycle)
94% sequence identity
1dcn
(eye lens)
Structure not always unique
Prion
protein
SCOP classes



http://scop.berkeley.edu/
All alpha proteins (138)
All beta proteins (93)
Alpha and beta proteins (a/b) (97)
– Mainly parallel beta sheets (beta-alpha-beta units)






Alpha plus beta proteins (a+b) (184)
– Mainly antiparallel beta sheets (segregated alpha and beta regions)
Multi-domain proteins (28)
Membrane and cell surface proteins and
peptides (11)
Small proteins (54)
Coiled coil proteins (5)
Peptides (77)
?
~30%
Comparative Protein
Modeling


Aim - To gain structural insights for a
new protein sequence before
experimental elucidation takes place
Method - Extrapolation of the new
structure from that of related family
members
Alternative: ab initio (or de novo ) modeling
Sequence + theory  model
A range of techniques; mostly energy based
Very difficult to apply
Folds, families and motifs
Evolutionary patterns are
critical for successful
prediction of function
Fold
Assignment
Template
Selection
Alignment
Model
Building
Evaluation
Templates




Atomic coordinates from X-ray or NMR
Highest sequence homology
Relevant domain fragment
SWISS-MODEL “first approach”:
– Can the structure be modeled?
Fold
Assignment
Template
Selection
Alignment
Model
Building
Evaluation
Target to template
alignment


Should consider (2º) structure: domain
boundaries, motifs, location of loops,
active site residues, SS bonds...
Can’t recover from incorrect alignment!
Fold
Assignment
Template
Selection
Alignment
Model
Building
Evaluation
Comparative modeling
methods



Manual model
building
Satisfaction of
spatial restraints
Template based
fragment assembly
Fold
Assignment
Template
Selection
Alignment
Model
Building
Evaluation
Model Evaluation





Does the model match the template(s)?
Is the stereochemistry good? Energy ok?
Are amino acids in reasonable
environments?
What parts are conserved in the sequence
alignment?
What information can the model provide?
Fold
Assignment
Template
Selection
Alignment
Model
Building
Evaluation
All-Atom
SmallProbe
Contact
Surface
Analysis
Contact score:
score
=
e
–(gap/err)2
[van der Waals contacts]
dots
+ 4 Vol(Hbonds)
[hydrogen bonds]
- 10 Vol(Overlaps)
[atomic clashes]
Clash score:
clsc = number(clashes > 0.4Å)/1000 atoms
MolProbity
Structure
validation
server
 Add H’s,
analyze
contacts

http://kinemage.
biochem.duke.edu/
CASP


Critical Assessment of Techniques
for Protein Structure Prediction
Biannual contest to model proteins
of unknown structure
– While experimental structure
determination is still in progress
Evaluates manual to completely
automated structure prediction
 http://predictioncenter.llnl.gov

Acknowledgements

Richardson Lab:
Dave & Jane,
Laura Weston,
Ian Davis, Bryan
Arendall, Shuren
Wang, Jeremy
Block, Michael
Prisant, Simon
Lovell, Thomas
LaBean, Mike Zalis

GlaxoSmithKline
Protein
Bioinformatics:
Nicolas Guex,
Kristin Koretke
NIH GM15000

GlaxoSmithKline
