Computational Redesign of Endonuclease DNA Binding and Cleavage Specificity

1
Computational Redesign of Endonuclease DNA
Binding and Cleavage Specificity

• Nature Vol 441 1 June 2006
• J Ashworth1, JJ Havranek1, CM Duarte1, D
  Sussman3, RJ Monnat2, BL Stoddard3, D Baker1
• 1Howard Hughes Medical Institute, Department of
  Biochemistry
• 2U. of Washington, Departments of Pathology and
  Genome Sciences
• 3Fred Hutchinson Cancer Research Center, Division
  of Basic Sciences
• Ahmet Bakan
• Department of Computational Biology
• University of Pittsburgh School of Medicine

2
Outline

• Introduction
• Protein-DNA interactions, studied case
• Methodology
• Computational model and procedure
• Experiments
• In vitro assay, X-ray
• Conclusion

3
Introduction protein-DNA
Hin recombinase DNA complex

• DNA-binding proteins
• Physiological role is determined by the affinity
  and specificity of the interaction
• Affinity and specificity depends on precise
  interactions between amino acids and nucleotides
• Redesign of biomolecules are important for
  biotechnology and medicine, e.g. gene therapy

NIH, NIDDKChiu, TK
4
Introduction protein-DNA

• Homing Endonucleases
• Generate strand breaks in cognate alleles
• Recognize long DNA target sites, so high binding
  specificity
• Constitute excellent model systems to understand
  the protein-DNA interactions
• In this paper, binding specificity of homing
  endonuclease I-MsoI is redesigned

5
Introduction I-MsoI
I-MsoI DNA complex

• Homing endonuclease I-MsoI
• Belongs to the LAGLIDADG family of homing
  endonucleases
• Homodimer of 170 residue monomers
• Cleaves 20-24 base pairs (bp) long target sites
• Structure reveals side-chain-nucleotide contacts
  within the DNA major groove

Base pair numbering

-
6
Methodology motivation

• Redesign of Biomolecules
• Requires realistic models and detailedknowledge
  of the specific interactionsin targeted complex
• Computational Tools
• Atomic-level force field
• Rotamer library for amino acid side chains
• Monte Carlo sampling algorithm

Pitt, Bio. Sci., Engler, L
H-bonding betweenamino acids and base pairs
7
Methodology model and procedure

• Energy model
• Packing and vdW LJ potential Solvation and
  electrostatics generalized Born model
  hydrogen bonding orientation dependent
• In silico screen
• Every single base pair substitution in DNA is
  modeled
• Monte Carlo search procedure is used to sample
  alternative amino acid identities and side-chain
  conformations
• Final complex structure is optimized using Powell
  method

8
Methodology test case
Wild-type blue and modeled yellow

• Test of Model
• Performed on IMso-I-DNA complex structure
• Ability of the model to reproduce side-chain
  conformations were tested
• Most side-chain dihedral angles were reproduced
• All direct h-bonding contacts were preserved

9
Methodology picking a model

• Specificity change plots
• Aim is to pick the redesigned enzyme and DNA pair
  that has the largest specificity distance to
  wild-type protein redesigned DNA complex
• Measure of distance is the predicted free energy
  chance upon bp substitution and amino acid
  mutation and optimization
• Energy change upon base-pair change
  versusEnergy change upon optimization of
  protein is plotted

10
Total Energy (- half of DNA)
Increasing specifity of redesigned enzyme to
redesigned DNA target site
-6G
-6A
Decreasing specifity of wild-type IMso-I to
redesigned DNA target site
11
Hydrogen Bonding Energy (- half of DNA)
Increasing specifity of redesigned enzyme to
redesigned DNA target site
-6A
-6G
Decreasing specifity of wild-type IMso-I to
redesigned DNA target site
12
Methodology selected model

• Selected complex model
• DNA base pair substitution-6CG and 6AT
  gtgtgt -6GC and 6CG
• Enzyme amino acid mutations28Lys and 38Thr gtgtgt
  28Leu and 38Arg

13
Methodology selected model

• Comparison of cognate and non-cognate binding
  complexes

14
Experiments in vitro assay

• In vitro competitive cleavage assay
• Relative activities of enzymes were determined on
  bothwild-type and designed target

Designed
Wild-type
Wild-type
Designed
15
Experiments in vitro assay

• In vitro competitive cleavage assay
• Wild-type enzyme favors WT site by gt
  27-foldDesigned enzyme favors designed site by gt
  25-fold
• Specificity is greater than 212-fold (gt4000)

27-fold
25-fold
16
Experimental Design x-ray
Refined x-ray structureDesigned structure

• X-ray crystallography
• Crystal structure ispredicted at 2 Å resolution
• Electron density mapand refined
  structuresuperimpose well with the predicted
  structure confirmingthe design accuracy

Electron density map Refined x-ray
structureDesigned structure
17
Conclusion

• Efficacy of atomic level model is shown in
  redesigning DNA-binding protein specificity
• The method is suggested to be applicable to any
  protein-DNA interface design
• The method can be improved to design protein and
  DNA backbone flexibility and water mediated
  interactions

18
Acknowledgements

• Thanks
• Dr. Bahar
• Dr. Cohen
• B. Lee

19
Homing Endonucleases