Electronic Properties of biomolecules: Theoretical studies of DNA in solution and biological environments

1
Electronic Properties of biomolecules
Theoretical studies of DNA in solution and
biological environments
CTBP

• D.L. Cox, Department of Physics, UC Davis
• Collaborations with J.C. Lin Thirumalai group
  at U. Md.), R.R.P. Singh (UCD), R.G. Endres
  (Imperial College), A. Huebsch, M.S. Swaroop and
  S.K. Pati (JNCASR Bangalore)
• Support NSF (Center for Theoretical Biological
  Physics and I2CAM), DOE
• Computer Support CTBP, JNCASR

2
Two short stories about electronic properties of
DNA in solvation environment
Au
Au
DNA water counterions Au Metallization of
Gs band gap engineering
DNA MutY repair protein Damage sensing for
repair?
3
Basic structure of biological (wet) B-DNA
4
Common themes

• Need to rationalize diverse set of data!
• Complexity of DNA (need for stabilization by
  water, counterions, plus fluctuations) ? not
  amenable to ab initio quantum MD
  (Carr-Parrinello)
• Use a combination of classical MD ab initio
  approximate (DFT) electronic structure to get at
• conformation dependence of electronic
  structure/tunneling
• contribution of solvation energy to
  electron transfer energetics
• range of conductance behaviors

5
Is the field alive and kicking?Some ISI evidence
?BB?AB?
?BB?AB?
?BB?AB?
Search on DNA and Electron/hole transfer
Search on DNA and Electronic structure
Search on DNA and Conduct

• For the last one good fit to exponential (R2
  0.99) with doubling time of 2 years assuming
  kickoff at year 0 AB (after Barton)

6
Why is it growing?

• Although DNA is unlikely to be used as a
  conductor itself (the mobility is low.), it
  remains a great tool for nanoscaffolding (Seeman,
  Mirkin, Kawai.) and when dressed up with other
  molecules or atoms it can be useful for optical
  and molecular electronic technologies (modified
  bases do increase mobility by 1-2 orders of
  mag-Kawai group)
• DNA does have the chance to be electronically
  active in its own right, unlike proteins, and
  represents a hydrogen atom for studying the
  role of emergent properties in conformational and
  atomic heterogeneity on conducting molecules in
  complex environments
• Rationalizing the wealth of data from a variety
  of experiments is a great and interesting
  intellectual challenge!

7
Nanostructures from DNAno controversy here! Use
hybridization as design control
3 and 4-way junctions (Niemeier lab)
Self Assembled Nanoparticle Networks (Mirkin
group)
Programmed Self Assembled Cube Structure (N.
Seeman lab)
8
So what kind of conductance can you get for DNA?
Not terrific, but at best close to Bechgaard
salts (review Endres, Cox, Singh, Rev. Mod.
Phys. 76, 195 2004)
High conductance
Semiconducting
?
Flatliners
Note in these latter flatline experiments,
emerging consensus is that DNA partially unwinds
and/or flattens, yielding Anderson localization.
9
K. Yoo
(biological)
NO CONTACTS! similar results in free hanging
bundles
10
So what is the best structure for conduction and
how does it depend upon water?
The Winner!
11
Whats up? Competition between p and s bonding
- near cancellation in A-DNA
12
Dependence upon twist and stretch(see also
recent work by Senthilkumar et al., JACS 2005)
??stack
Note 0s above measured to ambient twist
separation of DNA
13
Wetting your appetite influence of water on
conductance

• Expt exponential increase of
• conductance with humidity (Kleine-Ostmann et al,
  App. Phys. Lett. 2006)
• Theory evidence from combined QM/MM of water
  assisted hole conduction from waters in minor
  grooves linking oxygens of bases (Tsukamoto et
  al., Chem. Phys. Lett. 2007)

14
Short Story 1 Metallization of DNA by Au
electrodes and band gap engineering

• Interesting single molecule experiments II Xu
  et al, Nano Lett 2004 in water -- DNA metal???

15
Study again with AMBER SIESTA(Mallajosyula et
al., PRL 2008)

• Evolve DNA 10-mers with water and counter ions
  via AMBER8 (18 Na and 3000 TiP3P waters)
• Take average structure and prune waters and DNA
  to hexamers
• Attach to model Au electrodes (each 48 atoms)
  with thiol linkers (on hollow site of Au111)
• Carry out SIESTA PBE GGA functional with double
  zeta polarized for Au, P, counterions, double
  zeta for DNA bases, single zeta for water

16
Result DNA almost metallized by Au
With 0 Au Fermi Energy Gap 0.0006 ev Gap
0.05 eV Gap 0.03 eV Gap 0.4 eV
Homos are extended For GCn case AT
Intermediate breaks this
17
Simple picture - G is the most oxidizable base
(highest HOMO)
18
Further borne out by transmission and tunneling
estimate

• Surprise higher trans-
• mission through
• GGATGG than
• GCGCGC
• Depends upon detail
• of cross-strand hopping
• Using superexchange
• theory gives reason-
• able estimate of decay
• and transmission co-
• efficients (decay rate
• 0.54/angs. vs. expt.
• value of 0.42/angs.
• TGGATGG 1000 TGCATGC
• TGCGCGC 40 TGCATGC )

19
Role of water

• Stabilizes more highly conducting B-DNA structure
• Screens DNA and reduces oxidation potentials
  allowing proximity of G-levels to Au Fermi energy
  

20
Short Story II electrons in DNA damage sensing
for repair? JC Chin, DL Cox, RRP Singh
Biophysical J, 2008
21
Relevance to biology?
G-G hot spot

• Oxidative damage can lead to oxidized GG dimer.
  Damage site can be long distance from oxidation
  site (Barton et al Giese et al), via direct
  electron transfer (tunneling) at distances lt 20
  angstroms, electron hopping past that.
• Chemical attack can modify a G to an oxoG with
  extra O attached, which subsequently can mismatch
  with A on replication
• Intervening damage disrupts DNA
  conductance/damage/repair at a distance
  Numerous experiments by Barton group have
  illustrated this basic principle

22
On repair and damage proteins

• MutY glycosylase found in bacteria (e. coli)
  with homologues in yeast, mammals. Locates and
  excises As which are mismatched to 8-oxyguanines
  (oxidatively damaged Gs)
• Fe4S4 active cluster which is highly
  conserved-and remains intact-what is that for?

Structure of MutY monomer (Y. Guan et al, Nature
Struc. Biol. 5, 1058 (1998))
23
BIG QUESTIONS How do proteins locate damage
sites along DNA? Is Diffusion enough (Berg-von
Hippel)? Can there be remote sensing of damage
by use of electron transfer or migration
disrupted by lesions?

• Against other sensing models Diffusion may be
  enoughthere are lots of open questions (1d?
  Biased or nonbiased? 1d-3d combined? Time scales?
  Parallelization (lots of searchers)?
• Electronic detection (1) Protein-Protein redox
  couples , or (2) redox sensitive lesions.
• Redox modulation of search protein must slow in
  vicinity of binding site to facilitate
  recognition. Redox coupling could facilitate
  this.

24
Direct Evidence for electron assisted damage
recognition? (E.M. Boon et al, PNAS 100, 12543
(2003))

• Scenario
• Reduced MutY acts as transmitter (e- from Fe4S4
  cluster), oxidized MutY as receiver.
• Once reduced, MutY detaches.
• Damage blocks e- transmission and MutY processes
  to damage site, recruits repair complex
• Experimental Evidence current from MutY to end
  electrode, blocked by deliberate damage, altered
  by mutation at Fe-S site
• Theory order of magnitude or more enhancement of
  search rate (K.E. Ericksen, arXiv.orgq-bio.BM/031
  1033, preprint, Nov. 2003)

25
Theory Strategy

• For active regions of MutY (Fe-S cluster) and DNA
  (oxoG surrounding bases) use SIESTA based
  quantum mechanics to compute energy changes
• For passive regions, use AMBER MD to compute
  energy changes via free energy perturbation
  analysis (linear variable interpolating between
  MutY(2)-OxoG() to MutY(3)-OxoG(0)
• Add these contributions to get free energies of
  rearrangement and free energy differences -
  schematically
• ???????MD,tot - ??MD,in ??QM,in

26
A little math for the MD/QM

• Free energy perturbation
• H(?) (1-??HMutY()-OxoG()
  ?HMutY()-OxoG(0)
• Free energy difference
• ??GMD ?d????H?????(integral from 0 to 1)
• Reorganization energy
• ??????? (1/2)??H?????????????H????????
• Combination of energy differences
• ?G ?GMD,tot - ?GMD,in ?GQM,in
• ?????

27
Estimation of HDA

• Use the pathways algorithm of Beratan and Onuchic
  implemented through the HARLEM program
• HARLEM searches for optimal matrix element over
  all paths with the approximation

• ?????prefactor depending upon D-A bonds (energy
  units)
• ????c through covalent bond 0.6
• ?H through H-bond .36 e-1.7(R-2.8)
• ?S through solvent 0.6 e-1.7(R-1.4)
• Rbond separation in angstroms

28
Wild Type MutY-DNA

• Preference of electron transfer from MutY to
  oxidized oxoG enhances binding of 3 MutY in
  vicinity of oxoG
• Most probable rate from MD QM 2.1 x 106 sec-1

29
R149W mutation (kills MutY efficacy)

• R is right on optimal electron transfer
  pathway-losing hydrogen bond to DNA hurts HAD
• Estimate ketR149W/ketWT 1/8

30
L154F mutation

• More subtle - extra F size expands MutY and
  increases DA distance
• Factor of 2 decrease in optimal rates

31
L154F mutation

• More subtle - extra F size expands MutY and
  increases DA distance
• Factor of 2 decrease in optimal rates

32
Conclusion

• Preferential binding of MutY(3) in vicinity
• Of oxidized oxoguanine
• Enhanced Binding allows faster finding of damage
  site.

33
Summary

• Au can metallize G-rich n-mers explaining ohmic
  behavior of GCGC.. DNA AT insert induces
  tunneling
• Potential relevance of electron transfer in MutY
  damage detection
• References
• R.G. Endres, D.L. Cox, R.R.P. Singh, Rev. Mod.
  Phys. 76, 195 (2004)
• A. Huebsch, R.G. Endres, D.L. Cox, R.R.P. Singh,
  Phys. Rev. Lett. 94, 178102 (2005)
• R.G. Endres, D.L. Cox, R.R.P. Singh,
  cond-mat/0201404
• SS Mallajosyula et al. PRL 101 176805 (2008)
• JC Lin, DL Cox, RRP Singh Biophys J. 95,3259
  (2008)