CONTENT •Doing Chemistry with Computers •Description of the tools - classical and quantum models - dynamics - QM/MM: computation in a complex envirnoment • Applications
Dec 21, 2015
CONTENT
•Doing Chemistry with Computers•Description of the tools
- classical and quantum models
- dynamics
- QM/MM: computation in a complex envirnoment• Applications
Doing Chemistry with ComputersDoing Chemistry with ComputersI.I.
•investigate unusual temperature/pressure regionsinvestigate unusual temperature/pressure regions•simulate dangerous experimentssimulate dangerous experiments•find alternative for hazardous chemicalsfind alternative for hazardous chemicals•ggain an atain an atoomistic description of mistic description of a reactiona reaction•save lab costssave lab costs
Why Do Computer Experiments?Why Do Computer Experiments?
Complement and Alternative to Lab ExperimentsComplement and Alternative to Lab Experiments
Understanding of Reaction Understanding of Reaction MechanismMechanism• characterize reactive intermediatescharacterize reactive intermediates
• identify rate determining or stereoselective stepsidentify rate determining or stereoselective steps
optimization of catalystsoptimization of catalysts rational drug designrational drug design
controlling and tuning of chemical reactionscontrolling and tuning of chemical reactions
Per
form
ance
Per
form
ance
(G
FL
OP
S)
(G
FL
OP
S)
0
100100
200200
300300
400400
500500
9494 9595 9696 9797 98989393
YearYear
COMPUTER PERFORMANCECOMPUTER PERFORMANCE(10th fastest computer)(10th fastest computer)
Cray Y-MPCray T3D
Fujitsu VPP
Cray T3E
diatomics triatomics
1970s1970s
Per
form
ance
Per
form
ance
(G
FL
OP
S)
(G
FL
OP
S)
0
100100
200200
300300
400400
500500
9494 9595 9696 9797 98989393
YearYear
COMPUTER PERFORMANCECOMPUTER PERFORMANCE(10th fastest computer)(10th fastest computer)
Cray Y-MPCray T3D
Fujitsu VPP
Cray T3E
aspirin
caffeineup to 30 atoms
main group elements
Early 1990sEarly 1990s
Per
form
ance
Per
form
ance
(G
FL
OP
S)
(G
FL
OP
S)
0
100100
200200
300300
400400
500500
9494 9595 9696 9797 98989393
YearYear
COMPUTER PERFORMANCECOMPUTER PERFORMANCE(10th fastest computer)(10th fastest computer)
Cray Y-MPCray T3D
Fujitsu VPP
Cray T3E
100-1000 of atoms
PresentPresent
heavy elementsdynamics
IBM 650IBM 650
CDC 7600CDC 7600
IBM7094IBM7094
CRAY Y-MPCRAY Y-MP
CDC 205CDC 205
Rel
ativ
e C
ost
per
MF
LO
PR
elat
ive
Co
st p
er M
FL
OP
Relative Cost of the Most Powerful Relative Cost of the Most Powerful Commercial ComputerCommercial Computer
SGI/CRAY T3ESGI/CRAY T3E
Computer experiments needComputer experiments needmodels and theories models and theories
to describe nature lawsto describe nature laws
with the language ofwith the language ofmathematicsmathematics
•environmental sciencesenvironmental sciences•biologybiology•chemistrychemistry•physicsphysics•……..
II.II.
When Newton meets Schrödinger...
maF H
Sir Isaac Newton(1642 - 1727)
Erwin Schrödinger(1887 - 1961)
Computational Chemistry and BiologyComputational Chemistry and Biology
Electronic StructureMethods
Classical MDSimulations
• parameter-free MD• ab initio force field• no transferability problem• chemical reactions
• improved optimization• finite T effects• thermodynamic & dynamic properties• solids & liquids
Computational Chemistry and BiologyComputational Chemistry and Biology
Electronic StructureMethods
Classical MDSimulations
Force field approach Ab-initio approach
Walter Kohn and John Pople
Nobelprize in chemistry 1998
Schrödingers equations made easy with DFT !
Traditional QCMethods
First-Principles Car-Parrinello
MD
Classical MDSimulations
When Quantum Chemistry Starts to Move...When Quantum Chemistry Starts to Move...
Mixed Quantum-Classical
Our needs for a virtual labOur needs for a virtual lab
•ElectronsElectrons
•Eq. Eq. oof Motionf Motion
ReactionsReactions
•AtomsAtoms
Density functional theory & Car-Parrinello Molecular Dynamics
Mixed Quantum-Classical in a complex environment - QM/MM
Main idea
Partitioning the system into
1. chemical active part treated by QM methods
2. Interface region
3. large environment that is modeled by a classical force field
QM
interface
Classical MM
Mixed Quantum-Classical in a complex environment - QM/MM
Main idea
Partitioning the system into
1. chemical active part treated by QM methods
2. Interface region
3. large environment that is modeled by a classical force field
QM
interface
Classical MM
APPLICATIONS
Phys.Rev.Lett. 72, 665 (1994)
III.III.
Improved Optimization Techniques:Improved Optimization Techniques:(simulated annealing)(simulated annealing)
Nanoscale Silicon Nanoscale Silicon ClustersClusters
Si4545
2C2Li H2
Li
Li
C C
H
H
Phys.Rev.Lett. 72, 665 (1994)
Li
Li
C CH
H
J. Am. Chem. Soc. 177, 42 (1995)
In SituIn Situ Simulation of Chemical Reactions Simulation of Chemical Reactions
Chem. Phys. Lett. 297, 205 (1998)
.OH + .NO2 ONOOH
Gas Phase
Aqueous Solution
J. Phys. Chem. A,104, 6464 (2000)
Cis/trans isomerization ONOOH
ONOOH + NO2-
HNO3 + NO2-
Aqueous Solution
ONOO- NO- + 1O2
PNAS 97 , 10307 (2000)
ONOO- + CO2 ?
In collaboration with W. Koppenol, ETH Zurich
PdSi
SiCl
Cl
Cl
Cl
Cl
Cl
PFe
Homogeneous CatalysisHomogeneous Catalysis
OrganoLithium
Pd-Phosphine Ta2-Hydride
Pd(ll)-bis(trichlorosilyl)
Re,Tc-Thioether
W2Cl2(PMe3)2(NHR)2
Structure Determination ofStructure Determination of
Collaboration with Prof. D. Tilley, University of California, Berkeley, U.S.A.
Ta Cp (Si(HPh)N(Ar)) - HTa Cp (Si(HPh)N(Ar)) - H22 22
**22 22
NMR suggests NMR suggests asymmetric Ta’sasymmetric Ta’s
TaH 11.63, -1.00 ppmTaH 11.63, -1.00 ppm
Lowest Energy Structure:Lowest Energy Structure:
Collaboration with Prof. D. Tilley, University of California, Berkeley, U.S.A.
• one bridged andone bridged and
• excellent agreementexcellent agreement
with Xray and NMRwith Xray and NMR
one terminal Hone terminal H
Excitation spectra of molecules in Excitation spectra of molecules in solution;solution;
Solvent Shift in AcetonSolvent Shift in Aceton
n -> *
Solvent Shift: Solvent Shift: 0.21 eV 0.21 eV (exp) (exp) 0.23eV 0.23eV (ROKS, QM = Solute) (ROKS, QM = Solute) = 0.03-0.04eV= 0.03-0.04eV (ROKS, QM = Solute (ROKS, QM = Solute + 12 H+ 12 H22O)O)
U. Röhrig, A. Laio, J. VandeVondele, J. Hutter, I. Frank, U.R. (in preparation)
Anti-AIDS: Anti-AIDS: HIV-1 ProteaseHIV-1 Protease
PrionsPrions and and
Metal IonsMetal Ions
DNA-Repair: DNA-Repair: Endonuclease IVEndonuclease IV
Photoisomerization Photoisomerization in Rhodopsinin Rhodopsin
Molecular MechanismsMolecular Mechanismsof Apoptosis:Caspase-3of Apoptosis:Caspase-3
Selectivity of Selectivity of KcsA ChannelKcsA Channel
Ab initio Modelling of EnzymesAb initio Modelling of Enzymes Rational Design of Biomimetics & Enzyme-EngineeringRational Design of Biomimetics & Enzyme-Engineering
Biomimetics• easy preparation• easy handling• easy tuning
Engineering• inhibitors• metal centers• new residues
Modelling &
Understanding
Rational Design of Biomimetics
Galactose Oxidase
O
N
O
NCu
R5
R3R3
R5
Synthetic Compound
t i R3 = SPh, SPr , Bu , BrR5 = Bu , Brt
R- H2 C - OH + O2 R-C + H2O2
O
H=
\Stack et al., Science (1999)
QM/MM Hybrid Car-Parrinello Modeling of GOase
Cu
Tyr495
Cys228
His496
U.R, P. Carloni, K. Doclo and M. Parrinello JBIC 5, 236 (2000)
Parallel Modeling of the Catalytic Cycle
resting stateresting stateinactiveinactive
GOaseGOase MimicMimic GOaseGOase
resting stateresting stateactiveactive
MimicMimic
after protonafter protontransfertransfer
transition statetransition state H-abstractionH-abstraction
16 kcal/mol16 kcal/mol
21 kcal/mol21 kcal/mol
Biomimetic
Goase
U.R, P. Carloni Intl. J. Quant. Chem. 73, 209 (1999)
GOaseMimeticStack
New BiomimeticsM1: 16 kcal/molM2: 16 kcal/molM3: 18 kcal/molM4: 14 kcal/mol
16 kcal/mol 21 kcal/mol
HIV- Virus (AIDS)
HIV- I Protease
Asp25Asp25
Asp25’Asp25’Gly27Gly27
Gly27’Gly27’
Thr26Thr26
Thr26’Thr26’
HIV-PR is essential for the formation of infective viruses
Immature, non-infective Immature, non-infective Viral particlesViral particles
HIV-1 PRHIV-1 PR
Infective virusesInfective viruses
Viewing Enzymes at work
HIV- I Protease
Prion ProteinsPrion Proteins
Fatal NeurodegenerativeFatal NeurodegenerativeDiseases:Diseases: Mad Cow Disease (BSE)Mad Cow Disease (BSE)• ScrapieScrapie• Creutzfeldt-Jakob Creutzfeldt-Jakob • caused by abnormalcaused by abnormal isoform PrP(Sc)isoform PrP(Sc)Human Prion ProteinHuman Prion Protein
(Wuthrich et al. PNAS 97, 145 (2000))(Wuthrich et al. PNAS 97, 145 (2000))
http:\\ www.mad-cow.org
Localization of Possible Binding Sites via Localization of Possible Binding Sites via a Parallel Statistical and QM Approach a Parallel Statistical and QM Approach
• 111 PDB structures 111 PDB structures 2.0 Å resolution2.0 Å resolution• 216 copper binding sites216 copper binding sites• 928 donor atoms 928 donor atoms
0
10
20
30
40
50
60
70
H C M G D Y E Q S
Cu-coordinatedCu-coordinatedNatural Natural
abundanceabundance
His 187
Met 206
Tyr 157
R = 24
His 140
Asp 147
Asp 144
R = 21His 140
Asp 147
Asp 144
R = 21
His 140
Asp 147
Asp 144
R = 21
Secondary structure changes inducedSecondary structure changes inducedby external factors by external factors
(pH, temperature, [Cu++]) (pH, temperature, [Cu++])
Method: Enhanced sampling techniques
Metal Ion / DNA Interactions
Cis-Pt anticancer drugs
Metal Ion / DNA Interactions
Cis-Pt anticancer drugs
A piano chair to fight cancer
Organoruthenium anticancer drugs(in collaboration with
Prof. Paul Dyson, EPFL)
Cis/Trans Photoisomerisation in Rhodopsin:Cis/Trans Photoisomerisation in Rhodopsin:The First Steps of VisionThe First Steps of Vision
200fs0.67
Cis/Trans Photoisomerisation in Rhodopsin:Cis/Trans Photoisomerisation in Rhodopsin:The First Steps of VisionThe First Steps of Vision
200fs0.67
10ns classical MD simulations10ns classical MD simulations
RMS backbone: 0.9ÅRMS backbone: 0.9Åtotal # of atoms: 24000total # of atoms: 24000
Photoisomerisation in the Excited StatePhotoisomerisation in the Excited State
Dynamics in the first excited singlet stateDynamics in the first excited singlet state
(in collaboration with I. Frank, Univ. Munich, C. Molteni, (in collaboration with I. Frank, Univ. Munich, C. Molteni, Univ. Cambridge, M. Parrinello, CSCS Manno)Univ. Cambridge, M. Parrinello, CSCS Manno)
Not all chemists wear white coats...Not all chemists wear white coats...
Computer ExperimentsComputer Experiments• provide atomistic picture of (bio)chemical systemsprovide atomistic picture of (bio)chemical systems• help to characterize and understand reaction mechanismshelp to characterize and understand reaction mechanisms
planning of laboratory experimentsplanning of laboratory experiments computational modelling of catalysts and enzymescomputational modelling of catalysts and enzymes rational design of drugs and biomimeticsrational design of drugs and biomimetics
Current Limits and Future PerspectivesCurrent Limits and Future Perspectives• accuracy of electronic structure method• system size• limited time scale
improved QM/MM methodsimproved QM/MM methods long time scale techniqueslong time scale techniques