Israelاسرائيل

Theoretical Studies on Cytochrome P450cam

Walter Thiel

Max-Planck-Institut für Kohlenforschung

Symposium of The Lise Meitner - Minerva Center

Haifa, 19 December 2004

Happy birthday Yitzhak...

QM/MM approach: Overview

QMQM: density functional theory (B3LYP)

MMMM: force field (CHARMM22)

QM – MM interactionsQM – MM interactions:

„electronic embedding“

Border regionBorder region:

• hydrogen link atoms L

• charge shift for q(M1)

Codes:

ChemShell, Turbomole, DL-POLY

M1

M2

M2

M3

Q1

Q2

Q3

Q2

L

D. Bakowies and W. Thiel, J. Phys. Chem. 100, 10580 (1996).P. Sherwood at al, J. Mol. Theochem 632, 1 (2003).

J,A AJ

AJ

AJ

AJ

J,i AJ

AJ

J,i iJ

JO,IMMQM R

B

R

A

R

Zq

r

qH

612

ChemShell: A modular QM/MM package

GAUSSIAN98

GAMESS-UK

Chemshell

Tcl scripts

GROMOS96

CHARMm26MSI

Integratedroutines:

datamanagement

geometryoptimisation

moleculardynamics

genericforce fields

QM/MMcoupling

MNDO99

MOPAC

QM codes MM codes

DL_POLY

TURBOMOLE

CHARMM27academic

GULP

P. Sherwood et al, J. Mol. Struct. Theochem 632, 1-28 (2003).

MOLPRO

Cytochrome P450Cam (Pseudomonas Putida)

• heme protein, thiolato ligand

• completely buried active site

• soluble - extensively characterized by biochemical / biophysical techniques

• X-ray structures for various intermediates of the catalytic cycle

• natural substrate camphor, also other compounds

• biohydroxylation of non-activated C-H bonds

O

H5exo

H5endo

P450cam+ O2 + 2 e- + 2H+ + H2O

O

OH

H5endo

camphor

CYP450: Catalytic cycle

Mechanistic features:

• electrons from NADPH (2 3, 4 5)

• binding of molecular oxygen (3 4)

• active species 6 (Compound I) not observed experimentally

• hydroxylation mechanism 6 8 under dispute (rebound mechanism assumed)

N

N

N

N

Fe

COO

COO

N

N

N

N

Fe

FeIII

OH2

S

N

N

N

N

Cys

FeIII

S

N

N

N

N

Cys

R-H

FeII

S

N

N

N

N

Cys

R-H

FeII

S

N

N

N

N

Cys

R-HO

O

FeIII

S

N

N

N

N

Cys

R-HO

O

FeIV

S

N

N

N

N

Cys

R-HO

FeIII

S

N

N

N

N

Cys

ROH

FeIII

ROH

S

N

N

N

N

Cys

R-H

H2O

e

O2

e2-

2H+

H2O

Habs

H2O

ROH

1

2

3

4

5

6 (I)

7

8 2

3

4

5

6

7

8

1

Resting state

Initial coordinates: based on PDB structures 1DZ4 [1] and 1PHC [2]

[1] I. Schlichting, J. Berendzen, K. Chu, A. M. Stock, S. A. Maves, D. A. Benson, R. M. Sweet, D. Ringe, G. A. Petsko, S. G. Sligar, Science 287, 1615 (2000).

[2] T. L. Poulos, B. C. Finzel, A. J. Howard, Biochemistry 25, 5314 (1986).

P450cam: Iron(III)-aqua complex

Tyr96

Heme

Cys357

Resting state: QM/MM calculations

N N

NNFe

O

SR

MM

HNHN

C(H)

O

O

(H)

(H)

MM

MM

(H) MM

R1w: R =

R3w: R =

Leu356

Leu358

Cys357

H

H H

QM: UKS-DFT, B3LYPBasis: LACVP (ECP) + 6-31G (B1); ligand atoms: 6-31+G*, water protons: 6-31++G**(B2)

MM part: CHARMM22 force field

QM/MM: - electrostatic embedding scheme,

- hydrogen link atoms and charge shift model

QM regions: R1w (42 atoms), R3w (59 atoms)

J. C. Schöneboom and W. Thiel, J. Phys. Chem. B 108, 7468-7478 (2004).

Resting state: Comparison with experiment

B3LYP/CHARMM22 BLYP/CHARMM22 exp.

2A 4A 6A 2A 4A 6A

E(QM/MM) 0 2.37 3.30 0 8.45 18.27 2A

E(QM) 0 0.12 4.13 0 6.07 18.63 -

rFe–O [Å] 2.141 2.475 2.467 2.191 2.605 2.611 2.28(0.2)

rFe–S [Å] 2.269 2.491 2.418 2.239 2.490 2.421 2.25(0.2)

rFe–N [Å] 2.038 2.037 2.095 2.050 2.051 2.113 2.02(0.2)

rFe–H [Å] 2.633 2.928 2.905 2.655 2.980 2.961 2.62

Large QM region R3w, basis set B2:

Hyperfine coupling constants: Resting state

[1] D. Goldfarb, M. Bernardo, H. Thomann, P. M. H. Kroneck, V. Ullrich, JACS 118, 2686 (1996).[2] Y.-C. Fann, N. C. Gerber, P. A. Omulski, L. P. Hager, S. G. Sligar, B. M. Hoffman, JACS 116, 5989 (1994).

P450cam: Iron(III)-aqua complex B3LYP/CHARMM22

[MHz] Aiso Adx Ad

y Adz

H(1) 2.0 4.7 5.0 9.4

H(2) 0.4 4.6 4.4 9.0

[MHz] Atotx Atot

y Atotz

N(B) 6.8 6.1 5.9

N(A) 6.8 6.2 6.0

N(C) 6.7 5.8 5.2

N(D) 6.2 5.5 4.7

exp. [1] 1.5–2.0 4.2 – 4.5 8.4–9.0

exp. [2] 5.0 – 6.4

H(1) H(2)

Theoretical approaches to Compound I

Gas phase model calculations

• [FeO(porph)(SMe)] Antony et al. 1997; Green 1999: S character

• [FeO(porph)(SH)]

Harris & Loew 1998; Filatov et al. 1999:

A2u character*(FeO)

dxz dyza2u

p(S)

4A2u

*(FeO)

dxz dyz

a2u

p(S)

4S

N

N

N

N

Fe

mesoO

Fe

S

R

O

Fe

S

R

*(FeO) a2u

p(S)

p(S)

• Doublet and quartet states close in energy (2,4A2u normally below 2,4S)

• Electronic nature sensitive to substituent at sulfur (Ogliaro, Shaik et al. 2000)

QM region and basis set

N N

NNFe

O

SR

CH2 (H)MM

(H)MM

HN

HN

CH2

(H)MM

O

O

(H)MM

R1 =

R2 =

R3 =

R1

(SH)

R2

(SMe)

R3

(cys)

QM atoms 40 43 57

NAO B1 (LACVP,6-31G) 286 299 383

NAO B2 (LACVP, 6-31G, 6-31G*) 316 329 413

NAO B3 (LACVP, 6-31G, 6-31+G*) 340 353 437

QM/MM vs. isolated state (gas phase): Unpaired spin densities

0.000

0.100

0.200

0.300

0.400

0.500

0.600

0.700

0.800

sp

in d

en

sit

y

QM/MM QM (optimized)

QM (enzyme conformation)

(S)

gr(porph)

av(Cmeso )

av(N)

R1:SH

B3LYP,

basis B3

H-bond interactions within the Cys357 loop

X-ray structure and MD snapshots

[1] I. Schlichting et al, Science 287, 1615 (2000).

X-ray +) snapshot 29 ps snapshot 40 ps

snapshot 40 ps *) snapshot 50 ps superimposition

*) side chain of Gln360 manipulated

1)

H-bonds within Cys357 loop: N···S and N···O distances (Å)

• H-bonds with Leu358 and Gly359 conserved in all structures

• H-bonds involving Gln360 more flexible

• Conformation of H-bonds less optimal than assumed in previous (QM) model studies

backbone N-H ··· S(Cys357)

Leu358 Gly359 Gln360

side chain Gln360 amide

N-H ··· O=C(Cys357)

X-ray(1DZ9) 3.51 3.23 3.31 3.00

snapshot 29 3.46 3.32 3.78 4.97

snapshot 40 3.45 3.40 3.57 3.15

snapshot 40a 3.49 3.33 3.56 2.77

snapshot 50 3.48 3.35 3.78 4.22

a manipulated side chain

Unpaired spin densities for different snapshots

29 50 40 40, man.snapshot

gr(Porph)

(S)

0.70

0.65

0.25

0.20

RH-X (Å) snap 29 snap 50 snap 40 snap 40 man.

Gln360 backbone

N-H ··· S(Cys357)

3.81 3.65 3.47 3.43

side chain Gln360 amide

N-H ··· O=C(Cys357)

5.44 4.50 2.86 1.77

N N

NNFe

O

SR

HNHN

CH2

(H)

O

O(H)

R =IV

meso

AB

C D

Spectroscopic predictions - P450cam: Compound I

4Aprotein 0.64 0.13 -18 -6 +12isolated 1.33 0.09 -18 -6 +12

JEQ (57Fe) Aiso (57Fe) Adx,y (57Fe) Ad

z (57Fe)) [cm –1] [mm/s] [mm/s] [MHz] [MHz] [MHz]

2Aprotein -16 0.67 0.13 -29 -14 +23isolated -27 1.34 0.09 -29 -13 +23

Protein: B3LYP/CHARMM22, isolated: B3LYP, large flexible Fe basisOptimized geometries

[MHz] Aiso Adx Ad

y Adz

17Ooxo protein -21.5 -43.9 -40.6 +84.5

isolated -21.7 -43.5 -41.3 +84.8

14NBpyrr protein -2.1 +2.1 +1.5 -3.6

isolated -0.3 +1.2 +0.5 -1.6

1Hmeso protein +3.4 -1.6 -0.1 +1.7

isolated +1.9 -0.9 +0.1 +0.8

1Hcys- protein -7.6 -0.6 0.0 +0.6

isolated -17.0 -1.1 0.0 +1.1

N N

NNFe

O

SR

HNHN

CH2

(H)

O

O(H)

R =

IV

meso

AB

C D

2A StateProtein: B3LYP/CHARMM22, isolated: B3LYP, decontracted SVP basis for ligandsOptimized geometries

Spectroscopic predictions - P450cam: Compound I

Doublet-quartet splitting in Compound I : Method

• Ab initio MR-CI as implemented in the SORCI procedure [1]: spectroscopy oriented configuration interaction using the DDCI2 concept (difference-dedicated CI, Malrieu).

• Basis: Fe Wachters [14s11p6d3f]/(8s6p4d2f), O (oxo) aug-cc-p-VDZ, S and N TZVP, other atoms SV, 504 basis functions.

• Technical details for CI: Spin restricted RI-BP86 MOs for quartet, 491 active MOs, CAS (3,3) reference space with 8/1 CSFs for doublet/quartet, DDCI2 space of 961022/528664 CSFs with variational treatment of 643279/209828 CSFs for doublet/quartet.

• B3LYP: Unrestricted Kohn-Sham calculation, broken symmetry (BS) solution for doublet, high-spin (HS) solution for quartet, Heisenberg exchange coupling constant J from Yamaguchi formula [2].

[1] F. Neese, J. Chem. Phys. 119, 9428 (2003).[2] K Yamaguchi, F. Jensen, A. Dorigo and K. N. Houk, Chem. Phys. Lett. 149, 537 (1988).

Doublet-quartet splitting in Compound I : Results

Ab initio MR - CI :

UDFT (UB3LYP) :

QM/MM calculations with inclusion of protein environment at optimized UB3LYP/CHARMM geometries:

Ab initio MR - CI : J = -13 cm-1, |ΔE| = 39 cm-1

UDFT (UB3LYP) : J = -16 cm-1, |ΔE| = 48 cm-1

Doublet ground state, antiferromagnetic coupling

J. C. Schöneboom, F. Neese and W. Thiel, submitted

3

AEAEJ

24 )()(

BS

2

HS

2 SS

BSEHSEJ

)()(

J

J

Zero-field splitting in the quartet state of Compound I

• Contributions from direct electron-electron spin-spin coupling and spin-orbit coupling (SOC).

• SOC terms dominant in Compound I (Fe).

• Sum-over-states evaluation of SOC contributions from doublet, quartet, and sextet configurations at the DFT-BP86 level.

• Results in protein environment:D = 25 cm-1, E/D = 0.32

• Results in gas phase:D = 34 cm-1, E/D = 0.27

• Zero-field splitting D/3 affected by protein environment.

Spin Hamiltonian and eigenstates of Compound I

Computed parameters:J = -13 cm-1 (CI); DFe0 = 25 cm-1, EFe0 = 8 cm-1 (BP86).

Eigenstates (cm-1) : 0, 36, 58 <S2> values : 1.05, 3.56, 3.63

Population of lowest eigenstate:100% at 4 K, 54% at 77 K, 38% at 298 K.

Almost equal populations of three lowest eigenstates at 298 K.

2FeOy

2FeOxFeO3

22FeOzFeOPFeO SSESDSSJ2H ;;;

ˆˆˆˆˆˆ

Possible spin couplings in Compound I

Excited states of Compound I

• Lowest doublet/quartet pair from coupling between FeO triplet (SFeO = 1) and

porphyrin doublet (SP = ).

• Excited quartet/sextet pair from coupling between FeO quintet (SFeO = 2) and

porphyrin doublet (SP = ), in the protein 0.52 eV above the ground state with

J = -26 cm-1 (B3LYP).

• Low-lying quintet states in [FeO(NH3)4(H2O)]2+ confirmed by single-point ab

initio calculations at B3LYP optimized triplet geometry which yield the following relative energies of the quintet state:MR-CI between -0.11 and +0.07 eV depending on active space and basis, UCCSD 0.02 eV, UCCSD(T) 0.22 eV,B3LYP 0.75 eV.

2

1

2

1

Summary : Properties of Compound I

• Protein environment tunes electron and spin density destribution ("chameleon").

• Protein environment affects Mössbauer parameters (Fe) and hyperfine coupling constant (ligands).

• Ab initio and DFT calculations predict antiferromagnetic doublet ground state (40 - 50 cm-1 below quartet).

• Spin-orbit coupling causes large zero-field splitting in the quartet (D > 20 cm-1) and mixes the lowest doublet and quartet states such that the ground state has no well-defined spin multiplicity.

• There are low lying excited states which may give rise to multi-state reactivity.

Acknowledgement

Ahmet AltunIris AntesDirk BakowiesSalomon BilleterMarco BocolaHai LinNikolaj OtteJan SchöneboomHans Martin SennFrank TerstegenStephan ThielAlexander TurnerJingjing Zheng

Richard Catlow

Shimrit Cohen

Karl-Erich Jaeger

Christian Lennartz

Frank Neese

Manfred Reetz

Ansgar Schäfer

Sason Shaik

Paul Sherwood

Wilfred van Gunsteren

Support from

European Commission (ESPRIT/QUASI)

German-Israeli Foundation for Scientific Research

Spectroscopic properties: Formulas for J and HFC

)()34()( 1NNz

iso RPSNA

Isotropic Fermi contact coupling constant for nucleus N:

Electron–nucleus magnetic dipole coupling constant for nucleus N:

kllNNNNkklN

d rrrrPNA |)3(|)( 25 ρ

NeNeN ggP

Basis sets with high flexibility in core region required, e.g., standard basis with decontracted inner s-functions.

Second order spin-orbit contribution (only Fe) to HFC:

ao

aoaoaoaoao

oi

ioioioiooi

ANN

SO LLLLLLLLgS

NA,

31131,

,3113

1,

)(

4)(

A

jA

Aiij lrL |)(|Im1 ji

ij lL ||Im2 jAA

iij rlL ||Im 33

214 SS

EEJ BSHS

Heisenberg exchange coupling constant

Spectroscopic properties: Formulas for Mössbauer parameters

zz

yyxx

V

VV

2/12

31

2

1

zzQ VeQE

Field gradient tensor:

Asymmetry parameter:

Quadrupole splitting:

NC ND

NANB

FeIV

O

SCys

Isomer shift: )0()0()(5

400

220

SA

R

RReZZS

ba A )0(0

F. Neese, Inorg. Chim. Acta 337, 181 (2002).

)3(

|)3(|)(

25

25

ANANANNA

ANA

kllNNNNkkl

RRRRZ

rrrrNV

P

Spectroscopic properties: Model compounds

N N

NNFe

O

Ph

Ph

PhPhIV

L

J EQ (57Fe) Aiso (57Fe) Adx,y (57Fe) Ad

z (57Fe) [cm –1] [mm/s] [mm/s] [MHz] [MHz] [MHz]

B3LYP:L = - +56 2.21 0.08 -18 -6.5 +13.0L = H2O +26 1.16 0.11 -18 -5.9 +11.7

exp. [1]: [FeO(TDCPP)]+ +38 1.48 0.06 -18.3 -8.5 +17.0[FeO(TMP)]+ +43 1.62 0.08 -18.3 -9.2 +18.4

[1] D. Mandon, R. Weiss, K. Jayaraj, A. Gold, J. Terner, E. Bill, A. X. Trautwein, Inorg. Chem. 31, 4404 (1992).

Singly occupied orbitals of Compound I

Natural orbitals from spin unrestricted B3LYP calculations on

the quartet state with (A) and without (B) MM point charges.

Orbitals of (FeO)2+ motif in upper valence region

Computed for triplet ground state of [FeO(NH3)4(H2O)]2+ and

assigned under approximate C4v symmetry

Resting state: Electronic structure

d(xz)d(yz)

d(x2-y2)

d(xy)

d(z2)

N NNN

Fe

O

S R

H H

x

z

2A

d(xz)d(yz)

d(x2-y2)

d(xy)

d(z2)

4A

d(xz)d(yz)

d(x2-y2)

d(xy)

d(z2)

6A

*

(kcal/mol) QM/MM gas phase

D(2A) Q(4A) S(6A) D(2A) Q(4A) S(6A)

BLYP 0.00 3.94 19.75 0.00 10.90 21.80

B3LYP 0.00 -2.41 4.50 0.00 5.48 6.92

BH-LYP 0.00 -10.78 -18.05 0.00 -0.84 -13.50

Relative single point energies obtained with different density functionals.

Small QM region R1w, small basis set B1 (LACVP, 6-31G).

Geometries optimized at the B3LYP/B1 level (QM/MM: snapshot 195 ps).

Resting state: Relative energies

-4

-2

0

2

4

6

8

0 50 150 175 195 200

Snapshot [ps]

E

[kca

l/mol

]

E(4A-2A) =E(6A-2A) =

protein: QM/MM energyprotein: QM contribution to QM/MM energygas phase: QM energy

small QM region R1w, basis set B1

Different snapshots from MM-MD trajectory, QM/MM optimized

Resting state: Effect of the protein

Protein Gas phase

(*-dxz) / eV 0.848 1.057

PA(S) / e 16.452 16.202

EQM(4A-2A) / kcal mol-1 -2.41 5.48

Stabilization of the 4A state

- Kohn-Sham orbital energy difference

- Mulliken electron population at sulfur

- QM energy doublet-quartet gap

OH2

Fe

SR

pz(S)

d(xz) d(yz)

d(x2-y2)

d(xy)

d(z2)

Fe

*

*

*

pZ(S)

pZ(S)

gas phaseenzyme

Pentacoordinated ferric and ferrous complexes

• B3LYP relative energies (kcal/mol), R2/B2W, snapshot 93, for enzyme (QM/MM) and gas phase (QM).

• High-spin ground state for both complexes, in agreement with experiment.• Quintet of (3) with double occupancy of in enzyme (gas phase).)d(d 22 yxxz

Mössbauer spectrum of pentacoordinated ferrous complex

Isomer shift (mm/s), quadrupole splitting |EQ| (mm/s), and

asymmetry parameter .

QM/MM using B3LYP and a large uncontracted basis set at Fe

Quintet: Two electromers with double occupancy of dxz

Only the ground-state quintet electromer matches the experimental data

22 yxd

[1] P. M. Champion et al, Biochemistry 14, 4151 (1975).

X-ray structure 1DZ9Only one monomer from asymmetric unitTRIS buffer deletedPotassium ion included

H-atom positions constructed with CHARMM

1. Optimize crystallographic water H-atoms

2. Optimize all H-atoms

MM Setup

Add water layer of 16 Å thickness (InsightII, MSI)

Equilibrate inner 8 Å of water layer

Energy minimization, constrain backbone (100 kcal mol-1 Å-2) and sidechains (50 kcal mol-1 Å-2), scale down constraints every 60 steps by 0.65. Final GRMS: 0.04 kcal mol-1 Å-1

Molecular dynamics: 15 ps heating dynamics, 200 ps equilibration dynamics, T = 300K, integration step 1 fs, SHAKE

Heme, Cys357 and outer 8 Å of solvent layer fixed:

Minimize several snapshots from equilibration trajectory

Only MM calculations!

16965 atoms solvent24394 atoms total

Setup (Compound I)

QM/MM vs. isolated state (gas phase): Geometries

J. C. Schöneboom, H. Lin, N. Reuter, W. Thiel, S. Cohen, F. Ogliaro, S. Shaik, J. Am. Chem. Soc. 124, 8142 (2002).

R1(SH) / B3R1(SH) / B3 R3(cys) / B3R3(cys) / B3

44AA 22AA 44AA 22AA

protein (QM/MM) = B3LYP/CHARMM22isolated (QM) = B3LYP

protein 1.627 1.626isolated 1.624

1.622

1.626 1.6251.618 1.617

rFe–O [Å]

protein 2.560 2.589isolated 2.566

2.581

2.585 2.6092.678 2.697

rFe–S [Å]

protein 111.5 111.2isolated 97.5

97.5

112.0 111.8115.5 115.1

Fe–S – C [o]

QM/MM vs. isolated state (gas phase): QM energies

J. C. Schöneboom, H. Lin, N. Reuter, W. Thiel, S. Cohen, F. Ogliaro, and S. Shaik, JACS 124, 8142 (2002).

108.7 109.0 150.6 150.6

R1(SH) / B3R1(SH) / B3 R3(cys) / B3R3(cys) / B3

44AA 22AA 44AA 22AA

0 0 0 0QM/MM (protein)

QM (gas phase)

QM (protein)E [kcal/mol]

Evert

Ead

QM/MM = B3LYP/CHARMM22, QM = B3LYP

99.5 99.7 129.7 129.6

Energy differences E = E(4A2u) - E (2A2u)

Energy a Geometry b R3 (cys) c

QM/MM (p) QM/MM (p) +0.03 (+0.08)

QM (p) QM/MM (p) +0.04

QM (g) QM/MM (p) +0.06

QM (g) QM (g) +0.08

a p/g = evaluated with / without protein environment

b p/g = optimized for protein / gas phase

c Adiabatic (vertical) energy differences in kcal/mol

B3LYP, basis B3

QM/MM vs. isolated state (gas phase)

Proposed mechanisms for C–H hydroxylation by Compound I

Contradictory experimental findings:

• Product analysis, KIE- measurements:

rebound mechanism

• Radical clock experiments [1]:

apparent lifetimes of possible intermediates too short ( = 80 – 200 fs)

competing reaction channels, e.g. oxene insertion [2] ?

high-spin and low-spin states involved (two-state-reactivity) [3] ?

influence of protein pocket ?

FeIV

S

R

O

R H

FeIII

S

R

OR H

oxene insertion

concerted

FeIII

S

R

OHR

H-abstractionrebound step

stepwise or

effectively concerted, nonsynchronous

[1] M. Newcomb and P. H. Toy, Acc. Chem. Res. 33, 449 (2000).[2] M. Newcomb, M.-H. Le Tadi-Biadatti, D. L. Chestney, E. S. Roberts, and P. F. Hollenberg, JACS 117, 12085 (1995).[3] S. Shaik, M. Filatov, D. Schröder, and H. Schwarz, Chem. Eur. J. 4, 193 (1998).

Investigation of the rebound mechanism

QM subsystems Basis States Environment

R1pro B1:LACVP+ECP (Fe)

6-31G (others)

1.LS (2A)

2.HS (4A)

1.QM/MM (enzyme)

2.Gas phase

R3cam B4: B1 +

6-31+G* on ligand atoms,

C5, H5exo

1.LS (2A)

2.HS (4A)

1.QM/MM (enzyme)

2.Gas phase

N N

NNFe

O

S(H) MM

H

H

(H)

H(H)

(H)

(H)H

N N

NNFe

O

SR

HNHN

CH2

(H)MM

O

O(H)MM

H3C CH3

CH3O

H5exo

H5endo

R =

O

1

23

4

7

8 9

6

5

10

Fe

O

S-prot

R

FeIII

O

S-prot

H

Fe

O

S-prot

RHR

HYD TSR

FeIV

O

S-prot

RH

Cpd I

FeIII

O

S-prot

R H

PRODTSH

H

Rebound mechanism: Computational procedure

R1pro/B1:

• PES scan

• Full geometry optimizations of minima and saddle points along the reaction coordinate

• Finite-difference Hessian to characterize saddle points

R3cam/B4:

• Full geometry optimizations of minima and saddle points obtained from R1pro/B1 calculations

red: QM region R3cam

yellow: optimized MM atoms

green: fixed MM environment

Hydrogen abstraction: PES scan R1pro/B1

LS (2A) state

reference point:reactive complex

HS (4A) state

reference point:reactive complex

-2

3

8

13

18

2.69 2.54 2.39 2.24 2.09 1.94 1.79 1.64 1.49 1.34 1.19 1.04

R(O-H5exo) [A]

En

erg

y [k

cal/m

ol]

E(QM/MM) E(QM) E(MM)

-2

3

8

13

18

2.68 2.54 2.39 2.23 2.09 1.94 1.78 1.64 1.49 1.34 1.19 1.04

R(O-H5exo) [A]

En

erg

y [k

cal/m

ol]

E(QM/MM) E(QM) E(MM)

Transition state of hydrogen abstraction, 2A state, R1pro/B1

TSH

EA = 19.5 / 20.4 kcal/mol

( 2A / 4A )

Mechanism of C–H hydroxylation: Energy profile

J. C. Schöneboom, S. Cohen, H. Lin, S. Shaik, and W. Thiel, J. Am. Chem. Soc. 126, 4017-4034 (2004).

QM/MM geometry optimizations, R1pro/B1Two-state reactivity confirmed

O

Fe

R H

IV

L

O

Fe

RH

L

O

Fe

RH

L

O

Fe

R H

L

O

Fe

R H

L

III

Energy

4A2A,

4A

2A

20/ 2114/ 15

14/ 17

-38/ -43

[kcal/mol]

0/ 0

+ +

C–H hydroxylation: Energy profile in the gas phase

QM geometry optimizations, R1pro/B1

O

Fe

R H

IV

L

O

Fe

RH

L

O

Fe

RH

L

O

Fe

R H

L

O

Fe

R H

L

III

Energy

IV

4A2A,

4A2A

[kcal/mol]

0/ 0

20/ 20

12/ 13

-50/ -45

- / 16

Rebound barriers for quartet / doublet: MO diagram

N. Harris, S. Cohen, M. Filatov, F. Ogliaro, and S. Shaik, Angew. Chem. Int. Ed. 39, 2003 (2000).

Electronic situation during rebound step:

a) MO diagram

rebound barrier in HS state due to occupation of antibonding orbital

b) electron counting diagram

filling of „porphyrin hole“

Barriers for hydrogen abstraction: Model system

EA (kcal/mol), basis AE1 (508 basis functions)

BLYP BP86 B97 B3LYP PBE025.2 22.4 20.1 19.5 15.1

EA (kcal/mol), B3LYP functional: dependence on basis set (no. of basis functions)

SV (122) B1 (123) SVP (211) TZVP (271) B4 (186) AE1 (508) AE2 (892)25.4 22.7 21.7 21.4 19.5 19.5 19.2

Product release: Three minima

R. Davydov, T. M. Makris, V. Kofman, D. E. Werst, S. G. Sligar, and B. M. Hoffman, JACS 123, 1403 (2001).

4A state4PROD

rFe-O = 2.843 Å

Erel(R1/B1) = 0 kcal/mol

2A state2PROD1

rFe-O = 2.262 Å

Erel(R1/B1) = 4.87 kcal/mol

2A state2PROD2

rFe-O = 2.819 Å

Erel(R1/B1) = 5.18 kcal/mol

Product complex: X-ray vs. QM/MM (2PROD2)

[1] H. Li, N. Shakunthala, L. M. Havran, J. D. Winkler, and T. Poulos, JACS 117, 6297 (1995).

white: X-ray structure 1NOO [1]; yellow: B3LYP/CHARMM (R1/B1), 2PROD2

Cam-OH

Heme

Tyr96

Cys357

Tyr96

Cam-OH

Heme

Cys357

O

O

H

O

Hydroxylation mechanism: Summary of results

• Two-state reactivity

• Pre-organization of substrate within the pocket (linear arrangement C–H···O) lower entropic cost compared to gas phase

• Protein pocket essential for enantioselectivity in rebound step orientation of substrate through hydrogen bond to Tyr96

• Small influence of protein environment on abstraction barrier

• Protein influences relative stability of redox electromers (FeIII/FeIV) multi-state scenario ?

• Three product minima in agreement with EPR observations