The Mechanism of the Hydride Transfer between Anabaena Tyr303Ser FNRrd/FNRox

and NADP+/H

A Combined Pre-Steady-State Kinetic/Ensemble-Averaged Transition State Theory

with Multidimensional Tunneling Study

The cyanobacterium Anabaena PCC 7119.FNR:NADP+

PDB 2BSA

Isaias Lans, José Ramón Peregrina and Milagros Medina Departamento de Bioquímica y Biología Molecular y Celular, and Institute of Biocomputation and Physics of Complex Systems (BIFI), Universidad de Zaragoza, E-50009, Zaragoza, Spain.

José M. Lluch, Mireia Garcia-Viloca and Àngels Gonzàlez-LafontDepartament de Química and Institut de Biotecnologia i de Biomedicina, Universitat Autònoma de Barcelona, 08193 Bellaterra (Barcelona), Spain.

Darmstadt September  05 - 09, 2010.

The cyanobacterium Anabaena PCC 7119 and its photosynthetic chain

2 Fdrd + NADP+ + H+ 2 Fdox + NADPH

FNR

Fdox + PSIrd Fdrd + PSIox

CO2 FixationN2

Metabolism

PSI

PSII

ATPase

Phycobilisoms

Cytb6f

Cytosol

2e- + H+

1e-

Fe2S2

NADP+

Electron Transfer Chains in Biological Systems

FAD

1

2

3

4

NADP+two Fdrd

molecules

+

H+

2 Fdrd + NADP+ + H+ 2 Fdox + NADPH

FNR

Hydride Transfer Mechanism

Reduction of NADP+ to NADPH is proposed to take place by hydride transfer from the N5 position of the flavin ring of FNR to the C4 position of the nicotinamide ring.

Tyr303 seems not to be involved in the hydride transfer process but preventsthe direct interaction between the flavin ring of FNR and the nicotinamide ringof NADP+. Tyr303 side-chain must be displaced from its position during hydride transfer.

A B

FAD

Y303

NADP+

NADP+

FAD

FNR(1GJR) Y303S FNR(2BSA)

N5-C4N7.78 Å

N5-C4N3.4 Å

N10-N1N4.6 Å

Rings30º

OBJECTIVES

A better understanding of how FNR enzyme works, that is, of how the hydride transfer process that FNR catalyzes takes place in

Tyr303Ser FNR

To shed light on the role of C-terminal Tyr303 in WT FNR

By means of two complementary studies:

Theoretical study based on EA-VTST/MT

Experimental stopped-flow pre-steady-state kinetic study

Lans, I. et al. J. Phys. Chem. B 2010, 114, 3368 – 3379

FNRrd + NADP+ FNRox + NADPH

kB C

[FNRrd-NADP+] [FNRox-NADPH]

Dead time

Dead time

CTC-2 CTC-1

WT FNR Time Dependent Observed Process

kHT

kHT-1

FNRrd + NADP+ FNRox + NADPH

kB C

[FNRrd-NADP+] [FNRox-NADPH]

Dead time

Dead time

CTC-2 CTC-1

Tyr303Ser FNR

kHT

Time Dependent Observed Process

T=279 KEXPERIMENTAL RESULTS

BUT: Tyr303 in the WT must be displaced for the hydride transfer to take place !!

Tyr303 substitution by a Ser practically deactivates the capacity of the enzyme to reduce NADP+

’The equilibrium mixture is displaced towards NADPH production, consistent with the physiological main role of the enzyme.

285 s-1

270 s-1

190 s-1

NADP+

BUT: The structural disposition in the mutant between the flavin and the nicotinamide rings seems favorable to produce the hydride transfer.

Anabaena Tyr303Ser FNR:NADP+ complex2BSA

OBJECTIVES of the theoretical approach

Provide very subtle details of the mechanism that could be unavailable with experimental techniques.

By means of a fully microscopical simulation of the hydride transferin the complete solvated FNR:NADP+ system.

Calculate the macroscopic rate constants and kinetic isotope effectsof the hydride transfer in FNR that could be compared with the

experimental data.

By means of Ensemble-Averaged Transition State Theory with Multidimensional Tunneling.

C7

C8

C9

C9a

C5aC6

N10

C4aN5

N1

C2

N3

C4

O4

O2

C7M

C8M

CH3

C5N

C6N C2N

C3NC4N

CH3

C7N

O7N

N7N

N1N

H H

Gas-phase model hydride-transfer reaction: stacking

AM1 fails to describe the dispersion energy contribution to stacking favoring instead intermolecular hydrogen-bonding.

At the MPWB1K/6-31+G(d,p)level lumiflavinand 1-methylnicotinamide are in a vertical stacked configuration.

lumiflavin (FADH-/FAD)1-methylnicotinamide(NADP+/NADPH)

MPWB1K/6-31+G(d,p) accounts betterfor stacking interactions than B3LYP (6 kcal/mol of energy difference in complexation energies).

Model of the biological system

Initial Cartesian coordinates PDB file 2BSA.Protonation states: PROPKA. H atoms added with HBUILD in CHARMM.System neutralized with 7 Na+

Cubic box of water molecules. Total number of atoms: 41896 atoms (4800 protein atoms)

Equilibration of the solvated protein/cofactor/coenzyme system

NPT molecular dynamics (MD) simulations with periodic boundary conditions (PBC) at 279 K and 1 atm. Particle Mesh Ewald method.

SOFTWARE: CHARMM35

CHARMM22 force field +TIP3P

1-D QM/MM Potential Energy Profile

COMPUTATIONAL DETAILS

2)(2

1 oRESDRESD zzkV

ABNR algorithmMobile part: all atoms within a sphere of 20 Å.Distinguished coordinate z.

QM H-

N

N

N-

NH

O

O

C

CHO

C

C

CH2

HO

HO

H

H

H

O

P

O

P

O

O

-O

-O O CH2O

HO OH

N

N

N

N

NH2

N+

NH2

O

O

OHHOPO2

-

O

PO2-

CH2O

HO OPO2H-

N

N

N N

NH2

1 2

34aC

O

4

10a

5

10

5a

9a

6

98

7

1' HH

H H 12

34

5

6

7

5'

H

HN5

FADH- NADP+

Total number of QM atoms: 58Frontier GHO atoms: 2Total number of MM atoms : 41838

Definition of the QM region (atoms in orange) used for the QM/MM calculations

FNRrd-NADP+

SP

FNRox-NADPH

1D-POTENTIAL ENERGY SURFACE: AM1(QM)/CHARMM22-TIP3P(MM)

The hydride transfer HT-1 in the complete solvated enzymatic system results in:

An endoergic process (7.8 kcal/mol) with a potential energy barrier of 36.7 kcal/mol.

The flavin and the nicotinamide rings are set out in a roughly parallel configuration.

RD RARA RDzRPAM1

C7

C8

C9

C9a

C5aC6

N10

C4aN5

N1

C2

N3

C4

O4

O2

C7M

C8M

CH3

C5N

C6N C2N

C3NC4N

CH3

C7N

O7N

N7N

N1N

H H

RPAM1

AM1cQM

RPAM1

MPWB1KcQM

DLcorr zVzVΔV

Dual level single-point energy correction

,

.

Low level RPAM1

AM1cQM zV

High Level RPAM1

MPWB1KcQM zV

NPT molecular dynamics (MD) simulations with periodic boundaryconditions (PBC) at 279 K and 1 atm. Particle Mesh Ewald method.

Potential of mean force (PMF) and classical free energy curve with the umbrella sampling technique along the reaction coordinate z and the weighted histogram analysis method (WHAM).

.

COMPUTATIONAL DETAILS

MPWB1K//AM1(QM)/CHARMM22-TIP3P(MM)

)exp( */

RT

zG

h

TkTk

QCQCactBMTVTSTEA

CHARMMRATE =CHARMM + POLYRATE

RPAM

AMcQM

RPAM

KMPWBcQM

CMAMCMDL zVzVzWzW 11

111

The effect of thermal and entropic contributions is small.

CLASSICAL POTENTIAL OF MEAN FORCE

34.9 kcal/mol 35.8 kcal/mol The introduction of the MPWB1K

energy correction only changes significantly the endergonicity of the reaction.

3.6 kcal/mol

14.9 kcal/mol

CMDLz *

= 0.04 Ả CMAMz 1*

= -0.07 Ả

The location of the transition state moves slightly towards products.

RA RDz

It is worth noting that the vibrational contribution lowers the relative freeenergy of the transition state around 2.5 kcal/mol.

QUASICLASSICAL POTENTIAL OF MEAN FORCE

33.3 kcal/mol

14.7 kcal/mol

05.0* QCz

zWzWzW vibCMQC

Quantized-vibration correction

))(max()( * zGzG CMact

CMCMact

HT-1 (physiological) 35.04 31.87

HT (reverse) 20.08 17.02

CMFTRR

CMCMCMact GzWzWzG ,,

RvibCM

FTRRCMQCQC

act WGzWzWzG ,,,

zGzG QCact

QCQCact max*

)exp( */

RT

zG

h

TkTk

QCQCactBMTVTSTEA

Quasiclassical activation

free energy profile

Classical mechanical activation free energy profile

Quantal vibrational corrections clearly reduce the free energy barriers. The quassiclassical free energy barrier for the physiological reaction HT-1 is almost twice as big as the corresponding value for the reverse reaction HT.

EA-VTST/MT rate constants: activation free energy barriers

(kcal/mol)

Structural analysis along the reaction coordinateN5

C4N

H

N10

N1N

3.4 Ả

4.6 Ả

2.75 ẢN5 – C4N

N10 –N1N

The N5 – hydride – C4N angle hasto approach to 180º, and the hydride donor and acceptor atoms have to come closer.

Deformation of the contact ion pair FNRrd-NADP+

and partial loss of π stacking interaction

An energy penalty that increases the free energy cost of the transition state

R

TS

R SP PZ1

Z2

Static-Secondary Zone approximation(Frozen Bath)

Average net transmission coefficient γ

For each variational transition state configuration For each variational transition state configuration i i at at zz**QCQC::

1) AM1/CHARMM22 SP location2) AM1/CHARMM22 MEP3) Dual-level MEP energy correction (interpolation using ISPE).4) Quassiclassical transmission factor Γi

5) Semiclassical transmission coefficient κi

i(T) = exp{-Gi}

Potential of mean force

Vi (s)for configuration i.

Reaction Coordinate z

Gi

tunnelingI(T)

Quassiclasical Transmission Factor Γ

Semiclassical transmission coefficient

= P quantum

P classical

Pcl(E)

E

1

VAG

Veff

Veff,*

Pqu(E) )(2 Ee

0qu dEeP E

=

s

Action integral along the tunneling path

s

s

sEsVE d)(2)( 21eff

1

1/2Pqu(E)

s< s>

0cl dEeP E

E

Tunneling path

• MEP

ZPE)()( MEPG

Aeff sVsVV

s

s

sEsVE d)(2)( 21eff

1

• Assuming vibrational adiabacity of F-1modes orthogonal to s

s

= 0 = 1

• Zero-Curvature (ZCT) tunneling

ERAB

RBC

Reaction-path curvature

Coupling of vibrations k to the reaction coordinate s

v

“corner cutting”

211

1

2

d

d)(

F

kk

TkK

BCurv

s

vLsB

Small-curvature (SCT):

s

s

sEsVE d)(2)(21

eff1

eff ,GAV

)exp( */

RT

zG

h

TkTk

QCQCactBMTVTSTEA

HT-1 31.87 0.998±0.003 335.70±142.79 334.82±142.44

HT 17.02 0.998±0.003 335.70±142.79 334.82±142.44

ii

EA-VTST/MT rate constants: dynamic effectsCHARMMRATE

Average net transmission coefficient

Quassical transmission factor

Semiclassical transmission coefficient: SCT

i ≡ configuration of VTSE

ii ii)( *zG QCQC

act(kcal/mol)

13.8092.8334.82±142.44335.70±142.790.998±0.00317.0220.08HT

28.642.2X10-10334.82±142.44335.70±142.790.998±0.00331.8735.04HT-1

The hydride transfer from Tyr303Ser FNRrd to NADP+ hardly occurs in agreementwith our stopped-flow kinetic measurements.

Conversely, the reverse reaction, HT, does happen, with a reaction rate constant of 92.8 s-1 in very good agreement with our stopped-flow kinetic measurements (kHT = 190 s-1).

EA-VTST/MT rate constants: ii)( *zG QCQC

actMTVTSTEAk /

ln*/ RTzGG QCQC

actMTVTSTEA

act

(s-1)

(kcal/mol)

Phenomenological free energy ofactivation

3.44 3.46 3.48 3.50 3.52 3.54 3.56 3.58 3.602.0

2.5

3.0

Ln

(KIE

)103/T

3.44 3.46 3.48 3.50 3.52 3.54 3.56 3.58 3.602

3

4

5

6

Ln

(ko

bs)

A

B

H

D

Hydride and deuteride transfer in the Tyr303Ser FNR:NADP+ reactant complex: Kinetic Isotopic Effect in the non-photosynthetic direction

-2 -1 0 1-5

0

5

10

15

20

25

30

35

PM

F(k

cal/

mo

l)

z(Å)

H/D

TUNNELHYDRIDE TRANSFER

Parameter Theoretical Experimental

kHT (279k) 92.8 s-1 190 s-1

kDT (279 K) 3.5 s-1 16.3 s-1

KIE 26.3 11.6

AH/AD12.5

FAD:NADPH

(C4-H)FADH-:NADP+ (N5-H)

QM/MMLn(kobs) = lnA - Ea/RT

10.73 kcal/mol

10.68 kcal/mol

Vibration-driven tunneling

Conclusions• The physiological hydride transfer from Tyr303Ser FNRrd to NADP+ is not possible. The reverse reaction, the hydride transfer from NADPH to Tyr303Ser

FNRox, does occur.

• The experimental and theoretical reverse reaction rate constants are in very good accordance.

• At the reactant region the N5-C4N distance might be compatible with the hydride transfer, but the N5-hydride-C4N angle is very far from collinearity, and therefore, the hydride shift is quite inefficient. In going from the reactant region to the transition-state region the N5-hydride-C4N angle approaches 180º to make the hydride transfer easier, and the hydride donor and acceptor atoms come closer.

• Since the width of the hydride transfer reaction path is small, the hydride transfer involves an important degree of quantum mechanical tunneling.

• The H/D KIE is essentially temperature-independent.

• All those geometric deformations, including a partial loss of the π stacking interaction, involve a large free energy penalty to reach the transition state. The difference between the direct and the reverse reaction rates comes from the important positive reaction free energy (15 kcal/mol). Such a difference must be due to an important stabilization of the close contact ionic pair FADH-:NADP+ versus the situation corresponding to the product, where, after the hydride transfer, the two moieties are neutral.

Any factor able to distort the formation of that close contact ionic pair would destabilize the reactant, so increasing the rate constant of the direct reaction. Very interestingly, the affinity for the

coenzyme in the WT FNR (Kd = 5.7 μM), is much lower than in the mutant Tyr303Ser FNR (Kd < 0.01 μM), this fact being consistent with the feasibility of the direct hydride transfer corresponding to the physiological main role of the enzyme. Could this lower affinity in the WT FNR be attributed to the presence of Tyr303 distorting in some way the formation of a close contact ionic pair as stable as in the Tyr303Ser mutant, with the corresponding thermodynamic and kinetic consequences?

Possible role of Tyr303 in WT FNR?

WT FNR

Tyr303Ser FNR

Chair Persons: Manuel YáñezOtilia MóCo-chair: Saulo Vázquez

-2011

Contact: watoc2011@uam.es

Catalytic mechanism of hydride transfer. Role of transient charge transfer interactions.

– charge-transfer complex• An electron-donor–electron-acceptor complex, characterized by

electronic transition(s) to an excited state in which there is a partial transfer of electronic charge from the donor to the acceptor moiety.

– charge-transfer (CT) transition• An electronic transition in which a large fraction of an electronic

charge is transferred from one region of a molecular entity, the electron donor, to another, the electron acceptor (intramolecular CT) or from one molecular entity to another (intermolecular CT).

Mulliken, R. S. (1952), J. Am. Chem. Soc. 74, 811l

PSI

NADPH

NADP+

electrons

FNRhq

FNRsq

FNRqnFdrd (Fldhq)

Fdox (Fldsq)

Transient Interactions in the photosynthetic electron transfer from PSI to NADP+

Corrección Vibracional

-4 -2 0 2

-4

-2

0

2

B

A

B Gauss Fit of B

-4 -2 0 2-3

-2

-1

0

1

C

A

C Gauss Fit of C

Corrección vibracional H Corrección vibracional D

Ajuste a una curva gaussiana

z

E(k

cal/m

ol)

z