EXPERIMENTAL AND COMPUTATIONAL
STUDIES OF THE CATALYTIC MECHANISM OF
YEAST CYTOSINE DEAMINASE
BY
JIFENG WANG

A DISSERTATION
Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Biochemistry and Molecular Biology
2011

ABSTRACT
EXPERIMENTAL AND COMPUTATIONAL
STUDIES OF THE CATALYTIC MECHANISM OF
YEAST CYTOSINE DEAMINASE
By
Jifeng Wang

Yeast cytosine deaminase (yCD) catalyzes the conversion of the prodrug 5fluorocytosine (5FC) to 5-fluorouracil (5FU), which is a widely used drug in the
treatment of cancers. yCD is of significant biomedical interest because of its potential
application in combination with 5FC in gene-directed enzyme prodrug therapy (GDEPT).
In this thesis, the catalytic mechanism of yCD was elucidated by both
experimental and computational methods. First, the transition state structure was accessed
experimentally by kinetic isotope effects (KIEs) studies at natural abundance by

13

C

nuclear magnetic resonance (NMR) spectroscopy. The results indicate that the reaction
proceeds by a stepwise AN+DN mechanism, with the transition state in the C4-N4 bond
cleavage step. Second, the catalytic mechanism of the deamination of 5-fluorocytosine
(5FC) was investigated by an ONIOM (B3LYP/6-31G*:Amber)-electronic embedding
scheme and was compared side-by-side with that of cytosine. The higher catalytic
efficiency (kcat/Km) for 5FC is likely due to the energetically more favorable
intermediates formed during the reaction. The method alone was not suitable to elucidate
Zn

the O -Zn bond breakage because relatively large rearrangements of the active site

residues are required. Hence, Quantum Mechanics/Molecular Mechanics (QM/MM) MD
simulations were set up to incorporate configurational sampling into chemical
transformation. In the reaction pathway proposed, it is the C4-N4 bond cleavage step, not
Zn

the O -Zn bond cleavage step, that is the rate-limiting step for the overall reaction.
Functions of conserved residues around the active site were investigated by site-directed
mutagenesis, biochemical, NMR, and computational methods. Glu64 is found to be
essential in catalysis due to its role as a general base/acid. Steady-state kinetic studies
show that mutating Glu64 to Ala causes a dramatic decrease in kcat by a factor of >3000.
Glu64 also participates in the stabilization of the intermediates and transition state,
suggested by

19

F NMR experiments and ONIOM calculations. Asn51 and Asp155

interact with the substrate through hydrogen bonding, by which the substrate is anchored
in the right position during the initial binding and at the transition state. These residues
are important for substrate binding and catalysis.
yCD belongs to the cytidine deaminase (CDA) family of purine/pyrimidine
deaminases. Enzymes in this family have similar three-dimensional structures with a
zinc-containing catalytic apparatus. Studies on yCD not only elucidate the enzymatic
mechanism of yCD itself, but also provide significant insights into the catalytic
mechanism of enzymes in this family. Moreover, these findings can facilitate
antimicrobial drug design and gene directed therapy against cancers.

Copyright by
JIFENG WANG
2011

To Confucius (551 B.C. – 479 B.C.)

v

ACKNOWLEDGEMENTS

Am I the first PhD ever in my house? Errrr… no. But I am proud to be the second
one, in a BIG family of three. Hei, son, you need to catch up faster.
Well, I am lucky that my son still regards me as his “best friend” although I could
not play with him too often since he was born.
But why he likes me more? (This line is invisible to the wife)
Because I always bought toys for him? I am not sure. Hei, son, if you get a chance
to read this, come and tell me. But I will tell you this: It is you mom who took care of you
when I was not around. Let’s show our great appreciations to you mom, my best friend!
Of course, I would like to thank my own mom. And dad. Every little success,
from my first step, to this thesis, has endless supports and encouragements from them.
I want to express my sincere appreciations to my advisor, Dr. Honggao Yan, for
his guidance in my PhD studies. Dr. Yan gave me a lot of freedom to construct my
research and taught me to think independently. I felt more and more confident over these
years. The time spent in Room 313 of the Biochemistry building definitely prepared me
well to be a good researcher in the future.
Well, Room 313 is not the only room I frequents. In the conference room, I met
with my Guidance Committee regularly, from whom, I received great advices and critical
reviews. I want to thank all members: Dr. Zachary F. Burton, Dr. Robert I. Cukier, Dr.
Robert P. Hausinger, Dr. Aizhuo Liu, and Dr. Claire Vieille. I want to thank Dr. Michael
Feig and Dr. William J. Wedemeyer too for their supports and guidance.

vi

Room 313 is and will always be a special place for me. During these years, people
working in this room are very supportive with each other. I learnt a little bit this or that
from different person. I want to express my gratitude to my people: Dr. Yue Li, Yan Wu,
Dr. Lishan Yao, Dr. Zhenwei Lu, Dr. Yi Wang, Dr. Jaroslaw Blaszczyk, Dr. Bin Chen,
Rahul Banerjee, Yichen Tang and David Achila. It has been a lot of fun to work with
you!

vii

TABLE OF CONTENTS

LIST OF TABLES .................................................................................... x
LIST OF FIGURES ..................................................................................xi
ABBREVIATIONS ................................................................................. xv
Chapter 1. Introduction ............................................................................ 1
Biomedical Significance and Basic Kinetic Properties of Cytosine Deaminase ...............1
X-ray Crystallographic Analysis of yCD .......................................................................3
Reaction Pathway by ONIOM Calculations....................................................................5
Zinc-uracil Bond Breakage Calculations.........................................................................6
Product Release is Rate-limiting in the Activation of 5FC by yCD .................................8
yCD Dynamics ...............................................................................................................8
Cytosine Dissociation Pathway by MD Simulations .......................................................9
The Scope of this Dissertation ........................................................................................9
Figures ......................................................................................................................... 15
References.................................................................................................................... 18

Chapter 2. Probing the Transition State of Yeast Cytosine Deaminase
13
by NMR Measurement of C Kinetic Isotope Effects ........................... 24
Introduction.................................................................................................................. 24
Materials and Methods ................................................................................................. 26
Results and Discussion ................................................................................................. 32
Conclusions.................................................................................................................. 39
Tables and Figures........................................................................................................ 41
References.................................................................................................................... 56

Chapter 3. Reaction Pathway of the Deamination of 5FC: ONIOM
Calculation Revisited ............................................................................... 62
Introduction.................................................................................................................. 62
Computational Methods ............................................................................................... 64
Results and Discussion ................................................................................................. 65
Conclusions.................................................................................................................. 76
Tables and Figures........................................................................................................ 79
References.................................................................................................................... 91

Chapter 4. QM/MM MD Simulations Shed New Insight on the Catalytic
Mechanism................................................................................................ 94
Introduction...................................................................................................................94
viii

Experimental Procedures...............................................................................................96
Results and Discussion ................................................................................................100
Concluding Remarks ...................................................................................................113
Tables and Figures.......................................................................................................117
References...................................................................................................................134

Chapter 5. Role of Glutamate 64 in the Activation of the Prodrug 5FC
................................................................................................................. 139
Introduction.................................................................................................................139
Experimental Procedures.............................................................................................141
Results ........................................................................................................................147
Discussion...................................................................................................................156
Tables and Figures.......................................................................................................162
References...................................................................................................................172

Chapter 6. Contributions of Conserved Residues in the Active Site to
Substrate Binding and Catalysis ........................................................... 176
Introduction.................................................................................................................176
Experimental Procedures .............................................................................................178
Results ........................................................................................................................184
Discussion...................................................................................................................193
Conclusions.................................................................................................................196
Tables and Figures.......................................................................................................198
References...................................................................................................................204

Chapter 7. Summary and Future Work ............................................... 207
Summary.....................................................................................................................207
Future Work................................................................................................................209

ix

LIST OF TABLES

Table 2-1. Reaction extents for yCD catalyzed reaction.................................................41
Table 2-2. Intensities of the carbon peaks in the NMR samples for the original cytosine41
Table 2-3. Intensities of the carbon peaks in the NMR samples for the residual cytosine
after deamination by yCD..............................................................................................41
Table 2-4. Experimental
Table 2-5. Calculated

13

13

C KIEs.................................................................................42

C KIEs.....................................................................................42

Table 2-6. Bond distances for cytosine and the transition states .....................................43
Table 3-1. Comparison of key coordinates of structures from ONIOM calculations with
crystal structures............................................................................................................79
Table 4-1. Comparison of changes of key coordinates during the conversion of cytosine
to uracil with crystal structures ................................................................................... 116
Table 5-1. Kinetic constants of the wild-type yCD and the E64 mutant enzymes for the
activation of 5FC........................................................................................................ 162
Table 5-2. Inhibitory and binding properties of 5FPy.................................................. 162
Table 6-1. Kinetic constants of the wild type yCD and mutant enzymes for the activation
of 5-fluorocytosine ..................................................................................................... 198
Table 6-2. Binding constants of 5FPy with the wild type yCD and mutant enzymes ... 198
Table 6-3. Internal energy and solvation energy components for the interactions between
residues and Py........................................................................................................... 199
Table 6-4. Individual energy components by MM-GBSA energy decomposition ........ 199

x

LIST OF FIGURES
Figure 1-1. Reaction catalyzed by yCD..........................................................................15
Figure 1-2. Crystal structure of yCD..............................................................................16
Figure 1-3. Interactions between the inhibitor DHP and surrounding residues in the active
site of yCD ....................................................................................................................17
Figure 2-1. Proposed reaction pathway for yCD-catalyzed deamination of cytosine to
uracil .............................................................................................................................44
Figure 2-2. Experimental KIEs mapped on the cytosine structure ..................................45
Figure 2-3. Possible transition state structures ...............................................................46
Figure 2-4. Model system for ONIOM calculations .......................................................47
Figure 2-5. ONIOM optimized transition state structure in the active site of yCD..........48
Figure 2-6. Molecular electrostatic potential (MEP) surfaces for cytosine, tetrahedral
intermediate, uracil, Py, DHP, and three transition states ...............................................49
Figure 2-7. Proposed reaction mechanism based on KIEs ..............................................50
Figure 2-8. Coordinates of optimized cytosine...............................................................51
Figure 2-9. Coordinates of optimized tetrahedral intermediate. ......................................52
Figure 2-10. Coordinates of optimized transition state 1 ................................................53
Figure 2-11. Coordinates of optimized transition state 2 ................................................54
Figure 2-12. Coordinates of optimized transition state 3 ................................................55
Figure 3-1. Superposition of the ONIOM optimized yCD-DHP complex structure with
the crystal structure .......................................................................................................80
Figure 3-2. Proposed catalytic pathway for the conversion of cytosine (5FC) to Zncoordinated uracil (5FC)................................................................................................81
Figure 3-3. ONIOM optimized structures for complex 1-6 and transition state structures
during the deamination of 5FC ......................................................................................82
Figure 3-4. ONIOM optimized structures for complex 1’-6’ and transition state structures
during the deamination of cytosine ................................................................................83

xi

Figure 3-5. Energy profiles for the conversion of cytosine to the Zn-coordinated uracil
and the conversion of 5FC to the Zn-coordinated 5FC ...................................................84
Figure 3-6. Rotation of the 5FC ring during the deamination of 5FC .............................85
Figure 3-7. Motion of the active site residues during the deamination of 5FC................86
Figure 3-8. Proposed catalytic pathway for the replacement of the Zn-coordinated uracil
or 5FU with water .........................................................................................................87
Figure 3-9. ONIOM optimized structures for Complex 7, 8 and the transition state
structure for the deamination of 5FC .............................................................................88
Figure 3-10. ONIOM optimized structures for Complex 7’, 8’ and the transition state
structure for the deamination of cytosine .......................................................................88
Figure 3-11. Energy profiles for replacement of the Zn-coordinated uracil or 5FU with
water .............................................................................................................................89
Figure 3-12. Conformational change caused by the access of the water towards the zinc
atom ..............................................................................................................................90
Figure 4-1. QM partition in the simulation systems for the yCD-DHP complex (top) and
with the yCD water/cytosine complex (bottom) .......................................................... 117
Figure 4-2. RMSDs of the backbone C! atoms relative to the crystal structure during the
QM/MM MD simulation for the yCD-DHP complex.................................................. 118
Figure 4-3. Superposition of a representative snapshot during the yCD-DHP complex
simulation with the crystal structure ........................................................................... 119
Figure 4-4. Bond distances of snapshots along the MD trajectory of the yCD-DHP
complex simulation .................................................................................................... 120
Figure 4-5. Energy profiles for proton transfer............................................................ 121
Figure 4-6. Proposed reaction pathway for the conversion of cytosine to uracil in the
QM/MM MD simulations........................................................................................... 122
Figure 4-7. Snapshots of reaction species along the reaction pathway for the conversion
of cytosine to uracil in the QM/MM MD simulations.................................................. 123
Figure 4-8. RMSDs of the backbone C! atoms relative to the crystal structure during the
QM/MM MD simulation for the yCD water/cytosine complex ................................... 124
Figure 4-9. Rotation of the sidechain of Glu64 ("CB-CG-CD-OE2) in the free form yCD
and the yCD substrate complex simulations................................................................ 125

xii

Figure 4-10. Potential of mean force for the proton transfer from the Zinc-coordinated
!2

water to O

of Glu64 and to N3 of cytosine............................................................... 126

Figure 4-11. Potential of mean force for nucleophilic attack on C4 of cytosine........... 127
Figure 4-12. Superposition of the average structure of snapshots in the MD trajectory for
the complex 4 simulation onto that of the yCD-DHP simulation ................................. 128
Figure 4-13. Potential of mean force for the C4-N4 bond cleavage ............................. 129
Zn

Figure 4-14. Reaction pathway for the O -Zn bond breakage.................................... 130
!2

Figure 4-15. Potential of mean force for proton transfer from N3 to O ..................... 131
Zn

Figure 4-16. Potential of mean force for the O -Zn bond breakage ........................... 132
Figure 4-17. Overall energy profile for the conversion of cytosine to uracil by yCD... 133
19

Figure 5-1. F-NMR spectra of the E64A-5FPy complex and the wild-type yCD-5FPy
complex...................................................................................................................... 163
Figure 5-2. Proposed reaction pathway from Py to DHP catalyzed by the wild type yCD
................................................................................................................................... 164
Figure 5-3. ONIOM optimized structures for the conversion of Py to DHP by the wild
type yCD.................................................................................................................... 165
ONIOM

Figure 5-4. Schematic E
and E(High,model) energy profile for the conversion of
Py to DHP catalyzed by the wild-type yCD ................................................................ 166
Figure 5-5. Proposed reaction pathway from Py to DHP catalyzed by the E64A mutant
enzyme....................................................................................................................... 167
Figure 5-6. ONIOM optimized structures for the conversion of Py to DHP by the E64A
mutant enzyme ........................................................................................................... 167
ONIOM

Figure 5-7. Schematic E
and E(High,model) energy profile for the conversion
from Py to DHP catalyzed by the E64A mutant enzyme ............................................. 168
Figure 5-8. Conversion of cytosine to the zinc-coordinated uracil by the E64A mutant
enzyme....................................................................................................................... 169
Figure 5-9. ONIOM optimized structures for the conversion from cytosine to the zinccoordinated uracil....................................................................................................... 170

xiii

ONIOM

Figure 5-10. Schematic E
and E(High,model) energy profile for the conversion
from cytosine to uracil catalyzed by the E64A mutant enzyme ................................... 171
19

Figure 6-1. F-NMR spectra of the mutant-5FPy complex and wild type yCD-5FPy
complex...................................................................................................................... 200
Figure 6-2. RMSDs of the backbone C! atoms relative to the starting structure for the
yCD Py complex during the 8 ns production MD simulations..................................... 201
Figure 6-3. Ligand-residue interaction spectrum of the yCD Py complex by MM-GBSA
energy decomposition................................................................................................. 201
Figure 6-4. 2D energy profiles for proton transfer from the Zinc-coordinated water to N3
of Py........................................................................................................................... 202
Figure 6-5. Superposition of representative snapshots for the yCD Py complex, N51A Py
complex and D155A Py complex................................................................................ 203

xiv

ABBREVIATIONS

5FC

5-fluorocytosine

5FDHP

5-fluoro-4(R)-hydroxyl-3,4-dihydropyrimidine

5FPy

5-fluoro-pyrimidi-2-one

5FU

5-fluorouracil

ADA

Adenosine Deaminase

Amber

Software Package for Molecular Dynamic
Simulation

B3LYP

Becke three parameter, Lee-Yang-Parr

CD

Cytosine Deaminase

CDA

Cytidine Deaminase

CHARMM

Software Package for Molecular Dynamic
Simulation

DHP

4(R)-hydroxyl-3,4-dihydropyrimidine

eCD

E. coli Cytosine Deaminase

FID

Free Induction Decay

G03

Software Package for Quantum Mechanics

GBSA

Generalized Born Surface Area

GDA

Guanine Deaminase

GDEPT

Gene Directed Enzyme Prodrug Therapy

IC50

Half Maximal Inhibitory Concentration

kcat

Turnover Number
xv

Kd

Dissociation Constant

Ki

Inhibition Constant

KIE

Kinetic Isotope Effect

Km

Michaelis Constant

MD

Molecular Dynamics

MM

Molecular Mechanics

MM-PBSA

Molecular Mechanics Poisson Boltzmann Surface
Area

NMR

Nuclear Magnetic Resonance

NOE

Nuclear Overhauser Effect

ONIOM

Our own N-layered Integrated molecular Orbital
and molecular Mechanics

PARM99SB

An Amber force field

PMF

Potential of Mean Force

Py

2-pyrimidinone

QM

Quantum Mechanics

QM/MM

Quantum Mechanics/Molecular Mechanics

RMSD

Root Mean Square Deviation

RMSF

Root Mean Square Fluctuations

TadA

tRNA Adenosine Deaminase

TS

Transition State

WT

Wild Type

xvi

yCD

Yeast Cytosine Deaminase

xvii

CHAPTER 1

INTRODUCTION

Biomedical Significance and Basic Kinetic Properties of Cytosine
Deaminase
Cytosine deaminase (CD) (EC: 3.5.4.1) catalyzes the deamination of cytosine to
uracil and plays an important role in the pyrimidine salvage pathway (1) in bacteria and
fungi. The enzyme is absent in humans, in which cytidine deaminase is utilized for the
pyrimidine conversion instead.
Bacterial CDs, represented by the E. coli enzyme (eCD), and fungal CDs,
represented by the yeast enzyme (yCD), have no sequence homology and utilize different
metal ions for catalysis, an example of convergent evolution (2). yCD was discovered by
Hahn and Schafer in 1925 (3, 4). The basic kinetic properties of yCD were analyzed by
Kream and co-workers on a partially purified preparation of the enzyme (5). It was found
that yCD could use both cytosine (Figure 1-1) and 5-methylcytosine as substrate. yCD
was first purified by Ipata and co-workers (6). Their kinetic studies further demonstrated
that yCD specifically catalyzes the deamination of cytosine and 5-methylcytosine
(Km=2.5 mM for both) among a number of nucleosides and nucleotides. The reaction rate
was pH-dependent, with the optimum at pH 7.0. The reaction was inhibited by cytidine,
thymidine, guanosine, and their nucleotides in a noncompetitive manner, whereas many

1

bases and other nucleosides and nucleotides tested had no inhibitory effects at the
concentration of 0.5 mM, including adenine, adenosine, uracil, uridine, guanine,
xanthine, hypoxanthine, thymine, AMP, ADP, ATP, UTP, IMP, and ITP. The inhibitory
effects by cytidine, thymidine, guanosine, and their nucleotides were suggested to play
important roles in the regulation of pyrimidine nucleotide biosynthesis in yeast (6).
Both yCD and eCD also catalyze the deamination of 5-fluorocytosine (5FC) to 5fluorouracil (5FU) (Figure 1-1). 5FU is a widely used drug to treat breast, rectal,
stomach, colon, and pancreatic cancers (7, 8). However, 5FU is highly toxic. The
undesired side effects can be minimized (9) by using the prodrug 5FC in combination
with a CD in gene-directed enzyme prodrug therapy (GDEPT). The procedure generally
includes two steps (10). First, a gene coding a CD is delivered to the tumor cells. Second,
the prodrug 5FC is injected into the tumor cells and activated by the CD. Because
humans do not have their own CD, 5FC is converted to 5FU only in tumor cells (2).
Thus, in principle, only the tumor cells are exposed to the toxicity of 5FU and are killed,
and normal cells are not affected.
Both eCD and yCD have been used in GDEPT. eCD is chosen because of its
thermal stability (11). However, its application is limited, due to its inefficiency in
converting 5FC to 5FU. yCD has higher catalytic efficiency towards 5FC than eCD,
though it has a relatively lower thermal stability (12). The kcat of yCD for 5FC is 4 times
lower than that of eCD and the Km for 5FC is 21-fold lower than that of eCD (11, 13-15).
Consequently, the catalytic efficiency, kcat/Km, of yCD is 5 times that of eCD. Its
advantage over eCD in GDEPT was demonstrated by in vivo experiments, in which

2

complete tumor regression was observed in 6 of 13 HT29/yCD tumors when treated with
5FC, while none of 10 HT29/eCD tumors were cured (11).
Efforts have been made towards improving the catalytic efficiency and
5

-1 -1

thermostability of yCD (16, 17). yCD has a relatively low kcat/Km (1.1!10 M s
9

for

-1 -1

5FC), compared with that of the most proficient enzymes, which is around 10 M s

(18). By directed evolution or structure-based protein engineering, variants with high
efficiency and thermal stability may be obtained. For example, Stolworthy and coworkers
(16) have identified a yCD variant (yCD/D92E) with increased thermostability and
comparable activity by substituting the conserved residue Asp92 to a glutamate. Our
laboratory also identified several variants which may be more effective than the wild-type
enzyme for GDEPT (unpublished data).
In addition to developing improved candidates for gene therapy, a number of
kinetic, structural, and computational studies were reported to elucidate the catalytic
mechanism of yCD (2, 5, 17, 19-34). yCD belongs to the cytidine deaminase (CDA)
family of purine/pyrimidine deaminases. Enzymes in this family have similar structures
with a zinc-containing catalytic apparatus (2, 19). Studies on yCD not only elucidate the
enzymatic mechanism of yCD itself, but also provide significant insights into the
catalytic mechanism of enzymes in this family. Moreover, these findings can facilitate
antimicrobial drug design and gene-directed therapy against cancers. Such studies are
presented below.

X-ray Crystallographic Analysis of yCD

3

Structural information at the atomic level is almost a necessity for understanding
of the catalytic mechanism of an enzyme. Crystal structures of yCD have been
determined at high resolution, including the substrate-free form (PDB id: 1ox7) and the
complex form (PDB id: 1p6o and 1uaq) with 2-pyrimidinone (Py), a transition state or
reaction intermediate analog (2, 19).
yCD is a homodimeric metalloenzyme (Figure 1-2). Each monomer is tightly
packed into a mixed !/" fold, which consists of a central five stranded "-sheet
sandwiched by !-helices on both sides. A catalytic zinc atom is found in the active site,
#1

$

$

covalently bonded with His62 N , Cys91 S , Cys94 S and a water molecule in the free
form yCD. The water molecule is believed to serve as a nucleophile and is hydrogen
bonded to the conserved residue Glu64, which is likely to serve as a proton shuttle during
the reaction. The inhibitor is bound in the hydrated form 4(R)-hydroxyl-3,4dihydropyrimidine (DHP) (Figure 1-1) and coordinates with the zinc atom through the 4hydroxyl group (Figure 1-3). In both the free form and the complex form, the active site
is completely covered by the same cluster of residues from the C-terminal !-helix and the
loop between "4 and !D (F114 loop) and appears to be inaccessible to the substrate.
Structural alignment shows that the crystal structures of the free form and the complex
form are essentially the same with respect to protein conformation, with a root mean
square deviation (RMSD) of 0.23 Å (20) for the backbone C! atoms, which suggests that
dynamics may play a role in catalysis because either substrate binding or product release
needs the movement of the C-terminal !-helix or the F114 loop, if not both.

4

The crystal structures are invaluable to computational protein engineering studies
(17). A double mutant variant (A23L/I140L) and a triple mutant variant
(A23L/I140L/V108I) have been successfully designed (17), possessing higher apparent
melting temperature Tm and increased half-life of activity at 50 °C without sacrificing
catalytic efficiency.

Reaction Pathway by ONIOM Calculations
An ONIOM approach has been employed to elucidate the catalytic mechanism of
yCD for the conversion of cytosine to uracil (20). The ONIOM method in this study is a
two-layered hybrid method developed by Morokuma and co-workers (35) and
implemented in the Gaussian package (Gaussian 03) (36). Twenty-two amino acid
residues were included to model the enzyme in the molecular system. The molecular
system was divided into two layers, which were treated with different levels of theory.
The inner layer included the substrate cytosine and the zinc-coordinated water, the
catalytic zinc ion, the potential proton shuttle Glu64, and residues coordinated with the
catalytic zinc, His62, Cys91, and Cys94, and treated by the density functional method
B3LYP with the 6-31G** basis set. The remaining residues were included in the outer
layer and treated by the semiempirical PM3 method. A reaction pathway was proposed
based on the ONIOM calculations. The substrate cytosine is activated via two proton
transfer steps. First, the proton is transferred from the zinc-coordinated water to the
carboxyl group of Glu64. Second, the same proton is transferred to N3 of cytosine. The
energy barriers for these steps are negligible. The complex with the activated substrate is
more stable than the original complex by 7.1 kcal/mol at the ONIOM energy level. The

5

zinc-coordinated hydroxide is well positioned for the nucleophilic attack on C4 of
cytosine. The energy barrier is calculated to be 1.0 kcal/mol at the ONIOM level for the
formation of the tetrahedral intermediate. The C4-N4 bond is cleaved in the next step,
forming two products, ammonia and uracil. It is most likely that ammonia rapidly
diffuses into solution and is replaced with a water molecule from the solution, as
suggested by binding free energy calculations (21). The other product, uracil, is
covalently bonded with the zinc atom. The C4-N4 bond cleavage is suggested to be a
rate-limiting step for the deamination of cytosine, with an energy barrier of 7.1 kcal/mol
at the ONIOM level.
It is challenging to understand how the bond between Zn and O4 of uracil is
cleaved. Two possible pathways were proposed in this ONIOM study. In one pathway,
the bond is directly cleaved. However, the results indicate that this pathway is unlikely,
due to the high energy barrier (18.3 kcal/mol). In the second hypothesis, a new
mechanism is proposed, in which the oxygen of the water molecule exchanges into uracil.
Two enzymes in the CDA family of purine/pyrimidine deaminases, cytidine deaminase
and adenosine deaminase, adopt this strategy to break the same bond (37). The ratelimiting step is found in the formation of the gem-diol intermediate from the zinc bound
uracil and the water molecule. The energy barrier is 9.3 kcal/mol at the ONIOM level.
Although the energy barrier is much lower than that proposed in the first mechanism, it is
still the rate-limiting step for the overall reaction of cytosine deamination catalyzed by
yCD.

Zinc-uracil Bond Breakage Calculations

6

In the first hypothesis of the zinc-uracil bond breakage mentioned above, the
activation energy seems too high. However, our recent experiments (unpublished data)
did not support the second mechanism either, because this oxygen exchange rate is too
slow. Direct breakage of the Zn-O4 bond was re-examined by a two-stepped ONIOMMolecular Dynamics (MD) calculation (22). First, the protein in complex with uracil was
simulated by a series of standard MD simulations, with a restraint on the bond distance
between Zn and O4 of uracil. The bond distance increased between successive
simulations in 0.1 Å increments from 2.0 Å to 3.8 Å. Second, a number of snapshots (5 or
10) were selected from the equilibrated MD ensembles and optimized using the ONIOM
scheme as described above (20). The results show that the bond is cleaved with the help
%1

of O

%1

of Glu64. O

replaces O4 of uracil to maintain the zinc coordination. The

potential of mean force (PMF) along the bond-breaking process was calculated. The
transition state is located at the Zn-O4 distance of approximately 2.6-2.7 Å. The energy
barrier was calculated to be only 2.9 kcal/mol. It is much lower than that calculated in the
previous ONIOM calculation, in which the energy barrier is 18.3 kcal/mol. The
difference is rooted in the fact that in the previous ONIOM calculation, the outer layer is
frozen throughout the bond cleavage process, whereas, in this new strategy, the outer
layer is allowed to move in response to the bond-breaking process. The Zn-O4 bond
cleavage thus is unlikely to be the rate-limiting step in the overall deamination reaction of
cytosine.
To explore the dynamic effects on the catalysis, Matsubara and co-workers (23)
performed an ONIOM-MD study on the process of the zinc-uracil bond breakage. The
results show that the Zn-O4 bond abruptly breaks at 11 ps of simulation. The authors

7

suggest that the thermal motion of Ile33, which sandwiches uracil with His62, perturbs
the dynamics of uracil through steric contact and contributes to the bond breakage.

Product Release is Rate-limiting in the Activation of 5FC by yCD
The deamination of the anticancer prodrug 5FC by yCD has been investigated by
transient kinetic and NMR studies (25). The results show that product release is the ratelimiting step in the activation of 5FC, with the rate constants of the chemical steps being
at least 8 times that of the product release step. NMR spectra for yCD in the free form
and in the complex form with 5FU were acquired and analyzed. The Kd of 5FU with yCD
was estimated to be ~20 mM, indicating that the binding affinity of 5FU with yCD is low.
NMR saturation transfer experiments show that 5FU is in slow exchange with its yCD
complex on the NMR time scale. The association and dissociation rate constants are 0.6
-1 -1

-1

mM s and 14 s , respectively.

yCD Dynamics
As revealed by the crystal structures, the active site is completely covered by the
C-terminal !-helix, the F114 loop and loop 1 (residues 53-61). Protein dynamic
properties were studies by NMR relaxation and H/D exchange experiments and MD
simulations (Yao, unpublished data). 5-fluoro-pyrimidi-2-one (5FPy) is an intermediate
analog which can be converted to the hydrated 5-fluoro-4(R)-hydroxyl-3,4dihydropyrimidine (5FDHP) (Figure 1-1).

15

N NMR relaxation experiments show that

the F114 loop and the C-terminal helix are more rigid in the 5FPy bound form than in the

8

free form on the ps to ns timescale, whereas the dynamic properties of the 5FU bound
form are similar to those of the free form. MD simulations show that the root mean
square fluctuations (RMSFs) for the C! atoms are generally larger in these regions than
other regions. H/D exchange experiments show that binding of 5FPy increases exchange
times of the F114 loop, loop 1 and the C-terminal helix, while binding of 5FU does not
increase exchange time too much. These results suggest that these regions are important
for the access of the active site.

Cytosine Dissociation Pathway by MD Simulations
To understand how a ligand accesses the sealed active site of yCD, a series of MD
simulations were performed on the dissociation process of cytosine (27). Two exit paths
were suggested based on the simulations. One path (Path 1) is in between the C-terminal
helix and the F114 loop, and the other (Path 2) is in between the C-terminal helix and
loop 1. To test which exit path is preferred, two MD simulations were performed;
cytosine was pushed out of the active site along Path 1 in one simulation and Path 2 in the
other, respectively. In Path 1, the motion associated with cytosine exit is small and rather
local, involving only the C-terminal helix and the F114 loop. In Path 2, the
conformational changes are much more extensive. The average force is larger for pushing
cytosine out along Path 2 than Path 1, suggesting that cytosine and other ligands may exit
the active site preferably through Path 1.

The Scope of this Dissertation

9

In this dissertation, starting from what we have learned from the literature and
accumulated work in the laboratory, I carried out a series of experiments and
computations to further explore the catalytic mechanism of yCD. The work is presented
in detail in Chapter 2-6 and summarized below.
In the first project, presented in Chapter 2, the transition state of yCD was probed
by kinetic isotope effects (KIEs) at natural abundance by

13

C NMR spectroscopy (38).

KIEs reflect the change of reaction rate due to isotopic substitution at a given position.
Combining KIEs at different positions, it is possible to understand the reaction
mechanism. Instead of being isotopically labeled, the substrate cytosine at natural
abundance was used in this study. To determine the transition state structure, a series of
conformations was constructed that represented different stages of C4-O4 bond formation
or C4-N4 bond cleavage (39), by fixing the C4-O4 bond or the C4-N4 bond at different
distances, respectively. Vibrational frequencies were calculated, which, in turn, were
used to calculate KIEs. The results indicate that the reaction proceeds by a stepwise
AN+DN mechanism, with a transition state in the C4-N4 bond cleavage. The late
transition state established in this study is similar to that reported for E. coli cytidine
deaminase (39) and TadA RNA deaminase (40), but different from the early transition
states for adenosine deaminases from human, bovine and Plasmodium falciparum (41)
and AMP deaminase from Saccharomyces cerevisiae (42).
In the second project, presented in Chapter 3, the catalytic mechanism of the
deamination of 5FC by yCD was investigated by an ONIOM(QM:MM) method with the
electronic embedding (EE) scheme (43). Different from the previous work, the whole
protein was included in the model system. The inner layer was treated by the DFT

10

method B3LYP with the 6-31G* basis set and the outer layer by the Amber force field.
The results show that the reaction pathway for the deamination of 5FC is similar to that
of cytosine. However, energetically more favorable tetrahedral intermediates are found,
which may account for the relatively low Km of 5FC. This finding may facilitate rational
design of enzyme variants specifically towards 5FC, instead of cytosine. In addition, one
possible Zn-O4 bond cleavage pathway was proposed. However, configurational
sampling of proper geometries along the pathway was not promising. Hence, the reaction
pathway was re-investigated by QM/MM MD simulations in Chapter 4.
In the third project, presented in Chapter 4, the reaction mechanism of the
deamination of cytosine by yCD was investigated by combined QM/MM MD and free
energy simulations (44, 45). First, the substrate cytosine is activated by abstracting one
%1

proton from the zinc-coordinated water via O

of Glu64. The zinc-coordinated

hydroxide is well positioned for the nucleophilic attack on C4. Second, the tetrahedral
intermediate is formed after the nucleophilic attack. The energy barrier is quite low, with
a value of 1.4 kcal/mol. Third, the C4-N4 bond is cleaved, which is believed to be the
rate-limiting step with an energy barrier of 19.7 kcal/mol. The product ammonia can
easily diffuse into the solution, while the other product uracil is covalently bonded with
the zinc atom. We propose that one solvent water enters the active site and replaces uracil
in coordination with Zn. The energy barrier is about 8.1 kcal/mol. The results show the
overall transition state is located in the N4-C4 bond cleavage step, which is in agreement
with our KIE experiments.
In the fourth project, presented in Chapter 5, the functional role of Glu64 in the
activation of the prodrug 5FC was investigated by site-directed mutagenesis and
11

biochemical, NMR, and computational studies. In both ONIOM calculations and
QM/MM MD simulations, Glu64 was found to play a critical role in yCD catalysis,
serving as a proton shuttle in multiple steps during the catalytic cycle. A better
understanding of the role of Glu64 will provide more insight not only on the catalytic
mechanism of yCD but also on other members of the CDA family. Steady-state kinetic
studies show that substituting Glu64 with Ala (E64A) causes a dramatic decrease in kcat
and a dramatic increase in Km, with kcat reduced by a factor of >3000 and Km increased
by a factor of >100. The substitutions also cause a dramatic decrease in the affinity of the
reaction intermediate analogue 5FPy. After initial binding with the wild type enzyme, the
inhibitor is hydrated, as suggested by X-ray crystallography.

19

F-NMR experiments show

that binding of 5FPy causes an upfield shift, indicating that 5FPy is hydrated when bound
in the active center of the wild-type enzyme. However, binding with the E64A mutant
enzyme causes a downfield shift, indicating that 5FPy remains in the unhydrated form in
the complex. ONIOM calculations show that the wild type yCD complex with the
hydrated 5FPy is more stable than the initial binding complex, and in contrast, with the
E64A mutant enzyme, the hydrated inhibitor is no longer favored and the activation
energy is elevated. To further explore the function of Glu64 in catalysis, the deamination
of cytosine catalyzed by the E64A mutant enzyme was investigated by ONIOM
calculations. The results show that, without the help of Glu64, both proton transfers
before and after the formation of the tetrahedral intermediate become partially ratelimiting steps. Taking together all results, we conclude that Glu64 plays critical roles in
both binding and catalysis.

12

In the fifth project, presented in Chapter 6, the functional roles of the active site
residues, particularly Asn51 and Asp155, were investigated by site-directed mutagenesis
and biochemical, NMR, and computational studies. Steady-state kinetic studies show that
both the N51A and D155A variants cause a dramatic decrease in kcat and a dramatic
increase in Km, with kcat reduced by a factor of >1000 and Km increased by a factor of
>20. The substitutions also cause a decrease in the binding affinity to 5FPy by a factor of
>100. The MM-PBSA approach (46, 47) was employed to investigate the initial binding
of Py with the yCD mutant enzymes. The results show that the binding affinities of Py
were reduced by 10.6 and 7.6 kcal/mol with the N51A and D155A enzymes, respectively.
Binding of 5FPy with the N51A and D155A mutant enzymes causes a downfield shift in
the

19

F NMR signal, indicating that the inhibitor is in the unhydrated form. QM/MM MD

simulations show that, with the N51A and D155A mutant enzymes, the hydrated form is
no longer favored and the activation energy is elevated. The results suggest that similar to
Glu64, Asp51 and Glu155 are important not only for the initial binding of the ligand, but
also for the stabilization of the transition state.
In conclusion, the catalytic mechanism of yCD was investigated both
experimentally and computationally. First, the transition state structure was accessed
experimentally by KIE studies. The results indicate that the reaction proceeds by a
stepwise AN+DN mechanism, with the transition state in the C4-N4 bond cleavage.
Second, an ONIOM (B3LYP/6-31G*:Amber)-EE scheme was employed to elucidate the
catalytic mechanism of the deamination of 5FC and to compare with that of cytosine. The
calculations on the pathway for the formation of the zinc-coordinated 5FU are successful.

13

However, the pathways for the Zn-5FU bond breakage step cannot be properly
demonstrated, because a relatively large rearrangement of the active site is required.
Hence, QM/MM MD simulations were set up to incorporate configurational sampling.
The results show that the C4-N4 bond cleavage is the rate-limiting step for the overall
reaction. Functional roles of conserved residues around the active site were investigated
by site-directed mutagenesis, biochemical, NMR, and computational methods. Glu64 is
found to be essential for catalysis due to its role in proton transfer. It also participates in
the binding of ligands. N51 and D155 are important for ligand binding. They also
participate in catalysis by stabilizing the transition state.

14

FIGURES

Figure 1-1. Reactions catalyzed by yCD. (1) Deamination of 5FC to 5FU. (2)
Deamination of cytosine to uracil. (3) Conversion of 5FPy to 5FDHP. (4) Conversion of
Py to DHP.
15

Figure 1-2. Crystal structure of yCD (PDB id: 1p6o). Protein dimer is displayed by
ribbon representation with helices colored blue, beta strands colored yellow and the rest
colored gray. The zinc atoms are shown as spheres and are colored green. The inhibitor
DHP is colored purple. The residues coordinated with the zinc atom are displayed as balland-stick. The active site is covered by the C-terminal !-helix and F114 loop, displayed
in the red circle. Residues Phe114, Trp152 and Ile156 (displayed in ball and stick) are on
top of the inhibitor, covering the active site. For interpretation of the references to color
in this and all other figures, the reader is referred to the electronic version of this
dissertation.
16

Figure 1-3. Interactions between the inhibitor DHP and surrounding residues in the
active site of yCD (PDB id: 1p6o). Hydrogen atoms were added using InsightII. The zinc
atom is coordinated with His62, Cys91, Cys94 and the 4-hydroxyl group of DHP.

17

REFERENCES

18

REFERENCES
1. Kurtz, J., Exinger, F., Erbs, P., and Jund, R. (1999) New insights into the pyrimidine
salvage pathway of Saccharomyces cerevisiae: requirement of six genes for cytidine
metabolism, Curr Genet 36, 130-136.
2. Ko, T., Lin, J., Hu, C., Hsu, Y., Wang, A., and Liaw, S. (2003) Crystal structure of
yeast cytosine deaminase - Insights into enzyme mechanism and evolution, J Biol
Chem 278, 19111-19117.
3. Hahn, A., and Schafer, L. (1925) Uber das Verhalten von Pyrimidinderivaten in den
Organismen, Z Biol 511-514.
4. O'Donovan, G. A., and Neuhard, J.
microorganisms, Bacteriol Rev 34, 278-343.

(1970)

Pyrimidine metabolism

in

5. Kream, J., and Chargaff, E. (1952) On the cytosine deaminase of yeast, J Am Chem
Soc 74, 5157-5160.
6. Ipata, P., Marmocch, F., Magni, G., Feliciol, F., and Polidoro, G. (1971) Bakers yeast
cytosine deaminase - Some enzymic properties and allosteric inhibition by
nucleosides and nucleotides, Biochemistry 10, 4270-4276.
7. Aghi, M., Hochberg, F., and Breakefield, X. (2000) Prodrug activation enzymes in
cancer gene therapy, J Gene Med 2, 148-164.
8. Greco, O., and Dachs, G. U. (2001) Gene directed enzyme/prodrug therapy of cancer:
Historical appraisal and future prospectives, J Cell Physiol 187, 22-36.
9. Erbs, P., Regulier, E., Kintz, J., Leroy, P., Poitevin, Y., Exinger, F., Jund, R., and
Mehtali, M. (2000) In vivo cancer gene therapy by adenovirus-mediated transfer of a
bifunctional yeast cytosine deaminase/uracil phosphoribosyltransferase fusion gene,
Cancer Res 60, 3813-3822.
10. Dachs, G., Hunt, M., Syddall, S., Singleton, D., and Patterson, A. (2009) Bystander
or no bystander for gene directed enzyme prodrug therapy, Molecules 14, 4517-4545.
11. Kievit, E., Bershad, E., Ng, E., Sethna, P., Dev, I., Lawrence, T., and Rehemtulla, A.
(1999) Superiority of yeast over bacterial cytosine deaminase for emzyme/prodrug
gene therapy in colon cancer xenografts, Cancer Res 59, 1417-1421.
12. Senter, P., Su, P., Katsuragi, T., Sakai, T., Cosand, W., Hellstrom, I., and Hellstrom,
K. (1991) Generation of 5-fluorouracil from 5-fluorocytosine by monoclonalantibody cytosine deaminase conjugates, Bioconjugate Chem 2, 447-451.
13. Kievit, E., Nyati, M., Ng, E., Stegman, L., Parsels, J., Ross, B., Rehemtulla, A., and
Lawrence, T. (2000) Yeast cytosine deaminase improves radiosensitization and
bystander effect by 5-fluorocytosine of human colorectal cancer xenografts, Cancer
19

Res 60, 6649-6655.
14. Mahan, S., Ireton, G., Knoeber, C., Stoddard, B., and Black, M. (2004) Random
mutagenesis and selection of Escherichia coli cytosine deaminase for cancer gene
therapy, Protein Eng Des Sel 17, 625-633.
15. Mahan, S. D., Ireton, G. C., Stoddard, B. L., and Black, M. E. (2004) Alaninescanning mutagenesis reveals a cytosine deaminase mutant with altered substrate
preference, Biochemistry 43, 8957-8964.
16. Stolworthy, T., Korkegian, A., Willmon, C., Ardiani, A., Cundiff, J., Stoddard, B.,
and Black, M. (2008) Yeast cytosine deaminase mutants with increased
thermostability impart sensitivity to 5-fluorocytosine, J Mol Biol 377, 854-869.
17. Korkegian, A., Black, M., Baker, D., and Stoddard, B. (2005) Computational
thermostabilization of an enzyme, Science 308, 857-860.
18. Fersht, A. (1999) Structure and mechanism in protein science: A guide to enzyme
catalysis and protein folding. W. H. Freeman, New York.
19. Ireton, G., Black, M., and Stoddard, B. (2003) The 1.14 angstrom crystal structure of
yeast cytosine deaminase: Evolution of nucleotide salvage enzymes and implications
for genetic chemotherapy, Structure 11, 961-972.
20. Sklenak, S., Yao, L., Cukier, R., and Yan, H. (2004) Catalytic mechanism of yeast
cytosine deaminase: An ONIOM computational study, J Am Chem Soc 126, 1487914889.
21. Yao, L., Sklenak, S., Yan, H., and Cukier, R. (2005) A molecular dynamics
exploration of the catalytic mechanism of yeast cytosine deaminase, J Phys Chem B
109, 7500-7510.
22. Yao, L., Yan, H., and Cukier, R. (2006) A combined ONIOM quantum chemicalmolecular dynamics study of zinc-uracil bond breaking in yeast cytosine deaminase,
J Phys Chem B 110, 26320-26326.
23. Matsubara, T., Dupuis, M., and Aida, M. (2008) An insight into the environmental
effects of the pocket of the active site of the enzyme. Ab initio ONIOM-molecular
dynamics (MD) study on cytosine deaminase, J Comput Chem 29, 458-465.
24. Hsu, Y., Hu, C., Lin, J., and Liaw, S. (2003) Crystallization and preliminary
crystallographic analysis of yeast cytosine deaminase, Acta Crystallogr D 59, 950952.
25. Yao, L., Li, Y., Wu, Y., Liu, A., and Yan, H. (2005) Product release is rate-limiting
in the activation of the prodrug 5-fluorocytosine by yeast cytosine deaminase,
Biochemistry 44, 5940-5947.

20

26. Kornblatt, M., and Tee, O. (1986) Inhibition of yeast cytosine deaminase by 5bromo-2-pyrimidinone and its covalent hydrate, Eur J Biochem 156, 297-300.
27. Yao, L., Yan, H., and Cukier, R. (2007) A molecular dynamics study of the ligand
release path inyeast cytosine deaminase, Biophys J 92, 2301-2310.
28. Liu, A., Li, Y., Yao, L., and Yan, H. (2006) Simultaneous NMR assignment of
backbone and side chain amides in large proteins with IS-TROSY, J Biomol NMR 36,
205-214.
29. Liu, A., Yao, L., Li, Y., and Yan, H. (2007) TROSY of side-chain amides in large
proteins, J Magn Reson 186, 319-326.
30. Liu, A., Wang, J., Lu, Z., Yao, L., Li, Y., and Yan, H. (2008) Hydrogen-bond
detection, configuration assignment and rotamer correction of side-chain amides in
large proteins by NMR spectroscopy through protium/deuterium isotope effects,
ChemBioChem 9, 2860-2871.
31. Liu, A., Lu, Z., Wang, J., Yao, L., Li, Y., and Yan, H. (2008) NMR detection of
bifurcated hydrogen bonds in large proteins, J Am Chem Soc 130, 2428-2429.
32. Xu, Q., Guo, H., Gorin, A., and Guo, H. (2007) Stabilization of a transition-state
analogue at the active site of yeast cytosine deaminase: Importance of proton
transfers, J Phys Chem B 111, 6501-6506.
33. Yao, L., Liu, A., Li, Y., Cukier, R., and Yan, H. (2004) Dynamic properties of yeast
cytosine deaminase as revealed by NMR spectroscopy and molecular dynamic
simulations, Faseb J 18, C280.
34. Yao, L., Cukier, R., and Yan, H. (2007) Catalytic mechanism of guanine deaminase:
An ONIOM and molecular dynamics study, J Phys Chem B 111, 4200-4210.
35. Svensson, M., Humbel, S., Froese, R. D. J., Matsubara, T., Sieber, S., and
Morokuma, K. (1996) ONIOM: A multilayered integrated MO + MM method for
geometry optimizations and single point energy predictions. A test for Diels-Alder
reactions and Pt(P(t-Bu)3)2 + H2 oxidative addition, J Phys Chem 100, 19357-19363.
36. Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A.,
Cheeseman, J. R., Montgomery, J., Vreven, T., Kudin, K. N., Burant, J. C., Millam, J.
M., Iyengar, S. S., Tomasi, J., Barone, V., Mennucci, B., Cossi, M., Scalmani, G.,
Rega, N., Petersson, G. A., Nakatsuji, H., Hada, M., Ehara, M., Toyota, K., Fukuda,
R., Hasegawa, J., Ishida, M., Nakajima, T., Honda, Y., Kitao, O., Nakai, H., Klene,
M., Li, X., Knox, J. E., Hratchian, H. P., Cross, J. B., Bakken, V., Adamo, C.,
Jaramillo, J., Gomperts, R., Stratmann, R., Yazyev, O., Austin, A. J., Cammi, R.,
Pomelli, C., Ochterski, J. W., Ayala, P. Y., Morokuma, K., Voth, G. A., Salvador, P.,
Dannenberg, J. J., Zakrzewski, V. G., Dapprich, S., Daniels, A. D., Strain, M. C.,
Farkas, O., Malick, D. K., Rabuck, A. D., Raghavachari, K., Foresman, J. B., Ortiz, J.

21

V., Cui, Q., Baboul, A. G., Clifford, S., Cioslowski, J., Stefanov, B. B., Liu, G.,
Liashenko, A., Piskorz, P., Komaromi, I., Martin, R. L., Fox, D. J., Keith, T., AlLaham, M. A., Peng, C. Y., Nanayakkara, A., Challacombe, M., Gill, P. M. W.,
Johnson, B., Chen, W., Wong, M. W., Gonzalez, C., and Pople, J. A. (2004) Gaussian
03, Revision D.01. Gaussian, Inc., Wallingford CT.
37. Shih, P., and Wolfenden, R. (1996) Enzyme-substrate complexes of adenosine and
cytidine deaminases: Absence of accumulation of water adducts, Biochemistry 35,
4697-4703.
38. Singleton, D. A., and Thomas, A. A. (1995) High-precision simultaneous
determination of multiple small kinetic isotope effects at natural abundance, J Am
Chem Soc 117, 9357-9358.
15

39. Snider, M. J., Reinhardt, L., Wolfenden, R., and Cleland, W. W. (2002) N kinetic
isotope effects on uncatalyzed and enzymatic deamination of cytidine, Biochemistry
41, 415-421.
40. Luo, M., and Schramm, V. L. (2008) Transition state structure of E. coli tRNAspecific adenosine deaminase, J Am Chem Soc 130, 2649-2655.
41. Luo, M., Singh, V., Taylor, E. A., and Schramm, V. L. (2007) Transition-state
variation in human, bovine, and Plasmodium falciparum adenosine deaminases, J Am
Chem Soc 129, 8008-8017.
42. Merkler, D. J., Kline, P. C., Weiss, P., and Schramm, V. L. (1993) Transition-state
analysis of AMP deaminase, Biochemistry 32, 12993-13001.
43. Vreven, T., Byun, K. S., Komáromi, I., Dapprich, S., Montgomery, Morokuma, K.,
and Frisch, M. J. (2006) Combining quantum mechanics methods with molecular
mechanics methods in ONIOM, J Chem Theory Comput 2, 815-826.
44. de M Seabra, G., Walker, R. C., Elstner, M., Case, D. A., and Roitberg, A. E. (2007)
Implementation of the SCC-DFTB method for hybrid QM/MM simulations within
the amber molecular dynamics package, J Phys Chem A 111, 5655-5664.
45. Elstner, M., Porezag, D., Jungnickel, G., Elsner, J., Haugk, M., Frauenheim, T.,
Suhai, S., and Seifert, G. (1998) Self-consistent-charge density-functional tightbinding method for simulations of complex materials properties, Phys Rev B 58,
7260-7268.
46. Kollman, P. A., Massova, I., Reyes, C., Kuhn, B., Huo, S., Chong, L., Lee, M., Lee,
T., Duan, Y., Wang, W., Donini, O., Cieplak, P., Srinivasan, J., Case, D. A., and
Cheatham, T. E. (2000) Calculating structures and free energies of complex
molecules: Combining molecular mechanics and continuum models, Acc Chem Res
33, 889-897.
47. Srinivasan, J., Cheatham, T. E., Cieplak, P., Kollman, P. A., and Case, D. A. (1998)
22

Continuum solvent studies of the stability of DNA,
phosphoramidate!DNA helices, J Am Chem Soc 120, 9401-9409.

23

RNA,

and

CHAPTER 2

PROBING THE TRANSITION STATE OF YEAST
CYTOSINE DEAMINASE BY NMR MEASUREMENT OF
13

C KINETIC ISOTOPE EFFECTS

INTRODUCTION
Yeast cytosine deaminase (yCD) catalyzes the conversion of cytosine to uracil in the
salvage pathways of pyrimidine ribonucleotides. yCD is of great biomedical interest
because it is capable of catalyzing the conversion of the prodrug 5FC to the anti-cancer
drug 5FU (1, 2). Combination of yCD and 5FC is a promising strategy in gene directed
enzyme prodrug therapy (GDEPT) (1-11), in which the administrated nontoxic 5FC is
converted to the toxic and lethal 5FU in cancer cells.
X-ray crystal structures of yCD have been solved at high resolution either in the
free form or the complex form with the inhibitor Py (12, 13). The crystal structure of
yCD in complex with Py (Figure 1-3) reveals that the inhibitor is in the hydrated form
DHP (12, 13), suggesting that during deamination of cytosine, the substrate may also go
through a similar tetrahedral intermediate. Based on this assumption, a complete reaction
pathway for deamination of cytosine was proposed by a two-layered quantum mechanics
calculation using the ONIOM method (14). The calculations suggest that deamination of
cytosine is a step-wise reaction via a tetrahedral intermediate, which is stabilized by

24

Glu64 and other surrounding residues by a set of hydrogen bonds and !-packing
interaction. The formation of the tetrahedral intermediate is facile and the C4-N4 bond
cleavage during the decomposition of the tetrahedral intermediate is the rate-limiting
step. The transition state was located when the C4-N4 bond distance is 2.10 Å.
The computational study stimulated our interests in finding the transition state of
yCD experimentally by analysis of kinetic isotope effects (KIEs). KIEs reflect rate
changes of a reaction for isotopically different substrates and have been a powerful probe
to understanding of enzymatic mechanisms (15-24). Combing multiple KIEs makes it
possible to determine the transition state structure. For example,

15

N KIEs have been

measured for the deamination of cytidine catalyzed by E. coli cytidine deaminase (CDA)
(25), one homolog of yCD in the CDA family of purine/pyrimidine deaminases. The
results indicate a mechanism involving formation of a tetrahedral intermediate followed
by cleavage of the C6-N6 bond. E. coli tRNA-specific adenosine deaminase (26), another
homolog of yCD, was also found by KIE studies to adopt a stepwise reaction with a late
transition state in the N-C bond cleavage. On the contrary, adenosine deaminases from
human, bovine and Plasmodium falciparum (27) and AMP deaminase from
Saccharomyces cerevisiae (28), as revealed by KIE studies, have transition states in the
formation of the tetrahedral intermediate. Thus, although these enzymes share the same
catalytic apparatus, the mechanisms could be different.
In this study,

13

C KIEs were measured by NMR spectroscopy for yCD-catalyzed

deamination of cytosine at natural abundance (ca. 1.1% for

13

C), using the method

developed by Singleton and co-workers (29). Hence, the measured KIEs reflected the
competitive nature of substrates with different isotopes on the specificity constant

25

kcat/Km (30). This method has been successfully applied on non-enzymatic catalyzed (3137) and enzymatic catalyzed reactions (15, 38). The transition state of yCD was
determined by combining

13

C KIEs and quantum mechanics calculations. The results

indicate that the reaction proceeds by a stepwise AN+DN mechanism, with a transition
state during the C4-N4 bond cleavage.

MATERIALS AND METHODS
Materials. Cytosine (>99%) and uracil (>99%) were purchased from Sigma.
Dimethyl sulfoxide-D6 (DMSO-D6) (99.9%) and D2O (99%) were purchased from
Cambridge Isotope Laboratories. Chromium(III) acetylacetonate (Cr(acac)3) (99.99%)
was purchased from Aldrich. yCD was produced in E. coli and purified as previously
described (39).
Reaction Setup for KIE Measurement. Four individual reactions for the
deamination of cytosine by yCD were set up on a 300 mL scale. For each reaction, the
system underwent the following steps. First, cytosine was dissolved in 300 mL 50 mM
potassium phosphate buffer (pH 7.0) to a concentration of 15.0 mM. The reaction was
initialized by adding 30 µl of 60 µM yCD and kept at 25 °C until the extent of reaction
reached about 90%. The reaction was quenched by boiling for 10 min and adjusting the
solution pH to 2.5.
Extents of Reaction. Extents of reaction were monitored by following the
absorption change of cytosine at 270 nm (molar extinction coefficient of cytosine = 6 !
-3

10

-1

-1

mol cm ) on an ultraviolet (UV) spectrometer. The accurate extent of reaction was
26

1

determined by H NMR spectroscopy when the reaction was about 90% complete. The
NMR sample was made by collecting and lyophilizing 5 mL reaction solution, and
dissolving the powder in 1 mL D2O. The extent of reaction (F) was calculated according
to Equation 2-1:

F=

INTcyt

(2-1)

INTcyt + INTura

where INTcyt and INTura are areas under the H5 peak of cytosine and uracil on the
spectrum, respectively.
Purification of Residual Cytosine. After the reaction was quenched, the pH of
the solution was adjusted back to 7.0 and the solution was dried by lyophilization. The
powder was a mixture of uracil, residual substrate cytosine and phosphate salt. The
powder was dissolved in 100 mL methanol to remove the large amount of phosphate.
Methanol was removed by centrifuge under vacuum using a Savant SpeedVac system. A
solution was made by dissolving the powder in 20 mM phosphate buffer, pH 2.5, and was
loaded onto a Vydac 218TP 101522 (22 mm ! 250 mm) C18 peptide and protein column
in reverse phase HPLC. 20 mM phosphate, pH 2.5 with 1% methanol was used as elution
buffer and the elution was monitored by UV absorption at 256 nm. The retention times
for cytosine and uracil were 3.6 and 5.8 min, respectively. The cytosine fraction was
collected and adjusted to pH 7.0 before being dried by lyophilization. The powder was
dissolved in methanol again to remove the large amount of phosphate. Methanol was
removed by centrifuge under vacuum using the Savant SpeedVac system. The final
product was the pure residual cytosine, which was not consumed during the reaction. The

27

residual cytosine and original cytosine from the same package were subjected to

13

C

NMR analysis described in the next section.
NMR Experiments. NMR samples were prepared using both the original
cytosine and residual cytosine. Cytosine was dissolved in 0.7 mL DMSO-D6 and sealed
in 5 mm NMR tube. The concentration of cytosine was ~ 200 mM. Cr(acac)3 was used as
a relaxation reagent to reduce the collection time. The concentration of Cr(acac)3 was 20
mM.

13

C spectra were acquired on a Varian 600 MHz spectrometer at 25 °C.

The spectra were acquired with a spectral width of 32,000 Hz and acquisition time
1

of 0.5 s. H inverse-gated decoupling was applied only during acquisition to suppress the
Nuclear Overhauser effect (NOE). During acquisition, the sample tubes were not spun.
The 90° pulses were calibrated and spin-lattice relaxation time (T1) was measured by the
inversion-recovery method to make sure that the relaxation delay (d1) is at least 5 times
T1. The longest T1 for carbon atoms of cytosine was reduced to 0.8 s in the presence of
20 mM Cr(acac)3. Hence d1 was set at 5.0 s. 48,000 scans were recorded per free
induction decay (FID) and an additional 96,000 points were added by zero-filling the
FID. Peaks were assigned according to the deposited

13

C NMR spectrum of cytosine in

DMSO-D6 from the spectral database for organic compounds (SDBS) (40).

13

C chemical

shifts were referenced to the solvent peak of DMSO-D6 (! = 39.52 ppm) (41). The
chemical shifts were 156.910, 166.674, 92.544 and 142.630 ppm for the C2, C4, C5 and
C6 atom of cytosine, respectively.

28

Peak intensity was obtained by integrating the area under a given carbon peak. 10
Hz line broadening was used to improve the signal-to-noise ratio. Baseline correction was
achieved using the Whittaker smoother algorithm (42) embedded in the NMR processing
and analysis package Mnova (43). Peak intensities of C2, C4 and C5 were normalized
relative to that of C6, which was set with a value of 100.
Experimental KIEs. Peak intensities of the carbon atoms in the original cytosine
and residual cytosine were obtained by the procedure described above. The enrichment of
13

C at a given position was measured by the ratio of the intensity of the peak in the

residual cytosine spectrum (Ri) to that in the original cytosine spectrum (R0). KIE for
each carbon atom was calculated by Equation 2-2 (29).
KIE =

ln(1 ! F)

(2-2)

ln #(1 ! F) " Ri / R0 %
$
&

where F is the extent of reaction calculated by Equation 2-1. Ri and R0 for C6 in
both original and residual cytosine were the same because C6 was used as the internal
standard. Four individual reactions were performed and the KIEs were averaged for each
carbon atom.
ONIOM Calculations. A two-layer ONIOM(QM:MM) method (44-46) as
implemented in the Gaussian program (Gaussian 03) (47) was employed for the
calculations on the yCD-transition state complex. The model system is based on the highresolution crystal structure of yCD in complex with DHP (PDB code:1p6o) (12).
The system was divided into two layers: inner layer and outer layer (Figure 2-1,
top). The whole protein is termed as the real system. The inner layer is termed as the

29

model system. The overall ONIOM energy E

ONIOM

is obtained by combining three

independent calculations (44-46):

E ONIOM = E real,MM + E model,QM ! E model,MM
where, E

real,MM

(2-3)

is the energy for the entire system at the MM level, E

the energy for the inner layer at the QM level and E

model,MM

model,QM

is

is the energy for the inner

layer at the MM level. Hence, the ONIOM energy is defined as an extrapolation scheme.
The inner layer is composed of residues Glu64, His62, Cys91, Cys94, Asn51,
Asp155, zinc, and the zinc-bound transition state (Figure 2-1, bottom). The inner layer
was treated at a high energy level and fully optimized. The rest of the residues are in the
outer layer. The outer layer was treated at a low energy level. In the previous study (14),
the entire outer layer was restrained during optimization. Freezing of the residues in the
outer layer, especially those that have direct interaction with the substrate, may result in
unrealistic configurations and/or energy profiles. In this study, instead of restraining all
residues in the outer layer, only residues outside of a radius of 10 Å of the substrate were
restrained. Residues within a radius of 10 Å of the substrate include residues 18, 21-22,
25, 30-36, 49-50, 52-55, 58-61, 63, 65-68, 84-90, 92, 93, 95-98, 106-111, 113-119, 122,
137, 141, 144-145, 152-158, 223, 251, 254-255 and 257-259, and water molecules 326,
333, 338-339, 341, 353, 359, 376, 378, 380-381, 416, 448, 450, 454, 505, 600, 603, 629,
636, 645, 669, 684, 690-691, 726. Thus, both active site residues and surrounding
residues are expected to adopt proper conformation during optimization. Hydrogen atoms
were used as link atoms to saturate the dangling bonds for the interface between the inner
layer and the outer layer.

30

The optimization procedure includes three steps. First, the protein was dissolved
in a 2-Å water shell and the structure was optimized using the Sander program
implemented in the Amber 10 package (48) to remove unfavorable interactions within the
protein. The optimized protein was used as the starting structure for the ONIOM
calculation. The whole system included 316 amino acids from the protein, 2 zinc atoms,
the transition state and 286 additional water molecules. The total number of atoms is
5734. Second, the system was optimized by an ONIOM(B3LYP/6-31G*:Amber) mechanical embedding (ME) scheme. Third, the electric embedding (EE) formalism (49)
was applied in the ONIOM scheme to calculate the interaction between QM and MM.
With ME, the electrostatic interaction between the inner layer and the outer layer is
treated at the MM level. With EE electronic embedding, the electrostatic interaction is
incorporated in the quantum mechanical Hamiltonian (49). Thus, the interaction is better
described than in the ME formalism.
B3LYP/6-31+G** is a better basis set to describe a system containing anions by
introduction of diffusion functions. However, the optimization is very time-consuming.
Optimization was not performed at this energy level; instead, a single-point ONIOM
energy was calculated for the optimized structure, using the ONIOM(B3LYP/631+G**:Amber) - EE scheme. The results were found to be similar to those presented
here.
Potential transition state structures were obtained by systematically increasing the
C4-N4 bonding distance at an interval of 0.05 Å. The complete reaction pathway by this
ONIOM scheme can be found in chapter 3.
Transition State Structure. Using the experimental KIEs as restraints, the
transition state structure was obtained by quantum mechanics calculations in the

31

Gaussian 03 package (47). According to previous ONIOM calculations (14), the
deamination of cytosine is likely to proceed via a tetrahedral intermediate. To determine
the transition state structure, a series of potential transition state structures was built in
the processes of tetrahedral intermediate formation and decomposition. First, a cluster of
structures was built to search for the potential transition state during the tetrahedral
intermediate formation by varying the bond distance between O4 and C4. Similarly, a
cluster of structures was built to search for the potential transition state during the
decomposition of the tetrahedral intermediate by varying the bond distance between C4
and N4. All structures as well as the substrate cytosine were subjected to optimization
using the density functional method B3LYP with the basis set 6-31G**. Geometries of
the potential transition state optimized by ONIOM calculations were exacted and reoptimized at the same energy level mentioned above, but with restraints on the dihedral
angles !N4-C4-N3-C2 and !N4-C4-C5-C6. Another potential transition state was
calculated using the conformations after H3 transfers to O4, as in the tautomer of uracil.
Vibrational frequencies were calculated for the optimized structures. All 3N-6 vibrational
modes, except the trivial ones, were used to calculate KIEs using the ISOEFF98 package
(50). The KIEs of C2, C4, and C5 were normalized relative to that of C6, as applied for
the experimental KIEs.

RESULTS AND DISCUSSION
Deamination of Cytosine. yCD stock solutions with different concentrations
were tested for the enzyme catalyzed reaction. Four reactions were set up with yCD in a
final concentration of 6 nM. Under this condition, the reactions were 90% complete after
1

ca. 16 h. The exact extent of reaction was measured by H NMR spectroscopy. The peaks
32

for H5 of cytosine (! = -5.83 ppm) and H5 of uracil (! = -5.66 ppm) were well separated
1

on the H NMR spectrum. The reaction extents for the four independent reactions were in
the range of 86% to 90% (Table 2-1). The residual cytosine was successfully purified
1

from uracil by reverse phase HPLC. The H NMR spectrum confirmed that there was no
uracil contamination in the final product.
13

C NMR Spectra. The NMR Spectra were acquired for three samples of the

original cytosine. Four independent deamination reactions on cytosine from the same
bottle were set up and spectra of the residual cytosine were acquired. Peak intensity for
each carbon atom was calculated by integrating the area under the peak and normalized
relative to that of the C6 atom (Table 2-2 and 2-3). The natural abundance of

13

C is

around 1.1%, but they may differ from batch to batch. Hence, it is important to compare
spectra of residual and starting material from the same batch.
Experimental KIEs.

13

C enrichment at a given position was evaluated by

comparing the peak intensity of the atom in the residual and original cytosine

13

C NMR

spectra. KIEs were calculated for C2, C4 and C5 relative to C6 according to Equation 22. Hence, by definition, KIE of C6 is unity and KIEs of other carbons are relative.
Because C6 is two bonds away from C4, which is involved in the bond formation and
breakage during the reaction, KIEs relative to C6 are expected to be close to true KIEs.
Four individual reactions were performed and KIEs were averaged for each carbon atom.
KIEs were calculated to be 1.002(9), 1.023(4), 1.000(3) for C2, C4, C5, respectively
(Table 2-4 and Figure 2-2). The numbers in the parentheses are the third digit of the
standard derivation. A relatively large value of C4 KIE (1.023) indicates that the

33

tetrahedral intermediate formation or decomposition is the first irreversible step (Figure
2-3).
In vacuo Transition State Structure. Crystal structures have been determined at
high resolution for yCD with the inhibitor Py (12, 13), which was found to be the
hydrated form DHP. The authors proposed that the deamination reaction proceeds
through formation of a tetrahedral intermediate. Similarly, crystal structures have been
solved for enzymes in the CDA family of purine/pyrimidine deaminase, such as E. coli
cytidine deaminase (51-53), murine adenosine deaminase (54), E. coli tRNA deaminase
(55) and Bacillus subtilis guanine deaminase (56). All of these structures have transition
state based analogs bound in the active site. KIEs (25-28) were measured for these
enzymes, based on which, the transition states were accessed. The results showed that the
transition state is either in the formation or the decomposition of a tetrahedral
intermediate.
These studies suggest that the deamination catalyzed by the enzymes using the
same catalytic apparatus is a stepwise reaction via a tetrahedral intermediate, instead of
being concerted. The transition state of yCD could be located in either the formation or
decomposition of the tetrahedral intermediate (Figure 2-3). Thus, to determine the
transition state structure, a series of conformations of potential transition state structures
was constructed starting from the tetrahedral intermediate. These conformations represent
different states during C4-O4 bond formation or C4-N4 bond cleavage by fixing the C4O4 bond or the C4-N4 bond at different distances, respectively. Vibrational frequencies
were calculated for the optimized structures as well as the substrate cytosine. There was
no imaginary frequency for cytosine, suggesting that the optimized structure is a true

34

ground state minimum. For the transition state, it had one imaginary frequency
corresponding to the reaction coordinate.
The KIEs were calculated using all 3N-6 vibrational modes, except the trivial
ones. The KIEs of C2, C4, and C5 were normalized relative to that of C6. The primary
13

C KIE of C4 was found to match the experimental KIE when the C4-N4 bond distance

is 1.88 Å. The optimized transition state (TS1) is presented in Figure 2-4, A. A proton
atom is located in the middle between O4 and N4, with a distance of 1.21 and 1.31 Å,
respectively. It represents the transition state for a concerted reaction including the proton
transfer from O4 to N4 and the C4-N4 bond cleavage.
The primary

13

C KIE of C4 reflects the change of hybridization of C4 between

the substrate and the transition state. The protonation state of O4 and N4 cannot be
determined solely based on the
and

15

13

C KIE of C4. Additional deuterium KIEs of the proton

N KIE of N4 are desirable to correctly decipher the transition state structure, which

was not measured in this study. Instead, we tried to refine the transition state in the
context of the enzyme environment by ONIOM calculations as presented in the section
below.
ONIOM Calculations. The complete reaction path for the deamination of
cytosine catalyzed by yCD has been proposed by the previous ONIOM calculation (14)
and discussed in detail in later chapters (see Chapter 3). The reaction pathway mainly
consisted of three events, formation of a tetrahedral intermediate, decomposition of the
tetrahedral intermediate and Zn-uracil bond breakage.
The calculated

13

C KIE of C4 at the transition state (TS1) matched well with the

experimental KIE. However, within the enzyme context, the proton may not be directly
35

transferred to N4 from O4. One study showed that the energy barrier is extremely high
for a similar transition state in H2O-catalyzed deamination of cytosine (57).
A conserved glutamate residue Glu64 was found to be critical to catalysis.
8

Substituting Glu64 with Ala dramatically reduced the catalytic efficiency by 10 times
(see Chapter 5). During decomposition of the tetrahedral intermediate, Glu64 abstracts a
proton from O4 and transfers to N4. Both the previous and new ONIOM calculations
have proposed that the transition state was located in the process of the C4-N4 bond
cleavage. To incorporate the input of the enzyme on the transition state, a series of
structures was constructed starting from the tetrahedral intermediate, with different
distances between C4 and N4. ONIOM optimizations were performed on yCD in
complex with these structures. With the increase of the C4-N4 distance, C4 gradually
switches from sp3 to sp2 hybridization and the proton is transferred to N4 from the
protonated carboxyl group of Glu64. Geometries of these potential transition states
optimized by ONIOM calculations (Figure 2-5) were extracted and re-optimized at the
same energy level, but with constraints on the dihedral angles !N4-C4-N3-C2 and !N4C4-C5-C6. The purpose of imposing the dihedral angle constraints is to reproduce the
configuration of the leaving ammonia. Similarly KIEs were obtained using the vibrational
frequencies as input.
By comparing the calculated

13

C KIE of C4 with the experimental KIE, the

transition state (TS2) structure was found to be in the process of C4-N4 bond cleavage,
with a C4-N4 bond distance at 1.93 Å (Figure 2-4, B). In TS2, the proton is already
transferred to N4 with a distance of 1.02 Å. However, in the enzyme environment, the
N4-H bond distance is 1.06 Å. Similarly, the N3-H3 bond distance also decreased, from

36

1.06 to 1.02 Å. Comparing with TS1, the C4-N4 bond increased slightly to 1.93 Å. The
bond distance between C4 and O4 decreased greatly, from 1.33 to 1.24 Å. N3 adopts
more towards sp3 from sp2, indicated by the increased bond distance between N3 and
C4.
Surprisingly, the calculated

13

C KIE of C2 was reversed and needs to be

explained (Table 2-5). From the QM/MM MD simulation on yCD described in Chapter 4,
the newly generated uracil loses a proton in H3 to Glu64 after the C4-N4 bond cleavage.
Another potential transition state was calculated using the conformations after H3
transfers to O4, as in the tautomer of uracil. The new transition state (TS3) (Figure 2-4,
C) was successfully constructed by matching the calculated KIEs with experimental KIEs
(Table 2-5). The

13

C KIE of C2 was close to the experimental KIE too. The bond

distance between C4 and N4 increased to 1.99 Å. Although it matches with the
experimental KIEs better, one should be cautious to conclude that TS3 is the true
transition state, because the calculated secondary KIEs may not reflect true KIEs (18) in
an enzyme environment. Furthermore, the Glu64 sidechain would need to undergo a large
motion to reach the conformation in the transition state (TS3).
Commitment to Catalysis. It is important that the measured KIEs really reflect
the process of chemical transformation, instead of substrate binding or conformational
change. When an enzyme is so efficient that all substrates bound are consumed rapidly to
product, instead of being in equilibrium with the free substrate in solution, substrate
binding becomes the first irreversible step. Forward commitment can be measured by
isotope trapping experiments (58), in which the equilibrium between isotope labeled
substrate and enzyme is disrupted by addition of an excessive amount of unlabeled
substrate. For example, commitment factors for ADA (27) were measured. However, the
37

commitment is very small, because the product turnover rate is much slower than the
dissociation rate of the substrate. In the case of yCD, the Km for cytosine is in the range
of millimolar and the turnover rate is slow. The enzyme is not a very efficient enzyme
4

-1 -1

with a kcat/Km of 8.2 ! 10 M s . The most efficient enzymes, such as catalase, can
turn over the substrate so quick that the association of substrate becomes the rating
limiting step (18, 30).
The transient kinetics parameters were determined for the better substrate 5FC
(39). The dissociation rate constant (k-1) of the Michaelis complex (ES) is only 3 times
lower than the forward rate constant k2. Most likely, the forward commitment is very
small. Reverse commitment is not likely either, because the product ammonia can rapidly
diffuse into solution, making the reaction irreversible. Even for the deamination of 5FC,
although the rate-limiting step is the product release of 5FU (39), the first irreversible
step is prior to it because ammonia has already diffused into solution before the release of
5FU. Hence, the measured KIEs should represent the intrinsic KIEs on kcat/Km.
Based on the transition state analysis, a reaction mechanism was proposed as
shown in Figure 2-6. The substrate can go over a lower energy barrier, if any, and form
the tetrahedral intermediate. The first irreversible step is located in the decomposition of
the tetrahedral intermediate.
Molecular Electrostatic Potential Surface. The molecular electrostatic potential
surfaces were plotted for cytosine, the tetrahedral intermediate, the transition states and
uracil (Figure 2-7). The coordinates for cytosine, the tetrahedral intermediate, the
transition states are listed in Figure 2-8 to 2-12, respectively. Prior to the O4-C4 bond
formation, N3 is protonated and passes more positive charge to the C4 atom, facilitating
38

the nucleophilic attack on C4. The N4 atom in the tetrahedral intermediate gains some
negative charge compared with that of cytosine and is likely to accept a proton. In the
first transition state (TS1), the proton is trapped between O4 and N4, which may
dramatically increase the energy level needed for the departure of ammonia. In the
second transition state (TS2), the negative charge on N4 is neutralized and NH3 is ready
to be released. Similarly for the third transition state (TS3), the C4-N4 bond is ready to
be cleaved. In the product uracil, C4 returns to sp2 hybridization. The primary

13

C KIE

of C4 arises from the change of hybridization state. If the first irreversible step is the
substrate binding or the product release, in which there is no hybridization change, the
KIE should be close to unity.
Molecular electrostatic potential surfaces were also plotted for the inhibitor 2Py
and its hydrated form DHP (Figure 2-7). 2Py is like a substrate mimic and hence cannot
effectively inhibit yCD. DHP is more like a tetrahedral intermediate mimic because, like
the tetrahedral intermediate, H3 of DHP and HO of DHP form hydrogen bonds with the
carboxyl group of Glu64. According to the transition state theory, the protein
environment is used to lower the energy of the transition state. If the interactions
involved in the transition state stabilization can be immobilized by an inhibitor, the
inhibitor is more effective. Although DHP is already a very good inhibitor, it is still
possible to design a transition state analog with higher inhibition efficiency.

CONCLUSIONS
Measurement of

13

C KIEs for the deamination of cytosine catalyzed by yCD at

natural abundance was achieved by NMR spectroscopy. The transition state structure was

39

characterized by quantum chemical calculations. To get a more accurate transition state
in the context of enzyme, ONIOM calculations were performed on yCD in complex with
transition state structures. The transition state structure was found to be in the process of
the C4-N4 bond cleavage, with a C4-N4 bond distance at 1.93 Å. Previous ONIOM
calculations on the deamination of cytosine by yCD proposed a reaction pathway in
which a similar transition state was obtained. The results indicate that the reaction
proceeds by a stepwise AN+DN mechanism, with a transition state during the C4-N4
bond cleavage. The late transition state established in this study is similar to that reported
for E. coli cytidine deaminase and TadA RNA deaminase, but different from the early
transition states for adenosine deaminases from human, bovine and Plasmodium
falciparum and AMP deaminase from Saccharomyces cerevisiae.

40

TABLES AND FIGURES

Table 2-1. Reaction extents for yCD catalyzed reactions.
Reaction 1

Reaction 2

Reaction 3

Reaction 4

87.74%

87.94%

86.49%

89.39%

Reaction
Extent

Table 2-2. Carbon peak intensities for the original cytosine.
Original 1

Original 2

Original 3

Average

Standard
Deviation

C2

101.37

101.32

98.67

100.45

1.54

C4

102.48

101.39

101.25

101.71

0.67

C5

100.36

101.98

98.77

100.37

1.61

C6

100

100

100

100

0

a

a: Intensity of C6 was set at 100 and other carbon intensities were relative to C6.

Table 2-3. Carbon peak intensities for the residual cytosine.
Carbon

Reaction 1

Reaction 2

Reaction 3

Reaction 4

C2

101.46

98.16

102.19

101.37

C4

105.75

106.03

107.25

107.5

C5

99.42

100.11

100.76

100.97

C6

100

100

100

100

a

a: Intensity of C6 was set at 100 and other carbon intensities were relative to C6.

41

Table 2-4. Experimental

13

C KIEs.

Reaction 1

Reaction 2

Reaction 3

Reaction 4

Average

Standard
Deviation

C2

1.005

0.989

1.009

1.004

1.002

0.009

C4

1.019

1.020

1.027

1.025

1.023

0.004

C5

0.995

0.999

1.002

1.003

1.000

0.003

1

1

1

1

0

C6

1

a

a: KIEs were relative to C6.

Table 2-5. Calculated

13

C KIEs.

Carbon

KIEexp

TS1 KIEcalc

TS2 KIEcalc

TS3 KIEcalc

C2

1.002(9)

0.987

0.987

0.998

C4

1.023(4)

1.023

1.023

1.022

C5

1.000(3)

0.995

0.998

0.998

C6

1

a

1

1

1

a: KIEs were relative to C6.

42

Table 2-6. Bond distances for cytosine and the transition states.
Cytosine

Transition
State 1

Transition
State 2

Transition
State 3

N1-C2

1.425

1.392

1.392

1.429

C2-N3

1.372

1.381

1.382

1.368

C2-O1

1.220

1.219

1.222

1.223

N3-C4

1.332

1.415

1.452

1.330

N3-H3

NA

1.011

1.013

NA

C4-O4

NA

1.329

1.236

1.366

C4-N4

1.349

1.880

1.93

1.99

O4-H4

NA

1.206

NA

0.972

N4-H4

NA

1.314

1.019

1.019

C4-C5

1.442

1.474

1.492

1.457

C5-C6

1.358

1.344

1.342

1.351

C6-N1

1.356

1.385

1.387

1.366

43

Figure 2-1. Model system for ONIOM calculations. (Top) Partition of the inner and outer
layers for the ONIOM calculations. The inner layer is composed of residues Glu64,
His62, Cys91, Cys94, Asn51, Asp155, zinc and the TS (displayed by ball and stick in the
middle of the active site). The outer layer is composed of the rest of the components.
Residues in the outer layer but within 10 Å of the ligand are allowed to move freely (lines
in red) and the rest are frozen (lines in blue) during optimization. (Bottom) Schematic
drawing of the inner layer.
44

Figure 2-2. Experimental KIEs mapped on the cytosine structure. Number in the
parentheses is the third digit of the standard deviation. C6 KIE is used as internal
standard.

45

Figure 2-3. Proposed reaction pathway for yCD-catalyzed deamination of cytosine to
uracil. As a general base, Glu64 abstracts a proton from the zinc-coordinated water. The
hydroxide attacks C4 of cytosine to form a tetrahedral intermediate. Two possible
transition state structures were proposed, one in the formation of the tetrahedral
intermediate and the other in the decomposition of the tetrahedral intermediate.

46

Figure 2-4. Possible transition state structures. A: Transition state for the proton transfer
from O4 to N4 with C4-N4 distance of 1.88 Å. B: Transition state with constraints on the
N4-C4-N3-C2 and N4-C4-C5-C6 dihedral angles, the values of which were adapted from
the ONIOM optimized structure. C: Transition state after H3 transfers from N3 to O4.

47

Figure 2-5. ONIOM optimized transition state structure in the active site of yCD.

48

Figure 2-6. Proposed reaction mechanism based on KIEs.

49

Figure 2-7. Molecular electrostatic potential (MEP) surfaces for cytosine, tetrahedral
intermediate, uracil, Py, DHP, and three transition states.

50

N

-1.279358

0.896720

0.002476

C

-1.187607 -0.529956 -0.000089

O

-2.223508 -1.173897

0.001017

N

0.081818 -1.053172

0.002868

C

1.131126 -0.253206 -0.003479

C

1.052295

1.185972 -0.002115

C

-0.199374

1.715173

0.001656

N

2.358192 -0.845124 -0.041488

H

-2.221649

1.261800

0.005276

H

3.186568 -0.319468

0.183021

H

2.377873 -1.845160 0.091603

H

1.933523

1.814016 -0.013667

H

-0.391450

2.783130

0.002793

Figure 2-8. Coordinates of optimized cytosine.

51

N

-1.599229

0.881741

0.022234

C

-1.445535 -0.497695 -0.021928

O

-2.396864 -1.264444

N

-0.139585 -0.905555 -0.174018

C

1.073877 -0.121820

0.018930

O

1.651511 -0.299420

1.303213

C

0.734951

1.339690 -0.104803

C

-0.534206

1.763635 -0.073670

0.021043

N

2.042968 -0.649870 -0.941939

H

-2.552252

H

-0.025887 -1.910527 -0.184581

H

2.910914 -0.125155 -0.859440

H

1.692812 -0.528363 -1.888981

H

2.078347 -1.171024

H

1.558224

2.040990 -0.153054

H

-0.802946

2.813089 -0.124607

1.204832

0.075162

1.286337

Figure 2-9. Coordinates of optimized tetrahedral intermediate.

52

N

1.603991

0.852013

0.224228

C

1.460325 -0.516794

0.001499

O

2.408808 -1.285229

0.018672

N

0.160231 -0.894918 -0.251952

C

-1.023077 -0.079281 -0.250487

O

-2.017052 -0.532690 -1.011433

C

-0.671874

1.367665 -0.258483

C

0.587908

1.764135 -0.022908

N

-1.911794 -0.436782

1.191983

H

2.560422

0.324449

H

0.032033 -1.849285 -0.561520

H

-2.247455

0.374598

1.709908

H

-1.479083 -1.099839

1.834785

H

-2.513146 -0.774011

0.165942

H

-1.462138

2.072626 -0.480089

H

0.888629

2.806120 -0.018929

1.156598

Figure 2-10. Coordinates of optimized transition state 1.

53

H

-2.321350 -1.431980

1.014592

N

1.616375

0.870920

0.171906

H

2.570046

1.190265

0.242591

C

1.472422 -0.507881

0.050070

O

2.430534 -1.265663

0.072293

N

0.162149 -0.922234 -0.092934

H

0.082255 -1.872979 -0.433825

C

-0.991128 -0.095243 -0.399089

O

-1.787719 -0.487004 -1.259097

C

-0.699134

1.345207 -0.143437

H

-1.503138

2.051249 -0.307932

C

0.564973

1.762689

0.023476

H

0.837689

2.810878

0.082558

N

-2.083146 -0.451247

1.151930

H

-1.827778 -0.316275

2.131648

H

-2.930681

0.962364

0.079474

Figure 2-11. Coordinates of optimized transition state 2.

54

H

-1.643690 -1.461771

1.459898

N

1.568048

0.832763

0.175106

H

2.526941 1.106920

0.332236

C

1.379391 -0.562229 -0.069774

O

2.351185 -1.299316

N

0.112153 -0.947459 -0.411065

H

-1.832686 -1.352316 -1.334939

C

-0.889512 -0.072817 -0.370520

O

-2.020667 -0.445590 -1.038944

C

-0.682683

1.365937 -0.269190

H

-1.497580

2.060466 -0.418920

C

0.579712

1.761061

0.006277

H

0.868235

2.802947

0.101626

N

-1.645961 -0.443588

1.432375

H

-1.034963 -0.120265

2.180497

H

-2.591528 -0.120460

1.633947

0.025193

Figure 2-12. Coordinates of optimized transition state 3.

55

REFERENCES

56

REFERENCES
1. Aghi, M., Hochberg, F., and Breakefield, X. (2000) Prodrug activation enzymes in
cancer gene therapy, J Gene Med 2, 148-164.
2. Greco, O., and Dachs, G. U. (2001) Gene directed enzyme/prodrug therapy of cancer:
historical appraisal and future prospectives, J Cell Physiol 187, 22-36.
3. Hamstra, D., Rice, D., Fahmy, S., Ross, B., and Rehemtulla, A. (1999)
Enzyme/prodrug therapy for head and neck cancer using a catalytically superior
cytosine deaminase, Hum Gene Ther 10, 1993-2003.
4. Kievit, E., Bershad, E., Ng, E., Sethna, P., Dev, I., Lawrence, T., and Rehemtulla, A.
(1999) Superiority of yeast over bacterial cytosine deaminase for emzyme/prodrug
gene therapy in colon cancer xenografts, Cancer Res 59, 1417-1421.
5. Erbs, P., Regulier, E., Kintz, J., Leroy, P., Poitevin, Y., Exinger, F., Jund, R., and
Mehtali, M. (2000) In vivo cancer gene therapy by adenovirus-mediated transfer of a
bifunctional yeast cytosine deaminase/uracil phosphoribosyltransferase fusion gene,
Cancer Res 60, 3813-3822.
6. Kievit, E., Nyati, M., Ng, E., Stegman, L., Parsels, J., Ross, B., Rehemtulla, A., and
Lawrence, T. (2000) Yeast cytosine deaminase improves radiosensitization and
bystander effect by 5-fluorocytosine of human colorectal cancer xenografts, Cancer
Res 60, 6649-6655.
7. Nyati, M., Symon, Z., Kievit, E., Dornfeld, K., Rynkiewicz, S., Ross, B., Rehemtulla,
A., and Lawrence, T. (2002) The potential of 5-fluorocytosine/cytosine deaminase
enzyme prodrug gene therapy in an intrahepatic colon cancer model, Gene Ther 9,
844-849.
8. Nyati, M., Sreekumar, A., Li, S., Zhang, M., Rynkiewicz, S., Chinnaiyan, A.,
Rehemtulla, A., and Lawrence, T. (2002) High and selective expression of yeast
cytosine deaminase under a carcinoembryonic antigen promoter-enhancer, Cancer
Res 62, 2337-2342.
9. Warmann, S., Armeanu, S., Heigoldt, H., Ruck, P., Vonthein, R., Heitmann, H.,
Seitz, G., Lemken, M., Bitzer, M., Fuchs, J., and Lauer, U. (2009) Adenovirusmediated cytosine deaminase/5-fluorocytosine suicide gene therapy of human
hepatoblastoma in vitro, Pediatr Blood Cancer 53, 145-151.
10. Dachs, G., Hunt, M., Syddall, S., Singleton, D., and Patterson, A. (2009) Bystander
or no bystander for gene directed enzyme prodrug therapy, Molecules 14, 4517-4545.
11. Xing, L., Deng, X., Gade, T., Ling, C., Koutcher, J., and Li, G. (2007) Combination
of cytosine deaminase with uracil phosphoribosyl transferase with 5-fluorocytosine
radiosensitizes rat prostate cancer cells, Int J Radiat Oncol 69, S617-S618.

57

12. Ireton, G., Black, M., and Stoddard, B. (2003) The 1.14 angstrom crystal structure of
yeast cytosine deaminase: Evolution of nucleotide salvage enzymes and implications
for genetic chemotherapy, Structure 11, 961-972.
13. Ko, T., Lin, J., Hu, C., Hsu, Y., Wang, A., and Liaw, S. (2003) Crystal structure of
yeast cytosine deaminase - Insights into enzyme mechanism and evolution, J Biol
Chem 278, 19111-19117.
14. Sklenak, S., Yao, L., Cukier, R., and Yan, H. (2004) Catalytic mechanism of yeast
cytosine deaminase: An ONIOM computational study, J Am Chem Soc 126, 1487914889.
15. Lee, J. K., Bain, A. D., and Berti, P. J. (2004) Probing the transition states of four
13
glucoside hydrolyses with C kinetic isotope effects measured at natural abundance
by NMR spectroscopy, J Am Chem Soc 126, 3769-3776.
16. Berti, P. J., and Tanaka, K. S. E. (2002) Transition state analysis using multiple
kinetic isotope effects: Mechanisms of enzymatic and non-enzymatic glycoside
hydrolysis and transfer, Adv Phys Org Chem 37, 239-314.
17. Northrop, D. B. (1981) The expression of isotope effects on enzyme-catalyzed
reactions, Annu Rev Biochem 50, 103-131.
18. Berti, P. J. (1999) Determining transition states from kinetic isotope effects, in
Enzyme kinetics and mechanism Part E: Energetics of Enzyme Catalysis, 355-397.
Academic Press.
19. Schimerlik, M. I., Rife, J. E., and Cleland, W. W. (1975) Equilibrium perturbation by
isotope substitution, Biochemistry 14, 5347-5354.
20. Cook, P. F. (2007) Enzyme kinetics and mechanism. Garland Science, London.
21. Schramm, V. L. (2007) Enzymatic transition state theory and transition state
analogue design, J Bio Chem 282, 28297 -28300.
22. Cleland, W. W. (2007) Use of isotope effects to determine enzyme mechanisms, J
Label Compd Radiopharm 50, 1006-1015.
23. Schramm, V. L. (2007) Binding isotope effects: Boon and bane, Curr Opin Chem
Biol 11, 529-536.
24. Cook, P. F., Oppenheimer, N. J., and Cleland, W. W. (1981) Secondary deuterium
and nitrogen-15 isotope effects in enzyme-catalyzed reactions. Chemical mechanism
of liver alcohol dehydrogenase, Biochemistry 20, 1817-1825.
15

25. Snider, M. J., Reinhardt, L., Wolfenden, R., and Cleland, W. W. (2002) N kinetic
isotope effects on uncatalyzed and enzymatic deamination of cytidine, Biochemistry
41, 415-421.

58

26. Luo, M., and Schramm, V. L. (2008) Transition state structure of E. coli tRNAspecific adenosine deaminase, J Am Chem Soc 130, 2649-2655.
27. Luo, M., Singh, V., Taylor, E. A., and Schramm, V. L. (2007) Transition-state
variation in human, bovine, and Plasmodium falciparum adenosine deaminases, J Am
Chem Soc 129, 8008-8017.
28. Merkler, D. J., Kline, P. C., Weiss, P., and Schramm, V. L. (1993) Transition-state
analysis of AMP deaminase, Biochemistry 32, 12993-13001.
29. Singleton, D. A., and Thomas, A. A. (1995) High-precision simultaneous
determination of multiple small kinetic isotope effects at natural abundance, J Am
Chem Soc 117, 9357-9358.
30. Fersht, A. (1999) Structure and mechanism in protein science: A guide to enzyme
catalysis and protein folding. W. H. Freeman, New York.
31. DelMonte, A. J., Haller, J., Houk, K. N., Sharpless, K. B., Singleton, D. A., Strassner,
T., and Thomas, A. A. (1997) Experimental and theoretical kinetic isotope effects for
asymmetric dihydroxylation. Evidence supporting a rate-limiting “(3+2)”
cycloaddition, J Am Chem Soc 119, 9907-9908.
32. Singleton, D. A., Hang, C., Szymanski, M. J., Meyer, M. P., Leach, A. G., Kuwata,
K. T., Chen, J. S., Greer, A., Foote, C. S., and Houk, K. N. (2003) Mechanism of ene
reactions of singlet oxygen. A two-step no-intermediate mechanism, J Am Chem Soc
125, 1319-1328.
33. Meyer, M. P., DelMonte, A. J., and Singleton, D. A. (1999) Reinvestigation of the
isotope effects for the claisen and aromatic claisen rearrangements: The nature of the
claisen transition states, J Am Chem Soc 121, 10865-10874.
34. Nowlan, D. T., Gregg, T. M., Davies, H. M. L., and Singleton, D. A. (2003) Isotope
effects and the nature of selectivity in rhodium-catalyzed cyclopropanations, J Am
Chem Soc 125, 15902-15911.
35. Singleton, D. A., Merrigan, S. R., Kim, B. J., Beak, P., Phillips, L. M., and Lee, J. K.
13
(2000)
C kinetic isotope effects and the mechanism of the uncatalyzed
decarboxylation of orotic acid, J Am Chem Soc 122, 3296-3300.
36. Keating, A. E., Merrigan, S. R., Singleton, D. A., and Houk, K. N. (1999)
Experimental proof of the non-least-motion cycloadditions of dichlorocarbene to
alkenes: Kinetic isotope effects and quantum mechanical transition states, J Am Chem
Soc 121, 3933-3938.
13

37. Frantz, D. E., Singleton, D. A., and Snyder, J. P. (1997) C kinetic isotope effects
for the addition of lithium dibutylcuprate to cyclohexenone. Reductive elimination is
rate-determining, J Am Chem Soc 119, 3383-3384.
38. Brecker, L., Kögl, M. F., Tyl, C. E., Kratzer, R., and Nidetzky, B. (2006) NMR study
59

13

13

of C kinetic isotope effects at C natural abundance to characterize oxidations and
an enzyme-catalyzed reduction, Tetrahedron Lett 47, 4045-4049.
39. Yao, L., Li, Y., Wu, Y., Liu, A., and Yan, H. (2005) Product release is rate-limiting
in the activation of the prodrug 5-fluorocytosine by yeast cytosine deaminase,
Biochemistry 44, 5940-5947.
40. Spectral Database for Organic Compounds, SDBS,
http://riodb01.ibase.aist.go.jp/sdbs/cgi-bin/cre_index.cgi
41. Gottlieb, H. E., Kotlyar, V., and Nudelman, A. (1997) NMR chemical shifts of
common laboratory solvents as trace impurities, J Org Chem 62, 7512-7515.
42. Carlos Cobas, J., Bernstein, M. A., Martín-Pastor, M., and Tahoces, P. G. (2006) A
new general-purpose fully automatic baseline-correction procedure for 1D and 2D
NMR data, J Magn Reson 183, 145-151.
43. Claridge, T. (2009) Software review of MNova: NMR data processing, analysis, and
prediction software. J Chem Inf Model 49, 1136-1137.
44. Humbel, S., Sieber, S., and Morokuma, K. (1996) The IMOMO method: Integration
of different levels of molecular orbital approximations for geometry optimization of
!
large systems: Test for n-butane conformation and SN2 reaction: RCl+Cl , J Chem
Phys 105, 1959-1967.
45. Svensson, M., Humbel, S., Froese, R. D. J., Matsubara, T., Sieber, S., and
Morokuma, K. (1996) ONIOM: A multilayered integrated MO + MM method for
geometry optimizations and single point energy predictions. A test for Diels!Alder
reactions and Pt(P(t-Bu)3)2 + H2 oxidative addition, J Phys Chem 100, 19357-19363.
46. Dapprich, S., Komáromi, I., Byun, K. S., Morokuma, K., and Frisch, M. J. (1999) A
new ONIOM implementation in Gaussian98. Part I. The calculation of energies,
gradients, vibrational frequencies and electric field derivatives, J Mol Struct:
THEOCHEM 461-462, 1-21.
47. Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A.,
Cheeseman, J. R., Montgomery, J., Vreven, T., Kudin, K. N., Burant, J. C., Millam, J.
M., Iyengar, S. S., Tomasi, J., Barone, V., Mennucci, B., Cossi, M., Scalmani, G.,
Rega, N., Petersson, G. A., Nakatsuji, H., Hada, M., Ehara, M., Toyota, K., Fukuda,
R., Hasegawa, J., Ishida, M., Nakajima, T., Honda, Y., Kitao, O., Nakai, H., Klene,
M., Li, X., Knox, J. E., Hratchian, H. P., Cross, J. B., Bakken, V., Adamo, C.,
Jaramillo, J., Gomperts, R., Stratmann, R., Yazyev, O., Austin, A. J., Cammi, R.,
Pomelli, C., Ochterski, J. W., Ayala, P. Y., Morokuma, K., Voth, G. A., Salvador, P.,
Dannenberg, J. J., Zakrzewski, V. G., Dapprich, S., Daniels, A. D., Strain, M. C.,
Farkas, O., Malick, D. K., Rabuck, A. D., Raghavachari, K., Foresman, J. B., Ortiz, J.
V., Cui, Q., Baboul, A. G., Clifford, S., Cioslowski, J., Stefanov, B. B., Liu, G.,
Liashenko, A., Piskorz, P., Komaromi, I., Martin, R. L., Fox, D. J., Keith, T., AlLaham, M. A., Peng, C. Y., Nanayakkara, A., Challacombe, M., Gill, P. M. W.,
60

Johnson, B., Chen, W., Wong, M. W., Gonzalez, C., and Pople, J. A. (2004) Gaussian
03, Revision D.01. Gaussian, Inc., Wallingford CT.
48. Case, D. A., Darden, T. A., Cheatham, I., Simmerling, C. L., Wang, J., Duke, R. E.,
Luo, R., Crowley, M., Walker, R. C., Zhang, W., Merz, K. M., Wang, B., Hayik, S.,
Roitberg, A., Seabra, G., Kolossv¨¢ry, I., Wong, K. F., Paesani, F., Vanicek, J., Wu,
X., Brozell, S. R., Steinbrecher, T., Gohlke, H., Yang, L., Tan, C., Mongan, J.,
Hornak, V., Cui, G., Mathews, D. H., Seetin, M. G., Sagui, C., Babin, V., and
Kollman, P. A. (2008) AMBER 10. University of California, San Francisco.
49. Vreven, T., Byun, K. S., Komáromi, I., Dapprich, S., Montgomery, Morokuma, K.,
and Frisch, M. J. (2006) Combining quantum mechanics methods with molecular
mechanics methods in ONIOM, J Chem Theory Comput 2, 815-826.
50. Anisimov, V., and Paneth, P. (1999) ISOEFF98. A program for studies of isotope
effects using Hessian modifications, J Math Chem 26, 75-86.
51. Frick, L., Yang, C., Marquez, V. E., and Wolfenden, R. (1989) Binding of pyrimidin2-one ribonucleoside by cytidine deaminase as the transition-state analogue 3,4dihydrouridine and the contribution of the 4-hydroxyl group to its binding affinity,
Biochemistry 28, 9423-9430.
52. Betts, L., Xiang, S., Short, S. A., Wolfenden, R., and Carter, C. W. (1994) Cytidine
deaminase. The 2.3 Å crystal structure of an enzyme: Transition-state analog
complex, J Mol Biol 235, 635-656.
53. Xiang, S., Short, S. A., Wolfenden, R., and Carter, C. W. (1995) Transition-state
selectivity for a single hydroxyl group during catalysis by cytidine deaminase,
Biochemistry 34, 4516-4523.
54. Wilson, D. K., Rudolph, F. B., and Quiocho, F. A. (1991) Atomic structure of
adenosine deaminase complexed with a transition-state analog: understanding
catalysis and immunodeficiency mutations, Science 252, 1278-1284.
55. Kim, J., Malashkevich, V., Roday, S., Lisbin, M., Schramm, V. L., and Almo, S. C.
(2006) Structural and kinetic characterization of Escherichia coli TadA, the Wobblespecific tRNA deaminase, Biochemistry 45, 6407-6416.
56. Liaw, S., Chang, Y., Lai, C., Chang, H., and Chang, G. (2004) Crystal structure of
Bacillus subtilis guanine deaminase: the first domain-swapped structure in the
cytidine deaminase superfamily, J Biol Chem 279, 35479-35485.
57. Almatarneh, M. H., Flinn, C. G., and Poirier, R. A. (2008) Mechanisms for the
!
!
deamination reaction of cytosine with H2O/OH and 2H2O/OH : A computational
study, J Chem Inf Model 48, 831-843.
58. Rose, I. A. (1980) The isotope trapping method: Desorption rates of productive ES
complexes, Meth Enzymol 64, 47-59.

61

CHAPTER 3

REACTION PATHWAY OF DEAMINATION OF 5FC:
ONIOM CALCULATION REVISITED

IINTRODUCTION
yCD is of great biomedical interest because of its potential application in genedirected enzyme prodrug therapy (GDEPT). yCD catalyzes the conversion of 5FC to
5FU, which is widely used in the treatment of breast, rectal, stomach, colon, and
pancreatic cancers (1, 2). In GDEPT, cancer cells gain cytosine deaminase activity after a
yCD gene is introduced. When the nontoxic prodrug 5FC is administered, it is converted
to the active and lethal form 5FU. This strategy takes advantage of the fact that there is
no cytosine deaminase expressed in mammalian cells (3). Thus, in principle, only the
cancer cells are exposed to the toxic 5FU and are killed, and the normal cells are not
affected.
Therapeutic efficacy of GDEPT is affected by many factors, such as efficiency of
gene delivery, and stability and half-life of the enzyme, among which the kinetic activity
of enzyme is particularly important. The kinetic properties of yCD have been studied by
steady-state kinetic experiments for both cytosine and 5FC as substrates (4). 5FC is a
better substrate than cytosine regarding binding affinity. The Km value for 5FC is 7-fold

62

lower than that for cytosine. However, the turnover rate of 5FC is 5-fold slower than that
for cytosine and the catalytic efficiency (kcat/Km) of 5FC is only slightly higher than that
for cytosine. Hence, enzyme variants with improved catalytic efficiency for 5FC and/or
substrate preference towards 5FC are desirable candidates in GDEPT.
Efforts have been attempted to address substrate specificity towards 5FC by
random mutagenesis, directed evolution, and protein engineering. In one such study(5), a
yCD variant (D92E) was successfully designed with relative specificity for 5FC over
cytosine, although the variant is not better than the wild-type enzyme in cell sensitivity
assay. More promising variants have been found for Escherichia coli cytosine deaminase
(eCD), the counterpart of yCD used in combination with 5FC in GDEPT (6-8). For
example, the relative specificity of substrate preference towards 5FC by the eCD/D314A
increases 19-fold compared with that by the wild-type eCD.
To facilitate design of potential candidates in GDEPT, the reaction mechanism for
the deamination of 5FC and cytosine have been investigated. In Chapter 2, we have
shown that the deamination of cytosine by yCD proceeds by a stepwise AN+DN
mechanism, with the transition state involving C4-N4 bond breakage. The detailed
reaction pathway was proposed by previous two-layered ONIOM(B3LYP:PM3)
calculations (9). The results suggest that the C4-N4 bond breakage step is the ratelimiting step in generating the zinc-coordinated uracil. The Zn-O4 of uracil bond
breakage was further investigated by a two-stepped ONIOM MD calculation (10), which
shows that the activation energy for the bond breakage is only 2.9 kcal/mol, indicating
that product release is not the rate-limiting step for the overall reaction. The catalytic
mechanism of the deamination of 5FC by yCD is different as revealed by transient kinetic

63

and NMR studies (4). The results show that product release is indeed the rate-limiting
step. The rate constants for the forward reaction involving the chemical step are at least 8
times that of the product release step. NMR spectra for yCD in the free form and in the
complex form with 5FU were acquired and analyzed. The results show that the Kd of
5FU with yCD is estimated to be around 20 mM, indicating a low affinity of 5FU with
yCD. NMR saturation transfer experiments show that 5FU is in slow exchange with its
yCD complex on the NMR time scale. The association rate constant and dissociation rate
-1 -1

-1

constant are 0.6 mM s and 14 s , respectively.
In this study, an ONIOM electronic embedding (EE) scheme (11) was employed
to investigate the reaction mechanism of the deamination of 5FC. The same strategy was
applied on cytosine as well to enable direct comparison between the reaction
mechanisms. By doing so, we expect to understand the difference between the catalytic
mechanisms for 5FC and cytosine as substrate and hopefully to provide a guidance for
rational protein design of effective enzyme variants specific towards 5FC.

COMPUTATIONAL METHODS
The reaction pathways for conversion of 5FC to 5FU were calculated by the
ONIOM method (12-14) as implemented in the Gaussian program (Gaussian 03) (15).
The initial structure is from the high-resolution crystal structure of yCD in complex with
DHP (PDB code:1p6o) (16). Instead of only twenty-two amino acids being included, as
in the previous ONIOM calculation (9), the whole protein was used to construct the
molecular system to incorporate the protein environmental effects (17). The detailed
methodology has been described in Chapter 2. In the following paragraphs, I will only
64

describe briefly how to the calculations were set up.
Two-layered ONIOM (B3LYP/6-31G*:Amber) - mechanical embedding (ME)
calculations were performed for the yCD inhibitor complexes, yCD substrate complexes,
yCD product complexes and a series of intermediate complexes. The structures were
further optimized by an ONIOM (B3LYP/6-31G*:Amber) - electronic embedding (EE)
scheme and the ONIOM energies were calculated. We used scans to find approximate
geometries of the transition states rather than optimization. Consequently, the calculated
values of the barriers are their upper bounds.

RESULTS AND DISCUSSION
yCD-DHP Complex. ONIOM calculations were first performed on the yCDDHP complex. Key coordinates in the ONIOM optimized structure for the yCD-DHP
complex were summarized and compared with those of the crystal structures in Table 31. The results show that both ME and EE schemes are sufficient to describe the geometry
of the molecular system. Since the zinc atom and its coordination were completely
included in the inner layer, they were described at the B3LYP energy level. The bond
distances and angles between Zn and its coordination are consistent with those in the
crystal structures and those calculated in the previous calculations. Similarly, the
conformation of DHP was well described too. DHP adopts a tetrahedral geometry in
which the C4 atom shows obvious sp3 characteristic. The dihedral angles !O4-C4-N3C2 and !O4-C4-C5-C6 are 111° and -107°, respectively, in close agreement with those
in the crystal structure (Table 3-1). O4-C4 is clearly a single bond with a distance of 1.45
Å.

65

DHP is stabilized by a number of hydrogen bonds. O4 of DHP forms a strong
!1

hydrogen bond with O

of Glu64 with a distance of 2.49 Å. The hydrogen bond is

believed to be a low barrier hydrogen bond (LBHB). Indeed, a peak at 16.2 ppm was
observed on a 1D proton NMR spectrum of the yCD-DHP complex (unpublished data).
Normally, a hydrogen bond showing a proton peak in the range between 16 and 22 ppm
is considered a LBHB. Additional experiments are being constructed to further
investigate the nature of this hydrogen bond. The bond distances of other hydrogen bonds
between DHP and the surrounding residues are in close agreement with those
experimentally determined.
Superposition of the optimized structure onto the crystal structure (Figure 3-1)
shows that most residues around the active site are well aligned. Different from the
previous ONIOM calculation, in which all residues in the outer layer were fixed during
optimization, the residues in the outer layer and within a radius of 10 Å of DHP were not
restrained. The results show that removal of the restraints did not lead to unreasonable
conformation. On the contrary, these residues might adopt a more realistic conformation.
For example, the three residues, Trp152, I156 and F114, which cover the active site,
move slightly away from their positions in the crystal structure. These residues sit in
either the C-terminal helix or the F114 loop, both of which are very flexible fragments,
suggested by NMR relaxation experiments (Yao, unpublished data). Binding of DHP
may rigidify these fragments through hydrogen bonding, electrostatic and Van der Waals
interactions.
Deamination of 5FC. yCD catalyzes the deamination of 5FC to 5FU, as well as
the deamination of cytosine to uracil. The latter has been analyzed by previous ONIOM

66

calculations and a reasonable pathway has been proposed. In this work, we studied the
deamination pathway for 5FC by yCD using a different ONIOM scheme. The
deamination for cytosine was investigated as well in order to compare with the reaction
mechanism of the deamination of 5FC.
ONIOM calculations were performed on a series of yCD substrate, intermediate
and product complexes, including (1) water/5FC complex, (2) hydroxide/Glu64H 5FC
complex, (3) hydroxide/Glu64 5FC-H complex, (4) protonated tetrahedral intermediate
complex, (5) unprotonated tetrahedral intermediate complex, (6) 5FU/ammonia complex,
(7) 5FU/water complex and (8) water/5FU complex. Similarly, the calculations were
performed for the deamination of cytosine on the same complexes with cytosine and its
derivatives. The reaction schemes along the reaction pathways are proposed in Figure 3-2
and the optimized structures for each species are given in Figure 3-3 and 3-4. Same as the
previous ONIOM calculation, we used scans to find approximate geometries of the
transition states rather than optimization. Consequently, the calculated values of the
barriers are their upper bounds.
The reaction pathway starts from the yCD water/5FC complex (1). In 1, 5FC is
stabilized by a number of hydrogen bonds. At one end of the pyrimidine ring, N1 forms a
!1

hydrogen bond with O

!2

of D155 and O2 forms a hydrogen bond with N
"1

the other end, N4 forms a hydrogen bond with O

of Asn51. At

of Glu64. It is also stabilized by

hydrophobic interaction between, C5 and C6 of 5FC, and the surrounding nonpolar
"2

residues. The zinc-coordinated water forms a hydrogen bond with O

of Glu64. The
$

%

!

"2

Glu64 side chain can actually adapt two conformations with different #C -C -C -O
67

"

dihedral angles. In the first conformation (1), the Glu64 side chain dihedral angle (!C #

$

%2

C -C -O ) is 3.5°. To adapt to the other conformation (1A), this angle needs to rotate
clockwise by 83°. The energy barrier for the rotation is trivial.
The substrate 5FC is activated by protonating its N3 atom via a two-step proton
%2

transfer. First, a proton of the zinc-coordinated water is transferred to O

of Glu64 to

form the hydroxide/Glu64H 5FC complex (2). In 2, the Glu64 side chain rotates 36° to
form a stable hydrogen bond with N3 of 5FC. 2 is less stable than 1 by 3.9 kcal/mol at
ONIOM (Figure 3-5). The E

model,QM

energy for the model system (the inner layer plus

the link atoms) was calculated to be 5.5 kcal/mol higher than 1. The transition state
%2

(TS12) is found in the stage where the hydrogen bond between O
%2

and a new hydrogen bond between O

and O

Zn

is breaking

and N3 is forming (Figure 3-2). The energy

barrier was calculated to be 7.7 kcal/mol, suggesting the proton transfer is not a trivial
event. The difference of E

model,QM

between 1 and TS12 is 11.0 kcal/mol. The energy

difference (3.4 kcal/mol) indicates that the model system is stabilized by the protein
environment (the outer layer).
%2

In the next step, the same proton is transferred from O

to N3 to form the

complex 3. 3 is relatively more stable than 2 by 0.3 kcal/mol at ONIOM. E

model,QM

of 3

is 1.8 kcal/mol lower than 2. The transition state (TS23) is found when the proton is in
the middle of the two heavy atoms. The energy barrier is 2.9 kcal/mol at ONIOM and 0.8

68

kcal/mol on the E

model,QM

energy surface. The results indicate that the environment does

not contribute too much for this step.
Zn

The distance between O

(hydroxide) and C4 is reduced from 2.94 Å, to 2.66 Å,

to 2.53 Å from 1, to 2, to 3. In 3, the zinc-coordinated hydroxide is well positioned for
the nucleophilic attack on C4. The transition state (TS34) is found when the distance
between O4 and C4 is reduced to 1.88 Å. The corresponding energy barrier is calculated
to be 2.9 kcal/mol at ONIOM and 6.7 kcal/mol on the E

model,QM

energy surface. The

energy difference of 3.8 kcal/mol manifests the contribution from the protein
environment for the nucleophilic attack process. Complex 4 (Figure 3-3) is formed after
!1

nucleophilic attack. In 3, O

of Glu64 is involved in the hydrogen bond with N4 as a
!1

hydrogen acceptor. In 4, the hydrogen bond is disrupted and O

forms a new hydrogen

bond with O4. 4 is found to be more stable than 3 by 5.6 kcal/mol at ONIOM and by 4.5
kcal/mol on the E

model,QM

energy surface. Again, the protein stabilizes the model system

by 1.1 kcal/mol.
!1

Complex 5 is formed after O

of Glu64 abstracts the proton from O4 and forms a

new hydrogen bond with N4 as hydrogen donor (Figure 3-3). The proton is already
!1

transferred to O

at the transition state. The energy barrier is mainly contributed to the
!1

hydrogen bond breakage between O

and O4 and the rotation of the H41-N4 bond. The

energy barrier is calculated to be 3.9 kcal/mol at ONIOM and 5.2 kcal/mol on the

69

E

model,QM

energy surface. The difference between 4 and 5 is the protonation state of the

intermediate. In 4, the intermediate is protonated at O4, and a hydrogen bond is presented
!1

between O

of Glu64 and O4. In 5, the intermediate is unprotonated and a hydrogen
!1

bond is present between O

and N4. 5 is more stable than 4 by 5.2 kcal/mol at ONIOM.

However, the model system in 5 is less stable than that in 4 by 5.3 kcal/mol. The
difference of 10.5 kcal/mol is due to the stabilization effect of the protein environment.
The bond cleavage between N4 and C4 is initialized by the proton transfer from
!1

O

to N4. After the proton transfer, N4 keeps moving apart from C4. The transition state

(TS56) is found when the bond distance is about 2.05 Å. 6 is formed after the N4-C4
!1

bond is completely cleaved. The product ammonia is hydrogen bonded with O

of

Glu64. The other product 5FU is covalently bonded with the zinc atom. The energy
barrier for this step was calculated to be 5.1 kcal/mol at ONIOM. The energy barrier on
the E

model,QM

energy surface is 3.0 kcal/mol. However, it is not for the same transition

state. The highest point is found on the energy surface during the proton transfer from
!1

O

to N4.
Dynamics of the Active Site. Throughout the deamination of 5FC, one end of

5FC ring (the N1 and O2 atoms) is anchored by Asn51 and Asp155 via two hydrogen
bonds with N1 and O2, respectively. The other end (the C4 atom) rotates towards the
zinc-coordinated water molecule and eventually forms a covalent bond (Figure 3-6). The
ring rotates about 39° from the initial position. Obviously, the anchoring residues can

70

stabilize the enzyme-substrate complex. The transition state structure is stabilized as well,
as suggested by mutagenesis studies which show that substituting Asn51 and Asp155
results in great reduction of turnover rate (Chapter 6). Gln156 in eCD functions in a
similar way, which forms hydrogen bonds with N1 and O2 by the single side-chain amide
group (18). The function of these two residues is further investigated and presented in
Chapter 6.
Dynamic effect of the surrounding residues was investigated. Since only energy
optimizations were performed in the ONIOM calculations, the dynamic effects cannot be
directly observed and evaluated. Instead, it was implied by the movements of the
surrounding residues along the reaction pathway. Three fragments, including residues 32
to 34, 62 to 65 and 88 to 91, in the vicinity of the substrate, oscillate gently to
accommodate the conformational changes of the reaction species (Figure 3-7). These
motions may represent the minimum dynamic efforts required for the reaction, since
these fragments are the most rigid regions according to

15

N NMR relaxation experiments

(Yao, unpublished data) and molecular dynamics simulations. There are close contacts
between these fragments and the substrate. The main chain amide group of Gly63 forms a
hydrogen bond with O2. The main chain C=O group of Ser89 forms a hydrogen bond
with N3. The main chain amide of Cys91 forms a hydrogen bond with the zinccoordinated water. I33 sandwiches the 5FC ring with the side chain of His62 on the other
side. With the progress of the reaction, the position and conformation of these residues
change accordingly.
Motions of side chains of F114 and W152 were also observed. These residues sit
in the C-terminal helix and the F114 loop, respectively. The conformational change of

71

these two residues did not prolong to the C-terminal helix and the F114 loop. These two
regions are quite flexible according to the

15

N NMR relaxation experiments and MD

simulations. The movements of these regions are required for the substrate binding and
product release (10). The dynamics may not be equally important for the chemical steps
in the catalysis. However, their presence is definitely important for maintaining the
hydrophobic environment near the C5-C6 side of the pyrimidine ring. Substituting either
Phe114 or Trp152 results in reduced turnover rate (unpublished data).
Comparison the Deamination of 5FC and Cytosine. The first transition state
(TS12) from complex 1 and 2 for the deamination of 5FC is surprisingly high, with a
value of 7.7 kcal/mol at ONIOM. The approximate geometry of TS12 is found by
Zn

scanning the distance between O

!1

and HW (which is the proton transferred to O

of

Glu64). To examine whether it is only for the deamination of 5FC or not, we also
performed calculations on the deamination of cytosine using the same scheme. The
results show that the energy barrier is almost the same (7.6 kcal/mol). And the E

model,QM

energy is quite similar too (11.0 kcal/mol for 5FC and 10.1 kcal/mol for cytosine). It
basically means the energy barrier for the first proton transfer is not determined by the
Zn

substrate. However, the presence of the substrate may prevent the other O -HW bond
!2

from rotating and forming a hydrogen bond with O

of Glu64 and, hence, results in an

unrealistically high energy barrier.
The energy barrier for the proton transfer from 2 to 3 is very similar for these two
reactions too. This is reasonable since the only difference between 5FC and cytosine is

72

the fluorine substitution on C5, which is two bonds away from N3. Although fluorine has
strong electronegativity, it has little effect on the protonation state of N3.
On the other hand, more pronounced effect of fluorine substitution on the
tetrahedral intermediate formation in the next step is evidenced by the energy profile. The
energy barrier (TS34) from 3 to 4 for the deamination of 5FC is only 2.9 kcal/mol, 2.4
kcal/mol lower than that for cytosine. Fluorine pulls electrons towards itself, leading the
C4 atom more liable to nucleophilic attack by the zinc-coordinated hydroxide. Under the
same logic, 4 is more stable than 3 by 5.6 kcal/mol for 5FC. While for cytosine, 4 is only
more stable than 3 by 1 kcal/mol.
+

For the C4-N4 bond cleavage, a positively charged leaving group (-NH3 ) is more
likely to depart from 5FC, because C4 is more positive due to the presence of fluorine on
C5. The energy barrier is 5.1 kcal/mol for this step for the deamination of 5FC, while it is
6.3 kcal/mol for cytosine.
Steady-state kinetic constants have been measured for both 5FC and cytosine (4).
The Km for cytosine (1.1 mM) is about 7 times higher than that for 5FC (0.16 mM). For a
simple Michaelis-Menten reaction, Km is highly related to the binding affinity between
the enzyme and substrate. While for a reaction involving multiple intermediates, like the
deamination of 5FC by yCD, Km is not just that of the enzyme-substrate complex, but
rather reflects the overall binding affinity with all intermediate complexes (19). In the
deamination of 5FC, relatively stable intermediate complexes 4 and 5 are formed during
the reaction (Figure 3-5), which results in a smaller Km for 5FC. This finding is
interesting for protein engineering of effective yCD variants specific towards 5FC. If an

73

enzyme variant can lead to strong interaction with these intermediates, it would probably
be a better candidate for GDEPT.
For the deamination of 5FC, the energy barriers for each step are smaller when
compared with the same step for the deamination of cytosine. If the C4-N4 cleavage step
is the rate-limiting step for both reactions, the overall turnover rate kcat of 5FC would be
-1

higher than that of cytosine. However, kcat of 5FC is 17 s , 5 times lower than that of
-1

cytosine (91 s ). This means this step cannot be rating-limiting for the deamination of
5FC. Indeed, transient kinetic experiments and NMR experiments have shown that
product-release is the rate-limiting step for the deamination of 5FC. KIE studies have
suggested that the C4-N4 cleavage step is the rate-limiting step for the deamination of
cytosine. A two-step ONIOM MD simulation has suggested product release is not the
rate-limiting step for the deamination of cytosine. The energy barrier is calculated to be
2.9 kcal/mol. However, the energy barrier for the C4-N4 cleavage step is not compared
using the same protocol. In the next section, the Zn-5FU bond breakage is evaluated
using the same scheme as shown above.
Zn-5FU Bond Breakage. Ammonia can rapidly diffuse into the solution once it
is generated (20). However, direct cleavage of the bond between 5FU and Zn is
surprisingly hard according to the previous ONIOM calculation. An alterative mechanism
involving oxygen exchange was proposed and is energetically better than direct cleavage.
However, oxygen exchange experiments show the exchange rate with 5FU is too slow to
be the main route for the 5FU release. A further study combining ONIOM and MD
simulation shows rearrangement of the active site residues is needed for the Zn-O4 bond

74

cleavage (10), which cannot be achieved in the previous ONIOM calculation due to the
restraints on the outer layer atoms. In that study, the author proposed that 5FU is released
with the help of the carboxyl group of Glu64, which replaces O4 of 5FU to coordinate
with Zn.
In this study, we tried to investigate if the proper transition state can be located by
the ONIOM scheme alone if the active site residues are set free. Moreover, we want to
see if direct replacement of 5FU with a water molecule is a possible pathway (Figure 38).
Complex 7 was constructed by replacing the ammonia with a water molecule. The
reaction pathway was constructed by a series of calculations with different distances
between Zn and OW of water. The optimized structures are presented in Figure 3-9 and
3-10. The energy profile is plotted in Figure 3-11. When the water is approaching Zn, the
ONIOM energy increases accordingly. The energy reaches a plateau when the Zn-OW
distance decreases to ~ 4.0 Å, with a value of more than 20 kcal/mol. The energy of the
system stays at a high level until the Zn-OW distance reduces to 2.6 Å, when the Zn-OW
bond is about to form and the Zn-O4 bond is about to break. Complex 8 is formed when
the water completely replaces 5FU in coordination with Zn. 8 is less stable than 7 by 2.0
kcal/mol.
At first glance, the high energy barrier is reasonable because product release is
supposed to be the rate-limiting step for the overall reaction of deamination of 5FC as
revealed by transient kinetics and NMR experiments. However, the energy profile is too
high to be true. Moreover, a similar energy barrier was obtained for the bond cleavage of
Zn-uracil, which is not the rate-limiting step for the deamination of cytosine as suggested

75

by our KIE studies and the two-stepped ONIOM MD simulations (Figure 3-11).
The E

model,QM

energy is much lower than the ONIOM energy after the Zn-OW

distance reduces to 4.2 Å, suggesting that the protein environment does not help, but
impedes the reaction. In other words, although the surroundings are set free during the
optimization, proper arrangement is not achieved. Figure 3-12 shows that the water
accesses Zn through the space between Glu64 and Pro90. Glu64 and Pro90 are pushed
away from their original positions. Other regions should move accordingly to reach the
conformation of the global minimum along the reaction pathway. However, the
conformation is settled in a nearby local minimum, which results in an unrealistically
high energy at ONIOM. This suggests that for a large rearrangement, a simple ONIOM
optimization cannot properly sample the right reaction path.
Hence, proper sampling technique is needed to assess the Zn-5FU bond breakage
pathway. The two-stepped ONIOM MD approach is a good method and has been
successfully applied to uracil release with the assist of Glu64 carboxyl group. Such an
approach can also be applied on the 5FU release.

CONCLUSIONS
The deamination of 5FC catalyzed by yeast cytosine deaminase was investigated
using the ONIOM (B3LYP/6-31G*:Amber) - EE scheme. In the molecular system, the
whole protein was included in order to properly estimate contributions from the protein
environments.
yCD catalyzes the deamination of 5FC via a sequential mechanism, including
substrate activation, tetrahedral intermediate formation and C4-N4 bond cleavage. 5FC is
!2

activated by protonation at N3 via O

of Glu64. The zinc-coordinated water is converted

76

to a hydroxide after losing a hydrogen atom and is well positioned for nucleophilic attack
on C4. The tetrahedral intermediate is formed after nucleophilic attack and the
!1

deprotonated tetrahedral intermediate is formed after O of Glu64 abstracts the hydrogen
on O4 and forms hydrogen bond with N4. In the next step, the C4-N4 bond is cleaved, in
which ammonia and zinc-coordinated 5FU are generated.
Throughout the deamination of 5FC, one end of the 5FC ring (the N1 and O2
atoms) is anchored by Asn51 and Asp155 through two hydrogen bonds. The other end
(the C4 atom) moves towards the zinc-coordinated water molecule and eventually forms
a covalent bond. From the initial 5FC to the zinc-coordinated 5FU, the ring rotates about
38°. Dynamics of the three fragments, including residues 32 to 34, 62 to 65 and 88 to 91,
in the vicinity of the substrate, may facilitate the progress of the reaction. Interestingly,
these fragments are the most rigid regions in the protein according to NMR relaxation
experiments and molecular dynamics simulations. These motions may represent the
minimum dynamic effects required for the reaction.
The energy profile shows that the tetrahedral intermediate complexes are more
stable than the initial complex during the deamination of 5FC. However, the stabilization
of these intermediates during the deamination of cytosine is relatively weak. It could be
at least part of the reason why Km for cytosine is higher than that of 5FC.
For the Zn-5FU bond breakage, we proposed that a water molecule attacks the
zinc atom as a nucleophile and replaces 5FU in coordination with Zn. The access of the
zinc atom by the water is through the space between Glu64 and Pro90. The close contacts
disturb the conformation of these residues and thus results in unrealistic high energies at

77

ONIOM, though the E

model,QM

energy surface is reasonable. The results suggest that a

sole energy optimization starting from a static structure cannot sample the reaction
pathway if large rearrangements are involved.

78

TABLES AND FIGURES
Table 3-1. Comparison of key coordinates of structures from ONIOM calculations with
a
crystal structures
Internal Coordinate

X-ray
Structure
(1uaq)

!1

2.07, 2.06
2.35, 2.31

Cys91-S

ONIOM
EE

2.03

2.01

2.29, 2.30

2.32

2.36

2.27, 2.25

2.27, 2.27

2.30

2.29

2.06, 2.08
3.02, 3.10
2.69, 2.70

2.05, 2.06
3.02, 3.03
2.69, 2.72

2.05
3.11
2.73

2.03
3.13
2.73

#2

2.60, 2.45

2.49, 2.52

2.48

2.49

#1

2.82, 2.81

2.80, 2.78

2.94

2.93

!2

2.84, 2.79
2.96, 2.95

2.89, 2.87
2.92, 2.90

2.89
3.07

2.84
3.03

!1

"

ONIOM
ME

#1

His62-N

X-ray
Structure
(1p6o)
2.01, 2.00

2.67, 2.70

2.69, 2.68

2.73

2.70

!2

2.95, 2.98

2.92, 2.90

2.89

2.91

1.44, 1.44
1.42, 1.41
1.45, 1.45
1.37, 1.37
1.39, 1.39
1.38, 1.38
1.34, 1.35
1.21, 1.21
110, 109
110, 109
113, 119
-125, -116

1.48, 1.48
1.46, 1.45
1.47, 1.46
1.31, 1.31
1.32, 1.33
1.43, 1.42
1.31, 1.33
1.30, 1.30
107, 108
112, 109
113, 114
-113, -114

1.44
1.45
1.49
1.37
1.38
1.38
1.34
1.24
111
111
105
-110

1.45
1.45
1.49
1.36
1.37
1.38
1.34
1.25
111
111
107
-111

1.07
1.40

1.06
1.43

··· Zn
··· Zn

"

Cys94-S ··· Zn
DHP -O4 ··· Zn
Cys91-N ··· O4-DHP
Glu64-N

··· O -Glu64

DHP-O4

··· O -Glu64

DHP-N3
DHP-O2

··· O -Glu64
… N-Gly63

DHP-O2

… N -Asn51

DHP-N1

… O -Asp155
!2

Asn51-N … O -Asp155
DHP-O4 ··· C4-DHP
DHP-N3 ··· C4-DHP
DHP-C5 ··· C4-DHP
DHP-N3 ··· C2-DHP
DHP-N1 ··· C2-DHP
DHP-N1 ··· C6-DHP
DHP-C5 ··· C6-DHP
DHP-O2 ··· C2-DHP
$O4-C4-N3
$O4-C4-C5
$O4-C4-N3-C2
$O4-C4-C5-C6

b

DHP-HO4 ··· O4-DHP
#2

DHP-HO4 ··· O -Glu64

a: Distances between the paired atoms are in Å. Angles and dihedral angles are in degree.
b: Distances in subunits A and B, respectively.

79

Figure 3-1. Superposition of the ONIOM optimized yCD-DHP complex structure with
the crystal structure (PDB id: 1p6o). The backbone is displayed as a gray tube. DHP and
several surrounding residues are displayed as ball and stick. The residues are colored
yellow for the crystal structure.

80

Figure 3-2. Proposed catalytic pathway for the conversion of cytosine (5FC) to Zncoordinated uracil (5FC). (1) water/cytosine complex, (2) hydroxide/Glu64H cytosine
complex, (3) hydroxide/Glu64 cytosine-H complex, (4) protonated tetrahedral
intermediate complex, (5) unprotonated intermediate complex, (6) uracil/ammonia
complex.
81

Figure 3-3. ONIOM optimized structures for complex 1-6 and the transition state
structures during the deamination of 5FC.

82

Figure 3-4. ONIOM optimized structures for complex 1’-6’ and the transition state
structures during the deamination of cytosine.

83

Figure 3-5. Energy profiles for the conversion of cytosine to the Zn-coordinated uracil
and the conversion of 5FC to the Zn-coordinated 5FC.

84

Figure 3-6. Rotation of the 5FC ring during the deamination of 5FC. The initial substrate
is colored yellow. N1 and O2 are anchored by N51 and D155. The other end (C4) rotates
towards the zinc-coordinated water and forms a covalent bond with O4. The arrow shows
the direction how the ring rotates along the reaction coordinate.

85

Figure 3-7. Motion of the active site residues during the deamination of 5FC. Several
residues need to move around to accommodate the change of the intermediate (pointed by
red arrow). These residues are within the most rigid regions. Motion of W152 and F114
is also observed. Residues in the initial complex are colored yellow.

86

Figure 3-8. Proposed catalytic pathways for the replacement of the Zn-coordinated uracil
or 5FU with water.

87

Figure 3-9. ONIOM optimized structures for Complex 7, 8 and the transition state
structure for the deamination of 5FC.

Figure 3-10. ONIOM optimized structures for Complex 7’, 8’ and the transition state
structure for the deamination of cytosine.

88

Figure 3-11. Energy profiles for replacement of the Zn-coordinated uracil or 5FU with
water. ONIOM energies for 5FU and uracil are represented by black and blue curves,
model,QM
respectively. E
energies for 5FU and uracil are represented by red and green
curves, respectively.

89

Figure 3-12. Conformational change caused by the access of the water towards the zinc
atom. (top) the water molecule enters the space between Pro90 and Glu64 to access the
zinc atom. (bottom) Motion of fragments to allow the access of the water.

90

REFERENCES

91

REFERENCES
1. Greco, O., and Dachs, G. U. (2001) Gene directed enzyme/prodrug therapy of cancer:
Historical appraisal and future prospectives, J Cell Physiol 187, 22-36.
2. Aghi, M., Hochberg, F., and Breakefield, X. (2000) Prodrug activation enzymes in
cancer gene therapy, J Gene Med 2, 148-164.
3. Ko, T., Lin, J., Hu, C., Hsu, Y., Wang, A., and Liaw, S. (2003) Crystal structure of
yeast cytosine deaminase - Insights into enzyme mechanism and evolution, J Biol
Chem 278, 19111-19117.
4. Yao, L., Li, Y., Wu, Y., Liu, A., and Yan, H. (2005) Product release is rate-limiting
in the activation of the prodrug 5-fluorocytosine by yeast cytosine deaminase,
Biochemistry 44, 5940-5947.
5. Stolworthy, T., Korkegian, A., Willmon, C., Ardiani, A., Cundiff, J., Stoddard, B.,
and Black, M. (2008) Yeast cytosine deaminase mutants with increased
thermostability impart sensitivity to 5-fluorocytosine, J Mol Biol 377, 854-869.
6. Mahan, S., Ireton, G., Knoeber, C., Stoddard, B., and Black, M. (2004) Random
mutagenesis and selection of Escherichia coli cytosine deaminase for cancer gene
therapy, Protein Eng Des Sel 17, 625-633.
7. Mahan, S. D., Ireton, G. C., Stoddard, B. L., and Black, M. E. (2004) Alaninescanning mutagenesis reveals a cytosine deaminase mutant with altered substrate
preference, Biochemistry 43, 8957-8964.
8. Fuchita, M., Ardiani, A., Zhao, L., Serve, K., Stoddard, B., and Black, M. (2009)
Bacterial cytosine deaminase mutants created by molecular engineering show
improved 5-fluorocytosine-mediated cell killing in vitro and in vivo, Cancer Res 69,
4791-4799.
9. Sklenak, S., Yao, L., Cukier, R., and Yan, H. (2004) Catalytic mechanism of yeast
cytosine deaminase: An ONIOM computational study, J Am Chem Soc 126, 1487914889.
10. Yao, L., Yan, H., and Cukier, R. (2006) A combined ONIOM quantum chemicalmolecular dynamics study of zinc-uracil bond breaking in yeast cytosine deaminase,
J Phys Chem B 110, 26320-26326.
11. Vreven, T., Byun, K. S., Komáromi, I., Dapprich, S., Montgomery, Morokuma, K.,
and Frisch, M. J. (2006) Combining quantum mechanics methods with molecular
mechanics methods in ONIOM, J Chem Theory Comput 2, 815-826.
12. Humbel, S., Sieber, S., and Morokuma, K. (1996) The IMOMO method: Integration
of different levels of molecular orbital approximations for geometry optimization of
large systems: Test for n-butane conformation and SN2 reaction: RCl+Cl-, J Chem
92

Phys 105, 1959-1967.
13. Svensson, M., Humbel, S., Froese, R. D. J., Matsubara, T., Sieber, S., and
Morokuma, K. (1996) ONIOM: A multilayered integrated MO + MM method for
geometry optimizations and single point energy predictions. A test for Diels-Alder
reactions and Pt(P(t-Bu)3)2 + H2 oxidative addition, J Phys Chem 100, 19357-19363.
14. Dapprich, S., Komáromi, I., Byun, K. S., Morokuma, K., and Frisch, M. J. (1999) A
new ONIOM implementation in Gaussian98. Part I. The calculation of energies,
gradients, vibrational frequencies and electric field derivatives, J Mol Struct:
THEOCHEM 461-462, 1-21.
15. Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A.,
Cheeseman, J. R., Montgomery, J., Vreven, T., Kudin, K. N., Burant, J. C., Millam, J.
M., Iyengar, S. S., Tomasi, J., Barone, V., Mennucci, B., Cossi, M., Scalmani, G.,
Rega, N., Petersson, G. A., Nakatsuji, H., Hada, M., Ehara, M., Toyota, K., Fukuda,
R., Hasegawa, J., Ishida, M., Nakajima, T., Honda, Y., Kitao, O., Nakai, H., Klene,
M., Li, X., Knox, J. E., Hratchian, H. P., Cross, J. B., Bakken, V., Adamo, C.,
Jaramillo, J., Gomperts, R., Stratmann, R., Yazyev, O., Austin, A. J., Cammi, R.,
Pomelli, C., Ochterski, J. W., Ayala, P. Y., Morokuma, K., Voth, G. A., Salvador, P.,
Dannenberg, J. J., Zakrzewski, V. G., Dapprich, S., Daniels, A. D., Strain, M. C.,
Farkas, O., Malick, D. K., Rabuck, A. D., Raghavachari, K., Foresman, J. B., Ortiz, J.
V., Cui, Q., Baboul, A. G., Clifford, S., Cioslowski, J., Stefanov, B. B., Liu, G.,
Liashenko, A., Piskorz, P., Komaromi, I., Martin, R. L., Fox, D. J., Keith, T., AlLaham, M. A., Peng, C. Y., Nanayakkara, A., Challacombe, M., Gill, P. M. W.,
Johnson, B., Chen, W., Wong, M. W., Gonzalez, C., and Pople, J. A. (2004) Gaussian
03, Revision D.01. Gaussian, Inc., Wallingford CT.
16. Ireton, G., Black, M., and Stoddard, B. (2003) The 1.14 angstrom crystal structure of
yeast cytosine deaminase: Evolution of nucleotide salvage enzymes and implications
for genetic chemotherapy, Structure 11, 961-972.
17. Torrent, M., Vreven, T., Musaev, D. G., Morokuma, K., Farkas, O., and Schlegel, H.
B. (2002) Effects of the protein environment on the structure and energetics of active
sites of metalloenzymes. ONIOM study of methane monooxygenase and
ribonucleotide reductase, J Am Chem Soc 124, 192-193.
18. Ireton, G. C., McDermott, G., Black, M. E., and Stoddard, B. L. (2002) The structure
of Escherichia coli cytosine deaminase, J Mol Biol 315, 687-697.
19. Fersht, A. (1999) Structure and mechanism in protein science: A guide to enzyme
catalysis and protein folding. W. H. Freeman and Company, New York.
20. Yao, L., Sklenak, S., Yan, H., and Cukier, R. (2005) A molecular dynamics
exploration of the catalytic mechanism of yeast cytosine deaminase, J Phys Chem B
109, 7500-7510.

93

CHAPTER 4

QM/MM MD SIMULATIONS SHED NEW INSIGHTS ON
THE CATALYTIC MECHANISM

INTRODUCTION
Yeast cytosine deaminase (yCD) is of significant biomedical interest because it
catalyzes the conversion of the prodrug 5FC to 5FU which is widely used in the treatment
of cancers (1, 2). Combination of yCD and 5FC is a promising strategy in gene directed
enzyme prodrug therapy (GDEPT), through which the administrated nontoxic 5FC is
converted to the toxic and lethal 5FU in cancer cells.
Crystal structures of yCD determined at high resolution revealed that yCD is a
homodimeric metalloenzyme (3, 4). The complete reaction path for the deamination of
cytosine catalyzed by yCD has been proposed by the previous ONIOM calculations (5)
and in Chapter 3. First, the substrate cytosine is activated by protonation at N3 with the
help of the carboxyl group of Glu64. After transferring a proton to N3, the zinccoordinated hydroxide is well positioned for nucleophilic attack on C4 to form a more
stable tetrahedral intermediate, which is further stabilized by Glu64 and other
surrounding residues via a set of hydrogen bonds and hydrophobic interactions. Glu64 is
also involved in proton transfer during the C4-N4 bond cleavage. It abstracts the proton
Zn

from O

and then transfers it to N4 to form ammonia and the zinc-coordinated uracil.

The N4-C4 bond cleavage is suggested to be the rate-limiting step up to the formation of
94

Zn

the zinc-coordinated uracil. The O -Zn bond breakage was studied by the same ONIOM
Zn

calculation. The results show that the activation energy is quite high and the O -Zn
bond breakage is suggested to be the overall rate-limiting step.
A revised ONIOM scheme was employed in Chapter 3 to compare the reaction
mechanism of the deamination of 5FC by yCD with that of cytosine. In the new system,
the whole enzyme was included to incorporate the protein environmental effects (6).
Residues within a radius of 10 Å of the substrate were not restrained in order to
accommodate the rearrangement of the active site along the reaction coordinate. Both
molecular systems were well described before the formation of the zinc-coordinated
Zn

uracil/5FU. However, the energy barrier of the O -Zn bond cleavage was still too high
to be true. The E

model,QM

energy was low enough, suggesting that the protein

environments did not help much in sampling the proper configurations along the
proposed pathway.
Hence, configurational sampling needs to be incorporated while the chemical
transformations are investigated to properly describe a process in which large
conformational changes happen. A number of QM/MM protocols (7-14) have been
developed either at the cost of accuracy to permit more configurational sampling or the
opposite to precisely describe the chemistry.
A two-stepped ONIOM-MD calculation developed by Yao et al (15) was also
Zn

applied on the O -Zn bond breakage. The results show that the activation energy for the
bond breakage is only 2.9 kcal/mol, indicating that product release is not the rate-limiting
step for the overall reaction. Fixing atoms in the outer layer in the first ONIOM
calculation causes unrealistic effects, which were minimized by incorporating
95

configurational sampling in the MD scheme (15). Matsubara and co-workers developed a
method (16) to directly couple the ONIOM method with the molecular dynamics (MD)
method and applied this method to the process of zinc-uracil bond breakage (17). The
Zn

results show that the O -Zn bond abruptly breaks at 11 ps during the simulation. The
authors suggest that the thermal motion of Ile33, which sandwiches uracil with His62,
perturbs the dynamics of uracil through steric contact and contributes to the bond
breakage. A QM/MM MD method implemented in the CHARMM MD package (18) was
performed on the yCD-DHP complex (19) to investigate the origin of the stabilization of
the tight binding of DHP. The authors suggest that the inhibitor is converted to the
alkoxide-like transition-state analogue and is stabilized.
In this chapter, the reaction mechanism of the deamination of cytosine by yCD
was investigated by QM/MM MD simulations, using the fast semiempirical densityfunctional approach with self-consistent charge density functional tight-binding (SCCDFTB) (8, 20) for QM and the Amber PARM99SB force field (21) for MM. The reaction
mechanism was proposed starting from the substrate binding, the tetrahedral intermediate
Zn

formation, and the C4-N4 bond cleavage to the O -Zn bond cleavage. Events were
explored in details and analyzed, and the activation energy for each step was calculated
and incorporated into the energy profile.

EXPERIMENTAL PROCEDURES
Molecular Dynamics Simulation. MD simulations were carried out on the free
form yCD, the yCD-DHP complex and the enzyme complexes with the reaction species,
including (1) water/cytosine complex, (2) hydroxide/Glu64H cytosine complex, (3)
hydroxide/Glu64 cytosine-H complex, (4) protonated tetrahedral intermediate complex,
96

(5) unprotonated intermediate complex, (6) uracil/ammonia complex, (7) uracil/water
complex and (8) water/uracil complex. The starting structures were obtained from the
1.14-Å resolution crystal structure of yCD in complex with the inhibitor DHP (PDB ID:
1p6o) (4). The free form yCD was built by replacing DHP with a zinc-coordinated water
molecule. The water/cytosine complex (1) was built by replacing DHP with a zinccoordinated water and a cytosine. The hydroxide/Glu64H cytosine complex (2) was built
by replacing DHP with a zinc-coordinated hydroxide and a cytosine and changing Glu64
to the protonated form Glu64H. The hydroxide/Glu64 cytosine-H complex (3) was built
by replacing DHP with a zinc-coordinated hydroxide and a protonated cytosine (cytosineH). The protonated tetrahedral intermediate complex (4) was built by replacing DHP with
the zinc-coordinated protonated tetrahedral intermediate. The unprotonated tetrahedral
intermediate complex (5) was built by replacing DHP with the zinc-coordinated
unprotonated tetrahedral intermediate. The uracil/ammonia complex (6) was built by
replacing DHP with a zinc-coordinated uracil and an ammonia molecule. The
uracil/water complex (7) was built by replacing DHP with a zinc-coordinated uracil and a
water molecule. The water/uracil complex (8) was built by replacing DHP with a zinccoordinated water and a uracil molecule. Ten molecular systems were constructed, in
each of which, only the subunit A was modified. Subunit B remained the same as in the
crystal structure.
All crystal water molecules were removed. All ionizable residues were assigned
to their normal protonation state at pH 7.0 except for His50, which was set as doubly
protonated because a carboxyl group presents at each side of the imidazole ring. The
catalytic zinc ion is tetrahedrally coordinated with His62, Cys91, Cys94, and a water
molecule. The sulfur atoms of Cys91 and Cys94 were set as deprotonated and His62 was

97

singly protonated at the epsilon nitrogen. Explicit bonds between the zinc atom and its
coordination were used. The force constants for bonds and bond angles and the charges
of Zn and its ligands have been described previously (22). The protein was placed in a
+

periodic box of 12.5 Å TIP3P water molecules. Four Na ions were placed to counteract
the negative charge of the molecular system. The parallel version Sander in Amber 10
(23) with the PARM99SB (21) force field was employed for the MD simulations.
The energy minimization of the molecular system is divided into three steps. First,
the solvent was minimized with restraints on the protein for 15,000 steps, using the
steepest descent method for the first 7,500 steps followed by the conjugate gradient
method for another 7,500 steps. Second, the protein was minimized with restraints on the
solvent for 5,000 steps, using the steepest descent method for the first 2,500 steps
followed by the conjugate gradient method for another 2,500 steps. Third, the whole
system was minimized without restraints, using the steepest descent method for the first
15,000 steps followed by the conjugate gradient method for another 15,000 steps. During
the minimization, nonbonded interactions were calculated using the nonbonded pair list
cutoff at 12.0 Å, and the nonbonded pair list was updated every 25 steps.
The molecular system was heated up gradually from 0 to 300 K at 50 K intervals
for a total of 240,000 steps using a 1 fs time step. Temperature was regulated by
-1

Langevin dynamics (24) with ! = 1.0 ps . Another 40,000 steps were run at 300 K before
the simulation was switched to an NPT (atom number (N), pressure (P) and temperature
(T)) ensemble. During the heating process, nonbonded interactions were calculated using
the nonbonded pair list cutoff at 12.0 Å, and the nonbonded pair list was updated every
25 steps.

98

Equilibration simulation was performed using 1 fs as the time step. The SHAKE
algorithm (25) was turned on for all bonds involving hydrogen. The pressure was set at 1
atm and the temperature was controlled at 300 K by Langevin dynamics (24) with ! = 1.0
-1

ps . Nonbonded interactions were calculated using the nonbonded pair list cutoff at 12.0
Å, and the nonbonded pair list was updated every 25 steps. Production simulations were
performed under the same condition as the equilibration simulation. All of the MD results
were analyzed with the Ptraj module of Amber 10 (23).
QM/MM MD Simulations. The self-consistent charge density functional tightbinding (SCC-DFTB) method is a semi-empirical method developed by Elstner et al (20,
26), which has been applied on a number of enzymatic systems (7, 8, 27-29). In this
study, QM/MM MD simulations were performed for yCD using the SCC-DFTB method
implemented in the amber molecular dynamics package (8). The molecular systems for
the yCD-DHP complex, the yCD free form complex and the yCD water/cytosine
complex (1) were built as described in the previous section. The active site of subunit A
was modeled in the QM region (Figure 4-1), including the side chains of His62, Glu64,
Cys91, and Cys94 and the catalytic zinc atom, and the ligands for the respective
complexes. The rest of the protein was modeled in the MM region. SCC-DFTB and
PARM99SB force field were used for the QM and MM regions, respectively. The
SHAKE algorithm was turned on for all bonds except those in the QM region to allow
proton transfer.
The molecular systems were minimized and heated up to 300 K using the same
procedure as in the conventional MD simulation, expect that the residues in the QM
region were calculated by the SCC-DFTB method. The structures after being heated up to
300 K were used as the starting structures for the equilibration and production MD
99

simulations. As an alternative, representative structures from the conventional production
MD simulations were tested as the starting structures. The results were similar.
During the equilibration and production MD simulations, temperature was
-1

regulated by Langevin dynamics with ! = 1.0 ps . Tight convergence on density matrix
in the SCF was used. The electrostatic cutoff was set at 12.0 Å for QM/MM electrostatic
interaction. Long-range QM-QM and QM-MM electrostatic interactions were calculated
by Ewald sum (28, 30) and a QM compatible PME approach (7). The time step is 1 fs.
MD simulations were carried out using the parallel version Sander module in Amber 10.
All of the MD results were analyzed with the Ptraj module of Amber 10.
The free energy profile along a reaction coordinate was obtained from the
potential of mean force (PMF) by adopting an umbrella sampling technique (29). Proper
amount of windows were used to sample enough space along the reaction coordinate. The
force constants were set in the range of 100 to 600 kcal/mol for simulations during a bond
2

formation and the spring constants were set at 0.02 kcal/mol•deg for simulations during
the rotation of the side chain of Glu64. The data from the simulations were analyzed by
the weighted histogram analysis method (WHAM) (31, 32) to compute the PMF. Data
from a 200 ps trajectory were collected after 200 ps of equilibration for each window.

RESULTS AND DISCUSSION
Two crystal structures of the yCD-DHP complex have been solved (3, 4) at high
resolution. These structures not only serve as great starting conformations for the MD
simulation, but also can be used to estimate how well the QM/MM method performs
when compared with the MD simulations.

100

QM/MM MD Simulation on the yCD-DHP Complex. In the conventional MD
simulation on yCD, the residues coordinated with Zn were constrained by explicit bonds
as described elsewhere (22). The zinc coordination stayed stable and matched well with
the geometry of the crystal structure. However, the chemical change cannot be captured
due to the inherited nature of this methodology. A QM/MM MD simulation on the yCD!1

"

DHP complex was carried out instead. Since the residues (His62 N , Cys91 S , Cys94
"

S and DHP-O4) coordinated with zinc were included in the QM core, it is not necessary
to constrain the bond distances by explicit forces. Moreover, chemical changes, such as
proton transfer, bond formation and bond cleavage, can be studied directly.
A 1.3 ns simulation trajectory was collected and analyzed. Figure 4-2 shows the
time evolution of the C# atom root mean squared deviation (RMSD) of the snapshots
along the trajectory relative to the initial crystal structure. The average RMSDs for
subunit A and subunit B are 0.54 and 0.58 Å, respectively, indicating that the structure is
very rigid during the simulation course. The first 0.5 ns was regarded as the equilibrium
stage and discarded and key internal coordinates are summarized in Table 4-1 based on
the last 0.8 ns trajectory of the simulation. The structural similarity is shown in Figure 43 by superimposing the average structure of the snapshots in the MD trajectory with the
crystal structure.
The results show the zinc coordination was described very well. The average
"

bond distance between Zn and S of Cys91 is 2.37 Å; the average bond distance between
!1

"

Zn and S of Cys94 is 2.32 Å; the average bond distance between Zn and N

of His62 is

2.00 Å; the average bond distance between Zn and O4 of DHP is 2.12 Å. The bond
101

lengths are comparable with those in the crystal structure and previous ONIOM
calculations in Chapter 3 (Table 4-1).
The majority of the inhibitor in solution is in the unhydrated form 2Py. Only less
than 0.05% of the inhibitor is in the covalent hydrate form DHP (33). The crystal
structure suggests that the inhibitor in the active site of yCD is DHP. The simulation
captures the tetrahedral feature of the inhibitor in which the C4 atom shows obvious sp3
hybridization characteristics. The dihedral angles !O4-C4-N3-C2 and !O4-C4-C5-C6
are 120° and -117°, respectively. The bond distances are also in close agreement with
those in the crystal structure (Table 4-1). The O4-C4 bond is a single bond with a bond
distance at 1.47 Å and the N3-C4 bond is a single bond too, with a bond distance at 1.45
Å. When the inhibitor is in its unhydrated form, the N3-C4 bond is clearly a double bond
with a bond distance at 1.32 Å and C4 adopts a sp2 hybridization.
DHP is stabilized by a number of hydrogen bonds. N1 is hydrogen bonded with
the carboxyl group of Asp155 with a distance of 2.80 Å. O2 is hydrogen bonded with the
side chain amide group of Asn51 and the backbone amide group of Gly63 with distances
of 2.91 and 2.86 Å, respectively. DHP forms two hydrogen bonds with the carboxyl
group of Glu64, through O4 and N3, with a distance of 2.48 and 2.88 Å, respectively. O4
is also hydrogen bonded with the backbone amide of Cys91 with distances of 3.14 Å.
Other than the hydrogen bonds, hydrophobic interactions contribute to the stabilization
too. DHP is surrounded with a cluster of nonpolar residues, such as Ile33, Leu88,
Phe114, Trp152, and Ile156. The imidazole ring of His62 interacts with the pyrimidine
ring of DHP through a "-stacking interaction. The docking surface is well defined for
DHP and is rigidified by hydrogen bonds and hydrophobic packing. For example,

102

hydrogen bonds are formed between the amide group of Asn51 the carboxyl group of
!2

Asp155 and between N

of His62 and the backbone oxygen of Asp155.
!1

Characteristic of the O4…O
!1

of DHP and O

Hydrogen Bond. Hydrogen bonds between O4
!2

of Glu64, and between N3 of DHP and O

of Glu64 may be

particularly important for the stabilization of DHP because these two hydrogen bonds are
not present in the yCD-2Py complex. The former is a hydrogen bond with a very short
bond distance at 2.48 Å. The hydrogen does not exclusively form a covalent bond with
either heteroatom. It is located in the middle of the two heteroatoms, with an average
!1

distance between HO4 and O4 at 1.31 Å and between HO4 and O

at 1.17 Å in the 0.8

ns trajectory (Table 4-1 and Figure 4-4). The hydrogen bond may have only one potential
energy well or two wells separated by a low energy barrier.
A 2D umbrella sampling technique was employed to illustrate the potential of
mean force for the proton transfer between the two heavy atoms. The reaction coordinate
!1

is based on the bond distance between HO4 and O4 of DHP and between HO4 and O of
Glu64 (Figure 4-5). The results show that the energy profile indeed has a very flat well
where the hydrogen atom tends to sit in the middle between the two heteroatoms. There is
no obvious energy barrier. This calculation suggests that the hydrogen bond is very likely
a low barrier hydrogen bond (LBHB) (34).
!1

In comparison, a simulation on the yCD-2Py complex was performed. O

of

Glu64 forms a hydrogen bond with the zinc-coordinated water with a distance of 2.60 Å.
The distance is rather short too when compared with the average O…O type hydrogen
103

WAT

bond. However, the hydrogen in the hydrogen bond (H
WAT

the oxygen of the zinc-coordinated water (O
!2

O

) is covalently bonded with
WAT

) and the distance between H

and

of Glu64 is 1.8 Å. The energy profile for the proton transfer from the zinc!1

coordinated water to O

of Glu64 was obtained using the same protocol (Figure 4-5,
WAT

bottom). The reaction coordinate is set based on the distance between H
WAT

!1

and between H

and O
WAT

more stable when H

WAT

and O

of Glu64. The energy profile shows that the system is much
WAT

stays with O

. The proton transfer is not a preferred process
!1

probably due to the pKa mismatch between O

of Glu64 and the zinc-coordinated water.

The pKa of the zinc-coordinated water is 8-9 (27), while the pKa of Glu64 is normally
between 2 and 5.5 in protein (35). Preferably, the hydrogen will stay with the water
instead of transferring to Glu64. However, the pKa of the water would probably be
reduced to somewhere close to that of Glu64 when it is incorporated into the inhibitor.
1

The strong hydrogen bond in the yCD DHP complex was inferred by a H jump return
NMR experiment in which a proton peak at 16.2 ppm was observed on the NMR
spectrum (unpublished data).
!1

In Chapter 3, ONIOM calculations also suggested that this O4…O

hydrogen

bond is an LBHB. In the ONIOM calculation, the most essential elements were treated at
the semiempirical DFT method B3LYP with the 6-31G* basis set. B3LYP is a higher

104

energy level than DFTB used in this study. The ability to capture this important feature
also indicates the validation of this simulation protocol.
!1

The same O4···O

hydrogen bond in the yCD-DHP complex was investigated

recently by QM/MM MD simulations (19), using SCC-DFTB for QM and the CHARMM
!1

force field for MM. The authors suggest the proton is completely transferred to O

and

this proton transfer may play an important role in stabilizing the inhibitor. In the
calculation of the simple model in that study (19), the hydrogen is covalently bonded to
!2

!2

O , but the bond length is 1.0 Å and the distance between O

and O4 is 2.7 Å,

characteristic of a conventional H-bond. It is noted, however, that the zinc coordination
was not reproduced well because the distances of three out of the four ligands to the zinc
are significantly different from those measured by X-ray diffraction.
There have been debates (36-40) about whether the LBHB exists or not, and if it
does, what is its function. By using high-resolution neutron crystallography, a LBHB was
indeed discovered by Yamaguchi et al (41) in photoactive yellow protein (PYP) between
the chromophore of PYP, p-coumaric acid (pCA), and the carboxyl oxygen of Glu46. The
bond distance is 2.56 Å. The proton sits in the middle between the two heavy atoms, with
distances of 1.37 Å to pCA and 1.21 Å to Glu46. It was suggested to stabilize the buried
charge in the hydrophobic environments.
!1

For the O4…O

hydrogen bond in the yCD-DHP complex, the LBHB may play

an important role in the stabilization of the inhibitor. Further experiments are needed to
investigate the nature of this hydrogen bond and its function in catalysis.

105

Substrate Activation. To investigate the deamination of cytosine by yCD, a
series of QM/MM MD simulations were performed starting from the yCD water/cytosine
complex (1). 1 was constructed by replacing DHP with a cytosine and a zinc-coordinated
water molecule in the active site of yCD (Figure 4-6 and 4-7). Cytosine is placed in the
same orientation as DHP.
A 0.8 ns QM/MM MD simulation trajectory after equilibration (0.5 ns) was
collected and analyzed. Figure 4-8 shows the time evolution of the C! atom RMSDs of
the snapshots along the trajectory relative to the initial crystal structure. The average
RMSDs for subunit A and subunit B are 0.65 and 0.54 Å, respectively, indicating that the
structure is quite rigid during the simulation course.
Key internal coordinates are summarized in Table 4-1. Similarly, cytosine is
stabilized by a set of hydrogen bonds and hydrophobic interactions (Table 4-1). Two
additional hydrogen bonds were formed because of the presence of the amide group (N4).
One is between N4 and the main chain carbonyl of Ser89 and the other is between N4
"2

and O

"1

of Glu64. O

of Glu64 forms a hydrogen bond with the zinc-coordinated water,

with an average distance of 2.58 ± 0.10 Å.
Glu64 is believed to serve as a general base/acid during multiple steps. The side
chain rotation is critical to interact with different hydrogen donor/acceptor species during
different stages. The information on side chain rotation is extracted from the MD
trajectory (Figure 4-9). Moreover, we performed a simulation on the free form yCD. The
$

%

&

"2

dihedral angle of the side chain (defined as (#C -C -C -O ) was analyzed and
compared with that in the yCD cytosine complex.

106

For the free form yCD, the 180° rotation of the side chain of Glu64 was observed
in the 0.8 ns trajectory (Figure 4-9, C). The other rotation mode is between two states,
"

#

$

%2

between which the difference of the side chain dihedral angle (!C -C -C -O ) is 50°
(Figure 4-9, C). The hydrogen bond between the zinc-coordinated water and the carboxyl
group of Glu64 can switch rapidly from one carboxyl oxygen of Glu64 to the other
(Figure 4-9, A). Rotation of Glu64 side chain was investigated by the umbrella sampling
technique. The energy profile is plotted along the reaction coordinate based on the
"

#

$

%2

dihedral angel (!C -C -C -O ) (Figure 4-9, D). Two shallow wells were found at -20
and -70°, indicating the rotation is indeed facile. The 180° rotation costs about 2.0
kcal/mol, again suggesting that the rotation of the side chain of Glu64 is a common event
for the free form of yCD on the MD time scale.
After the binding of cytosine, the side chain of Glu64 adopts a slightly different
"

#

$

%2

configuration, with the dihedral angle (!C -C -C -O ) oscillating around -120° (Figure
4-9). Furthermore, the presence of cytosine obscures the 180° rotation of the side chain of
Glu64, suggested by the PMF, in which the activation energy is increased to 4.5 kcal/mol.
"

#

Two potential energy wells were found, corresponding to two stable states with (!C -C $

%2

C -O ) at -100° and -150°, respectively. However, the rotation did not break the
hydrogen bond between the carboxyl oxygen and the zinc-coordinated water (Figure 4-9,
B).
The other carboxyl oxygen of Glu64 forms a hydrogen bond with N4 of cytosine
and the hydrogen bond pulls cytosine closer to the zinc-coordinated water. The distance
107

Zn

between the water oxygen (O ) and C4 of cytosine is 2.90 Å. In the complex with 2Py,
the distance is 3.07 Å. However, the water is not likely to attack C4 directly. The water
needs to be activated by losing a proton to the carboxyl group of Glu64, which, in turn,
activates cytosine by transferring a proton to cytosine at N3.
This process was investigated by a 2D umbrella sampling calculation along two
reaction coordinates. One is the distance between HO1 (hydrogen of the zinc-coordinate
!2

water) and N3 and the other is the distance between HW and O of Glu64 (Figure 4-10).
!2

First, the proton (HO1) transfers to O

to form the yCD hydroxide/Glu64H cytosine

complex (2). But there is no noticeable energy barrier. 2 is ~ 5 kcal/mol less stable than
complex 1. It suggests that although the water is coordinated with the zinc atom, its pKa
is still not low enough to match that of Glu64. The energy surface around 2 is very flat.
!2

Zn

When the hydrogen bond between O
!2

between O

and O

breaks, a new hydrogen bond is formed

and N3, which compensates for the energy loss of the former hydrogen

bond. 2 in this form is ~ 8 kcal/mol less stable than 1. The system is stabilized after the
proton (HO1) transfers to N3 to form the complex 3. The transition state (TS2) is located
!2

when the proton is in the middle of O

and N3. The activation energy for the overall

reaction from 1 to 3 is about 14.9 kcal/mol. 3 is less stable than 1 by 10.9 kcal/mol. In 3,
the energy level is elevated because two charged groups are generated, one is the zinccoordinated hydroxide and the other is the positive charged cytosine (Cytosine-H).
However, this special configuration and charge distribution facilitate the nucleophilic
attack in the next step. The addition of the proton atom on N3 draws elections from C4.
108

Zn

The distance between O

and C4 decreases to 2.10 Å, from 2.45 Å in 2 and from 2.90 Å
Zn

in 1. With the shortened distance between O

Zn

and C4, O

is well positioned to attack

C4.
Nucleophilic Attack. The zinc-coordinated hydroxide is well positioned for the
nucleophilic attack on C4. The energy profile was obtained by umbrella sampling along
Zn

the reaction coordinate by shortening the distance between O

and C4 (Figure 4-11).

Zn

The transition state was found when the O -C4 distance is around 1.82 Å. The energy
barrier is only about 1.4 kcal/mol and the tetrahedral intermediate (4) is slightly more
stable than the hydroxide/Glu64 Cytosine-H complex (3) by 1.0 kcal/mol.
Zn

With the decrease of the distance between O
"

#

$

and C4, the side chain of Glu64

%2

(!C -C -C -O ) rotates from 128° to 104°, leading to the formation of a hydrogen bond
%1

between O

Zn

%1

of Glu64 and O . At the same time, the hydrogen bond between O

%1

Glu64 and N4 of cytosine is broken. The newly formed hydrogen bond between O
Zn

O

of

and

is quite strong, with a distance of 2.50 Å. It is comparable with that found in the

yCD-DHP complex, which is believed to be important for the stabilization of the
inhibitor (also see Chapter 3). Proton transfer within this hydrogen bond was investigated
by umbrella sampling. The energy profile shows that the energy barrier is negligible.
The average structure of the snapshots of 4 in the MD simulation trajectory is
superimposed onto that of the yCD-DHP complex, shown in Figure 4-12, from which, we
can see that the structures align very well. All hydrogen bonds between DHP and yCD

109

are preserved for the intermediate and yCD. This may be the origin where the tightbinding of DHP comes from.
C4-N4 Bond Cleavage. From 3 to 4, the hybridization state of C4 changes from
sp2 to sp3. The bond distance between N4 and C4 is increased to 1.44 Å, from 1.37 Å in
3. 4 is stabilized by a network of hydrogen bonds and hydrophobic interactions. To
investigate how the C4-N4 bond is cleaved, first, a 1D umbrella sampling calculation was
carried out along the reaction coordinate based on the distance between C4 and N4. The
results show that with the increase of the C4-N4 distance, the hydrogen of the hydroxyl
!1

group (HO4) transfers to O

of Glu64. The transition state was found with a distance

between C4 and N4 at 1.91 Å. The energy barrier is 16.9 kcal/mol. This is quite
consistent with the transition state obtained from the KIE studies. However, whether the
proton transfer to N4 and the breakage of C4-N4 bond occur in a concerted or stepwise
manner is not clear from this energy profile.
Hence, a 2D umbrella sampling calculation was carried out along two reaction
coordinates. One is the distance between HO4 and N4 and the other is the distance
between C4 and N4. The energy profile is plotted in Figure 4-13. HO4 is first transferred
!1

to O

of Glu64 by crossing over a low energy barrier. We have already seen this kind of

hydrogen bond in the yCD-DHP complex. This strong hydrogen bond may be important
for the stabilization of the tetrahedral intermediate and hence increase the catalytic
efficiency (kcat/Km). With the increase of the distance of the C4-N4 bond, HO4 prefers to
!1

Zn

stay with O , instead of O , and forms a hydrogen bond with N4. It turns out that the
proton needs to transfer to N4 before the C4-N4 bond completely breaks. The transition

110

state is found for the proton transfer when C4-N4 bond distance is increased to 1.65 Å
and the energy barrier is 19.7 kcal/mol. It is also the overall activation energy for the C4N4 bond cleavage. After the proton transfers to N4, the C4-N4 bond keeps increasing and
the ammonia is formed. The remaining part of the intermediate forms the other product
uracil, which is still covalently bonded with the zinc atom.
!2

A surprising feature is that H3 is transferred to O

of Glu64 when the C4-N4

bond distance is increased to 1.95 Å. The protonated side chain of Glu64 forms a
Zn

hydrogen bond with O

as a proton donor. The proton transfer was investigated by

umbrella sampling calculations along the reaction coordinate, in which the N3-H3 bond
distance is gradually increased. Ammonia is believed to diffuse out of the active site
relatively quickly. Instead, a water molecule enters the active site and binds at the
ammonia binding site. Hence, a new molecular system is set up, in which the ammonia is
!2

replaced by a water molecule, to investigate the proton transfer from N3 to O of Glu64.
The results show the activation energy is quite low, with a value of 2.4 kcal/mol (Figure
4-14). Furthermore, the deprotonated uracil with its cognate enzyme is ~ 5.0 kcal/mol
more stable than the protonated uracil complex.
The pKa of N3 of uracil in solution is 9.5 (42). Binding with Zn at the O4 position
can lower the pKa of N3 to match that of the carboxyl group of Glu64. Hence the proton
!2

can transfer to O . Once the proton is transferred, the hydrogen bond between Glu64 and
N3 is broken, which may facilitate the release of uracil.

111

O

Zn

-Zn Bond Cleavage. Uracil is covalently bonded with the zinc atom after the
Zn

Zn

C4-N4 bond cleavage. It cannot diffuse out unless the O -Zn bond is cleaved. The O Zn bond cleavage has been proposed by previous ONIOM calculations. A two-stepped
Zn

ONIOM-MD calculation developed by Yao et al (15) was also applied on the O -Zn
bond breakage. The results show that the activation energy for the bond breakage is only
2.9 kcal/mol, indicating product release is not the rate-limiting step for the overall
reaction. Fixing atoms in the outer layer in the first ONIOM calculation causes unrealistic
effects, which were minimized by incorporating configurational sampling in the MD
scheme (15). Matsubara and co-workers developed a method (16) to directly couple the
ONIOM method with the MD method and applied this method on the process of the zincZn

uracil bond breakage (17). The results show that the O -Zn bond abruptly breaks at 11
ps during the simulation. The authors suggest that the thermal motion of Ile33, which
sandwiches uracil with His62, perturbs the dynamics of uracil through steric contact and
contributes to the bond breakage.
In this study, a pathway was proposed, in which a solvent water attacks the zinc
atom and releases uracil (Figure 4-15). It has been attempted by ONIOM calculations in
Chapter 3, but the right path was not sampled. A series of MD simulations were carried
out to push the water towards the zinc atom and replace uracil and the PMF was obtained
(Figure 4-16). The transition state was located when the distance between water and zinc
Zn

was at 2.9 Å. The energy barrier is 8.1 kcal/mol. The O

atom is expected to be quickly

released from Zn afterwards. The zinc atom forms a new coordination with the water.
The uracil would now be ready to diffuse out of the active site.

112

!1

Proton transfer from N3 to O

Zn

of Glu64 plays an important role in the O -Zn
!1

bond cleavage. The hydrogen bond between N3 an O

is disrupted because the side
Zn

chain of Glu64 rotates away to form a new hydrogen bond with O

or the water

molecule. Hence the zinc-coordinated uracil is not as rigidified as the intermediates
occurring in the prior steps. When the water moves close to Zn, the pyrimidine ring can
shift away a bit to decrease the steric hindrance. Once water moves into the cavity close
!1

to Zn, the O -H group swings back to form a hydrogen bond with N3 and eventually
Zn

transfers the proton back to N3. This also facilitates the release of O

from Zn.

The overall energy profile for the enzyme reaction is plotted in Figure 4-17. Even
with the help of Glu64, the proton transfer from the zinc-coordinated water to N3 of
cytosine needs a large activation energy to cross over the transition state. Once the
complex 3 is formed, it can be easily converted to the tetrahedral intermediate (4). 4 is
stabilized by a hydrogen bond network and hydrophobic interactions. The hydrogen bond
!1

between O4 and O

of Glu64 may be particularly important. For the C4-N4 bond

cleavage, the activation energy is very high. It is the overall rate-limiting step. The
energy barrier is 19.7 kcal/mol. The C4-O4 bond is cleaved during the water attacks Zn.
The energy barrier is only 8.1 kcal/mol, which is much lower than that in the C4-N4 bond
cleavage step.

CONCLUDING REMARKS
In this study, a series of QM/MM MD simulations were performed to explore the

113

catalytic mechanism of the deamination of cytosine by yCD. Different from the ONIOM
calculations in Chapter 3, dynamic effects were taken into consideration in the MD
simulation. The essential elements, composed of the side chains of Glu64, His62, Cys91
and Cys94, the catalytic zinc and the ligands, were included in the QM part and simulated
by the fast semiempirical density-functional approach SCC-DFTB. The remainder of the
protein and the solvent were simulated by the PARM99SB force field. The method was
applied on the yCD-DHP complex to validate the methodology on the yCD system. The
key coordinates of the resulting structures in the equilibrium ensemble are in close
agreement with those of the crystal structure, especially those for the catalytic zinc
apparatus. The results also suggest that the hydrogen bond between the carboxyl group of
Glu64 and O4 of DHP is possibly a low barrier hydrogen bond, which may play an
important role in the stabilization of the tetrahedral intermediate.
The reaction pathway of the deamination of cytosine was proposed. First, the
substrate cytosine is activated by abstracting one proton from the zinc-coordinated water
!1

via O

of Glu64. The energy is elevated due to charge redistribution and geometry

change. However, the zinc-coordinated hydroxide is well positioned for the nucleophilic
attack on C4. Second, the tetrahedral intermediate is formed after the nucleophilic attack.
The energy barrier is quite low, with a value of 1.4 kcal/mol. The intermediate is
stabilized by a hydrogen bond network and hydrophobic interactions. The hydrogen bond
!1

between O

Zn

of Glu64 and O

is very strong, with a distance of 2.51Å, which is quite

similar to the hydrogen bond in the yCD-DHP complex. This hydrogen bond is likely to
play an important role in stabilizing the intermediate. Third, the C4-N4 bond is cleaved.
Several proton transfer events were observed during this process. HO4 first transfers

114

Zn

from O

!1

to O

of Glu64. With the increase of the distance between C4 and N4, HO4
!2

transfers to N4 to form the product ammonia. H3 transfers to O

of Glu64 when the

distance C4-N4 increases to 1.95 Å, probably because the pKa of N3 is lowered due to
Zn

the interaction between Zn and O . The transition state for the C4-N4 step is found
when the C4-N4 distance is about 1.65 Å, where HO4 already transferred to N4 and H3
still bonds with N3. It is also the overall transition state, which is believed to be the ratelimiting step with an energy barrier of 19.7 kcal/mol. The product ammonia can easily
diffuse into the solution. The other product uracil is covalently bonded with the zinc
atom. We propose that a solvent water molecule enters the active site and replaces uracil
in coordination with Zn. The energy barrier is about 8.1 kcal/mol, which is lower than
that for the C4-N4 bond cleavage step.

115

TABLES AND FIGURES
Table 4-1. Comparison of changes of key coordinates during the conversion of cytosine
a
to uracil with crystal structures .
Internal Coordinate
X-ray
X-ray
yCDComplex Complex
Structure
Structure
DHP
1
4
b
(1p6o)
(1uaq)
1
!
2.07, 2.06 2.01, 2.00 2.00
2.00
2.00
His62-N
···Zn
···Zn

2.35, 2.31

2.29, 2.30

2.37

2.35

2.38

···Zn

2.27, 2.25

2.27, 2.27

2.32

2.30

2.32

O4
Cys91-N

···Zn
···O4

Glu64-N

···O -Glu64

#1

2.06, 2.08
3.02, 3.10
2.69, 2.70

2.05, 2.06
3.02, 3.03
2.69, 2.72

2.12
3.14
2.78

2.14
3.23
3.05

2.14
3.13
2.80

O4

···O -Glu64

#2

2.60, 2.45

2.49, 2.52

2.48

2.58

2.51

N3
O2

···O -Glu64
…N-Gly63

#1

2.82, 2.81

2.80, 2.78

2.88

3.64

2.87

O2

!2

…N -Asn51

2.84, 2.79
2.96, 2.95

2.89, 2.87
2.92, 2.90

2.86
2.91

2.91
2.82

2.83
2.93

!1
…O -Asp155 2.67, 2.70
!2
!2
Asn51-N …O -Asp155 2.95, 2.98

2.69, 2.68

2.80

2.81

2.81

2.92, 2.90

3.02

2.90

3.04

2.67, 2.70

2.89

2.86

2.89

3.48
2.94

3.41
3.24

2.90
1.34
1.45
1.37
1.40
1.36
1.37
1.27
91
94
99
-91
-120

1.51
1.46
1.50
1.37
1.38
1.38
1.35
1.27
106
108
119
-115
-78

Cys91-S
Cys94-S

"
"

c

N1

#2

His62-N
N4

…O-Asp155
…O-Ser89
#1

N4
…O -Glu64
O4
···C4
N3
···C4
C5
···C4
N3
···C2
N1
···C2
N1
···C6
C5
···C6
O2
···C2
$O4-C4-N3
$O4-C4-C5
$O4-C4-N3-C2
$O4-C4-C5-C6
%

"

2.74, 2.76

!

#2

1.44, 1.44
1.42, 1.41
1.45, 1.45
1.37, 1.37
1.39, 1.39
1.38, 1.38
1.34, 1.35
1.21, 1.21
110, 109
110, 109
113, 119
-125, -116
-65, -68

1.48, 1.48
1.46, 1.45
1.47, 1.46
1.31, 1.31
1.32, 1.33
1.43, 1.42
1.31, 1.33
1.30, 1.30
107, 108
112, 109
113, 114
-113,-114
-68, -68

1.47
1.45
1.50
1.36
1.38
1.38
1.35
1.28
108
110
120
-117
-72

$C -C -C -O (Glu64)
a: Distances between the paired atoms are in Å. Angles and dihedral angles are in
degrees. b: Distances in subunits A and B, respectively. c: O4 is the oxygen atom in the
water, or the hydroxide oxygen of DHP, cytosine or other intermediates. Atom names are
presented in the same manner for C1, C2, C4, C5, C6, N1, N3, N4 and O2.
116

Figure 4-1. QM partition in the simulation systems for the yCD-DHP complex (top) and
with the yCD water/cytosine complex (bottom).

117

Figure 4-2. RMSDs of the backbone C! atoms relative to the crystal structure during the
QM/MM MD simulation for the yCD-DHP complex. Black curve is for subunit A and
red curve is for subunit B. The first 500 ps was regarded as the equilibrium stage. The
data analysis was based on the last 800 ps trajectory.

118

Figure 4-3. Superimposition of a representative snapshot during the yCD-DHP complex
simulation with the crystal structure. The backbone was displayed as a red tube (crystal
structure) and gray tube (MD). DHP and several surrounding residues were displayed as
ball and stick type.

119

Figure 4-4. Bond distances of snapshots along the MD trajectory of the yCD-DHP
!2

complex simulation. (A) The hydrogen bond distance between Glu64-O
!2

(B) The bond distance between Glu64-O
DHP-O4 and HO4.

and DHP-O4.

and HO4. (C) The bond distance between

120

WAT

Figure 4-5. Energy profiles for proton transfer. (top) H
!2

coordinated water to O

!2

from O4 of DHP to O

is transferred from the zinc-

of Glu64 in the free from yCD. (bottom) HO4 is transferred

of Glu64 in the yCD-DHP complex.
121

Figure 4-6. Proposed reaction pathway for the conversion of cytosine to uracil in the
QM/MM MD simulations. (1) yCD water/cytosine complex. (2) yCD hydroxide/Glu64H
cytosine complex. (3) yCD hydroxide/Glu64 cytosine-H complex. (4) yCD protonated
tetrahedral intermediate complex. (5) yCD deprotonated tetrahedral intermediate
complex. (6’) yCD zinc bound uracil/ammonia complex. (6) yCD Glu64H deprotonated
zinc bound uracil/ammonia complex.

122

Figure 4-7. Snapshots of the reaction species along the reaction pathway for the
conversion of cytosine to uracil in the QM/MM MD simulations.

123

Figure 4-8. RMSDs of the backbone C! atoms relative to the crystal structure during the
QM/MM MD simulation for the yCD water/cytosine complex (1). Black curve is for
subunit A and red curve is for subunit B. The first 500 ps was regarded as the equilibrium
stage. The data analysis was based on the last 1000 ps trajectory.

124

"

#

%2

$

Figure 4-9. Rotation of the sidechain of Glu64 (!C -C -C -O ) in the free form of yCD
%1

and the yCD substrate complex simulations. (A) Distance between O
Zn

%2

or O

of Glu64

and the zinc-coordinated water (O ) in the yCD free form simulation. The curve for the
%1

distance between O

Zn

and O
%1

%2

is colored blue and between O
%2

Zn

and O

is colored green.
Zn

(A) Distance between O or O of Glu64 and the zinc-coordinated water (O ) in yCD
substrate form simulation. The curves are colored as the same as in (A). (C) Rotation of
"

#

$

%2

the side chain of Glu64 (!C -C -C -O ). The curve is colored black for the free form
simulation and colored red for the yCD substrate complex. (D) PMF for rotation of the
"

#

$

%2

side chain of Glu64 (!C -C -C -O ). The curves are colored as the same as in (C).

125

Figure 4-10. Potential of mean force for the proton transfer from the zinc-coordinated
!2

water to O

of Glu64 and to N3 of cytosine.

126

Figure 4-11. Potential of mean force for nucleophilic attack on C4 of cytosine. The
tetrahedral intermediate complex (4) is 1.0 kcal/mol more stable than the hydroxide/
Glu64 cytosine-H complex (3). The energy barrier is calculated to be 1.4 kcal/mol.

127

Figure 4-12. Superposition of the average structure of snapshots in the MD trajectory for
the complex 4 simulation onto that of the yCD-DHP simulation.

128

Figure 4-13. Potential of mean force for the C4-N4 bond cleavage.

129

!2

Figure 4-14. Potential of mean force for proton transfer from N3 to O . The transition
state is found when C3-H3 distance is 1.3 Å. The energy barrier is 2.4 kcal/mol.

130

Zn

Figure 4-15. Reaction pathway for the O -Zn bond breakage.

131

Zn

Figure 4-16. Potential of mean force for the O -Zn bond breakage. The transition state
Zn
is found when O -Zn distance is 2.9 Å. The energy barrier is 8.1 kcal/mol.

132

Figure 4-17. Overall energy profile for the conversion of cytosine to uracil by yCD. The
overall rate-limiting step is found to be the C4-N4 bond cleavage step, coupled with
multiple proton transfers.

133

REFERENCES

134

REFERENCES
1. Greco, O., and Dachs, G. U. (2001) Gene directed enzyme/prodrug therapy of cancer:
Historical appraisal and future prospectives, J Cell Physiol 187, 22-36.
2. Aghi, M., Hochberg, F., and Breakefield, X. (2000) Prodrug activation enzymes in
cancer gene therapy, J Gene Med 2, 148-164.
3. Ko, T., Lin, J., Hu, C., Hsu, Y., Wang, A., and Liaw, S. (2003) Crystal structure of
yeast cytosine deaminase - Insights into enzyme mechanism and evolution, J Biol
Chem 278, 19111-19117.
4. Ireton, G., Black, M., and Stoddard, B. (2003) The 1.14 angstrom crystal structure of
yeast cytosine deaminase: Evolution of nucleotide salvage enzymes and implications
for genetic chemotherapy, Structure 11, 961-972.
5. Sklenak, S., Yao, L., Cukier, R., and Yan, H. (2004) Catalytic mechanism of yeast
cytosine deaminase: An ONIOM computational study, J Am Chem Soc 126, 1487914889.
6. Torrent, M., Vreven, T., Musaev, D. G., Morokuma, K., Farkas, O., and Schlegel, H.
B. (2002) Effects of the protein environment on the structure and energetics of active
sites of metalloenzymes. ONIOM study of methane monooxygenase and
ribonucleotide reductase, J Am Chem Soc 124, 192-193.
7. Walker, R. C., Crowley, M. F., and Case, D. A. (2008) The implementation of a fast
and accurate QM/MM potential method in Amber, J Comput Chem 29, 1019-1031.
8. Seabra, G. D. M., Walker, R. C., Elstner, M., Case, D. A., and Roitberg, A. E. (2007)
Implementation of the SCC-DFTB method for hybrid QM/MM simulations within
the Amber molecular dynamics package, J Phys Chem A 111, 5655-5664.
9. Hu, H., Lu, Z., and Yang, W. (2007) QM/MM minimum free energy path:
Methodology and application to triosephosphate isomerase, J Chem Theory Comput
3, 390-406.
10. Okamoto, T., Yamada, K., Koyano, Y., Asada, T., Koga, N., and Nagaoka, M. (2011)
A minimal implementation of the AMBER-GAUSSIAN interface for ab initio
QM/MM-MD simulation, J Comput Chem 32, 932-942.
11. Torras, J., Seabra, G. D. M., Deumens, E., Trickey, S. B., and Roitberg, A. E. (2008)
A versatile AMBER-Gaussian QM/MM interface through PUPIL, J Comput Chem
29, 1564-1573.
12. Woodcock, H. L., Hodoscek, M., Gilbert, A. T. B., Gill, P. M. W., Schaefer, H. F.,
and Brooks, B. R. (2007) Interfacing Q-Chem and CHARMM to perform QM/MM
reaction path calculations, J Comput Chem 28, 1485-1502.
13. Riccardi, D., König, P., Prat-Resina, X., Yu, H., Elstner, M., Frauenheim, T., and
Cui, Q. (2006) “Proton Holes” in long-range proton transfer reactions in solution and
enzymes: A theoretical analysis, J Am Chem Soc 128, 16302-16311.
135

14. Gao, J., and Truhlar, D. G. (2002) Quantum mechanical methods for enzyme
kinetics, Annu Rev Phys Chem 53, 467-505.
15. Yao, L., Yan, H., and Cukier, R. (2006) A combined ONIOM quantum chemicalmolecular dynamics study of zinc-uracil bond breaking in yeast cytosine deaminase,
J Phys Chem B 110, 26320-26326.
16. Matsubara, T., Dupuis, M., and Aida, M. (2007) The ONIOM molecular dynamics
method for biochemical applications: Cytidine deaminase, Chem Phys Lett 437, 138142.
17. Matsubara, T., Dupuis, M., and Aida, M. (2008) An insight into the environmental
effects of the pocket of the active site of the enzyme. Ab initio ONIOM-molecular
dynamics (MD) study on cytosine deaminase, J Comput Chem 29, 458-465.
18. Brooks, B. R., Bruccoleri, R. E., Olafson, B. D., States, D. J., Swaminathan, S., and
Karplus, M. (1983) CHARMM: A program for macromolecular energy,
minimization, and dynamics calculations, J Comput Chem 4, 187-217.
19. Xu, Q., Guo, H., Gorin, A., and Guo, H. (2007) Stabilization of a transition-state
analogue at the active site of yeast cytosine deaminase: Importance of proton
transfers, J Phys Chem B 111, 6501-6506.
20. Elstner, M., Porezag, D., Jungnickel, G., Elsner, J., Haugk, M., Frauenheim, T.,
Suhai, S., and Seifert, G. (1998) Self-consistent-charge density-functional tightbinding method for simulations of complex materials properties, Phys Rev B 58,
7260-7268.
21. Hornak, V., Abel, R., Okur, A., Strockbine, B., Roitberg, A., and Simmerling, C.
(2006) Comparison of multiple Amber force fields and development of improved
protein backbone parameters, Proteins 65, 712-725.
22. Yao, L., Sklenak, S., Yan, H., and Cukier, R. (2005) A molecular dynamics
exploration of the catalytic mechanism of yeast cytosine deaminase, J Phys Chem B
109, 7500-7510.
23. Case, D. A., Darden, T. A., Cheatham, I., Simmerling, C. L., Wang, J., Duke, R. E.,
Luo, R., Crowley, M., Walker, R. C., Zhang, W., Merz, K. M., Wang, B., Hayik, S.,
Roitberg, A., Seabra, G., Kolossv¨¢ry, I., Wong, K. F., Paesani, F., Vanicek, J., Wu,
X., Brozell, S. R., Steinbrecher, T., Gohlke, H., Yang, L., Tan, C., Mongan, J.,
Hornak, V., Cui, G., Mathews, D. H., Seetin, M. G., Sagui, C., Babin, V., and
Kollman, P. A. (2008) AMBER 10. University of California, San Francisco.
24. Loncharich, R. J., Brooks, B. R., and Pastor, R. W. (1992) Langevin dynamics of
peptides: The frictional dependence of isomerization rates of N-acetylalanyl-N'methylamide, Biopolymers 32, 523-535.
25. Ryckaert, J., Ciccotti, G., and Berendsen, H. J. C. (1977) Numerical integration of the
cartesian equations of motion of a system with constraints: molecular dynamics of nalkanes, J Comput Phys 23, 327-341.

136

26. Elstner, M. (2005) The SCC-DFTB method and its application to biological systems,
Theor Chem Acc 116, 316-325.
27. Riccardi, D., and Cui, Q. (2007) pKa analysis for the zinc-bound water in human
carbonic anhydrase II: Benchmark for "multiscale" QM/MM simulations and
mechanistic implications, J Phys Chem A 111, 5703-5711.
28. Nam, K., Gao, J., and York, D. M. (2005) An efficient linear-scaling Ewald method
for long-range electrostatic interactions in combined QM/MM calculations, J Chem
Theory Comput 1, 2-13.
29. Cui, Q., Elstner, M., Kaxiras, E., Frauenheim, T., and Karplus, M. (2001) A QM/MM
implementation of the self-consistent charge density functional tight binding (SCCDFTB) method, J Phys Chem B 105, 569-585.
30. Darden, T., York, D., and Pedersen, L. (1993) Particle mesh Ewald: An N!log(N)
method for Ewald sums in large systems, J Chem Phys 98, 10089-10092.
31. Roux, B. (1995) The calculation of the potential of mean force using computer
simulations, Comput Phys Commun 91, 275-282.
32. Grossfield, A. WHAM: the weighted histogram analysis method, version 2.0,
http://membrane.urmc.rochester.edu/content/wham.
33. Tee, O. S., and Paventi, M. (1980) Kinetics and mechanism of the reaction of 2(1H)pyrimidinone with bromine. Reaction via its covalent hydrate, J Org Chem 45, 20722076.
34. Cleland, W., and Kreevoy, M. (1994) Low-barrier hydrogen bonds and enzymic
catalysis, Science 264, 1887-1890.
35. Fersht, A. (1999) Structure and mechanism in protein science: a guide to enzyme
catalysis and protein folding. W. H. Freeman and Company, New York.
36. Warshel, A., and Papazyan, A. (1996) Energy considerations show that low-barrier
hydrogen bonds do not offer a catalytic advantage over ordinary hydrogen bonds,
Proc Natl Acad Sci U S A 93, 13665-13670.
37. Perrin, C. L., and Nielson, J. B. (1997) "Strong" hydrogen bonds in chemistry and
biology, Annu Rev Phys Chem 48, 511-544.
38. Tuckerman, M. E., Marx, D., Klein, M. L., and Parrinello, M. (1997) On the quantum
nature of the shared proton in hydrogen bonds, Science 275, 817-820.
39. Cleland, W. W., Frey, P. A., and Gerlt, J. A. (1998) The low barrier hydrogen bond
in enzymatic catalysis, J Bio Chem 273, 25529 -25532.
40. Schutz, C. N., and Warshel, A. (2004) The low barrier hydrogen bond (LBHB)
proposal revisited: The case of the Asp ··· His pair in serine proteases, Proteins 55,
711-723.

137

41. Yamaguchi, S., Kamikubo, H., Kurihara, K., Kuroki, R., Niimura, N., Shimizu, N.,
Yamazaki, Y., and Kataoka, M. (2009) Low-barrier hydrogen bond in photoactive
yellow protein, Proc Natl Acad Sci U S A 106, 440 -444.
42. Kimura, E., Kitamura, H., Koike, T., and Shiro, M. (1997) Facile and selective
electrostatic stabilization of uracil N(1) anion by a proximate protonated amine:! A
chemical implication for why uracil N(1) is chosen for glycosylation site, J Am Chem
Soc 119, 10909-10919.

138

CHAPTER 5

ROLE OF GLUTAMATE 64 IN THE ACTIVATION OF
THE PRODRUG 5FC

INTRODUCTION
A major problem in cancer treatment is the toxicity of anticancer drugs to normal
cells. Gene directed enzyme prodrug therapy (GDEPT) is aimed to address this critical
problem by activating nontoxic prodrugs in tumors, and thereby killing cancer cells while
minimizing the side effects of anticancer drugs (1, 2). One of enzyme prodrug
combinations frequently used in GDEPT is cytosine deaminase (CD) and 5fluorocytosine (5FC). CD is a metalloenzyme that catalyzes the deamination of cytosine
to generate uracil as well as the deamination of 5FC to generate 5-fluorouracil (5FU).
5FU is an anticancer drug used for the treatment of colorectal, breast, stomach, and
pancreatic cancers, but, like many anticancer drugs, the drug is highly toxic, causing side
effects such as myelosuppression, mucositis, dermatitis, and diarrhea. 5FC is fairly nontoxic, as CD is absent in human. By activating the prodrug 5FC in the tumor, the
CD/5FC-based GDEPT minimizes the side effects of 5FU.
Two CDs have been used as the activating enzyme in CD/5FC-based GDEPT.
One is from E. coli (eCD), and the other from yeast (yCD). The two enzymes have no

139

detectable homology and arise apparently by convergent evolution. The E. coli enzyme is
a homohexameric protein with a molecular weight of 47.5 kDa per subunit (3), whereas
the yeast enzyme is a homodimeric protein with a molecular weight of 17.5 kDa per
subunit (4-6). Although eCD was used in early GDEPT studies, more recent studies
showed that yCD is preferable for GDEPT, because yCD has a much higher catalytic
efficiency for 5FC than eCD (7-9).
Because of its biomedical significance, there have been quite a few experimental
and computational studies of the structure and catalytic mechanism of yCD. The structure
of yCD has been determined by X-ray crystallography at high resolution, including the
substrate-free form (10) and the complex form with the inhibitor 2-pyrimidinone (Py)
(10, 11). The bound Py is in the hydrated form, 4(R)-hydroxyl-3,4-dihydropyrimidinone
(DHP), mimicking the transition state for the deamination of cytosine. The key elements
of the catalytic apparatus of yCD are believed to be the zinc ion coordinated with DHP
and Glu64 with two hydrogen bonds to DHP (Figure 1-3). The catalytic mechanism of
yCD has been investigated by both experimental and computational approaches (12-16).
Transient kinetic and NMR studies have shown that product release is rate-limiting in the
activation of 5FC by yCD (17). A complete reaction pathway for the deamination of
cytosine by yCD has been proposed by two-layered quantum mechanics calculations
using the ONIOM method (16) and combined molecular dynamics simulations and
ONIOM calculations (14). In the proposed reaction pathway, Glu64 plays a critical role
in catalysis, serving as a proton shuttle for both the generation of the tetrahedral reaction
intermediate from substrate cytosine and the conversion of the intermediate to the product

140

uracil. However, the role of Glu64 in the enzymatic reaction has not been investigated
experimentally.
In the present work, we have investigated the role of Glu64 in the activation of
5FC by combined experimental and computational studies. The results show that Glu64
is important not only for chemical transformation but also for substrate binding in the
activation of 5FC. The role of Glu64 in substrate binding is also of great importance for
pyrimidine salvage, the physiological role of yCD.

EXPERIMENTAL PROCEDURES
Materials. Cytosine, uracil, 5-fluorocytosine and 5-fluorouracil were purchased
3

from Sigma. [6- H]-5-fluorocytosine was purchased from Moravek. 5-fluoropyrimidinone was synthesized in our lab. Restriction enzymes and T4 ligase were
purchased from New England Biolabs. Pfu DNA polymerase and the pET-17b vector
were purchased from Strategene and Novagen, respectively.
Cloning and Site-Directed Mutagenesis. The pET-17bHR vector was
constructed by adding DNA encoding a His tag and a TEV protease cleavage site to the
pET-17b vector (pET17bHR). The expression construct of the His-tagged yCD with a
TEV protease cleavage site (pET17bHR-yCD) was made by PCR. The primers for PCR
were 5’-GGG ATC CAT ATG GCA AGC AAG TGG GAT CAG-3’ and 5’-GGA ATT
CTA CTC ACC AAT ATC TTC AAA CC-3’. The PCR product was digested with NdeI
and EcoRI restriction enzymes and ligated with the vector pET-17bHR digested with the
same restriction enzymes. The ligation mixture was transformed into E. coli strain DH5!.
The correct coding sequence of the cloned yCD gene was verified by DNA sequencing.
141

The DNA of the expression construct (pET17bHR-yCD) was then transformed into E.
coli strain BL21(DE3) pLysS.
The expression construct encoding the E64A mutant protein (pET17b-E64A) was
made by PCR-based site-directed mutagenesis using the construct pET17b-yCD
described elsewhere (17). The forward and reverse primers for the E64A mutant
construct are 5’-CC ACA CAT GGT GCG ATC TCC ACT TTG GAA AAC-3’ and 5’GTT TTC CAA AGT GGA GAT CGC ACC ATG TAG TGT GG-3’. The construct was
transformed into E. coli strain DH5!. The mutations were verified by DNA sequencing.
The mutant DNA constructs were transformed into E. coli strain BL21(DE3) pLysS.
The expression construct encoding the His-tagged E64A mutant enzyme with a
TEV protease cleavage site (pET17bHR-E64A) was made by ligating the pET17b-E64A
construct and the pET-17bHR vector, both digested with NdeI and EcoRI restriction
enzymes. The construct was transformed into E. coli strain DH5!. The mutations were
verified by DNA sequencing. The mutant DNA constructs were transformed into E. coli
strain BL21(DE3) pLysS.
The expression construct encoding the His-tagged E64D mutant enzyme with a
TEV protease cleavage site (pET17bHR-E64D) was made by PCR-based site-directed
mutagenesis using the construct pET17bHR-yCD. The forward and reverse primers for
the E64D mutant construct were 5’-CC ACA CTA CAT GGT GAC ATC TCC ACT
TTG GAA AAC-3’ and 5’-GTT TTC CAA AGT GGA GAT GTC ACC ATG TAG TGT
GG-3’. The construct was transformed into E. coli strain DH5!. The mutations were
verified by DNA sequencing. The mutant DNA constructs were transformed into E. coli
strain BL21(DE3) pLysS.
142

Expression and Purification. yCD and E64A and E64D mutant enzymes were
expressed using the same protocol implemented for the wild type protein, as described
elsewhere (17). Three liters of E. coli cells in LB medium were grown for each protein.
The cells were harvested by centrifugation, washed once with 50 mM phosphate buffer
(pH 8.0) and kept at -20 °C until they were used.
The E64A mutant protein without the His tag was purified using the same
protocol for the wild type yCD protein as described elsewhere (17). The His-tagged yCD
and E64A and E64D mutant enzymes were purified as described below. The frozen
bacterial paste was thawed at room temperature and suspended in 100 mL of precooled
buffer A [50 mM potassium phosphate buffer with 300 mM NaCl (pH 8.0)]. DNase I and
MgCl2 were added to reach a final concentration of 1 µg/ml and 20 mM, respectively.
The cells were disrupted with a French press. The resulting lysate was centrifuged for 20
min at ~27,000g and 4 °C. Polyethyleneimine was added to the pooled supernatant to a
final concentration of 0.1% to precipitate RNA. The solution was centrifuged
immediately after mixing at 15,000g for 30 min. The supernatant was loaded onto a
nickel-nitriloacetic acid (Ni-NTA) column equilibrated with 10 mM imidazole in buffer
A. The column was washed with 10 mM imidazole in the same buffer until the OD280 of
the effluent was <0.05. The column was eluted with a linear imidazole gradient (from 10
mM to 250 mM) in buffer A. Fractions containing the yCD protein were identified by
SDS-PAGE and concentrated to ~15 mL with an Amicon concentrator using an YM10
membrane. The protein solution was then subjected to TEV protease digestion for 4 h at
room temperature in the presence of 0.5 mM EDTA and 1 mM DTT. The digestion
mixture was dialyzed against buffer A and loaded onto another Ni-NTA column
143

equilibrated with 10 mM imidazole in buffer A. The column was eluted with the same
buffer. The yCD protein was washed out because it cannot bind with the resin after the
His tag was removed. The flow-through fraction was collected and dialyzed against 2
mM phosphate buffer, pH 8.0 and lyophilized. The powder was kept at -80 ºC until it was
used.
Kinetic Measurements. Steady-state kinetic parameters were measured by a
radiometric assay or an HPLC assay. The reaction was set up in a total volume of 40 µl in
50 mM phosphate buffer with 50 mM NaCl at pH 7.5. The final enzyme concentration
was 1 µM, 500 µM, and 34 µM for the wild-type, E64A, and E64D yCDs, respectively.
5FC concentrations were 0.08 - 0.68 mM for the wild-type yCD, 10-100 mM for E64A
3

yCD, and 6-64 mM for E64D yCD, respectively. A trace amount of [6- H]-5FC was used
to quantify the conversion of 5FC to 5FU. The reactions were initiated at 20 °C by the
addition of the enzymes. The reaction time was 10 s, 30 h and 8 h for the wild type, E64A
and E64D yCDs, respectively. The reactions were quenched and 5FC and 5FU were
separated by TLC as previously described (17). Kinetic parameters were obtained by
nonlinear least squares fitting of the data to the standard Michaelis-Menten equation. The
kinetic parameters of the E64A enzyme were also measured by an HPLC assay. In the
HPLC assay, the reactions were quenched by the addition of an equal volume of 0.1 N
HCl. The quenched solutions were injected onto a reverse-phase C18 column (22 mm !
250 mm) and the column was eluted with 20 mM phosphate buffer, pH 2.0. The elution
was monitored by UV absorption at 256 nm. The retention times for 5FC and 5FU were
3.6 and 5.8 min, respectively. The peak area of 5FU was integrated and compared with
standard samples of known concentrations.
144

Fluorometric Measurements. The dissociation constant for the binding of 5FPy
to the wild-type yCD was measured by fluorometry using a FluoroMax-2 fluorometer. A
series of 5FPy solutions of different concentrations were made in 50 mM phosphate
buffer, pH 7.5. A yCD stock solution in the same buffer was added to these 5FPy
solutions and the mixtures were incubated at room temperature for 10 min in order for
them to reach equilibrium. yCD and 5FPy concentrations were 1 µM and 0 - 6.4 µM,
respectively. Fluorescence intensities were measured at an emission wavelength of 386
nm with a slit of 4 nm. The excitation wavelength and slit were 320 nm and 4 nm,
respectively. A set of controls was obtained in the same manner by adding the buffer
instead of the protein stock solution. The Kd value was obtained by nonlinear leastsquares fitting of the difference data to the following equation:
!F =

df "(Pt + L t + K d #

2

(Pt + L t + K d ) # 4Pt L t )

(5-1)

2Pt

where !F is the fluorescence change upon ligand binding, df is the fluorescence
constant difference between the bound form and the free form ligand, Pt and Lt are the
total concentration of the enzyme and 5FPy, respectively and Kd is the dissociation
constant of 5FPy with the yCD enzyme.
Inhibition Assay. The inhibition of the 5FC deamination reaction by 5FPy was
measured by a radiometric assay as described earlier. IC50 values were obtained by
nonlinear least squares fitting of the inhibition assay data to the equation below:

145

v = kmin +

kmax ! kmin
[I]
1+
IC50

(5-2)

where v is the initial reaction rate, kmin the minimum rate, kmax the maximum
rate, and [I] the concentration of the inhibitor. Ki values were calculated by the ChengPrusoff equation (18):
Ki =

IC50

(5-3)

1 + [S] / K m

where [S] is the concentration of the substrate 5FC.
19

F NMR Spectroscopy. NMR samples were prepared by dissolving the wild-

type yCD or the E64A enzyme in 100 mM potassium phosphate buffer (pH 7.5) made in
D2O. An aliquot of a 5FPy stock solution was added to the protein solutions and the
mixtures were incubated for 10 min to reach equilibrium.

19

F NMR experiments were

performed at 10 or 20 °C on a Varian Inova 600 MHz NMR spectrometer. NMR spectra
were acquired with a spectral width of 10,000 Hz, 20,000 complex data points, and 1,024
transients and processed with a 10 Hz of line broadening. Relaxation delay was 6 s. The
19

F chemical shifts were referenced to CF3C6H5, which was set to -63.73 ppm.
ONIOM Calculations. The ONIOM method (19) as implemented in the

Gaussian program (Gaussian 03) (20) was used to investigate the effects of the E64A
substitution on yCD catalysis. The molecular set up and protocol for the ONION
calculations were essentially the same as previously described (16). Briefly, the
molecular system for the calculations was based on the 1.14-Å crystal structure of the

146

yCD-DHP complex (PDBID: 1p6o) (10). The enzyme is represented by the catalytic Zn
and the active site and neighboring residues, including Ile33, Asn51, Thr60, Leu61,
His62, Gly63, Glu64, Ile65, Leu88, Ser89, Pro90, Cys91, Asp92, Met93, Cys94, Thr95,
Phe114, Trp152, Phe153, Glu154, Asp155, and Ile156. The system was divided into two
layers. The inner layer is composed of the sidechains of residues Glu64, His62, Cys91,
Cys94, Asn51, Asp155, zinc and the zinc-bound water or substrate and its derivatives.
The outer layer composed of the rest of the molecular system. The inner layer was treated
by the density functional method B3LYP and the 6-31G** basis set and the outer layer
by the semi-empirical PM3 method. The total ONIOM energy E

ONIOM

is given by (19,

21):
E

ONIOM

= E(high, model) + E(low, real) - E(low, model)

(5-4)

where E(high, model) and E(low, model) are the energies of the inner layer at the
high level of theory (B3LYP and the 6-31G** basis set) and at the low level of theory
(PM3), respectively, and E(low, real) is the energy of the entire system at the level of
PM3.

RESULTS
Biochemical Analysis. To investigate the role of Glu64 in the activation of the
anticancer prodrug 5FC, we replaced Glu64 with alanine and aspartate. Preliminary
biochemical characterization of the mutant enzymes without a His tag indicated that the
purified mutant enzymes contained a minute amount of E. coli CD, even though the E.
coli enzyme is not homologous to the yeast enzyme, and differs in both size and
quaternary structure. To get rid of the E. coli CD contamination, genes encoding the
147

wild-type yCD and mutant proteins were sub-cloned into a vector for the production of
His-tagged proteins. The His-tagged proteins were purified with a Ni-NTA column and
the His-tag was removed by TEV protease digestion. Our analysis was focused on the
effects of the E64A substitution, which completely removed the carboxyl functional
group. The steady-state kinetic parameters of the wild-type yCD and the E64A and E64D
enzymes are summarized in Table 5-1. Because of the relatively large errors of the
radiometric assay in measuring a small amount of 5FU in the presence of a large amount
of 5FC, the kinetic parameters of the E64A mutant enzyme were also determined by
HPLC. Both substitutions caused a dramatic decrease in kcat and a dramatic increase in
Km. The E64A substitution caused a decrease in kcat by about 5 orders of magnitude and
an increase in Km by about 2 orders of magnitude, resulting in a decrease in kcat/Km by
about 8 orders of magnitude. The effects of the E64D substitution were only slightly
milder. The kinetic results indicated that Glu64 is involved in both binding and catalysis
in the activation of 5FC by yCD, and the longer side-chain is important, as Glu64 cannot
be replaced with aspartate with the same functional group but a shorter side-chain.
To further investigate the role of Glu64 in the activation of 5FC, we measured the
binding and inhibitory activities of the putative transition state analogue 5FPy. The
results are summarized in Table 5-2. The inhibitory activity of this compound on yCD
was measured for the E64D mutant enzyme but not for E64A protein, because the
activity of the latter mutant enzyme was too low for such an analysis. The E64D
substitution caused a decrease in Ki by about 4 orders of magnitude, indicating that the
mutant enzyme has a much lower affinity for 5FPy than the wild-type yCD. The Kd value
148

for the binding of 5FPy could be measured by fluorometric titration for the wild-type
yCD but not for either of the two mutant enzymes, because the method is not suitable for
the weak binding of 5FPy to the mutant enzymes. The Kd value for the binding of 5FPy
to the E64A mutant enzyme was estimated by NMR as described below.

19

F NMR Analysis. The bound Py in the crystal structure of yCD is in the

hydrated form DHP (10, 11). 5FPy differs from Py in having a fluorine atom at position 5
instead of a hydrogen. In order to determine whether the bound 5FPy is hydrated like Py
and to examine the role of Glu64 on the hydration,

19

F-NMR spectra were acquired for

5FPy in absence of and the presence of the wild-type yCD or the E64A mutant enzyme
(Figure 5-1). In the absence of the enzymes, 5FPy showed a sharp peak at -160.6 ppm. In
the presence of a lower concentration of the wild-type yCD, 5FPy showed an additional
broad peak due to the bound 5FPy at around -161.1 ppm, upfield of the peak of free
5FPy. In the presence of the same concentration of the E64A mutant enzyme, 5FPy also
showed a broad but much smaller peak due to the bound 5FPy at -158.1 ppm, downfield
of the peak of free 5FPy. Using model compounds, it has been shown that the

19

F NMR

signal is a good indicator of the hydration of 5FPy (22). The hydration of 5FPy results in
a change in the bonding of C4 from sp2 to sp3, causing a ~1 ppm upfield shift of the
NMR signal. This upfield shift of the

19

19

F

F NMR signal has been used for the

determination of the hydration state of 5-fluoropyrimidin-2-one ribonucleoside (22). The
dramatic opposite shift of the

19

F NMR signal of 5FPy in Figure 5-1 indicates that 5FPy

149

is hydrated when it binds to the wild-type yCD and that it is not hydrated when it binds to
the E64A mutant enzyme. In either case, however, free 5FPy is in slow exchange with the
bound 5FPy on the NMR chemical shift timescale. In such an exchange regime, chemical
shift does not change with protein or ligand concentration and cannot be used for the
estimation of Kd values. However, the relative peak intensities of the NMR signals of the
free form and the bound form can be used to estimate the Kd value if the protein and
ligand concentrations are the same order of magnitude as the Kd value. The protein and
5FPy concentrations were suitable for the estimation of the Kd value for the E64A mutant
enzyme but not for the wild-type yCD. The Kd value for the binding of 5FPy to the E64A
mutant enzyme was estimated to be ~10 mM based on the

19

F NMR data, about 5 orders

of magnitude higher than that for the binding of 5FPy to the wild-type enzyme. As 5FPy
is the hydrated form when it binds to the wild-type yCD and the unhydrated form is
3

favored over the hydrated form by a factor of 2.9!10 , the Kd for the binding of the
hydrated 5FPy, a transition state analogue, to the wild-type CD was estimated to be 2.3
-10

!10

M.

Computational Analysis. Because 5FPy is hydrated when bound to the wild-type
yCD, but not hydrated when bound to the E64A mutant enzyme, we investigated the
hydration by yCD by ONIOM calculations taking advantage of the availability of the
high resolution crystal structures of the enzyme in complex with DHP, the hydrated form
of Py. Py was docked into the active site by superposition with DHP in the X-ray crystal
150

structure (10). After optimization, Py shifted slightly from the initial position to give
Complex 1 (Figure 5-2 and Figure 5-3). Complex 1 is stabilized through a hydrogen bond
network and a !-stacking interaction. One new interaction is present between the zincZn

coordinated water O

and N3 of Py. Glu64 is in the deprotonated form and forms a

strong hydrogen bond with the other hydrogen of the zinc-coordinated water. The
carbonyl oxygen (O2) of Py is hydrogen bonded to the side chain amide of Asn51 and the
backbone amide of Gly63. The amide group at position 1 (N1) of Py is hydrogen bonded
to the carboxyl group of Asp155. The latter is also hydrogen bonded to the side chain
"2

amide of Asn51. The hydrogen bond between the amide group (N ) of His62 and the
#1

backbone carbonyl oxygen (O ) of Asp155 helps in maintaining appropriate orientation
of the imidazole ring of His62 that stacks with the cytosine ring. Before the hydration
occurs, the water is converted to a hydroxide with the assistance of Glu64. First, the
Zn

hydrogen bond between O

Zn

and N3 is broken and a new hydrogen bond between O

and Glu64 is formed (not the one already present). The new conformation is given in
Zn

Complex 2 (Figures 5-2 and 5-3). In this complex, the hydrogen bond O …N3 is not
completely disrupted. One important feature of this conformation is that the distance
Zn

between O

OZn

and C4 is shortened because the change of the H

position reduces the

steric hindrance. Complex 2 is calculated to be less stable than 1 by 2.9 kcal/mol (Figure
5-4). However, 2 is more stable than 1 by 5.4 kcal/mol at the E(high, model) energy
level. The difference of 8.3 kcal/mol indicates the surrounding protein destabilizes the
complex. The corresponding barrier is 3.2 kcal/mol and there is no barrier found at the
151

OZn

E(high, model) energy surface. Second, H

Zn

is transferred from O

to Glu64, as seen
Zn

in Complex 3 (Figures 5-2 and 5-3). In this process, the hydrogen bond between O

and

N3 is completely broken. After the proton transfer, a strong hydrogen bond is formed
Zn

between the Zn-coordinated hydroxide and Glu64-OH. The distance between O

and C4

is further reduced. Complex 3 is slightly more stable than Complex 2. The energy barrier
is calculated to be 5.6 kcal/mol at the ONIOM energy level. Again, there is no barrier
Zn

found at the E(high, model) level. In the next step, the hydrogen bond between O

and

Glu64-OH is broken and a new hydrogen bond is formed between Glu64-OH and N3,
Zn

which can be seen in Complex 4 (Figures 5-2 and 5-3). The distance between O

and C4

is shortened to 2.17 Å. The energy barrier is 0.9 kcal/mol at the ONIOM energy level and
1.7 at the E(high, model) energy level. Complex 4 is more stable than the initial Complex
1 by 7.5 kcal/mol at the ONIOM energy level and by 11.8 kcal/mol at the E(high, model)
energy level. The Zn-coordinated hydroxide in Complex 4 is well positioned for the
nucleophilic attack on C4. The distance between the hydroxide and C4 decreases from
3.01 Å in Complex 1 to 2.66 Å in Complex 2, to 2.44 Å in Complex 3, and 2.17 Å in
Complex 4. The nucleophilic attack on C4 is concomitant with the proton transfer from
Glu64-OH to N3, resulting in the formation of the tetrahedral product (DHP) as shown in
Complex 5 (Figures 5-2 and 5-3). The energy barrier is 1.0 kcal/mol at the ONIOM level
and 0.7 kcal/mol at the E(high, model) level. The bond lengths, bond angles, and dihedral
angles of DHP in complex 5 are in close agreement with those of DHP in the crystal
structure. The yCD-DHP complex is more stable than the initial yCD-Py complex by
23.2 kcal/mol at the ONIOM level and 30.9 kcal/mol at the E(high, model) level. The
152

difference of 7.7 kcal/mol indicates that the model system is relatively destabilized by the
outer layer. 5 is considerably more stable than 1 because of the intrinsic stability of the
model system, which suggests that the tight binding of the inhibitor is mainly due to the
high affinity of DHP for the wild-type yCD.
To investigate the effects of the E64A substitution on the hydration, an E64A-Py
complex was constructed in the same manner as for the wild-type yCD complex with Py
(Complex 6 in Figures 5-5 and 5-6), assuming that the structure of the E64A mutant
enzyme is the same as that of the wild-type yCD. The conformation of the optimized
complex 6 is similar to that of the corresponding wild-type yCD complex. The Zncoordinated water, however, forms only one hydrogen bond (with N3 of Py), and the
other hydrogen bond in the yCD-Py complex is absent, as Glu64 is not present in the
mutant enzyme complex. Because of the lack of Glu64, the Zn-coordinated water cannot
be converted to hydroxide before the nucleophilic attack on C4. The distance between
Zn

O

and C4 is 3.08 Å, longer than that in the yCD-Py complex, making it even harder to

attack C4. The nucleophilic attack of the Zn-coordinated water on C4 of Py is
concomitant with the transfer of one of its protons to N3 of Py to form the tetrahedral
product as in Complex 7 (Figures 5-5 and 5-6). The approximate geometry of transition
OZn

state TS67 was found by scanning the H

Zn

-N3 and O -C4 distances from complex 6
Zn

to Complex 7. The transition state was found where the O -C4 distance is 2.58 Å. The
energy barrier is 14.4 kcal/mol at the ONIOM energy level and 36.1 kcal/mol at the
E(high, model) level (Figure 5-7), indicating that the protein environment (the outer
layer) compensates for the barrier by 21.7 kcal/mol. The energy barrier is much higher

153

than the corresponding energy barrier for the hydration catalyzed by the wild-type yCD.
Zn

It is noted that the proton is transferred slightly ahead of the formation of the O -C4
bond. Complex 7 is calculated to be less stable than 6 by 1.7 kcal/mol at the ONIOM
level, indicating that Py is predominantly in the unhydrated form.
To investigate further the effects of the E64A substitution on the deamination
reaction, a complex of the E64A mutant enzyme with the substrate cytosine was
constructed as for the construction of the wild-type enzyme-substrate complex (16),
assuming again that the structure of the mutant enzyme is the same as that of the wildtype enzyme. A water molecule was added in place of the carboxyl group of Glu64. After
optimization, cytosine shifted slightly from the initial position to give Complex 8
(Figures 5-8 and 5-9). The interactions between the substrate and enzyme in complex 8
are similar to those in the corresponding wild-type enzyme complex (16) except for the
interaction involving the Zn-coordinated water and Glu64. In the wild-type enzyme
complex, Glu64 forms a hydrogen bond with the Zn-coordinated water and another with
the amino group of cytosine. In the mutant enzyme complex, the Zn-coordinated water
forms a hydrogen bond with N3 of cytosine, and the amino group of cytosine forms a
hydrogen bond with the additional water molecule. As the Zn-coordinated water is
hydrogen bonded to N3 and cannot be deprotonated by the additional water molecule, the
Zn-coordinated hydroxyl is generated by a direct proton transfer to N3. The energy
barrier for the proton transfer is 7.7 kcal/mol at the ONIOM energy level and 13.6
kcal/mol at the E(high, model) level. But the hydroxide is not as well positioned for the
nucleophilic attack on C4 as in the corresponding wild-type complex (16). The distance
Zn

between O

and C4 is only 2.24 Å in the wild-type enzyme complex but is 2.9 Å in the
154

Zn

mutant enzyme complex. The transition state is found when the distance between O

and C4 is 2.48 Å. The energy barrier for the nucleophilic attack is calculated to be 11.2
kcal/mol at the ONIOM energy level and 22.1 kcal/mol at the E(high, model) level. For
the wild-type yCD-catalyzed reaction, the formation of the tetrahedral intermediate is
through three steps, in which the highest energy barrier is only 1.9 kcal/mol (16). The
stabilization of the tetrahedral intermediate is also affected, because the energy level of
the tetrahedral intermediate is 6.1 kcal/mol higher than that of the substrate complex in
the mutant enzyme-catalyzed reaction. However the energy level of the same
intermediate is 10.8 kcal/mol lower than that of the substrate complex during the wildtype yCD catalyzed reaction (16).
Zn

Once the tetrahedral intermediate is formed, another proton transfer from O

to

N4 is needed before the C4-N4 bond is cleaved (16). Two possible pathways were
explored. In pathway one, the proton is transferred directly to N4 via a four-member ring
transition state. In pathway two, the proton transfer is mediated by the additional water
Zn

molecule. The water donates a proton to N4 and extracts a proton from O

in a

concerted manner. The C4-N4 bond distance increases to 1.64 Å after the proton is
transferred to N4. The energy barrier in second pathway (12.1 kcal/mol at the ONIOM
energy level) is dramatically lower than that in the first pathway (22.3 kcal/mol at the
Zn

ONIOM energy level), indicating that the O

proton is transferred to N4 via the second

pathway. The energy barrier is high when compared with the wild-type yCD catalyzed
deamination, in which the energy barrier is only 7.8 kcal/mol at the ONIOM energy level.

155

DISCUSSION
The results of the combined experimental and computational studies reported here
indicate that Glu64 is important for both binding and chemical steps in the activation of
the anticancer prodrug 5FC. Assuming that the activation of 5FC follows a mechanism
similar to that of the deamination of cytosine, it involves the generation of a tetrahedral
reaction intermediate from the substrate and the conversion of the intermediate to the
product. A critical role of Glu64 in the chemical steps is supported by the results of the
biochemical analysis of the mutant enzymes. First, the kcat of the E64A mutant enzyme
decreases by 5 orders of magnitude relative to that of the wild-type yCD. Our previous
transient kinetic and NMR studies showed that product release is rate-limiting in the
activation of 5FC by the wild-type yCD (17). Consequently, the kcat of the wild-type
enzyme is significantly smaller than the rate constant for the chemical transformation. On
the other hand, the rate-limiting step is most likely the chemical transformation in the
activation of 5FC by the mutant enzyme, because of its severely impaired catalytic
apparatus. Therefore, the kcat of the E64A mutant enzyme should be compared with the
corresponding rate constant of the wild-type yCD for the chemical transformation, and
thus the E64A mutation slows down the chemical transformation by 6 orders of
magnitude. Glu64 contributes to the stabilization of the transition state of the activation
of 5FC by ~9 kcal/mol. Second, inhibition and binding studies indicate that the Ki of the
putative transition state analogue 5FPy for the inhibition of the E64D mutant enzyme
increases by 3 orders of magnitude relative to that for the inhibition of the wild-type yCD
and the Kd of 5FPy for the binding of the E64A mutant enzyme increases by 4 orders of
156

magnitude relative to that for the binding of the wild-type yCD, corresponding to ~6
kcal/mol.
A more precise role of Glu64 in the activation of 5FC is revealed by the NMR
analysis of the binding of 5FPy to the wild-type and mutant enzymes and the
computational analysis of the hydration of Py by the wild-type yCD and the deamination
of cytosine by the E64A mutant enzyme. The NMR analysis shows that 5FPy is hydrated
when bound to the wild-type yCD but is unhydrated when bound to the E64A mutant
enzyme, suggesting that Glu64 plays a critical role in the formation and/or stabilization of
the hydrated 5FPy. The ONION calculations of the hydration of Py indicate that Glu64
functions as a proton shuttle for the transfer of a proton from the Zn-coordinated water to
N3 of Py and the hydration follows a sequential mechanism with N3 protonated first
followed by the nucleophilic attack of C4 by the Zn-coordinated hydroxide. Furthermore,
the complex of the hydrated Py is much more stable than that of the unhydrated Py, based
on the dramatic difference in the ONION energies of the two complexes. Without the
proton shuttle Glu64 in the E64A mutant enzyme, the hydration follows a concerted
mechanism with a significantly higher energy barrier and the ONION energy of the
complex of the hydrated Py is higher than that of the complex of the unhydrated Py. The
results of the ONION calculations suggest that Glu64 facilitates the hydration of Py and
stabilizes the hydrated Py. Since the hydration reaction mimics the formation of the
reaction intermediate in the deamination reaction and 5FPy is predominantly in the
unhydrated form in the active center of the E64A mutant enzyme, Glu64 may play a
critical role in the formation of the reaction intermediate in the activation of the
anticancer prodrug 5FC.

157

The ONION calculations of the deamination reaction catalyzed by the E64A
mutant enzyme show that Glu64 is important for both the formation of the tetrahedral
reaction intermediate from the substrate and the conversion of the reaction intermediate
to the product. Without Glu64 shuttling protons, the barriers for both the formation of the
tetrahedral reaction intermediate and its conversion to the product are much higher than
those in the wild-type enzyme reaction. Furthermore, the ONION energy of the reaction
intermediate complex is 10.8 kcal/mol lower than that of the substrate complex in the
wild-type enzyme reaction but is 6.4 kcal/mol higher in the mutant enzyme reaction. The
combination of these effects makes the mutant enzyme a very sluggish catalyst.
Glu64 is also important for the binding of the anticancer prodrug 5FC. The Km of
the E64A mutant enzyme for 5FC increases more than 200-fold in comparison with that
of the wild-type enzyme. Because the rate-liming step in the activation of 5FC by the
wild-type yCD is product release, the Km of the enzymatic reaction is a complex
parameter, smaller than the Kd for the binding of 5FC (17). On the other hand, the ratelimiting step for the activation of 5FC by the E64A mutant enzyme must be the chemical
transformation. Consequently, 5FC is in rapid equilibrium with its complex with the
E64A mutant enzyme and thus the Km and Kd values should be similar. Therefore, the
Km value of the mutant enzyme is a good measure of the affinity of the mutant enzyme
for 5FC and should be compared with the Kd value of the wild-type enzyme for the
binding of 5FC, which was determined to be 0.19 mM. The ratio of the mutant enzyme’s
Km to the wild-type enzyme’s Kd is ~170, indicating that Glu64 contributes to the

158

binding of 5FC by ~3 kcal/mol. The biochemical result is consistent with the complex of
the enzyme with cytosine (16) built based on the crystal structure of the complex of yCD
with the hydrated Py (10, 11). After energy minimization, Glu64 forms a hydrogen bond
with the amino group of cytosine. Disruption of the hydrogen bond by mutagenesis is
likely to decrease the affinity of the enzyme for the substrate.
yCD belongs to a family of purine/pyrimidine deaminases with an active center
featuring a catalytic zinc coordinated with cysteine and histidine residues and a strictly
conserved glutamate. Interestingly, yCD is not homologous to eCD, which has a very
different size, a different quaternary structure, and an active center containing a ferrous
iron ion coordinated with four histidine residues, an aspartate residue, and a water
molecule. Among the family of purine/pyrimidine deaminases, yCD is most similar to E.
coli cytidine deaminase (CDA), having a catalytic Zn of the same coordination chemistry
and a superimposable glutamate (10, 11). However, the functional role of Glu64 of yCD
is not the same as that of the corresponding residue Glu104 of E. coli CDA, based on our
studies on yCD presented here and the published mutagenesis study of E. coli CDA (23).
Both glutamate residues are critical for catalyzing their respective reactions. The E104A
mutation of E. coli CDA decreases the kcat of the enzyme by 8 orders of magnitude (23).
The effect is more dramatic than that of the E64A mutation of yCD, which decreases the
kcat of yCD by 5 orders of magnitude and the rate constant for the chemical
transformation by 6 orders of magnitude. Similarly, the fluorinated transition-state
analogue 5-fluoro-pyrimidin-2-one riboside is in the hydrated form when bound to the
wild-type E. coli CDA but is in the unhydrated form when bound to the E104A mutant

159

enzyme (22), suggesting that Glu104 also plays a critical role in the generation of the
reaction intermediate in the deamination reaction catalyzed by E. coli CDA.
The roles of the two glutamate residues in substrate binding are different. Glu64
of yCD contributes to the stability of the yCD complex with 5FC by ~3 kcal/mol, as
discussed earlier. In contrast, Glu104 of E. coli CDA destabilizes the substrate complex
of the enzyme E. coli CDA (23). Thus, the Km of the E104A mutant enzyme decreases by
a factor of ~30. Interestingly, the Kd for the binding of the fluorinated transition-state
analogue 5-fluoro-pyrimidin-2-one riboside to the E104A mutant enzyme is even lower
than that for its binding to the wild-type enzyme, although the fluorinated transition-state
analogue is in the unhydrated form when bound to the mutant enzyme, suggesting that
Glu104 of E. coli CDA dramatically destabilizes the complex of the unhydrated form of
the analogue with the enzyme. It has been proposed that Glu104 of E. coli CDA
decreases the energy barrier of the deamination reaction not only by stabilizing the
transition state but also by destabilizing the enzyme-substrate complex (22). The
destabilization of the enzyme-substrate complex is probably due to the unfavorable
interaction between the negatively charged Glu104 and the lone pair of electrons of N3 of
cytidine, which becomes a favorable interaction when N3 is protonated during the
deamination reaction. E. coli CDA can afford this sacrifice of substrate binding energy
for the reduction of the energy barrier of the deamination reaction, because the ribose
moiety of cytidine contributes to the binding energy for the formation of the enzymesubstrate complex and consequently the enzyme is still able to maintain its Km in the submillimolar range even with the unfavorable interaction between the carboxyl group of

160

Glu104 and N3 of cytidine. On the other hand, without a ribose moiety, cytosine needs a
favorable interaction with the carboxyl group of Glu64 of yCD, i.e. a hydrogen bond
between the amino group of cytosine and the carboxyl group of Glu64 as described
earlier, to keep its Km in the millimolar range. Without this favorable interaction, the Km
increases by more than an order of magnitude as shown this mutagenesis study and a CD
with a Km in the ten millimolar range would not be useful for pyrimidine salvage.
In conclusion, the combined experimental and computational studies show that
Glu64 plays a critical role in the activation of the anticancer prodrug 5FC. It is important
not only for the deamination reaction but also for the binding of the substrate. In the
deamination reaction, Glu64 facilitates both the formation of the tetrahedral intermediate
and its conversion to the product by shuttling protons. Furthermore, it stabilizes the
tetrahedral intermediate and the hydrated form of the inhibitor 5FPy, which mimics the
transition state and/or the reaction intermediate. In contrast to Glu104 of E. coli CDA,
which destabilizes the enzyme-substrate complex, Glu64 of yCD is also important for
substrate binding. Its contribution to substrate binding is of great physiological
significance. Without Glu64, the Km of the enzyme would be too high for pyrimidine
salvage.

161

TABLES AND FIGURES

Table 5-1: Kinetic constants of the wild-type yCD and the E64 mutant enzymes for the
activation of 5FC.
-1

-1 -1

Protein

Km (mM)

kcat (s )

kcat/Km (mM s )

WT

0.16 ± 0.01

17 ± 0.4

106

E64A

89 ± 30
33 ± 6

E64D

a

b

19 ± 6

-4 a

1.4 ! 10

-6 a

-5 b

2.2 ! 10

-3

2.6 ! 10

(1.3 ± 0.3) ! 10

-6 b

(7.3 ± 0.6) ! 10

-4

(5.1 ± 1.0) ! 10

a: Determined by a radiometric assay.
b: Determined by HPLC.

Table 5-2: Inhibitory and binding properties of 5FPy.
Protein

Ki (µM)

Kd (µM)

WT

0.86 ± 0.16

0.65 ± 0.08
4b

E64A
E64D

1.0 ! 10
1060 ± 60

a: Measured by fluorometric titration.
19
b: Estimated by F NMR.

162

a

19

Figure 5-1. F-NMR spectra of the E64A-5FPy complex and the wild-type yCD-5FPy
complex. All samples were in a D2O buffer containing 100 mM potassium phosphate, pH
7.5. (A) Spectrum of 5FPy (4.0 mM) in the absence of the enzyme; (B) spectrum of 5FPy
(3.0 mM) in the presence of the wild-type yCD (1.5 mM); (C) spectrum of 5FPy (3.0
mM) in the presence of E64A (1.5 mM).

163

Figure 5-2. Proposed reaction pathway from Py to DHP catalyzed by the wild type yCD.

164

Figure 5-3. ONIOM optimized structures for the conversion of Py to DHP by the wild
type yCD. (1) complex 1, (2) TS12 between complex 1 and 2, (3) complex 2, (4) TS23
between complex 2 and 3, (5) complex 3, (6) TS34 between complex 3 and 4, (7)
complex 4, (8) TS45 between complex 4 and 5, (9) complex 5.

165

ONIOM

Figure 5-4. Schematic E
and E(High,model) energy profile for the conversion of
Py to DHP catalyzed by the wild-type yCD.

166

Figure 5-5. Proposed reaction pathway from Py to DHP catalyzed by the E64A mutant
enzyme.

Figure 5-6. ONIOM optimized structures for the conversion of Py to DHP by the E64A
mutant enzyme. (1) complex 6, (2) TS67 between complex 6 and 7, (3) complex 7.

167

ONIOM

Figure 5-7. Schematic E
and E(High,model) energy profile for the conversion
from Py to DHP catalyzed by the E64A mutant enzyme.

168

Figure 5-8. Conversion of cytosine to the zinc-coordinated uracil by the E64A mutant
enzyme.

169

Figure 5-9. ONIOM optimized structures for the conversion from cytosine to the zinccoordinated uracil. (1) complex 8, (2) TS89 between complex 8 and 9, (3) complex 9, (4)
TS910 between complex 9 and 10 and (5) complex 10.

170

ONIOM

Figure 5-10. Schematic E
and E(High,model) energy profile for the conversion
from cytosine to uracil catalyzed by the E64A mutant enzyme.

171

REFERENCES

172

REFERENCES
1. Greco, O., and Dachs, G. U. (2001) Gene directed enzyme/prodrug therapy of cancer:
historical appraisal and future prospectives, J Cell Physiol 187, 22-36.
2. Aghi, M., Hochberg, F., and Breakefield, X. (2000) Prodrug activation enzymes in
cancer gene therapy, J Gene Med 2, 148-164.
3. Ireton, G. C., McDermott, G., Black, M. E., and Stoddard, B. L. (2002) The structure
of Escherichia coli cytosine deaminase, J Mol Biol 315, 687-697.
4. Ipata, P., Marmocch, F., Magni, G., Feliciol, R., and Polidoro, G. (1971) Bakers yeast
cytosine deaminase - Some enzymic properties and allosteric inhibition by
nucleosides and nucleotides, Biochemistry 10, 4270-4276.
5. Hayden, M., Linsley, P., Wallace, A., Marquardt, H., and Kerr, D. (1998) Cloning,
overexpression, and purification of cytosine deaminase from Saccharomyces
cerevisiae, Protein Expres Purif 12, 173-184.
6. Hsu, Y., Hu, C., Lin, J., and Liaw, S. (2003) Crystallization and preliminary
crystallographic analysis of yeast cytosine deaminase, Acta Crystallogr D 59, 950952.
7. Kievit, E., Bershad, E., Ng, E., Sethna, P., Dev, I., Lawrence, T., and Rehemtulla, A.
(1999) Superiority of yeast over bacterial cytosine deaminase for emzyme/prodrug
gene therapy in colon cancer xenografts, Cancer Res 59, 1417-1421.
8. Kievit, E., Nyati, M., Ng, E., Stegman, L., Parsels, J., Ross, B., Rehemtulla, A., and
Lawrence, T. (2000) Yeast cytosine deaminase improves radiosensitization and
bystander effect by 5-fluorocytosine of human colorectal cancer xenografts, Cancer
Res 60, 6649-6655.
9. Hamstra, D., Rice, D., Fahmy, S., Ross, B., and Rehemtulla, A. (1999)
Enzyme/prodrug therapy for head and neck cancer using a catalytically superior
cytosine deaminase, Hum Gene Ther 10, 1993-2003.
10. Ireton, G., Black, M., and Stoddard, B. (2003) The 1.14 angstrom crystal structure of
yeast cytosine deaminase: Evolution of nucleotide salvage enzymes and implications
for genetic chemotherapy, Structure 11, 961-972.
11. Ko, T., Lin, J., Hu, C., Hsu, Y., Wang, A., and Liaw, S. (2003) Crystal structure of
yeast cytosine deaminase - Insights into enzyme mechanism and evolution, J Biol
Chem 278, 19111-19117.
12. Xu, Q., Guo, H., Gorin, A., and Guo, H. (2007) Stabilization of a transition-state
analogue at the active site of yeast cytosine deaminase: Importance of proton
transfers, J Phys Chem B 111, 6501-6506.
13. Yao, L., Sklenak, S., Yan, H., and Cukier, R. (2005) A molecular dynamics
exploration of the catalytic mechanism of yeast cytosine deaminase, J Phys Chem B
173

109, 7500-7510.
14. Yao, L., Yan, H., and Cukier, R. (2006) A combined ONIOM quantum chemicalmolecular dynamics study of zinc-uracil bond breaking in yeast cytosine deaminase,
J Phys Chem B 110, 26320-26326.
15. Yao, L., Yan, H., and Cukier, R. (2007) A molecular dynamics study of the ligand
release path inyeast cytosine deaminase, Biophys J 92, 2301-2310.
16. Sklenak, S., Yao, L., Cukier, R., and Yan, H. (2004) Catalytic mechanism of yeast
cytosine deaminase: An ONIOM computational study, J Am Chem Soc 126, 1487914889.
17. Yao, L., Li, Y., Wu, Y., Liu, A., and Yan, H. (2005) Product release is rate-limiting
in the activation of the prodrug 5-fluorocytosine by yeast cytosine deaminase,
Biochemistry 44, 5940-5947.
18. Cheng, Y., and Prusoff, W. H. (1973) Relationship between the inhibition constant
(Ki) and the concentration of inhibitor which causes 50 per cent inhibition (I50) of an
enzymatic reaction, Biochem Pharmacol 22, 3099-3108.
19. Svensson, M., Humbel, S., Froese, R. D. J., Matsubara, T., Sieber, S., and
Morokuma, K. (1996) ONIOM: A multilayered integrated MO + MM method for
geometry optimizations and single point energy predictions. A test for Diels-Alder
Reactions and Pt(P(t-Bu)3)2 + H2 oxidative addition, J Phys Chem 100, 1935719363.
20. Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A.,
Cheeseman, J. R., Montgomery, J., Vreven, T., Kudin, K. N., Burant, J. C., Millam, J.
M., Iyengar, S. S., Tomasi, J., Barone, V., Mennucci, B., Cossi, M., Scalmani, G.,
Rega, N., Petersson, G. A., Nakatsuji, H., Hada, M., Ehara, M., Toyota, K., Fukuda,
R., Hasegawa, J., Ishida, M., Nakajima, T., Honda, Y., Kitao, O., Nakai, H., Klene,
M., Li, X., Knox, J. E., Hratchian, H. P., Cross, J. B., Bakken, V., Adamo, C.,
Jaramillo, J., Gomperts, R., Stratmann, R., Yazyev, O., Austin, A. J., Cammi, R.,
Pomelli, C., Ochterski, J. W., Ayala, P. Y., Morokuma, K., Voth, G. A., Salvador, P.,
Dannenberg, J. J., Zakrzewski, V. G., Dapprich, S., Daniels, A. D., Strain, M. C.,
Farkas, O., Malick, D. K., Rabuck, A. D., Raghavachari, K., Foresman, J. B., Ortiz, J.
V., Cui, Q., Baboul, A. G., Clifford, S., Cioslowski, J., Stefanov, B. B., Liu, G.,
Liashenko, A., Piskorz, P., Komaromi, I., Martin, R. L., Fox, D. J., Keith, T., AlLaham, M. A., Peng, C. Y., Nanayakkara, A., Challacombe, M., Gill, P. M. W.,
Johnson, B., Chen, W., Wong, M. W., Gonzalez, C., and Pople, J. A. (2004) Gaussian
03, Revision D.01. Gaussian, Inc., Wallingford CT.
21. Humbel, S., Sieber, S., and Morokuma, K. (1996) The IMOMO method: Integration
of different levels of molecular orbital approximations for geometry optimization of

-

large systems: Test for n-butane conformation and SN2 reaction: RCl+Cl , J Chem
Phys 105, 1959-1967.
174

22. Carlow, D. C., Short, S. A., and Wolfenden, R. (1996) Role of glutamate-104 in
generating a transition state analogue inhibitor at the active site of cytidine
deaminase, Biochemistry 35, 948-954.
23. Carlow, D. C., Smith, A. A., Yang, C. C., Short, S. A., and Wolfenden, R. (1995)
Major contribution of a carboxymethyl group to transition-state stabilization by
cytidine deaminase: mutation and rescue, Biochemistry 34, 4220-4224.

175

CHAPTER 6

CONTRIBUTIONS OF CONSERVED RESIDUES IN THE
ACTIVE SITE TO SUBSTRATE BINDING AND
CATALYSIS

INTRODUCTION
Yeast cytosine deaminase (yCD), an enzyme participating in the pyrimidine
salvage pathway in fungi, catalyzes the deamination of cytosine to uracil. yCD is of great
biomedical interest because it also catalyzes the conversion of the prodrug 5fluorocytosine (5FC) to 5-fluorouracil (5FU), which is a widely used drug in the
treatment of breast, rectal, stomach, colon, and pancreatic cancers (1, 2). 5FU has high
gastrointestinal and hematological toxicities. The undesirable side effect is minimized (3)
by the combination of yCD/5FC in gene-directed enzyme prodrug therapy (GDEPT)
through which the administrated nontoxic 5FC is converted to the toxic and lethal 5FU in
the cancer cells by yCD. This strategy takes advantage of the fact that there is no cytosine
deaminase expressed in mammalian cells (4). Thus, in principle, only the tumor cells are
exposed to the toxic 5FU and killed, while the normal cells are not affected.
Crystal structures of yCD have been determined at high resolution, including the
substrate-free form (PDB id: 1ox7) and the complex form (PDB id: 1p6o and 1uaq) with
2-pyrimidinone (Py), a transition state analog (4, 5). yCD is a homodimeric

176

metalloenzyme. Each monomer is tightly packed into a mixed !/" fold, which consists of
a central five stranded "-sheet sandwiched by !-helices on both sides. A catalytic zinc
#1

$

$

atom is found in the active site, covalently bonded with His62 N , Cys91 S , Cys94 S

and a water molecule in the free form yCD. The inhibitor is bound in the hydrated form
4(R)-hydroxyl-3,4-dihydropyrimidine (DHP) and coordinates with the zinc atom through
the 4-hydroxyl group. In both the free form and the complex form enzyme, the active site
is completely covered by the same cluster of residues from the C-terminal !-helix and the
loop between "4 and !D and appears to be inaccessible to the substrate.
The yCD-DHP complex is stabilized by a hydrogen bond network. At one end of
#1

the pyrimidine ring of DHP, N1 forms a hydrogen bond with the carboxyl group (O ) of
#2

Asp155 and O2 forms hydrogen bonds with the side chain amide (N ) of Asn51 and
with the backbone amide of Gly63, respectively. At the other end, the zinc-coordinated
hydroxide (O4) is hydrogen bonded with the backbone amide group of Cys91. Glu64 is
believed to be critical for the stabilization of DHP (see Chapter 5) through two hydrogen
bonds, one with O4 and the other with N3. DHP is stabilized by hydrophobic interactions
as well. A cluster of residues, including Phe114, Trp152, Ile156, Trp152, Ile33 and
Leu88, protect DHP from being exposed to the solvent.
In the previous chapter, the function of Glu64 was investigated. The results show
that Glu64 does not merely function in catalysis, but also contributes significantly to
substrate binding. In the closely related enzyme E. coli cytidine deaminase, the
counterpart of Glu64 seems not to affect substrate binding (6). Instead, the authors
suggest that the glutamate may destabilize the substrate binding. In this chapter, I
177

continue investigating functions of conserved residues in the active site. A series of
mutant proteins, including substitutions involving N51A, D155A, E154G, E154D, R73A
and D151A, were constructed. Their functions were investigated by kinetic, NMR, and
computational methods.

EXPERIMENTAL PROCEDURES
3

Materials. Cytosine, uracil, 5FC and 5FU were purchased from Sigma. [6- H]5FC was purchased from Moravek. 5-fluoro-pyrimidinone was synthesized in our lab.
Restriction enzymes and T4 ligase were purchased from New England Biolabs. Pfu DNA
polymerase and the pET-17b vector were purchased from Strategene and Novagen,
respectively.
Site-Directed Mutagenesis and Protein Expression. The pET-17bHR and
pET17bHR-yCD vectors were constructed as described in Chapter 5. The expression
constructs encoding the His-tagged N51A, D155A, E154G, E154D, R73A, and D151A
proteins with a TEV protease cleavage site were made by PCR-based site-directed
mutagenesis using the construct pET17bHR-yCD. The forward and reverse primers for
the N51A construct were 5’-CTC GGT CGT GGT CAC GCC ATG AGA TTT CAA
AAG GG-3’ and 5’-CCC TTT TGA AAT CTC ATG GCG TGA CCA CGA CCG AG3’, respectively. The forward and reverse primers for the D155A construct were 5’-CT
CAG GAT TGG TTT GAA GCT ATT GGT GAG TAG AAT TCC-3’ and 5’-GGA ATT
CTA CTC ACC AAT AGC TTC AAA CCA ATC CTG AG-3’, respectively. The
forward and reverse primers for the E154G construct were 5’-GA CCT CAG GAT TGG
TTT GGT GAT ATT GGT GAG TAG AAT TC-3’ and 5’-GAA TTC TAC TCA CCA
ATA TCA CCA AAC CAA TCC TGA GGT C-3’, respectively. The forward and reverse
178

primers for the E154D construct were 5’-GA CCT CAG GAT TGG TTT GAT GAT
ATT GGT GAG TAG AAT TC-3’ and 5’-GAA TTC TAC TCA CCA ATA TCA TCA
AAC CAA TCC TGA GGT C -3’, respectively. The forward and reverse primers for the
R73A construct were 5’-CT TTG GAA AAC TGT GGG GCA TTA GAG GGC AAA
GTG TAC-3’ and 5’-GTA CAC TTT GCC CTC TAA TGC CCC ACA GTT TTC CAA
AG-3’, respectively. The forward and reverse primers for the D151A construct were 5’-C
GAT GAA AGA CCT CAG GCT TGG TTT GAA GAT ATT GG-3’ and 5’-CCA ATA
TCT TCA AAC CAA GCC TGA GGT CTT TCA TCG -3’, respectively.
The constructs were transformed into E. coli strain DH5!. The mutations were
verified by DNA sequencing. The mutant DNA constructs were transformed into E. coli
strain BL21(DE3) pLysS. The His-tagged mutant enzymes were purified as described in
Chapter 5.
Kinetic Measurements. Steady-state kinetic parameters for 5FC were determined
by radiometric assays. The reaction was set up in a total volume of 40 µl in 50 mM
phosphate buffer with 50 mM NaCl at pH 7.5. The final enzyme concentrations were 1
µM, 40 µM, 40 µM, 5 µM, 5 µM, 5 µM and 0.1 µM for the wild-type, N51A, D155A,
E154G, E154D, R73A, and D151A enzymes, respectively. 5FC concentrations were 0.08
- 0.68 mM for the wild-type enzyme, 2.0 - 24 mM for N51A, E154G, E154D and R73A
enzyme, 2.0 -67 mM for D155A enzyme and 0.05 – 1.0 mM for D151A enzyme. A trace
3

amount of [6- H]-5FC was used to quantify the conversion of 5FC to 5FU. The reactions
were initiated at 20 °C by the addition of the enzymes. The reaction times were 10 s, 80
min, 300 min, 60 s, 60 s, 60 s, and 30 s for the wild type, N51A, D155A, E154G, E154D,
R73A and D151A enzymes, respectively. The reactions were quenched and 5FC and 5FU

179

were separated by TLC as previously described (7). Kinetic parameters were obtained by
nonlinear least squares fitting of the data to the standard Michaelis-Menten equation.
Dissociation Constant of 5FPy. The dissociation constant of 5FPy with yCD was
obtained by measuring the change in fluorescence intensity upon binding, using a
FluoroMax-2. A series of 5FPy solutions at different concentrations were made by
dissolving in 50 mM phosphate buffer, pH 7.5. yCD solution in the same buffer was
added into these solutions and incubated at room temperature for 10 min in order to reach
equilibrium. Fluorescence intensities were measured at an emission wavelength of 386
nm with a slit width of 4 nm. The excitation wavelength and slit width were 320 nm and
4 nm, respectively. A set of controls was obtained in the same manner by adding buffer
instead of protein solution. The Kd value was obtained by nonlinear least-squares fitting.
The same protocol was applied to the binding of 5FPy with the N51A and D155A mutant
enzymes, however, no fluorescence change was observed.
19

F NMR Spectroscopy. NMR samples were prepared by dissolving the wild

type yCD and the mutant enzymes in 100 mM potassium phosphate buffer (pH 7.5) made
in D2O. 5FPy was added to the protein solutions and incubated at room temperature for
10 min to reach equilibrium.

19

F NMR experiments were performed at 20 °C on a Varian

Inova 600 MHz NMR spectrometer. The carrier frequency was 564.342 Hz. The NMR
spectra were acquired with a spectral width of 10,000 Hz, 20,000 data points, exponential
multiplication (10 Hz) and 1,024 transients for each spectrum. Relaxation delay was set
as 2s. The

19

F chemical shifts were referenced to CF3C6H5 (chemical shift of -63.73

ppm).

180

Molecular Dynamics Simulations. Molecular dynamics (MD) simulations were
carried out on the yCD-Py complex, N51A yCD-Py complex, D155A yCD-Py complex,
yCD cytosine complex, N51A yCD cytosine complex, and D155A yCD cytosine
complex. The starting structure was taken from the 1.14 Å resolution crystal structure
(PDB code: 1p6o) (5). The mutant enzymes were made by replacing Asn51 or Asp155
with Ala in InsightII. The ligand Py or cytosine was docked in by superimposing with
DHP in the active site.
The procedure for the conventional MD simulation was described in Chapter 4.
For each complex, 8 ns MD simulation trajectory was obtained and analyzed in the
molecular mechanics Poisson Boltzmann surface area (MM-PBSA) calculations.
MM-PBSA Computations. The molecular mechanics Poisson Boltzmann surface
area (MM-PBSA) technique (8, 9) was used to estimate the free energy changes of initial
binding of cytosine or Py with the wild type, N51A, and D155A yCD enzymes. Since
there are no chemical steps involved after the initial binding of the ligand in the MD
simulation, the calculated binding free energy only reflects the initial binding of the
ligand. According to the binding process

!Gbinding
!!!!!!!!!! [EL] #!!!!!!!!!!! [E*L* ]
"
!!!!!!!!!!!!
"
[E] + [L] #!!!!!!!!!
Initial binding !
Chemical reaction !
the initial binding free energy

!Gbinding = G(EL) " G(E) " G(L)

(6-1)

where G(EL), G(E) and G(L) are the free energies for the initial complex, the
receptor and the ligand, respectively. The single MD trajectory protocol was utilized, i.e.
the snapshots of complex, the receptor and the ligand were extracted from an MD
trajectory of the complex. Hence, the binding free energy was estimated according to

181

!Gbinding = !EMM + !Gsol " T !SMM

(6-2)

where !EMM is the difference of the molecular mechanical energies, including
electrostatic and van der Waals interaction energies, since the internal energies are
canceled out. !Gsol was estimated by summing up the contributions from the electrostatic
solvation energy !Gele and the nonpolar solvation energy !Gnp. The electrostatic
contribution was calculated by solving the Poisson-Boltzmann equations (10) with the
grid spacing set at 0.5 Å and the internal and external dielectric constant set at 1.0 and
80.0 respectively. The nonpolar contribution (11) was calculated according to
!Gnp = " SABA + b

(6-3)

where SABA is the solvent accessible surface area and " was set at 0.00542
2

kcal/Å and b was set at 0.92 kcal/mol. The entropy term was not considered in the
calculation because it is computationally time-consuming. So, strictly the calculated free
energy is not the true binding free energy. But in this study, because we are comparing
the binding free energy between very similar systems, the entropy term should be close to
!G and cancels out each other when !!G is calculated.
QM/MM MD Simulations. Self-consistent charge density functional tightbinding (SCC-DFTB) method is a semi-empirical method developed by Elstner et al (12,
13), which has been applied on a number of enzymatic systems (14-18). In this study,
QM/MM MD simulations were performed for yCD using the SCC-DFTB method
implemented in the amber molecular dynamics package (15). The molecular systems for
the yCD water/cytosine complex, N51A water/cytosine complex and D155A
water/cytosine complex were built as described in Chapter 4. The active site of subunit A

182

was modeled in the QM region, including the side chains of His62, Glu64, Cys91 and
Cys94 and the catalytic zinc atom, and the ligands for the respective complexes. The rest
of the protein was modeled in the MM region. SCC-DFTB and PARM99SB force field
was used for the QM and MM region, respectively. SHAKE was turned on for all bonds
except those in the QM region to allow proton transfer.
The molecular systems were minimized and heated up to 300 K using the same
procedure in the conventional MD, expect that the residues in the QM region were
calculated by the SCC-DFTB method. The structures heated up to 300 K were used as the
starting structures for the equilibration and production MD simulations. As an alternative,
representative structures from the conventional production MD simulations were tested as
the starting structures. The results were similar.
During the equilibration and production MD simulation, temperature was
-1

regulated by Langevin dynamics with ! = 1.0 ps . Tight convergence on the density
matrix in the SCF was used. The electrostatic cutoff was set at 12.0 Å for QM/MM
electrostatic interactions. Long-range QM-QM and QM-MM electrostatic interactions
were calculated by Ewald sum (17, 19) and a QM compatible PME approach (14). The
time step was 1 fs. MD simulations were carried out using the parallel version Sander
module in Amber 10. All of the MD results were analyzed with the Ptraj module of
Amber 10.
The free energy profile along a reaction coordinate was obtained from the
potential of mean force (PMF) by adopting an umbrella sampling technique (20). The
proper amount of windows was used to sample enough space along the reaction
coordinate. The force constant was set in the range of 100 to 200 kcal/mol. The data from
the simulations were analyzed by the weighted histogram analysis method (WHAM) (21,
183

22) to compute the PMF. Data from a 200 ps trajectory were collected after 200 ps of
equilibration for each window.

RESULTS
Steady State Kinetics. By site-directed mutagenesis, plasmids encoding the
N51A, D155A, E154G, E154D, R73A and D151A proteins were constructed and the
proteins were expressed and purified. Steady-state kinetic parameters of the conversion of
5FC to 5FU by the wild type yCD and the mutant enzymes were measured by initial
velocity analysis as described in the methodology section.
Asn51 forms a hydrogen bond with O2 of DHP in the yCD DHP complex,
indicating it may contribute to substrate binding. The results show that substituting
Asn51 with Ala greatly reduces the catalytic efficiency. As shown in Table 6-1, the
7

kcat/Km value of N51A yCD decreases about 1 ! 10 -fold, relative to that of wild-type
-1 -1

yCD, from 106 mM s

-5

-1 -1

to 1.1 ! 10 mM s . The reduced activity is not only due to

weakened affinity to the substrate 5FC (Km = 4.1 mM), but it is also a result of the low
-3 -1

turnover rate (kcat = 4.6 ! 10 s ).
Similar behavior was observed for the D155 enzyme. Asp155 forms a hydrogen
bond with N1 of DHP in the yCD DHP complex. Replacing Asp155 with Ala reduces
7

kcat/Km about 2 !10 -fold. The Km of 5FC increases to 17 mM and the kcat decreases to
-3 -1

9.8 ! 10

s . The results suggest that both Asn51 and Asp155 are important for ligand

binding. They also participate in catalysis, probably by stabilizing the transition state.

184

E154G, E154D, and R73A proteins were constructed to explore the function of
the C-terminal !-helix. The C-terminal !-helix is suggested to serve as a “gate”
controlling access to the active site. Glu154 in this region forms a strong salt bridge with
Arg73 from the other subunit. Removal of the interaction may increase the flexibility of
the C-terminal !-helix and hence increase the product release rate. Three mutant
enzymes, with E154G, E154D, and R73A substitutions, were made and steady state
kinetics experiments were performed. The results show that the values of Km of all three
enzymes increase, relative to that of the wild type enzyme. The value of Km for E154G,
E154D and R73A proteins is 4.2, 2.4 and 3.2 mM, respectively. This finding could be an
indication of a more opened active site; however, the alterations may reduce the stability
of the transition state. The kcat of these enzymes decreased several fold to 8.7, 2.6, and
-1

1.7 s for E154G, E154D and R73A proteins, respectively.
Asp151 is also a residue in the C-terminal !-helix. The negatively charged side
chain does not interact with the inhibitor in the yCD DHP complex. Rather, it points out
of the active site and is likely surrounded by solvent molecules. The activity of the
D151A enzyme is close to that of the wild type enzyme, with a similar Km. The kcat only
-1

decreases 2-fold to 6.3 s .
Dissociation Constants of 5FPy. The dissociation constants of 5FPy with the
wild type and the N51A and D155A enzymes were measured. It turned out that 5FPy
bound yCD very tightly with an apparent Kd value of 0.65 µM (Table 5-2 in Chapter 5 or
Table 6-2). However, the change in fluorescence signal upon ligand binding was not

185

observed in the binding assays for either mutant enzyme, suggesting that they have low
binding affinity with 5FPy.
The dissociation constants of 5FPy with the mutant enzymes were estimated by
comparing the peak area of 5FPy in the bound form and in the free form by
spectroscopy.

19

19

F NMR

F-NMR spectra were acquired for the wild-type yCD (1.5 mM) in the

presence of 3.0 mM 5FPy and the N51A and D155A enzymes (1.5 mM) in the presence
of 20.0 mM 5FPy, as shown in Figure 6-1. All enzyme spectra had a peak as the free
form inhibitor at around -160.6 ppm, because the 5FPy concentration was much higher
than that of the enzyme. Both mutant enzyme spectra also had a broad peak
corresponding to the bound inhibitor. The peaks were well separated because the
exchange rate is slow on the NMR time scale. However, the Kd of 5FPy with these two
mutant enzymes is harder to obtain because the intensity of the peak is so low, even when
more 5FPy was added. These results indicate that both residues are critical for the
binding of the inhibitor. This result is consistent with the steady state data in which Km
for both enzymes are increased due to the low binding affinity with 5FC.
Binding Free Energy Analysis by MM-PBSA/GBSA. To understand how
Asn51 and Asp155 contribute to ligand binding, MM-PBSA/GBSA was performed to
estimate the binding free energy of Py. Contributions from other conserved residues in
the active site were analyzed as well.
An 8 ns production MD trajectory for the yCD Py complex simulation was
collected. The stability of the system was analyzed by plotting the root-mean-square
displacements (RMSDs) of the backbone C! atoms along the reaction coordinate,

186

relative to the starting structure (Figure 6-2). RMSDs show that the system was stable
and finally reached the equilibrium state after 1.5 ns.
The interactions between Py and the surrounding residues include a set of
!2

hydrogen bonds and hydrophobic interactions. O2 of Py forms a hydrogen bond with N

of Asn51 and with the backbone amide of Gly63 with distances of 2.88 and 3.14 Å,
!1

respectively. N1 of Py forms a hydrogen bond with O
!2

2.73 Å. O
"2

Å. N

!2

of Asp155 forms a hydrogen bond with N

of Asp155 with a distance of

of Asn51 with a distance of 2.88

of His62 forms a hydrogen bond with the backbone carbonyl oxygen of Asp155

with a distance of 2.89 Å. One new interaction is between the zinc-coordinated water
Zn

O

and N3 of Py with a distance of 3.09 Å. Glu64 is in the deprotonated form and forms

a strong hydrogen bond with the other hydrogen from the zinc-coordinated water with
2.60 Å. Residues such as Ile33, His62, Leu88, Phe114, Trp152 and Ile156 contribute to
the interaction by hydrophobic effects.
100 snapshots in the MD trajectory between 5 ns and 8 ns, at 30 ps intervals, were
extracted and analyzed. The data in Table 6-3 show the energy terms accounting for the
binding free energy by the MM-PBSA technique. !EMM represents the change of
molecular mechanics interaction energy upon inhibitor binding, including electrostatic
energies and van der Waals interaction energies. The negative value (-51.6 kcal/mol)
suggests the binding is favored by the solutes in the gas phase. !Gsol represents the
change of solvation energies, including polar and nonpolar energy. The positive value
(33.2 kcal/mol) indicates that the binding is not favored due to the presence of the

187

solvent. The polar solvation energy is dominant compared with the nonpolar solvation
energy. The overall binding free energy suggests the binding of Py with yCD is favored
by 18.4 kcal/mol. The binding free energy was also calculated by MM-GBSA. The same
conclusion can be drawn, though the value is more negative. Entropy was not considered
because it is computationally time consuming and the results normally have large
deviation.
Free Energy Decomposition. Per-residue free energy decomposition was
performed to evaluate the contributions from individual residues. The energy
contributions were plotted in Figure 6-3. A number of residues, including Val31, Ile33,
Asn51, His62, Gly63, Leu88, Trp152, Asp155 and Ile156, contribute to the binding of
Py.
Both Asn51 and Asp155 make a significant contribution (Table 6-4) to the total
free energy. The individual energies are 2.7 and 0.7 kcal/mol for Asn51 and Asp155,
respectively. The contribution is mainly from the gas phase energy of the side chain. The
solvation effects are quite different for these two residues. The Asp155 side chain
contributes to the solvation energy by 5.3 kcal/mol and hence the overall contribution
from Asp155 decreases to 0.7 kcal/mol. Similarly the side chain of Glu64 contributes to
solvation energy (1.47 kcal/mol). The presence of Glu64 is an unfavorable force for the
inhibitor binding. Interestingly, the E64A enzyme does not increase the affinity to Py
(see Chapter 5). Instead, the binding affinity is greatly reduced. Per-residue energy
reflects the contribution of a given residue in the wild type protein complex with Py.
When Glu64 is changed to Ala, the behavior of the mutant enzyme was quite different.
MM-PBSA on the E64A-Py complex showed that the enzyme indeed loses affinity to Py
(data not shown).

188

Gly63 forms a hydrogen bond with O2 of Py through the main chain amide, by
which Gly63 contributes to the gas phase energy (-2.5 kcal/mol). Similarly, since the
backbone NH is exposed to solvents in the active site, the solvation energy cost is
relatively high too. The overall effects favor the inhibitor binding.
His62 interacts with Py on one side by !-stacking interaction and residues Val31,
Ile33, Thr86, Leu88, Ser89, Pro90, Cys91, Trp152 and Ile156 interact on the other side
by hydrophobic effects. His62 is far more important than the other nonpolar residues
because of the strong !-stacking interaction. This residue contributes to the gas phase
energy with 4.5 kcal/mol and a total binding free energy of 3.5 kcal/mol. The individual
contribution from the other hydrophobic residues is small. However, the collective effects
are not negligible, contributing roughly 3.9 kcal/mol to the total binding energy.
Binding Free Energy with Mutant Enzymes. Kinetics and binding assays
suggest that the substitutions greatly reduce the binding affinity of 5FC/5FPy with the
mutant enzymes. Energy decomposition analysis shows that Asn51 and Asp155
contribute to the total binding free energy with 2.7 kcal/mol and 0.7 kcal/mol,
respectively. To get a better estimation of the free energy, MM-PBSA analysis was
performed on the mutant enzyme Py complexes.
An 8 ns production MD trajectory was collected for each mutant enzyme using
the same procedure as for the wild type yCD Py complex. Similarly, stabilities of the
systems were analyzed by plotting the RMSDs for the backbone C" atoms in the MD
trajectory of the production MD simulation, relative to the starting structure (Figure 6-2).
RMSDs show that the systems were not as stable as the wild type complex but eventually
reached equilibrium after 3 ns. For the MM-PBSA analysis, 100 configurations from the
MD trajectory starting from 5ns to 8 ns, at 30 ps intervals, were extracted and analyzed.
189

Similar results were obtained when different time windows were used between 3 to 8 ns
or more snapshots were analyzed (data not shown).
The results are summarized in Table 6-2. Both GB and PB methods were used to
calculate the free energy difference !Gbinding. The value calculated from PBSA is a little
larger than that from GBSA, but in both cases binding of the inhibitor is favored. The
binding free energy of Py with the wild type yCD is calculated to be -18.4 kcal/mol.
When Asp51 is replaced by Ala, !G increases to -7.8 kcal/mol, which means that the
N51A enzyme Py complex is 10.6 kcal/mol less stable than the wild type yCD Py
complex. When Asp155 is changed to Ala, !G increases to -10.8 kcal/mol, which also
means the mutant enzyme in complex with Py is less stable than that of the wild type
yCD enzyme. These results explain the behavior of the enzymes in the binding assays,
which revealed a great increase of Kd for both mutant enzymes.
Since Py may be present in the active site as either an unhydrated form or a
hydrated form, the initial binding with Py (with no chemical transformation) alone may
not account for the overall binding affinity with these two enzymes. To get a better
understanding,

19

F NMR was performed, as presented in the previous section. The results

suggest that there is no hydrated form inhibitor formed in the complex with either mutant
enzyme.
Formation of DHP by the Mutant Enzymes. To explain why the hydrated
inhibitor is not preferred by the N51A and D155A enzymes, QM/MM MD simulations
and free energy calculations were performed to obtain the energy profiles for the
formation of DHP from Py. We proposed the formation of DHP through three steps.
First, a proton (HW) from the zinc-coordinated water transfers to the carboxyl group
190

!1

!1

(O ) of Glu64. Second, N3 is protonated by O

of Glu64. Third, the zinc-coordinated

hydroxide attacks C4 and DHP is formed.
A 2D umbrella sampling technique was performed for the first two proton transfer
steps. The PMF was obtained along two reaction coordinates, one is the distance between
!1

HW and O

and the other is the distance between HW and N3. The energy profiles for

the N51A and D155A enzymes, in Figure 6-4, are different from that for the wild type
yCD. In complex with yCD, Py is stabilized by a hydrogen bond network, including
those with Asn51 and Asp155. It costs about 7.7 kcal/mol for the proton transfer from the
!1

zinc-coordinated water to O!1 of Glu64. After HW is transformed to O

of Glu64, it still

forms a hydrogen bond with N3. The conformation sits in a very shallow energy well. It
!1

costs another 4.9 kcal/mol for the proton (HW) transfer from O

of Glu64 to N3 of Py.

The transition state is found when the distances between HW and N3 and between HW
!1

and O

are 1.15 and 1.47 Å, respectively. O4-C4 bond is already formed in a concerted
!1

manner with the proton transfer from O

of Glu64 to N3 of Py. The yCD DHP complex

is found to be 1.8 kcal/mol more stable than the yCD Py complex.
The behaviors of the N51A and D155A enzymes in complex with Py are different
from that of the wild type yCD. For the N51A enzyme, Py can still form hydrogen bonds
"1

with O

of Asp155 at a distance of 2.79 Å and with the backbone amide group of Gly63

at a distance of 3.01 Å. However, these interactions cannot hold Py in the same position
as in the complex with yCD. The pyrimidine ring of Py rotates around 30° to form a

191

hydrogen bond with the zinc-coordinated water with a better configuration. Hence, extra
!1

energy is needed to transfer a proton from the zinc-coordinated water to O

of Glu64.
!1

The overall energy transition state was found when the distances between HW and O

and between HW and N3 are 1.50 and 1.12 Å respectively, with an activation energy of
17.1 kcal/mol. For Py in complex with the D155A enzyme, its pyrimidine ring shifts
further away from where it is in the complex with yCD, although O2 still forms a
"2

hydrogen bond with N

of Asn51 with a distance of 2.88 Å. However, the hydrogen

bond between the backbone amide of Gly63 and O2 is completely disrupted. N1 forms a
new hydrogen bond with the backbone carbonyl oxygen of Asp155 with a distance of
2.89 Å. This hydrogen bond completely disrupts the one between the backbone carbonyl
!2

oxygen of Asp155 and N

of His62, which is presented in the yCD Py complex to

maintain the scaffold of the active site. The energy profile shows that the overall energy
barrier is 26.1 kcal/mol, which is much larger than what is needed in the same process by
yCD.
Transition state structures for three reactions catalyzed by the wild type, N51A
and D155A yCDs were extracted and superpositioned, as shown in Figure 6-5. Although
Py orientations in the initial structures are quite different, the transition state structures
are very similar. Even the imidazole ring of His62 rotates back to the position, where it
stays in the yCD complex. However, without the presence of Asn51 or Asp155 in these
two mutant enzymes, the transition state is less stable than the wild type enzyme.
The N51A and D155A enzymes have low affinity for the inhibitor Py and the
substrate 5FC, as revealed by steady state kinetics studies and binding assays. Their

192

contributions were explored by free energy decomposition, MM-PBSA calculations and
QM/MM MD simulations from an energetic point of view. Py binds with the mutant
enzymes in an alternative orientation compared to the complex with wild type yCD and
the activation energy is elevated.

DISCUSSION
To understand the functions of conserved residues in the active site, we examined
the N51A, D155A, E64A, G154G, E154D, R73A and D151A variant proteins. In Chapter
5, we already discussed the function of Glu64. The results show that Glu64 indeed
participates in substrate binding and transition state stabilization. The other mutant
proteins were successfully expressed and purified. These mutants were initially designed
for different purposes.
N51A and D155A. Asn51 and Asp155 form direct hydrogen bonds with the
ligand. These hydrogen bonds, with the hydrogen bond between the main chain amide
group of Gly63 and O2, are important for substrate binding, and probably for transition
state stabilization. The steady state kinetics data show that the Km of the N51A and
D155A enzymes for 5FC is greatly increased, indicating their function in substrate
binding. However, the variant proteins have much lowered kcat, which means that these
residues are not merely used for substrate binding. Their presence may affect the catalysis
by stabilizing the transition state.
These residues function in stabilization of the transition state as implied by
NMR studies.

19

19

F

F NMR spectra were acquired to detect what form of the inhibitor was

bound in the active site of the wild type yCD and the mutant enzymes. As discussed in
193

Chapter 5, the inhibitor in the active site of the wild type enzyme is the hydrated form
5FDHP. The results show that the inhibitor is the unhydrated form in complex with the
N51A and D155A enzymes. The observed Kd is most likely the true dissociation constant
of the unhydrated form of the inhibitor.
A reasonable explanation for why the hydrated form inhibitor cannot be formed
by the mutant enzymes was obtained by QM/MM MD simulations and free energy
calculations. For the unhydrated inhibitor Py, the pyrimidine ring adopts a different
orientation, rotating about 30 degrees along the axis of N1-C4, when it binds with the
N51A enzyme. Accordingly, the !-stacking interaction with His62 is attenuated. The
energy needed for the formation of the hydrated inhibitor in the variant protein is 3.4
kcal/mol greater than that by the wild type enzyme. In complex with the D155A mutant
enzyme, Py shifts further to a new position in which N1 forms a new hydrogen bond with
the main chain carbonyl oxygen of Asp155. Hence, extra energy is required to shift Py
back to a suitable position for the proton transfer from Glu64 to N3. The overall
activation energy is 26 kcal/mol and the hydrated form is 15 kcal/mol greater than the
unhydrated form. Thus, the inhibitor most likely stays in the unhydrated form, when
binding with both mutant enzymes.
Py is a substrate mimic and shares a similar interaction pattern with the
surrounding residues as for cytosine, except that Glu64 forms a hydrogen bond with N4
of cytosine. It is no surprise that substituting Asn51 and Asp155 with Ala results in a
lower binding affinity for cytosine. Kd for Py is much lower than Km for cytosine
because Py is converted to DHP in the active site of yCD. For the initial binding, cytosine
should be a better ligand due to the additional interaction between Glu64 and the amide
group of cytosine. This result was supported by MM-PBSA calculations, which show
194

that the binding free energy of cytosine with yCD is 7.4 kcal/mol lower than that of Py.
Similarly, cytosine is a better ligand for the N51A and D155A mutant enzymes. The
reaction rate of the deamination is so slow that Km is close to the dissociation constant of
cytosine with yCD. The same results should hold for the 5FC and 5FPy pair, because the
steady-state kinetics and binding assays indicate that Km for 5FC is lower than Kd for
5FPy for the mutant enzymes.
E154G, E154D and R73A. Besides the N51A and D155A substitutions, other
mutations in the C-terminal !-helix were introduced. Glu154 forms a strong salt bridge
with Arg73 from the other subunit, which rigidifies the C-terminal !-helix. Previous
studies show that product release is the rate-limiting step for the deamination of 5FC.
E154G, E154D and R73A were designed to increase the flexibility of the C-terminal !helix and hopefully to increase the rate of product release. However, the results show that
the turnover rates (kcat) were further reduced. The reduced rates are likely due to rate
changes of the chemical steps because the flexibility of the C-terminal !-helix is indeed
increased, as suggested by MD simulations. Py binds with these mutant enzymes in the
hydrated form DHP, as shown by

19

F NMR spectra (data not shown), which suggests that

the transition state is still stabilized by the mutant enzymes. However, the transition state
may not bind as tightly as in the wild type enzyme.
Besides the hydrogen bonds that stabilize DHP, hydrophobic interactions are
important too. A cluster of conserved residues Ile33, Leu88, Phe114 and Trp152 and
Ile156 interact with DHP through hydrophobic interactions. Phe114, Ile156 and Trp152,
separate the active site from being accessed from the solvent. The functions of these

195

residues need to be investigated to get a clear picture for how the enzyme interacts with
its ligand and to understand the catalytic mechanism of enzymes in the CDA family of
purine/pyrimidine deaminase.

CONCLUSIONS
Functions of conserved residues in the active site were investigated by sitedirectly mutagenesis, kinetics, NMR, and computational methods. A series of plasmids
encoding mutant proteins were constructed, including the substitutions N51A, D155A,
G154G, E154D, R73A, and D151A.
Asn51 and Asp155 form hydrogen bonds with the ligand. Substituting Asn51 or
7

Asp155 with Ala greatly reduces the catalytic efficiency by a factor of 10 . The low
activity is not only from reduced turnover rate, but also from weakened binding affinity.
4

The kcat value of N51A enzyme decreases about 1 ! 10 -fold, relative to that of wild-type
-1

yCD, from 17 s

-3 -1

to 4.6 ! 10 s . The Km value of N51A protein increases about 25-

fold, relative to that of wild-type yCD, from 0.16 mM to 4.1 mM. Similarly, for the
-3 -

D155A enzyme, the Km of 5FC increases to 17 mM and the kcat decreases to 9.8 ! 10 s
1

. The results suggest both Asn51 and Asp155 are important for ligand binding, as well

as catalysis.
Initial binding of ligand to these enzymes was investigated by MM-PBSA and
free energy decomposition. The binding free energies of Py with the mutant enzymes are
much lower than that with the wild type enzyme. Formation of hydrated inhibitor by the
mutant enzymes was investigated by

19

F NMR spectroscopy and QM/MM MD

196

simulation. The hydrated form was not observed in

19

F NMR spectra of the mutant

proteins and the activation energy for the formation of the hydrated inhibitor is elevated,
compared with the wild type enzyme. These results suggest that Asn51 and Asp155
participate in catalysis by stabilizing the transition state.

197

TABLES AND FIGURES
Table 6-1: Kinetic constants of the wild type yCD and mutant enzymes for the activation
of 5-fluorocytosine.
Protein

-1

-1 -1

yCD WT
N51A

Km (mM)
0.16 ± 0.01
4.1 ± 1.0

kcat (s )
17 ± 0.4
-3
(4.6 ± 0.6)!10

kcat /Km (mM s )
106
-5
1.1 ! 10

D155A

17 ± 2.0

E154D
E154G
R73A
D151A

2.4 ± 0.3!
4.2 ±0.7!
3.4 ± 0.7!
0.15 ± 0.03

(9.8 ± 1)!10
2.6 ± 0.2!
8.7 ± 0.9!
1.7 ± 0.2!
6.3 ± 1.5

5.4 !10
1.0 !
2.1 !
0.5 !
45

-3

-6

Table 6-2: Binding constants of 5FPy with the wild type yCD and mutant enzymes.
Protein
yCD WT
N51A
D155A

Kd, 5F-Py2O
0.65 " 0.08 µM
>100 mM
>100 mM

198

Table 6-3. Internal energy and solvation energy components for the interactions between
a
residues and Py .

!Eele !Evdw !EMM !Gpolar !Gnp

!Gsol !Gbinding

!!G

yCD2Py
-35.3 -16.3
-51.6
35.6
-2.4
33.2
-18.4
N51A2Py
-26.7 -16.4
-43.1
37.9
-2.6
35.3
-7.8
10.6
D155A
-2Py
-25.4 -18.0
-43.5
35.2
-2.5
32.7
-10.8
7.6
yCDcyt
-48.1 -18.7
-66.9
43.6
-2.6
41.0
-25.8
N51Acyt
-42.2 -19.2
-61.4
44.0
-2.6
41.4
-20.0
5.8
D155A
-cyt
-41.6 -19.1
-60.7
43.1
-2.6
40.5
-20.2
5.6
a: !EMM = !Eele + !Evdw; !Gsol = !Gpolar + !Gnp; !Gbinding = !EMM + !Gsol.
Table 6-4. Individual energy components (kcal/mol) by MM-GBSA energy
a
decomposition .
SGAS
BGAS
SGBsol BGBsol SGBtot BGBtot TGBtot
Val31
-0.4
0.47
0.12
-0.6
-0.2
-0.1
-0.3
Ile33
-1.3
0.46
0.22
-0.7
-1.1
-0.2
-1.3
Asn51
-3.3
-0.7
0.87
0.48
-2.4
-0.3
-2.7
His62
-2.1
-2.4
0.07
0.89
-2.0
-1.5
-3.5
Gly63
0
-2.5
0
1.29
0
-1.2
-1.2
Glu64
-0.7
0.1
1.47
-0.2
0.79
-0.2
0.59
Thr86
-0.2
-0.2
-0.1
0.09
-0.3
-0.1
-0.3
Leu88
-0.4
0.14
-0.1
-0.2
-0.5
-0.1
-0.6
Ser89
0.19
-1.2
-0.2
1.03
0
-0.2
-0.2
Pro90
0.19
-0.4
-0.2
0.29
-0.1
-0.1
-0.2
Cys91
-1
0
0.6
0
-0.4
0
-0.4
Trp152
-0.5
-0.3
0.17
0.34
-0.3
0.01
-0.3
Asp155
-6.0
-0.4
5.31
0.38
-0.7
-0.01
-0.7
Ile156
0.11
-0.3
-0.4
0.22
-0.2
-0.1
-0.3
Zn
0.7
0
-0.3
0
0.39
0
0.39
-0.3
0
0.5
0
0.23
0
0.23
H2O
a: SGAS and BGAS are the gas phase energy for the sidechain and backbone,
respectively. SGBsol and BGBsol are the solvation energy for the sidechain and
backbone, respectively. SGBtot and BGBtot are the total binding free energy for the
sidechain and backone, respectively. TGBtot is the total binding free energy for the
whole residue.
199

19

Figure 6-1. F-NMR spectra of the mutant enzyme-5FPy complexes and wild type
yCD-5FPy complex. (A) 1.5 mM 5FPy in solution. (B) 5FPy (3.0 mM) in complex with
yCD (1.5 mM). (C) 5FPy (20.0 mM) in complex with N51A (1.5 mM). The insert shows
a small peak at -157 ppm, corresponding to the inhibitor in the hydrated form. (D) 5FPy
(200 mM) in complex with D155A (1.5 mM). The insert shows a small peak at -157 ppm,
corresponding to the inhibitor in the hydrated form.

200

Figure 6-2. RMSDs of the backbone C! atoms relative to the starting structure for the
yCD Py complex during the 8 ns production MD simulations.

Figure 6-3. Ligand-residue interaction spectrum of the yCD Py complex by MM-GBSA
energy decomposition.

201

Figure 6-4. 2D energy profiles for proton transfer from the zinc-coordinated water to N3
of Py. (A) The yCD Py complex. (B) The N51A Py complex. (C) The D155A Py
complex.

202

Figure 6-5. Superposition of representative snapshots for the yCD Py complex, N51A Py
complex and D155A Py complex. (A) Initial structures. (B) Transition state structures.
The residues of the yCD Py complex are colored by atom names; the residues of the
N51A Py complex are colored yellow; and the residues of the D155A Py complex are
colored green.

203

REFERENCES

204

REFERENCES
1. Aghi, M., Hochberg, F., and Breakefield, X. (2000) Prodrug activation enzymes in
cancer gene therapy, J Gene Med 2, 148-164.
2. Greco, O., and Dachs, G. U. (2001) Gene directed enzyme/prodrug therapy of cancer:
Historical appraisal and future prospectives, J Cell Physiol 187, 22-36.
3. Erbs, P., Regulier, E., Kintz, J., Leroy, P., Poitevin, Y., Exinger, F., Jund, R., and
Mehtali, M. (2000) In vivo cancer gene therapy by adenovirus-mediated transfer of a
bifunctional yeast cytosine deaminase/uracil phosphoribosyltransferase fusion gene,
Cancer Res 60, 3813-3822.
4. Ko, T., Lin, J., Hu, C., Hsu, Y., Wang, A., and Liaw, S. (2003) Crystal structure of
yeast cytosine deaminase - Insights into enzyme mechanism and evolution, J Biol
Chem 278, 19111-19117.
5. Ireton, G., Black, M., and Stoddard, B. (2003) The 1.14 angstrom crystal structure of
yeast cytosine deaminase: Evolution of nucleotide salvage enzymes and implications
for genetic chemotherapy, Structure 11, 961-972.
6. Carlow, D. C., Short, S. A., and Wolfenden, R. (1996) Role of glutamate-104 in
generating a transition state analogue inhibitor at the active site of cytidine
deaminase, Biochemistry 35, 948-954.
7. Yao, L., Li, Y., Wu, Y., Liu, A., and Yan, H. (2005) Product release is rate-limiting
in the activation of the prodrug 5-fluorocytosine by yeast cytosine deaminase,
Biochemistry 44, 5940-5947.
8. Kollman, P. A., Massova, I., Reyes, C., Kuhn, B., Huo, S., Chong, L., Lee, M., Lee,
T., Duan, Y., Wang, W., Donini, O., Cieplak, P., Srinivasan, J., Case, D. A., and
Cheatham, T. E. (2000) Calculating structures and free energies of complex
molecules:! Combining molecular mechanics and continuum models, Acc Chem Res
33, 889-897.
9. Srinivasan, J., Cheatham, T. E., Cieplak, P., Kollman, P. A., and Case, D. A. (1998)
Continuum solvent studies of the stability of DNA, RNA, and
phosphoramidate!DNA helices, J Am Chem Soc 120, 9401-9409.
10. Carlow, D. C., Smith, A. A., Yang, C. C., Short, S. A., and Wolfenden, R. (1995)
Major contribution of a carboxymethyl group to transition-state stabilization by
cytidine deaminase: Mutation and rescue, Biochemistry 34, 4220-4224.
11. Gilson, M. K., Sharp, K. A., and Honig, B. H. (1988) Calculating the electrostatic
potential of molecules in solution: Method and error assessment, J Comput Chem 9,
327-335.
12. Elstner, M., Porezag, D., Jungnickel, G., Elsner, J., Haugk, M., Frauenheim, T.,
Suhai, S., and Seifert, G. (1998) Self-consistent-charge density-functional tight205

binding method for simulations of complex materials properties, Phys Rev B 58,
7260.
13. Elstner, M. (2005) The SCC-DFTB method and its application to biological systems,
Theor Chem Acc 116, 316-325.
14. Walker, R. C., Crowley, M. F., and Case, D. A. (2008) The implementation of a fast
and accurate QM/MM potential method in Amber, J Comput Chem 29, 1019-1031.
15. Seabra, G. D. M., Walker, R. C., Elstner, M., Case, D. A., and Roitberg, A. E. (2007)
Implementation of the SCC-DFTB method for hybrid QM/MM simulations within
the Amber molecular dynamics Package, J Phys Chem A 111, 5655-5664.
16. Riccardi, D., and Cui, Q. (2007) pKa analysis for the zinc-bound water in human
carbonic anhydrase II: Benchmark for "multiscale" QM/MM simulations and
mechanistic implications, J Phys Chem A 111, 5703-5711.
17. Nam, K., Gao, J., and York, D. M. (2005) An efficient linear-scaling Ewald method
for long-range electrostatic interactions in combined QM/MM calculations, J Chem
Theory Comput 1, 2-13.
18. Cui, Q., Elstner, M., Kaxiras, E., Frauenheim, T., and Karplus, M. (2001) A QM/MM
implementation of the self-consistent charge density functional tight binding (SCCDFTB) method, J Phys Chem B 105, 569-585.
19. Darden, T., York, D., and Pedersen, L. (1993) Particle mesh Ewald: An N!log(N)
method for Ewald sums in large systems, J Chem Phys. 98, 10089-10092.
20. Torrie, G. M., and Valleau, J. P. (1977) Nonphysical sampling distributions in Monte
Carlo free-energy estimation: Umbrella sampling, J Comput Phys 23, 187-199.
21. Roux, B. (1995) The calculation of the potential of mean force using computer
simulations, Comput Phys Commun 91, 275-282.
22. Grossfield, A. WHAM: the weighted histogram analysis method, version 2.0,
http://membrane.urmc.rochester.edu/content/wham

206

CHAPTER 7

SUMMARY AND FUTURE WORK

SUMMARY
My efforts have been devoted to understanding the catalytic mechanism of yeast
cytosine deaminase by both experimental and computational methods.
I tried to explore the enzymatic mechanism by probing the transition state by
kinetic isotope effects and quantum mechanics calculations.

13

13

C

C KIEs were measured at

natural abundance by NMR spectroscopy. The results indicate that the reaction proceeds
by a stepwise AN+DN mechanism, with a transition state in the C4-N4 bond cleavage
step. To obtain more detailed information about the reaction pathway and facilitate
protein design of variant enzymes with increased substrate specificity towards 5FC, a
two-layered ONIOM(QM:MM) method was employed to study the deamination of 5FC
and the reaction pathway was compared with that of cytosine. The results show that the
reaction pathway for the deamination of 5FC is similar to that of cytosine. Energetically
favorable tetrahedral intermediates were found for 5FC, which may account for the
relatively low Km for 5FC. However, configurational sampling of proper geometries
during the Zn-O4 bond breakage was not promising. Hence, the pathway was reinvestigated by QM/MM MD simulations and free energy calculations. The results show

207

that the C4-N4 bond cleavage is indeed the rate-limiting step with an activation energy of
19.7 kcal/mol. The activation energy for the Zn-O4 bond breakage is only 8.1 kcal/mol.
Functional roles of conserved residues in the active site were investigated by sitedirected mutagenesis, biochemical, NMR, and computational methods. Glu64 is believed
to serve as a proton shuttle during multiple steps. Steady-state kinetic studies show that
changing Glu64 to alanine causes a dramatic decrease in kcat and a dramatic increase in
Km. The substitution also causes a dramatic decrease in the affinity of the reaction
intermediate analogue 5FPy.

19

F NMR experiments show that inability to form a

hydrated inhibitor is one of the causes for the low binding affinity. ONIOM calculations
also demonstrate that, without the help of Glu64, both proton transfers before and after
the formation of the tetrahedral intermediate become partially rate-limiting steps. The
results indicate that Glu64 is not only directly involved in catalysis, but also important for
substrate binding.
The functional roles of the active site residues, particularly Asn51 and Asp155,
were investigated as well. Steady-state kinetic studies show that both the N51A and
D155A proteins exhibit a dramatic decrease in kcat and a dramatic increase in Km. The
substitutions also cause a decrease in the binding affinity of 5FPy by a factor of >100.
The lowered binding affinity is partially due to weak initial binding, suggested by
MM-PBSA and binding free energy calculations, which show that changing Asn51 and
Asp155 to alanine reduces the binding affinity of Py by 10.6 and 7.6 kcal/mol,
respectively. The binding of 5FPy to the N51A and D155A enzymes causes a downfield
shift in the

19

F NMR signal, indicating that the inhibitor is in the unhydrated form.

208

QM/MM MD simulations show that, with the N51A and D155A enzymes, the hydrated
form is no longer favored and the activation energy is elevated. The results suggest that
similar to Glu64, Asp51, and Glu155 are important not only for the initial binding of the
ligand, but also for the stabilization of the transition state.

FUTURE WORK
Starting from what we have learned, other researchers will be able to continue to
work on the catalytic mechanism of yCD. Meanwhile, researchers could try to construct
thermostable variants of yCD, which are highly desirable in the gene directed enzyme
prodrug therapy (GEDPT).
Deuterium and

15

N KIEs.

13

C KIE studies provide a relatively rapid and

convenient probe to detect the transition state of yCD. While it alone may not precisely
describe the conformation of the transition state, the use of ONIOM calculations can
provide additional information. Deuterium and

15

N can be employed to get a better

description of the transition state. Deuterium KIEs can be measured by proton inventory
analysis of reactions in buffers with different D2O content.

15

N KIEs can be measured for

substrate at natural abundance by isotope ratio mass spectrometric analysis. More details
in the transition state are expected. For example, the protonation state of N3 can be
determined from the
15

N and

13

15

N KIE of N3 and deuterium KIE of H3. By combining deuterium,

C KIEs, one can also determine if the proton transfer from Glu64 to N4 and

the C4-N4 bond cleavage are concerted or stepwise.

209

Function of other residues in the active site. The conserved Glu64 and the zinc
apparatus define the enzyme as a deaminase. Asn51 and Asp151 provide specific
interactions with the substrate and the transition state. However, these residues may not
offer enough driving forces for catalysis. The other residues surrounding the substrate,
such as Ile33, Phe114, Leu88, and Trp152, etc, also contribute through hydrophobic
interactions. For example, changing Trp152 to leucine, which has a shorter side chain,
reduces the catalytic efficiency (kcat/Km) by 2-fold (data not shown). How exactly these
residues contribute to catalysis is worth further study.
Protein engineering. Besides understanding the catalytic mechanism, researchers
are interested in designing thermostable variants of yCD with high catalytic efficiency.
Computational approaches and random mutagenesis approaches have been developed to
design stable protein with point mutation in the enzyme core region or the active site. In
our laboratory, we identified such variants showing better catalytic efficiency and
thermostability in in vitro enzymatic assays. New variants can be designed by combining
two or more mutations into a single construct. Cell assays or animal models need to be
designed to evaluate their performance.

210