r in.

...,., .33....
. 3.... 4
2. .4 .
:. .3.
{3.13:

a .

. “ ....—.. 2 ES 1 2.; .. r...
.

u‘ .1
r’

I: '

4mm

,- w

 

nu: 3.22.3.1).

 

llnlmunll‘l‘lfilllilllillnliliumlml

3 1293 01789 3144

 

LIBRARY

Michigan State
University

 

 

 

This is to certify that the

thesis entitled
1H NMR SPECTROSCOPIC STUDIES OF DINUCLEAR TRANSITION
METAL CARBOXYLATE ADDUCTS OF DNA OLIGONUCLEOTIDES

presented by

Elizabeth Ursula Lozada Carrasco

has been accepted towards fulfillment
of the requirements for

Masters Chemistry

 

 

degree in

fi/fl %4fi24 £42,“ /

Major professor

Date 7/07 j/QJ

0-7639 MS U is an Aflirmative Action/Equal Opportunity Institution

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1M clam-9659.14

1H NMR SPECTROSCOPIC STUDIES OF DINUCLEAR TRANSITION
METAL CARBOXYLATE ADDUCTS OF DNA OLIGONUCLEOTIDES

By

Elizabeth Ursula Lozada Carrasco

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Chemistry

1998

ABSTRACT

1H NMR SPECTROSCOPIC STUDIES OF DINUCLEAR TRANSITION
METAL CARBOXYLATE ADDUCTS OF DNA OLIGONUCLEOTIDES

By

Elizabeth Ursula Lozada Carrasco

Cisplatin, cis-Pt(NH3)2Clz, is in clinical use as an effective treatment
for cancer. Its ability to bind covalently at the N7 position of guanine bases
and form intrastrand cross-links with DNA is thought to be responsible for
its biological activity. Similar metal-DNA interactions involving other
transition metal compounds have been implicated, but none have been
elucidated to date. Studies in our laboratories revealed an unprecedented
bridging mode for dimetal carboxylate compounds of 9-ethylguanine and 9-
ethyladenine involving the N7 and O6/N6 atoms. These results led us to
investigate the type of interactions that occur with oligonucleotides. 1H
NMR spectroscopy was used to determine the solution structure of the
duplex d(5 ’-CCTCTGGTCTCC-3 ’) - d(5 ’-GGAGACCAGAGG-3 ’) before
and after reaction with [Rh2(OzCCH3)2(CH3CN)6][BF4]2. Differences in
chemical shifts and intensities of the cross-peaks provided information about
the perturbations of the oligonucleotide structure caused by covalent

interactions with the metal at the GG sites of d(5’-CCTCTGGTCTCC-3’).

TABLE OF CONTENTS .

LIST OF TABLES V.“ V

 

 

CHAPTER 1
Mechanism of Action of Cisplatin ...... 4
Dirhodium Carboxylate Compounds as

Potential Anticancer Agents 14
NMR Analysis of Cisplatin-DNA Adducts 18

 

 

DJ

 

CHAPTER 2

APPLICATION OF 1H NMR SPECTROSCOPY TO THE

STRUCTURAL ELUCIDATION OF METAL-DNA

ADDUCTS 21
Solvent Suppression Techniques .._.._..._ w 23
Two Dimensional 1H NMR Spectroscopy..-.._._,.__.-.-..-- 29

 

CHAPTER 3

EXPERIMENTAL SECTION 39
DNA Purification 40
NMR Sample Preparation .-__._ 45
1H NMR Analysis of a DNA Dodecanucleotide-.......-.-.............,.-.....--.-...._.-. 46

 

 

iii

Reaction of the DNA Dodecanucleotide with a

 

 

 

 

Dirhodium Acetate Complex 70
1H NMR Analysis of a RhZ-DNA Adduct 74
CHAPTER 4
MOLECULAR MODELING STUDIES 88
CONCLUSIONS 98
LIST OF RERERENCES 102

 

iv

 

--.'!'}

Table 1 .

Table 2 .

Table 3 .

Table 4 .

LIST OF TABLES

Chemical Shifts for Exchangeable and

 

Non-Exchangeable Protons in DNA

DNA 12-mer : Relaxation and Calculated

 

Correlation Times for AH8 Protons

Base-to-Base and Base-to-Hl ’ Intra and
Inter-residue Distances

 

ha-DNA Adduct : Relaxation and Calculated
Correlation Times for AH8 Protons

 

47

59

64

76

 

 

LIST OF FIGURES

 

 

 

 

 

 

 

 

 

Figure l . Schematic of the Cisplatin Molecule 3
Figure 2 . Nucleotide Structure Showing Labeling Scheme .................................... 5
Figure 3 . Schematics Depicting The Basic Structure of DNA ............................ 7
Figure 4 . Cisplatin-DNA Inter— and Intrastrand Adducts ........................................... 9
Figure 5 . Biologically Relevant Pt Compounds 13
Figure 6 . Comparison of Cisplatin and Dimetal
Carboxylate Binding Sites 15

Figure 7 . X-Ray Structures of Dirhodium Compounds

with Guanine Bases ~ 17
Figure 8 . DNA Octamer Sequence Studied by 1H NMR

Spectroscopy 19
Figure 9 . Sugar Conformations 19
Figure 10 . NMR Principle : Magnetization Vector 22
Figure 11 . NMR Pulse Experiment 23
Figure 12 . Binomial 1-1 Pulse Sequence 25
Figure 13 . Solvent Magnetization Vector for 1-1 Sequence .......................... 25

Figure 14 . Signal Intensity as a Function of Frequency Offset ..................... 26

vi

 

 

Figure 15 .
Figure 16.

Figure 17.

Figure 18.

Figure 19 .

Figure 20 .

Figure 21

Figure 22 .

Figure 23 .

Figure 24 .

Figure 25 .

Figure 26 .

Figure 27 .

Figure 28 .

Figure 29 .

Figure 30 .

Gradient Pulse Effect on Magnetization

27

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

vii

WATERGATE Pulse Sequence 27
Spin State Observable and Non Observable
Transitions ' 3O
NOESY Pulse Sequence 32
T1 Measurements : Inversion
Recovery Experiment 33
TOCSY Pulse Sequence 36
. 2Q COSY Pulse Sequence 38
HPLC Chromatogram of DNA Strands 42
DNA 12-mer 1H NMR Spectrum
in 90% H20 at 25 °C 48
Imino Region of DNA Duplex at
Different Temperatures 50
Base-to-Hl ’ NOE Connections 52
DNA Duplex WATERGATE NOESY
Spectrum at 25 °C 54
2D Imino Region of DNA Duplex 56
DNA 12-mer 1H NMR Spectrum
in 99% D20 at 25 °C 57
DNA 12-mer NOESY Spectrum
in 99% D20 at 25 °C 60
DNA 12-mer 1H NMR Spectrum
in 99% D20 at 30 0C 62

 

 

Figure 31 . DNA 12-mer NOESY Spectrum

Figure 32

Figure 33

Figure 34 .
Figure 35 .

Figure 36 .

Figure 37 .

Figure 38 .

Figure 39

Figure 40

Figure 41 .

Figure 42 .

Figure 43

Figure 44 .

 

 

 

 

 

 

 

 

 

 

 

 

in 99% D20 at 30 °C 63
. DNA Duplex TOCSY Spectrum
in 99% D20 at 30 °C 65
. DNA Duplex 2Q COSY Spectrum
in 99% D20 at 25 °C 66
Base-to-Hl ’ NOESY Walk 68
Base-to-H3’ NOESY Walk 69
1H NMR Spectrum of
[Rh2(02CCH3)2(CH3CN)6][BF412 in D20 71
HPLC Chromatogram of ha-DNA Adduct ...................................... 73
Imino Region of ha-DNA Adduct
at Different Temperatures 75
. ha-DNA Adduct 1H NMR Spectrum
in 99% D20 at 30 °C 77
. ha-DNA Adduct NOESY Spectrum
in 99% D20 at 30 °C 80
ha-DNA Adduct 1H NMR Spectrum
in 99% D20 at 20 °C 81
ha-DNA Adduct NOESY Spectrum
in 99% D20 at 20 °C 82
. Base-to-Hl ’ Region of RhZ-DNA Adduct 83
Imino Region of ha-DNA Adduct in
Buffer Solution 87

 

viii

Figure 45 . DNA Backbone Showing Torsion Angles ............................................ 91

Figure 46 . DNA Structure Showing NOE-Distance

 

Distance Restraints 94
Figure 47 . Superimposed Structures of
d(CCTCTGGTCTCC)-d(GGAGACCAGAGG)
96

 

using Simulated Annealing

ix

 

LIST OF ABBREVIATIONS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

cis-DDP cis-Diamminedichloroplatinum (II)
RNA Ribonucleic Acid

DNA Deoxyribonucleic Acid

A Adenine

C Cytosine

G Guanine

T Thymine

p Phosphate Group

HMG High Mobility Group

SRY Sex-determining Region Y protein
LEF-l Lymphoid Enhancer Binding Factor 1
tsHMG Murine Testis Specific High Mobility Group
trans-DDP trans-Diamminedichloroplatinum
TBP TATA Binding Protein
NMR Nuclear Magnetic Resonance

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

JM2 1 6 Bis-Acetato-Ammine-Dichlorocyclohexylamine-
platinum (IV)

NOE Nuclear Overhauser Effects

FID Free Induction Decay

M0 Magnetization Vector

My Transverse Magnetization Vector

Bo Magnetic Field

B1 External Magnetic Field

WATERGATE Water Suppression by Gradient Tailored

Excitation

NOESY Nuclear Overhauser Effect Spectroscopy

W1 First Order Transition

Wo Zero Order Transition

W2 Second Order Transition

S Spin

T1 Longitudinal Relaxation Time

T2 Transverse Relaxation Time

Tc Correlation Time

rm Mixing Time

t1 Delay Time

t2 Acquisition Time

xi

 

 

F1

 

First Dimension

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

F 2 Second Dimension

FID Free Induction Decay

LP Linear Prediction

TOCSY Total Correlation Spectroscopy
ms milliseconds

ns nanoseconds

ps picoseconds

2Q COSY Two Quantum COSY

1H Proton

d Deoxy

HPLC High Performance Liquid Chromatography
nm nanometers

min minutes

mL milliliters

g grams

cm centimeters

M Molar

mM millimolar

kDa kiloDaltons

 

xii

 

 

 

PPm

 

SA

 

RMSD

 

MD

Revolutions per minute

parts per million

Simulated Annealing

Root Mean Square Deviation

Molecular Dynamics

 

xiii

 

 

INTRODUCTION

Inorganic compounds have been used extensively as medicinal agents
for centuries. Presently, metal compounds are in clinical use for the
treatment of a number of diseases such as arthritis, hypertension, bacterial
infections and cancer. Regarding the latter, a great deal of research has been
conducted in the last 30 years due to the discovery that cisplatin and other
related complexes exhibit remarkable efficacy in the chemotherapeutic
treatment against several types of cancer.

Although the success of cisplatin as a chemotherapeutic agent has
been well established and, indeed, it is currently in widespread use, its toxic
side effects represent a considerable health risk for patients being treated
with this drug. This disadvantage has provoked the search for new
compounds with the same or higher activities but with lower toxicities than
cisplatin.

The use of other transition metals as a possible alternative to platinum
has been proposed in a number of studies over the years. In this dissertation,

a structural study of the principal binding target of cisplatin, namely, the

intrastrand GpG cross-link is being investigated for dirhodium compounds.
The main objective of this research is the elucidation of the mode of action
of dimetal antitumor active compounds by understanding the structural
distortions of DNA after metal binding has occurred. It is hoped that a better
understanding of the DNA perturbations at the molecular level will lead to
the discovery of new compounds that will satisfy the conditions necessary

for inhibiting DNA replication in cancer cells.

Chapter 1

1. BACKGROUND

The history of the anticancer properties of the compound cisplatin,
PtC12(NH3)2 or cis-DDP (Figure 1), dates back to 1965 when Professor
Barnett Rosenberg and his research group observed filamentous growth of
E.coli in experiments designed to study the effects of an electric field on the
properties of these bacteria.1 The bacterial cells, which normally divide very
rapidly, grew up to 300 times their size without undergoing division. This
observation was eventually recognized to be due to the presence of
platinum-ammine complexes formed by in situ electrolysis of the Pt
electrodes used in the experiment. One of these complexes is cis-
diamminedichloroplatinum(II) or cisplatin which has been shown to be
effective in treating testicular, ovarian, lung, head and neck, bladder and

cervical cancers.

H 3 N $531
’0... Pt . \\‘
/ \
H 3 N Cl

Figure 1. Schematic of the Cisplatin Molecule

A. Mechanism of Action for Cisplatin

After many decades of investigations, researchers have compiled a
large database of information about how cisplatin is metabolized. The
chloride ligands of cisplatin are stable with respect to substitution at high
chloride concentrations such as those found in the extracellular matrix, but
when the compound diffuses into the cell, the lower chloride concentration
causes loss of chloride ligands and replacement with water.2 The resulting
mono and bis-aqua complexes are reactive with nucleophilic sites of cellular
macromolecules. The effect of this inorganic compound on the synthesis of
RNA, DNA, and proteins has been evaluated, and it was found that at high
concentrations and long incubation periods, the synthesis of these three
biomolecules is inhibited.3 DNA synthesis is preferentially inhibited at all
concentrations, thus DNA is thought to be the most likely target of cisplatin
with respect to its chemotherapeutic properties.

In order to understand the suppression of DNA synthesis by cisplatin,
it is important to consider the nature of this macromolecule. DNA is a
polynucleotide chain composed of nucleotides that are phosphate esters of
pentose with nitrogenous bases linked to them (Figure 2). The standard

nomenclature for nucleic acids establishes numbering from 1 to 9 for the

 

*IIZ'LAK' “or. -. "'1' .

 

base and numbers from 1’ to 5’ for the sugar moieties. The hydrogen atoms
retain the same label of the carbon to which they are attached. In the case of
where there is more than one hydrogen atom on the same carbon atom, ie,
CX’ (X=H), the second hydrogen atom is labeled as CX”. The base is
attached at the 1’ position of the sugar residue, which, in this case, is a
deoxyribose since it has only one hydroxyl group located in the 3’ position.
The bases are derivatives of purines and pyrimidines, and are responsible for
the specificity of the sequence of DNA. This sequence defines the structure
and coding for further transcription of DNA into RNA and final translation
into proteins. The purine bases are guanine (G) and adenine (A) while the
pyrimidine bases are cytosine (C) and thymine (T). The polynucleotide
chain presents an orientation from the 5’ end to the 3’ end which are the

positions where the phosphate group binds to the deoxyribose.

N1

 

Figure 2. Nucleotide Structure Showing Labeling Scheme

The most common form of DNA consists of two strands wound about
a common axis in an antiparallel and helical fashion forming a right handed
5’ to 3’ helix. The strands are held together by hydrogen bonds formed
between the bases of each strand, thereby forming base pairs. The most
stable base pairs are those formed between guanine (G) and cytosine (C) and
between adenine (A) and thymine (T). These bases are called the Watson-
Crick base pairs. The double stranded DNA is then formed with the
phosphate backbone at the periphery (hydrophilic region) in contact with the
solvent, keeping the bases in the core (hydrophobic region) of the helix.
Because the base-pairs are not exactly in the center of the helical axis, the
DNA helix possesses two grooves with different widths and depths; these
are the major and minor grooves (Figure 3).

Double helical DNA can be classified into three types, namely, A, B,
and Z-DNA, which exhibit fundamental differences in base stacking along
the strand and in the conformation of the sugar-phosphate backbone.
Among these three types, B-DNA is considered the native form because its

X-ray pattern resembles that of DNA in intact sperm heads.4

.7-4——-—__m—_ a ‘_—\__‘ .. - v-

 

IL ,H

N 0
N ONIH N'H
/ H / Bar
H—< I NAS- H:l‘< N/l N/LO
Guanine Adenine Thymine Cytosine

  

Watson-Crick Minor
base pairs groove

Figure 3. Schematics Depicting The Basic Structure of DNA

The biological action of cisplatin is linked to the formation of stable
Pt-DNA adducts. Major products of Pt-DNA binding involve the (N7,N7)-
didcntate cisplatin cross-link between two intrastrand adjacent guanines
(GpG) or adenine-guanine (ApG) (Figure 4a,b). Minor products are
intrastrand cross-links between two guanine residues separated by one
nucleotide residue and interstrand cross-lian between two guanine residues
(Figure 4c,d).5 These cross-links have been implicated for the biological
activity of cisplatin because they produce distortions in the conformational
structure of DNA.6 Distortions observed in an X-ray structure of a DNA
oligonucleotide of twelve base pairs bound to cisplatin was published in
1996 by S. J. Lippard and co-workers. In this structure, the DNA
oligonucleotide exhibits a bent conformation with an angle of 35-40".7
Unwinding of the duplex was also observed.

It is presumed that the distortions caused by cisplatin have deleterious
effects on replication and transcription, thereby causing mutation.8 It is
hypothesized that these lesions are recognized by certain proteins whose
binding prevents the lesions from being repaired, a situation that ultimately
leads to cell death. Such recognition is specific and requires a “kinked”

DNA conformation.9

 

 

 

Figure 4. Cisplatin-DNA Inter- and Intrastrand Adducts

10

Some of the binding proteins have been identified as the high mobility
group (HMG) domain proteins. Recently, it has been discovered that
proteins responsible for mediating the biological processes of
recombination, replication and transcription require the presence of DNA in
an unwound and bent form and contain one or more HMG domains.10

HMG domains are a-helical structures that contain a high proportion
of proline, aromatic and basic amino acids. Some of these HMG domain
proteins, such as SRY and LEF-l, bind to specific DNA sequences via their
hydrophobic residues. These domains behave as intercalators, causing
unstacking of several base pairs, which leads to unwinding and bending of
the DNA.11 Similar structural distortions have been observed when cisplatin
binds to a DNA duplex.12 One of the latest reports on this subject,
published in 1997, established the high binding affinity and specificity of the
murine testis-specific high mobility group protein, tsHMG, to cisplatin-
modified DNA.”

It has been shown that the trans isomer of cisplatin, namely trans-
DDP, enters the cells and binds to DNA like cisplatin.l4 Experiments
performed in vitro have shown that, initially, trans-DDP binds more rapidly
than cisplatin, but, with time, the quantity of bound trans-DDP diminishes

5

while the quantity of bound cisplatin continues to rise.1 These results

 

11

suggest that preferential repair of trans-DDP adducts occurs over those of
cisplatin, although the details of how trans-DDP is removed are still not well
understood.

Interestingly, the HMG domain proteins studies also helped to explain
the inactivity of the trans isomer. Now it is believed that the inactivity of
trans-DDP is due to the specific binding of the HMG-domain proteins to
cisplatin modified DNA. Although trans-DDP is capable of binding to only
one strand in duplex DNA, it does so by forming a 1,3-intrastrand cross-link
of guanine bases, trans-[Pt(NH3)2{ d(G*prG*)}]16, which leads to different
structural alterations that are thought to be repaired more efficiently than the
distortions produced by cisplatin.

More recently, experiments using in vitro transcription challenge
competition assays demonstrated the capability of the promoter recognition
factor involved in transcription, the TATA binding protein (TBP), to
preferentially bind cisplatin damaged DNA in favor of normal binding sites.
The consequence of this is the prevention (or at least reduction) of
transcription.17

Despite its high anticancer activity, the use of cisplatin as a drug is
mediated due to its highly toxic side effects. Patients who receive the drug

as part of their therapeutic treatment are likely to suffer from renal toxicity,

 

12

hematologic deficiencies, hearing loss, nausea, neurotoxicity and cardiac
abnormalities. In response to these problems, other platinum compounds
with reduced toxicity such as carboplatin have been developed (Figure 5a).
Results indicate, however, that carboplatin may be less effective against
testicular cancer with respect to long term cures.18 Presently, other
compounds such as JM216 19 (Figure 5b) and oxaliplatin20 (Figure 5c) are in
various stages of clinical testing. The former has been shown to exhibit
lower toxicities than cisplatin and carboplatin. The latter appears to be
promising for colon cancer treatment, but its use might be limited due to its
high levels of neurotoxicity.

In spite of the promise of metal-based drugs, the battle against cancer
continues, with studies being focused on both improved platinum
compounds and on non-platinum metal compounds. New compounds may
exhibit reduced toxicities, improved activities against cisplatin-resistant

tumors or efficacy against cancers that are currently untreatable.

13

Cl In, \\\ N
” ...... ‘\ 2 OIO> Pt/ ""2
CI NH3 0 O \ NH2
JM216 Oxaliplatin

O

(b) (c)
2+
0 I’m", Pt,.\\\NH 3 NH 3/01.. Pt..\\\\Cl
NHa’ ‘NH2(CH2).H2N/ Na,

(d)
2+
NH3I/,,..Pt...\\\Cl Cllm..,Pt,..\\NH3
NH,’ ‘NH2(CH2),,H2N/ ‘NH3

(8)

Figure 5. Biologically Relevant Platinum Compounds

. d—--.—l-—r‘l—‘ — -
1
‘

 

14

B. Rhodium Carboxylates as Potential Anticancer Agents

In the search for new chemotherapeutic agents, studies were initiated
with antitumor active compounds that may be capable of binding to bases in
a manner similar to cisplatin. Dinuclear platinum and rhodium compounds
fall into this category. Because of their wider “bite angle” for binding as
compared to cis-DDP, there is a greater possibility that dinuclear compounds
can form both intrastrand and interstrand cross-links.21 Indeed, dinuclear
platinum complexes structurally related to cisplatin such as the cis- and
trans- isomers of [{PtCl(NH3)2}2H2N(CH2),,NH2]2+ (Figures 5d and Se), have
been tested for anticancer activity in mice and were shown to be effective
against cisplatin resistant tumors. 22 The DNA binding mechanism of these
compounds may be different from that of cisplatin, since it was found that in
a d(GCGC) sequence, the preferred binding sites are interstrand over the
favorite intrastrand -GG- cross-link of cisplatin.23 Some of these dinuclear
complexes promote a change in the conformation of DNA from the B type to
Z-DNA.24 Like cisplatin, they unwind DNA but they do not bend DNA in
the same way. This may play a role in their efficacy towards cisplatin-

resistant cells.

15

Among the group of non-platinum compounds that may exhibit
binding modes similar to cisplatin are dimetal carboxylates. Instead of one
square planar metal unit, they are composed of two approximate square
planar units joined by a metal-metal bond (Figure 6). Over the years, it has
been shown in our laboratories and others that the dimetal tetracarboxylate
compounds react with incoming bases by displacing two cis RCOz' ligands,

which means that they could react in a similar manner to cisplatin.

/><’\O

H3N/ NH3 0 O O \
.Pt’ — L
Cl Cl 0 ’ 0
90° 900

Figure 6. Comparison of Cisplatin and Dimetal Carboxylate Binding Sites

During the 1970’s, studies of dirhodium carboxylate compounds
revealed their potential as chemotherapeutic agents against certain tumors
including Ehrlich ascites and L1210 tumor cells in mice. The strong
inhibitory effect of the compounds on DNA but not on RNA synthesis in
viva was also demonstrated.25 In these experiments, the anticancer activity
of the compounds Rh2(OzCCR)4L2 (R = CH3, C3H7, C4H9) was explored and

it was found that the propionate and butyrate compounds exhibited higher

16

potencies than the acetate complex. The level of toxicity however, was
invariant. Among the dinuclear rhodium carboxylates, the compound
Rh;(DTolF)2(OzCCF3)2(HzO)2 (DTolF = N,N’-di-pftolylfonnamidinate) was
also tested for antitumor activity with promising results against Yoshida
sarcoma and T8 sarcoma of Guerin being obtained.26 This compound
exhibits reduced toxicity as compared to Rh2(OzCCH3)4 and cisplatin.

The early literature regarding the DNA binding studies of dirhodium
carboxylates reported that these compounds were unreactive towards poly G
DNA sequences but reactive towards poly A sequences and single-stranded
DNA. This conclusion was based on the absence of color change during the
reaction with guanine bases, while in the reactions with adenine bases a
dramatic color change from blue-green to pink was observed. Reactions
performed with 9-ethylguanine in our laboratories revealed that gaunine
does indeed bind and in an unprecedented manner wherein two guanine
bases are bound in a bridging mode. X-ray structures were determined and
the compounds were found to contain cis- guanines with the N7 and 06
atoms coordinated in a “head-to-head” bridging mode (Figure 7a) or a
“head-to-tail” bridging mode (Figure 7b).27 The reactions were performed
with 9-ethyladenine as well and it was found that N7,N6 bridges formed.

These results support the conclusion that dirhodium tetraacetate has

 

 

l7

0 V
- .' 1" . ll

3" "’ :’ H8
06’ ‘ . sf. 7C8 H89
07 ' Rh2
"' os
01 1'3- 1004 ‘a
. 2
,’__ 0 &4

(b) Rh2(02CCH3)2(9-EIG)2(MCOH)2.ZMCOH

NHz

F N
HN
f .s\\\\\\

T
M‘— T

u

(y

Head-to-Head

N/

e .
HN /)

\\\\\\\

‘3

(“—07
V

Head-to-Tail

T
T.

 

Figure 7. X-Ray Structures of Dirhodium Compounds with Guanine Bases

 

18

equatorial sites available for binding as well as axial sites.

C. N MR Analysis of Cisplatin-DNA Adducts

The aforementioned studies of dimetal compounds, which included
single cystal X-ray structures of model purine reaction products, were the
starting point for our studies of the DNA binding of dimetal carboxylate
compounds. We began our current investigation from the perspective of
what is known about cisplatin in order to make comparisons and contrasts to
the behavior of dirhodium carboxylates under the same conditions.

Determination of the structural distortions of cisplatin-DNA adducts
at the molecular level is important for the eventual understanding of
cisplatin’s activity. In addition to X-ray crystallography, NMR spectroscopy
is an extremely valuable tool for probing DNA-adducts of cisplatin. The full
mechanism of action of cisplatin as an anticancer agent may eventually be
understood on the basis of both solution and solid-state structures of
DNA/cisplatin/protein interacions.

Initial analyses of the distortions created by platinum complexes
bound to DNA were conducted with di- and tn'nucleotides. The small size

of these adducts renders them excellent candidates for study by NMR

 

l9

spectroscopy, molecular mechanics and X-ray crystallography.28 Distortions
upon platination of DNA duplexes of ten, eleven and twelve base pairs were
also investigated by 1D and 2D NMR spectroscopy.29 In 1995, a more
detailed 2D NMR analysis of an intrastrand cis-DDP-GG cross-link structure
of an octameric DNA duplex was performed.30 The DNA sequence selected
for these spectroscopic studies was d(CCTGGTCC) d(GGACCAGG),

represented in Figure 8.

1234 s 678
5’CCTG*G*TCC 3’
3’GGACCAGG5’

1615 14 13 12 11 1o 9

G*G* = Platinated site
Figure 8. DNA Octamer Sequence Studied by 1H NMR Spectroscopy

Some of the results of this work are the following:
(1) The conversion of the sugar conformation of the 5’ platinated
guanosine from C2’ endo to C3’ endo was confirmed (Figure 9), although
the remainder of the sugar residues of the double helix were in the original B

type DNA conformation (C2’ endo).

cs' 2' N C5' 3'
0 w
4' 3' 1' 4' 2. 1'

C2' endo; S type C3' endo; N type

Figure 9. Sugar Conforrnations

20

(2) The connectivities of both strands were generally uninterrupted with
some variations in the intensity of the NOE cross-peaks from protons near
the lesion site (Pt-G4pG5).
(3) The chemical shifts of G4*H8 and G5*H8 resonances were downfield
in comparison with the other G resonances, indicating that both bases are
platinated at their respective N7 positions.
(4) The G4 imino proton was not observed, presumably due to its rapid
exchange with water.

The NOE restrained, refined structure of the modified duplex showed
that the helix is kinked approximately 58° and unwound by —21°.

It should be noted that the structural features of this DNA sequence
were obtained by using 1H NMR spectroscopic data only, underscoring the

importance of this analytical technique for these types of structural studies.

Chapter 2

1. APPLICATION OF NMR SPECTROSCOPY TO STRUCTURAL

ELUCIDATION OF METAL-DNA ADDUCTS

The application of NMR spectroscopy to the study of biological
macromolecules involves the analysis of NMR spectra in two or more
dimensions. In these experiments, it is possible to observe through-bond
couplings for nuclei separated by two, three or sometimes four bonds. Also,
one can observe through-space connections between resonances of nuclei
that are close to each other in space (55 A), and exchange processes between
resonances of the same nucleus in two different chemical environments.
The structural information is principally obtained from measurements of the
Nuclear Overhauser Effect (NOE), which provides constraints on
internuclear distances, especially for small oligonucleotides. In addition to
the determination of complete three-dimensional structures for
macromolecules in solution, NMR provides valuable information on local
structure, conformational dynamics and structural aspects of interactions

with small molecules.

21

22

The principles of NMR spectroscopy are based on the behavior of the
nuclear spins when they are placed in a static magnetic field (Bo). Each
nucleus precesses around the field axis and by adding up all the 2
components of the nuclear magnetic moments, a macroscopic magnetization
vector (Mo) along the field direction is obtained (Figure 10). In an NMR
experiment, transitions are induced between the energy levels by irradiating
the nuclei with a external magnetic field (Bl) originated from an

electromagnetic wave of the suitable radiofrequency.

Mo= Magnetization

 

 

 

a
vector
I AB = y(h/21t)Bo
m = +1/2
[3
m = - 1/2

 

Figure 10. NMR Principle : Magnetization Vector

In the pulse experiment, a radio frequency pulse of duration 1,, (Figure

11a) is obtained by applying the magnetic field (B1) along a specific

23

direction. This can be observed in Figure 11b where the x and y axes are
rotating at the same frequency of B1. The magnetization vector M0 is tipped
away from the z axis by an angle that is dependent on the duration of the
pulse and the strength of the magnetic field applied. The resultant vector
(My) is called the transverse magnetization vector. This vector is important
because it represents the NMR signal since the detector contains the receiver

coil aligned along the y direction.

 

(a) *
—-> "p <—-

    

 

 

My=Mo

 

Bl = external magnetic field applied

Figure 11. Pulse Experiment

A. Solvent Suppression Techniques

For a thorough understanding of NMR processes in nucleic acids, it is
important to differentiate between two types of protons in these
biomolecules. This classification stems from the ability or inability of these

protons to exchange with protons of the solvent. The exchangeable types

24

include the protons A (NH2)6, G (NH2)2 and G(NH)1, T(NH)3, and
C(NH2)4 with the remainder of the protons present in the molecule being
non-exchangeable. In order to detect the exchangeable protons, the sample
must be dissolved in a protonated, rather than deuterated, solvent.
Unfortunately, this presents the problem of dealing with a very high
concentration of water compared with the sample concentration, diminishing
tremendously the dynamic range of the instrument and the intensity of the
resonances of interest. For this type of analysis either selective excitation or
solvent suppression NMR techniques must be performed in order to
overcome this problem.

Among the selective excitation methods, the binomial method is
commonly used. The method uses a selective excitation pulse sequence
which produces a non-excited region around the solvent chemical shift while
causing significant excitation in the region of interest. The magnetization
process can be explained by looking at one of the simplest pulse sequences,
the 1-1 sequence (Figure 12). In this sequence, the excitation pulses are
placed at one end of the spectrum, separated by N Hz, where N represents
the distance from the solvent resonance frequency to the region to be
excited. All the pulses have the same phase and are separated by a delay

time (2N'l ) in seconds.

25
1c/4x n/4 ,

1
tD Acquire tD = 2N sec

 

 

 

 

 

 

Figure 12. Binomial 1-1 Pulse Sequence

By observing the magnetization process for the solvent signal using
the vector diagram shown in Figure 13, one can observe that after the two
pulses, the magnetization vector returns to its initial position along the z

axis, therefore yielding no net excitation of this signal.

112 ‘1
3.?!L/<%\}P,%\flf />\

r
‘4 \4/ \q/ ’\</

Figure 13. Solvent Magnetization Vector for 1-1 Sequence

 

 

 

 

 

The magnetization vectors that do not precess at frequencies multiples
of N Hz however, will not be aligned in the yz plane before the second
pulse; instead, they will experience a net excitation. According to their
precession frequency, they will resonate before or after the solvent signal,

resulting in a 180° dephased spectrum. The intensity of the signal as a

26

function of frequency offset with respect to the solvent resonance is shown
in Figure 14. From this picture, it can be noted that it is important to set the
frequency offset appropriately for maximum excitation of the region of
interest. Other pulse sequences are called 1-2-1 and 1-3-3-1 which provide a

wider zero excitation around the solvent signal.

I l I l l A =freq(sample signal)-freq(watersignal)
I l I I I I l I

Hzo"||||||||I||'Aaiz)

 

freq

Figure 14. Signal Intensity as Function of Frequency Offset

Water suppression techniques are also very useful and effective for
NMR analysis of biomolecules. The development of actively shielded
probes that can produce high-power field gradient pulses has resulted in new
experiments that make use of these field gradients instead of phase cycling
for selection of coherences. A field gradient pulse can be defined as a period
during which the magnetic field B0 is made intentionally inhomogeneous. In

this field, the nuclear spins along the NMR sample experience different

27

magnetic fields, and thus transverse magnetization and other coherences
dephase across the sample. (Figure 15). The coherence can be refocused by

another appropriately applied gradient to generate gradient echoes.

 

¢ = Yszt

 

Trfl -- .

Figure 15. Gradient Pulse Effect on Magnetization

An example of this technique is the Water Suppression by Gradient
Tailored Excitation (WATERGATE),31 whose pulse sequence is shown

schematically in Figure 16.

(TI/2) x “Y
(rt/2) ., (11/2) -y

 

1H

 

t Receiver

 

t Selective Selective

6' f) H

 

 

 

 

Figure 16. WATERGATE Pulse Sequence

28

The gradient pulses (G2), shaped as a smooth function of time, are used to
remove the unwanted coherences. Following the initial non-selective
pulses, a field gradient dephases the solute and solvent magnetizations. The
former is not affected by the selective pulses. The non selective 1t pulse
inverts the coherence order of the solute magnetization and, in this way, the
second gradient pulse rephases the solute magnetization to form a gradient
echo. On the other hand, the combination of selective and non selective
pulses leaves the coherence order of the solvent magnetization unchanged.
The consequence of this is complete dephasing of this signal after the second
pulse gradient is applied and no echo is formed. This technique provides

highly efficient suppression of the water signal.

 

29

B. Two-Dimensional NMR

Two-dimensional NMR analysis of biomolecules, provides a variety
of techniques for obtaining information on a molecule in three dimensions.
One of the most important techniques is Nuclear Overhauser Effect
Spectroscopy, or NOESY which correlates protons that are separated by
distances as great as 5 A. This technique is based on the Nuclear Overhauser
Effect, or NOE, which is the change in intensity of a resonance due to the

perturbation of transitions of another resonance.

n1: (I-IO)/109

where Io represents the equilibrium intensity. This change is a consequence
of the modulation of the dipole-dipole coupling between different nuclear
spins produced by the motion of the molecule in solution. In a spin network
contained in a macromolecule, spin diffusion by two or several subsequent
relaxation steps among nuclear spins in proximity, can also take place and

influence the observed NOE intensities.

30
For two 1= 1/2 uncoupled spin nuclei with different chemical shifts, the
energy level system is represented in Figure 17. The a0 and [301 levels are

not at the same energy because the proton nuclei do not have the same

chemical shift which renders them inequivalent.

 

f/ BB A
Y I
—— 8,1 = Observable transitions
“B VX\\‘
I Ba X,Y = Non observable transitions
S
___L_
(10.

Figure 17. Spin State Observable and Non Observable Transitions

There are two observable transitions, indicated by I and S, as well as
two non-observable transitions, represented by X and Y, which may
contribute to the relaxation processes. The relaxation through observable
transitions is a first order relaxation process (W1) while the relaxation due to
non-observable transitions is denoted as W0 for X, which is the zero-order
transition, and W; for Y, which is the second-order transition. Therefore the

relaxation process is a result of the combination of W0, W1, and W2, because

31

now a saturation of spin S will affect the intensity for spin 1. These zero and
second order transitions are responsible for the detection of an NOE.

The local field that one spin experiences due to the presence of
another spin will depend on the orientation of the molecule. In solution,
rapid molecular motion averages the dipolar interaction, generating
fluctuating fields which stimulate longitudinal relaxation (T1). Furthermore,
the strength of these interactions depends on the internuclear distance.
Therefore we can say that the intensity of an NOE will depend indirectly on
the distance r between pre-irradiated and observed spins, and directly with

respect to the correlation time, ‘tc, which is the average time for the molecule

to rotate through one radian. The NOE is a population effect; therefore, it
will influence the intensities of the resonances since they are proportional to
the difference in population of the energy levels between which the nuclear
transition occurs.

NOE a (l/r6).tc

The pulse sequence for NOESY is shown in Figure 18. A variable
delay time t1 is inserted after the first pulse, thereby allowing the different
nuclear spins to precess according to their frequencies in the XY plane. The

second pulse brings the magnetization vector back to the z axis.

32

Magnetization transfer due to dipolar coupling takes place during the mixing
time, rm, before observable transverse magnetization is created by the final
90° pulse. Finally, the signal is detected and recorded during the acquisition
time t2. Fourier transform with respect to t2, corresponds to the first
dimension of the spectrum while Fourier transform with respect to the delay
time, t1, corresponds to the second dimension of the spectrum. The cross-
peaks will consist of responses corresponding to the size of the NOE that
was built up during the mixing time. By looking at the intensity variation as
a function of this time, distance information can be extracted from the

spectrum.

 

 

113/ 2x 75/24; 113/2,‘
t2
t1 Tm [—
Acquire
evolution mixing acqu'sition

Figure 18. NOESY Pulse Sequence

One important preliminary experiment for any DNA sample prior to a
NOESY experiment is the measurement of T1, or longitudinal relaxation

time. The value of T1 is defined as the time required for the magnetization

33

vector representing a certain spin to return to its equilibrium state along the z
axis. This measurement can be obtained by performing an Inversion
Recovery experiment. This experiment records a series of spectra using the
following sequence:

180°, — t — 90°x - Aquisition

The value of t changes for each spectrum. The behavior of the

magnetization vector is illustrated in Figure 19.

 

 

Figure 19. T1 Measurement : Inversion Recovery Experiment

As t varies, the amplitudes of each signal changes in different ways.
The next 90° pulse will give an observable transverse magnetization. A
return to equilibrium will take place through relaxation associated with the
transfer of energy from the spin system to the surroundings (spin-lattice
relaxation). This relaxation will occur at a rate determined by T1 based on

the following equation,

34

sz / dt = (-Mz- M0) / T1

where M2 = -M0 at t=0; resolving this equation:

ln(Mo-Mz) = 1n 2M0 -— t/Tl
Therefore, by plotting ln(M0-Mz) against t, the value of T1 can be

calculated from the slope of the straight line obtained.
Fully deoxygenated samples of duplex DNA exhibit T1 values for the
H2 protons of adenosine in excess of 3 sec. The long T1 values for these
isolated protons require long recycle delays to be implemented between
transients during acquisition of NOESY data to allow for full recovery of
their signals. Another important factor for NOESY is obtaining data of
adequate signal-to-noise ratio and sufficient digital resolution. Acquisition
of NOESY data with good digital resolution in t; is possible'with the help of
computers. However, the acquisition of NOESY data with sufficient digital
resolution in t1 requires making a compromise between high signal-to-noise
in each increment when the value of t1 is varied, or acquiring more
increments with reduced signal-to-noise. By using linear prediction (LP) as
part of the data processing, the need for this compromise can be eliminated.
In many cases insufficient t1 points are acquired during NOESY to permit an
interferogram signal to decay to zero. Forward LP can be used in these

cases to extend the data.

35

Linear prediction is based on singular value decomposition (LPSVD).
This is a digital processing technique that assumes each interferogram in the
F1 dimension or free induction decay (FID) in the F 2 dimension is a series of
exponentially decaying cosine or sine functions. Therefore, linear prediction
allows one to extend these mathematical functions to longer acquisition (t2)
or evolution (t1) times. The frequencies, amplitudes and phases for the
signals can be calculated from any portion of the data. Furthermore, with
LP, the amplitudes of the signals of the FID are determined only from the
experimental data, having unaltered intensities for the NOESY experiment
extended with LP. It also provides resolution enhancement, allowing for the
evaluation of volumes for cross- peaks that are closely spaced, which would
be poorly resolved in the NOESY spectra without LP. Therefore, the utility
for using linear prediction as a data processing tool has become very
important in terms of obtaining better spectra and reducing acquisition time.

Among NMR techniques involving scalar coupling spins, Total
Correlation Spectroscopy or TOCSY is a very useful technique for
elucidating direct coupling in biomolecules. TOCSY uses a pulse sequence
in which an isotropic mixing time is included. In this sequence, the
chemical shift terms are eliminated over a significant frequency range while

the spin locking radiofrequency field is applied. Magnetization transfer

36

under the influence of such a sequence, results in a continuous mixing
process with the magnetization moving periodically among all the spins in
the scalar coupled network. In this way, all the spins that belong to this
network can be connected. The isotropic mixing time can be varied in order
to obtain either efficient transfer of magnetization through a single three-
bond scalar coupling only or maximum transfer of magnetization between
resonances at extreme ends of a spin system. Both of these aims cannot be
accomplished by using a single mixing time, therefore experiments using
different rm values are recommended. The range of variation of rm is from
50 to 100 ms.

The basic pulse sequence for a TOCSY experiment is presented in
Figure 20. After frequency labeling during t1, magnetizatiOn is returned to
the z axis for the isotropic mixing process to take place. After that the

magnetization is rotated to the transverse plane for detection.

11/ 2x 111/ 2.x fl/zx
isotmpic mixing t2
t1 |

 

Tm Acquire

 

Figure 20. TOCSY Pulse Sequence

37

The observation of the cross-peak between two spins in the TOCSY
experiment does not necessarily indicate that the spins are directly coupled;
the magnetization is transferred between them by. a series of steps through

two- or three-bond scalar couplings between mutually coupled spins.

In most cases, the assignments of the majority of the proton chemical
shifts of a complex DNA molecule are possible by using a combination of
different 2D NMR experiments. Usually NOESY and TOCSY are the
most important and useful ones. However, sometimes these methods are not
sufficient to allow for information to be obtained from overlapping areas of
the spectrum, especially those close to the diagonal. Furthermore, with
TOCSY, magnetization transfer through the H3’ protons can be poor due to
the unfavorably low values of the H3’-H4’ coupling constants found in B-
form DNA. Furthermore, the use of the isotropic mixing pulse for a long
period of time can produce heating of the sample. These problems can be
eliminated by using an adjusted phase-sensitive 2Q COSY experiment.32
Here, direct correlations are present as pairs of cross-peaks at mirror-image
locations about the double quantum diagonal, where the chemical shift in F1
is the sum of the single-quantum chemical shifi in F 2. One of the advantages

of the 2Q COSY experiment is the absence of diagonal peaks in the

i1.
'iUre

38

spectrum. It also provides direct and relayed correlations if an adequate
pulse-phase is applied. Furthermore, these correlations are easy to separate
since the direct ones are symmetrical about the double-quantum diagonal
while the relayed correlations are not; therefore, the separation can be made

by symmetrization about the double-quantum diagonal.

The pulse sequence of 2Q COSY is presented in Figure 21. The first
three pulses are dephased by 45° with respect to each other. Following the
initial 90° pulse, an antiphase coherence is developed during a fixed spin-
echo sequence. This is in contrast to COSY experiments in which antiphase
coherence is developed during t1. The 180° pulse re-focuses the chemical
shift evolution. Multiple-quantum coherence is generated by the second 90°
pulse and the precession of the desired 2Q state is monitored during the t1
period. The final 90° pulse creates observable single-quantum

magnetization which is recorded during t2.

it'll all

1t/2 it
in 1:/2 t1 ‘2
Acquire

Figure 21. 2Q COSY Pulse Sequence

 

 

Chapter 3

1. EXPERIMENTAL SECTION

The work in this chapter was developed along similar lines to
previous studies carried out with cisplatin. Samples of dirhodium
carboxylate compounds and DNA oligonucleotides were reacted and the
products analyzed by 1H NMR spectroscopy. In order to establish a
comparison between the native and dirhodium-bound DNA, a detailed
analysis of the native DNA oligonucleotide sequence was first undertaken.

The DNA sequence selected for these studies is d(5’-
CCTCTGGTCTCC-3’) -d(3’-GGAGACCAGAGG-5’). There are several
characteristics of this sequence that render it desirable for this study. The
presence of two consecutive guanine bases in the middle of one strand
allows for only one binding site for the dirhodium carboxylate molecules on
this oligonucleotide. Furthermore, the 2D 1H NMR analysis of a very
similar sequence, although with fewer base-pairs, has been performed with

the cisplatin work. A summary of these findings was presented earlier. This

39

4O

NMR analysis will allow us to make an adequate comparison of the previous
studies with our work. Additionally, since the platinated adduct of this 12-
mer sequence was structurally characterized, there is good potential for

obtaining a crystal structure of the dirhodium-DNA adduct as well.

A. DNA Purification

In order for a sample to be analyzed by NMR spectroscopy, it must be
very pure. Any contamination from residual oligomers of shorter lengths
leads to complications in the NMR spectra from the presence of more than
one DNA population. Contamination due to the presence of metal di-cations
is also a concern since they bind strongly to DNA. Procedures used in the
elimination of these two contaminants will enrich the sample with salts
which have to be removed as well.

Obtaining purified DNA oligomers is limited by the available
technology for the synthetic process. Synthetic oligonucleotides of x-bases
are contaminated by varying proportions of “failures”, i.e., oligonucleotides
of lengths x-l, x-2, ..., x-(x-l). Separation of the 12-mer oligonucleotide
from the mixture was performed by ionic exchange chromatography using a

Perkin Elmer LC 235 HPLC instrument. A Source Q15 of 15 microns bead

41

size (Pharmacia) was used as the stationary phase. The eluants used for the
mobile phase were a combination of low and high salt concentration
solutions labeled as A and B:
EluantA : 0.1 MNaOAc, 20% CH3CN
Eluant B : 0.1 M NaOAc, 20% CH3CN , 1 M KCl

For the separation to be effective, the process was performed by

using the following gradient :

 

O 100 0 4
5 70 30 4
65 65 35 4
70 0 100 4
75 O 100 4
80 100 0 4

 

 

 

In this gradient, the salt concentration is increased rapidly during the
first five minutes in order to allow for the shorter-length, less highly
charged, oligonucleotides to elute first. The salt gradient then becomes more
shallow which allows for separation of the oligomers of higher charge,
(especially the ll-mer and 10-mer oligonucleotides) from the one of interest,

namely the purified 12-mer. This process was performed for both strands.

42

The HPLC chromatograms are presented in Figure 22. Each of the residues
was observed by using UV detection with a wavelength of 260 nm. The

retention time for each 12-mer strand was typically 30 minutes.

 

12

d(CCTCTGGTCTCC)
R. = 29.49 min

 

 

 

 

65 t(min)

 

d(GGAGACCAGAGG)
R. = 34.35 min

 

0 5 65 t(min)

 

 

 

Figure 22 - HPLC Chromatogram of DNA Strands

After purification of each 12-mer strand, the sample was concentrated

to reduce the volume of the fractions. In order to eliminate metal ions, a

43.
chelating ion exchange resin called CHELEX (BIORAD) was added in
amounts of 1 g per each 30 mL of sample. CHELEX consists of a sodium
salt of a styrene divinylbenzene copolymer that contains paired
iminodiacetate ions which act as chelating groups in binding polyvalent ions
with a selectivity for divalent over monovalent ions of 5000 to l. The
sample was incubated with this resin overnight with constant stirring during
which time the divalent ions were exchanged for Na+ ions. The DNA was
separated from CHELEX by passing the sample through a small column.
The resin residue was rinsed with three 1 mL aliquots of a l M NaCl
solution to extract residual DNA. The sample was then passed through a
Sephadex G-25 fine pore size exclusion column (Pharmacia) in aliquots of
17 to 22 mL for desalting. The size of the columns was 75 x 2.5 cm in order
to provide efficient separation. The desalted sample was evaporated to
dryness, giving the pure DNA 12-mer strand as a transparent film. This
process was performed for each strand. The purification yield for each
strand was determined by measuring the absorbance at 260 nm using the
experimentally determined extintion coefficient, 8, of 10,133 M'1 cm’1 for
d(CCTCTGGTCTCC) (GG strand) and 13,467 M" cm" for

d(GGAGACCAGAGG) (CC strand). The yield was between 30-50% for

44

both strands. Later, this yield was improved by changing the eluant
composition to:

Eluant A : 10 mM NaOH , 200 mM NaCl

Eluant B : 10 mM NaOH , 1 M NaCl

The gradient used with these eluant times was similar to the one
employed earlier, although the composition of A ranged from 64% to 59%
during the 60 minute elution period. This change in eluant composition and
elution profile increased the purified oligomer yield to 60%.

Once purified, both strands were mixed in an equimolar ratio. The
mixture was allowed to stabilize for about 30 minutes and then heated to 60
°C and slowly cooled to room temperature to improve annealing. In order to
eliminate any excess single strand, the sample was spun on a centriprep 3
centrifugal concentrator (Amicon), having a membrane with a molecular
weight cutoff of 3 kDa which is slightly smaller than the double-stranded
12-mer. Excess single strand passes through the membrane, with the

purified 12-mer oligonucleotide being retained in the upper sample chamber.

45

B. NMR Sample Preparation

The NMR sample was first prepared in H20 to allow for detection of
the exchangeable protons. An appropriate stock NMR buffer was added to
the sample to obtain a final concentration of 10 mM NazHPO4/NaH2PO4, 50
mM NaCl and 0.1 mM NaC2H6Ast (sodium cacodylate). The pH of the
sample is 7 in 0.25 mL final volume.

The NMR tube used for all of the experiments was a Shigemi tube
that allows for the analysis of a small volume of sample, therefore
minimizing problems with the signal-to-noise ratio for low concentration
samples. In a Shigemi tube, the sample is enclosed between two glass plugs
that have been treated in such a way that their pennitivity is equal to that of
the solvent used for all the experiments which in our case is water. In this
way, the sample volume can be reduced in order to equalize the length of the
radiofrequency coil of the instrument. When the sample is introduced into
the tube, an air bubble is produced at the sample/upper glass plug interface.
This bubble can be eliminated by placing the sample in a centrifuge and

spinning it for 5-10 minutes at 1500 rpm.

46

C. 1H N MR Analysis of DNA Dodecanucleotide

The NMR spectra were collected on Varian VXR 500 and UNITY
INOVA 600, 500 and 600 MHz spectrometers. A WATERGATE
experiment in 90% HzO/10% DzO was performed at 25 °C on the INOVA
600 which is programmed with the pulse field gradients needed for this
experiment. This spectrum allows for the non-exchangeable and
exchangeable protons to be observed. Only 10% of the deuterated solvent is
necessary for locking the radiofrequency signal. The spectrum obtained is
shown in Figure 23. It is important to point out that the water resonance at
4.769 ppm has been completely suppressed without distortion of the baseline
or elimination of the sample resonances demonstrating the effectiveness of
the WATERGATE pulse sequence for solvent suppression.

Using this first 1D spectrum, the basic proton regions in nucleic acids
were assigned. The protons chemical shifts were identified by using the

ranges presented in Table l.33

47

Table 1. Chemical Shifts for Exchangeable and Non-Exchangeable

 

 

 

 

 

 

 

 

Protons in DNA

Label 8(ppm)
2’H,2”H 1.8—3.0
4’ H , 5’ H, 5” H ........................................ 3.7.45
3’ H 4.4— 5.2
1’ H 5.3 - 6.3
CH3 ofT 1.2—1.6
5 H ofC 5.3 — 6.0
6HofCandT 7.1—7.6
8HofGandA,2HofA ----- 7.3—8.4
NH2 of A, C, and G ................................... 6.6 — 9.0
NHofGandT 10-15

 

 

 

 

48

o. 2 a on: $8 a 8:8st ~32 P 58-2 <75 - 2 25E

 

 

 

 

 

 

 

 

 

 

8% o N e e w 2 .2 E 2
LPPPLbLEF»FrwrprhhppplrphPFPLPPle-ppphler_hnhrpltprbL-pprp-prPbrpphkpppbpppp_bpbrP-rpr
l A! _),.. f s, £2 3
_ ...,,
5.....,\
¢ L Eamon :58—
than

:3on eta—=95.

 

3:80 3532

 

 

2.395 haw—5

49

A series of 1D experiments at different temperatures were performed
using the binomial 1331 sequence; these spectra are presented in Figure 24.
The differences in chemical shifts and intensities are remarkably obvious as
the temperature increases by 5 °C for each spectrum.

The A-T imino proton resonances of the duplex DNA are located
further downfield with respect to the G-C imino proton resonances. The
observation of the hydrogen-bonded imino protons in a Watson-Crick base-
paired system represents direct evidence for the formation of the double-
helical structure of the DNA dodecanucleotide. By examining these spectra,
it can be noticed that, at room temperature, only ten out of twelve resonances
are observed due to fraying of the ends of the duplex which allows for a
rapid exchange with water. At lower temperatures, the line-width of the
resonances increases due to a decrease of the rate of tumbling of the
molecule; the rate of fraying also diminishes. As a consequence, it is
possible to observe the imino protons at both ends of the duplex. For this
sequence the ends consist of G-C base pairs for which resonances can be
observed at 12.6 and 12.8 ppm at 5 °C. As the temperature is increased, the
line-widths of some of the resonances become sharper in such a way that at
20 °C the four A-T imino base pairs are clearly observed while some of the

G-C imino base pairs are still overlaped. The highest temperature used in

50

A-

T G-C
5 °C /\
l
’ Jw/V‘
L

Yfi—Vfivfir—vif

 

f v~v ifiy fifiYV—f

14.5 14.0 13.5 13.0 12.5 12.0 ppm

10°C

 

rfij

14514.0 13.5 13.0 12.5 12.0 ppm

15°C

 

fiVfi—f v—fi v Y—rw j

14.5 14.0 13.5 13.0 12.5 12.0 ppm

20 °C

 

 

 

v—v fi—r v 7

14.5 14.0 13.5 13.0 12.5 12.0 ppm

25 °C

11

1' Vi TV #7 fir

14.5 14.0 13.5 13.0 12512.0 ppm

 

30 °C

1

Y vv—fifvfivvvv—rVVfiVfirvvafr vv‘

14.5 14.0 13.5 13.0 12.5 12.0 ppm

 

40 0C

 

f f f T

14.5 14.0 13.5 13.0 12.5 12.0 ppm

49 °C

 

 

vav r

14.5 14.0 13.5 13.0 12.5 12.0 ppm

Figure 24. Imino Region of DNA Duplex at Different Temperatures

51

this experiment was 49 °C, and even at this temperature some of the
resonances are still intense. This confirms the high stability of this duplex.
For all the experiments, the water resonance was adjusted to the correct
chemical shift as the temperature was varied.

In a NOESY spectrum, we can observe the connections between
protons that are close in space. The spectrum can be analyzed by dividing it
into small regions. Each region will consist of the chemical shifts for a
group of protons of the oligonucleotide structure along the horizontal axis
(W1) and the chemical shifts of another group of protons along the vertical
axis (W2). Thus, we can follow, by alternating vertical and horizontal lines,
the cross-peaks that represent internucleotide and intranucleotide
connections between different proton groups along the DNA sequence. This
procedure is called a NOESY walk. The regions that provide the most
information through NOESY walks are the aromatic region (H8 of A and G,
H6 of C and T) and the H1’ and H3’ regions of the sugar residues. The
connections between the sugar H1 ’ protons and aromatic H6/H8 protons of
C,T/A,G can be observed in Figure 25 where each of the arrows represents a

cross-peak in the 2D spectrum.

Figure 25. Base-to-Hl ’ NOE Connections

 

5'

 

 

 

3'

53

A 2D WATERGATE NOESY spectrum was acquired in 90%
HzO/10% D20 at 25 °C in order to examine the connections between imino
protons. The pulse sequence is the basic NOESY sequence with a
WATERGATE pulse sequence inserted after the first 90° pulse. The
spectral width was 14500 Hz in F1 and F2 dimensions with a mixing time of
250 msec and a recycle delay time of 5 seconds. Each F ID was the average
of 64 scans using 512 t, increments. A total of 2048 complex points out of
a total of 4096 real points were used. Data processing was performed by
using NMR Pipe (Molecular Simulations Incorporated). This program
contains a collection of NMR tools for the UNIX workstation environment
It also contains a conversion program that allows for the implementation of
processing schemes directly from the spectrometer data format.

During the data processing, solvent subtraction was performed in the
time domain. Baseline correction was obtained by using a square
polynomial function in the frequency domain. A 0.4 shifted cube cosine bell
window was used as the apodization function. Zero-filling was also applied
for the next 4096 points. Forward and backward linear prediction (LP) was
used for respective prediction and replacement of data at the head and tail of
the FID. The full spectrum is presented in Figure 26. By looking at the

expanded region of the spectrum that connects the imino-to-imino protons

54

 

2.

 

 

 

 

Figure 26. DNA Duplex WATERGATE NOESY Spectrum at 25 °C

F2 (Ppm)

Hl

55

in Figure 27, it can be observed that each A-T imino proton is connected to
two G-C imino protons, thereby illustrating the communication between
hydrogen-bonded imino protons of the duplex.

Following the same reasoning, it can be assumed that the bases of the
next lowest intensities will correspond to the bases next to the ends of the
duplex. It is also important to note the existence of a complete overlap
between C4-G21 and C9-G16 imino protons. Therefore, these assignments
could only be made by combining this portion of the spectrum with regions
connecting the aromatic and H1 ’ protons and also with the amino-imino
connections. As stated before, the ends were not observable and the least
intense peaks corresponded to G23-C2 and G14-C11 base pairs.

1D and 2D NOESY 1H NMR spectra in D20 at 25 °C were performed
for the same sample using the Varian 600 MHz spectrometer. The sample
was concentrated to dryness, dissolved in 99.9% D20 and concentrated
again in vacuo. This process was repeated three times with the sample being
back-filled with argon each time in order to reduce exposure to atmospheric
H20 as much as possible. Once again, a Shigemi NMR tube was used
following the same procedure for the loading of the sample and elimination
of the air bubble enclosed between the glass plugs. A 1D 1H NMR spectrum

of this sample is depicted in Figure 28.

56

 

1 2 34 5 6 78 9101112
cc'rc'roc'rc'rcc
oaaoaccaoaoo
24 23 22 21 20 19 18 17 16 15 14 13
201522 17 2314181921/16

o 0

o o

o

o ’

 

 

' o ....... o

a ,

 

 

 

 

 

 

 

 

 

 

14.0 13.8 13.6 15.4 13.2 15.0 12’s 15.5 12'.4 12°.2
Hl (ppm)

Figure 27. 2D Imino Region of DNA Duplex

13.6 13.4 13.2 13.0 12.8 12.6

13.8

(ppm)

Hl

57

can O

it’ll;

.11].

 

 

_\\IIA\J__ _a. ‘3
_>.. \ ., T _. ..
_ L F 2 .1
.1 T
.1... ‘

0. mm a 03.x? a 6.58% E22 E. 58$ <za an 2:3...

‘_—.———

 

_ __M

 

' fi_._______-

.2
:.

 

...1 f. E
r. .__._
. E

.. \1./_ s 2111111111111. ) 5 a N)
._ .1 . : ..., .3 .
. ...,. i. :3 .2 2.3.2.1.; ..._ __. Z
n i: (1:. 2.. f... 1.: _._ 21:. i.

A:

58

An inversion recovery experiment was first performed in order to
obtain the T1 relaxation values for the AH2 protons. This allows for the
selection of the best recycle delay time, d1, for obtaining a better 2D data.
The relaxation time T1 observed for these protons was 4 see which is
relatively long compared to the values for other protons, e.g. the H8’s which
exhibit values between 1-2 sec. Therefore the delay time was set to three
times this value, which represents a compromise between obtaining fully
relaxed signals before each cycle, and increasing the amount of time
required for the experiment. Along with T1 experiments, measurements of
transverse relaxation time (T2) were performed. With these two
measurements, the correlation time 1c of the DNA duplex that describes the
overall tumbling rate of the molecule can be calculated by using an equation
that relates the T1 and T2 relaxation times which were measured
experimentally for individual protons.34

rc = (2/(11)[T1/3T2]”2 where a) is the Larmor frequency

The correlation time was calculated for the base protons AH8. The

results are presented in Table 2 with an average Tc value of 2.19 i 0.03 ns.

59

Table 2. DNA 12mer : Relaxation and Calculated Correlation Times

for AH8 Protons

 

ins) T2(S) - 1. x109 6")
A20H8 2.358 0.0666 2.188
A15H8 2.427 0.0695 2.173
A17H8 2.431 0.0707 2.156
A22H8 2.451 0.0670 2.223

 

 

 

It is important to note that although the calculation of the correlation
time was made with only a few protons, the reliability of the value obtained
is acceptable since these protons are isolated in the spectrum without
interference from coupling to other protons. This fact allows for an accurate
measurement of their relaxation times to be obtained.

1 The NOESY experiment was set with the same parameters used for the
sample in H20 at 25 °C. The processing parameters were similar to those
used in earlier experiments, but in this case a polynomial cube function in the
frequency domain was used for baseline correction and only forward LP was

needed. The full spectrum is shown in Figure 29.

6O

 

:0

l

8.0 7.0 6:0 5:0 410 3:0 210

T

l

 

 

 

 

l

810 770 610 5:0 4:0 3:0 210 1.0
H1 F1 (ppm)

 

0 Regions of Strong Overlap

 

 

 

Figure 29. DNA Duplex NOESY Spectrum in 99% D20 at 25 °C

F2 (ppm)

H1

61

The assignment process was performed following an established
protocol.35 The assignments were initiated in the base to H1 ’ region. The
AH8 protons were easily recognized by their shape and chemical shifts.
They were easily differentiated from AH2 protons because the latter are
sharper and, in most cases, they are associated with some T1 noise due to
their slow relaxation rates. This T1 noise is represented by bands of random
fluctuations running parallel to the F1 dimension at the position of strong and
sharp diagonal peaks which tend to distort the spectra. A problem observed
during the assignment procedure was the presence of regions in the spectrum
with severe overlap. Some of the cross-peaks coincide in the same chemical
shift region rendering them undistinguishable. Due to this fact, 1D
experiments at different temperatures were performed in order to obtain
better resolution of the overlapping resonances to continue the assignments
and confirm the ones already made. The temperature selected was 30 °C due
to the better separation and sharper line-widths for the resonances. A
NOESY experiment was performed at this new temperature eliminating the
majority of overlapped regions. The 1D and 2D spectra obtained at 30 °C

are presented in Figures 30 and 31 respectively.

62

Sam o _

 

 

 

0. cm a can 38 a 8:8st ~52 m. 88$ <zm .8 8:5

m v m o N. w a

1 L 1 L L , L L L 1. 1. .
L123... ...,.L. ...L..,,._........ ...L L...) L.....L... ._.. L
E: ‘L L L L___...LL L L LLL: __ c : L .
2 .L L_. . LL. : ...L L LLL __L _. LL LLL
L . L... L... L... 5 LL

_ ._ L. L L L ..., .L.

L. ,.

 

 

 

63

 

 

 

o O
O. 0' ‘
o .0 o
.8‘ 7. l-'
d. w. .35- a
GI. O " 0
M a... . ..
o o o
'0 o.
o O

 

7.0 6.0 5.0 4.0 3.0 2.0
F2 (Ppm)

Tm

 

 

Figure 31. DNA Duplex NOESY Spectrum in 99% D20 at 30 °C

F1

.0 3O 2.

(ppm)

Hl

64

A TOCSY experiment was also performed at the same temperature.
The spectrum is presented in Figure 32. This experiment is designed to
locate the H5-H6 scalar coupling cross-peaks of the C residues.
Furthermore, some of the directly coupled proton resonances were
confirmed from the 2Q COSY spectrum taken at 25 °C (Figure 33). This
spectrum, together with the 2D NOESY data, led to the observation that
there were two consecutive C bases connected to each other that were
connected to A bases as well. From this information, these residues were
assigned to the C residues located in the middle of one of the strands. For a
standard B form DNA oligonucleotide, there are many possible
internucleotide and intranucleotide proton distances. Some of the distances

that were useful for our initial assignments are presented in Table 3.3°

Table 3. Base-to-Base and Base-to-Hl ’ Intra and Interresidue Distances

 

H8 - Hl’ 3.9 7.7
H1’- H8 2.8
H5 — H6 5.0

H6 — H5 3.7

 

 

 

65

 

 

 

 

T
.O

8.0 7.0 5.0 510 4:0 3:0 2:0 1

l

l

T

 

 

8:0 7:0 6:0 5:0 4:0 3:0 20
H1 F1 (ppm)

1 .

l

0

Figure 32. DNA Duplex TOCSY Spectrum in 99% D20 at 30 °C

F2 (ppm)

Hl

F1
(ppm)

10
12
14

16

Figure 33. 2Q COSY Spectrum of DNA Duplex at 25 °C

66

 

 

 

...9,
‘, o
”"139: °. 3.
. .. - ' o
' I...“ ‘4';
5.7;." if»
. i . I
o w. Y
.6! -
‘8 \ 0'. '
Direct Coupling CH5,CH6
8 7 6 5 4

 

F2 (ppm)

67

From the information in Table 3, one can observe that the cross-peak
that represents a distance between the aromatic proton (H6/H8) and the sugar
proton H1 ’ within the same residue is less intense than the cross-peak
representing the same distance between consecutive residues. This
observation is what one would expect for the B form of DNA.

Furthermore, from the intemucleotide distances between H6 and H5
of the C bases, the distinction between C18 and C19 was possible, since in
the B form of DNA, the cross-peaks that correspond to this distance going
from 5’ to 3’ are more intense than those corresponding to the distance going
from 3’ to 5’. Therefore, only one cross-peak that connects these two base
protons is observed. From this starting point, the NOESY walk was
continued until both ends were located. The intensity pattern agrees for
most of the assignments with the intensity of the cross-peaks assigned as
intra or intemucleotide connections. The expanded version corresponding
to this region can be observed in Figure 34.

The assignment process was continued with the region that
corresponds to the connections between the sugar protons H1’ and H3’
(Figure 35). During these assignments the chemical shifts for each Hl’
protons were confirmed by examining the region already assigned. The

aromatic-aromatic region was assigned as well.

68

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

H1 F1 (ppm)

 

CCTCTGGTCTCC Strand
GGAGACCAGAGG Strand

 

 

 

Figure 34. Base-to-Hl ’ NOESY Walk

69

 

 

 

 

 

 

 

 

1--r

i

   
     

GZ3H3’,G24H8

 

 

:
s
1:”-
\ \ t‘
I?
. ‘I

i \g;

1.1,: C12H3’,C12H6
G1 I 3’.G13H8

.\V/, 1

L4 C1H3’.C1H6

 

 

 

 

 

 

 

I

I

l

7.9 7.7 7.5 7.3

H1

1

8.1

 

 

5305 4°90 4375 4f60
H1 F1 (ppm)

 

CCTCTGGTCTCC Strand

 

 

 

GGAGACCAGAGG Strand

 

 

Figure 35. H3’-to-Base NOESY Walk

(ppm)

F2

70

D. Reaction of the DNA Dodecanucleotide with Dirhodium Acetate

Complex

The metallated DNA oligomer was synthesized by reacting the
purified - GG- 12-mer with the compound [Rh2(OzCCH3)2(CH3CN)6][BF4]2
in a 1:1 ratio. The 1H NMR spectrum of this starting material in D20 is
presented in Figure 36. Once the purple solution of the compound is added
to the DNA sample, the color immediately changes to bright orange. The
reaction was continued by incubation in a thermocycle at 70 °C for 72 hours,
and small aliquots of the reaction mixture taken at regular intervals were
injected into an HPLC to monitor the progress of the reaction. This method
revealed that the reaction had gone to completion in ~72 hours. The color of
the solution after incubation is dark green. Once this product was purified,
the complementary strand was added to form the duplex.

The purification process was performed in the same manner as for the
native DNA oligonucleotide by HPLC chromatography using acetate buffers
in the following concentrations:

Eluant A : 0.02 M NaOAc , 20% CH3CN

Eluant B : 0.02 M NaOAC, 20% CH3CN , 1.2 M KCl

71

05 8 NEELAZUEBALEUNQNE .8 8.88% ”.22 a. 8n 2.5....

 

L

8.528 08.5 L

8.... . N m .. m e a w a o. :
1111111111112... 11; 11111111,; 1 1
L ,
8.8 L. .. 1.11

20.5 .8

 

0Q:

 

72

The gradient that was used is as follows:

 

Time(min) %A %B Flow Rate (mL/min)
O 100 O 4
5 7O 3O 4
65 6O 4O 4
70 O 100 4
75 O 100 4
80 100 O 4

 

 

 

The ionic exchange column was also used for this particular
separation, but the product did not elute from the column until eluant B was
in high percentage in the elution mixture (86%). This may be due to steric
factors that begin to play an important role because the oyerall shape of the
DNA strand is affected by metal binding. The detection of the adduct was
possible by using a diode-array UV visible detector. The DNA absorbance
was detected at 260 nm while the ha chromophore was detected at 365 nm.
The sensitivity of the Rh detector was increased ten times more than for the
DNA due to the low extintion coefficient for the metal-based transition (8 =
46 M'1 cm") at 365 nm compared to the extintion coefficient of DNA at 260

nm (8 =1.18 x 104 M'lcm’l). The Chromatogram is presented in Figure 37.

73

 

[ha(02CCH3)2(CHsCI\I)6]2+
Chromophore
365 nm . 0.1 AU

 

 

 

 

 

 

m A

DNA Chromophore
260 nm , 1.0 AU

 

 

ill

I
o 5 65 7075 80 t(min)

 

 

 

 

Figure 37. HPLC Chromatogram of ha-DNA Adduct

74

E. 1H NMR Analysis of ha-DNA Adduct

The NMR analysis of the dirhodium-DNA adduct was performed on a
sample whose concentration was 1.3 mM. Buffer solution in the same
concentration used for the unmetallated duplex (10 mM NazHPOJNaH2P04,
50 mM NaCl and 0.1 mM NaC2H6Ast)) was added to a final volume of
0.250 mL in a Shigemi tube.

The experiments were initiated by measuring 1D spectra of the imino
region in 90% H20 using the binomial method. Experiments at different
temperatures were performed to observe variations in chemical shifts or
resonance intensities as the duplex was denatured (Figure 38). By
examining the spectra it is obvious that after metallation and addition of the
complementary strand, the duplex has formed, since the imino protons are
observed. The spectra reveal some changes of the resonances as a result of
metallation, for example at 25 °C one of the A-T imino protons is no longer
observed. This proton corresponds to the A20-T3 base pair which is next to
the metallation site. Furthermore, the temperature range over which the
metallated duplex is stabilized is dramatically different as compared to the
native oligonucleotide. The duplex is practically denatured at 49 °C while in

the native duplex some resonances can still be observed at this temperature.

6 °C
(11 n ,
1,111 .111 01
1 1‘” 1*
J1 1 , 1 1
I' ‘ 11 ,1 1
I L 1
.// 1J1 / V 111
/ 1‘u/I 1‘\.., .V.,~m~. N
14.5 14.0 13.5 13.0 12.5 12.0 ppm
1
15 °C 1
[1 1111'1
1 1.1111
1111 11111 1|
/ ,. ' 1

MM/‘a/ me—N-

14.5 14.0 13.5

 

13.0 12.5 12.0 ppm
20°C ,1
1»
1 1
. l n
1 1.111211 1.
1111 1 11:1 1'11 11
f111 1 111 .1 1\
’1 «J 1’
WMIM’J; H f NJ \‘M‘M’AW-ritm
14.5 14.0 13.5 13.0 12.5 12.0 ppm
25 °c ,
1
1 1 1
I. 111 1
11 '11 1
11 1 1.
111 1 1“
V1 pf “Wu-1A“
“WWW - . 5W“ h . “15%
14.5 14.0 13.5 13.0 12.5 12.0 ppm

Figure 38 Imino Region of ha-DNA Adduct at Different Temperatures

75

30 °C

$2,

L~l
Wuflwyml‘ww LIVVQMIAAJLW 1%WW
14.5 14.0 13.5

13.0 12.5 12.0 ppm

40°C 1

' 1 . 1 . ,1"
11111111111111 11111111111111 1.111111r.1_1‘.4.-11 .

14.5 14.0 13.5

.1111'11113111111._ W. 1111 11111111111

13.0 12.5 12.0 ppm

45 °C

1 1 111

11111me
1: I ‘ l
kl 11111111111i111'11111.111111111I1’111111t111'fi11111 r1111lk‘111‘lfl W 11111111111111’111111111

14.5 14.0 13.5 13.0 12.5 12.0 ppm
49°C
1. 11
1111111111
11 \I'11'u11111r
1H1- 1 :11111 I! 1111 (1.11111 n 1'11
1fi1111f'1 11111 11 ‘1
111111111111 1 111111
14.5 14.0 13.5 13.0 12.5 12.0 ppm

76

A 1D 1H NMR spectrum of the metallated DNA product in 99% D20
was measured at 30 °C which was the temperature used for the native DNA
12-mer (Figure 39), and the residual HDO resonance was suppressed. A
very intense resonance at 1.8 ppm was observed which corresponds to the
acetate group of the metal compound.

Correlation time calculations were also performed following the same
procedure used for the unmetallated DNA 12-mer. The results are presented

in Table 4 with an average value of to of 1.10 i- 0.05 ns.

Table 4. ha-DNA Adduct : Relaxation and Calculated Correlation Times

 

for AH8 Protons
T1(S) T2(S) rcx10'9(s")
A20H8 l .787 0.1999 1.099
A15H8 1.882 0.1976 1.134
A17H8 1.664 0.1773 1.132
A22H8 1.612 0.2023 1.038

 

 

 

77

a
o

S
a

w

n

v m o ‘

m

m

Ema _

 

. I w p. ,y A .‘l \J
, i \ 4212/. .3: _u: __ ._ .

. ‘ , .l . _, J _ .

x I) 3 x 5 D .:_ /. 2 _\ . k f,“ a. r
fl/R _‘ /J *2) \i r. \ 5:: F... ..2.‘ .1.) , F
x. t 2 , K

‘ . \
K ~ r).

l,
v_

_ ‘
. ¥ ... — — _ 2 f
,. . c: ... ...,, _. J
. . H ,, fl
. A _, .._ 7, H c _
.. : . ..._ \ . x
, g ,7 I .

 

 

 

 

78

The 1D spectrum clearly shows the distortions experienced by the
DNA duplex upon metallation. Regarding chemical shifts, differences can
be observed in the entire spectrum. In the aromatic region, the resonances of
the C residues in the middle of the complementary strand have shifted
downfield for C19H6 and upfield for C18H6. The methyl groups of the T
residues are not observed, most likely because of shifting of these groups
and subsequent overlap with the methyl of the acetate groups. Most of the
resonances of the GO strand are not observed which implies strong
instability of the duplex as a consequence of metal binding.

Another important point is the correlation time value which is also
indicative of the severe effects on the overall conformation of the duplex
caused by the binding of the metal complex. This value has decreased
considerably compared to the native DNA oligonucleotide, implying a
distortion that favors a rapid tumbling of the molecule in solution. Although
the value of Tc has not been calculated for a large number of protons of the
biomolecule, it represents a measurable comparison between unmetallated
and metallated duplexes since we are taking into consideration the same

number and type of protons for the calculation.

79

After acquiring a 2D NOESY spectrum of this sample at the same
temperature, namely 30 °C, the instability of the adduct was continued
(Figure 40). Chemical exchange effects could be very strong if we consider
that according to the imino proton resonances, the melting or denaturing
point of the duplex is around 45 °C. Furthermore, by comparing 2D
NOESY spectra of metallated and non-metallated duplexes at the same
temperature, it is obvious that the line broadening has increased in such a
manner that the splitting of the resonances corresponding to C18H6 coupled
to C18H5 at 7.18 ppm has been eliminated due to line-broadening. The
same phenomenon is observed for the C19 residue as well.

It was decided at this point to decrease the temperature and to
measure the 2D spectrum at 20 0C. At this temperature, sharper and more
well-resolved resonances were obtained (Figure 41). In the 2D NOESY
spectrum measured at 20 °C, severe line broadening was also observed
(Figure 42). Focusing specifically on the base-to-Hl’ region of this
spectrum (Figure 43), the cross-peaks corresponding to the non-metallated
strand are clearly observed. However the metallated strand cross-peaks
coalesce in a region between 7.5 and 8.0 ppm. During chemical exchange, if
an ensemble of spins is precessing at a certain frequency, and then some of

the spins briefly exchange into another environment with a different

8O

 

 

 

 

392.
<5 '.
T

l

l—O
O o H
...m... ...mxar, m, , m _ O
O. ‘0! '. o. . N
...W. ' ?.‘ ‘ .. o
' O
W)
. 0
r3. " V
L, . o
.. / 0" O. m
8 /‘ b . o O
(O
O
-o O O N
o o c _,
..., -:-’ 3':-
. J .. - ”gas- 0
Z .' CD

 

 

 

l l

50 7.0 5:0 5:0 40 3:0 210 1.0
HI F1 (ppm)

Figure 40. ha-DNA Adduct NOESY Spectrum in 99% D20 at 30 °C

F2 (ppm)

 

H1

81

 

o. om a 05 $3 a 886on «22 m. .263 <zo-N§ .=. 95$...—

a 2

 

 

 

m
w A l Al Al r i l
\/ \llJ/
l%/ A... A. j \ A131.) .. ..sAW
AAA 3. A AA... A): AA A\ A AA AAA _AAAAA A___AAA AA
AAA 5 A. . AA AAA... AADADA AA. AAAAAAAA AAAA A.
..AAAAA A . AA A _
A A AA A
r A AAA/A AA (A AAA
AAA . AA.
A _
A AA
AAA AA AAAAA
AA AA A
A AA
AA’.
AA \sllllz/ \ i l i i will.» x
AA ,/.A W)\l 1A... \28/ A\)_ A\\ \).\
_A\ / , \ .AAA AAA
fA C AA AA} A.\ AA AAA AAAA AAAA AA AAA
A A. A AAA AA AAA AA
AA AA A AA. AA AA: _
A AAA AAAAAAAAA r
A AA, A _AA
AA; A
.AAAA
AAAAA

82

 

 

 

14.

9.0 8.0 7.0 6.0 5.0 4.0 3.0 2.0

110

l

l

j

l l l

l

 

 

9:0 8:0 7H36§t>5flo 4:0 3:0 2:0 1
HI Fl (ppm)

Figure 42. ha-DNA Adduct NOESY Spectrum at 20 °C

l

. 0

F2 (Ppm)

Hl

83

 

 

 

’43:}:

 

 

 

 

 

H8H

0H

 

 

 

 

 

 

g.

g
-“l

 

i‘k’ -‘

a 2'1!

 

Inge” 5

 

- 3‘:
- ”)3! 9:: -6 f
q ”A rq‘,. ._ _ P‘ :
* AL¢§LP 5
he?“ 5.29"} .A‘ Q "
G24l~l8,Hl’ “ >~

 

 

Figure 43. Base-to-Hl ’ Region of ha-DNA Adduct

84

frequency before returning to their original chemical shifi, they will be out
of phase with the spins that did not exchange. The accumulation of these
phase errors over time will lead to destructive interference in the FID and
loss of the signal. As a result, the FID decays rapidly and the Fourier
transformation produces a broad line. This is known as exchange
broadening. The greater the chemical exchange rate, the more frequently the
spin visits another site with different precession frequency, and the more
rapidly will be the loss of phase coherence and therefore the greater the line
broadening will be.

Several attempts to overcome the severe instability of the ha-DNA
adduct were performed. One of the parameters that was modified to remedy
this situation is temperature. Due to the fact that it was possible to obtain a
NOESY walk of the H1’-to-base region of one of the strands when the
temperature was lowered from 30 °C to 20 °C, it was decided to lower the
temperature even further, down to 5 °C. Focusing on the same H1 ’-to-base
region, it was observed that the breadth of the lines due to the decrease in the
correlation time of the molecule at this temperature negatively affected the
cross-peaks of the complimentary strand while causing only a small
improvement in the resolution of resonances corresponding to the metallated

strand.

85

A second factor that was modified is the ionic strength of the solution.
By increasing the salt concentration of the sample, the negative repulsion of
the phosphate backbone of the oligonucleotide is diminished. The objective
is to minimize the instability of the metallated DNA as much as possible.
By decreasing the electrostatic repulsion, it is possible to stabilize the adduct
such that a 2D spectrum with more defined cross-peaks may be obtained.

A buffer solution of 1M NazHPO4/NaH2PO4 (pH = 6.4) and 1M NaCl
was added to the D20 sample to obtain a final salt concentration of 50 mM
NazHPO4/NaH2PO4, 150 mM NaCl and 0.1 mM NaC2H6As02 at pH = 6.5 in
a final volume of 0.250 mL. It is important to note that by lowering the pH
of the solution, chemical exchange may be diminished as well.

A 2D NOESY spectrum in at 5 °C was measured under these new
buffer conditions. After data processing, the H1 ’-to-base region of the
spectrum showed the cross-peaks of both strands mixed together in the same
ill-defined region. We take this to mean that the degree of stability of the
sample was dramatically changed to the point of causing aggregation which
would lead to poor resolution of the 2D spectrum. By adding the same
buffer solution to a sample in 90% H20, the change in stability of the sample
was confirmed by observing the imino region at different temperatures

(Figure 44). The breadth of the resonances increased and, as a consequence,

86

the signal-to-noise ratio diminished; however, the stability of the metallated
duplex increased. This is clearly observed in the spectra. Even at 40 °C,
some of the imino proton resonances could still be observed.

From these aforementioned experiments, it was determined that 20 °C
is the best temperature for measuring the 2D NOESY spectrum. The
NOESY spectrum at 20 °C, however, revealed that the molecule is quite
unstable especially in the region of the metallated strand. This instability is
exacerbated by the decrease in signal-to-noise ratio, which we were able to
partially overcome by acquiring twice the number of scans. Unfortunatly
this only increased the signal—to-noise ratio by 40 percent. For this reason,
more assignments of the cross-peaks of the spectrum are not possible to
make at this time.

It is evident that further manipulation of the sample is required if we
are to obtain a better resolved spectrum with more dispersed cross-peaks in
both dimensions. The attempts that have been made to date lead us to
conclude that this will not be a straightforward task, and that additional
factors such as concentration and use of buffer solutions will have to be

taken into account.

5°C

fl
1
r I g
i . ' {l
f. i.
;' “ a.
[ .
r} l
g
/ «l M \
f l
. lwn F \l-A
l _ .
~wkm‘ufiw 1* ”M “flail-w “‘40 "‘1‘;

14.5 14.0 13.5 13.0 12.5 12.0 ppm

15 °C

,
.7) ‘
,1. .':
,"V V l
,1 l
g 'I
A ,I '
/ "‘1 M ‘
I l \ f ‘.
N” l r \
V)- . ‘
WWL‘Jx‘V W QK'IVI N‘WMI M”. 19"“; —'\ {k 1
14.5 14.0

13.5 13.0 12.5 12.0 ppm

20 °C '

.' H
.. ,1 W
\J V
r» l \
I l
f l j ,
If . I“. \\
("V V l f" V)‘
NM‘MWW _ MW" 'MJM‘WYW‘Y

14.5 14.0 13.5 13.0 12.5 12.0 ppm

25 °C

.1
:1
W!
IV"
R) lb!)
l l
,1 \
(l ) 1
1V! 0 \‘n
Wim‘i’fi V '..~~f". _,.___“Tv~"f¢ ."W
14.5 14.0 13.5 13.0 12.5

12.0 ppm

87

30 °C

7 1"
WMM'J/ )VA‘NM/ A- . .

14.5 14.0 13.5 13.0

40 °C ll

‘-
W"
‘J ¢.‘)V~‘A~1FVJI‘.N Y‘ k’kfl‘fhfi'fil‘fi ‘A‘VWFJV’WM

14.5 14.0 13.5 13.0125 12.0 ppm

45 °C

1.1"
H)
V
II) I
it 1'
v x .
l x. I

l i ‘N r N LN
M.-wt~.~i‘-m~rw‘»"*f*f .‘Ww-N, , . , _ .Wwwfi

14.5 14.0 13.5 13.0 12.5 12.0 ppm

49 °C

14‘
J» l H
A ‘ A I If “‘4 JY .-
W‘s"A‘i‘.’--i-’v~xt‘r’t7in:it";. .“‘. MW: “5*. , ... WMMH
14.5 14.0 13.5 13.0 12.5 12.0 ppm

Figure 44. Imino Region of ha-DNA Adduct in Buffer Solution

MM

12.5 12.0 ppm

Chapter 4
1. MOLECULAR MODELING STUDIES

The ultimate goal of 1H NMR spectroscopic measurements on large
biomolecules is the modeling of structural features that will satisfy the NMR
data as well as chemical information. The structure of the DNA double
helix varies according to its sequence and physical environment.
Intemuclear distances, torsion angles, bond lengths and bond angles are
characteristics of the molecule that provide important restraints for the
determination of its structure. Structural determination involves taking the
restraints together with the covalent structure and inputting into them a
computational algorithm that searches the conformational space. This
method provides a series of possible or feasible structures that satisfy the
assigned structural restraints.

Determination of DNA structures by 1H NMR spectroscopy is an
alternative to X-ray crystallography, and also provides valuable
complementary information about solution structure. The main difference
between these two techniques is that in X-ray crystallography typically only

one structure with a minimum energy in the solid state is obtained.

88

89

However, with NMR spectroscopy, according to the quality and number of
restraints that can be extracted from the multidimensional experiments, a set
of similar structures of varying energies are obtained as expected for
molecules in the solution state. New techniques and computational
programs are continually being developed in order to establish more
accurate restraints and, consequently, to narrow the set of possible structures
for the molecules.

As stated earlier, the intensity of the NOE cross-peaks is inversely
related to the distance between two nuclei. Therefore, an accurate
measurement of the volume of each of the cross-peaks will give rise to an
array of inter-proton distance restraints. Interestingly, the two protons that
give rise to a cross-peak are not alone in the molecule.» They belong to a
coupled network. However, assuming that each cross-peak is due to two
isolated nuclear spins is a good approximation; this is referred to as the two
spin or Isolated Spin Pair Approximation (ISPA) represented by the
following equation:37

r ab: r..r(a ref/a ab)1/6
where rah is the interproton distance between nuclei a and b that needs to be
determined, aab is the corresponding NOE cross-peak intensity and ad and

ab are the known interproton distance and cross-peak intensity respectively.

90

The reference distances can be obtained from fixed covalent distances
within the molecule under investigation. These distances can be related to
the cross-peak intensities in the NOE spectrum. However, it is important to
note that the mixing time (rm) of the NOE experiment plays a critical role in
the accuracy of the NOE measurements that can be obtained. A limitation
of ISPA is the omission of factors such as spin diffusion that may give rise
to cross-peaks of greater intensities than those expected, which would lead
to an inaccurate assessment of distances.

Structural information can be obtained by using torsion angle
restraints and NOE distance restraints. To obtain data of the former,
correlation spectroscopy (COSY) in any of its variations must be performed.
From these experiments, measurements of coupling constants provide
torsion angle restraints in the sugar deoxyribose rings 38 which consequently

help to define the conformation of the DNA sugar phosphate backbone

(Figure 45).

 

91

 

Figure 45. DNA Backbone Showing Torsion Angles

Regarding NOE distance restraints, simulated annealing (SA), which
is a form of restrained molecular dynamics, relates the experimental
structural restraints with energy considerations. In this approach, the
potential energy is calculated for a set of atomic coordinates using a force
field whose limits are determined by the values established for the restraint
bounds and standard covalent bond parameters (lengths, angles). The farther
the NOE distance restraints are from these bounds, the higher the potential
energy of the molecule and therefore, the greater the degree of instability of

the structure.

 

92

In simulated annealing, an initial structure is selected. This structure
is allowed to search conformational space. After that, the set of restraints,
(torsion and/or distance) is applied. These restraints will be responsible for
the final structure obtained. By repeating the (process several times, the
structures produced at the end, although not identical, should be very similar
with a small atomic root mean square deviation (RMSD) between them
(<1A). The different structures are averaged and a final structure is
reported. F urtherrnore, several structures can be considered as initial
structures, e.g., A and B-type DNA, with the expectation that similar
structures will be obtained if the restraint set is sufficient.

For the dodecanucleotide under study, the NOESY data obtained from
the spectrum measured at 30 °C gives rise to assignments of 718 NOE cross-
peaks from both sides of the diagonal. These assignments were introduced
in a table under FELIX 97.0 (Molecular Simulations Inc.) which generated
the volumes table that represents the intensity of each cross-peak. The
reference distance and cross-peak intensity were obtained from the C19
residue, specifically, Cl9:H5-C19zH6 covalent bond distance (2.464 A) and
C l9:H2’-Cl9:H2” distance (1.785 A). From these tables, the restraints file

was obtained with a total of 320 restraints calculated. Some of the restraints

 

93

were ignored due to overlapping problems that would leads to erroneous
distance information.

The restraints file was exported to the structural calculation program
InsightII (Molecular Simulation Inc.). Within InsightII, the simulated
annealing module was selected for the refinement calculations. Together
with the NOE distance restraints obtained from the NOE spectrum in D20,
some additional restraints regarding the imino-imino cross-peaks from the
NOESY spectrum in HZO were included. To enforce proper base pairing,
hydrogen bonding restraints were also used for all Watson-Crick base pairs
in the sequence. The starting structure selected was the sequence in standard
B form conformation. The set of restraints was displayed on this molecule
and this result is presented in Figure 46.

The simulated annealing protocol contains a Molecular Dynamics
schedule (MD_Schedule) edit command that provides a flexible and easy-
to-use interface for designing new protocols according to the requirements
of the molecule under study. For the DNA oligonucleotide, the restrained
molecular dynamics protocol begins with a minimization of the B-form
DNA duplex with gradual application of the restraints and the force
constants by 10 percent increments. The molecule was allowed to adopt

randomized velocities from the beginning of the protocol. The simulated

94

 

wuflabmvm + Ohio—Cum

 

fiafibmom 853me was oozSflQflOZ wagonm BBQ—Em <75 .3. 0.53m

33.38% 8:8me 28 ooqfimaéoz Baogm <75

 

N,“ V

C u—A‘.
b

‘7»-..3

‘0.!‘- I

9......“

a A A...

"‘2‘;

u A...

"‘1‘\.

'4“le

...4"

r'.’~'1.|

 

95

annealing Quartic system applies the non-bond repulsion energy values
gently. This first step in the protocol is followed by two dynamic processes
starting with a steepest descent minimization algorithm. The molecular
structure is heated from O to 300 K for 1000 ps under fully restrained
conditions. The second dynamics process involves a conjugate gradient
minimization algorithm that uses the cartesian coordinates of the atoms as
the space variables, keeping the temperature at 300 K. These conditions led
to sufficiently randomized starting geometries that, after a final conjugate
gradient minimization, gave rise to a structure that lies well within the
standard B-form DNA family. Some structural distortions are observed as
were expected, especially at the ends of the molecule. The final structure
obtained had a total energy of approximately 2202 Kcal. Energy increments
due to NOE distance violations were lower than 200 Kcal comprising less
than 10% of the overall energy of the molecular structure.

Figure 47 shows superimposed structures resulting from ten
simulations, each starting from a different trajectory. Convergence to an
atomic RMSD of 0.87 A was achieved using this protocol. The ten
structures overlapped with fairly good reproducibility. Most of the
differences among the structures are observed near the ends of the duplex.

This is due to the lower number of restraints in these regions of the

 

96

 

<m wEms AOO<O<UU<0<8vv AUUHUHOOHUHUUVU we mouBBEm cowogomsm .5. 0&3

8:58 com

 
 

97

molecule, and it reflects the greater freedom of motion in solution. It is clear
that the restraints for the ends of the molecule, which are the same
throughout the entire DNA duplex, are insufficient for defining the structure
at the ends with the same degree of reproducibility as for the, more stable,
middle part of the strand.

In conclusion, for this oligonucleotide, it was demonstrated that the
assignment of the majority of the non-exchangeable protons and of some of
the exchangeable protons were sufficient for the restrained molecular
dynamics refinement, starting from randomized models to produce
converged structures. Future studies will be directed towards the structural

determination of the ha-DNA adduct in solution by using this protocol.

 

CONCLUSIONS

Although major advances have been made regarding the mechanism
of action of the anticancer drug cisplatin, the molecular aspects of metal-
DNA covalent interactions are still not well understood. In the case of
antitumor active dirhodium carboxylate compounds, DNA binding studies
were initiated in the early 1970’s, but no conclusive data were obtained. Our
group has been the first to investigate the structures that these dinuclear
compounds form with DNA. Regarding this matter, the work described in
this thesis represents a promising first step. Structural studies of
unmetallated and metallated oligonucleotides are presently being performed
with good results.

lH NMR spectroscopic studies of the unmetallated DNA duplex were
very important to perform in order to clearly understand the behavior of the
oligonucleotide in solution. Duplex stability was determined by the
observation of the imino region and the type of conformation that this

sequence preferentially adopts by assigning the NOE cross-peaks of the two

98

 

 

99

dimensional spectrum and further application of the NOE distance restraints
to the molecule in a molecular modeling routine.

One of the most important aspects of such a study is the sequence,
which, in this case, consists of one purine-rich strand and one pyrimidine-
rich strand. From the base-stacking point of view, each strand exhibits a
different degree of stability since the pyrimidine base stacking is not as
efficient as the purine base stacking due to the absence of the imidazole ring
in the former. Although this issue is not important for the stability of the
unmetallated duplex, it indeed contributes to the deviations of this
oligonucleotide from the standard B-form DNA. This instability, however,
becomes quite evident in the metallated duplex, due to the fact that the
strand directly affected by the dirhodium carboxylate compound is the one
that is composed mostly of pyrimidine bases. This is most likely the reason
for the high degree of instability of the DNA duplex. The correlation time
measurements showed a difference between the unmetallated and metallated
duplexes values of 0.8 ns. This difference implies that the ha-DNA adduct
molecule tumbles faster in solution. A bent conformation may be a cause of
this difference, because it would allow for rapid motion. In spite of the
instability, it can be concluded that after the reaction and further annealing

of the complementary strand, the DNA oligonucleotide remains as a

 

100

permanent duplex as was observed in the imino region. The degree of
stability of the metallated duplex however, in comparison to the native DNA
duplex has decreased by about 20 0C. The formation of the ha-DNA
duplex demonstrated that dirhodium carboxylate compounds remain bound
to double stranded DNA at different temperatures, pH and salt
concentrations.

Further 1H NMR spectroscopic studies must be accomplished in order
to confirm these initial observations about the metallated sample. Sample
concentration is one of the parameters that is important for determining the
quality of the data. By increasing the concentration of the ha-DNA adduct
by 4 to 5 times, the problem of signal-to-noise ratio would be diminished
which may open up the possibility for the acquisition of a better spectrum
without the collection of many scans during each t1 increment. With respect
to the stability of the sample, the use of different stabilizing agents will be
the next variable to alter. It is known that spennine (N ,N’-bis[3-
Aminopropyl]-1,4-butanediamine) is a polar molecule that is very useful for
stabilization of DNA oligonucleotides in solution. The permitivity of the
sample will not greatly be altered since the ionic strength of the solution will
not be drastically increased as with the high salt concentration buffers used

during these first experiments. Furthermore, measurements of chemical

 

101

exchange rates of the imino protons with water molecules is a
straightforward experiment that will provide important information about the
stability of the base pairs of the metallated duplex, especially at the site of
the reaction site and adjacent base pairs nearby. .

A combination of the experiments described in this thesis may allow
for a two-dimensional spectrum of the quality required for further
assignment and determination of the structure of the dirhodium-DNA
adduct. The results will be compared to X-ray crystallographic studies that
are being performed at the present time. These studies represent the initial
step towards understanding, at the molecular level, the DNA binding of

dinuclear antitumor agents and may help us to design more effective drugs.

LIST OF REFERENCES

LIST OF REFERENCES

. Rosenberg, B.; Van Camp, L.; Krigas, T. Nature, 1965, 205, 698-699.

. (a) Lim, M. C.; Martin, R. B.; J. Inorg. Nucl. Chem. 1976, 38, 1911. (b)
Roberts J. J .; Thomson A. J. Prog. Nucleic Acid Res. Molec. Biol. 1979,
22, 71-133.

. (a) Harder, H.C.; Rosenberg B.; Int. J. Cancer, 1970, 6, 207. (b) Howle,
J. A.; Gale, G. R.; Biochem. Pharmacol. 1970, I9, 2757.

. Voet, D.; Voet, J. G.; Biochemistry, 1990, John Wiley & Sons. Inc.

. (a) Bancroft, D-P.; Lepre, C.A.; Lippard 8.]. J. Am. Chem. Soc. 1990,
112, 6860. (b) Fichtinger-Schepman, A. M. J .; van der Veer, J .L.; den
Hartog, J. H. J .; Lohman, P. H. M.; Reedijk, J. Biochemistry, 1985, 24,
707-713. (c) Sip, M.; Schwartz, A.; Vovelle, F.; Ptak, M.; Leng, M.
Biochemistry 1992, 31, 2508-2513. ((1) Shwartz, A.; Leng, M. J. Mol.
Biol. 1994, 236, 969-974.

. (a) Bellon, S. F.; Coleman, J. H.; Lippard, S. J. Biochemistry, 1991, 30,
8026-8035. (b) Brabec V.; Kleinwachter V.; Butour J-L.; Johnson N.
Biophysical Chemistry, 1990, 35, 129-141. (c) Rice J .A.; Crothers D.
M.; Pinto A. L.; Lippard, S. J. Proc. Natl. Acad. Sci. USA, 1988, 85,
4158-4161.

. Takahara. P.M.; Frederick, C. A.; Lippard, S. J. J. Am. Chem. Soc. 1996,
118,12309-12321.

. Corda, Y.; Anin M.-F.; Leng, M.; Job, D. Biochemistry, 1992, 31, 1904-
1908. (b) Anderson, K.S.; Mutat. Res. 1979, 67, 209-214. (c) Bradley,
L. J. N.; Yarema, K. J .; Lippard, S.J.; Essigmann, J. M. Biochemistry,

1993, 32, 982-988.

103

104

9. (a) Chao, C. C—K.; Huang, S.-L.; Lee, L-Y.; Lin-Chao, S. Biochem. J.
1991, 277, 875-880. (b) Calson, D.; Frit, D.; Salles, B. Nucleic Acids
Research, 1992, 20, 6363-6368. (c) Chu, G. J. Biol. Chem. 1994, 269,
787-790. ((1) Churchill, M. E. A.; Travers, A. A. Trends Biochem. Sci.
1994, 19, 185-187.

10. Grossched1,R.; Giese, K.; Pagel, J. Trends. (Genet. 1994, 10, 94-100

11. (a) Love, J .J .; Li, X.; Case, D. A.; Giese, K.; Grosschedl, R.; Wright,
RE. Nature, 1995, 376, 791-795. (b) Werner, M. H.; Huth, J. R.;
Gronenbom, A. M.; Clore, G. M. Cell, 1995, 81, 705-714.

12. Takahara. P.M.; Rosenzweig, A.C.; Frederick, C. A.; Lippard, S. J.
Nature, 1995, 377, 12309-12321.

13. Ohndorf U-M.; Whitehead, J. P.; Raju, N. L.; Lippard, S. J.
Biochemistry, 1997, 36, 14807-14815.

14. Pasco, J. M.; Roberts, J. J. Biochem. Pharmacol. 1974, 23, 1345.

15. Ciccarelli, R. B.; Solomon, M. J .; Varshavsky, A.; Lippard, S. J.
Biochemistry, 1985, 24, 7533.

16. (a) den Hartog, J. H. J .; Altona, C.; van Boom, J. H.; Marcelis, A. T. M.;
van der Marel, G. A.; Rinkl L. J .; Wille-Zazeleger, G.; Reedijk, J. Eur. J.
Biochem, 1985, 14 7, 371-379. (b) Gibson, D.; Lippard, S. J. Inorg.
Chem. 1987, 26, 2275-2279.

17. Vichi, P.; Coin, F.; Renaud, J-P.; Venneulen, W.; Hoeijmakers, J. H. J .;
Moras, D.; Egly, J-M. The EMBO Journal, 1997, I6, 24, 7444.

18. (a) Hambley, T. W. Coordination Chemistry Reviews, 1997, 166, 181-
223. (b) Howell, S. B., Ed.; Platinum and Other Metal Complexes in
Cancer Chemotherapy Plenum Press, New York, 1991. (c) Knox, R
Friedles, F.; Lydall, D.; Roberts, J. Cancer Res. 1986, 46, 1972. (d)
Blommaert, F. A.; van Kijt-Knijnenburg, H. C. M.; Dijt, F. J .; den
Engelse, L.; Baan, R. A.; Berends, F.; Fichtinger-Schepman, A. M. J.
Biochemistry, 1995, 32, 11676.

105

19. (a) Weiss, R.B.; Christian, M. C. Drugs, 1993, 46, 360. (b) Mckeage,
M. J.; Morgan, S. E.; Boxall, F. E.; Hard, G. C. Annals of Oncol. 1992,
111. (c) Mckeage, M. J .; Boxall, F. E.; Jones, M.; Harrap, K. R. Canc.
Res. 1994, 54, 1. (d) Casazza, A. M.; Rose, W. C.; Comereske, C.;
Oleson, R.; Fairchild, C. Proc. Am. Assoc. Canc. Res. 1992, 33, 536.
(e) Judson, 1.; Mckeage, M. J .; Raynaud, F.-; Hanwell, J .; Berry, C.;
Burrer, B.; Crabeels, D.; Harrap, K. Abstracts, 70' International
Symposium on Platinum and Other Metal Coordination Compounds in
Cancer Chemotherapy, 1995, 5054.

20. (a) Levi, F.; Misset, J. L.; Brienza, 8.; Adam, R.; Metzger, G.; Itzakhi,
M.; Caussanel, J. P.; Kunstlinger, F.; Lecouturier, S.; Descorps-Declere,
A.; Jasmin, C.; Bismuth, H.; Reinberg, A. Cancer, 1991, 69, 893. (b)
Extra, J. M.; Espie, M.; Calvo, F .; Ferme, C.; Mignot, L.; Marty, M.
Cancer Chemotherapy Reports, 1990, 25, 299.

21. Manzotti, C.; Pezzoni, G.; Giuliani, F .; Valsecchi, M.; Farrell, N.;
Tognella, S. Proc. AA CR, 1994, 35, 2628.

22. Bloemink, M. J .; Reedijk, J .; Farrell, N.; Qu, Y.; Stetsenko, A. I. J.
Chem. Soc. Chem. Commun. 1992, 1002.

23. (a) Zou, Y.; Van Houten, B.; Farrell, N.; Biochemistry, 1994, 33, 5404.
(b) Wu, P. K.; Qu, Y.; Van Houten, B.; Farrell, N. J. Inorg. Biochem.
1994, 54, 207. (c) Farrell, N.; Qu, Y.; Feng, L.; Van Houten, B.
Biochemistry, 1990a, 29, 9522.

24. Johnson, A.; Qu, Y.; Van Houten, B.; Farrell, N. Nucleic Acids Res.
1992, 20, 1697.

25. (a) Bear, J. L.; Gray, H. B. Jr.; Rainen, L.; Chang, I. M.; Howard, R.;
Serio, G.; Kimball, A. P. Cancer Chemother. Rept. 1975, 50, 611. (b)
Bear, J. L.; Erck, A.; Sherwood, E.; Kimball, A. P. Cancer Research,
1976, 36, 2204. (c) Rainen, L.; Howard, R. A.; Kimball, A. P.; Bear, J.
L. Inorg. Chem. 1975, Vol 14, 11, 2752.

26. Fimiani, V.; Ainis, T.; Cavallaro, A.; Piraino, P. J. Chemother. 1990, 2,
319.

106

27. Dunbar, K. R.; Matonic, J. H.; Saharan, V. P.; Crawford, C. A.;
Christou, G. J. Am. Chem. Soc. 1994, 116, 2201.

28. (a) Fichtinger-Schepman, A. M. J .; van der Veer, J. L.; den Hartog, J. H.
J.; Lohman, P. H. M.; Reedijk, J. Biochemistry, 1985, 24, 707. (b)
Kozelka, J.; Petsko, G. A.; Lippard, S. J. J. Am. Chem. Soc. 1985, 107,
4079. (c) Sherman, S. E.; Gibson, D.; Wang, A. H-J.; Lippard, S. J.
Science, 1985, 230, 412. (d) Bellon, S. F .; Coleman, J. H.; Lippard, S. J.
Biochemistry, 1991, 30, 8026. (e) Van Hemelryck B.; Guittet, E.;
Chottard, G.; Girault, J-P.; Huynh-Dinh, T.; Lallemand J-Y.; Igolen, J .;
Chottard, J-C. J. Am. Chem. Soc. 1984, 106, 3037. (f) Bemers-Price, S.
J.; Ranford, J. D.; Sadler, P. J. Inorg. Chem. 1994, 33, 5842.

29. (a) den Hartog, J. H. J .; Altona, C.; van Boom, J. H.; van der Marel, G.
A.; Haasnoot, D. A. G.; Reedijk, J. J. Am. Chem. Soc. 1984, 106, 1528.
(b) Parnham, K.; Bemers-Price, S. Angew. Chem. Int. Ed. Engl. 1995,
34, 17. (c) Huang, H.; Zhu, L.; Reid, F. R.; Drobny, G. R; Hopkins, P.
B. Science, 1995, 270, 1842. (d) den Hartog, J. H. J .; Altona, C.; van den
Elst, H.; van der Marel G. A.; Reedijk, J. Inorg. Chem. 1985, 24, 986.
(e) Iwamoto, M.; Mukundan, S. Jr.; Marzilli, L. G. J. Am. Chem. Soc.
1994, 116, 6238. (f) den Hartog, J. H. J.; Altona, C.; van Boom, J. H.;
van der Marel, G. A.; Haasnoot, D. A. G.; Reedijk, J. J. Biomol. Struct.
Dynm. 1985, Vol. 2, 6, 1137. .

30. Yang, D.; van Boom, S. G. E.; Reedijk, J.; van Boom, J. H.; Wang, A. J.
H. Biochemistry, 1995, 34, 12912.

31. (a) Callihan, D.; West, J .; Kumar, S.; Schweitzer, B. 1.; Logan, T. M. J.
Magn. Reson. 1996B, 112, 82. (b) Piotto, M.; Sandek, V.; Sklenar, V.
J. Biomol. NMR, 1992, 2, 661. (c) Sklenar, V.; Piotto, M.; Leppik, R.;
Sandlek, V. J. Magn. Reson. 1993A, 102, 244.

32. Bishop, K. D.; Borer, P. N.; Pelczer, I. J. Magn. Reson. 1996B, 110, 9.

33. Wuthrich, K. NMR of Proteins and Nucleic Acids, 1986, John Wiley and
Sons. Inc. USA

34. Suzuki, E-I.; Pattabiraman, N.; Zon, G.; James, T. L. Biochemistry,
1986, 25, 6854.

107

35. (a) Broido, M. S.; Zon, G.; James, T. L. Biochem. Biophys. Res.
Commun. 1984, 119, 663. (b) Feigon, J .; Denny, W. A.; Leupin, W.;
Keams, D. R. Biochemistry, 1983, 22, 5930. (c) Scheek, R. M.; Russo,
N.; Boelens, R.; Kaptein, R.; van Boom, J. H. J. Am. Chem. Soc. 1983,

105, 2914.

36. Croasman, W. R.; Carlson, R. M. K. Two Dimensional NMR
Spectroscopy. Applications for Chemists and Biochemists, 1984, VCH

Publishers Inc.

37. James, T. L.; Bishop, K. D.; Blocker, F.; Gonzalez, C.; Kumar, A.; Liu,
H.; Mujeeb, A.; Schmitz, U.; Ulyanov, N. B.; Yuan, Y-C.; Weisz, K.

NATO ASI Series, 1994, 87, 335.

38. a) Rinkel, L. J.; Altona, C. J. Biomol. Struct. Dyn. 1987, 4, 621.
b) Weisz, K.; Shafer, R. H.; Egan, W.; James, T. L. Biochemistry, 1992,

31, 7477.

"I11111111111111“