f H, i1” [5

Llichiga'n C112? '
Umvcrsriy F

q man-:- a“ rmsqn 'IIR'V~ .3- $17"

  

This is to certify that the
thesis entitled
SODIUM S'-GUANOSINE MONOPHOSPHATE AGGREGATION:
STOICHIOMETRY, MODES OF SPECIFIC SODIUM ION
COMPLEXATION, AND BINDING T0 ETHIDIUM

presented by

CHRISTOPHER L . MARSHALL

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Chemistry

 

  

442726”; ‘Zér ¢__,
Maj professor

Date flflZ/yfl

0-7639

    
   

    
 

V |i'

Llois Il-J
'\ .un .

21.4563. 5

'LI flaw!" .'

 
   
 

 

mm}: 'nues:
25¢ per day bar it.

mymgus' mam MATERIALS:

'Nacc in'book return to renew
chum from circulation records

 

SODIUM 5"GUANOSINE MONOPHOSPHATE AGGREGATION:
STOICHIOMETRY, MODES OF SPECIFIC SODIUM ION
COMPLEXATION, AND BINDING TO ETHIDIUM

BY

Christopher L. Marshall

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1980

L!

0/

M/é‘V ‘ "

ABSTRACT

SODIUM 5'-GUANOSINE MONOPHOSPHATE AGGREGATION:
STOICHIOMETRY, MODES OF SPECIFIC SODIUM ION
COMPLEXATION, AND BINDING TO ETHIDIUM

BY
Christopher L. Marshall

The dianion of guanosine-S'-monophosphate (5'-GMP=)
forms regular structures in aqueous solution that are slow
to exchange on the 1H NMR.time scale. The self-assembly
process is dramatically dependent on the nature of the
alkali metal counterion, which is believed to direct struc-
ture formation through a size—selective coordination
mechanism. Previous work has not been successful in deter-
mining the exact metal ion and nucleotide stoichiometries
of the aggregated units, in part because the metal to
nucleotide ratio in the pure salts is fixed at 2:1.

The discovery that the tetramethyl ammonium ion (TMA+)
acts as neither a structure directing nor inhibiting ion
has facilitated the study of S'—GMP= self-assembly at
varying ratios of metal to nucleotide. The objective of
this thesis was to determine the stoichiometry of the Na+/

5'-GMP= aggregates. The present study examined the con-

centration dependence of Na+/5'-GMP= self-aggregates at

Christopher L. Marshall

constant free Na+ ion concentration using 1H NMR. Several
equilibrium.models were investigated as possible fits to
the experimental data using a computer curve-fitting rou-
tine. The models selected were: (1) monomer in equili-
brium with aggregated n-mer, (2) monomer in equilibrium
with stacked tetramers, (3) monomer in equilibrium with
tetramer and tetramer in equilibrium with stacked tetramer,
and (4) tetramer in equilibrium with stacked tetramer. On
the basis of closeness of fit to the experimental data and
symmetry arguments, an equilibrium involving monomer
5'-GMP= aggregating to form two stacked tetramer units is
preferred.

A second series of experiments examined the Na+ depen-
dence of the aggregation process. An equation which
assumes an octameric nucleotide stoichiometry was fit to
the experimental data. Closest fit of this equation to
the experimental data was for the Na+ stoichiometry with a
value x = 4.1(i0.1) which results in an overéall stoichio—
metry for the self-structures of Na4(5'-GMP)%2-.

Model building studies in conjunction with symmetry
considerations show that the stacking of two tetramers can
explain the three line H(8) 1H NMR spectrum characteristic
of the ordered form. The directional character of the
hydrogen bonding of the tetramer allows two distinctly

different stacking patterns upon octamer formation.

Christopher L. Marshall

Stacking of two tetramers with the same hydrogen bond
directionality creates an octamer with C4 symmetry which
is believed responsible for the appearance of the outer
H(8) structure lines. The inner H(8) structure line is
believed to arise from one of the diastereomers with D4
symmetry created from the stacking of two tetrameric
plates with opposite hydrogen bond directionality.
Specific Na+ coordination and interplatelet hydrogen
bonding are also believed to be important factors in the
stabilization of the structures. The four Na+ ions are
most likely coordinated in two different positions. Two
Na+ ions are placed in the cavity defined by the four 0(6)
donors of the tetramer unit. A water molecule may also be
coordinated to the Na+ at this position. The remaining
two Na+ ions are believed to be chelated by phosphates on
the tOp and bottom tetramer units. Between four (OH(2')-OP

or OH(3')-OP) and eight (NH(2)-OP) interplate hydrogen

 

bonds are possible.

The intercalation drug ethidium bromide (EtdBr) was
found to interact very strongly with structured forms of
both Na25'-GMP and K25'-GMP. Evidence provided by 1H NMR
chemical shift changes indicates that the octamers of
Na25'—GMP form a strong 2:1 (EtdBr:octamer) structure and

a slightly weaker 1:1 complex. The stacking of two and one

drug molecules respectively to the exterior aromatic faces

Christopher L. Marshall

is postulated. The stoichiometry of the structure:drug
complex supports the over-all stoichiometry of the aggre-
gated nucleotides.

The drug appears to break up the proposed hexadeca-
meric structures of K25'-GMP and to stack in a 2:1 and 1:1
fashion with the octameric fragments. As in the sodium

case, the 2:1 complex is stronger than the 1:1 complex.

To Susan

9.
P.

ACKNOWLEDGEMENTS

It is hard in a page to properly thank all of those
who have been of help over the past few years of graduate
school. Those listed below were especially helpful through-
out; for those not mentioned, a general thank you is ex-
tended.

First, I would like to thank the M.S.U. Chemistry
Department for both financial support and the use of the
facilities. Financial support received through the
National Institutes of Health is gratefully acknowledged.

I would also like to thank Dr. Carl H. Brubaker for acting
as my second reader. Assistance received from Dr. James
L. Dye and Mr. John Leckey with the KINFIT program, Ms.
Kathryn Nyland with figures, and Dr. Thomas Atkinson with
computer graphics is noted. Dr. Alexander I. Popov is
thanked for both serving on my guidance committee and for
his assistance with the ethidium structure fitting.

The help and love of my parents, Kay and Lyle Marshall,
and my wife's parents, Betty and Bob Bolda, has always been
a source of guidance for me. Their encouragement of my
graduate career helped make the going a little easier.

Dr. Thomas J. Pinnavaia has been a continual source of

guidance, wisdom, and encouragement, for which I am

iii

sincerely thankful. He is an able and patient teacher to
whom I shall always be indebted. I would also like to
thank Mr. Rasik Raythatha, Mr. Richard Barr, Mr. Steven
Christiano, and Dr. Elene Bouhoutsos-Brown whose collabora-
tion and fruitful discussions (often late into the night)
were extremely helpful. Dr. Bouhoutsos-Brown is also
thanked for providing the spectra used in the sodium ion
stoichiometry study.

It is not a simple task to thank my wife for all the
help that she has always been to me. Susan not only did
the typing and graphics that made this thesis possible;
her love and encouragement always provided the driving force
to go on when I felt down. Her faith, patience, and love

for me cannot be compared to any other source of help.

iv

TABLE OF CONTENTS

INTRODUCTION
General
Guanine Nucleotides
Intercalation Drugs
Statement of Research Aims
EXPERIMENTAL
A. Materials, Techniques, and Preparation
B. Instrumentation
Nuclear Magnetic Resonance Spectroscopy

Ultraviolet/Visible Absorption
Spectroscopy

Computation
C. Theoretical
GMP= Stoichiometry
Na+ Stoichiometry
Drug Binding Constants
RESULTS

A. Stoichiometry of 5'-GMP= Aggregation:
Na+ Salt

B. Model Building Investigation of Octameric
Structures

C. Metal Ion Stoichiometry: Na+ Salt

Page

13
17

19
20

21
21
22
22
28
30

35

45
62

D. Interaction of Intercalation Compounds
with Structured Na25'-GMP and K25'-GMP
E. Interaction of Ethidium Bromide with
Structured and Unstructured Forms of
Na25'-GMP
F. Interaction of Ethidium Bromide with
Structured and Unstructured Forms of
K25'-GMP
DISCUSSION
A. Present Study
B. Other Proposed Structures
C. Stability of the Base Overlap
D. Drug Intercalation
CONCLUSION
BIBLIOGRAPHY

vi

67

69

82

91
100
111
114
121
124

Table

Table

Table

Table

Table

Table

Table

2.

LIST OF TABLES

Fraction of Structure of 5'-GMP= with
Changing Total Nucleotide Concentration

Results of Stoichiometry Fitting of
5'-GMP= in NaCl

Results of Curve Fitting Equation
Equation (la) to the Data in Table 1
with n Held to a Constant Value

Number of Isomers and H(8) Lines
Expected for Three Exchange Regions

Torsional Angles and Notations for
5'-Nucleotides

Results of the Model Building Study
of the Octameric Forms of 5'-GMP

Fraction of Structure of 5'-GMP=
with Changing 5'-GMP"/Na+

vii

Page
38

41

46

58

59

61

64

Figure

Figure

Figure

Figure

Figure

LIST OF FIGURES

a) Structure and numbering sequence
for pyrimidine and purine.

b) Structure and numbering sequence
for B-D-ribose (RNA) and 2-deoxy-
B-D-ribose (DNA).

Ring current effects in aromatic
systems.

a) Structure of 5'-guanylic acid

(5 ' -GMP) .
b) Structure of 3'-guany1ic acid
(3'-GMP).

Proposed structure of the hydrogen
bonded tetramer in the Na2(5'-GMP)
solution and NaH(3'-GMP) gel (R =
ribose-phosphate group).3 '37

Schematic diagram of the preposed
size selective coordination sites
in the structuring of a neutral
aqueous solution of 5'-GMP. a)
Coordination into the center of
the etrameric plate by Na+ (r =
0.99 ). b) Coordination between
the lates by + and Rb+ (r =
1.51 and 1.60 respectévely).
Note that Li+ (r = 0 59 for
4-coordination, 0.74% for 6-
coordination) would be too small
to substitute for Na+ and that
Cs+ (r = 1.78g) is too large to
replace K+ or Rb+. (Ionic radii
from ref. 51)

viii

Page

12

Figure 6.

Figure 7.

Figure 8.

Figure 9.

Figure 10.

Figure 11.

Schematic of the mode of inter- 14
calation binding to nucleic acids.

a) Normal double stranded poly-

nucleotide with ribose-phosphate

backbone (striped) and hydrogen

bonded base pairs. b) Same

sketch as in a) with three drug

molecules (shaded region) inter-

calated.

Chemical structures of a) actino- 15
mycin D; b) ethidium bromide; c)

daunorubicin (also called dauno-

mycin) and andrianmycin; and d)

platinum (II) 2,2',2"-terpyridyl

chloride.

Effect on the H(8) NMR of adding 36
(TMA)25'—GMP into 0.85M_NaCl in

D20 at 15°C. Structure lines are

labelled A, B, and Q, and the

monomer line is labelled g.

Plots of the computed curves for 44
equation (la) when a) n and K' are

both varied, and b) K' is varied

and n is held constant at either

8, 12, or 16. Experimental condi-

tions are the same as in Table 1.

Error bars represent the experi-

mental points. Errors in n and K'

are reported at the 95% confidence

level.

Schematic diagram of a) normal stacking 48
and b) the two enantiomeric reversed

stacking modes of the octamers of

5 ' -GMP .

Overlap pattern of the guanine bases 51
for the two enantiomeric forms in the

normal stacking mode of two tetramers.

The dark tetramer is the top layer and

the light tetramer is the bottom layer.

a) C4(CW,CW,+30°); b) C4(CW,CW,-30°).

a) is the enantiomer of b) in the

absence of the chiral ribose-phosphate

group.

ix

Figure

Figure

Figure

Figure

Figure

Figure

Figure

12.

13.

14.

15.
16.

17.

18.

Overlap pattern of the guanine bases 54
for two of the diastereomeric forms

in the reversed stacking mode of two
tetramers. The dark tetramer is the

top layer and the light tetramer is

the bottom layer. a) D4(CW,CCW,+30°);

b) D4(CW,CCW,+60°).

Overlap pattern of the guanine bases 56
for the remaining two diastereomeric
forms in the reversed stacking modes
of two tetramers. The dark tetramer
is the t0p layer and the light tetra-
mer is the bottom layer. a)
n4(ccw,cw,-600); b) D4(CCW,CW,-30°).
In the absence of the chiral ribose
phosphate group, Figure 13a is the
enantiomer of Figure 12b, and Figure
13b is the enantiomer of Figure 12a.

Plots of the best fits for equation 66
(5a) with a) the values of x and K

allowed to vary and b) x held con-

stand at 6 and K allowed to vary.

Experimental conditions are the same

as in Table 7. Error bars represent

the experimental points. Errors in

x and K are reported at the 95%

confidence level.

Structure of hemat0porphyrin. 67

Effect on the 1H NMR of adding 71
Na25'-GMP into 0.034M EtdBr at

5°C. Nucleotide H(8T resonances

are denoted by shading.

Effect of the fraction of NaZS'GMP 76
structured in the presence ([3) and

in the absence (A) of EtdBr. Experi—

mental conditions are the same as in

Figure 16.

Plot of the chemical shift of the 77
EtdBr methyl resonance as a function

of the structured Na25'-GMP/Drug

ratio. Extrapolation of the linear

portions of the curve included.

Figure

Figure

Figure

Figure

Figure

Figure

Figure

19.

20.

21.

22.

23.

24.

Plot of the best fit curve for
equations (6a) and (6c). 60 s is
the observed chemical shift in ppm
of the EtdBr methyl resonance and
Mg/D is the ratio of the structured
Na25'-GMP to total drug. X's repre-
sent experimental points. Errors in
K's and 6's are standard deviations.

Plots of the best fits for equations
(6a) and (6c) in which K1 is held
constant at a) 6.4 x 103 and b)

6.4 x 107 and only Kg, 61 and 62 are
allowed to vary. Legends are the
same as in Figure 19. Errors in

K2 and 6's are standard deviations.

Plot of the best fit for equations
(6a) and (6c) in which 6 and 62
are held constant at 0.6 and 0.87
ppm and K1 and K2 are allowed to
vary. Legends the same as in
Figure 19. Errors in K1 and K2
are standard deviations.

Effect on the 1H NMR of addingo
K25'-GMP into 0.29M_EtdBr at 5 C.
Nucleotide H(8) resonances denoted
by shading.

Plot of the chemical shift of the
EtdBr methyl resonance as a function
of the structured K25'-GMP/Drug
ratio. Extrapolation of the linear
portions of the curve included.

Plots of the best fits for a) equa-
tions (6a) and (6c) and b) equations
(10) and (11). 5obs is the observed
chemical shift in ppm of the EtdBr
methyl resonance and a) M3/D or b)
M1 /D is the ratio of structured
Kzg'-GMP to total drug. Errors are
standard deviations.

Schematic diagram of the proposed

structure for the self-assembled
forms of Na25'-GMP.

xi

80

81

83

85

88

90

101

Figure

Figure

Figure

Figure

Figure

26.

27.

28.

29.

30.

Structures proposed by Bprzo and
Laszlo46 to explain the H NMR spec-
trum of the self-assembled forms of
Na25'-GMP. Open circles represent
phosphate groups pointing down while
closed circles represent phosphate
groups pointing up. Na ions repre—
sented by striped circles (from ref.
46).

Structure proposed by Detellier and
Laszlo9 3to explain the NMR results
for the K /Na+ -5'-GMP= self-structure.
Only two of the four chelated Na+

ions are illustrated (from ref. 98).

Expected up ield shift (in ppm) of a
nucleus 3.5 above or below the
aromatic plane due to the ring
current and diamagnetic suscepti-
bility effects (from ref. 82).

Schematic diagram of the proposed
equilibrium involved in the formation
of a 2:1 and 1:1 complex between the
intercalation drug EtdBr and the
octameric forms of Na25'-GMP.

Schematic diagram of the proposed
equilibrium involved in the formation
of a 2:1 and 1:1 complex between the
intercalation drug EtdBr and struc-
tured forms of K25'-GMP.

xii

103

107

113

117

120

INTRODUCTION

General

Since the discovery of nucleic acids in the late nine-
teenth century1 the study of their chemical and biochemical
properties has become a highly active area of research. In
1978 alone, over 2,000 separate listings were indexed under
the general heading of "nucleic acids."2 As the building
blocks in the formation of ribonucleic acids (RNA's) and
deoxyribonucleic acids (DNA's), mononucleotides are of
fundamental importance in the molecular mechanisms by
which genetic information is stored, replicated, and trans-
cribed throughout the cell. They have also been shown to
participate in intermediary metabolism, to act as factors
during certain enzymatic processes, and to take part in
energy transforming reactions in biological systems.
Naturally occurring mononucleotides are composed of three
characteristic components: 1) a nitrogenous aromatic base,
2) a five carbon ribose sugar, and 3) a phosphoric acid
group.3

The nitrogenous bases are aromatic heterocyclic com-
pounds which are classified as either pyrimidines (six
membered rings) or purines (pyrimidines with fused imi4

dazole rings) (Figure la). The two major naturally

I

a)

   

N?
K,
N

PYRIMIDINE PURINE

 

   

OH OH OH H

B'D-RIBOSE 2-DEOXY-B'D-RIBOSE

Figure l. a) Structure and numbering sequence for

pyrimidine and purine.
b) Structure and numbering sequence for

B-D-ribose (RNA) and 2-deoxy-B-D-ribose
(DNA).

occurring purines, adenine (6-amin0purine) and guanine
(2-amino-6-oxypurine) occur in both ribonucleic acids and
deoxy-ribonucleic acids. There are three major naturally
occurring pyrimidines: thymine (5-methy1-2,4-dioxypyri-
midine), which occurs only in DNA, uracil (2,4-dioxypyri-
midine), which occurs only in RNA, and cytosine (2-oxy-4-
aminopyrimidine), which occurs in both forms of cellular
polynucleotides.

Nucleosides are N-glycosides of the pyrimidine or
purine, in which carbon 1 of a B-D-ribose sugar is linked
to N(l) of a pyrimidine or N(9) of a purine (Figure lb).
In RNA the sugar is a simple B-D-ribose whereas in DNA the
sugar is 2-deoxy-B-D-ribose. Naturally occurring nucleo-
sides always have the glycosidic linkage in the B epimeric
form and the pentose is always found in the cyclic furanose
form. The most common ribonucleosides; adenosine, guano-
sine, cytidine, thymidine, and uridine are more water
soluble than their parent bases.

Nucleotides are phosphoric acid esters of nucleosides
in which one or more mono-, di-, or tri-phosphoric acids
are esterified to hydroxyl groups of the pentose. Although
these phosphate groups may be attached at the 2', 3', or
5' hydroxyl positions, the 5'-isomers are by far the most
abundant in the cell. Mononucleotides (nucleotides con—
taining only a single base) are strong acids with two

dissociable protons on the phosphoric acid group having

pKa values of approximately 1.0 and 6.2. Therefore, at
neutral pH the free mononucleotide is completely
dissociated.

Mononucleotides are able to form hydrogen bonded com-
plexes between complementary base pairs as first prOposed
by Watson and Crick for the double stranded helix formation

of DNA.4'5

This base pairing is highly specific and has
been studied for nucleosides in non-aqueous solvents by
infrared spectroscopy.6'-9 In both aqueous and non-aqueous
solution, NMR of nucleosides has also shown evidence for
complementary hydrogen bonded base pairing by the observa-
tion of chemical shift changes in both amino and aromatic

ring protons.]'0"13

The hydrogen bonded amino protons
exhibit both concentration and temperature dependent down-
field chemical shifts in comparison to those of the
monomer.

While hydrogen bonding between base pairs is signifi-
cant, base stacking effects have also been shown to pro-

14

vide ordered stabilization. Stacking is the vertical

aligning of the planar aromatic bases one atop the other.
The stabilizing effect of base stacking is usually des-

cribed in terms of hydrophobic forces,14 special van der

15

Waals interactions,15 n-electron polarizability, dipole-

dipole induced forces,16 or electrostatically stabilized

ionic (charge transfer) complexes.15

It is important to note that base stacking has been

observed only in aqueous solutions.17

This dependence on
the solvent indicates that the hydrophobic description of
base stacking is appropriate. It is suggested17 that
stacking is a demonstration of the ordering of water mole-
cules around the dissolved solute allowing short-ranged,
attractive forces within the bases to become operative.18
Proton magnetic resonance (1H NMR) is very useful for
determining the stacking tendencies of various aromatic
molecules. Because of the mobile n electrons in an aro-
matic system, a large diamagnetic current is induced in
the plane of the ring by an external magnetic field. This
induced ring current gives rise to a small secondary mag-
netic field which reinforces the primary field at positions
outside and in the plane of the ring. However, at posi—

tions directly above and below the plane of the ring, the

primary field is reduced (Figure 2). Therefore, the

 

 

 

Figure 2. Ring current effects in aromatic systems.

chemical shifts of protons for molecules which stack ver-
tically above the aromatic plane should appear shielded
from their normal positions. This interaction should
easily be distinguished from hydrogen bonding which induces
a downfield shift.

19 have shown that the chemical shifts of

Chan, 33 a1,
aromatic protons of purine and 6-methylpurine in aqueous
solution move significantly to higher field as the concen-
tration is increased. However the same solutes in either
dimethyl sulfoxide or dimethylformamide exhibit aromatic
chemical shifts which are virtually independent of con-
centration. This change was therefore attributed to
hydrOphobic stacking of the aromatic rings. Ts'o and his
coworkers showed a similar upfield shift of aromatic pro-
tons as a function of concentration for aqueous solutions

20,21 20,21

of nucleic acid bases, nucleosides, and mono-

22 They found that base stacking was stronger

nucleotides.
for purine bases than for pyrimidine bases and the ring
current effects of the latter are not very substantial.
Furthermore, they reported that the nucleosides had a far
greater tendency to form vertical stacks than the corres-

ponding mononucleotides which carry some negative charge

on the phosphate moiety.

Guanine Nucleotides

 

Derivatives of the mononucleotide, guanylic acid,

have shown the unique ability to spontaneously form

regular ordered structures in aqueous solution. This
ability has not been observed to date for any other mono-
nucleotide component.

As early as 1910, Bang, in work involved with iso-

lating and characterizing guanine nucleic acids, reported

that acidic solutions of guanylic acid formed clear gels.23

This phenomenon has been observed in aqueous solutions of
both the nucleoside and the mononucleotide derivatives.

Several authors have investigated by infrared spectrose

c0py,24-26 optical rotatory dispersion (0RD),27'28

27'28 Ihman spectroscopy,29-31

metry,32 and x-ray fiber diffraction-53'-36 the structure of

circu-
lar dichroism (CD), calori-
these gels, which are formed at pH S 5. Investigation

of the fibrillar structures upon drying of the gels indi-
cates that,at pH = 5, two different hydrogen bonded self-
structured forms exist. One of these structures, repre-
sented by fibers of sodium 5'-guanosine monOphosphate
(NaH(5'-GMP)) (Figure 3a) at pH = 5, consists of a long
chain, single stranded continuous helix with 15 nucleo-

36 On the other

tide units for every four helical turns.
hand, Na2(5'-GMP) fibers grown from concentrated solution
at neutral pH in the presence of NaCl consist of hydrogen
bonded planar tetramer units stacked in a helical array
(Figure 4).37 The stacked planar tetramer structure has
also been found for NaH(3'-GMP) fibers (Figures 3b and

33

4). In both the helical and the stacked tetramer

 

O O
N
// N-H
H l I
ll
HO-T-o-H C 0
OH H H
H )4
O OH
a) b)

Figure 3. a) Structure of 5'-guanylic acid (5'-GMP).
b) Structure of 3'-guanylic acid (3'-GMP).

pr0posed structures, nitrogen (l) and nitrogen (2) are
hydrogen bond donors, while oxygen (6) and nitrogen (7)
act as hydrogen bond acceptors.

In 1972, Miles and Frazier reported IR evidence that

5'-GMP in neutral or slightly alkaline solution (pH = 8)

38

can exist in a regular ordered form. In this pH range,

both acidic protons on the phosphate group are dissociated

(pK = 6.1).39 They reported that in going from the

a2
unstructured to the structured form, three major IR

changes occur. First, the carbonyl stretch at 1665 cm"1

 

Figure 4. Proposed structure of the hydrogen bonded
tetramer in the Na2(5'-GMP) solution and
NaH(3'- 3 gel (R = ribose-phosphate
group).33r 7

10

1

moves to 1672 cm- and intensifies. Second, the

C(4)=C(5) ring vibration at 1580 cm-1 decreases in inten-

sity and shifts to 1593 cm-1. Finally, a new ring vibra-

tion at 1540 cm-1 was found to grow in. This IR pattern
differs from that of the structured form of NaH(5'-GMP).
The neutral solution also shows no evidence for gelation
as is observed in the acidic solution. Based on the simi-

larities between the IR spectra of 3'-GMP in acidic media24

and 5'-GMP in neutral or slightly alkaline media,38

Miles
and Frazier postulated that the structure for 5'-GMP= in
this medium is based on the same stacked planar tetramer
units (Figure 4). Zimmerman's 1976 x-ray fiber diffraction
study37 supported this assumption of stacked planar tetra—
mer units. His work indicated that the planar tetramer
units were stacked on a common axis at a distance of 3.43
with each unit rotated 300 with respect to the preceding
unit.

A 1H NMR study of Na2(5'-GMP) by Pinnavaia, Miles,
and Becker40 showed that the structures of Na2(5'-GMP)
were slow to exchange and had an energy barrier to exchange
with monomers of greater than 15 kcal-mol-l. Previously
all H-bonded mononucleotides, including those with
Watson—Crick complementarity, were reported as exhibiting
only time average 1H NMR spectra. Subsequent 1H NMR work
showed that structure formation is highly dependent upon

41-43

the alkali metal counter-ion. These studies showed

11

that structure formation was based on a size selective
coordination of the alkali metal cation. The thermal
stability of the ordered forms decreased in the order

K+ > Na+ 2 Rb+ >> Li+

, Cs+. The proposal was made that
the Na+ ion coordinates to the four carbonyl 0(6) positions
which form a hole in the center of the tetrameric plate
(of: Figure 4) (Figure 5). The ionic radii of Li+ and K+
would be too small and too large, respectively, to coor-
dinate adequately in this position. A second coordination
site, bounded by eight carbonyl oxygens, was proposed as
existing halfway between two stacked tetrameric plates
(cf, Figure 5). The size of this hole could adequately
coordinate with either K+ or Rb+ but would be too large to
coordinate Na+ properly and too small to coordinate Cs+
without sacrificing base-stacking interactions. Due to
similarities in IR spectra upon coordination, all struc-
tures were assumed to be based on stacked tetramer plates.
This was the first reported case of selective alkali metal
ion coordination by any nucleic acid component. Pre-
viously the role of alkali metal ions was relegated to
neutralization of phosphate charge.

Since that time other groups have confirmed the alkali
metal ion specificity in 5'-GMP44-47 and also in the poly-
meric nucleic acids, polyinosinic acid48 and polyxanthylic
acid.49 Recent work by Bouhoutsos-Brown50 sought to

determine the metal ion and nucleotide stoichiometry of

12

 

 

Figure 5.

 

Schematic diagram of the proposed size selec-
tive coordination sites in the structuring of
a neutral aqueous solution of 5'-GMP. a)
Coordination into the enter of the tetrameric
plate by Na+ (r = 0.99 ). b) Coordination
between §he plates by K+ and Rb+ (r = 1.51

and 1.60 respectively). Note tha Li+

(r = 0.593 for 4-coordination, 0.74 for
6-coordination) would be too small 0 substi-
tute for Na+ and that Cs+ (r = 1.78 ) is too
large to replace K+ or Rb+. (Ionic radii from
ref. 51)

13

the 5'-GMP self-structures. She postulated that the
Na25'-GMP structure is octameric with respect to the
nucleotide and requires six Na+ ions per octamer; one each
in the centers of the two tetramer plates and the remaining
four chelated between the top and bottom tetramers. Fur-
thermore, she noted the existence of two structurally
different self-assembled forms with either K+ or Rb+ as
the counter-ion. These structures she termed "simple" (at
ratios of 5'-GMP=/M+ 2 l) and "complex" (5'-GMP=/M+ 2 0.5)
and proposed the stacking patterns with respect to the
5'-GMP to be octameric and hexadecameric respectively. It
was also discovered that five other GMP derivatives
(2'-GMP, 3'-GMP, 2',3'-cGMP, 3',5'-cGMP, and 5'-dGMP) were

able to form metal ion specific self-assembled structures.

Intercalation Drugs

 

The determination of the structures of nucleic acids

has been aided in recent years by the use of intercalation

drugs.52 The word intercalation comes from the Greek and

 

literally means "insert between." Thus an intercalating
drug is one in which the planar portion of the drug mole-
cule is inserted between the adjacent base pairs of a
double-stranded region of a DNA or RNA helix (Figure 6).
Figure 7 shows some of the more commonly employed inter-
calation drugs. All are antibiotics and some are known

52

antitumor agents. The intercalation mode of binding was

14

 

a) b)

Figure 6. Schematic of the mode of intercalation binding
to nucleic acids. a) Normal double stranded
polynucleotide with ribose-phosphate backbone
(striped) and hydrogen bonded base pairs.

b) Same sketch as in a) with three drug mole-
cules (shaded region) intercalated.

first proposed by Lerman53

to explain the binding of the
aminoacridines to DNA. The physiological activity of
intercalating drugs is generally attributed to their
ability either to produce mutations or to interfere with
DNA or RNA polymerase.52 There is also a wide range of
compounds in which only a portion of the molecule may
intercalate (for example, the intercalation of aromatic
amino acids is probably one of the modes involved in
protein-nucleic acid recognition54).
Ethidium bromide (EtdBr; Figure 7b) is biologically

active, possessing trypanocidal,55 antiviral56 and

 

15

 

 

 

 

0 I 1
A .N, .-
9 9 \Otah
C. CH) CU: ,C "—'—'—-w
’CH.'{\ gH‘HC\
(WK (MCW fiTM mem
IO oc sue:
:N'CH3
0 do
/H7C‘CH
2K C’N‘
cm” ;0
/CH-HC\
cm ya
00
-—;CH-—;H
CH, HN‘
co
N
o
(H,
R = H DAUNORUBICIN
R = OH ADRIAMYCIN
Figure 7. Chemical structures of a) actinomycin D;

b) ethidium bromide; c) daunorubicin (also
called daunomycin) and andriamycin; and d)
platinum (II) 2,2',2"-terpyridyl chloride.

16

antibacterial56 properties. EtdBr inhibits DNA synthesis

_i_gvivo;57 and, in_vitro, EtdBr interferes with DNA-

dependent DNA and RNA polymerases of Escherichia coli.58’59

EtdBr-DNA complexes exhibit several prOperties which are

indicative of intercalation. These include a diminished

60 O I O I
enhanced intrinSic Vis-

61,62

sedimentation coefficient,
cosity,60 and a greatly augmented fluorescence.
Recent interest in EtdBr has centered on its special
ability to unwind closed circular DNA.”-66
Proton NMR has proven to be a useful tool in studying
the interaction of intercalation drugs with short nucleic
acid polymers. The purpose is to gain an understanding of
the binding mode of the drug in polynucleotides. Krugh,

67'68 and Chan and coworkers69 have used dinucleo-

£11.13;-
side monophosphates with intercalated EtdBr as models.
Their findings were that as the drug is intercalated, the
aromatic protons shift upfield due to the change in ring
currents felt by the nucleus. The change for each indi-
vidual line was not clear due to overlap with other drug
lines and with nucleotide lines. However, the resonance
of the side chain methyl group lies upfield of all other
resonances and its change in chemical shift as a function
of nucleotide concentration provided stoichiometric in-
formation. In all cases it was found that one EtdBr in-

tercalates between two stacks of hydrogen bonded base

pairs.

17

Statement of Research Aims

To date, the proposed structures of the ordered alkali
metal salts of 5'-GMP= in neutral solution have been based
on intuition and indirect evidence. This is primarily due
to the fact that most previous work was done at a fixed
ratio between the nucleotide and the metal ions
(M+/5'-GMP= = 2). Direct evidence about the exact stoi—
chiometry requires a method in which the metal and nucleo-
tide concentrations may be varied independently. This must
be accomplished, however, without adding ions which will
act as competitors with the species of interest. What is
required is a counter-cation which will allow the addition
of 5'-GMP= and will act neither as a structure directing
ion nor as a structure inhibiting ion.

Bouhoutsos-Brown50 has shown that the tetramethyl
ammonium cation (TMA+) acts as neither a structure inhibi-
tor nor director. This makes it an ideal "dummy ion" for
use in directly determining the stoichiometry of the
structures in solution.

The present study will use (TMA)25'-GMP as a source
of nucleotide in order to vary the 5'-GMP=/M+ ratio. This
should yield more quantitative information than has pre-
viously been available about the exact stoichiometries of

both the metal and the nucleotide in the structures.

18

It is also hoped that intercalation compounds added
into solutions of the structured nucleotide will interact
strongly enough with the structured forms of 5'-GMP= to
allow collaborating evidence concerning the stoichiometries

of the structured forms.

EXPERIMENTAL

A. Materials, Techniques, and Preparation

5'-guanosine monophosphate was purchased as the di-
sodium salt monohydrate from Calbiochem, Inc. and as the
free acid from Sigma Chemical Co. The free acid form was
converted to the dipotassium salt or the ditetramethyl-
ammonium (TMA) salt by titration with KOH (Mallinkroft)
or (TMA)OH (Matheson, Coleman and Bell). Dilute solutions
(~0.0EM) of the free acid were titrated to a pH of ~7.8
with the use of a Fisher model 370 pH meter fitted with a
Markson model E-884 combination pH electrode. All dis-
tilled water was passed through an Illinois Water Treat-
ment Company purification system to remove any paramag-
netic ion impurities and organic residues. All nucleo-
tide solutions were lyophilized three times from 99.8%
D20 (Stohler Isotope Chemicals or Norell Chemical Co.,
Inc.) then prepared to the pr0per concentrations. NMR
spectra were obtained from samples in 10mm tubes (Wilmad
Glass Co., Inc.) and stoppered with silicone rubber serum
caps. Sodium 2,2,3,3,-d4-3-trimethylsilylpropionate
(TSP), purchased from Merck and Company was used as an

internal reference (~0.l wt%) for chemical shift

19

20

measurements. In solutions containing counter-ions other
than sodium, the Na+ ions of the reference were replaced
by passing the NaTSP solution through the appropriate
Dowex 50W-X8 (100-200 mesh) cation exchange resin. The
resin has an exchange capacity of 1.7 meq/mL The resin
was prepared by the following wash sequence: (1) 10% HCl,
(2) H o and EDTA, (3) 10% NaOH, (4) H

O, (5) 1:1 H O-CH OH,

2 3
(9) 1:1

2
(6) CH

2

OH, (7) 1:1 CH3OH-CH2C12, (8) CH2C12,

3OH, (11) 1:1 HZO-CH3OH, (12) 10% HC1,

O and EDTA, (14) lg solution of desired metal ion

3

(13) H

(10) CH

2

hydroxide and (15) H O.

2
Ethidium bromide (EtdBr), which was purchased from
Sigma Chemical Company, was 1y0phi1ized three times from
99.8% D20 to remove ethanol and water of crystallization.
Hematoporphyrin (Sigma Chemical Company) was titrated
(pH = 8) with the apprOpriate metal hydroxide to obtain
the salt form.
Platinum 2,2',2"-terpyridyl chloride was prepared
according to the procedure of Intille.70
Model building studies used CPK space filling models

(Ealing Corporation). The models had a scale of

12.5 cm/nm.

B. Instrumentation

 

Nuclear Magnetic Resonance Spectroscopy-~NMR spectra
were obtained on either a Bruker WH-180 interfaced to a

Nicolet 1180 computer with 16K of memory or a Bruker

21

WM-250 interfaced to an Aspect 2000 computer with 32K of
memory. Probe temperatures were maintained to :10 C by a
Bruker temperature control unit coupled to a probe-mounted
thermocouple calibrated at low temperature in liquid N2
and high temperature in ice water.

Ultraviolet/Visible Absorption Spectroscopy--Concen-
trations of all 5'-GMP and EtdBr solutions were determined
by UV/Vis absorption using a Beckman DU spectrometer
equipped with a Gilford Photometer 252. Solution concen-
trations were determined by assuming that Beer's Law71
holds. The molar absorptivity of 5'-GMP= at 252 nm is

72

13,700, and the absorptivity for Etd+ at 480 nm is

5450.66

Computation--Computer curve fitting was accomplished

 

on a Control Data Corporation, Cyber 170 model 750 com-
puter by using the KINFIT program developed by Dye.73
This method uses Wentworth's74 minimization procedure.
Since the weighting procedure used in minimization and
therefore the final estimates of the parameters and
their standard deviations is dependent upon the errors in
integration, a description of the technique used in inte-
gration and error estimation is presented.

NMR integrations of the H(8) region of the spectra
were accomplished by cutting and weighing of photocopies

of the individual spectra. Triplicate measurements were

made and the average taken. Systematic errors were

22

determined by taking the average value obtained from inte-
gration by two separate investigators. Errors were also
estimated from the threshold concentrationcflfstructure that
can be observed and from the signal to noise ratio at the
given concentration. A11 errors in the variable parameters
that were used in curve fitting were the sum of random and
systematic errors.

The assumption was made for determining the fraction of
structure (see Figure 8 for characteristic spectra) that
the third highest field resonance (labelled B) which lies
under the monomer resonance (9) has an integral intensity
equal to the average of resonances B and B. At higher
concentrations where the monomer resonance (9) moves up-
field, the integral intensity of structure line (B) is equal
to that of either structure line (B) or (B). Integration of

the structure peaks for K 5'-GMP/EtdBr solutions was accom-

2
plished by integrating the entire aromatic region (6.5-8.5
ppm) and the drug methyl resonance. The portion of the

aromatic region due to the drug (10/3 x methyl integration)

was then subtracted out to give the H(8) integration.

C. Theoretical

GMP= Stoichiometry--The following is a summary of the

 

equations used in the KINFIT curve fitting routine to deter-
mine the stoichiometry of 5'-GMP= aggregation under condi-
tions where the ratio of the metal to structured 5'-GMP=

concentration is large. CM defines the concentration of

23

5'-GMP= in disordered monomer form and C defines the con-

8
centration of 5'-GMP= in the ordered form.

1. Formation of ordered n-mers, Mn' from monomers,

 

and

(la)

(lb)

(1c)

(1d)

(1e)

(1f)

If C = XX(1), C = XX(2), 1/n = U(1), and nK = U(2), this

S M
equation can be expressed in FORTRAN language as:

24

XX(Z) = (XX(1)/U(2)*U(1))**U(1)

2. Formation of stacked tetramers, (M4)n' from

monomers, M:

..)

[(M4) n]

[mm] = KM“

Substituting with CS and CM:

[M]=cM

O

_ s
[(M4)n] — E
and
_ 4n
cs - 4nK(CM)

If CS = XX(1), CM

equation can be expressed in FORTRAN language as:

= XX(2), n = U(1), and K = U(2), this

(lg)

(2a)

(2b)

(2C)

(2d)

(2e)

(2f)

25

XX(l) = 4.*U(1)*U(2)*(XX(2)**(4-*U(1))) (29)

3. Equilibria involving monomers (M), tetramers (M4)
and stacked tetramers ((M4)n). In this case CM = total
concentration for the disordered monomer and unstacked

tetramer, and X = concentration of disordered monomer:

4M 2 M4 (3a)

nM (M (3b)

4)n

[M4]

K1 = W (3C)
)

K2 = M (3d)
[”4]

[M4] = (MY/n w (3e)

K2

K' = K K = [”445] (3f)

[M14 [M4] n- 1

 

[man] = K'[M]4 [M.Jn'l <39)

26

 

Substituting with CS' CM, and X:
[M] = x (3h)
C - X .
_ M .
[M4] ‘ 4 (31)
C
[(M4)n] = 751- (3j)
(3e) becomes:
Cs l/n
X = CM "' 4(4—1'1—K; (3k)
and (3g) becomes:
n-l
c = K'(X4)(%M - f) (4n) (31)
S . -—-1r—-

If cS = XX(1), CM

and K2 = U(3), then (3k) can be eXpressed in FORTRAN

language as:

= XX(Z), X = C1, n = U(1), K' = U(2),

C1 = XX(Zl-(4-*((XX(ll/((4.*U(1))*U(3)))**(l-/U(1)))) (3m)

27

letting
C2 = ((XX(2) - C1)/4.)**(U(l) - l.)
and
C3 = 4.*U(l)
then, equation (31) in FORTRAN becomes
XX(l) = U(2)*((Cl)**4)*C2*C3

4. Formation of stacked tetramer, (M4)n, from

tetramer, M4:

run 2 (N1

4 4)n

= [(M4)n]
(“4]n

[M4] = ( BBQ )1/n

Substituting Wlth CS and CM:

K

(3n)

(30)

(3p)

(4a)

(4b)

(4c)

28

CM
[M4] = 4— (4d)
C
_ S
[(M4)n] - 4?: (4e)
and
CS l/n
6.. = 4 m- me

If CS = XX(1), CM = XX(Z), n = U(1), and K = U(2), this

equation can be expressed in FORTRAN language as:

XX(Z) = 4.*((XX(1)/(4.*U(l)*U(2)))**(lo/U(l))) (49)

Na+ Stoichiometry--The following is a summary of the

 

equations used in the KINFIT curve fitting to determine
the stoichiometry of the Na+ in the 5'-GMP= aggregate.
The number of 5'-GMP= units per aggregate was assumed to
be eight. CM defines the concentration of 5'-GMP= in the

disordered form and C defines the concentration of

S
5'-GMP= in the ordered form. CNa is the total concen-
tration of Na+.
xNa+ + 8M2- Z NaxM8(16—x)- (5a)

29

 

Na M (l6-x)-]
K = r 33% 2:]:
Na M
NaXM8(16Tx)-] = K[Na+]x[M2-]8 (5c)
Substituting with CS, CM' and Cgaz
[MZ'] = CM (5d)
[NaxM8(16'x"] = E—S- (5e)

0

[Na+] = CNa- x<§§> (5f)

and

O

X
_ t _ _____S_ 8
CS -— 8K [CNa x<8 )] (CM) (59)

If c§a= XX(1), cS = xx<2), cM = xx<3), x = U(1), K = U(2),

and C1 =[Naf], then equation (Sg) can be expressed in

FORTRAN language in the following way:

30

let

C1 = XX(1) - (U(1)*XX(2)/8.) (5h)

then

XX(2)== 8.*U(2)*((Cl)**U(1))*(XX(3)**8) (5i)

Drug Bindinngonstants-—The following is a summary of

 

the equations used in the KINFIT curve fitting to deter-
mine the binding constants of the structured forms to the
intercalation drug ethidium bromide using the procedure
developed by Popov and coworkers.75 The number of 5'-GMP=
units per aggregate was assumed to be 8. An octamer

(M8) binding to two drug molecules (D) gives the following

expressions,

K1
+ 0
M8 + 2D + M8 D2 (6a)
K = _[_._-__._J_ (.b.
1 2
[Mb]
M8°D2 can react with more M8 to give,
K2
M8°D2 + M8 2 2M8°D (6c)

31

K = [M8°D] 2

2 [M8 - 02] [M8]

 

(6d)

The total concentration of drug (CE) and the total concen-

tration of octamer (Cfie) can be defined as

= [D] + 2[M3'Dz] 4' [M343] (6e)
= [D] + 2K1[M31[D]2 + K1K2[M8] [D] (Gen

0 = (2K1[M8])[D]2 + (1 + ’K1K2|[M8])[D] - c; (sen)

= [:48] + [M.DZJ + [en]

c; =[M.l+K[M8][v]2+Jfi3[naB 1m

Therefore, the free drug concentration, [D], and the free

octamer concentration, [M8], are defined as

 

[1):] -(1+‘/K‘K'2[M8 ]) + \flEN/KIKHMBZP +4(2K1[_M8])C;

2 (2K1 [148]) (69)

32

C;
= 8
[M8] 1 + K1[D]2 + le2 [D] (6h)

 

Defining X as the mole fraction of drug in M8°D2 and 51

1

as its corresponding chemical shift, X2 as the mole frac-

tion of drug in M °D and 5 as its corresponding chemical

8 2

shift, and X as the mole fraction of free drug and 5f

f
as its corresponding chemical shift, the observed chemical

shift (Gobs) of the drug can be defined as

 

 

2
x1 = HID/1:] [D] (61)
CD
x2 = VK1K2£M8] [D] (61.)
CD
Xf = (E7113. (51")
and
5 = 5 x + 5 x + 5 x (6j)

obs f f l l 2 2

33

If CE = CONST(1), af= cousuz), c518 = XX(1), [D] = coma,
aobs = XX(Z), Kl = U(1), K2 = U(2), 51 = U(3), 52 = 0(4),

B = K1K2[D], Bl = 1 + K1[D]2, and C(N) = [M8], equation

(6h) can be written in FORTRAN language as

B = ((U(l)*U(2))**2)*COND (6k)
Bl = 1. + (U(1)*(COND**2)) (6k')
C(N) = XX(1)/(B + B1) (6k")

With D = 2K1[M8] and D1 = 1 + K1K2[M8] equation (69)

written in FORTRAN becomes

D = (2.*U(l)*C(N)) (61)
D1 = 1. + (SQRT(U(1)*U(2)))*C(N) (61')
c1 = -D1 + SQRT((D1**2) + (4.*D*CONST(1))) (61")
COND = Cl/(2.*D) (61“)

Based on the technique of reference 75, an estimate of [D]

is made and used in equations (6k) - (6k") to solve for

[M8]. This value is used in equations (61) - (61m) to

34

solve for a new value of [D] which is resubstituted into
the first set of equations. This loop is continued until
self-consistent values of [M8] and [D] are obtained.
KINFIT then uses these values to solve equation (6j).
If X

= F, X = COMPl, X = COMP2, and éobs = SOBS, then

f 1 2
equations (6i) - (6j) in FORTRAN become

F = COND/CONST(1) (6m)
COMPl = (2.*U(l)*C(N)*(COND**2))/CONST(1) (6m')
COMP2 = ((SQRT(U(l)*U(2)))*C(N)*COND)/CONST(1) (6m")

SOBS = (CONST(2)*F) + (U(2)*COMP1) + (U(3)*COMP2) (6n)

RESULTS

A. Stoichiometry of 5'-GMP= Aggregation: Na+ Salt

In order to determine the stoichiometry of 5'-GMP=
aggregation, it is necessary to investigate a system in
which the metal ion concentration is constant. For the

general reaction,

(Zn-x)-

+ = K
xNa + n5'-GMP 2 Nax(5'-GMP)n

(7)

if the concentration of Na+ is large with respect to the
concentrations of either the dissociated nucleotide monomer
or aggregated structure, then the Na+ concentration and
ionic strength can be assumed to remain nearly constant.
In this case, we can define an apparent equilibrium con-

stant, K', as

K. = K[Na+] = [Nax(5'-GMP)I(12n-X)-:) (8)

[5 ' -GMP=] n

Figure 8 shows the effect on the H(8) region of the

 

tnfizspectrum of adding TMA 5'-GMP into a 0.85M_solution of

2
NaCl in D20. The tetramethyl ammonium ion (TMA+) has been

35

36

 

 

 

 

 

 

B + C
0.138
A 1);,
0.17M
0.23M
[#1111111lllllLJlllll111111111111]
8.7 8.5 8.3 8.) ppm 7.7 7.5 7.3 7.:

Figure 8. Effect on the H(8) NMR of adding (TMA) 5'-GMP
into 0.85M.NaC1 in D20 at 15°C. Struc ure
lines are labelled B, B, and B, and the
monomer line is labelled Q.

37

shown by Bouhoutsos-Brown50

to be a structurally inert
cation with regard to Na+/S'-GMP= self-assembly. As the
concentration Of 5'-GMP= increases, an increase in the
total concentration of structure is observed. The reson-
ances are assigned to the monomer 5'-GMP= (labelled B),
the two observable structure lines (labelled B and B), and
a third structure line (labelled B) which lies under the
monomer resonance. At high concentrations of 5'-GMP=, the
B peak moves upfield making the third structure peak, B,
visible. At all concentrations where it is possible to
integrate the B peak, the integral intensities of all three
structure lines (B, B, and B) are identical within 1 3%
experimental uncertainty. In this study, therefore, at
concentrations where the Byline is not observable due to
overlap with the monomer resonance, its intensity is
assumed to be the average of the intensities of the B

and B peaks.

Table 1 lists the fraction of structured 5'--GMP= as
the total nucleotide concentration is varied from 0.13B
to 0.23B while the Na+ concentration is held conStant at
0.85B. Under these conditions the total Na+ to total
5'-GMP= ratio varied from ~ 6.5 to ~ 3.7. The Na+ concen-
tration is much larger than the concentration of structured
5'-GMP= (11 E [Nafltotal/[§'-GMP=]structured S 85).

An attempt was made to fit these results to a series

of different equilibrium equations (for a complete summary,

38

Table 1. Fraction of Structure of 5'-GMP= with Changing
Total Nucleotide Concentrationé

 

+
[Na ] total

 

 

['-GMPi]totalE' Fstructured. F'-GMP—]structured
0.13(i0.003) 0.078(i0.048) 85
0.l4(i0.003) 0.134(i0.049) 45
0.15(i0.003) 0.164(10.013) 35
0.16(10.003) 0.180(10.020) 30
0.17(i0.003) 0.220(i0.018) 23
0.18(i0.004) 0.236(i0.017) 20
0.19(i0.004) 0.272(i0.019) 16
0.20(i0.004) 0.287(i0.020) 15
0.21(i0.004) 0.299(i0.025) 14
O.22(i0.004) 0.322(i0.019) 12
0.23(i0.005) 0.328(10.024) 11

 

a. 0.85B_NaCl in D O; 15(il)°C; errors are the sums of
the systematic and random errors of two independent
integrations of H(8) resonances. The 5'-GMP' was
introduced as the TMA+ salt.

b. Concentration in moles/liter.

39

see Experimental section). The most general form that
such an equilibrium could take would be that of disordered
monomers of S'-GMP= (M) associating to form structured

n-mers (Mn) such that

nM 2 Mn (1a)

If the basic structural building block is assumed to
be a tetramer, then a more specific equilibrium can be

written as

4nM I (M4)n (2a)

Equation (2a) can be written in terms of a two-step
equilibrium process such that monomer (M) first forms
tetramer (M4) which then proceeds to form the stacked

tetramers ((M4)n). The resulting expressions are

4M 1 M4 . (3a)

+
Chantot and Guschlbauer76 using UV spectroscopy have

reported observing free tetramer (M4) for the gels of the

acidic nucleotides, KH(5'-GMP) and KH(3'-GMP) (pH 2 5.5).

40

Lee and Chan77 have determined by 1H nmr a tetramer forma-

tion constant of 2.5 x 107 B-3 for NaH(5'-GMP). If a
similar value is applicable for Na2(5'-GMP) at concentra-
tions where structure formation occurs,nearly all of the
monomer is expected to be in the tetrameric form. The
B_resonance (3B. Figure 8) would therefore be all tetramer
and a fourth equilibrium expression,

rm 2 (M Ma)

4 4)n
can be written.

Table 2 shows the results of fitting values of n and
K' in the above equations using the data from Table 1.

It is clear from an inspection of the results of the
curve fitting that equations (3a) and (3b), describing
tetramer formation from monomer and subsequent stacking
of tetramers, cannot be fitted to the data. The standard
deviation estimates (95% confidence) of the best values
are large. The K' values are too large to fit with the
available range of data. In addition, (3a) and (3b) allow

for negative values of n, a physical impossibility.

For reasons that will be discussed later, the n and
K' values obtained for equations (la) or (2a) are prefer-
able to those obtained for equation (4a).

Looking at the results for n-mer formation from

monomer (equation (1a)) it appears that between 4.9 and

41

.Hmema moemeameoo ema men
“ool.aasma asz.o u mH.e u fluazou.mg “2mm.o

“M COUMEflHmm OHM MHOHHG

u Tee; .m

 

 

 

 

 

eoa x 1.a~avm.m 1m.Havm.e lees glaze H as:
a
bad x Aoooommavm.a Inns lazy H ext
1m.mave.a a
msoa x loooooeaava.a lame z N 2e
. III . . II . c w IIY
Nos x I -+.~ a 1m e+vm A real I so + see
an
mes x 1H.aave.~ 1e.Havm.e luau z N 2:
.M : coflumsvm

 

maomz ea um2o1.m mo meauuam

manoeoflsoHODm mo muasmom .m manna

42

8.1 5'-GMP= units are contained in a structured unit with-
in experimental uncertainty. No single-stranded helical
structure containing between five and eight 5'-GMP= units
can explain a three line H(8) spectrum. If structure for-
mation is assumed to be based on tetramer units, this range
of values of n would indicate between 1.2 and 2.0 tetra-
meric 5'-GMP='s per structured form. Equation (2a) yields
a value between 1.2 and 1.8 tetramers per structured form.
Taking only integer values within this range allows the
structures to be composed of either one or two tetramers
of 5'-GMP=. As will be shown later, an octamer consisting
of two stacked tetramers can explain the observed H(8)
spectrum.

A single tetramer unit has a C4 axis perpendicular to
the aromatic plane (9B. Figure 4). All four H(8) positions
are therefore equivalent and cannot give rise to the three-
line H(8) nmr spectrum observed for the ordered nucleotide.

Based on the above observations, it is believed that
eight nucleotide units combine to form the aggregate in
the ordered form of the sodium salt. A later section
shall discuss how an octameric 5'-GMP= structure can
exhibit the three H(8) resonances by 1H NMR. Figure 9a
shows a plot of the best fit curve for (1a) over the
experimental data.

Besides the octameric form of the Na 5'-GMP, other

2
stoichiometries have been proposed in the literature.

Figure 9.

43

Plots of the computed curves for equation
(la) when a) n and K' are both varied, and
b) K' is varied and n is held constant at
either 8, 12, or 16. Experimental condi-
tions are the same as in Table 1. Error
bars represent the experimental points.
Errors in n and K' are reported at the

95% confidence level.

GMP.

'0

0

b

3

Q-l

0

L a
I.
2+1
0‘!)
la.

 

44
nMgMn
n-a.5(t1.0);K-2.7(t9.1)x1o3

 

 

 

 

 

 

 

 

 

0.4+ AAAAAAA 14-Hen-14“-.-.“lure“--.’
1 :
0.3{ t
v 3 :
o . _
b 4 ’
3 : :
OJ 1 ,
”0.2-1 .—
3 a .
+4 .. '
m . _
LL : :
l I
0.3-“ :—
1 l
‘ :
o.o' ....... fil..,.fie.-l...flWL
0.10 0.15 0.20 0.25 0.30
GMP-
cTotol
n=8(————-);K=4.o(to.6)x1o4
n=12( -------- );K=7.4(t1.1)x1o7
”=15(“""");K=9.6(t1.0)x101°
0.4 -.1-...A11L....11..I.111....1 ........ .
r
)
P
0.30

Figure 9.

45

40

Pinnavaia, Miles, and Becker have suggested the possi-

bility that the three tetramer plates stack to form a

78 in their most

dodecamer. Laszlo and his coworkers,
recent model, have introduced the possibility of stacking
four tetramers to form a hexadecamer.

In an attempt to compare the validity of the dodeca-
mer and hexadecamer models with the present octameric
model, equation (1a) was refitted to the data. This time,
however, instead of allowing n to vary, its value was
held constant at values of either 8, 12, or 16 and only
the value of K was allowed to vary.

Table 3 summarizes the results of the curve—fitting
of equation (la) at a constant value of n.

Figure 9b shows plots of the computed curves for the
best values of K' for octamer, dodecamer. and hexadecamer

formation. It is apparent from a comparison of the com-

puted curves and the experimental data that the value of

n 8 provides a closer fit to the observed points.
Therefore, all further work shall be based on the assump-
tion that sodium 5'-GMP forms octameric structures in

aqueous solutions.

B. Model Buildigg Investigation of Octameric Structures
Having proposed an octameric unit for the structured

form of Na 5'-GMP it is necessary to establish how the

2
simple stacking of two tetramers can yield a three line

46

Table 3. Results of Curve Fitting
Equation (la) to the Data
in Table l with n Held to
a Constant Value.3-

 

 

§l

nM + Mn (1a)

n K'
4
8 4.0(to.6) x 10
7
12 7.4(il.1) x 10

16 9.6(i1.0) x 1010

 

B. The experimental conditions are
the same as in Table 1. Errors
are at the 95% confidence level.

47

H(8) NMR spectrum. In order to address this question, a
model building study was undertaken. In addition to
answering questions concerning symmetry, it was also hOped
that additional stabilizing factors (e.g., additional
hydrogen bonding) for octamer formation could be found.
For the stacking of tetrameric plates of 5'-GMP=, the
symmetry of the stacking is very important. The hydrogen
bonding scheme for 5'—GMP= has a pair of hydrogen bond
donor atoms, N(l) and N(2), nearly orthogonal to a pair
of hydrogen bond acceptors, 0(6) and N(7) (9B. Figure 4).
Arrows bent at right angles, the heads representing hydro-
gen bond donor atoms and the tails representing hydrogen

bond acceptor atoms, can be used to illustrate a tetramer

rm
L11

Due to the directional character of this type of bonding,

unit of 5'-ch= (1).

the tetramer unit is prochiral independent of the ribose
group. When a pair of tetramers stack, stacking can occur
in either normal (head-to-tail) or reversed (head-to-head)
arrangements (Figure 10). A twist angle, 9, is defined as
the angle caused by the projection of a line joining the

center of a tetramer unit and 0(6) with a similar line on

48

R 7?W§_R
ate

Normal

e: are

 

Reversed

Figure 10. Schematic diagram of a) normal stacking mode
and b) the two enantiomeric reversed stacking
modes of the octamers of 5'-GMP-.

49

the tetramer plate below it. The comparison is made be-
tween carbonyls of overlapping base units. 0 is defined as
positive if the bottom plate twists clockwise with respect
to the top plate and negative for a counterclockwise twist.
Taking into account the symmetry of the possible
stacking patterns, one may predict the number of H(8) lines
expected for each structure. The normal stacking pattern
(Figure 11), in the absence of the chiral ribose phosphate
group, exists in three different symmetry states depending
upon the value of e. For a = 0°, the symmetry is °4h and
therefore the unit is achiral. When 0 = 45°, the symmetry
is $8 and the unit is again achiral. However, at 0° < e <
45°, the unit has C4 symmetry and, therefore, is chiral.
If only the values of 9 equal to +30° and -30° are con-
sidered (for reasons which will be described later), it is
found that the structure with 0 = +30° is the enantiomer of
the structure with e = -30° (designated c4(cw,cw,+3o°) and
C4(CW,CW,-30°) respectively). The addition of the chiral
ribose phosphate group causes the normal stacking pattern
to adopt C4 symmetry, regardless of the value of 0. For
0 = +30° and —30° there are therefore two diastereomers
(C4(CW,CW,+30°)-ribose phosphate and C4(CW,CW,-30°)-ribose
phosphate respectively). When 9 = 0° or 45°, these dia—
stereomers isomerize into a single species.

Since the point group for normal stacking is C4,

regardless of the value of e, the top and bottom units are

Figure 11.

50

Overlap pattern of the guanine bases for
the two enantiomeric forms in the nor-
mal stacking mode of two tetramers. The
dark tetramer is the top layer and the
light tetramer is the bottom layer. a)
C4(CW,CW,+30°); b) C4(CW.CW,-30°).

a) is the enantiomer of b) in the absence
of the chiral ribose-phosphate group.

52

inequivalent and two H(8) environments should be observed
for each value of 0. Consequently, for 0 = +30° and 9 =
-30° a total of four H(8) lines are predicted for normal
stacking. This would be the expected maximum at the slow
exchange limit. If the twisting of the plates from

0 = +30° to 0 = -30° is fast on the NMR time scale, isomer-
ization reduces the number of H(8) lines to two.

Figures 12 and 13 show the overlap of the purine
bases for the reversed stacking modes. For the reversed
stacking modes in the absence of the ribose phosphate,
the symmetry is D4 for all values of 0. This chiral
stacking pattern will give two enantiomers for a given
value of 0. One is formed by the hydrogen bonds going
clockwise on the t0p tetramer and counterclockwise on the
bottom tetramer (D4(CW,CCW)) and the other has the hydrogen
bond direction counterclockwise on top and clockwise on
the bottom (D4(CCW,CW)). With the attachment of the
chiral ribose phosphate group the symmetry remains D4 and
causes the enantiomeric forms to become diastereomeric.
Since in D4 symmetry the tetramer plates are equivalent,
each diastereomer should yield only one H(8) environment

+30° and e=

for a given value of 9. Considering only 0
-300 gives a total of four H(8) lines expected at the
slow exchange limit. Rapid twisting would collapse this

into a two line spectrum.

Figure 12.

Overlap pattern of the guanine bases for
two of the diastereomeric forms in the
reversed stacking mode of two tetramers.
The dark tetramer is the top layer and
the light tetramer is the bottom layer.
a) 04(cw,ccw,+3o°); b) 04(cw,ccw,+eo°).

Figure 13.

55

Overlap pattern of the guanine bases for
the remaining two diastereomeric forms
in the reversed stacking mode of two
tetramers. The dark tetramer is the
top layer and the light tetramer is the
bottom layer. a) D4(CCW,CW,-60°); b)
D4(CCW,CW,-30°). In the absence of

the chiral ribose phosphate group,
Figure 13a is the enantiomer of Figure
12b, and Figure 13b is the enantiomer
of Figure 12a.

57

Table 4 summarizes the isomers discussed along with
the number of H(8) lines expected for each exchange region.
It should be pointed out that for all isomers, rapid twist-
ing with isomerization of e is not hindered by the ribose
phosphate group.

This summary supports the assumption that the four
line H(8) spectrum for Na25'-GMP is due to the stacking of
tetramers to form octameric structures. On Figure 8, the
second highest field line (labelled B) is assigned to the
unassociated monomer. The lowest and highest field reson-
ances (B and B) have been reported40 as having equal inten-
sities and T1 values at all concentrations and temperatures
studied and relax at the same rate upon addition of paramag-

netic Mn2+ or Cu2+.

These two lines are therefore assumed
to arise from two inequivalent H(8) environments on the
same structured octamer. Only the normal stacking pattern
has inequivalent H(8) positions on the same structure,
therefore lines B and B_are assigned to this stacking mode.
Since only two lines with these characteristics exist (in-
stead of four) it is assumed that the twisting mechanism is
fast on the NMR time scale (BB. Table 4).

The second lowest field line (B, Figure 8) has a diff-
erent T1 than lines B and B and relaxes at a much faster

40 It therefore

rate upon addition of paramagnetic ions.
arises from an octameric structure in which H(8) environ-

ments on the top and bottom tetramer plates are equivalent.

58

 

“Auoaocosv Moemuuou
spas ouonuo>m mafia.

H ”nosed Amen
a «muoEOMH

Aoomn.zo.zoovea

AOOMI.3U.3UOV 9

++ e ”H

1000+.zuo.su.vo

Aoom+.3UU.30v¢G

Aoomu.3o.3ovvo
H. ++ e
Aoom+.3o.sov o

.coflumummom xomum ummm
.moflumwsu uwmm

 

v umwnwd Amvm
m "MHQEOMH

Accen.zo.zooveo
++ v
Aoomn.zo.3oov a
Aooo+.zoo.so.¢o
++ v
Aoom+.zoo.zo. o

N "nosed Amy:
N “muosoma

loom-.zu.sueso
++ s
Room+.3o.3ov o

m "modes Amer
H ”muoEOMH

.noaumummom xomum scam
.o no cofiumnwuoEOmH
has: meanness umnm

 

m“ “waned Amvm
o ”muoEDMH

Aooou.so.zuoc
loomn.eo.zoov
A oo+.3oo.3ov

O
A om+.300.30v Q

on
v
D
v0
v

o
v "nosed Amy:
v “MHQEOMH

AOOMI.3U.BUVWU
A00m+.zU.3UV 0

v "mocaa Amy:
N "muoBOMH

.cowunummom Momum 3on
.oouuomoum adamsuo
oomn u m can oom+ u m
spa: mcdumwsu 30Hm

 

Hove?

mcwxomuw oomuo>om

meexoeum Hesuoz

mcowuwocoo

 

mcowoom omcmnoxm cough Mom oouoomxm nosed vam can muoEOMH mo uonssz .w manna

59

Correspondingly this line is assigned to one of the in—

equivalent reversed stacking modes possessing D4 symmetry.
A number of crystal structures have been done on a

wide variety of nucleic acid constituents and have been re-

viewed previously.79-81

They illustrate that the bond
angles for the ribose moiety are normally confined to
limited values independent of the nucleotide. Table 5
defines the four critical bond angles along with listing
their most probable values. Along with these angles,
there are two predominant conformations for the ribose

sugar: C(3')-endo (BB) for the ribonucleic acids and

C(2')-endo (III) for the deoxyribonucleic acids.

Table 5. Torsional Angles and Notations for 5'-

 

 

Nucleotidesé
Torsion AnglesB Notation Most Probable Angle
O(l')-C(l')-N(9)-C(8) x 36° 2 19°
C(3')-C(4')-C(5')-O(S') w 53° a 7°
C(4')-C(5')-O(5')-P 5 187° 1 24°
C(5')-O(5')-P-O(3') w 69: a 43
294 a 13

 

3. From reference 80.
B. All angles A-B-C-D are measured clockwise from A to
D when viewed along B-C. An eclipsing D is 0°.

60

 

(In the technique employed, two tetramer units built
with CPK space filling models were set up with the most
common bond angles (EB. Table 5) and with C(3')-endo con-
formations of the ribose moiety. Both the normal and
reversed stacking patterns were investigated. The twist
angle, 0, was set at either +30° or ~30° (:300 or i60°)
corresponding to Zimmerman's model building work with four-
stranded poly(G)36 in which planar tetrameric guanine units
are present. The same helical parameters were also foundlnr
Zimmerman for fibers of Na25'-GMP grown from excess NaCl.37
The bond angles were then adjusted by the smallest amount
possible in order to give octameric structures with hydro—
gen bonds between the tetramer plates.

Table 6 shows the results of the model building
studies. The parameter A°H(8) is the intermolecular
shielding (in ppm) for an H(8) nucleus 3.4 A above or
below the ring plane due to the sum of the contributions
of the ring current and of the atomic diamagnetic suscep-
tibility anisotropies as calculated by Giessner-Prettre

and Pullman.82 Several important points should be noted

61

 

 

 

 

 

 

 

 

 

 

~.o
meson 8.1.28 96a angelic me can 0H3 08 oem Aoee+.zoo.3o.eo
~.o
meson 1.201.320 86m onediCo To can on: com com loanisooioaee
mmflxomum commo>mm
«.0 e
meson do 13% names 851.30 «S can on: 08 com rosetzoiooe e
~.o
mocoo mo.A~vmz usmflm ooomlfi.mvo ~.o 0mm oama Oom 0o~ “Conl.su.zuovvo
mmyxomum commo>om
v.o e
mocoo moim.mvmo noon oncolfi.mvo o.o 0mm ona com Com “Ooml.zu.zuv U
v.0
cocoa mo.*.mvmo Hsom oooola.mvu o.o 0mm oama 0om Com “Oom+.3U.SUV¢U
msfixomum HmEHoz
modem mcoflumEH0mcoo A cm
comouowm mmoowm AEmmv m e< 3 e e x e

 

mzwn.m mo meuom OHHoEmuoo on» no hosum maeoaflsm Homo: on» no muasmom .m manna

62

from this table. First, it is significant that in all
eight octamer structures studied, the nucleotide has to
undergo little variation from its "normal" bond angles in
order to gain at least four and in two cases eight addi-
tional hydrogen bonds between the top and bottom tetramer
plates. Second, it should also be noted that this variance

is self-consistent for all eight structures.

C. Metal Ion Stoichiometry: Na+ Salt

Having established the nucleotide stoichiometry at
eight for the solution ordering of Na25'-GMP, we now direct
our attention to the number of metal ions required per
structured unit. The equilibrium expression can be written

as

xNa+ + 8M2- I NaxMé16-x)- (5a)
with the equilibrium constant, K, defined as
[NaxMé16-x) -]
K = (5b)

 

PENN”?

In the experimental procedure for determining K and
x in equation (5b), two important criteria were met.
First, the total 5'-GMP= concentration was held constant

in order to determine only the effect of the Na+. This

63

was accomplished by adding incremental amounts of 0.78B
Na25'-GMP into 0.78B_(TMA)25'-GMP. Second, care was taken
to avoid changes in ionic strength. Data were collected
with 0.5B_TMAC1 to keep the ionic strength nearly constant
(u = 2.8(20.1)B). The NMR spectra used were provided by
Bouhoutsos-Brown.50
Table 7 lists the fraction of 5'-GMP= structured as a
function of the 5'-GMP= to Na+ ratio. These data were fit
to equation (5b) using KINFIT (see Experimental section for
full details). This resulted in a value of x = 4.1(iO.1)

and K = 1.7(iO.7) x 1010 if”

at the 95% confidence level.
This would indicate that four Na+ ions are required for
every octamer unit.

Bouhoutsos-Brown50 has proposed that each octamer re-
quires six Na+ ions in order for the structure to be

formed. The results in Table 6 were refitted using KINFIT

allowing only K to vary and holding x constant at six.

8 9

This calculation yields a value of K = 3.5(121.) x 10 BI
at the 95% confidence level. Due to the high standard
deviation in the value of K when x = 6,it appears that a
value of four is a more acceptable value.

Figure 15 shows plots of the ratio of 5'—GMP=/Na+
versus the fraction of 5'-GMP= structured showing the

fitted curves when x and K both are varied and when

x = 6 and K is varied.

64

Table 7. Fractign of Structure of
5'-GMP: with Changing
5'-GMP"/Na+ E

 

 

5'-GMP=
Na+ F
structured
2.0 0.805(10.035)
3.0 0.700(i0.035)
4.0 0.535(10.03l)
5.0 0.452(10.043)
6.0 0.358(i0.03l)
7.0 0.249(10.027)
9.0 0.202(10.022)

 

a. 0.78M 5'-GMP added into 0.7811

(TMAT25'-GMP and 0.5 B_TMACI;
0.3°C; ionic strength = 2.8(30.1)B,

Figure 14.

65

Plots of the best fits for equation (5a)
with a) the values of x and K allowed to
vary and b) x held constant at 6 and K
allowed to vary. Experimental conditions
are the same as in Table 7. Error bars
represent the experimental points. Errors
in x and K are reported at the 95% confi-
dence level.

memnw

FGIII
Structured

1 .0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1

0.0

1 .0
0.9
0.8

0.7

0.6 '

0.5

0.4

0.3

0.2

0.1

0.0

xNa

++8M

2

66

= NaxM

8

(16—x)-

x-4.1(to.1);K-=1.7(to.7)x101°

1.]

11,1.IL1

L

 

 

 

 

 

 

 

 

 

 

"_fillrl'l'lTfiI II
0.0 1.0 2.0 3.0 4.0 5.0 0.0 7.0 8.0 0.0 10.0
(cur/um
XNO+ + 8M2—'—_-3NQXM8(‘°-x)-
x=6:K=3.5(121.)x1o'0
‘11].l.l.l.l.ll.l
\‘ L

\
\
— ‘ —
\
\
\
\
H \
i
l
l'l'lfil'l'l'l'l‘l
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 0.0 10.0

GMT/ME)
Figure 14

67

D. Interaction of Intercalation Compounds with Structured
Na251-GMP and K25T-GMP

Due to the fact that the stacking of hydrogen bonded
tetramers of 5'-GMP= is structurally similar to the poly-
nucleotides, an attempt was made to insert intercalation
compounds either between the tetramers or to bind the com-
pounds on the external surfaces of the stacked tetramers.
Studies were done on the intercalation compounds ethidium
bromide (EtdBr) and platinum(II)2,2',2"-terpyridyl chloride
(Pt(terpy)C12) (9B, Figure 7). In addition, the disodium
and dipotassium salts of the water soluble porphyrin,

hematoporphyrin (Figure 15L.were investigated. Measurement

 

OOH C H3

Figure 15. Structure of hematoporphyrin.

of CPK models of the tetramer and the porphyrin show that
one aromatic porphyrin should be the appropriate size to

overlap with one of the tetramer units.

68

Unfortunately, the hematOporphyrin proved too insol-
uble in aqueous solution to give adequate NMR spectra at
concentrations where 5'-GMP= shows structure formation.

At high concentrations of either Na25'—GMP or K25'-GMP the
intensity of the chemical shifts due to the porphyrin were
too weak to be observed. No change in chemical shift or
intensity of either the structure or monomer H(8) reson-
ances was observed. These observations were taken to mean
that little or no interaction was occurring.

The addition of Pt(terpy)Cl2 to structured forms of
Na 5'-GMP and K

2 2
platinum salt and some of the nucleotide. Lippard and

5'-GMP caused precipitation of all the

coworkers,83 in work with Pt(terpy)C12 and calf thymus
DNA, have prOposed that in addition to intercalation, the
platinum-chloride bond is sufficiently labile to be

broken in aqueous medium and a new covalent bond is formed
to the DNA. They postulated that the platinum binds to
nitrogen donors on the aromatic bases thereby disrupting
the normal hydrogen bonding scheme. If a similar reaction
is expected with 5'-GMP=, the platinum would be expected
to bond to the N(7) position and block tetramer formation.
The precipitate was therefore assumed to be a platinum-5'-

GMP complex.

69

E. InteractiongBEthidium Bromide with Structured and
Unstructured Forms of Na25‘-GMP

 

The addition of the intercalation drug ethidium
bromide (EtdBr) to Na2(5'—GMP) does not cause precipitation.
A preliminary investigation of the change in chemical
shifts of both the drug and nucleotide resonances revealed
interactions were occurring. This led to a more quantita-
tive investigation of the interaction.

Figure 16 shows the effect of the addition of Na25'-GMP
into 0.034B_EtdBr at 5°C. The tOp spectrum represents the
spectrum of the "free" EtdBr with chemical shifts the same
as previously characterized by Kreishman and Chan.69 They
assigned the resonances as follows (see Figure 7 for num-
bering sequence): EEEEI and mgggfprotons on phenyl group
between 7.8 and 7.7 ppm; H(l) and H(lO) at 7.69 and 7.653mm1
respectively; the BBBBprosition of the phenyl group, H(2),
H(9), and H(4) between 7.20 and 7.06 ppm; H(7) at 6.26 ppm;
methylene resonance at 4.40 ppm;and methyl resonance at
1.40 ppm. Splitting between the methyl resonance and meth-
ylene resonance (J = 8.3 Hz) is incompletely resolved at
this concentration and temperature due to some base stack-
ing of the drug with itself.69

As the drug becomes intercalated, Kreishman and Chan
reported a general upfield movement of the aromatic reson-

69 3(1) and H(lO)

ances along with considerable broadening.
were observed to shift upfield to ~ 8.0 ppm; H(2), H(4),

and H(9) shift upfield to the region of ~ 6.6 to 7.2 ppm;

70

Figure 16. Effect on the 1H NMR.of adding Na 5'-GMP
into 0.034B EtdBr at 5°C. Nucleoéide
H(8) resonances are denoted by shading.

71

| 0.00B.Na25'-GMP
I" J 3 J.

,I
\ /~ I
\ .-r ~ - . , ’ \.
. ‘whn—h ___.-.__.__ v-“'-

0.02;! Na25'-GMP

 

 

' ,
J l. ,1 l I
. I :4 fl
h . 1 L/ JUFI
,' “ U
t/ w/ \
i. t f n I ; V 0.18M Na25'-GMP
) 1(‘1.
(j ‘ i
J \) 1
\J
I ‘ , ’ KL 1" (\
~ /J \I V u \‘ ~~~~~~ ‘__.__ ,/ K,
0.29M Na25'-GMP
1‘ 4 a
II a (I \1
2| r. E [J J 'v‘; \\_
u v \J ;——————-—— —s \
I ,1 .
l ’ \

,9.) ' 0.60B Na 5'-GMP

A i I)“ K/ x
l \1‘

Et- « » \»
I - \.
."‘\‘ v ‘M‘J f

lllllllLillllllillllllllllllllLLLlllJ
9 8 7 6 5ppm4 3 2 l o

\

\a‘

Figure 16.

72

and H(7) moves to a position under the nucleotide%;H(l')
resonance at ~ 5.9 ppm. The chemical shifts of the QEEEQ'I
TEES?! and pgggfhydrogens of the phenyl group on the drug
exhibit no change upon intercalation.

As a small amount (0.02B) of Na 5'-GMP is added into

2
the solution, below the concentration at which structure
formation begins to occur, dramatic spectral changes can be
observed. First, the aromatic region of the EtdBr spectrum
broadens substantially and generally shifts upfield. This
is the same general movement observed by Kreishman and Chan
upon intercalation of the drug.69 Due to overlap,most re-
sonances cannot be assigned; however, H(7) has shifted from
6.26 to 6.13 ppm. A small upfield shift of the methyl re-
sonance of the drug (~ 0.05 ppm) is also observed. The
drug methylene resonance has become obscurred by the ribose
protons of the 5'-GMP= (°H(1') = 5.63 ppm, 5H(2.) = 4.59
ppm, °H(3') = 4.44 ppm, °H(4') = 4.29 ppm, and °H(5') -
4.00 ppm). The most dramatic change, however, is that the
monomer H(8) resonance of the 5'—GMP=, which in the absence
of the drug has a chemical shift of 8.22 ppm at this con-
centration and temperature, has shifted upfield to 7.94gumn
Since no evidence of structure formation is apparent, the
shift of the H(8) line and the change in the drug reson-
ance indicate that the EtdBr is interacting with the 5'-GMP=

in some manner which perturbs the chemical shift of the

aromatic region of the spectrum. This is most likely due

73

to base stacking of the EtdBr and 5'-GMP= monomer. This
type of monomer nucleotide interaction with the drug has
been suggested by Waring84 to explain the change in the
extinction coefficient of EtdBr upon addition of mononucle-
otides or deoxymononucleotides.

As the concentration of monomer nucleotide increases,
the monomer H(8) resonance moves downfield until at 0.18B
it remains at 8.10 ppm, 0.07 ppm upfield of its position at
the same concentration and temperature in the absence of
the drug. At this concentration, structure formation of
the 5'-GMP= is observable. Two new H(8) resonances at 8.48
and 8.33 ppm can be observed. These are presumably the same
structure lines (B_and B, Figure 8) which have chemical
shifts of 8.55 and 8.26 ppm in the absence of added drug.
The change in their chemical shift is an indication thattflme
drug is also interacting with the structures in some way.
The highest field structure line (B, 7.25 ppm in absence of
drug) appears at 7.18 ppm, partially overlapping with some
of the aromatic resonances. The most notable change, how-
ever, is that the resonance associated with the methyl group
on the EtdBr has moved upfield to 1.07 ppm. As the total
concentration of Na25'-GMP in the system increases and the
concentration of structure increases, the drug methyl re-
sonance continues to move upfield. At 0.29B.Na25'-GMP the
methyl resonance has moved upfield to 0.69 ppm. Upfield

shifts of the methyl resonance have been reported previously

74

when EtdBr intercalates with dinucleoside monophos-

pihates.°7-'69

However, in these cases, the maximum change
in chemical shift in the upfield direction (A5) was only
about 0.50 ppm. The upfield shift of the drug methyl re-
sonance, relative to the uncomplexed state (A5 = 0.71 ppm),
is therefore 0.21 ppm higher field than has been observed
previously.

A further increase in the total concentration of
structured Na25'-GMP causes the drug methyl resonance to

shift downfield until at 0.60B Na 5'-GMP its chemical shift

2
reaches 0.85 ppm. Any further increase in the total con-
centration of Na25'-GMP causes no further change in this
methyl resonance. In addition, as the concentration of
structure is further increased, the H(8) structure lines
approach the chemical shifts that are observed in the
absence of the drug.

These observations indicate that when Na25'-GMP is
added into a solution of EtdBr, two modes of binding can be
observed. The first, denoted as "monomer binding," occurs
at low concentrations of Na25'-GMP and is characterized by
the upfield shift of the H(8) monomer resonance. The sec-
ond, denoted as "structure binding," is observed only at
concentrations of Na25'-GMP which are high enough for struc-
ture formation to be observed. Structure binding of the

drug to the 5'--GMP= is most easily characterized by the

large upfield shift of the methyl resonance of the EtdBr.

75

Binding of the drug to the structured nucleo-
tide does not appear to perturb the structure to a great
extent as can be noted by the small changes of chemical
shift of the H(8) structure lines (A5 = 0.07 ppm). It does
however have a pronounced effect on the total amount of
structure for a given concentration and temperature.
Figure 17 shows the fraction of 5'-GMP= structured as a
function of the total 5'-GMP= in solutions containing and
absent of EtdBr. It is readily apparent that the fraction
of structure is increased in the solution that contains
EtdBr. This is especially apparent at the lower concen-
trations of Na25'-GMP. This is probably attributable to
the drug acting as a template for structure formation. The
tricyclic aromatic portion can most efficiently overlap
with two hydrogen bonded base pairs. If this were to
occur in the case of Na25'-GMP this could be a precursor
to the formation of the tetramers and later the octamers.

Figure 18 shows the chemical shift of the EtdBr
methyl resonance as a function of the ratio of the amount
of structured Na25'-GMP to EtdBr. The upfield then down-
field chemical shift changes indicate that two separate
complexes are being formed between the EtdBr and the
structured Na25'-GMP. Extrapolation of the linear portions
of the curve indicates that these stoichiometries are
approximately four and eight 5'-GMP='s per drug molecule.

If it is assumed that the octamer is the basic unit of

76

Fraction of NoszP Structured With 8: Without Drug

(A-Without;D-With)

 

 

 

 

0.8 E
u DI E
0.7 —: =_
. a
0.6 -—f /' ’—
: u .0 :
E (‘0. t E
0.5 -f :—
: O I
E I: n .‘ E
: 9 . 5
Egg OA-fig /;// Q' F
0.3 -: .° 5-
: .5 E
I ° :
a / .- s
0.2 —‘: a o. E'-
: I .u E
0.1 -: .° E.
00 “"“T‘ I ' I I I l I I I ‘
0.10 0.20 0.30 0.40 050 0.00

Figure 17.

Effect of the fraction of Na25'-GMP structured
in the presence ([3) and in the absence (A)

of EtdBr. Experimental conditions are the
same as in Figure 16.

Structured 5'-GMP'

1.4

1.3

1.2

1.1

1.0

aI-wa)

0.9

007

0.6

Figure 18.

77

Binding to EtdBr
Sodium Salt

 

 

 

 

 

.I. I.I .l I I .l. I I I .I..I.I..II
3‘. z
2‘ :
a \ _
:\ :
a E
5 "vs-"F :0—5
1 _
i o E
5 ‘1 O 0 E
.1 ‘1 ° 0 :—
2 A,” ;
1’ \ E
'l'lfil‘l‘l'l'l'l'1'l'l'l'lfil‘
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0
(Ir-cue- Structured/EtdBr)
Plot of the chemical shift of the EtdBr methyl

resonance as a function of the structured

Na25'-GMP/Drug ratio.

Extrapolation of the

linear portions of the curve included.

78

structured Na25'-GMP this would indicate the binding
of two and one drug molecules per structured unit respec-
tively.

A two step equilibrium process can therefore be writ-

ten for the formation of the two structure-drug complexes.

 

M8 + 21) :1 M8'D2 (6a)
M8°D2 + M8 5’2 2M8°D (6c)
_ [M8°°2]
K = [M8'°]2 _ (6d)
2 [Me-D2108]
75

Popov and coworkers observed a similar change in the

133Cs chemical shift in the reaction of Cs+ with the macro-

cyclic polyether 18-crown-6 (l8C6). In their work, they
used KINFIT to determine the binding constants K1 and

K for the complexes Cs(l8C6)+ and Cs(18C6): respectively.

2
In the present work, their equations have been modified

(see Experimental section for details) to solve for values

of K1, K the chemical shift of the drug for M8°D2,

2' 61'

79

and 52, the chemical shift of the drug for MB-D, using the
data from.Figure 18.

This curve-fitting yields values of K = 6.4(i6.4) x

1
Mfz, K2 = 1.1(to.54) x 103, 51 = 0.47(rO.11), and 52 =

0.86(10.01). Figure 19 shows this theoretical curve

105

plotted against the experimental data. The curve seems to
diverge from the data in the first part of the curve where

the equilibrium defined by K is very large. Its large

1
standard deviation could be an indication that the tech-
nique is insensitive to changes in the data for large
values of K. To test this hypothesis, an attempt was made
to fit the data with K held constant at two orders of

1
magnitude higher and lower than the previously calculated

value and only K2, 51, and 52 allowed to vary.
Figure 20 shows the plots for Kl = constant = 6.4 x
103 and K1 = constant = 6.4 x 107. The values obtained

for the former case appear to diverge away from the experi-
mental points even further. It should also be noted that

the 51 value of -1.7(10.75) appears far too low in compari-

son to the extrapolated value of 0.63 (9:. Figure 19). The

curve obtained for K1 = constant = 6.4 x 107 fits the data
at low MB/D ratios better but the fit is unsatisfactory at
the intermediate values where the equilibrium defined by
K2 becomes important. The calculated value of 51 =

0.67(10.01) agrees well with the extrapolated value.

80

 

  
  
   
 

 

 

 

 

x1 I I I I I I I I I I I I
~ 19.
L2 K2 __
M8.DZ+MG=2M8.D
(Sodium System)
I.l K.-6.4(:6.4Ion5 —
Bobs x K2 - I.I (20.54)on3
(ppm) 3,-0.47(:0.II)
'-° " 32- 0.86 (a 0.0:) —
0.9— _.
L x :2
0.8— —
X
0.7II P‘I.I.I.I.I.I.IIILI.I—
.2 .4 .6 .8 Lo (.2 1.4 (.6 IE 2.0 2.2 24 2.6

Ma/o

Figure 19. Plot of the best fit curve for equations (6a)
and (6c). °obs is the observed chemical shift
in ppm of the EtdBr methyl resonance and M /D
is the ratio of the structured Na25'-GMP t
total drug. X's represent experimental points.
Errors in K's and 5's are standard deviations.

. - K. -6.4xl03 sconst. —'

 

81

 

K
no-zoéua-oz

K
M" 02* MaéZMa' D

E l I l l l l I l l 1 l I T
)—

X

" (Sodium System)

x K2'2.8 (. 20) n03

3. - -I.7 (a 0.75)
02- 0.87 (3 0.0I)

 

 

 

 

 

    

 

 

 

X
0-7 IL) 4441 L1 llllLlJlJ—LI—J—J—LLE
.2 .4 £5 .8 L0 Ii! h4 Ifii L8 2£)EMZ 24 215
MBID
xi 1 1 l l I i I F l l I I

M.+20:'L*M°'D2

12 K2 '°
_ “H'Q2*NMF=i2Nh’D

_ (Sodium System)
U “ K. -6.4x IO7 - cons). "

Bots - K2-3.5 (11.6)xl02
(Rpm)'o_ \ 8,4- o.67(t 0.0I) _

' 32- o.e7(:o.on
(19“ —‘
.— X X X "
0.8— ,. ..._.
X
07—11111 fifillilllrlllmlnlllnl:
.2 A) .6 £3 H3 L2 U1 L6 113120 21! 24 2K5
Ina/D
Figure 20. Plots of the best fits for equations (6a) and

(6c) in which K1 is held constant at a) 6.4 x
103 and b) 6.4 x 107 and only Kg, 51 and 52 are
allowed to vary. Legends are the same as in
Figure 19. Errors in K2 and 5's are standard
deviations.

82

It therefore appears that the value of K1 for equa-

tion (6a) lies somewhere between 105 7 (105 5 K1 5
7

10 ). As a final check on this assumption, the data were

and 10

again fitted allowing K1 and K2 to vary, and 51

held constant at their extrapolated values of 0.63 and 0.87

and 52 were

ppm (EB. Figure 19). The results of this calculation are

shown in Figure 21. The value of K1 = 5.1(il.8) x 106 lies

well within the expected value. Therefore, acceptable

values would be 105 M72 5 K1 5 107 M52, K2 = 1.1(10.54) x
3

10 , 51 = 0.47(10.11), and 52 = 0.86(10.01).

F. Interaction ogEthidium Bromide with Structured and
Unstructured Forms of K25'-GMP

Figure 22 shows the results of adding K25'-GMP into a
0.029M solution of EtdBr at 5°C. As in the case of the
addition of Na25'-GMP to EtdBr an interaction is
apparent between the monomer and the drug. This is char-
acterized as it was in the Na25'—GMP case by a strong up-
field shift of the H(8) monomer resonance. At 0.01B_this
resonance, which has a chemical shift of 8.22 ppm moves
upfield to 7.88 ppm. As the concentration of monomer

K 5'-GMP is increased, the monomer H(8) resonance moves

2
to lower field indicating that for the equilibrium

M + D Z M-D (9)

an excess of monomer (M) is being build up.

83

 

 

 

 

 

 

F I l I l l l I 1 l I l l l
L- K' -(
Me + 20 = M002
1.2 _K; '—
Me'°2*Me*-' 2Me'D
. (Sodium System)
'-' " K. - 5.l (cIeIKIo‘3 7
( “:3 - K2-4.3 (to.7omo2 ~
pp l.0 _ 8, - 0.63 - constant _
82 - 0.87 - constant .
(19- —‘
I— x ___x d
0.8 _ -—d
0.7 l l I 1 ’9‘ XI?) I I 1 l I l I l I I I L 1 l l 1 L ""
.2 .4 .6 .8 1.0 |.2 L4 1.6 LB 2.0 2.2 2.4 2.6

NIB/D

Figure 21. Plot of the best fit for equations (6a) and
(6c) in which 51 and 52 are held constant at
0.63 and 0.87 ppm and K1 and K2 are allowed to
vary. Legends the same as in Figure 19.
Errors in K1 and K2 are standard deviations.

84

Figure 22. Effect on the 1H NMR of adding K 5'-GMP
into 0.029B_EtdBr at 5°C. Nucleotide
H(8) resonances denoted by shading.

85

i J 0.00M'K25'-GMP
I I
)1, . f , . J
”I,)e/)qjyjqfluj/_B\JB\BMBBBM fLBBBBBL—
' I . °-°1§.K25'-GMP

,) I i I . w a
7.7 J/ \,4/.,II [I
,,// J K._./‘ V‘WI
r“ i -
I
!,t| 0.11B’K25'-GMP
H I) :1; I:
U W'I
NI '
\JI

 

 

I [I()hi O 21! K25"'GMP
I: ’ \ N |
i. \
Wm M \
..4- \1 i _. TIN: \~\v/\‘~Je
(I? 3 i I)“ U
I / d ; 0.37B_K25'—GMP

g, W i

I
W \l--. J

-1“. xi.-.)
L111]!)lJIllllillllJlllllllllllllllL]
9 8 7 6 5pm4 3 2 I O

wmb

Figure 22.

86

At 0.11B_K25'-GMP structure formation begins to occur
as evidenced by new lines appearing in the H(8) region of
the spectrum. In addition to the normal K25'-GMP structure

lines which appear between 8.01 and 7.23 ppm,42'43

two

new H(8) resonances appear at 8.56 and 8.32 ppm. These
lines were confirmed as H(8) resonances by recording the
spectrum with deuterium selectively substituted for H(8).85
These lines have never been observed in spectra of struc-
tured K25'-GMP. Their chemical shifts, however, are very
close to those of the two lowest field resonances (9:, B
and B, Figure 8) in the NMR spectrum of structured
Na25'-GMP. This would indicate that EtdBr, at least to
some extent, has structure directing or structure altering
abilities, similar to those of the alkali metal ions.

As was observed in the Na 5'-GMP system, the drug's

2
methyl resonance moves upfield with increasing concentra-
tion of structure. This indicates that EtdBr forms a
"strong" complex with structured K25'-GMP. The upfield
movement of this methyl resonance continues as the total
concentration and hence the structured concentration of
K25'-GMP increase. At 0.21B_total nucleotide concentra-
tion, the drug methyl resonance ceases moving upfield
with further increases in nucleotide concentration and be-
gins to move downfield, as was observed in the Na+ case.

Above a total concentration of nucleotide of 0.52B, the

drug's methyl resonance has a chemical shift of 0.90 ppm

87

independent of nucleotide concentration. It is interesting
to note that at concentrations of total nucleotide higher
than required for the drug methyl resonance to be at its
highest field position, the two new H(8) structure lines
(8.56 and 8.32 ppm) decrease in intensity and eventually
disappear altogether. This would indicate that although
the EtdBr has some structure perturbing properties, the
structuring ability at the K+ ion is far stronger. However,
EtdBr also destabilizes the K+/5'-GMP= self-structure. At

0.37B K 5'—GMP in the presence of the drug about 25% of the

2
nucleotide is in the monomer form (9;, Figure 22). In the

absence of drug at this concentration the monomer represents

less than 10% of the total 5'-GMP= concentration.42

Figure 23 shows a plot of the chemical shift of the
EtdBr methyl resonance as a function of the ratio of struc—

tured K25'-GMP to total EtdBr. The general shape of the

curve is the same as the shape observed in the case of

Na 5'-GMP/EtdBr (BB. Figure 18). Extrapolation of the

2
linear portions provides estimates of the chemical shifts

and stoichiometries of the two K25'-GMP/EtdBr complexes
(0.72 and 0.90 ppm and 4.0 and 11.6 respectively). Due to
the similarities between this system and the Na+ system,
an attempt was made to fit this data to equations (6a) and

(6c) using KINFIT. This calculation yields values of

K1 = 3.1(i1.5) x 104 5'2, K = 2.7(10.86) x 103 14:2:

2

51 = -0.29(i0.32) ppm, and 52 = 0.91(10.01) ppm. A.plot of

88

Structured 5'-0MP- Binding to EtdBr
Potassium Salt

1" ..l.l.l.l.l.l.J.l.l.l.l.l.L.l.

 

 

 

 

 

0‘ "T‘I'I'I‘I'I‘I'I‘I'I'I‘I'r‘I‘
0.0 20 4.0 6.0 8.0 10.0 120 14.0
(5'-cuP' Structured/ EtdBr)

Figure 23. Plot of the chemical shift of the EtdBr methyl
resonance as a function of the structured
K25'-GMP/Drug ratio. Extrapolation of the
linear portions of the curve included.

89

the best fit to the experimental data points is shown in
Figure 24a. The curve appears to fit the data very well.
The only questionable result is the chemical shift of the
M8'D2 species, 51, which at -0.29 ppm would be an enormous
upfield shift (A5 = 1.69 ppm versus the free drug). This
led to the search for an alternative model.
It has been pr0posed by Bouhoutsos-Brown that the

predominant species in solution for a structured K25'—GMP
system is a hexadecamer.50 Therefore an attempt was made

to fit the two step equilibrium

(10)

M -D (11)

+
16 2+M16*'2M

16°D

Figure 24b shows the result of fitting equations (10) and
(11).

It appears obvious from a comparison of the fit of
the curve to the experimental points and also the high
standard deviations for three of the four parameters that
a hexadecamer interacting with the drug molecules does
not fit nearly as well as an octamer interacting with the

drug in the case of K 5'-GMP.

2

90

 

 

 

 

 

 

 

 

 

 

 

 

 

LS‘ I I I I T’ I l l I I I I. l _-
__ "3*ZD=M.'DZ
'2 Ma'thMe=2Me'D
' (Potassium System)
- K. -3.I (215),: I0"
L, K2-2.7Ic0.86)xI03 ._
Bots _ 0.--0.29 (20.32)
(ppm) 02 - 0.9! (30.00
LOr- '—
(_ X
(19- ' '
(18- -‘
Ill lllllllllllJlJlllll
.2 .4 .6 .8 LO [.2 L4 16 LB 2.0 2.2 24 2.6
Mg/D
ISEJ I I’ I I I I l l I I I I I I I I_
K
-x M,.+20 =5 M..-20 -
K
1.2 MIG '02'1’ Mgaé ZM'B'D .—
_ (Potassium System) _
K. - I.2(:Ie.IxI0‘
I.) _x K2- [.4 (8 20.)"IO6 ""
(“so - a.--30(t450.) _
pp 82- 0.99 (20.00
L0- _.
_ x _
X
(ISF- = 4--
0.8_ X —
I.I.I.x:Im.I‘.I.I.I.I.I.I.I.I.I.I.I.I
.08 ‘ .24 .40 .56 .72 .88 1.04 l.20 I.36
"ha/E)
Figure 24. Plots of the best fits for a) equations (6a)

and (6c) and b) equations (10) and (11).

6obs

is the observed chemical shift in ppm of the
EtdBr methyl resonance and a) Mg/D or b) M15/D
is the ratio of structured K25'—GMP to total
drug. Errors are standard deviations.

DISCUSSION

A. Present Study

The results of the curve-fitting for the 5'-GMP= stoi-
chiometry in the presence of excess Na+ indicate that upon
structure formation, two tetramers are brought together to
form an octamer. The fits for both the monomer to n-mer
and monomer to stack tetramer equilbria (equations (la) and
(2a) respectively) indicate the requirement of two tetramers.
Although the standard deviations in the values of n are
fairly large (25% and 22% respectively at 95% confidence)
they provide a range of values from which to make a logical
choice. In contrast, the standard deviations of the values
of K' are extremely high (337% for (1a) and 237% for (2a)
at 95% confidence). Two features of the experimental tech-
nique may help to explain the large errors encountered.
First, although the total concentration of Na+ remains a
constant throughout the experiment, its ratio with respect
to the total 5'-GMP= concentration decreases. At the
beginning, the ratio [Na+]tot/[5'-GMP=]tot = 6.5; whereas
in the final measurement, the ratio = 3.7. This lowering
of the over-all ratio is probably not an important factor

. + =
due to the fact that the ratio of [ha ]tot/[§"GMP1]structo

91

92

which goes from 84 to 11 is far larger than the expected
stoichiometry of Na+. Even if the ratio of bound Na+ to
structured 5'-GMP= was 1:1, the amount of Na+ in the final
solution is still in a ten-fold excess.

A.more serious drawback is the limited concentration
range over which structure formation could be observed.
For the concentration of NaCl employed (0.85%), structure
formation does not occur at 15°C until the concentration of
(TMA)25'-GMP reaches ~ 0.13!. Above the highest concentra-
tion studied ([5'-GMP=] = 0.23%), precipitation occurs,
yielding unusable data. Therefore, the range of usable
5'-GMP= concentrations is only 0.10%. Extrapolation of the
curve in Figure 9a shows it to be sigmoid shaped. The
data used only provide information about slope and position
of the middle portion of the curve and provide no informa-

tion in the region where the [5'-GMP=] versus

structured

[5'-GMP=] curve is horizontal. This leads to uncer-

total
tainties in the K' and n values.

Lowering the Na+ concentration to 0.72g_still did not
permit observation of a large concentration range for struc-
ture formation. Precipitation occurred at the same fraction
of 5'-GMP=.

An examination of Table 2 shows that the tetramer to
stacked tetramer equilibrium (equation (4a)) yields errors

in n and K' that are comparable with those of either the

monomer-n-mer or monomer-stacked tetramer equilibria

93

(equations (la) and (2a) respectively). Within the reported
error limits of the values of n for equation (4a) there are
between five and eight tetramers per structured unit. This
would yield a minimum of three (n = 5 or 6, intensity ratio
2:2:1 or l:l:l respectively) or four (n = 7 or 8) H(8)
structure lines for the most symmetric stacking pattern.

If one considers the second highest H(8) resonance (line 9,
Figure 8) as being due to monomer (which is what was
assumed in the data treatment) then only three structure
lines are observed. Thus the stacking of seven or eight
tetramer units can be eliminated on the basis of the number
of lines observed. The highest and lowest field structure
lines (A_and 2) have been reported40 to arise from the

same structure, which is different from that giving rise to
the second lowest field line (g). A single structure with
five or six stacked tetramer units would be inconsistent
with these observations,thereby eliminating the results of
equation (4a). .

Chantot and Guschlbauer76 and Yee and Chan77 found
strong formation constants for tetramer formation in slight-
ly acidic (pH 3 5) aqueous solution. The lack of an ade-
quate fit for the monomer-tetramer-stacked tetramer equili-
bria (equations (3a) and (3b)) and the tetramer-stacked
tetramer equilibrium (equation (4a)) indicates that at the
pH value studied (~8) the amount of free tetramer in solu-

tion is apparently small. Thus,the results of Chantot and

94

Guschlbauer76 and Yee and Chan77 at acidic pH values do not

appear to be applicable in the present case. The major
difference between their work and the present work appears
to be the size of the charge on the monomer unit. In the
pH 2 5 form the monomer nucleotide unit is monoprotonated
and has a -1 charge, which would give a —4 total charge to
the tetramer. At pH = 8 the monomer has a -2 charge,
yielding a -8 charge to the tetramer. Thus, tetramer forma-
tion should be less favorable at pH = 8 than at pH 2 5
because of increased electrostatic repulsion. As further
indication of the importance of charge on tetramer forma-
tion, the uncharged nucleoside guanosine gels in very dilute

2

(~10- g) solutions in the presence of an alkali metal

76

salt. In this case there is no charge build up in the

aggregate as the tetramers are stacked.
The principal equilibrium in the structuring of
Na25'-GMP is believed to be

2

8M ’ 2 (M 16'

4)2 (12)
In regard to the stoichiometry ofthe‘Na+ in the oc-
tamers, the data presented in Figure 14a support a value of
x = 4.0(¢0.1) and K' = 1.7(i0.7) x 1010 gfll, In contrast
to the large errors encountered in the determination of n
and K' for 5'-GMP= aggregation, the standard deviations of

x and K' at the 95% confidence level are reasonable with

95

respect to this case. The salt concentration employed
allowed investigation over a larger concentration range and
allowed better definition of the data at extreme concentra—
tions of 5'-GMP== by the curve fitting program. The more
complete data set provides lower deviations and supports
the suggestions made earlier for the sources of the errors
in the determination of n and K' for the stoichiometry of
5'-GMP= aggregation.

These data, along with those obtained for 5'-GMP=

aggregation, support the following equilibrium:

12-

4Na+ + 8GMP= : Na4(GMP4)2

(13)

Model building studies make it possible to prOpose
reasonable solution structures. In order to do this, how-
ever, a firmer understanding of the type of binding between
the Na+ and the structured nucleotide is required.

Investigations of the effects of cationic binding to
polyanions have revealed three states in which the cation

86-88 These can be classified as "site bound,"

89

may exist.
"territorial bound," and "free" counterions. Manning
has recently reviewed work on the theory of polyelectrolyte
binding. He defines site binding as direct contact be-
tween the counterion and the polyion without intervening

solvent molecules. This would be analogous to "inner

96

sphere" complexation in coordination chemistry90 or contact
ion pairing. Bound counterions which are not site bound
are therefore considered to be territorially bound. As
defined by Manning, a territorially bound ion has at least
its first hydration sphere intact (i.e., solvent separated
ion pairing). For territorial binding the cation and its
hydration sphere will be strongly pulled to the polyanion,
but it will be free to wander the length of the polymer
chain. In territorial binding, the number of bound cations
is dependent only on the charge on the cation and the
charge density of the polyanion. Territorial binding is
therefore independent of both the ionic strength of the
solution and the specific cation. For example, K+ and Na+
should give the same amount of binding per phosphate charge
irrespective of the ionic strength of the solution. Thus,
a mass-action formulation does not hold for territorial
binding.

As stated earlier, changing the ionic strength, through
changing the total concentration of NaCl, has a pronounced
effect on the amount of structure and hence the amount of
sodium bound. This ionic strength effect implies that the
mode of alkali binding is mass action. Furthermore, the
stability and types of structures formed are highly depen-
dent upon the specific cation employed.41-43’44—47'50

The cation specificity for structure formation indi-

cates that Na+ coordinates to (5'-GMP)%6- via site binding.

97

Clearly, the tetramer hole defined by four 0(6) oxygens
defines a unique site for complexation of the alkali
metal ion which should be quite Specific. The metal ion
size dependence of aggregation should be based upon the
unhydrated metal ion to ligand bond length.

It has been proposed previously that Na+ occupies the
central hole in the tetramer due to a size selective coor-

42,43

dination mechanism (cf. Figure 5a). The stacking of

two tetramers to form an octamer would account for two of

50 has

the four required sodium ions. Bouhoutsos-Brown
pr0posed that the required ions not in the "ion specific
hole" are chelated between the octamer plates by a pair of
negatively charged phosphate groups. This proposal is
based on the fact that addition of excess salt to a struc-
tured system does not stabilize, but destabilizes the
structures. It was proposed that an excess of metal ion
causes ion pairing with the phosphate which would decrease
the number of phosphates available for chelation, thereby
destabilizing the structures. The requirement is to
determine which of the stacking patterns proposed (9;,
Table 6) would allow the best alkali metal ion chelation.
In the case of normal stacking,the two diastereomers
(C4(CW,CW,+30°) and C4(CW,CW,-30°)) have identical overlap
patterns and nearly identical interplate hydrogen bonding

schemes (four OH(3')-OP bonds in the former and four

OH(2')-0P bonds in the latter). Consequently, their

98

energies should be nearly equal, with little preference to
be shown for either in solution. As stated previously, the
highest and lowest field lines (A and 2) probably arise
from isomerization of C4(CW,CW,+30°) and C4(CW,CW,-30°).
This requires that the two remaining Na+ ions, which are
chelated to phosphates in this stacking pattern, should
have equal probabilities of being found on either of the
two diastereomers.

The reversed stacking modes for the D4(CCW,CW,0) and
D4(CW,CCW,O) isomers have markedly different interplate
hydrogen bonding schemes (Table 6). Each D4(CCW,CW,O)
isomer is capable of forming eight hydrogen bonds between
a phosphate group on one plate and an N(Z) position on the
plate above or below it. In the D4(CW,CCW,G) isomers,
only four interplate hydrogen bonds (four OH(2')-O(3')
hydrogen bonds for the D4(CW,CCW,+60°) isomer and four
OH(2')-OP hydrogen bonds for the D4(CW,CCW,+30°) isomer)
are expected per octamer (Table 6). Therefore, the
D4(CCW,CW,G) isomers are expected to be more stable than
the D4(CW,CCW,O) isomers. In addition, model building
studies show that the average phosphate oxygen-phosphate
oxygen distance is ~ 4,2fi in both D4(CCW,CW,9) isomers,
~ 6.52 in the o4(cw,ccw,+3o°) isomer, and ~ 8.32 in the
D4(CCW,CW,+60°). The phOSphate oxygens are ~ 4.83 apart
for the C4(CW,CW,+30°) isomer and ~ 7.88 apart for the

C4(CW,CW,-30°) isomer. Crystallographic Na-O bond

99

distances have been reported for Na ATP (2,3X),91 NaGpC

2
(2.3-2.4X),92 and NaApU (2.3-2.42).93 These distances in-
dicate that the best chelation of a sodium ion by two
phosphate groups occurs when the ioninzed oxygens are

~ 4.6—4.82 apart as they are in the C4(CCW,CW,G) isomers.

Consequently, the second lowest field structure line
(g) is believed to arise from the D4(CCW,CW,-30°) and
D4(CCW,CW,-60°) isomers, which are being interconverted on
the nmr time scale by rapid segmental twisting between
a = +30° and e = -60° about the o4 axis.

In summary, the two octameric structures which appear
in concentrated solutions of Na25'-GMP are assigned to
C4(CW,CW,0) and D4(CCW,CW,O) stacking patterns with rapid
isomerization of +300 and -30° (or -30° and ~600) of the
twist angle (a). The structure directing Na+ ions coordin-
ated to four 0(6) donors are believed to be situated slightly
above the plane of the top tetramer and slightly below the
plane of the bottom tetramer. Since Na+ ions bound to
donor atoms in nucleic acids are always found coordinated
to water94 a water molecule probably occupies a fifth
coordination site. In addition, in order to reduce the
coulombic repulsions between these two sodium ions, it is
proposed that a water molecule intercalates between the
tetramer plates by hydrogen bonding to the 0(6) donors.
Assuming a 3.52 stacking distance and an OH-O hydrogen bond

distance of 2.73,95 an intercalated water can hydrogen bond

100

to an 0(6) donor on one tetramer plate and to an 0(6) on the
second plate which is twisted by 9 = 600 with respect to
the first. The remaining two Na+ ions per octamer unit are
chelated between a phosphate group on the top tetramer and
a phosphate group on the bottom tetramer. The lifetime of
the metal ion at a specific site is small on the NMR time
scale (correlation time, To = 5—8n396). Therefore, the
symmetry as observed by 1H NMR is unaffected by the site
specificity of the Na+ binding. 0n the NMR time scale,
each phosphate will chelate an Na+ part of the time. In
addition to metal ion coordination, each of the two struc-
tures is expected to have either four (C4(CW,CW,@)) or
eight (D4(CCW,CW,O)) interplate hydrogen bonds increasing
the stability (Table 6).

A schematic representation of the structures proposed

is illustrated in Figure 25.

B. Other Proposed Structures

In the past five years, several papers have proposed
structures for the self-assembled form of Na25'-GMP. A
discussion of these in relation to the structures proposed
here should be informative with respect to the validity of
any or all of them.

Laszlo and coworkers have presented several different

models to explain the metal ion specific binding to

96

5'-GMP= by Na+. Their first model assumed the structures

23

to be micelles. Based on Na NMR lineshape analysis, they

101

 

r——| Totromor ‘ H Unit a h
"' o. O Na"

I'I , H\.
Totromor Ufli' ‘ ]'_'|R ibophosphote

 

 

 

Figure 25. Schematic diagram of the proposed structure
for the self-assembled forms of Na25'-GMP.

102

prOposed that the structures contain between 12 and 20
monomer units. A preference for a structure based upon 16
monomer units was stated but not substantiated. No infor-
mation concerning the nature of the structures was pre-
sented. In 1978, they pr0posed the coordination of the
Na+ ion into the central cavity of the tetramer plate.44
A week later they proposed the coordination of K+ at the
same position on the tetramer.4S Their model for the K+
structure was based on the observation that when KCl is

added to a dilute solution of Na 5'-GMP (0.1g, 31°C) the

23

2
Na linewidth shows a dramatic increase (Avl5 = 150 Hz)

indicative of structure formation. An analogous change in

39

the K linewidth (Av35 2 50 Hz) was taken as indication

that K+ and Na+ structure in the same position. The addi-
tion of LiCl, NaCl, RbCl, or CsCl had no effect on the
23Na linewidth under the same conditions. No explanation
was provided with regards to how the same coordination site
can accommodate both the K+ and the Na+ ions and no pro-
posals of nucleotide stoichiometry were advanced.

31

Based on 1H and P NMR of Na25'-GMP, Borzo and

Laszlo46 accepted the proposal made by Pinnavaia, et al.40
that the Na25'-GMP self-Structure contains two distinctly
different structures in solution. The structures they
proposed are illustrated on Figure 26. The structure
labelled A on Figure 26 was presumed by Borzo and Laszlo46

to represent the structure which gives rise to the highest

Figure 26.

103

 

Structures pEOposed by Borzo and Laszlo46 to

explain the H NMR spectrum of the self-
assembled forms of Na25'—GMP. Open circles re-
present phosphate groups pointing down while
closed circles represent phosphate groups
pointing up. Na+ ions represented by striped
circles (from ref. 46).

104

and lowest field H(8) resonances (A and 2, E£° Figure 8).
The highest and lowest field lines observed in the 31p
spectrum were also assigned to structure A. The Na+ ion
was presumed to be coordinated in the core position of the
tetramer with a fifth coordination site occupied by a
phosphate group on the adjacent tetramer unit. The rapid
equilibrium between states A} and A? (Figure 26) was pre-
sumed to impart two-fold symmetry to the octamer units with
the open and shaded circles representing phosphate groups
pointing down and up respectively. The rapid equilibrium
of Bf and BI would give the time averaged structure B?
which was assumed to give rise to the second lowest field
structure line (B, Figure 8).

Species A_(Figure 26) can be ignored on several
grounds. First, the authors completely ignore the hydrogen
bond directionality of the tetramers. It is clear from

previous arguments that in order for a C axis to exist

2
between the tetramer plates, the tetramers have to be
stacked in one of the reversed modes. Second, back and

forth slippage of the tetramer plates along only one direc—

tion requires a rigid glycosidic (x C(9)-C(l')) conforma-

tion, in which half of the nucleotide units in the octamer

are in a syn conformation (x = 215°)80 and half are in an
anti conformation (x = 35°).80 Even if this unlikely

conformation existed, the absence of a mirror plane in both

the normal and reverse stacking modes would cause this

105

motion to give rise to more than two lines. The model also
permits very little overlap of the purine bases, which is
an essential contributor to the stability of the struc-

tures.“’17

Finally, model building studies show that the
phosphates not chelated to the central Na+ ions are too far
apart (>128) to allow chelation of additional metal ions.

The model could have some merit, however, if facile
rotation between §y2_and anti conformations is allowed and
if segmental slippage is allowed along two directions to
permit migration of one tetramer unit around the periphery
of the other. Such a motion would result in a two-line
spectrum for normal stacking. Each of the two diastereo-
mers resulting from reversed stacking would give a one—
line H (8) spectrum.

In addition, no explanation was provided as to how the
B structure would give rise to only one H(8) resonance. It
is clear, however, from previous arguments, that the tetra-
mers would have to be stacked in a reversed stacking mode.

Again using 23Na NMR lineshape analysis, Deville,

Detellier, and Laszlo47

predicted structures for the
K25'-GMP self-structures in solution. This paper proposed
the same structure for the potassium self-structure as they
postulated previously for the sodium system (Figure 26).46
The structure labelled A (Figure 26) was changed to include
Rf in the two core positions of the tetramer. This time,

the A} I‘A" equilibrium was not considered and the single

106

"staircase" model was assumed to have C2h symmetry. The
authors state that this model gives rise to three of the
H(8) nmr resonances observed for the potassium salt with
an integral intensity of 1:2:1. They also state, with
no apparent justification, that structure A_is energet-
ically more favorable than structure B. If, however, the
hydrophobicity of the base stacking is such an important

14'17 structure

stabilizing factor in structure formation,
B should be the more favorable species. Model studies
have shown only limited overlap of two of the eight guanine
bases for structure B. The overlap patterns of B should
be nearly the same as in Figures 11, 12, and 13.

In the two most recent papers presented by Laszlo and

97,98 46,47

coworkers, the model A proposed previously

(Figure 26) is entirely abandoned without explanation. In
these two papers, lH, 23Na, 31P, 39K, and 87Rb NMR were
used to determine the size of the structures and the sites
of binding. Again using 23Na lineshape analysis, they
examined the mixed K+/Na+-5'-GMP= structured system. In
their model, they proposed that the tetrameric species is
by far the predominant form of the monomer. They infer
that this tetramer formation requires no metal ion inter-
action. This assumption is in conflict with the findings
76

of Chantot and Guschlbauer that the nucleoside guanosine

does not gel without the presence of an alkali metal salt.

107

98
Laszlo, et a1., attempted to fit a three step

equilibrium for a mixed Na+/K+-5'-GMP= system.

K1

G4 + K+ z G4,K+ (14)
K2
+ +
G4,K + G4 2 GB,K (15)
K3
G8,K+ + nNa+ z G8,K+,nNa+ (16)

where G = 5'-GMP= monomer unit.

2

From the linear dependence of pr against K+

(pB = fraction of Na+ bound, x = 23Na quadrupole coupling
constant) they found 2.1 < n < 5.6. They reasoned that
due to the four-fold symmetry of the tetramer a value

of n = 4 was the most reasonable. Their proposed structure

for the K+/Na+-5'-GMP= structure is shown on Figure 27.

Figure 27. Structure proposed by Detellier and Laszlo98 to

explain the NMR results for the K+/Na -5'-GMP=
self-structure. Only two of the four chelated
Na ions are illustrated (from ref. 98).

 

 

 

     

 

 

108

The value of n arrived at for the metal stoichiometry

in the K+/Na+-5'-GMP= was then applied to the structured

97

system with no added K+. Here, by using what they termed

a "phase separation model," they calculated thermodynamic
parameters based upon a stoichiometry of Na5(5'-GMP);1- in
which it was assumed that four Na+ ions were chelated
between phosphates and the fifth occupied the central core
position. Although claiming to have abandoned the micell-

96

ular model of a previous paper the "phenomenological

equations" used were based on those used for micelle for-

97

mation. In addition, the equations used were for a mass-

action process and not a phase separation process as they
claimed.99

With the basic Na5(5'-GMP)él- stoichiometry, Laszlo
and coworkers proceeded to propose a series of structures
to explain the spectra. They favored one in which two major
equilibria exist

26 (16)

++
63

2G (17)

Mr
0

16

The G species was presumed to give rise to the second

8
lowest field H(8) structure line (B, Figure 8) in which top
and bottom H(8) positions are equivalent. This would cor-

respond to a reversed stacking mode in the present study

109

(9;, Figures 12 and 13). Laszlo and coworkers state that
this type of stacking for the octamer has either C4h or S8
symmetry. However, these point groups are meaningless be-
cause they ignore the directionality of the hydrogen
bonding in the tetramer units and the presencecifthe chiral
ribose-phosphate. The two outer lines (B and B, Figure 8)
are claimed to be due to dimerization of the GB species to

form the G species. They presumed that G could lead

16 16
to a species with two-fold symmetry; the two "outer" tetra-
mers would be equivalent, and the two "inner" tetramers
would be equivalent. This would be equivalent to the
reversed stacking of a pair of reversed octamer stacks
(e.g., CW,CCW,CW,CCW) or reversed stacking of a pair of
normal octamer stacks with opposite hydrogen bond direc-
tionality (e.g., CW,CW,CCW,CCW). No explanation was offered
as to why other G16 and G8 stacking patterns were not ob-
served. They also stated that the B line (Figure 8) could
arise from free tetramer and the B and B lines could arise
from octamer.

In summary, the models that have been proposed by
Laszlo and coworkers are of only limited value due to their

total neglect of both the hydrogen bond directionality and

the ribose-phosphate chirality.

100 13

Fisk has recently investigated the C NMR of

structured Na25'-GMP. In her work, she determined the

longitudinal and transverse relaxation rates (T1 and T2)

110

and the Nuclear Overhauser Enhancement (NOE) for the car-
bons in the structured form. Using W'oessner'slo1 treatment
for determining the relaxation rates and NOE's of nuclei

in a rotating spheroid, she then calculated T1, T2, and NOE
values for a series of stacked tetramers between n = 4 and
n = 64. She found that four models would yield the same
values as were obtained experimentally. These were

i) a hydrated octamer; n = 0.074

ii) an unhydrated 44-mer; n = 0.074

iii) a hydrated 52-mer; n = 0.019

iv) an unhydrated 64-mer; n = 0.019
where n is the solution viscosity (n = 0.019 is the vis-
cosity of D20 at 5°C; n = 0.074 is the viscosity of the
solution of Na25'-GMP at 5°C). From this she supported the
hydrated octamer. The remaining structures could not ex-
plain the three H(8) structure lines observed.

Based on mixed salt experiments, Bouhoutsos-Brown50
has proposed a Na+/5'-GMP= structure with six Na+ ions for
every octameric structured unit. Each octamer was pre-
sumed to have a sodium ion in the center of each tetramer.
In addition, four other sodium ions were presumed to be
chelated between phosphate groups on the top and bottom
plates. The fact that Li+ tended to decrease structure
formation was taken to mean that the strong lithium phos-

phate bond would break up chelation and cause the formation

of "free" tetramers in the solution. The observation of a

111

fifth structure line (6 = 7.6 ppm) was taken as the chemi-
cal shift of the tetramer in solution. The same solution
showed a very strong "monomer" resonance indicating that
only under certain special conditions can a sizable amount
of tetramer be observed in solution, an assumption that the
present study supports.

Finally, a structurally similar polynucleotide, poly-
inosine, has recently been shown to have a melting tempera—
ture which is highly dependent upon the alkali metal salt
which is added into the solution.48 Inosine differs from
guanosine in only having a hydrogen at position 2 of the
base (9:. Figure 3). This means that poly(I) can still
form stacked tetramers of the bases from the tetramerization

of four strands,102'103

however, only one hydrogen bond is
expected per base unit. The metal ion dependence of the
melting profiles implies the size selective coordination

of the alkali metal ions to either the center of the
tetrameric plates or to octahedral holes between the plates.
The stability of the ordered polynucleotide was a direct
function of the fit of the ion into the coordination site.

This is analogous to the size selective coordination pro-

posed here for the 5'-GMP= self-structure.

C. Stability of the Base Overlap
With the stacking patterns pr0posed, the question of

stability of the overlap of the bases must be addressed.

104

Gupta and Sasisekharan have done theoretical calculations

112

to determine the stability of various overlap patterns of
the free bases. These calculations were based on monopole-
monopole interactions of a pair of overlapping bases.105
Their results indicated that the reversed stacking mode is
energetically more favorable than the normal stacking mode.
In addition, the enantiomeric pairs, D4(CW,CCW,+30°) and
D4(CCW,CW,-30°), are very close in overlap to the lowest
energy structure. This structure is pr0posed to be

12.0 kca1/2 moles more stable than the free bases. These
calculations do not rule out the normal stacking pattern,
however. A stacking pattern similar to the normal stacking
pattern proposed here would still be lower in energy than
the two free base units. It must be pointed out, moreover,
that the utility of using these calculated numbers is
extremely limited due to the fact that contributions from
the bulky ribose-phosphate groups have not been considered
in their work.

Table 6 lists the expected upfield shifts of the H(8)
protons from Giessner-Prettre and Pullman's theoretical
calculations considering only ring current and atomic
diamagnetic susceptibility effects.82 The expected upfield
shift of a proton depending on its position is shown on
Figure 28. A comparison of Table 6 with the spectra on
Figure 8 shows no correlation between the calculated shifts

and the observed shifts. However, the theoretical curves

(Figure 28) also would predict no downfield shift of

113

 

 

Figure 28. Expe§ted upfield shift (in ppm) of a nucleus
3 5 above or below the aromatic plane due to
the ring current and diamagnetic suscepti-
bility effects (from ref. 82).

114

structured protons as observed for peaks B_and B (2::
Figure 8). This indicates that other effects (e.g., prox-
imity to phosphate charge, electron withdrawal upon hydro-
gen bonding to N(7)) not taken into account in the theo-
retical calculation82 are also effecting the chemical

shifts of H(8).

D. Drug Intercalation

It has been clearly shown that EtdBr interacts with
both monomeric and structured forms of Na25'-GMP and
K25'-GMP. Binding to the monomer is characterized by a
large (A6 = 0.3 ppm) upfield shift of the H(8) monomer
resonance along with corresponding upfield shifts (~ 0.2
PPm) of the aromatic protons of the drug. This interaction
is most likely due to the base stacking of monomer S'-GMP=
and monomer drug.

The spectral changes assigned to the structure-drug
complex on the other hand are observed only at concentra—
tions of nucleotide sufficiently large for the observation
of structure formation. At this point a large upfield
shift (A6 2 0.50 ppm) is observed for the methyl resonance
of the drug molecule. Due to the fact that only small
changes in the chemical shifts of the structured H(8)
resonances (A6 < 0.05 ppm) are observed,it is assumed that
the mode of binding is for the drug molecules to stack on
the outside of the octamer units. If the mode of binding

were intercalation, the symmetries of the stacking modes

115

would be expected to undergo dramatic changes with corres-
pondingly dramatic changes in the H(8) chemical shifts.

The changes in the aromatic drug resonances would indicate
that the mode of binding occurs through aromatic base
stacking of the drug with the octamer. Simple ion pairing
of the anionic octamer with the cationic drug would not be
expected to induce the large shifts observed for the drug's
aromatic protons.

This mode of structure binding would add further evi-
dence to the belief that the B_resonance (Figure 8) is
monomer and not tetramer. If the mode of structure-drug
binding is through aromatic overlap to the outside of the
octamers, the surfaces available to the drug should be
nearly identical in the free tetramer. As such, a large
upfield shift of the drug's methyl resonance would be ex-
pected. Since only small perturbations of this methyl
resonance are observed at concentrations where only the B
resonance is observed, this resonance is assumed to be due
to free monomer.

The success in fitting the equilibria for the binding
of two drug molecules and then one drug molecule to the
octamer of Na25'-GMP (equations (6a) and (6c) respectively)
indicates two important points. First, the fit of these
two equations supports the evidence that an octamer is the
predominant species for structured Na25'-GMP.T Second, when

the binding of the drug to the structured form occurs, it

116

takes place via a two step process. First, two drug mole-
cules bind to the outside of the octamer. As more octamer
is built up, it then competes for some of the drug already
bound, allowing only a 1:1 (drug:structure) species to
remain. In addition, this binding is to the monomeric
drug molecule and not a stacked dimer. At the concentra-
tions of drug employed, the EtdBr is predominantly in the
monomeric form (>90%).106

The 2:1 (drug:structure) complex is expected to be
binding of both drug molecules to the same aromatic face of
the octamer unit. If binding of the two drugs were to
opposite faces of the octamer, when the 1:1 complex is
formed, little change would be expected in the orientation
of the remaining drug on the octamer due to the lack of any
direct interactions. The lack of direct interaction should
invoke no change in the chemical shift of the drug's methyl
resonance in going from the 2:1 to the 1:1 species, which
is not observed. Figure 29 shows schematically the type
of binding expected for the two complexes.

The case of the binding of the drug to structured forms

of K 5'-GMP is quite different from the sodium case. With

2
the potassium self—structure, a dramatic change in the
structure lines is observed. Most noteworthy is the
appearance of two new H(8) structure resonances (6 = 8.56

and 8.32 Ppm) downfield from the monomer resonance. This

indicates that EtdBr has some structure directing abilities

117

 

. 2 c: —4.__ MI 3,...
' Nd+

N2:'”

 

Hf DP P

 

 

 

 

 

 

 

 

N Na++ No }Na+ ‘__-K’—‘ - EL“
9.3+ as ' 43*

R] :1”

Figure 29. Schematic diagram of the proposed equilibrium

involved in the formation of a 2:1 and 1:1
complex between the intercalation drug EtdBr
and the octameric forms of Na25'-GMP.

118

of its own. The fact that the equations for the binding of
the drug molecules to octameric species (equations (6a) and
(6a)) provide a better fit to the experimental data than

the equations for the drug binding to a hexadecamer (equa—
tions (10) and (11)) supports the same binding mode for
K25'-GMP as was proposed for Na25'-GMP (Figure 29). As fur-
ther evidence of the octameric form of the structured potas-
sium-drug complexes, the new structured H(8) resonances
observed have nearly the same chemical shift as those

observed in the sodium case (6 = 8.56 and 8.32 ppm for

K25'-GMP-EtdBr and 6 = 8.48 and 8.33 ppm for Na25'—GMP-
EtdBr) which is proposed as having an octameric structure.
Bouhoutsos-Brown50 has proposed a structure for what

she terms the "complex" potassium structure as arising
from dimerization of octamers to form a hexadecamer. In
her model, a K+ ion occupies the core position of each
octamer coordinated to the eight 0(6) positions of the
two tetramers. The dimerization of the two octamers is
facilitated by chelation of K+ ions between phosphate
groups of each octamer.

It is proposed in this case that the drug, ethidium
bromide, is capable of disrupting the hexadecameric forms
of K25'-GMP in favor of an octameric form when the concen-
tration of the structured form is low. The mode of action

suggested is one in which the drug molecule attempts to

intercalate between the two octamers of the hexadecamer.

119

In separating the two octamers, the chelation of K+ ions
which holds them together is no longer possible. The newly
formed octamer-drug complex gives rise to the additional
H(8) structure lines (6 = 8.56 and 8.32 ppm). As the con-
centration of structure increases, however, the hexadeca-
meric structure becomes the predominant species and the
equilibrium favors the binding of the drug to the other
surface of the hexadecamer. A concomitant disappearance
of the two new H(8) resonances assigned to octamer occurs
as the equilibrium is shifted back towards the hexadecamer.
In addition, the chemical shifts of the H(8) structure
resonances approach those observed in the absence of added

42,43,50

drug. A schematic representation of the proposed

binding modes is shown in Figure 30.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 30. Schematic diagram of the proposed equilibrium
' involved in the formation of a 2:1 and 1:1
complex between the intercalation drug EtdBr
and structured forms of K25'—GMP.

CONCLUSION

The work presented in this thesis has shown evidence
that structure formation by Na25'-GMP can best be described
in terms of the stacking of two hydrogen bonded tetramer
units to form an octamer. The most likely structures
are of two types. One type, which gives rise to the outer
set of H(8) structure lines, arises from the normal
stacking of two tetramers to form a unit with C4 symmetry.
A second structure, which is reSponsible for the inner H(8)
structure line, has D4 symmetry due to the reversed
stacking mode in which the top tetramer has hydrogen bonds
in a counterclockwise fashion and clockwise on the bottom
tetramer. In both cases, twisting of the tetramer plates
is rapid on the NMR time scale while stack separation is
considered slow. The amount of free tetramer in solution
is believed to be small, based on both the inability to fit
the data to equations which allow a large tetramer build up
and the fact that no evidence of any tetramer—drug interac-
tion was found.

Stabilization of the octamers is possible through two
mechanisms. First, sodium ions are site bound to two posi-
tions of the octamer. Two sodiums are coordinated to four

carbonyl oxygens in the center of each tetrameric face. A

121

122

water molecule intercalated between the plates, is proposed
to decrease coulombic repulsions between these two ions.
The remaining two Na+ ions are chelated between a.pair of
phosphate groups on opposite tetrameric plates. The life—
time of a given sodium ion at a given site is considered
short on the NMR time scale. The second form of stabili—
zation is through interplate hydrogen bonding. Model
building studies predict a minimum of four and up to eight
hydrogen bonds per octamer unit.

Finally, the present study has shown that the struc—
tured forms of Na25'—GMP and K25'-GMP interact very strongly
with the intercalation drug, ethidium bromide.

In the sodium case, evidence has been shown for the
strong binding of two drug molecules to the octamer unit
and the less strong binding of a 1:1 (octamer:drug molecule)
complex. The binding is presumed to be to the exterior
aromatic faces of the octamer since only slight changes in
the chemical shifts of the H(8) structure lines are
observed.

The potassium self-structures, however, show dramatic
changes for the H(8) structure lines upon the addition of
the drug. It is presumed that this is evidence that the
drug molecules are able to disrupt the hexadecameric
structures of K25'-GMP. The drug then forms a strong
2:1 complex with the octamer and weaker 1:1 complex. A

build up in solution of the hexadecameric K25‘-GMP

123

structure shifts the equilibrium towards binding of the

drug to the hexadecamer.

BI BLIOGRAPHY

10.

11.
12.

13.

14.

BIBLIOGRAPHY

"Die Histochemischen und Physiologischen Arbeiten von
Friedrich Miescher," l and 2, F.C.W. Vogel, Leipzig
(1897).

Sci. Cit. Index, Perm. Subj. Index, 11, 42635(1978).
A. L. Lehninger, "Biochemistry: The Moelcular Basis
of Cell Structure and Function," World Publishers,
Inc., N.Y., 1970, Ch. 12, 25, and 28.

J. D. Watson and F. H. C. Crick, Nature, 171, 737
(1953).

 

J. D. Watson and F. H. C. Crick, Nature, 171, 964

 

J. Pitha, R. N. Jones, and P. Pithova, Can. J. Chem.,
44,1045(1966)-

 

Y. Kyogoku, R. Lord, and A. Rich, Science, 154, 518
(1966).

E. Kuchler and J. Derkosch, Z. Naturforsch., Bl,
209(1966).

 

R. M. Hamlin, Jr., R. C. Lord, and A. Rich, Science,
148, 1734(1965).

R. R. Shoup, H. T. Miles, and E. D. Becker, Biochem.
Biophys. Res. Comm., 2,194(1966).

L. Katz and S. Penman, J. Mol. Biol., 1B, 220(1966).

 

S. M. wang and No Co Li, J. Am. Chem. 5°C.] 19’
5069(1968).

 

M. Raszka and N. 0. Kaplan, Proc. Nat. Acad. Sci. USA,
BB, 2025(1972).

P. O. P. Ts'o, in "Basic Principles in Nucleic Acid

Chemistry," P. O. P. Ts'o, Ed., Academic Press, New

124

15.

16.

17.

18.

19.

20.

21.

22.

23.
24.

25.

26.

27.

28.

29.

30.

31.

125

W. Schimmack, H. Sapper, and W. Lohmann, Biophys.
Struct. Mechanism, l, 311(1975).

 

E. Plesiewicz, E. Stepien, K. Krystyna, and K. L.
Wierzchowski, Nucl. Acids Res., B, 1295(1976).

 

P. O. P. Ts'o, in "Basic Principles in Nucleic Acid
Chemistry," P. O. P. Ts'o, Ed., Academic Press, New
York, 1974, Vol. 1, p. 561.

D. M. Crothers and D. I. Ratner, Biochem., Z, 1823
(1968).

S. I. Chan, M. P. Schweizer, P. O. P. Ts'o, and
G. K. Helmkamp, J. Am. Chem. Soc., BB, 4182(1964).

 

M. P. Schweizer, S. I. Chan, and P. O. P. Ts'o,
J. Am. Chem. Soc., Bl, 5241(1965).

 

A. D. Broom, M. P. Schweizer, and P. O. P. Ts'o,
J. Am. Chem. Soc., BB, 3612(1967).

 

M. P. Schweizer, A. D. Broom, P. O. P. Ts'o, and
D. P. Hollis, J. Am. Chem. Soc., BB, 1042(1968).

 

I. Bang, Bioch.z., BB, 293(1910).

H. T. Miles and J. Frazier, Biochim. Biophy_. Acta,
12, 216(1964).

 

F. B. Howard and H. T. Miles, J. Biol. Chem., 240,
801(1965).

 

J. P.-Leicknam, C. Chavelier, J. F. Chantot, and
W. Guschlbauer, Biophys. Chem., B, 134(1973).

 

J. F. Chantot and W. Guschlbauer, FEBS Lett., 1,
173(1969).

 

J. F. Chantot, M. T. Sarocchi, and W. Guschlbauer,
Biochimie, BB, 347(1971).

 

E. W. Small and W. Peticolas, Biopolymers, BB, 1377
(1971).

 

V. Gramlich, H. Klump, R. Herbeck, and E. D. Schmid,
FEBS Lett., BB, 15(1976).

 

J. M. Delabar and W. Guschlbauer, Biopolymers, BB,
2073(1979).

32.

33.

34.

35.

36.

37.
38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

126
C. Becker and T. Ackerman, Ber. Bunsenges Phys. Chem.,
21, 203(1973).

M. Gellert, M. N. Lipsett, and D. R. Davies, Proc.
Nat. Acad. Sci. USA, BB, 2013(1962).

J. Iball, C. H. Morgan, and H. R. Wilson, Nature,
199, 688(1969).

P. Tougard, J. F. Chantot, W. Guschlbauer, Biochim.
Biophys. Acta, 308, 9(1973).

S. B. Zimmerman, G. H. Cohen, and D. R. Davies,
J. Mol. Biol.,BB, 171(1975).

S. B. Zimmerman, J. Mol. Biol., 100, 663(1976).

H. T. Miles and J. Frazier, Biochem. Biophys. Res.
Comm., BB, 199(1972).

 

R. M. Bock, N. S. Ling, S. A. Morell, and S. H.
Lipton, Arch. Biochem. Biophys., BB, 253(1956).

T. J. Pinnavaia, H. T. Miles,and E. D. Becker, 9;
Am. Chem. Soc., B1, 7198(1975).

C. M. Mettler, M. S. Thesis, Michigan State Univer-
sity, East Lansing, Michigan, 1977.

C. L. Marshall, M. S. Thesis, Michigan State Univer—
sity, East Lansing, Michigan, 1977.

T. J. Pinnavaia, C. L. Marshall, C. M. Mettler, C.
L. Fisk, H. T. Miles, and E. D. Becker, J. Am. Chem.
Soc., 100, 3625(1978).

 

 

A. Paris and P. Laszlo, C. R. Hebd. Seances Acad.
Sci., Ser. D, 286, 717(1978).

 

 

C. Detellier, A. Paris, and P. Laszlo, C. R. Hebd.
Seances Acad. Sci., Ser. D, 286, 781(1978).

 

M. Borzo and P. Laszlo, C. R. Hebd. Seances Acad.

———

Sci., Ser. C, 287, 475(1978).

 

A. Delville, C. Detellier, and P. Laszlo, J. Mag.
Reson., BB, 301(1979).

H. T. Miles and J. Frazier, J. Am. Chem. Soc., 100,
8037(1978).

49.

50.

51.

52.

53.
54.

55.

56.

57.
58.
59.
60.

61.

62.

63.

64.

65.

127

K. B. Roy, J. Frazier, and H. T. Miles, Biopolymers,
BB, 3077(1979).

 

E. Bouhoutsos-Brown, Ph.D. Thesis, Michigan State
University, East Lansing, Michigan,l980.

J. E. Huheey, "Inorganic Chemistry: Principles of
Structure and Reactivity,” Second Edition, Harper
and Row, N.Y., 1978, pp. 72-73.

E. F. Gale, E. F. Cundliff, P. E. Reynolds, M. H.

Richmonds, and M. J. Waring, "The Molecular Biology
of Antiobiotic Action," Wiley, New York, 1972.

L. S. Lerman, J. Mol. Biol., B, 18(1961).

 

T. R. Krugh and M. E. Nuss in "Biological Applica-
tions of Magnetic Resonance," R. G. Shulman, Ed.,
Academic Press, New York, 1979, pp. 113—175.

B. A. Newton in "Advances in Chemotherapy," A.
Goldin and F. Hawking, Eds., Academic Press, New
York, Vol. 1, 1964, p. 35.

L. Dickinson, D. H. Chantrill, G. W. Inkley, and
M. J. Thompson, Brit. J. Pharmacol., B, 139(1953).

 

J. F. Henderson, Cancer Res., BB, 491(1963).

 

W. H. Elliot, Biochem. J., BB, 562(1963).

 

M. J. Waring, Biochim. Biophys. Acta, BB, 358(1964).

 

J.-B. LePecq and C. Paoletti, C. R. Acad. Sci. Paris,
260, 7033(1965).

 

J.-B. LePecq and C. Paoletti, J. Mol. Biol., BB,
87(1967).

 

V. W. F. Burns, Arch. Biochem. Biophys., 133, 420
(1969).

 

J. Vinograd and W. R. Bauer, J. Mol. Biol., BB, 141
(1968).

 

J. Vinograd and W. R. Bauer, J. Mol. Biol., BB,
281(1970).

J.-B. LePecq in "Methods of Biochemical Analysis,"
D. Glick, Ed., Wiley Interscience, New York, Vol. 20,
pp. 42-86, 19710

66.

67.

68.

69.

70.

71

72.

73.

74.
75.

76.

77.

78.

79.
80.

81.

82.

128

L. P. G. Wakelin and M. J. Waring, Mol. Pharm.,
B, 544(1974).

 

T. R. Krugh, F. N. Wittlin, and S. P. Cramer, Bio-
polygers, BB, 197(1975).

T. R. Krugh and C. G. Reinhardt, J. Mol. Biol., BB,
133(1975).

 

G. P. Kreishman, S. I. Chan, and W. Bauer, J. Mol.
Biol., BB, 45(1971).

G. M. Intille, Ph.D. Thesis, Syracuse University,
Syracuse, New York, 1970.

D. A. Skoog and D. M. west, "Analytical Chemistry,
An Introduction,” Third Edition, Holt, Rhinehart,
and Winston, Inc., New York, 1979, p. 438.

E. Volkin and W. E. Cohn in "Methods of Biochemical
Analysis," D. Glick, Ed., Interscience, New York,
1954, p. 304.

J. L. Dye and V. A. Nicely, J. Chem. Educ., BB,
443(1971).

 

W. E. wentworth, J. Chem. Educ., BB, 96(1965).

 

E. Mei, J. L. Dye, and A. I. Popov, J. Am. Chem. Soc.,
BB, 5308(1977).

J. F. Chantot and W. Guschlbauer, Jerusalem Symp.
Quant. Chem. Biochem., B, 205(1972).

 

G. C. Yee and S. I. Chan, J. Am. Chem. Soc., BB,
3218(1972).

M. Borzo, C. Detellier, P. Laszlo, and A. Paris,
J. Am. Chem. Soc., 102, 1124(1980).

 

M. Sundaralingam, Biopolymers, B, 821(1969).

M. Sundaralingam, Jerusalem Symp. Quant. Chem.
Biochem., B, 417(1973).

M. Sundaralingam, Int. J.,guant. Biol. Symp., B, 81
(1974).

C. Giessner-Prettre and B. Pullman, Biochem. Biophys.
Res. Comm., BB, 578(1976).

 

83.

84.

85.

86.
87.
88.

89.
90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

129

K. W. Jennette, S. J. Lippard, G. A. Vassiliades,
and W. R. Bauer, Proc. Nat. Acad. Sci. USA, BB,
3839(1974).

 

M. J. waring, Biochim. Biophys. Acta, 114, 234(1966).

 

F. J. Bullock and O. Jardetsky, J. Org. Chem., BB,
1988(1964).

 

R. Zana and C. Tondre, Biophys. Chem., B, 367(1974).

 

P. Spegt and G. weill, Biophys. Chem., 4, 143(1976).

 

P. C. Karenzi, B. Meurer, P. Spegt, and G. Weill,
Biophys. Chem., B, 181(1979).

 

G. S. Manning. Acc. Chem. Res., BB, 443(1979).

 

F. Basolo and R. G. Pearson, "Mechanisms of Inorganic
Reactions, A Study of Metal Complexes in Solution,"
Second Edition, John Wiley and Sons, Inc., New York,
1967, PP. 34-35.

0. Kennard, N. W. Isaacs, W. D. S. Motherwell, J. C.
Coppola, D. L. wampler, A. C. Larson, and D. G.
Watson, Proc. R. Soc. London Ser. A., 325, 401 (1971).

 

J. M. Roseberg, N. C. Seeman, R. 0. Day, and A. Rich,
3. Mol. Biol., 104, 145(1976).

 

N. C. Seeman, J. M. Rosenberg, F. L. Suddath, J. J.
P. Kim, and A. Rich, J. Mol. Biol., 104, 109(1976).

V. Swaminthan and M. Sundaralingam, Crit. Rev. Bio-
Chem., B, 245(1979).

 

J. E. Huheey, "Inorganic Chemistry: Principles of
Structure and Reactivity," Harper and Row, New York,
1972, p. 194.

A. Paris and P. Laszlo, Am. Chem. Soc. §ym . Ser.,
BB, 418(1976).

M. Borzo, C. Detellier, P. Laszlo, A. Paris, J. Am.
Chem. Soc., 102, 1124(1980).

C. Detellier and P. Laszlo, J. Am. Chem. Soc., 102,
1135(1980).

 

L. R. Fisher and D. G. Oakenfull, Quart. Rev. Chem.
Soc., 1977, 25(1977).

 

130

100. C. L. Fisk, personal communication.

101. D. E. Woessner, J. Chem. Phys., B1, 647(1962).

 

102. S. B. Zimmerman, G. H. Cohen, and D. R. Davies,
J. Mol. Biol., BB, 181(1975).

 

103. S. Arnott, R. Chandrasekaran, and C. M. Marttila,
Biochem. J., 141, 537(1974).

104. G. Gupta and V. Sasisekharan, Nucl. Acids Res., B,
1639(1978).

105. A. Pullman and B. Pullman, Adv. Quant. Chem., B, 267
(1968).

106. D. H. Turner, R. Yuan, G. W. Flynn, and N. Sutin,
Biophys. Chem., B, 385(1974).