ELECTROCHEMICAL AND PHOTOCHEMICAL OXIDATION OF
TERESA :E. PETERSON
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
The number of people involved directly or indirectly with this
research project are too numerous to mention in this short space.
Everyone in the chemistry department (i.e., faculty, staff and
graduate students) has provided encouragement, support and technical
advice. Thanks are extended to all of them.
Many thanks go to Dr. Anna Brajter-Toth and the members of the
group for their guidance and friendship. I will miss all of them and
I wish them good luck in their future endeavors.
Heartfelt thanks go to my family for their continued
encouragement and especially to my father who never took life too
Most of all I would like to thank my husband, Mark, for showing
me that there is life after graduate school.
TABLE OF CONTENTS
LIST OF TABLES..................................................... vi
LIST OF FIGURES...................................................vii
1 INTRODUCTION ............................................. 1
2 ELECTROCHEMICAL STUDIES OF BIOLOGICAL OXIDATIONS.........9
3 INTRODUCTION TO METHODS USED IN THE STUDY OF
BIOLOGICAL OXIDATIONS................................... 18
3.1 Cyclic Voltammetry.................................18
3.1.1 Adsorption, Diffusion and Reversibility.....18
3.1.2 Homogeneous Chemical Reactions..............20
3.1.3 Determination of pKa........................22
3.1.4 Estimation of n.............................23
3.2 Determination of n-Values by Coulometry............ 24
3.3 Formation of Oxidation Products by Constant
3.4.1 Estimation of n.............................25
3.4.2 Determination of Electrode Area.............27
4.1 Cyclic Voltammetry................................. 28
4.2 Coulometry and Constant Potential Electrolysis.....31
4.3 Product Analysis...................................33
4.5 Photooxidation Reactions...........................38
4.6 Enzymatic Oxidation Studies........................39
4.7 Electrochemical Thermospray Mass Spectrometry
5 CYCLIC VOLTAMMETRY OF TUBERCIDIN-5'-MONOPHOSPHATE.......44
5.1 Oxidation and Reduction Behavior, Reversibility....44
5.2 Effect of pH.......................................47
5.3 Electrode Surface Effects..........................52
5.4 Behavior in Organic Solvents and Solvent
6 DETERMINATION OF n-VALUE OF TUBERCIDIN-5'-
6.1 Coulometry......................................... 64
6.2 Peak Current Comparisons...........................65
6.3 Chronocoulometry................................... 69
7 CONSTANT POTENTIAL ELECTROLYSIS OF TUBERCIDIN-5'-
7.2 Cyclic Voltammetry.................................80
7.3 UV Spectra......................................... 81
7.4 HPLC............................................... 83
8 ANALYSIS OF TUBERCIDIN-5'-MONOPHOSPHATE
ELECTROCHEMICAL OXIDATION PRODUCTS...................... 85
8.1 Separation by Gel Permeation Liquid
8.2 Analysis of Products with High Pressure Liquid
Chromatography, Cyclic Voltammetry and
8.3 Analysis of Products with Gas Chromatography
Mass Spectroscopy, Fast Atom Bombardment Mass
Spectrometry, Fourier Transform Infra-Red
Spectroscopy and Nuclear Magnetic Resonance
8.3.1 Product 1...................................95
8.3.2 Product 2...................................96
8.3.3 Product 3..............................104
8.4 Summary of Results................................ 105
9 PHOTOOXIDATION OF TUBERCIDIN-5'-MONOPHOSPHATE.......... 110
9.1 Oxidation with Ultraviolet Light..................110
9.2 Mechanistic Studies...............................113
10 ENZYMATIC OXIDATION OF TUBERCIDIN-5'-MONOPHOSPHATE.....121
11 ELECTROCHEMISTRY THERMOSPRAY MASS SPECTROSCOPY
(EC/TSP/MS) OF TUBERCIDIN-5'-MONOPHOSPHATE............. 128
12 SUMMARY AND FUTURE WORK ................................ 133
APPENDIX. .......................................................... 139
BIOGRAPHICAL SKETCH ............................................... 153
LIST OF TABLES
6-1 Estimation of the n-Values of TMP from Cyclic
Voltammetric Peak Currents................................ 68
6-2 Validity of Peak Current Comparisons for Estimation
of n-Values............................................... 68
6-3 Estimation of the n-Value of TMP Using Chronocoulometry...70
6-4 Validity of Chronocoulometric Comparison for
Estimation of n-Values....................................71
8-1 Analysis of Products Eluting Under GPLC Peaks A, B
and C by GC/MS and FABMS.................................. 92
8-2 Analysis of Products Eluting Under GPLC Peaks A, B
and C by FTIR and NMR.....................................93
11-1 Analysis of TMP and Its Oxidation Products by
EC/TSP/MS in 0.1M Ammonium Acetate Buffer pH 7...........130
LIST OF FIGURES
1-1 Structure of tubercidin and some adenosine analogs.........2
1-2 Structures of bases, nucleosides and nucleotides...........5
2-1 Proposed reaction scheme for the electrochemical
oxidation of uric acid at physiological pH................12
2-2 Proposed reaction scheme for the electrochemical
oxidation of adenine...................................... 14
3-1 Homogeneous chemical reactions that accompany
heterogeneous electron transfer processes.................21
4-1 Schematic of a saturated calomel electrode in an
electrochemical cell with analyte solution................29
5-1 Cyclic voltammogram of TMP at a rough PG electrode,
600UM solution in pH 7 phosphate buffer, y=0.5M...........45
5-2 Effect of concentration and scan rate on the oxidation
peak current of TMP in a pH 7 phosphate buffer,
p=0.5M: a) plot of peak current versus concentration
for the oxidation peak of TMP, scan rate 200mV/s;
b) plot of log peak current versus log scan rate for a
100M TMP solution (c.c. = correlation coefficient)....... 46
5-3 Plot of log peak current versus log scan rate for a
998UM TMP solution in pH 7 phosphate buffer, u=0.5M
(c.c. = correlation coefficient)..........................48
5-4 Cyclic voltammetric behavior of TMP at a PG electrode
in different buffers, i=0.5M: a) pH 2.8 phosphate;
b) pH 7 phosphate; c) pH 9.5 phosphate; d) pH 6.6
5-5 Plot of peak potential versus pH for the oxidation
peak of TMP in phosphate buffers, 4=0.5M. All
concentrations were ca. 300pM in TMP. Scan rate
5mV/s; c.c. = correlation coefficient.....................51
5-6 Spectral behavior of TMP (ca. 150pM) in phosphate
buffers, y=0.5M, of pH a) 2.8-4.6 and b) 6.8-11...........53
5-7 Cyclic voltammograms in a a) pH 7 phosphate buffer,
4=0.5M, at rough PG; b) 600UM TMP solution in pH 7
phosphate buffer at rough PG; c) same as a, at GC;
d) same as b, at GC....................................... 56
5-8 Cyclic voltammograms of a 575UM TMP solution in pH 7
phosphate buffer, y=0.5M, at a glassy carbon electrode:
a) before pretreatment and b) after pretreatment. Scan
rate 200mV/s.............................................. 59
5-9 Cyclic voltammograms in pH 7 phosphate buffer, p=0.5M,
at a GC electrode: a) before pretreatment and b) after
5-10 Plot of TMP oxidation peak current at PG and GC
versus % DMF added to a 600pM TMP solution in pH 7
phosphate buffer, P=0.5M.................................. 62
7-1 HPLC during the electrolysis of a 600pM TMP solution
in pH 7 phosphate buffer, y=0.5M, at a rough PG
electrode: a) before; b) 25min; c) 1hr; d) 2hr,
e) 3hr, 22min; f) 7hr.....................................74
7-2 UV spectra during the electrolysis of a 100jM TMP
solution in pH 7 phosphate buffer, 4=0.5M, at a
rough PG electrode........................................ 76
7-3 Cyclic voltammograms of a 6000M TMP solution in a
pH 7 phsophate buffer, y=0.5M, during electrolysis
at a rough PG electrode: a) before; b) 2hr, 2min;
c) 7hr, 2min ..............................................77
7-4 GPLC separation of TMP electrolysis products from a
600pM TMP solution in a pH 9.5, u=0.5M, phosphate
buffer after a) 4-12hr and b) 60hr of electrolysis........78
7-5 UV spectra during the electrolysis of a 100UM TMP
solution in pH 7 phosphate buffer, P=0.5M, at an
unroughened PG surface.................................... 82
8-1 HPLC of compounds under GPLC peaks a) A, b) B and c) C
(flow rate 1ml/min).......................................87
8-2 UV spectrum of separated product fractions under GPLC
peak A.................................................... 88
8-3 UV spectrum of separated product fractions under GPLC
peak B .................................................... 89
8-4 UV spectrum of separated product fractions under GPLC
peak C ....................................................91
8-5 Proposed structures for products 1, 2 and 3...............94
8-6 Standard NMR signals for a) inosine, b) guanosine and
c) adenosine-5'-dihydrogen phosphate......................98
8-7 Key to peak assignments for figures 8-6 and 8-8 to
8-10 ...................................................... 99
8-8 Standard NMR signals for a) 3-aminopyridine,
b) 2-aminopyrimidine and c) 2-aminopyridine.............. 100
8-9 Standard NMR signals for a) 2-pyridyl acetamide,
b) histamine and c) inosine triacetate...................101
8-10 Standard NMR signals for a) p-dimethylamino
benzaldehyde, b) 2-ethyl pyrrole and c) N,N-
8-11 Proposed structure of the second product eluting
under GPLC peak B........................................ 103
8-12 Structure of 5-hydroxyhydantoin 5-carboxamide............108
8-13 Products from the gamma irradiation of adenosine
(80) and the photosensitized oxidation of 3 methyl
indole (81) .............................................. 109
9-1 UV spectra during photooxidation of a 1004M TMP
solution in a pH 7 phosphate buffer, p=0.5M, with a
deuterium lamp ........................................... 111
9-2 HPLC during photooxidation with a deuterium lamp of
a 100uM TMP solution in p=0.5M, pH 7 phosphate
buffer: a) before; b) 2hr, 37min; c) 10hr, 15min........ 112
9-3 Spectral changes of a 67VM TMP solution in pH 7
phosphate buffer, y=0.5M, after 4hr exposure to a
deuterium lamp........................................... 114
9-4 Effect of NaN3 on the photooxidation of TMP with a
deuterium lamp, 67pM TMP/0.1M NaN3 in pH 7 phosphate
9-5 Effect of DABCO on the photooxidation of TMP with a
deuterium lamp, 80M TMP/1.1mM DABCO in pH 7 phosphate
buffer ................................................... 117
10-1 Cyclic voltammograms during the reaction of TMP with
liver microsomes: a) NADPH and MgCl2 in the absence
of microsomes; b) same as a, plus TMP (450pM); c) same
as b, plus 300ul microsomal pellet, 10min incubation;
d) same as c, 150min incubation..........................126
11-1 Proposed structures for the products from EC/TSP/MS
of TMP ................................................... 132
A-1 FTIR of separated product fractions under GPLC peak A....139
A-2 FTIR of separated product fractions under GPLC peak B....140
A-3 NMR of separated product fractions under GPLC peak B.....141
A-4 FTIR of separated product fractions under GPLC peak C....142
A-5 NMR of separated product fractions under GPLC peak C.....143
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
ELECTROCHEMICAL AND PHOTOCHEMICAL OXIDATION OF
TERESA E. PETERSON
Chairperson: Anna Brajter-Toth
Major Department: Chemistry
The purpose of this investigation was to use instrumental
methods to provide insight into the biological degradation pathways
of the purine drug tubercidin-5'-monophosphate (TMP). Tubercidin is
a highly cytotoxic antiviral and chemotherapeutic agent. In this
study the electrochemical and photochemical oxidation of TMP was
The electrochemical oxidation of TMP was monitored by cyclic
voltammetry, ultraviolet spectroscopy and high pressure liquid
chromatography. Oxidation products formed by constant potential
electrolysis were separated by gel permeation liquid chromato-
graphy. The products were analyzed by mass spectrometry, Fourier
transform infra-red spectroscopy, fast atom bombardment mass
spectrometry and nuclear magnetic resonance spectroscopy.
Three oxidation products were isolated and identified. One
product is a hydantoin-type structure resembling a product of the
electrochemical oxidation of uric acid, 2,6-diaminopurine and
hypoxanthine. The other two products resemble products obtained from
the photosensitized oxidation of indoles and the irradiation of
nucleic acid components with ionizing radiation.
To determine if the electrochemical oxidation of TMP provides
any insight into its biological reactivity, the enzymatic oxidation
of TMP was investigated with xanthine oxidase, chloroperoxidase and
cytochrome P450. The results indicate that TMP may be a substrate of
The oxidation of TMP by ultraviolet light indicates that water
and oxygen are necessary for the photooxidation to occur and that
singlet oxygen is produced during photooxidation. Similarities
between the electrochemical and photochemical oxidation are
The cyclic voltammetric behavior of TMP was studied in different
environments (i.e., different pH, electrode surfaces and solvents).
The results indicate that the environment (i.e., solvent and surface)
plays a major role in the electrochemical behavior of TMP and
possibly other biological compounds.
A preliminary study of TMP oxidation was also carried out using
electrochemistry thermospray mass spectroscopy. Molecular ions
corresponding to the products generated by bulk constant potential
electrolysis were not detected by this method.
Studies of the electrochemical oxidation of purines and purine
drugs can provide a great deal of information about the biological
oxidation of these compounds (1-9). In this study we have investi-
gated the electrochemical oxidation of tubercidin-5'-monophosphate
(TMP). The electrochemical oxidation was monitored with ultraviolet
(UV) spectroscopy and high pressure liquid chromatography (HPLC)
which provided additional information about the formation of inter-
mediates and products. On a preparative scale the products were
separated by gel permeation liquid chromatography (GPLC) and were
analyzed by gas chromatography mass spectrometry (GC/MS), fast atom
bombardment mass spectrometry (FABMS), Fourier transform infra-red
spectroscopy (FTIR) and nuclear magnetic resonance spectroscopy
(NMR). Additional studies of the photooxidation of TMP indicate that
most of the electrochemical and photochemical oxidation products are
Tubercidin is a naturally occurring, highly cytotoxic antiviral
and chemotherapeutic agent. It is structurally similar to adenosine
and belongs to a class of compounds known as 7-deazaadenosines, or
pyrrolopyrimidines (Fig. 1-1). Two other compounds in this class are
sangivamycin and toyocamycin (Fig. 1-1). Although their reactivity
is somewhat different these compounds are all cytotoxic to mammalian
Figure 1-1 Structure of tubercidin and some adenosine analogs.
cell lines in culture and inhibitory to the growth of bacteria,
fungi, ribonucleic acid (RNA) and deoxyribonucleic acid (DNA) viruses
(10). They are invaluable biochemical tools for studying cellular
and enzyme reactions because they can replace adenosine, adenosine
monophosphate (AMP), adenosine diphosphate (ADP) and adenosine
triphosphate (ATP) in cellular reactions (10). They are also useful
in determining the structural requirements for interaction with
Although much is known about the physical effects of these
drugs, little is known about their biological activation or degrada-
tion in the body. Metabolism studies provide a better understanding
of a drug's mode of action, toxicity and biochemical reaction-
mechanism so its safety and effectiveness can be evaluated. Modern
drugs are complex and can be metabolized by several reactions,
producing many metabolites which can be toxic, active or inert.
Drugs can be metabolized by a wide variety of enzymes in the
body. The enzymatic reactions responsible for metabolism are
classified as phase I and phase II reactions (11). Phase I reactions
usually result in oxidation, reduction or hydrolysis. In phase II
reactions the products of phase I reactions are conjugated to natural
compounds such as glucuronic acid. These reactions may transform the
drug into a more polar compound that can be rapidly removed from the
body or transform the drug into a toxic or therapeutically active
metabolite(s). Whether the metabolic reaction produces a toxic,
inert or therapeutically active metabolite is a function of the
drug's structural features (12).
Compounds which result from a slight modification in the
structure of naturally occurring purines or pyrimidines are known as
antimetabolites. Some of these antimetabolites exhibit antiviral and
antineoplastic activity. The most successful modifications usually
involve the base component rather than the sugar component (12). The
following types of modifications have been tried: replacement of
OH-groups with SH or NH2-groups, replacement of a ring C-atom with a
N-atom or a ring N-atom with a C-atom and introduction of halogen
groups at C-2 or C-6 (12). The numbering system for purines is given
in Figure 1-2. The majority of effective drugs are simply structural
modifications of purines and pyrimidines.
Purines and pyrimidines are present in cells as polymerized
nucleotides (DNA, RNA), free nucleotides (ATP, AMP), nucleosides
(adenosine) and bases (adenine) (Fig. 1-2). The numbering system for
purines is illustrated in Figure 1-2. Purines and pyrimidines serve
as precursors for nucleic acid synthesis; are involved in energy
metabolism, group transfer reactions, mediation of hormone action;
and act as metabolic regulators (13).
Antimetabolites can usually interact with the same enzymes as
the parent purine or pyrimidine. If this occurs, the enzyme may
become blocked and will not be able to fulfill its normal biological
functions or the antimetabolite may be converted to another compound
that disturbs the cell's metabolism.
Unlike other groups of therapeutic agents, nucleoside analogs
have the entire cell metabolism as a target. In contrast a folate
Structures of bases, nucleosides and nucleotides.
analog can only inhibit dehydrofolate reductase or closely allied
Essentially, all of the antimetabolites which act as
chemotherapeutic agents interfere with the synthesis and metabolism
of nucleic acids (12). Inhibition of nucleic acid synthesis
primarily affects those cells which have a high rate of mitosis.
This includes malignant cancer cells as well as other rapidly
dividing cells (e.g., bone marrow, hair and various epithelia
tissues) (12). Fortunately, cancer cells are more susceptible than
normal, rapidly dividing cells, to the actions of the antimeta-
bolites. The exact reasons are not fully understood but they stem
from physiological differences between cancer and normal cells (15).
Antimetabolite drugs are usually administered as nucleosides.
In the process of cell penetration they are converted to nucleotides
by adenosine kinase. An active compound will usually interfere with
nucleic acid synthesis. To interfere with nucleic acid synthesis the
compound must be phosphorylated, since the de novo synthesis of
purines proceeds entirely via phosphorylated intermediates. If the
compound is administered as a nucleotide, dephosphorylation usually
occurs during cell penetration, resulting in an inactive metabolite
Tubercidin was isolated from streptomyces tubercidus by Anzai,
Nakamura, and Suzuki (16). Subsequent studies by Suzuki and Marumo
determined its structure (17,18). This naturally occurring compound
enters the same anabolic pathways as adenosine, yet it is not
degraded by the enzymes that degrade adenosine. It has been shown to
inhibit mammalian and bacterial cell growth, RNA and DNA viruses, and
glycolysis (19). Glycolysis is one of the pathways by which cells
extract energy in the form of ATP by metabolizing glucose to
pyruvate. In addition to this, tubercidin interferes with
mitochondrial respiration, de novo purine synthesis, protein, RNA and
DNA synthesis, and transfer RNA processing (19).
Tubercidin showed promise as a drug for several reasons. First,
it was converted to the active nucleotide form during cell
penetration and, secondly, it was stable to deamination by adenosine
deaminase and to glycosidic bond cleavage by purine nucleoside
phosphorylases. These are the two major pathways which inactivate
purine drugs (19,20).
The therapeutic effects of tubercidin result from its wide range
of biological activity. Tubercidin was tested against a large number
of potential targets. It was first evaluated as an antitubercular
and antitumor agent. In both cases the drug showed no selectivity
and pronounced toxicity to the host (14). This toxicity manifested
itself in the form of nephrotoxicity and venous thrombosis at the
sites of injection (20). A nephrotoxin is a substance which is toxic
to kidney cells. Thrombosis refers to blood clotting. Toxicity is
not encountered when tubercidin is suspended in a petrolatum base and
applied to the skin. In this application, tubercidin is effective
against basal cell skin carcinomas (14).
Recent reports in the literature have shown that tubercidin,
administered in combination with nitrobenzylthioinosine (a potent
purine transport inhibitor) protects the host against some of the
cytotoxic effects of tubercidin (21,22). Earlier studies showed that
when tubercidin is added to animal or human blood "in vitro" it
enters erythrocytes and is maintained there in nucleotide form. When
the blood is transfused back into the subject nephrotoxicity and
venous thrombosis are avoided (20). The discovery that tubercidin
enters red blood cells led to an effective treatment for
schistosomiasis, a parasitic blood disease (14). A later report
shows that tubercidin, along with other deaza analogues of adenosine,
inhibit blood platelet aggregation (i.e., blood clotting) (23).
A study of the effects of various inhibitors of purine
metabolism showed that tubercidin may be useful against
trichomoniasis in humans (a parasitic infection). This study was
done using two parasites that synthesize nucleotides differently
(24). To understand which cellular reactions contribute to the
cytotoxic effects of tubercidin and toyocamycin genetic and
biochemical studies are being conducted on mutant cells resistant to
toyocamycin and tubercidin (25).
A great deal is known about the effects of tubercidin on
cellular processes such as mitochondrial respiration, synthesis of
purines, RNA and DNA and on biochemical targets, cancer cells,
viruses and parasites (10). However, it is not known which of these
effects is primarily responsible for the inhibition of cell growth
and cytotoxicity (20). Is the compound metabolized to another
compounds) which exerts the vast number of biological effects? Are
these metabolites toxic or inert? These are some of the questions
for which answers are being sought.
ELECTROCHEMICAL STUDIES OF BIOLOGICAL OXIDATIONS
Carbon electrodes are commonly used in electrochemical studies
of biological oxidations (1-9). These electrodes provide a wide
potential range (+1.2 to -1.7V versus SCE) and a low background
current in aqueous solutions (26,27,28). The mercury electrode
cannot be used at potentials more positive than about +0.4V versus
SCE (27). Platinum electrodes are troublesome because oxygen and
hydrogen adsorb onto the surface and high backgrounds are obtained
(29). The formation of an oxide layer on platinum inhibits many
oxidation processes including those of organic substances. The
potential range for platinum in aqueous solution is between +0.9 and
-0.6V versus SCE.
There are a number of carbon based electrodes in use. Carbon
paste electrodes were introduced by Adams and co-workers (27).
Glassy carbon (GC) electrodes were applied for the first time in
electroanalytical chemistry by Zittel and Miller (cited in 30). They
provide a smooth mirror-like surface with very low residual
currents. The pyrolytic graphite (PG) electrode has a much rougher
surface and usually provides for faster kinetics. It was introduced
by Beilby and co-workers and Miller and Zittel (cited in 27 and 30).
The electrochemical oxidation of a large number of biological
purines has been reviewed (1,2). The most extensively studied purine
is uric acid. Its electrochemical oxidation has been studied using
lead oxide, spectroscopic graphite and PG electrodes (1,2).
The electrochemical oxidations of a large number of purines
besides uric acid have been studied. These include adenine (3),
guanine (7), 6-thiopurine, 2-thiopurine and 2,6 dithiopurine (1).
More recently, there have been studies of 2,6 diaminopurine (4), 9-B-
D ribofuranosyluric acid (5), xanthine (6), 8-oxyguanine (7) and
In the majority of these studies, both the electrochemical and
enzymatic oxidations were carried out. For the most part, the
electrochemical and enzymatic oxidations produced similar products.
However, more mechanistic information could be obtained from the
The enzymatic oxidation of purines by xanthine oxidase has been
studied by Bergman and Krenitsky and co-workers (31,32). These
studies suggest that oxidation occurs initially at the C-6 position
then at the C-2 or C-8 positions. The numbering system for purines
was given in Figure 1-2. If oxidation occurs at C-2, the C-8
position will also be oxidized. The reverse is also true.
Dryhurst and co-workers made a number of generalizations about
the electrochemical oxidation of purines (1).
1. Purine itself is not electrochemically oxidizable. To be
oxidized, a purine must contain an amino or oxy group in the
pyrimidine ring of the purine molecule.
2. Susceptibility to oxidation increases as the purine ring is
3. Oxidation appears to begin initially at any unoxidized or
unsubstituted N=C bonds. The final site of electrochemical
attack is the C4=C5 bond.
4. An unstable diimine type structure appears to be the initial
oxidation product. This can occasionally be detected by fast
sweep voltammetry at a PG electrode.
5. If the purine is substituted at both N-3 and N-7 or N-9 by
methylation, a diiminium ion forms which is more reactive
than the diimine. This is usually not detected by cyclic
6. Hydrolysis of the diimine or diminium ion leads to a 4,5-
diol, which undergoes a number of secondary reactions to give
7. A large number of purines are adsorbed at PG.
These generalizations are supported by the electrochemical studies of
uric acid (33,34).
The primary reaction in the electrochemical oxidation of uric
acid involves a quasi-reversible 2e-2H+ process, leading to an
anionic diimine intermediate, II (Fig. 2-1) (33,34). This diimine,
II, undergoes a rapid (pseudo) first order hydration reaction, giving
an anionic imine-alcohol, III. This intermediate may be detected by
UV spectroscopy during thin layer electrochemical oxidation. It is
also the species responsible for a reduction peak at -0.9V versus SCE
in the cyclic voltammogram of uric acid. A (pseudo) first order
hydration of this imine alcohol, III, gives a diol, IV, which
H2N O > + CO2
H H H
Proposed reaction scheme for the electrochemical
oxidation of uric acid at physiological pH. Modified
from reference (34).
decomposes via an isocyanate to allantoin, VII. Cyclic voltammetry,
thin layer spectroelectrochemistry and GC/MS provided information to
support this mechanism (33,34).
The enzyme (peroxidase) oxidation of uric acid yields
intermediates that are identical, spectrally and electrochemically,
to the UV-absorbing intermediates generated upon electrochemical
oxidation (33,34). The ultimate reaction product formed by
electrochemical and enzymic oxidation is allantoin. These results
support the conclusion that the electrochemical and enzymic reactions
are, in a chemical sense, the same (34).
Since TMP is structurally similar to the purine nucleoside,
adenosine (Fig. 1-1), a similar type of electrochemical behavior may
be expected for both compounds. There are no reports on the
electrochemical oxidation of adenosine but there is a report on the
electrochemical oxidation of the parent purine adenine (3). Since
the electrochemical behavior of uric acid and uric acid riboside is
very similar (i.e., the same types of products are formed in both
oxidations) and the ribose group is not electroactive, the electro-
chemical behavior of adenosine may be similar to that of adenine (5).
The primary electrochemical oxidation of adenine is shown in
Figure 2-2. It was shown that six electrons were involved in the
oxidation (3). The number of electrons was determined using
coulometry and the majority of products were analyzed by paper and
thin layer chromatography. The proposed mechanism indicates a series
of 2e-2H+ oxidations to 2-hydroxyadenine, II, then 2,8-dihydroxy
adenine, III. Finally, oxidation of the 4,5 double bond leads to a
+ H20 ON 5 + 2H + 2e
H20 ---> O N C=0 2H* 2e .___
:=0 + 2e
O N JO
* H* + CO2 + C=0 + NH3 + e
O N NH2
+ NH3 + CO2
0 N N2
+ H' e H20
ON C=0 +" NH3 CO2. H*
Proposed reaction scheme for the electrochemical
oxidation of adenine. Modified from reference (3).
dicarbonium ion, IV, similar to that formed in the electrochemical
oxidation of uric acid (3).
Following the primary electrochemical oxidation, three distinct
chemical and electrolytic reactions occur (Fig. 2-2). Further
electrochemical oxidation of the dicarbonium ion, IV, leads to
parabanic, V, and oxaluric acids, VI, urea and ammonia. Some of IV
is also reduced to 4-amino-2,5,6 trihydroxy-, VII, and 5 amino-2,4,6-
trihydroxy pyrimidine, VIII, which condenses to form 4-aminopurpuric
acid, IX. Allantoin, X, forms from acid hydrolysis of IV, liberating
NH3 and CO2 (3).
Based on what is known about the electrochemical and enzymatic
oxidation of purines, we can hypothesize what might occur during the
electrochemical oxidation of TMP. Since the C-6 position contains an
NH2 group, oxidation will occur at the C-2 or C-8 positions or
both. The C4=C5 double bond may also be oxidized. It is already
known that TMP is resistant to deamination and glycosidic bond
cleavage (19,20). The oxidation products identified confirm this.
The purpose of this investigation was to provide insight into
the biological degradation of the purine drug tubercidin. The
nucleotide form of this drug (i.e., TMP) was investigated because it
is more soluble in water and because tubercidin is probably in the
nucleotide form inside the cell (12). Although much is known about
its biological activity, little is known about the metabolites which
may be the source of drug toxicity. For the most part, purines are
biologically degraded via oxidation reactions and purine drugs are
often activated in the process (1, p. 127). In this study electro-
chemistry was used as a tool for evaluating the oxidative degradation
A number of reports in the literature, including the above
description of the oxidation of uric acid, illustrate that the
electrochemical and enzymatic oxidation of purines can be chemically
the same (1,2,4,5,6). The experimental methods used in these types
of studies, which were described earlier, were also used in this
study. In addition, the electrolysis of TMP was also monitored by
HPLC and products were further analyzed by FTIR and NMR.
To determine if the results from electrochemical methods provide
any insight into the biological redox reaction of TMP the enzymatic
oxidation of TMP was also investigated. Tubercidin-5'-monophosphate
was not significantly oxidized by any of the enzymes that were
investigated although experimental results indicated that TMP may be
a substrate of cytochrome P450"
Studies of TMP oxidation as a function of pH, at different
electrode surfaces and in polar and nonpolar solvents were carried
out to provide more information on the reactivity of TMP. These
studies indicate that the environment (i.e., solvent and surface)
plays a major role in the oxidation of TMP.
The photooxidation of TMP, which has not been previously
reported, was also investigated. Similarities between the
electrochemical and photochemical oxidation are discussed.
Oxidation of TMP was also studied using electrochemistry
thermospray mass spectrometry (EC/TSP/MS). This technique allows
oxidation products to be analyzed on-line by mass spectrometry
without the need for separations. Preliminary results do not confirm
the products identified from bulk constant potential electrolysis.
INTRODUCTION TO METHODS USED IN THE STUDY
OF BIOLOGICAL OXIDATIONS
3.1 Cyclic Voltammetry
3.1.1 Absorption, Diffusion and Reversibility
Cyclic voltammetry is used quite often in the mechanistic
studies of redox systems (26, p. 86). The experimental set up
usually consists of a potentiostat, waveform generator and a three-
electrode cell containing counter, working and reference
electrodes. The current is measured at the working electrode as the
potential is scanned at different rates (26, p. 86). The resulting
electrochemical process can be adsorption or diffusion controlled and
the electron transfer may be reversible or irreversible. The peak
current equations for the different process are given below (29, pp.
218, 222, 522, 525).
Diffusion controlled systems:
reversible i = (2.69x105)n3/2 A D/2 v2 C (3-1)
5 1/2 1/2 1/2
irreversible i = (2.99x10 )n(an ) A C D v (3-2)
p a o o
Adsorption controlled systems:
reversible i vA (3-3)
p 4RT o
nan F 2Av *
irreversible i = a (3-4)
i = peak current, A
n = number of electrons transferred per molecule or ion,
a = transfer coefficient, measure of the symmetry of the
na = number of electrons involved in the rate determining
step, eq mol1
A = electrode area, cm2
D = diffusion coefficient, cm2sece
Co* = bulk concentration, mol cm-3
v = scan rate, V/sec-1
F = surface concentration of species before redox reaction,
T = temperature, K
R = gas constant, J mol1K-1
F = Faraday, 96,496 coul mole-1 of electrons
For diffusion controlled processes, ip is proportional to C *
and v/2 For adsorption controlled processes ip is proportional to
v. A broad peak usually indicates a diffusion process while a sharp
symmetrical peak indicates an adsorption controlled process (35).
In a diffusion controlled process reactant moves from one region
to another as the result of a concentration gradient. As the
concentration of an electroactive species, X, is depleted at the
electrode a concentration gradient is set up and more X diffuses from
the bulk solution to the electrode (26, pp. 9-49).
Adsorption of a species onto the electrode surface can occur as
a result of electrostatic attraction or covalent bonding between the
species and the electrode. Adsorption can also result from
hydrophobic interactions between a species in solution and the
electrode. For example, neutral organic molecules in aqueous
solutions may adsorb onto the electrode surface as a result of
hydrophobic interactions (26, pp. 43-48).
For reversible and irreversible adsorption and reversible and
irreversible diffusion controlled systems the equations for ip are
similar. One of the diagnostic criteria for a reversible diffusion
controlled system is a .059/n V separation between oxidation and
reduction peaks (26, pp. 90-92). This is observed when the voltage
scan rate is small compared to the electrode reaction rate con-
stant. If the reaction is quasi-reversible, the peak potentials will
be separated by more than .059/n volts (26, pp. 90-92). Typically, a
system is irreversible if only one peak, either cathodic or anodic,
is present in the cyclic voltammogram or if cathodic and anodic peaks
are present and they are separated by more than 0.10 volts.
3.1.2 Homogeneous Chemical Reactions
In many cases the heterogeneous electron transfer reaction at
the electrode is affected by homogeneous chemical reactions.
Consider the reaction 0 + ne -- R, where 0 is the oxidized form of
the standard system and R is the reduced form of the standard
system. Figure 3-1 summarizes some of the reactions that could
occur. The oxidized form of the standard system could be produced
from a preceding chemical reaction. In a following reaction R
b) Following Reaction (EC)
0+ ne -R
c) Catalytic Reaction
R + Z->0+ Y
Figure 3-1 Homogeneous chemical reactions that accompany
heterogeneous electron transfer processes.
could chemically react (e.g., with solvent) to form an electro-
inactive species X at potentials where 0 is reduced. If the product
of a following chemical reaction is electroactive where 0 is reduced,
a second electron transfer could take place. Additionally, a
catalytic reaction could occur where the product R undergoes a
chemical reaction with a nonelectroactive species Z in solution to
regenerate 0 (29, p. 431).
3.1.3 Determination of pKa
Cyclic voltammetry at slow scan rates allows application of the
Nernst equation to the determination of pKa's for compounds involved
in acid/base reactions. The Nernst equation is valid at slow scan
rates because there the system may be assumed to be at equilibrium.
mR <=> pP+ + rH+ + ne
0.0591 [P]P [H]r
Ep = E + log (3-5)
E = E 59(rpH -) (3-6)
p o n R]m
n = number of electrons transferred per molecule or ion,
Ep = measured peak potential, V
E = standard potential, V
The compound exists in different forms in the pH region above
and below the pKa. This should be reflected in the measured r/n
values for different acid/base forms (equation 3-6). Thus a plot of
Ep versus pH may be used to estimate the pKa. Since the slope of
this plot equals -.0591r/n the number of protons, r, and electrons,
n, involved in the oxidation or reduction of the electroactive
species can be estimated.
3.1.4 Estimation of n
The equations of ip for reversible and irreversible processes
(equations 3-1 through 3-4) show that ip is proportional to n.
Therefore, the n-value for a compound can be estimated by comparing
its ip to the ip of a compound with a known n-value. Equation 3-7
can be used to compare the i values of two electrochemically
irreversible diffusion controlled systems, where k is the system with
known n-value and u is the system with unknown n-value. The terms A,
v1/2, D 1/2 and Co* do not have to be included in the calculation if
these values are the same for both compounds. Since Do depends on
the size of the diffusing molecules in liquids, it will be similar
for structurally similar compounds under the same conditions (36,37).
i n (an )1/2
p uk \ak
n = (3-7)
u i 1/2
If the ip values for two irreversible processes of similar compounds
are compared, the ana values have to be estimated independently for
both compounds. The ana values can be estimated from equation 3-8 by
measuring Ep at two sweep rates v1 and v2 (27, pp. 135-137).
E E n (3-8)
p p2 ana- Fv
3.2 Determination of n-Values by Coulometry
Coulometry is the standard method used to determine n-values.
This experiment is usually carried out in a three-compartment, three-
electrode cell. The working electrode is maintained at a constant
potential where the reaction occurs at a maximum rate (i.e., ca.
100mV past the cyclic voltammetric peak). In coulometry the current
is integrated until the redox reaction is completed.
As the electrolysis proceeds the current decays exponentially as
a function of time (26, pp. 119-121). A coulometer integrates this
current over time to give the total charge (Q) passed during the
electrolysis. The total charge can be related to n through Faraday's
Law (equation 3-9).
Q = nFN (3-9)
Q = total charge passed, C
n = number of electrons transferred per molecule or ion, eq mol"1
F = 96,496 C eq-1
N = number of moles of substance being electrolyzed
Ideally the coulometer stops collecting charge when the electroactive
starting material is completely electrolyzed.
If the charge increases after the starting material has been
completely electrolyzed it is not possible to calculate an exact
value of n. Under such conditions products may be further oxidized
at the electrolysis potential or a catalytic oxidation or reduction
of water may be occurring (38). The presence of oxygen containing
functional groups such as carboxyl, hydroxyl, carbonyl, lactone and
quinones on carbon electrodes has been suggested (39). These groups
could participate in catalyzing the oxidation or reduction of water
3.3 Formation of Oxidation Products by Constant
Constant potential electrolysis can be used to generate products
of the electrode reaction which can be isolated and further
analyzed. The experimental set up is the same as that for
coulometry. When the electrode is held at a constant potential the
intermediates and products from the electrode reaction are generated
in solution. By periodically stopping the electrolysis, the solution
can be analyzed for these intermediates and products by a number of
methods. Cyclic voltammetry and UV spectroscopy are most commonly
In this investigation an HPLC method was developed and used to
monitor the electrolysis to show when TMP was completely
electrolyzed. Cyclic voltammetry and UV spectroscopy were not useful
in this respect because the intermediates and products have cyclic
voltammetric peaks and UV absorbance bands in the same region as TMP.
3.4.1 Estimation of n
The n-value can also be estimated using chronocoulometry. In a
chronocoulometric experiment the current associated with an
electrochemical process is integrated over time. The time scale of
the experiment is typically 50-500ms. The behavior can be analyzed
by plotting charge versus the square root of time (29, p. 200). This
method was popularized by Anson and co-workers and is sometimes
referred to as an Anson plot (41,42). The experiment is carried out
in an unstirred solution at a planar electrode held at an initial
potential (Ei) where no significant electrolysis takes place. The
potential is then stepped to Ef, a potential where electrolysis
occurs at a maximum diffusion controlled rate. This current is
described by the Cottrell equation (equation 3-10). The integrated
form of this equation gives the cumulative charge passed, Qd
(equation 3-11). The terms have their usual meaning.
i(t) = id(t) = 1/2 0 (3-10)
2nFAD 1/2C *t1/2
Qd = 1/2 (3-11)
The cumulative charge, however, is due only to the diffusing
material. Additional charge arises from double layer charging Qd1
and from the oxidation or reduction of adsorbed reactant
molecules (nFAr ). The total charge measured is
2nFAD 1/2C *t1/2
Q 0 1/2 + Qdl + nFAro (3-12)
The slope of a Q vs t1/2 plot is
Qd 2nFAD C 0
slope = /-- 1/2 (3-13)
Since the slope is proportional to n, the n-value of a compound can
be estimated by comparing its slope to that of a compound whose
n-value is known (equation 3-14). In equation 3-14 k is the known
and u is the unknown system.
slopes (n C o*)k
slope (n C *) (3-14)
u o u
In order to use this equation A must be the same and Do must be
known. If structurally similar compounds are chosen, Do is expected
to be the same for both compounds (36,37).
3.4.2 Determination of Electrode Area
From a chronocoulometric experiment using a standard solution of
K 3Fe(CN)3 (Do 7.6x10-6cm 2sec-1) electrode area can be determined
(equation 3-15). The terms have their usual meaning. This equation
A slope 2 (3-15)
was derived from equation 3-13.
4.1 Cyclic Voltammetry
All cyclic voltammetric experiments were carried out with an IBM
EC225 2A Voltammetric Analyzer (IBM Instruments Inc., Danbury,
Connecticut) with a Houston Instruments 2000 X-Y recorder (Bausch and
Lomb, Austin, Texas) or a model 173 Potentiostat and model 175
Universal Programmer from EG&G Princeton Applied Research (Princeton,
New Jersey). A 10 ml glass beaker contained the analyte solution and
working, counter, and reference electrodes. The PG working electrode
was prepared by sealing a piece of PG (1.5mm x 10mm) (Pfizer) into
glass tubing with inert epoxy (1C white, Dexter Corporation, Hysol
Division). In some experiments a GC working electrode was used.
These were prepared by sealing a rod (3mm diam x 10mm length) into a
glass tube with epoxy. Vitreous (glassy) carbon was purchased from
Electrosynthesis Company, Inc. (Amherst, New York). The platinum
wire counter electrode was encased in glass. Electrical contact for
these electrodes was made with a copper wire and mercury.
The saturated calomel (SCE) reference electrode was prepared
according to a standard procedure (43). The SCE was connected to the
analyte solution by a salt bridge prepared by filling one side of a
U-shaped glass tube with KC1 agar and the other side with analyte
solution (Fig. 4-1). The KC1 agar is prepared by warming 4g of agar
Schematic of a saturated calomel electrode in an
electrochemical cell with analyte solution.
and 90ml of water on a steam bath, then adding 30g of KC1. When the
salt has dissolved the gel is pipetted into the glass connection.
In order to obtain a reproducible surface the working electrodes
were resurfaced before each measurement. For the PG electrode,
resurfacing was accomplished by grinding the electrode on 600 grit
SiC paper using a Buehler Ecomet 1 Polisher-Grinder (Evanstown,
Illinois). The polishing paper was a Carbimet Abrasive disc (Buehler
Ltd., Lake Bluff, Illinois). GC electrodes were resurfaced by
polishing on Buehler billiard cloth with Gammal gamma alumina (Fisher
Scientific). Once polished the GC electrode was rinsed and placed in
an ultrasound bath for 10min to remove any residual alumina. Both
glassy carbon and pyrolytic graphite electrodes were rinsed
thoroughly with deionized water and carefully wiped with a Kimwipe
tissue to remove excess water and graphite particles.
The supporting electrolytes used for cyclic voltammetric
experiments were phosphate and ammonium acetate buffers of varying
pH, ionic strength 0.5M, prepared with doubly distilled deionized
water. All solutions were deaerated 5min with N2 before
measurement. A blanket of N2 was passed over the solution throughout
the experiment to keep the solution free of dissolved 02.
For the cyclic voltammetric experiments done in N,N-
dimethylformamide (DMF) and dimethylsulfoxide (DMSO) (Fisher
Scientific) the supporting electrolyte was 0.5M tetrabutylammonium
perchlorate (TBAP) (Kodak, Rochester, New York). When a cyclic
voltammogram was run in organic solvent the SCE was connected to the
analyte solution by a vicor tip and not by the salt bridge described
earlier. The vicor tip prevented the organic solvent from mixing
with the solution inside the SCE better than the salt bridge.
In some experiments the GC electrode was electrochemically
pretreated before use. Electrochemical pretreatment was carried out
by applying a constant positive potential of +1.8V for 8min to the
electrode in a solution of supporting electrolyte (44). This is
followed by 2-5 cyclic scans towards negative potentials first (+1.8
to -1.7V) in the supporting electrolyte at a sweep rate of 20mV/s.
These cyclic scans are carried out until the background peaks at ca.
-1 and 1V have disappeared.
Tubercidin-5'-monophosphate was obtained from Sigma Chemicals
(St. Louis, Missouri) and was used without purification.
4.2 Coulometry and Constant Potential Electrolysis
All coulometric and constant potential electrolysis experiments
were carried out with a model 173 potentiostat and model 179 digital
coulometer (EG&G and Princeton Applied Research). The current range
The experiments were carried out in a three-compartment, three-
electrode cell with compartments separated by cation exchange
membranes type P-1010 (RAI Research Corporation, Hauppauge, Long
Island, New York). The middle compartment contained the working
electrode and the electroactive species in a solution of supporting
electrolyte. The outer compartments contained a reference SCE and Pt
mesh counter electrode in equal volumes of supporting electrolyte.
Working electrodes were two rectangular plates of PG (6.3 x 1.8cm)
(Pfizer) which were cleaned by polishing with 600 grit SiC paper
followed by copious washing. All three-cell compartments contained
10ml of solution. Typically, 600iM TMP solutions in phosphate or
ammonium acetate buffers, ionic strength 0.5M, were electrolyzed.
Nitrogen (N2) was passed during the electrolysis through a
continuously stirred solution. Stirring was done because it
increased the magnitude of the current during electrolysis and
consequently decreased the time of an electrolysis by minimizing the
diffusion layer (26, p. 199).
To see the formation of intermediates and products the
electrolysis was monitored with cyclic voltammetry, UV spectroscopy
and HPLC. This was done by stopping the electrolysis every 10-
30min. Cyclic voltammograms were recorded using a PG electrode.
A Tracor Northern TN 6500 diode array spectrophotometer
(Middleton, Wisconsin) with a Hewlett Packard ink jet printer or a
Hewlett Packard 8450A diode array spectrophotometer with a Hewlett
Packard 7470A plotter were used for measuring UV spectra during
electrolysis. For UV spectra, the electrolysis solution was pipetted
into a 1cm quartz cuvette (4ml volume) and the absorbance spectrum
was recorded from 200 to 400nm. The reference quartz cuvette
contained the supporting electrolyte used for electrolysis (i.e.,
phosphate or ammonium acetate buffer).
For HPLC analysis an Altex 110A pump and solvent programmer were
coupled to a Kratos Spectroflow variable wavelength UV detector. A
Fisherbrand Resolvex C-18 (4.6mm x 25cm) column was used. The C-18
column was equilibrated with the aqueous mobile phase for 1hr at a
flow rate of 2ml/min. When the experiment was complete acetonitrile
was flushed through the column for 1hr at a flow rate of 2ml/min to
remove the aqueous buffer. The sample injection volume was 20pl.
The separation was isocratic with a 0.02M potassium hydrogen
phosphate solution, pH 4.7-5.1, as the mobile phase. This mobile
phase was used because TMP was retained for a reasonable length of
time on the column. Doubly distilled deionized water exposed to UV
light for 24hr was used to prepare the mobile phase. The mobile
phase was filtered through a 0.45pm filter before use. Exposure to
UV light photodecomposes many organic compounds which would interfere
with UV detection. The flow rate was 1ml/min. Separations of
electrolysis products were monitored at 220, 237, 260 and 305nm.
4.3 Product Analysis
Once the bulk electrolysis was complete the solutions were
quantitatively removed from the working electrode compartment with a
pipet and transferred to a clean glass vial. The solution in the
vial was immediately frozen in a dry ice acetone bath (-700C) and
lyopholized using a Freeze Dry-8 (model 75040, Labconco Corporation,
Kansas City, Missouri). The dry residue was dissolved in 1-2ml of
doubly distilled deionized water and separated on a liquid
chromatographic column (38cm x 3.0cm) packed with G-10 Sephadex size
exclusion gel (molecular weight cut off was 700) using doubly
distilled deionized water as the mobile phase. The G-10 resin was
obtained from Pharmacia Fine Chemicals (Piscataway, New Jersey). The
flow rate was maintained at approximately 0.065ml/min. An LC 200
Fraction Collector (HaakeBuchler Instruments, Saddlebrook, New
Jersey), was used to collect separated fractions. Ca. 80 drops
(3.4ml) were collected per fraction. Sixty fractions were
collected. All products eluted within this volume. Some of the
products coeluted approximately with each other.
It was found that in the separation of electrolysis products, as
the pH of the electrolysis solution increased, the region where
phosphate eluted in the separation decreased. For optimum separation
of products from phosphate, electrolysis was carried out in pH 9.5
phosphate buffer. It was verified that the same products formed at
every pH. Separation from phosphate was important because phosphate
can be an interference in GC/MS analysis of the silylated
Electrolysis was also attempted in ammonium acetate buffer
(NH4Ac) (pH 6.6). This buffer was chosen because it can be removed
by lyopholization which eliminates the need to separate products from
phosphate. However, it was still necessary to separate the products
from each other by GPLC so that they could be analyzed further by NMR
and FTIR. The results from such separations indicated that the major
products formed in phosphate buffers also formed in NH4Ac. However,
the GPLC retention times of the products formed in NH4Ac were so
similar that this method of product separation was not pursued
The presence of free phosphate in each fraction was detected by
a wet chemical test. Five drops of the sample were acidified with 2
drops of 0.01M nitric acid and a few crystals of ammonium molybdate
(Fisher Scientific) were added. The formation of a yellow
precipitate indicated the presence of phosphate. Phosphate bound to
the ribose group was not detected by this method. Only fractions
free of phosphate were lyopholized and analyzed.
The absorbance of each fraction was measured separately in a 1cm
quartz cuvette with the spectrophotometer described earlier. The
reference quartz cuvette contained doubly distilled deionized
water. A spectrum was recorded from 200 to 400nm for each
fraction. All products absorbed at 220nm.
The fractions which contained significant amounts of products as
indicated by UV spectroscopy were analyzed by HPLC. The HPLC
analyses were monitored at 237, 260 and 305nm.
Gas chromatographic/mass spectral analysis of the oxidation
products was carried out on a Finnigan Model 4021 GC/MS system with a
DB-5 (0.32mm id x 15m) capillary column (J & W Scientific, Rancho
Cordova, California). Electron impact, EI, spectra using a 70eV beam
and chemical ionization, CI, spectra using methane gas were
obtained. The injection volume was 5pl.
Since purines and their metabolites are often too polar and
thermally unstable to be volatilized, the oxidation products were
silylated before GC/MS analysis to insure volatility. The dry
samples were derivatized for 10-20min at 120C with 70ul
bis(trimethylsilyl)trifluoroacetamide (BSTFA) from Supelco Inc.
(Bellefonte, Pennsylvania) and 70pl silylation grade acetonitrile
from Pierce Chemical Co. (Rockford, Illinois) in 3ml reaction vials
(45). The silylations were carried out in a Reacti-Therm heating
module (Pierce Chemical Company). The reaction vials were tightly
sealed with teflon lined plastic caps (Supelco, Inc.).
Methane CI mass spectra give characteristic M+1, M+29 and M+41
ions corresponding to the addition of H+, C2H5+ and C H5+ to the
molecular ion, M+ (46). In a silylation acidic protons are replaced
by Si(CH3)3 groups (47). Typically, silylated samples give large
M-15 peaks corresponding to the loss of CH3 from the Si(CH3)3 group
(46, pp. 295-305;47). Silylation can result in overlapping mass
spectra of partially silylated derivatives which may not be separated
by GC. The resulting mass spectra then give sets of M and M-15 peaks
separated by 72 mass units; 72 corresponds to the molecular weight of
Si(CH3)3 minus the H it would replace. During silylation
decomposition of the oxidation products may occur (7).
A computer program developed by Toth et al. was used to analyze
the mass spectra of the silylated derivatives (48). This program
allowed determination of the number of silyl groups and the molecular
weight of the compound before silylation.
Fast atom bombardment mass spectrometry (FABMS) was also used to
confirm the molecular weight assignments. In FABMS there is little
fragmentation and polar samples can be analyzed without the need for
derivatization (49,50,51). The spectra were obtained with a Kratos
MS-50 mass spectrometer equipped with an RF magnet and DS-55 data
system. Fast xenon atoms (8eV) were used. The compounds were
dissolved in a thioglycerol matrix. Both positive and negative ions
were detected. The sensitivity of this technique is dependent on the
sample concentration and its susceptibility to ionization (50,51).
This analysis was carried out at the Middle Atlantic Mass
Spectrometry Facility at Johns Hopkins University School of Medicine.
To obtain additional structural information about the oxidation
products FTIR spectra were run. A Nicolet 5-DX spectrometer was
used. The samples were prepared in a Nujol mull and analyzed on KC1
or NaCl plates. Data were obtained after 126-256 scans. Nujol
absorbs in the region from 2850-3000cm-1 and 1400-1500cm-1 and the
spectra were background corrected. Because of the very small amount
of sample available it was not possible to correct for the background
100%; therefore, a spectrum of the background was overlayed on each
FTIR to distinguish between sample and Nujol absorptions.
Analysis of products by proton NMR was also used to obtain
structural information. The NMR spectra were run on a Nicolet NT-300
system. The samples were 0.2-0.5mg dissolved in l00l of deuterated
dimethyl sulfoxide (dmso-d6) (99.9 atom %D, MSD Isotopes, Montreal,
Canada) and pipetted into microcapillary NMR tubes. The NMR tubes
were rinsed with acetone and dried in an oven for 2 days at 900C to
remove water. The tubes were then flushed with N2 before capping.
Data were acquired for about 24hr. Since the oxidation products were
hygroscopic water was present in the samples and was evident in the
NMR spectra between 2-4.5 ppm. Presaturation of the water peak was
done to enhance resolution (52).
To determine the electrode area a chronocoulometric experiment
was carried out using a Bioanalytical Systems BAS 100 Electrochemical
analyzer (West Lafayette, Indiana). A 1mM solution of potassium
ferricyanide, K3Fe(CN)6 (Mallinkrodt) was prepared in a 0.50M KC1
solution and was used to determine the electrode area. The potential
was stepped from 359 to -172mV and the pulse width was 250ms. The
electrode area was determined from the slope of an Anson plot (i.e.,
charge versus square root of time) using equation 3-15 (41,42).
4.5 Photooxidation Reactions
The Tracor Northern spectrophotometer was used to record the
spectral changes during irradiation of various TMP solutions in a 1cm
quartz cell with a deuterium lamp (type L1637, Hamamatsu TV Co.,
Ltd., Japan). The operating current of the lamp was 0.3amps. All
solutions were less than 100UM in TMP. For solutions at TMP
concentrations greater than 100M the detector was saturated and
spectral changes were difficult to see. The spectral changes were
monitored every 15-30min. The photooxidation of TMP was also
monitored by HPLC using the same HPLC conditions described to monitor
the constant potential electrolysis of TMP.
To determine if oxygen is necessary for the oxidation to proceed
the photooxidation of TMP was carried out in a pH 7 phosphate buffer,
y=0.5M, exposed to air and bubbled with N2. The photooxidation was
also carried out in a nonaqueous solvent, DMSO, to determine if water
is necessary for the oxidation to proceed. The photooxidation of TMP
was also performed in water alone and the results indicate that the
presence of phosphate does not affect the spectral changes during
photooxidation in a pH 7 phosphate buffer.
Photooxidation experiments were also carried out to determine if
singlet oxygen was involved in the photooxidation. The singlet
oxygen quenchers sodium azide (0.01M) (Fisher Scientific) and
diazabicyclo[2.2.2]octane (DABCO) (.0011M) (Sigma Chemicals) were
dissolved in 60-80M solutions of TMP. Supporting electrolyte was a
pH 7 phosphate buffer, u10.5M. The UV spectrum in the region from
200 to 400nm was recorded after 4hr of irradiation with the deuterium
lamp. Photooxidations of sodium azide and DABCO solutions with no
TMP present were also carried out to determine if these compounds
were degraded by the UV light after 48hr.
A similar photooxidation experiment was carried out with Trolox
(6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid), a water
soluble antioxidant (Aldrich). This antioxidant is a quencher of
radical species (53). Spectral changes observed in a solution of
Trolox indicate that it is decomposed by UV light thus further
experiments were not performed.
4.6 Enzymatic Oxidation Studies
Enzymatic oxidations were carried out with xanthine oxidase
(E.C. 18.104.22.168, grade I from buttermilk, 0.42units/mg prot, 24mg
prot/ml, Sigma Chemical), chloroperoxidase (E.C. 22.214.171.124, purified
grade, 1556units/mg prot, 15mg prot/ml, Sigma Chemical) and the
cytochrome P450 monooxygenase system (E.C. 126.96.36.199) in a rat liver
microsomal pellet (54). All enzymatic oxidations were carried out in
pH 7 phosphate buffer, ionic strength 0.5M. The E.C. number for each
enzyme consists of four numbers separated by periods. The first
number defines the class (one of six reactions) to which the enzyme
belongs. The next two numbers indicate subclass and sub-subclass.
The fourth number is a serial number given to each enzyme in its sub-
In the chloroperoxidase oxidations 3ml of a 480pM TMP solution
in pH 7 phosphate buffer contained 1.6E-3M H202 and ca. 8.9E-8M of
chloroperoxidase (MW=42,000g/mole). In xanthine oxidase oxidations
the enzyme substrate molar ratio was 1:40. The typical concentration
of xanthine oxidase was 1pM. The molecular weight of xanthine
oxidase is ca. 260,000g/mol. All oxidations were monitored
spectrally and the oxidations were carried out in a 1cm quartz
Microsomal oxidations were performed with microsomal pellets
(obtained from Dr. James in the pharmacology department at the
University of Florida) which were prepared according to described
procedures (56). Briefly, a male rat liver is homogenized in a
buffer and a microsomal pellet is obtained after 1hr centrifugation
at 100,000 x g. The pellets were obtained from the livers of male
rats dosed with phenobarbitol to increase the cytochrome P450
concentration. The cytochrome P450 concentration in each pellet as
determined by standard procedures ranged from 2-44nmol/mg liver (57).
A UV difference spectrum was obtained to determine if TMP was a
substrate of cytochrome P450 (58,59). A difference spectrum reflects
changes in the electronic configuration of the iron porphyrin
prosthetic group of cytochrome P450 resulting from specific types of
interactions with a compound (58,59).
Since the microsomal fractions had been stored at -70C before
use it was necessary to verify that the enzymes were still active.
It was also important to test if a cyclic voltammetric assay, using a
rough PG working electrode, was practical. A cyclic voltammetric
assay was chosen because the microsomal solution is too opaque to
measure small absorbance changes. This was done by incubating (380C)
a 10ml solution containing 300l1 microsomal pellet, 9.8mM MgCl2,
564yM NADPH and 2004M uric acid in a pH 7 phosphate buffer, u=0.5M,
and recording a cyclic voltammogram every 5-10min of incubation. The
solution was shaken vigorously to dissolve the added uric acid. The
uric acid oxidation peak at 0.28V in the cyclic voltammogram
disappeared after 1hr incubation. This showed that the microsomal
enzymes were still active and that the cyclic voltammetric assay
could be used to monitor the oxidation of an electroactive species.
A similar experiment was performed on a 10ml solution containing
450PM TMP, 9.8mM MgCl2 and 280pM NADPH in phosphate buffer pH 7,
y=0.5M. Cyclic voltammograms were recorded every 10min of
incubation, up to 160min. After incubation the solution was
centrifuged for 10min and a UV spectrum of the supernatant was
recorded, from 200 to 400nm, to determine if any UV absorbing
products had been produced. The presence of any UV absorbing
products could not be determined because the solution was still too
opaque to give a clear spectrum.
4.7 Electrochemical Thermospray Mass Spectrometry (EC/TSP/MS)
Using the technique of EC/TSP/MS the electrolysis products can
be analyzed by mass spectrometry immediately after they are generated
electrochemically (60,61). In this technique the analyte is pumped
through an electrochemical cell held at the desired potential.
Electrolysis occurs and the products are forced into a heated
capillary tube of a thermospray ion source. Complete vaporization of
the liquid occurs to produce a superheated mist. This mist contains
molecules which are randomly positively and negatively charged. The
mass to charge ratio of these ions is then analyzed by a mass
The EC/TSP/MS experiments were carried out with a system
containing a coulometric cell (ESA, Inc.) which was coupled via a
Vestec thermospray interface to a Finnegan MAT TSQ45 triple
quadrupole mass spectrometer with an INCOS data system. Both counter
and reference electrodes were palladium. The working electrode was a
reticulated vitreous carbon block of large area (12cm 2), insuring ca.
100% electrochemical conversion efficiency. The cell volume was
5pl. The thermospray probe tip and source block temperature were
2500C and 2900C, respectively.
The conditions for TSP/MS were scan range m/z 125-300, electron
multiplier voltage 1000V and preamplifier gain 108 V A-1. For MS/MS,
the scan range was m/z 15 to 300, electron multiplier voltage 1600V,
preamplifier gain 108V A-1 collision energy 25eV and N2 collision gas
pressure 1.8mTorr. The mobile phase was 0.1M ammonium acetate
(pH 7), and the flow rate was 2.0ml/min. Twenty-five microliters of
a 300 yM solution was injected. The working electrode voltage was
varied from 0.8 to 1.2V. Both positive and negative ions were
analyzed by mass spectrometry.
CYCLIC VOLTAMMETRY OF TUBERCIDIN-5'-MONOPHOSPHATE
5.1 Oxidation and Reduction Behavior, Reversibility
The typical cyclic voltammogram of a 600M TMP solution in pH 7
phosphate buffer at a PG electrode shows an oxidation peak at ca.
0.90V and a broad shoulder at ca. 1.0V versus SCE (Fig. 5-1). Three
reduction peaks are evident at -0.56, -1.09, and -1.22V on the
reverse scan. These reduction peaks are present in the cyclic
voltammogram only after the oxidation peak has been scanned.
The sharp symmetrical peak shape of the oxidation peak suggests
an adsorption process. A study of the effect of concentration (Co)
of TMP on the oxidation peak current (i p) can give information about
the presence or absence of adsorption (section 3.1.1). In pH 7
phosphate buffer, P=0.5M, the plot of ip versus Co (Fig. 5-2a) shows
a change in the slope at 156gM. The slope changes from 1.17E+6 to
2.00E+5PA/M. Since the ip versus Co plot shows isotherm-like
behavior adsorption of TMP onto the electrode surface is indicated.
To further verify if the process is adsorption controlled the
dependence of ip on scan rate (v) was examined. For the TMP
oxidation peak in pH 7 phosphate buffer, y=0.5M, at a Co of 1001M
(before the change of slope of the ip versus Co plot), the plot of
logi versus logy has a slope of 0.8280.123 (Fig. 5-2b). At a C of
scan rate: 200mV/s
Figure 5-1 Cyclic voltammogram of TMP at a rough PG electrode, 600UM
solution in pH 7 phosphate buffer, y=0.5M.
I I I I
SI I I I
slope = .828
Effect of concentration and scan rate on the oxidation peak
current of TMP in a pH 7 phosphate buffer, u=0.5M: a) plot of
peak current versus concentration for the oxidation peak of TMP,
scan rate 200mV/s; b) plot of log peak current versus log scan
rate for a 100pM TMP solution (c.c. = correlation coefficient).
998yM (after the change in slope of the ip versus Co plot) the same
plot has a slope of 0.5120.0998 (Fig. 5-3). Thus as indicated by
the slope values, at low Co the TMP oxidation peak is primarily
adsorption controlled and at high C it is primarily diffusion
The oxidation appears to be irreversible since no corresponding
reduction peak is observed on the reverse scan. However, the peak
potential dependence (E ) on pH indicates that the oxidation is not a
simple irreversible process. For simple irreversible reactions Ep is
independent of pH. The absence of the reduction peak may be due to a
fast following chemical reaction occurring after electron transfer.
5.2 Effect of pH
The cyclic voltammetric behavior of TMP varies with buffer pH.
The cyclic voltammograms in Figure 5-4 illustrate this. As the
phosphate buffer pH increases from 2.8 to 9.5, the oxidation and
reduction peaks shift to more negative potentials and become
broader. At pH 9.5 the oxidation peak is a combination of two peaks
(Fig. 5-4c). In a pH 6.6 ammonium acetate buffer (Fig. 5-4d) the
oxidation peak is sharper than in a pH 7 phosphate buffer (Fig. 5-4b)
but the peak potentials are comparable. It is clear that the cyclic
voltammetric behavior of the oxidation peak is affected by buffer
The increased broadness of the oxidation peak in phosphate
buffer as the pH increases indicates that either the kinetics of the
. | I -
. I I I I I I I
(Vri)di 6o| m
scan rate: 200mV/s
I 50 IA
1.0 0.5 0
Figure 5-4 Cyclic voltammetric behavior of TMP at a PG electrode in
different buffers, p=0.5M: a) pH 2.8 phosphate; b) pH 7
phosphate; c) pH 9.5 phosphate; d) pH 6.6 ammonium acetate.
reaction decreases or adsorption has less of an effect on the
oxidation. It seems contradictory that the kinetics of the reaction
should decrease when the oxidation peak shifts to less positive
potentials which indicates that the rate of the reaction slows down
when the oxidation becomes easier. However, the shift of the peak
potential to less positive values may be a result of different
acid/base forms of TMP being involved in the redox reaction at
different pH (62). The dependence of Ep on pH indicates that protons
may be involved in the rate determining step or in a preceding
chemical reaction. The complex nature of this reaction prevented
application of existing theories to the elucidation of the reaction
kinetics and the mechanism of the first e/H+ transfer steps.
The dependence of Ep on pH is defined by the Nernst equation
(equation 3-6) for a fast electrochemical process in which protons
are involved in the rate determining step. The Ep versus pH
measurements were carried out at scan rates of 5mV/s because at this
slow scan rate the system may be assumed to be at equilibrium and the
Nernst equation applies.
From the slope of a plot of Ep versus pH (Fig. 5-5) the number
of protons (r) and electrons (n) involved in the oxidation of TMP at
a PG electrode were estimated. In the pH region from 2 to 5, the
slope is 68mV/pH suggesting an r/n ratio of 1:1 (Fig. 5-5). At a pH
of ca. 5.7 the peak potential dependence on pH changes slope. This
corresponds closely to the pKa of 5.3 for tubercidin reported in the
literature (10, p. 316). Above pH 5.7 the slope is 31mV/pH,
suggesting an r/n ratio of 1:2.
I I I p I I I I
The pKa value was also verified spectrally. In the pH region
from 2 to 4.6 the UV spectrum of TMP gives three absorption bands at
205, 225 and 270nm (Fig. 5-6a). In the pH region from 6.8 to 11
(Fig. 5-6b) the UV spectrum of TMP gives only two absorption bands at
210 and 270nm. The purine base of TMP probably exists in a
predominantly protonated form below the pH=pKa and a predominantly
neutral form above the pH=pKa since structurally similar adenine is
predominantly protonated below pH=pkal (4.2) and predominantly
neutral above pH=pKal (63). Crystallographic data show that the
cationic proton of the aminopurine is attached to the 1-nitrogen of
the purine ring (Fig. 1-2) (64). This may also be the case for the
structurally similar purine base of TMP. At a pH below the pKa of
TMP the phosphate group on TMP is negatively charged. The net result
may be a zwitterionic TMP molecule at pH below 5.7.
5.3 Electrode Surface Effects
The cyclic voltammetric behavior of TMP was studied at PG, GC
and pretreated GC electrodes. The cyclic voltammetric behavior was
affected by the nature and condition of the carbon-electrode surface.
Pyrolitic graphite is formed by depositing carbon from the vapor
phase on the surface of a substrate, usually a metal. This process
results in a highly oriented polycrystalline form of carbon. The
substrate will always contain minute flaws which serve as nucleation
sites for the growth of cones. These cones are the main structural
characteristics of PG (26, p. 303; 65,66). To activate the PG
240 260 280 300 320 340 360 380 400
220 240 260 280 300 320 340 360 380
Figure 5-6 Spectral behavior of TMP (ca. 150pM) in phosphate
buffers, P=0.5M, of pH a) 2.8-4.6 and b) 6.8-11.
surface it is usually polished by grinding on 600 grit silicon
carbide paper before voltammetry.
Glassy carbon is formed by slowly heating a polymeric resin in
an inert atmosphere. A carbonization process starts as the
temperature increases above 3000C where oxygen, nitrogen, hydrogen
and anything else is eliminated until only carbon is left. The
material obtained by this process is referred to as vitreous or
glassy carbon. The structure of GC is believed to be a network of
tangled aromatic ribbon molecules which are cross-linked by highly
strained carbon-carbon bonds (26, p. 308; 30). It is isotropic,
impermeable to gas, electrically conductive and resistant to chemical
attack (26, p. 308). The surface is polished to a smooth mirror-like
finish with polishing alumina before use as an electrode.
Figures 5-7b and d show the cyclic voltammetric behavior of a
600uM TMP solution in pH 7, p=0.5M, phosphate buffer at a PG and GC
electrode. Figures 5-7a and c show the cyclic voltammograms in
buffer at these electrodes. The TMP oxidation peak is sharp and
symmetrical at PG and broad at GC. This suggests that the surface
plays a role in the oxidation of TMP. It was determined that the TMP
oxidation is an adsorption controlled process at PG from studies of
the behavior of peak current on concentration and scan rate (Figs.
5-2 and 5-3).
The diffusion nature of the TMP oxidation peak at a GC electrode
was confirmed by a linear plot of ip versus concentration for TMP
solutions of 63-575gM in pH 7 phosphate buffer, u=0.5M.
Figure 5-7 Cyclic voltammograms in a a) pH 7 phosphate buffer,
0=O.5M, at rough PG; b) 600vM TMP solution in pH 7
phosphate buffer at rough PG; c) same as a, at GC; d)
same as b, at GC.
scan rate: 200mV/s
I I -1.0 I
0 -0.5 -1.0 -1.5
To provide additional information about the electrochemical
behavior of TMP at GC the dependence of i on v at GC was also
examined. Scan rates ranged from 20 to 500mV/s. The plot of logip
versus logv for a 106iM TMP solution in pH 7 phosphate buffer has a
slope of 0.5630.754. At a concentration of 575pM, the same plot has
a slope of 0.5110.231. Thus at both concentrations the TMP
oxidation peak at GC occurs predominantly by diffusion.
Cyclic voltammetry at fast scan rates shows no corresponding
reduction peak indicating that the oxidation peak at GC is also
Since ip is proportional to electrode area and the area of the
GC and PG electrodes are not the same; the peak currents cannot be
directly compared. The normalized peak current, pA/cm2, however, can
be directly compared. To determine the normalized peak current the
electrode area was estimated using chronocoulometry (section
3.4.2). The results indicate that the GC electrode is about twice as
large as the PG electrode (84E-4cm2 versus 49E-4cm2). Using these
values, the normalized peak current for the oxidation peak at GC and
PG in Figure 5-7 is 501 and 1708IA/cm respectively.
Electrochemical pretreatment in many cases improves the
performance of GC electrodes (i.e., enhances the peak currents and
improves reversibility) (44,67,68). It has been suggested that
electrochemical pretreatment introduces or alters functional groups
on the electrode surface. These functional groups may facilitate
electrode reactions in some way. Specific chemical interactions such
as adsorption or changes in hydration of the electroactive species
may also be occurring as a result of pretreatment (69,70).
Electrochemical pretreatment of GC was carried out to observe the
effects on the TMP oxidation peak.
Electrochemical pretreatment was performed by applying a
constant positive potential of +1.8V versus SCE for 8min to the GC
electrode, followed by 2-5 cyclic scans towards negative potentials
first (+1.8 to -1.7V) at 20mV/s (68). These cyclic scans are carried
out until the background peaks at ca. -1 and +1V disappear (section
Figure 5-8 shows the cyclic voltammograms of a TMP solution in
pH 7 phosphate buffer, y=0.5M, at GC before and after electrochemical
pretreatment. The cyclic voltammograms in a pH 7 phosphate buffer
alone (Fig. 5-9) illustrate how the background is affected by
pretreatment. The cyclic voltammogram of TMP in Figure 5-8b
illustrates that pretreatment does not improve the appearance of the
oxidation peak at ca. 1V which is difficult to discern from the
background. Thus electrochemical pretreatment did not enhance the
TMP oxidation peak current or improve the kinetics of the oxidation.
5.4 Behavior in Organic Solvents and Solvent Mixtures
The cyclic voltammetric behavior of TMP is also affected by the
composition of the solvent. TMP gives no oxidation or reduction
peaks in a solution of dimethylformamide (DMF) with 0.05M tetrabutyl
ammonium perchlorate (TBAP) added as the supporting electrolyte.
Adding successive amounts of organic solvent (DMF) to a solution of
E (UOLT I
Cyclic voltammograms of a 575uM TMP solution in pH 7
phosphate buffer, P=0.5M, at a glassy carbon electrode:
a) before pretreatment and b) after pretreatment. Scan
Cyclic voltammograms in pH 7 phosphate buffer, P=0.5M, at
a GC electrode: a) before pretreatment and b) after
TMP in pH 7 phosphate buffer, decreases the oxidation peak at a PG
electrode (Fig. 5-10). No shift in peak potential is observed.
After addition of 40% organic solvent, the peak has decreased 56%.
After addition of 30% organic solvent, the shape and the magnitude of
the TMP oxidation peak at PG is similar to that at GC for the same
concentration of TMP in pH 7 phosphate buffer (Fig. 5-7d).
Mixed water organic solvent systems are commonly used to prevent
adsorption of surface active organic compounds (70). This type of
medium may lead to an increase in solvent-solute interactions over
electrode solution interactions. The peak due to TMP at a PG
electrode may decrease in this medium because the organic solvent may
prevent adsorption of TMP onto the PG electrode surface.
At a GC electrode the oxidation is predominantly controlled by
diffusion (section 5.3). If adding organic solvent prevents
adsorption of TMP onto the PG surface, then it should not affect the
electrochemical behavior of TMP at GC. Figure 5-10 shows that
addition of organic solvent (DMF) to an aqueous solution of TMP has
little effect on the ip of TMP at GC. These results indicate that
addition of organic solvent may be preventing adsorption of TMP onto
the PG electrode surface. It is clear that TMP behaves differently
at GC and PG.
TMP gives no cyclic voltammetric peaks in an organic solvent.
If water is added to a solution of TMP in DMF with TBAP added as a
supporting electrolyte no oxidation or reduction peaks appear in the
cyclic voltammogram when 40% water is added. If phosphate buffer is
Plot of TMP oxidation peak current at PG and GC versus %
DMF added to a 600pM TMP solution in pH 7 phosphate
added to a similar solution of TMP instead of water, the poor
solubility of phosphate causes it to precipitate out of solution.
DETERMINATION OF THE n-VALUE OF TUBERCIDIN-5'-MONOPHOSPHATE
When the constant potential electrolysis of TMP was carried out
the coulometer continued to accumulate charge after TMP was
completely electrolyzed as shown by HPLC. Cyclic voltammograms and
UV spectra did not clearly show when the electrolysis of TMP was
complete because products which formed had oxidation peaks and UV
absorbance bands that overlapped with those of TMP.
When TMP is completely electrolyzed in a pH 7 buffer, u=0.5M, at
rough PG, as indicated by the disappearance of the TMP peak in the
HPLC, the n-value was determined to be 11 by coulometry (section
3.2). This n-value is independent of concentration and depends on
pH. When TMP is electrolyzed in pH 9.5 buffer, u=O0.5M, the n-value
was determined to be 5.2 by the same method. A background correction
was carried out at a typical electrolysis potential of ca. 0.97V
versus SCE. The values of charge that were obtained were ca. 1% of
the total charge value obtained during electrolysis of TMP. The
background corrected values are reported.
The accumulation of charge which is observed after TMP has been
completely electrolyzed indicates that there may be a catalytic
oxidation of water occurring at the typical electrolysis potentials
in solution in presence of TMP and/or its oxidation products in
solution. There are reports which suggest that functional groups on
the electrode surface can cause the catalytic oxidation or reduction
of species in solution (38,40). It is also possible that the initial
electrolysis products of TMP are being further oxidized. This is
supported by the product analysis (section 8). Thus the n-value
obtained by this method may be inflated.
6.2 Peak Current Comparisons
Since the n-value of TMP obtained from coulometry may be
inflated the n-value was also estimated by comparing the cyclic
voltammetric peak current of TMP to that of structurally similar
compounds. The ip for reversible and irreversible systems is defined
in equations 3-1 through 3-4. The cyclic voltammetric i values of
compounds with known n were compared to the ip value for TMP. By
assuming that both systems were diffusion controlled, the Do values
were the same, using the same electrode, concentration and scan rate,
equations 6-1 and 6-2 were used to determine the n-value of TMP.
Equation 6-1 compares the ip of a reversible (r) standard to the ip
of an irreversible (ir) system (TMP).
diffusion controlled systems
i (ir)u 1.11[C *n(na) /2](ir) u
i p(r)k [C *n3/2(r)k
i (ir) C *n(na) 1/2(ir)
i (ir)u 1na)/2(r) (6-2)
p u [C0*n(ana) ](ir)
o a. u
If both systems were assumed to be adsorption controlled (r *
values and electrode area the same), equations 6-3 and 6-4 were used
to determine the n-value of TMP.
adsorption controlled systems
i (ir) 1.47[nan a](ir)u
p u a-- u(6-3)
i (r) 2
p k n (r)k
i (ir) [nan ](ir)
p(ua u 6-4)
i (ir) [nan ](i r) k6
p k a k
Equations 6-1 and 6-3 compare the ip of a reversible (r)
standard to the ip of an irreversible (ir) system (TMP). From cyclic
voltammetry TMP oxidation appears to be irreversible and adsorption
controlled (section 5.1). Equations 6-2 and 6-4 compare the ip of
two irreversible systems (r), where k and u represent the known and
unknown systems, respectively.
Equations 6-5 and 6-6 could be used to compare the ip values of
an adsorption and diffusion controlled system but the values for ro*
and Do are not known. Therefore, to estimate n by comparing
Pads (nanaF 2o *)ads
S ( 731,142RT(n3/2 D012 C */2)
ipads (ir) (n 1/2 r 1/2 1/2V)ads (6-6)
idiff (r 812,682RT[n(an ) 1/2C *D v 2]
cyclic voltammetric ip values, both systems were assumed to be either
adsorption or diffusion controlled.
The compounds that were used for oxidation peak comparisons were
xanthine, guanine and uric acid. The oxidations of guanine and
xanthine are four-electron irreversible processes and the oxidation
of uric acid is a reversible two-electron process (6,7,33,34). These
oxidations are primarily adsorption controlled at a rough PG surface
(1,2,6,7,33,34). Equations 6-1 and 6-3 were used to determine the n
of TMP by comparison to uric acid and equations 6-2 and 6-4 were used
to determine the n of TMP by comparison to xanthine and guanine. The
ana values were estimated by measuring the shift in Ep at two sweep
rates using equation 3-8. Table 6-1 illustrates the results that
were obtained which estimate the n-value of TMP to be between 4.57
and 7.32 if both systems are assumed to be diffusion controlled and
between 4.68 and 10.90 if both systems are assumed to be adsorption
In order to determine the validity of this approach n-values of
compounds whose n-values have been reported in the literature were
determined in a similar manner. In this study, uric acid, guanine
and xanthine were compared with each other. The results in Table 6-2
show that estimating the n of xanthine by comparing it to uric acid
using equation 6-1 gives an n of 4.32 for xanthine (+8% error).
Estimating the n of uric acid by comparing it to xanthine using
equation 6-1 gives an n of 1.90 for uric acid (-5% error). The
errors of other comparisons range from -32 to +132%. It appears that
the % error is less when equations 6-1 and 6-2 (which assume a
diffusion controlled system) are used to estimate n.
Estimation of the n-Value of TMP from Cyclic Voltammetric
1 2 3 4
compound ana conc,pM 1p, A n nTMP
Guanine (ir) .546 101.6 24.453.97 4(7) 4.575,4.687
Xanthine (ir) .512 101.3 15.942.4 4(6) 6.775,6.747
Uric acid (r) 101.0 13.09.89 2(33,34) 7.326,10.908
TMP (ir) .515 101.0 27.02.33 ? --
1 an determined from equation 3-8. The scan rates were 100 and
2 All solutions were prepared in pH 9.5 phosphate buffer, u=0.5M,
working electrode was rough PG.
4 Determined at a scan rate of 100mV/s.
5 Literature value, references in ( ).
5 Determined from equation 6-2, u=TMP.
6 Determined from equation 6-1, u=TMP.
8 Determined from equation 6-4.
8 Determined from equation 6-3.
* Uric acid is a reversible reaction; therefore, it has no ana value.
Validity of Peak Current Comparisons of Estimation of
comparison1 ncalc. n6 %error ncalc. n6 %error
U vs G G=6.402,9.303 4(7) +61,+132 U=1.462,1.313 2(33) -28,-34
U vs X X=4.322,6.473 4(6) +8,+62 U=1.902,1.573 2(33) -5,-21
G vs X X=2.70 ,2.785 4(6) -32,-30 G=5.924,5.755 4(7) +48,+44
1 G-guanine, U=uric acid, X-xanthine.
2 Determined using equation 6-1.
3 Determined using equation 6-3.
Determined using equation 6-2.
4 Determined using equation 6-2.
5 Determined using equation 6-4.
6 Literature value, references in ( ).
Note: Values for i and an were given in Table 6-1.
Experimental conditions were the same as in Table 6-1.
The calculations summarized in Tables 6-1 and 6-2 were made
using equations 6-1 to 6-4. The calculations are based on the
assumption that both systems were diffusion controlled or both
systems were adsorption controlled. Uric acid, xanthine, guanine and
TMP are primarily adsorption controlled systems at the PG surface.
However, the results in Table 6-2 indicate that the % error is less
than when the systems are assumed to be diffusion controlled.
Diffusion coefficients and electrode area were assumed to be the same
which may not be valid. For example, the electrode area changes
every time the electrode is resurfaced. Typical precision of
measurements at a PG electrode is ca. 5-10%. The bulk concentration
and the concentration at the electrode surface may also be different.
Because of the large errors associated with the determination of
n-values by cyclic voltammetric peak current comparisons, the n-value
of TMP was also estimated using chronocoulometry. In a chronocoulo-
metric experiment the electrode potential is stepped from a potential
where no electrolysis takes place (Ei) to a potential where electro-
lysis occurs at a maximum diffusion controlled rate (Ef) (section
3.4.1). Chronocoulometry can be used to estimate the n-value of TMP
by comparing its slope to that of a compound whose n-value has been
reported. By using the same electrode and assuming the diffusion
coefficients are the same equation 3-14 can be used for this
purpose. Adenosine-5'-monophosphate (AMP) was used for this
comparison because it is structurally similar to TMP thus their
diffusion coefficients should be very similar.
Although the n of AMP has not been determined it was assumed
that the n of AMP equals that of adenine, which is 6 (3). This
assumption was based on the report that the presence of a ribose
group did not affect the value of n (5). The n-value of TMP that was
estimated using this approach was 3.32. The conditions and results
of this experiment are given in Table 6-3.
Table 6-3 Estimation of the n-Value of TMP Using Chronocoulometry
slopes conc,pM2 Ei,mV Ef,mV value
AMP 7.9167 1017 300 1250 6(3)3
TMP 4.3107 998 200 925 3.324
1 Determined from a chronocoulometric experiment (section 4.4).
2 All solutions were prepared in pH 7 phosphate buffer, y=0.5M.
3 Literature value, reference in ( ).
Determined from equation 3-14.
To validate this approach a similar comparison was carried out
to estimate the n of uric acid, xanthine, and guanine. Table 6-4
shows that this approach leads to an error of -36 to +57%. The error
may be due to the assumption that Do and A are the same for all of
these compounds when using equation 3-14 for the calculations.
Table 6-4 Validity of Chronocoulometric Comparison of Estimation of
compound slope2 conc,M3 k4 u5 nucalc6 nu7 %error
G .7517 101.9 G X 3.86 4(6) -3.5
X .7164 100.6 X G 4.14 4(7) +3.5
U .2286 96.96 U G 6.26 4(7) +57
G U 1.28 2(33) -36
U X 6.04 4(6) +51
X U 1.32 2(33) -34
2 G-guanine, X=xanthine, U=uric acid.
2 Determined from chronocoulometry (section 3.3) pulse width 500ms
for all compounds, for guanine and xanthine Ei=500mV and Ef=675mV,
for uric acid Ei=150mV and Ef=325mV.
3 All solutions were prepared in pH 7 phosphate buffer, u=0.5M.
4 k-known system in equation 3-14.
5 u=unknown system in equation 3-14.
6 Determined using equation 3-14.
Literature value, reference in ( ).
When the n-value is obtained by comparing the slopes of a plot
of charge versus t1/2 for xanthine and guanine (irreversible
processes) there is a small error, 3.5%, between estimated and true
n-values. When comparing an irreversible and reversible process
(e.g., xanthine to uric acid and guanine to uric acid), the error
ranges from -36 to +57%.
Using coulometry the n-value of TMP was estimated to be 11 when
the TMP peak disappeared in the HPLC. Under these conditions further
oxidation of initial electrolysis products or catalytic oxidation of
solvent may also be contributing to the n-value. The n-value of TMP
that was determined from cyclic voltammetric peak current comparisons
ranged from 4.57 to 7.32 when both systems were assumed to be
diffusion controlled. The n-value ranged from 4.68 to 10.90 when
both systems were assumed to be adsorption controlled. However, this
method of determination could be in error by as much as +132% as
indicated in Table 6-2. Chronocoulometric comparisons suggest that n
of TMP is ca. 3 assuming n of AMP is 6. This method of n
determination gives a small error (3.5%) when two irreversible
processes are compared and a significant error (-36 to +57%) when an
irreversible and a reversible process are compared. The discrepancy
between n-values determined by coulometry, cyclic voltammetry and
chronocoulometry probably results from the different time scale of
CONSTANT POTENTIAL ELECTROLYSIS OF TUBERCIDIN-5'-MONOPHOSPHATE
Using the technique of constant potential electrolysis the
stable intermediates and products from an electrode reaction can be
generated for further analysis (section 4.2). The electrolysis of
TMP was monitored by cyclic voltammetry, UV spectroscopy and HPLC to
provide information on the formation of intermediates and products.
It was concluded that TMP was completely electrolyzed when the
TMP peak disappeared in the HPLC (Fig. 7-1e). The UV spectrum at
this point resembles trace 4 in Figure 7-2 and the cyclic voltammo-
gram (Fig. 7-3c,d) shows a broad oxidation peak at potentials of TMP
oxidation. Separation of the products by GPLC gives two peaks (Fig.
7-4a). From the spectral changes observed (Figure 7-2) it is clear
that after TMP is completely electrolyzed oxidation of products must
continue. This is supported by cyclic voltammetry (Fig. 7-3) which
shows an increase in the reductions peaks at -1.09 and -1.2V and HPLC
(Fig. 7-1f) and GPLC (Fig. 7-4b) which show the formation of new
products after continued electrolysis at TMP oxidation potentials.
It is clear that cyclic voltammetry, UV spectroscopy and GPLC cannot
be used to determine when TMP is completely electrolyzed. Complete
electrolysis is easily determined using HPLC. If constant potential
electrolysis is continued after TMP is completely electrolyzed,
HPLC during the electrolysis of a 600pM TMP solution in
pH 7 phosphate buffer, u=0.5M, at a rough PG electrode:
a) before; b) 25min; c) 1hr; d) 2hr, e) 3hr, 22min; f)
\I I TMP i i
w 3HR. 22MIN.
u E 0.10-
Z 7 HOURS
tM E f
< i ,, i i
I I I ( I I ) I
0 5 10 15 20
. ... .,.'- "-.4H
**,, b .d A
....- l N H
*, 'I '. .
"Z /" / .
solution in pH 7 phosphate buffer, =05M, at a rough PG
solution in pH 7 phosphate buffer, v=0.5M, at a rough PG
Figure 7-3 Cyclic voltammograms of a 600VM TMP solution in a pH 7
phsophate buffer, 4=0.5M, during electrolysis at a rough
PG electrode: a) before; b) 2hr, 2min; c) 7hr, 2min.
e: 200mV/s -1.-1.2 before
1.0 0.5 0 -0.5 -1.0 -1.5
Volts vs SCE
E FREE PHOSPHATE
O A B
8 24 40 56 72 88 104 120 136 152 168
W 2.0 -
8 24 40 56 72 88 104 120 136 152 168
Figure 7-4 GPLC separation of TMP electrolysis products from a 600pM
TMP solution in a pH 9.5, P=0.5M, phosphate buffer after
a) 4-12hr and b) 60hr of electrolysis.
further oxidation of products occurs (Fig. 7-1f). This is clearly a
slow process since the GPLC (Fig. 7-4b) shows that after 60hr of
electrolysis only two new product peaks formed and that one of the
first formed products finally disappeared.
The time required to completely electrolyze TMP was a function
of the condition of the PG electrode surface. At a rough surface the
time ranged between 1-4hr and at a surface which was not roughened
the time ranged from 8-12hr. The electrolysis times at these
different surfaces varied because the degree of roughness was not
well controlled when the electrode was manually polished on silicon
carbide paper. Changes in the UV spectra show that the oxidation
pathway was also affected by the condition of the electrode surface
Electrolysis of TMP was carried out in phosphate buffer (pH 2.8,
7 and 9.5), u=0.5M, to optimize the separation of products from
phosphate by GPLC (section 4.3). The optimum pH for separation of
products from phosphate was 9.5. Since the objective of this
investigation was to provide insight into TMP reactivity at physio-
logical pH it was important to verify if the same products were
produced during electrolysis in both pH 9.5 and 7. Monitoring the
electrolysis of TMP in pH 9.5 and 7 phosphate buffers with HPLC
(section 7.4) verified that the same products were produced in both
pH 7 and 9.5.
The specific cyclic voltammetric, spectral and HPLC behavior of
TMP during electrolysis in pH 7 and 9.5 phosphate buffers will be
7.2 Cyclic Voltammetry
Figure 5-4 showed that the cyclic voltammetric behavior of TMP
before electrolysis depended on pH. The cyclic voltammetric behavior
also changed with pH during electrolysis. However, the same-major
products which were separated and analyzed formed at both pH 7 and
9.5. This was verified by HPLC (section 7.4).
Figure 7-3 shows typical cyclic voltammograms obtained during
electrolysis of TMP in a pH 7 phosphate buffer. The reduction peak
at -1.2V increases throughout the electrolysis and increase further
after TMP has been completely electrolyzed, according to HPLC. The
reduction peak at -0.56V disappears before TMP is completely
electrolyzed. A broad oxidation peak forms at 0.89V when the amount
of TMP has decreased by 77% as determined by HPLC. This new broad
oxidation peak does not increase further after TMP is completely
electrolyzed. The products) oxidized at this potential decreases
very slowly. The formation of this broad oxidation peak obscures the
TMP oxidation making it difficult to determine if TMP is completely
electrolyzed using cyclic voltammetry. Electrolysis at pH 7 was
carried out until no significant changes were observed by cyclic
voltammetry and UV spectroscopy.
Figure 5-4c showed the cyclic voltammogram of TMP in a pH 9.5
phosphate buffer before electrolysis. As the electrolysis proceeds
the reduction peak at -1.32V increases. Before TMP has been
completely electrolyzed a broad oxidation peak forms at slightly more
positive potentials than the TMP oxidation peak. This oxidation peak
does not increase as the electrolysis proceeds but rather slowly
decreases. The reduction peaks at -.66V and -1.22V disappear before
TMP is completely electrolyzed. This is similar to the behavior at
pH 7. When electrolysis in pH 9.5 is carried out for a very long
time, 60hr, all oxidation and reduction peaks in the cyclic
voltammogram disappear and the product absorbance at 305nm in the UV
spectrum disappears. The UV spectrum at this point shows no
Figures 7-2 and
electrolysis at a PG
has not. Since this
TMP at physiological
surface was confined
has been shown to be
7.3 UV Spectra
7-5 illustrate the spectral changes during
surface which has been roughened and one which
investigation was concerned with the behavior of
pH, the comparison of effect of electrode
to pH 7. The chromophore circled in Figure 7-2
responsible for the absorption peak at 270nm
The general behavior during electrolysis at a rough PG electrode
as determined by UV spectroscopy was similar regardless of pH. The
spectral changes are a function of electrolysis time as indicated in
Figure 7-2. During electrolysis the band of TMP at approximately
210nm decreases and a new band grows in at 230nm. The band due to
TMP at 270nm decreases and shifts to 265nm. Simultaneously a new
band forms at 305nm. The shift to 265nm is observed in both pH 7 and
The spectral changes during electrolysis at an unroughened PG
electrode (Fig. 7-5) are similar to those observed at a rough PG
280 300 320
UV spectra during the electrolysis of a 100pM TMP
solution in pH 7 phosphate buffer, p=0.5M, at an
unroughened PG surface.
electrode (Fig. 7-2) except that during electrolysis at an
unroughened surface isosbestic points are held at 217, 247 and
292nm. The presence of isosbestic points suggests that the reaction
proceeds through a simultaneous (A + B + C) rather than a consecutive
reaction pathway (A + B + C) (72).
Many of the results which were described are based on a reverse
phase HPLC assay that was developed and used to monitor the
electrolysis of TMP. This HPLC method is based on that developed by
P. Brown and co-workers for the separation of nucleosides,
nucleotides and bases (73,74). The conditions of this experiment
were described in section 4.2. Typically, in reverse phase HPLC
neutral compounds are retained while ionic compounds are only
minimally retained (73). As shown by our results, TMP was relatively
strongly retained using a pH 4.7-5.1 mobile phase and a nonpolar C-18
column. The purine base of TMP is believed to exist in a
predominantly protonated form in this pH region (section 5.2) but the
phosphate group of the nucleotide is negatively charged in this pH
region. The net result may be zwitterionic and possibly neutral
molecule in this pH region. For this reason the molecule may be
retained under these conditions (73).
A typical electrolysis in pH 7 phosphate buffer monitored by
HPLC is shown in Figure 7-1. Immediately after initiation of
electrolysis a major product with very short retention time formed,
peak I (Fig. 7-1b). In addition, three other peaks formed, peaks II,
III and IV (Fig. 7-1c). As electrolysis proceeds, peaks I, II, III,
and IV increase and a new peak, V, forms late in the electrolysis
when TMP is ca. 90% oxidized (Fig. 7-1d). When TMP is completely
oxidized (Fig. 7-ie) peak I has decreased ca. 41%, and peak III has
decreased ca. 66% from their maximum absorbance values. Peak II
continues to increase and peak VI forms after TMP is completely
oxidized. This indicates that peak VI must form from other products.
When electrolysis in a pH 9.5 phosphate buffer is monitored by
HPLC, peak V does not form but another product peak forms instead
which has a retention time between that of peak III and TMP. This
peak is the same height as peak II and decreases when TMP is
The HPLC results indicate that at least 6 UV absorbing products
form. Five peaks form before TMP is completely electrolyzed (peaks
I-V) and one forms after TMP is completely electrolyzed (peak VI).
The products corresponding to HPLC peaks I, II, IV and VI are the
major UV absorbing products at 225 nm after TMP has been completely
ANALYSIS OF TUBERCIDIN-5'-MONOPHOSPHATE ELECTROCHEMICAL
8.1 Separation by Gel Permeation Liquid Chromatography
Typical GPLC separations were shown in Figure 7-4 for different
electrolysis times. The experimental conditions were discussed in
section 4.3. Figure 7-4a shows a typical GPLC separation after 4-8hr
electrolysis of TMP at a rough PG electrode in pH 9.5 phosphate
buffer. At this point TMP has been completely electrolyzed and two
product peaks are present, A and B. Figure 7-4b shows a separation
after 60hr of electrolysis in a pH 9.5 phosphate buffer. The
fractions eluting under A and B in Figure 7-4a and 7-4b are the same
compounds) based on spectral analysis. The products) eluting under
peak A is still present while those eluting under peak B have
decreased considerably. Two new product peaks C and D have formed.
Peak D is a mixture of the product eluting under C and another minor
component based on analysis by GC. Since the other component under
peak D is a minor component in the mixture it was not analyzed
Separation of the solutions electrolyzed at intermediate times
(12-20hr) between the separations in Figure 7-4 showed one new very
broad product peak which eluted between A and B and coeluted with
phosphate thus it was not analyzed further. Since HPLC analysis of
the products eluting under GPLC peaks A and B verified that the
products did not decompose on the column it was concluded that the
products under GPLC peaks C and D were not decomposition products but
formed as a result of further electrolysis. Since GPLC peak B has
decreased considerably after 60hr electrolysis and peaks C and D have
formed it is quite possible that peak C or D formed from the
compounds) eluting under GPLC peak B.
8.2 Analysis of Products with High Pressure Liquid
Chromatography, Cyclic Voltammetry and
Analysis of the fractions eluting under GPLC peak A in Figure
7-4a using HPLC is shown in Figure 8-1a. This analysis indicates
that primarily one component elutes under this peak with a retention
time similar to HPLC peak I formed during electrolysis of TMP in
Figure 7-1b. A UV spectrum of the fractions under GPLC peak A is
shown in Figure 8-2. This product exhibits an absorption at 225nm.
Cyclic voltammetry of this compound shows no oxidation or reduction
A UV spectrum of the fractions eluting under GPLC peak B is
shown in Figure 8-3. The fractions under this GPLC peak give an
absorption maximum at 235 and 305nm. Analysis of the fractions
eluting under GPLC peak B using HPLC is shown in Figure 8-1b. This
analysis shows primarily two HPLC peaks which correspond to HPLC
peaks II and VI formed during the electrolysis of TMP in Figure
7-1f. Analysis of both peaks in Figure 8-1b, at 225nm, 260nm and
305nm, using HPLC indicates that the products corresponding to both
peaks have a significant absorbance at 225 and 305nm and a minimum
absorbance at 260nm. Thus the two UV absorbing components which
z E 0.02
HPLC of compounds under GPLC peaks a) A, b) B and c) C
(flow rate Iml/min).