Citation
The electrochemistry of ferritin using self-assembled monolayers on polycrystalline gold electrodes

Material Information

Title:
The electrochemistry of ferritin using self-assembled monolayers on polycrystalline gold electrodes
Creator:
Niichel, Robert Joseph
Publication Date:
Language:
English
Physical Description:
viii, 71 leaves : illustrations ; 29 cm

Subjects

Subjects / Keywords:
Biological transport ( lcsh )
Ferritin ( lcsh )
Voltammetry ( lcsh )
Biological transport ( fast )
Ferritin ( fast )
Voltammetry ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 69-71).
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Science, Department of Chemistry
Statement of Responsibility:
by Robert Joseph Niichel.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
36397575 ( OCLC )
ocm36397575
Classification:
LD1190.L46 1996m .N55 ( lcc )

Full Text
THE ELECTROCHEMISTRY OF FERRITIN
USING SELF-ASSEMBLED MONOLAYERS
ON POLYCRYSTALLINE GOLD ELECTRODES
Robert Joseph Niichel
B.S., Iowa State University, 1985
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Science
Chemistry

1996


1996 by Robert Joseph Niichel
All rights reserved.


This thesis for the Master of Science
degree by
Robert Joseph Niichel
has been approved
by
/Hap 6, me
Date


Niichel, Robert Joseph (M.S., Chemistry)
The Electrochemistry of Ferritin using
Self-Assembled Monolayers
on Polycrystalline Gold Electrodes
Thesis directed by Assistant Professor Donald C. Zapien
ABSTRACT
Cellular ferritin is the primary protein which sequesters excess iron in
biological systems, subsequently preventing damage to cells. Electron
transfer is known to be an important step in the sequestering process.
Electrochemical methods employing modified polycrystalline gold
electrodes were used to investigate the electron transfer properties of
ferritin. Adsorbed layers, formed at controlled electrode potential, of 3-
mercapto-propanoic add (3-MPA), 6-mercapto-hexananoic acid (6-MPA),
and 8-mercapto-octanoic add (8-MOA) were used to investigate the redox
chemistry of ferritin iron. The surface coverage of the adsorbed layers on
the eledrode increased as the chain lengths were increased. Cyclic
voltammetry with these electrodes indicates that ferritin exhibits slow
eledron transfer kinetics at low potentials with fairly well defined current
potential curves. The heterogeneous rate constants were determined
using rotating disc electrode voltammetry for each adsorbed layer. The
rate constants followed a trend of decreasing as the chain length was
increased. Controlled potential eledrolysis measurements using the 3-
MPA monolayer were used to obtain a n-value of 1910 eledrons
in


transferred per ferritin molecule. The induction of iron transport out of
the ferritin core by the electrochemical reduction of ferritin iron was
investigated using EDTA. Cyclic voltammetry of ferritin with
nitrilotriacetate in solution yielded no electrolytic currents at redox
potentials for the iron-nitrilotriacetate complex. Self-assembled
monolayers with varying chain lengths containing carboxylate end groups
effectively promoted the direct electron transfer of ferritin iron at gold
electrodes. These studies demonstrate that the electron transfer
mechanisms of ferritin can be probed electrochemically.
This abstract accurately represents the content of the candidates's thesis. I
recommend its publication.
Signed
Donald C. Zapien
IV


CONTENTS
FIGURES........................................................vii
CHAPTER
1. INTRODUCTION............................................1
Cellular Ferritin: Structure and Function............1
Pathways and Regulation of Cellular Ferritin
and Iron.............................................6
Cellular Ferritin, Iron, and Disease.................9
Horse Spleen Ferritin Structure.....................10
The Nemst Equation and Electrochemistry.............12
Linear Sweep Voltammetry and Cyclic Voltammetry.....13
Controlled Potential Electrolysis...................14
Rotating Disc Electrode Voltammetry.................17
Potential Step Chromaoamperometry...................18
Electron Transfer Mechanisms in
the Sequestering of Iron into Cellular Ferritin.....19
Cyclic voltammetry of Proteins......................21
Research Objectives.................................23
v


2. EXPERIMENTAL METHODS,
.24
Chemicals and Equipment................................24
Experimental Procedures................................28
3. EXPERIMENTAL RESULTS AND DISCUSSION.......................33
Instrumental Analysis of the Cellular
Ferritin Protein.......................................33
Voltammetry of Ferritin at a Clean Gold
Surface................................................33
Modification of the Gold Electrode Surface.............36
Cyclic Voltammagrams of Ferritin.......................38
Concentration Effects on the Electrochemistry
of Ferritin............................................43
Scan Rate Dependence on the Electrochemistry
of Ferritin............................................45
Ferritin Iron Versus Free Iron.........................47
Current Potential of Apoferritin.......................50
Electrochemically Induced Iron Transport
Studies Using EDTA.....................................50
Effects of Varying Mercaptan Chain Lengths on the
Electrochemistry of Ferritin...........................57
4. CONCLUSIONS...............................................67
BIBLIOGRAPHY.................................................69
vi


FIGURES
Figures
1.1 Schematic representation of ferritin....................2
1.2 Schematic representaion radical formaiton
and the Fenton reaction..................................5
1.3 Schematic representation of ferritin stem loop...........8
1.4 Ribbon diagram of the alpha-carbon backbone
of ferritin.............................................11
1.5 Redox reaction of hydroquinone..........................15
1.6 Cyclic Voltammagram of hydroquinone.....................16
3.1 Cyclic Voltammagram of clean gold electrode.............34
3.2 Cyclic Voltammagram of ferritin.........................35
3.3 Cyclic Voltammagram of monolayer of MPA.................37
3.4 Cyclic Voltammagram of ferritin with MPA................39
3.5 Cyclic Voltammagrams of varying cathodic limits
of ferritin.............................................40
3.6 Tafel plot of ferritin..................................42
3.7 Cyclic Voltammagrams of varying concentration
of ferritin.............................................44
3.8 Cyclic Voltammagrams of scan rates of ferritin..........46
3.9 Cyclic Voltammagrams of ferritin
and nitrilotriacetate...................................49
vii


3.10 Cyclic Voltammagrams of apoferritin..............51
3.11 Cyclic Voltammagrams of ferritin and EDTA.............54
3.12 Cyclic Voltammagrams of scan rates ferritin
and EDTA..........................................55
3.13 Cyclic Voltammagrams of ferritin and EDTA........56
3.14 Cyclic Voltammagrams of different purification
stages of ferritin................................58
3.15 Cyclic Voltammagrams of 3-MPA, 6-MHA, 8-MOA......59
3.16 Cyclic Voltammagrams of ferritin on 3-MPA,
6-MHA, and 8-MOA..................................60
3.17 Cyclic Voltammagrams of scan rates, ferritin at 3-MPA.62
3.18 Cyclic Voltammagrams of scan rates, ferritin at 6-MHA.63
3.19 Cyclic Voltammagrams of scan rates, ferritin at 8-MOA.64
viii


CHAPTER 1
INTRODUCTION
Cellular Ferritin: Structure and Function
Cellular ferritin is a protein that is responsible for the storage of
iron in cells of biological systems. Cellular ferritin is a large protein
having a molecular weight of approximately 450,000 Da which is 12 nm in
diameter. Cellular ferritin has the ability to sequester up to approximately
4500 molecules of iron for a total weight of 675,000 Da with the iron
binding to the cellular ferritin protein (Munro et al., 1978). The iron-
ferritin complex is formed as a crystalline core of hydrous ferric oxide
FeO(OH) with variable amounts of attached phosphate (Clegg et al., 1980).
Cellular ferritin contains 24 subunits which are composed of varying
ratios of two distinct types of subunits, heavy (H-type) subunits which
have a molecular mass of 21,099 Da (182 residues) and light (L-type) which
have a mass of 19,766 Da (174 residues) (Theil, 1987). Mammalian cells
encode the H and L-type subunits at separate genes on chromosomes 11
and 19, respectively (Caskey et al., 1983; Worwood et al, 1985). Horse
spleen cellular ferritin is composed mostly of L-type subunits which are
associated in 4:3:2 symmetry (Figure 1.1). This arrangement of the
subunits result in eight hydrophilic and six hydrophobic channels which
link the exterior with the interior of the molecule. Iron passes into the
protein shell through the hydrophilic channels.
1


8 Hydrophilic Channels
6 Hydrophobic Channels
Figure 1.1-Schematic representation of the 24 subunit ferritin
protein shell (Ford, 1984).
2


Iron is the most abundant trace element in the human and animal
body and one of the most abundant in nature. The adult human body
contains approximately 3.5 grams of iron with about 75% traveling in red
blood cells as a component of hemoglobin. A relatively smaller amount is
associated with electron transport and with several enzymatic functions.
The generation of ATP in biological structures relies on iron within the
heme-containing cytochromes and iron-sulfur proteins of electron
transport and oxidative phosphorylation in all cells. The iron containing
enzymes, cytochromes P450 and b$, which are primarily responsible for
drug metabolism in the liver are essential for eliminating toxins from
mammalian systems. Within the nucleus iron is involved in the
synthesis of deoxyribonucleotides for DNA via the enzyme ribonucleotide
reductase. Larger amounts of iron are located in myoglobin cells which
are found in muscle tissue. Highly variable amounts of iron are stored in
the cellular ferritin protein which is present in all cells, especially in the
liver, spleen, heart, and kidneys. Across each organ cellular ferritin has
varying amounts of heavy and light subunits. Iron is also stored in
hemosiderin which is thought to be a breakdown product of cellular
ferritin (Gutteridge et al., 1983).
The main functions of iron involve oxygen transport within blood
and muscle, and electron transfer which is the basis for energy production
and metabolism. Within the mitochondria of each cell, as the Kreb's cycle
produces ATP, iron is involved extensively by being continually reduced
and then oxidized within cytochrome c proteins. However the iron of
cytochrome-c transfers electrons instead of oxygen as in hemoglobin.
The importance of iron for biological systems is evident, however
3


the levels of iron must be carefully regulated since high levels are
extremely toxic (Theil, 1987). Iron can cause major damage as an oxidizing
and reducing agent. During the last step of the oxidative phosphorylation
process a super oxide radical (02') can be generated instead of bonding
with protons to form water. Biological systems can normally breakdown
this radical by the super oxide dismutase (SOD) enzyme into hydrogen
peroxide (H2O2). Subsequently hydrogen peroxide is normally
enzymatically catalyzed into H2O and CO2 by either catalase or glutathione
peroxidase. However in an environment of high iron hydrogen peroxide
can be broken down into a highly toxic peroxide radical (OH ), driven by
the Fenton reaction in which iron is oxidized from the ferrous (Fe2+)
form to the ferric (Fe3+) form (Figure 1.2). At the level of intercellular
interactions the peroxide radical can then readily oxidize
membrane lipids, altering membrane permeability. Additionally
intracellular damage can occur when the peroxide radical attacks the
proteins of cellular organelles and DNA. Cellular ferritin is the protein
which sequesters excess iron and maintains the iron atom in a non-toxic
form.
Cellular ferritin has been shown to be able to bind other toxic
metals. Beryllium binds to nucleic acids and proteins, and inhibits a
number of enzymes. Ferritin has been shown to bind and detoxify up to
1000 atoms of beryllium. Additionally cellular ferritin can bind Cu2+,
Zn2+, Cd2+, Tb3+, Cr^+, or VC)2+ (Price and Joshi, 1983; Treffry
and Harrison, 1982).
4


02
Ft
Figure 1.2-Schematic representation of potential radical formation and the
Fenton reaction (Zane, 1991).
5


Pathways and Regulation of Cellular Ferritin and Iron
Approximately 1% of red blood cells (lifespan 120 days) are degraded
and reformed daily resulting in a turnover of 19-24 mg of hemoglobin
iron per day in the adult. These aged red blood cells are phagocytosed by
reticuloendothelial cells located mainly in the spleen and liver. Iron
released from degraded hemoglobin and porphyrin rapidly appears on
transferrin and in cellular ferritin within minutes of damaged red blood
cell uptake (Siimes and Dallman, 1974). Transferrin transports the iron
back to the bone marrow for resynthesis of hemoglobin while any high-
iron ferritin in the serum is rapidly taken up by the liver hepatocytes
(Mack et al., 1981; Siimes and Dallman, 1974). If necessary, intracellular
ferritin iron may also be mobilized for transport to the bone marrow. For
mobilization, iron within the central FeO(OH)n core of cellular ferritin
must be reduced, chelated, and transferred to the plasma, where it must be
reoxidized to Fe^+ for transport on transferrin. Studies indicate that the
copper-containing plasma protein, ceruloplasmin, may play a role in iron
mobilization. However in free energy terms, it has been calculated that
the "pull" of erythropoiesis in the bone marrow should be sufficient to
cause the flow of iron out of storage and into the bone marrow (May and
Williams, 1977). Iron lost from red blood cells before their uptake by the
reticuloendothelial cells is transported as hemoglobin or porphyrin iron,
on haptoglobin and hemopexin, respectively, and enters liver
parenchymal cells for incorporation into hepatocyte ferritin. The release
of hemoglobin iron, incorporated into ferritin and hemosiderin in the
spleen may have a dependence on the availability of ascorbic acid.
Ascorbate deficiency reduces recycling of red blood cell iron which is
6


associated with an enhanced proportion of iron in spleen hemosiderin.
Furthermore an influx of ascorbate by iron-loaded individuals can result
in severe toxicity due to iron released into the blood (Lipschitz et al., 1988).
Once iron is in a biological system the iron will be carried
throughout the blood via transferrin. When reaching a cell which is
presenting a transferrin receptor, the iron-transferrin complex will bind.
The transferrin-iron complex will be taken into the cell via receptor-
mediated endocytosis. Once inside the cell, within the clathrin coated pit
at a pH of 7.4, a mechanism of pumping protons (H+) into the transferrin-
iron complex will increase the acidity to a level of 4.0 pH which will cause
the iron to be dumped into the cytosol of the cell. The transferrin is now
considered apotransferrin, without the iron binding to the complex and
will travel back to the bloodstream to bind additional iron. Once the iron
is intra-cellular the element can either be incorporated with the iron
containing proteins, e.g. cytochrome c, in the electron transport chain, be
used as an enzymatic co-factor, or if in excess bind with ferritin.
The production of cellular ferritin and the degradation of
transferrin are controlled at the level of translation. The main component
in this system is a binding protein, aconitase (Figure 1.3). During periods
of low iron concentration aconitase binds the 5' untranslated region of
ferritin mRNA inhibiting the translation of ferritin mRNA. Aconitase
binds at the 3' end of transferrin mRNA preventing the endonuclease
degradation of the transferrin mRNA (Klausner et al., 1993). This
situation results in continuous production of transferrin receptors on the
surface of the cell to import additional iron into the cell while cellular
ferritin is not being produced, therefore not sequestering any iron. When
7


FERRITIN mRNA
ACONITASE
AAAAAA
40S AUG
IRE occupied by ACONITASE
Inhibiting translation initiation.
ACONITASE
AAAAAA
Ono or more IREs occupied by ACONITASE
protecting mRNA from rate-determining
stop In mRNA degradation.
;igure 1.3-Schematic representation of the stem loop structure
occurring at the 5' end of ferritin mRNA and the 3' end of
transferrin receptor mRNA (Klausner et al., 1993).
8


iron increases to levels beyond what the cellular metabolism requires
aconitase will release from the mRNA sequences. Subsequently the
translation of ferritin and the degradation of transferrin mRNA will be
initiated. This sequence will not allow additional iron to be transported
into the cell by the transferrin receptors while additional cellular ferritin
will be produced to sequester the excess iron.
Cellular Ferritin, Iron, and Disease
Cellular ferritin is the main line of defense from iron toxicity in
biological systems. If cellular ferritin levels are not sufficient to sequester
excess iron efficiently then various complications may develop. In
infectious diseases, serum iron concentrations and transferrin iron
binding decrease while intestinal iron absorption is reduced. Most serum
iron enters the liver and spleen and (low iron) serum ferritin is released.
In bacterial infections these phenomena reflect an effort on the part of the
biological system to reduce the availability of iron to bacteria which
require iron for their own growth and proliferation (Weinberg et al., 1974).
Increases in serum ferritin concentrations are associated with several
disease processes (primarily infections), liver disease, cancer, and heart
disease. The releases of serum ferritin associated with other diseases may
reflect a leakage of cellular ferritin from damaged cells or a deliberate
secretion of cellular ferritin. Subsequently, disease processes can mask
evidence of iron deficiency by increasing serum ferritin. Increased serum
ferritin levels and high dietary intake of iron can be linked to an increase
in heart disease. The iron associated with serum ferritin may be able to
oxidize low density lipoproteins (LDL-cholesterol) resulting in further
9


proliferation of atherosclerotic plaque (Salonen et al, 1992).
Horse Spleen Ferritin Structure
Horse spleen ferritin is the most characterized ferritin protein with
its primary amino acid structure being sequenced. The composition of
horse spleen ferritin primarily (90%) consists of the L-chain (Theil et al.,
1985). Similar to other ferritin proteins horse spleen ferritin consists of 24
subunits with each subunit being composed of four closely packed major
helices with a shorter minor helix E located adjacent to the primary axis
and opposite the N-terminus (Figure 1.4). The four major helices are held
closely together due to hydrophobic interactions. Research on horse
spleen ferritin indicates that 129 of the 174 subunit residues are located in
the five helices (Heusterspreute and Crichton, 1981). These 24 subunits
form the ferritin protein by packing together in 4:3:2 symmetry. Each
subunit contains many hydrogen and ionic bonds giving a high level of
strength to the tertiary structure. These hydrogen and ionic bonds are also
present between the subunits making the ferritin protein difficult to
denature due to quaternary structure rigidity (Harrison et al., 1985). The
ferritin protein has eight hydrophilic channels which have been proposed
to be the sites where the metal binding capability exists. This indicates that
these are the sites where iron pass in and out of the core of ferritin (Theil,
1987).
10


Figure 1.4-Ribbon diagram of the alpha-carbon backbone of a ferritin
subunit (Ford, 1984).
11


The Nernst Equation and Electrochemistry
The Nemst equation relates the electrode potential to the analyte
concentration (Eqn. 1.1):
E = E' + ln CqVCr*
(1.1)
where (E) is the electrode potential, (E) is the formal potential, (R) is the
molar gas constant (R = 8.31441 J/mol K), (T) is the temperature, (n) is the
number of electrons transferred in a reduction/ oxidation reaction, (F) is
Faraday's constant (F = 9.64846 XI04 C/Equiv.), (C0*) is the bulk
concentration of the oxidized species, and (CR*) is the bulk concentration
of the reduced species.
Based upon this equation the electrode potential (E) will change
throughout the reaction as the bulk concentrations of the reduced and
oxidized species change. Cyclic voltammetry involves the continuous
variation of externally applied potential while measuring the current
response.
The peak current can be expressed in the following equation (Bard
et al., 1980):
where the peak current (ip) is a function of the number of electrons being
transferred during the redox reaction (n), the area of the electrode (A), the
ip = (2.69 X 105) n3/ 2
(1.2)
diffusion coefficient (D0), the rate at which the potential is scanned (v),
12


and the bulk concentration of the reactive species (C0*) and (CR*). For
systems with fast electron transfer kinetics (large k) the potential at which
currents flow are governed only by the ratio of reduced/oxidized species
concentrations; these systems follow the Nernst equation. Other systems
which exhibit slow electron transfer kinetics (small k) can be treated as
reactions with large activation energy barriers, requiring a higher
(potential) to overcome the barrier in order for the reaction to proceed
(electron flow). The linear scan voltammagrams of these systems are
characterized by not having currents flowing until large potentials are
applied away from equilibrium. Subsequently the cyclic voltammagrams
of systems with small k values exhibit large peak separations and systems
with exceedingly small rate constants exhibit no current.
Linear Sweep Voltammetry and Cyclic Voltammetry
Potential is applied between a working electrode and an auxiliary
electrode. The current which is due to redox reactions is measured
between these two electrodes. The potential of the working electrode is
controlled at a value or values with respect to the potential of a reference
electrode. In linear sweep voltammetry the potential at the working
electrode is ramped uniformly with time. Cyclic voltammetry is a linear
sweep voltammagram where the sweep direction is changed at designated
limits. The potential will reach a level which will induce dissolved
species to liberate electrons (oxidation) or take up electrons (reduction).
Electrons passed per unit time from a given reaction constitutes a current.
Therefore the higher the rate of the redox reaction the higher the current.
Hydroquinone (HQ), which exhibits fast electron transfer kinetics,
13


can be used as an example to demonstrate an electrochemically reversible
system. Hydroquinone, 1,4-dihydroxybenzene, which represents the
reduced form of this redox couple and the oxidized form is called
benzoquinone (BQ). The reduction of (BQ) involves the removal of two
electrons and two protons (Figure 1.5). Upon viewing the cyclic
voltammagram of hydroquinone it can be seen that at the applied
potential of 0.35 V the current rises rapidly until it reaches a peak (Figure
1.6). This rapid rise is due to the increasing rate at which HQ is oxidized to
BQ as the potential is scanned anodically. The rate of this redox reaction
reaches a maximum at 0.48 V. At this potential the rate of the reaction is
controlled by the rate at which hydroquinone diffuses to the electrode
surface. When the potential is scanned cathodically the rapid rise in
current is the reduction of BQ to HQ. The current-potential relationship is
mathematically described by the following equation (Bard et al., 1980) (Eqn.
1.3).
= FAk[C(0,t)e ^(E Eo ) CR(0,t)e'
(l-a)nf(E-E')
] (Eqn. 1.3)
For an electrochemically reversible system, a system which exhibits fast
electron transfer kinetics, this equation reduces to the Nemst equation.
Controlled Potential Electrolysis
This procedure involves an electrode being immersed into a
solution of known concentration and known volume. The potential of
the electrode is poised at a value at which electrolysis (oxidation or
reduction) is known to occur. The solution is stirred in order to effect
14


Figure 1.5-The redox reaction of hydroquinone, an electrochemically
reversible system
15


0.2
0.4
o.s
0.8
POTENTIAL (Volts vs Ag/AgO)
Figure 1.6-Cyclic voltammagram of hydroquinone in 1 M H2S04/
hydroquinone concentration 0.01 M, scan rate 20 mV/ sec,
2
electrode area 2.15 cm
16


efficient mass transport of the sample species to the electrode surface.
Using the following equation the n-value can be determined using the
Faraday relationship (Bard et al., 1980) (Eqn. 1.4)
n = Qo/FVC0* (Eqn. 1.4)
Where (n) is the number of electrons transferred, (QJ is the total
integrated charge, (F) is the Faraday constant (98,465), (V) is the volume,
and CD* is the concentration of the oxidized species. The number of
equivalents of electrons transferred per mole of the sample molecule (n) is
calculated when the reduction charge is integrated.
Rotating Disc Electrode Voltammetry
In an experiment using cyclic voltammetry the rate of electron
transfer depends on: a.) the rate at which diffusion brings the electroactive
species to the electrode surface (dependence on D0) and b.) the rate at
which electrons are transferred once they reach the electrode (dependence
on k). To measure k the mass transport limitations imposed by diffusion
must be eliminated by mechanically inducing convection, thus making
mass transport very fast compared with electron transfer. Ordinarily rapid
stirring is sufficient unless the rotating magnet induces extraneous
currents. A rotating disc electrode powered by a DC motor eliminates the
noise problem. As the solution is agitated the current potential curve is
scanned. The net current is measured at different potentials prior to the
plateau and log (i) is plotted vs. (r]) a Tafel plot. Where (i) is the current
and (ti) is the overpotential which is calculated by subtracting the
17


equilibrium potential from the applied potential (Emeas E^). The
equilibrium potential is estimated by the mid-point potential of the
current-potential curve of ferritin at a 3-MPA modified surface. The Tafel
equation (Eqn. 1.5) can be used to determine the exchange current iD for
systems which exhibit slow electron transfer kinetics (Bard et al., 1980).
n = a + b log(i) (Eqn. 1.5)
The exchange current is the y-intercept from the Tafel plot and the
heterogeneous rate constant is calculated using i0 log extracted from the
following equation (Bard et alv 1980) (Eqn. 1.6):
When the potential of the electrode is poised at a value where no
electron transfer of the system occurs, then stepped to a value where the
rate of electron transfer is high, the measured current will decay with a
iD = nFAk0C*
(Eqn. 1.6)
The other terms have their usual meanings.
Potential Step Chromoamperometry
1/2
t dependence as given by the Cottrell equation (Bard et al., 1980) (Eqn.
1.7).
(Eqn. 1.7)
Where i(t) is the current at time (t), n is the number of electrons
18


transferred per sample, A is the area of the electrode, F is Faradays
constant, D0 is the diffusion coefficient, and C0 is the concentration of the
oxidized species. From the Cottrell equation the diffusion coefficient can
be determined.
The Stokes-Einstein equation (Eqn. 1.8) allows the approximation of
the diffusion coefficient specifically for species moving through a liquid
medium.
kT
DoF--------- (Eqn. 1.8)
6:rcr)srF
Where ferritin molecules F are spherical with a radius rF, and are moving
through solvent S, is the viscosity of the solvent (in poise), k is
Boltzmanns constant, and T is the temperature. This equation is
formulated for cases in which rs -7
ferritin molecule is very close to being spherical with a radius of 6 X 10
cm.
Electron Transfer Mechanisms in Sequestering of Ferritin Iron
Ferritin is not considered a redox protein, however previous studies
indicate an electron transfer step is involved when iron is passed in and
out of the protein (Chasteen et al., 1991). Evaluating the electron transfer
is fundamental to understanding the mechanism by which iron is
sequestered by cellular ferritin. Ferritin has been reduced
electrochemically in micro-coulometric studies in which viologen
derivatives were used as electrochemical mediators. This redox reaction
19


of iron within ferritin has been observed as having one electron
transferred per iron atom (Watt et alv 1979). Studies using mediated
coulometry aided by ESR, have suggested that ferritin may have
electrochemical active sites which are possibly involved in the redox
activity of ferritin (Watt et alv 1992). Iron clusters such as undecairon (III)
oxo-hydroxo aggregates have been synthesized in an effort to model the
polyiron core of ferritin. These aggregates have been shown to be
electrochemically reversible through a one-electron transfer for every iron
atom at very negative equilibrium potentials (Goran et alv 1987).
In experiments using UV-Visible spectrophotometry, assisted by
EPR spectrometry, the oxidizing or reducing agents did not have to
interact directly with the mineral core of ferritin for the iron to be oxidized
or reduced. This UV-Visible spectrophotometry was used to determine
that long range electron transfer through the channels of ferritin is
probably what is instrumental in the rapid rate of the redox reaction
between reduced iron and ferrihemoproteins (Kadir et al., 1991). This
electron transfer to the ferritin core occurs through mediation of the
channel-bound Fe2+ or Fe3+ ions, or by electron tunneling through the
protein shell (Watt et al., 1988). In additional studies Mossbauer
spectroscopy has been used to indicate that phosphate on the mineral iron
core may be an important component in catalyzing the internal electron
transfer reaction from Fe2+ to Fe3+ (Jacobs et al., 1989). Highly critical
residues on the H-chain subunits may be responsible for the binding and
the formation of an Fe^ +-oxo-bridged dimer. The Fe^+ dimer formation
could be the product of the oxidation of two Fe^+ atoms by one oxygen
molecule, catalyzed by ferrioxidase activity (Treffry, 1992).
20


Cyclic Voltammetry of Proteins
Cyclic Voltammetry is a very effective approach in the study of
electron transfer kinetics of electroactive species. The kinetics and
thermodynamics of a system can be analyzed using peak potential
separation and scan rate dependencies. These processes can be used to
determine if the electron transfer step is fast compared with mass
transport, or if chemical reactions are coupled to the electron transfer step.
Proteins usually adsorb strongly onto bare metal electrode surfaces,
however this irreversible adsorption by redox proteins has often inhibited
electron transfer with proteins. The electron transfer step of certain
proteins can be catalyzed when the bare metal electrode is modified with a
monolayer of various adsorbed species. The formation of a well-ordered,
single molecular layer of an adsorbed species has been developed by
placing a clean gold electrode into a solution containing the promoter.
Promoters which bind to the bare metal electrode through a sulfur atom
allow for a variety of terminal functional groups to be oriented away from
the electrode surface and towards the protein in solution. Typically when
proteins exhibit direct electron transfer at the electrode surface adsorption
by the protein is involved to some degree. These studies suggest that
unfolding of the protein at the electrode surface prevents or inhibits
electron transfer with the electrode. It is possible that the adsorbed layer
acts to reduce the extent of unfolding of the protein (Sagara et al., 1991).
Due to the hydrophilic characteristics of certain bare metal electrodes the
direct electron transfer of certain heme proteins has been promoted. The
majority of adsorbed cytochrome-c is electroactive and adsorbs strongly to
21


tin oxide electrodes (Collinson et al., 1992). Other studies have shown that
while using indium oxide electrodes for analyzing cytochrome-c,
oligomeric, and deamidated forms of cytochrome c inhibit the
electrochemical reactions of certain purified proteins in solution.
Additionally indium oxide electrodes have been shown to be effective in
promoting the direct electron transfer of certain myoglobins (Taniguchi et
al., 1992). Horseradish peroxidase, a heme-containing enzyme, adsorbs
onto pyrolytic graphite and exhibits direct electron transfer (Tominaga et
al., 1993).
Thiols with carboxylate end groups have been used effectively in
promoting direct electron transfer in proteins (Taniguchi et al, 1982).
Straight chain thiol acids longer than five carbons assemble into highly
organized monolayers when adsorbed onto a bare gold electrode. In
contrast, short chain systems, such as 3-mercaptopropanoic acid (3-MPA),
have exhibited greater disorder than the longer chain systems (Malem et
al., 1993). Molecules consisting of a divalent sulfur moiety (sulfide or
thiol), spontaneously and irreversibly bond covalently to the gold surface
through the divalent sulfur atom. The bonding of 3-MPA onto gold
through the sulfhydryl moiety has been analyzed using infrared reflection-
absorption. Additionally secondary ion mass spectrometry (SIMS) has
been used to probe the structure and bonding of a 3-MPA monolayer
indicating the presence of covalent bonds between the sulfur atoms and
the gold atoms of the electrode surface (Leggett et al., 1993).
22


Research Objectives
Investigating the role of electron transfer is instrumental in
understanding the mechanism of how iron is sequestered by cellular
ferritin. To investigate this mechanism a method is required which can
probe the kinetics and thermodynamics of the electron transfer reaction
associated with the sequestering of iron by ferritin. The primary goal in
this work is to determine an effective promoter for the redox reaction of
cellular ferritin. Monolayers of short and long chain thiol adds were
adsorbed on clean gold electrodes which were used in cyclic voltammetric
experiments of solutions containing horse spleen ferritin. These layers
have been found to promote the direct electron transfer of ferritin iron
with the gold electrode. By varying the length of the adsorbed mercaptan
layers experiments were conducted to investigate whether chain length is
critical in the catalytic activity of the promoter layer. Form the measured
redox potentials the stability of the chelated iron core can be characterized.
In addition the current-potential behavior can be used to characterize the
electron transfer kinetics of cellular ferritin.
23


CHAPTER 2
EXPERIMENTAL METHODS
Chemicals and Equipment
The following compounds were purchased from Sigma Chemical
Company (St. Louis, MO): 3-mercaptopropanoic acid (3-MPA) (99.3%),
horse spleen apoferritin, horse spleen ferritin: type I, Bovine albumin
(fraction 5 powder), disodium hydrogen phosphate (Reagent Grade),
phosphoric acid (Reagent Grade), hydrochloric acid (Reagent Grade),
ethylenediaminetetraacetic acid (EDTA) (98.5%), ferrous sulfate
heptahydrate (99%), phenylmethylsulfonylfluoride (PMSF) (>99%), and
sodium dichromate (98%). Gold wire and gold foil (99.99%) were
purchased from Alfa-Johnson Matthey (Ward Hill, MA). Aldrich
Chemical company (Milwaukee, WI) was the commercial supplier of 6-
bromohexanoic acid (98%), 8-bromooctanoic add (97%), sodium
hydrosulfide hydrate (75% as hydrate), nitrilotriacetate (99%), and
octanethiol (>97%). The sulfuric acid (analytical reagent grade), nitric acid
(analytical reagent grade), and sodium hydroxide (analytical reagent grade)
were obtained from Baxter Sdentific Co., Inc. (McGraw Park, IL). The
commerdal supplier of sodium chloride (99.2%) and sodium azide (>99%)
was Mallinckrodt Inc. (Paris, KY). Tris(hydroxymethyl) aminomethane
(electrophoresis purity, 99.8%) was obtained through Bio-Rad Laboratory
(Hercules, CA).
The water was purified by distilling de-ionized water vapor through
a heated platinum catalyst in the presence of oxygen, subsequently distilled
24


again. Pyrolytically distilled water (PDW) was used in the preparation of
all solutions and in cleaning procedures.
The adsorbate compounds 6-mercaptohexanoic acid (6-MHA) and 8-
mercaptooctanoic acid (8-MOA) were synthesized according to published
procedures (Ivanovics et al., 1944). The 6-MHA was purified by vacuum
distillation. The (8-MOA) was purified by the following procedure: The
product mixture was dissolved in methanol, then gaseous HC1 bubbled
into the solution for approximately ten minutes. The mixture was
refluxed for two hours. After the purity of the methyl ester product was
determined by TLC it was refluxed in concentrated aqueous HC1 for one
hour. Upon cooling the acidic solution yielded 8-mercaptooctanoic add.
The pH 7 phosphate buffer was prepared by mixing 50 mL 0.1 M
disodium hydrogen phosphate with 29.1 mL 0.1 M sodium hydroxide then
titrating to pH 7 with concentrated phosphoric add. The pH 7 tris buffer
was prepared by mixing 50 mL 0.1 M tris (hydroxymethyl) aminomethane
with 46.6 mL of 0.1 M hydrochloric acid then titrating to pH 7 with
concentrated hydrochloric add (Handbook of Chemistry and Physics, 1995).
Prior to any analysis the excess free iron was removed from the
commerdally available horse spleen ferritin and apoferritin by eluting the
protein sample through a Sephadex G 25-M column (Pharmada) using
primarily pH 7 phosphate buffer. However for the atomic absorption
analysis PDW was used as the eluant, and tris buffer was used as the
eluant for the nitrilotriacetate analysis. The column specifications: 9.1 mL
swollen Sephadex G-25 media, bed height-5 cm, diameter-1.8 cm, initial
equilibration volume-25 mL, sample volume-2.5 mL, eluant volume-3.5
mL. A gravity column (25 cm X 2.5 cm) containing G-200 Sephadex
25


(protein fraction range, 5,000-600,000 Da) and G-200 buffer (20 mM pH 7.0
phosphate buffer, 0.9% NaCl, 0.2 mM PMSF, and 0.05% NaN3) was used to
purify the ferritin protein. A series of standards (cytochrome-c; 12,400 Da,
hemoglobin; 64,000 Da, and horses spleen ferritin; 550,000 Da) were applied
to the column using G-200 buffer as the eluant. The elution fractions were
collected in 1 mL increments and analyzed at 410 nm. A calibration curve
was constructed by plotting log (MW) vs. retention volume of the
standard. A 2 mL sample of horse spleen ferritin at 100 mg/mL was
applied to the column and eluted with G-200 buffer. The ferritin was
eluted in 1 mL increments, with the fractions being analyzed at 410 nm.
The corresponding ferritin fractions were combined with the final
concentration of ferritin being determined by projecting the absorbance of
the fractions at 280 nm onto the concentration axis of a bovine albumin
calibration curve. The post G-200 Sephadex column ferritin was eluted
through the G-25 Sephadex column to remove any free iron from the
sample.
The electrodes and all glassware were soaked in chromic acid for at
least four hours prior to use in all experiments. The chromic acid cleaning
solution was prepared by dissolving 92 grams of sodium dichromate into
458 mL of water followed by the addition of 800 mL of sulfuric add while
stirring. Following the cleaning procedure in chromic add the glassware
and electrodes were rinsed with PDW.
The electrochemical cells that were used in all cyclic voltammetry
experiments were of an H-cell configuration. A side chamber that
contained the reference (silver/ silver chloride) and auxiliary (platinum)
eledrodes, and the main chamber which contained the electrochemical
26


solution under study. The two chambers were separated by a 1 cm sintered
glass fritted disc which allowed ions to flow between the compartments
while keeping contaminants from the side compartment from entering
the main chamber.
The metal electrodes consisted of 0.1 mm thick gold foil fused to
gold wire 0.1 mm in diameter. The electrode was inserted into an 8.0 mm
O.D. glass sheath connected to a nitrogen source. By applying a constant
positive pressure of nitrogen through the glass sheath an inert
environment is produced which will prevent any oxidation of the
electrode surface during transfer from one cell to another. While in
solution the flow of nitrogen expels any solution from the electrode. Once
the flow of nitrogen is stopped fresh solution will flow into contact with
the electrode cavity.
Voltammetric scans were performed with a Cypress Model Omni 90
potentiostat (Lawrence, KS) and a BioAnalytical Systems Model-R XY
recorder (West Lafayette, IN). The determination of the n-value used a
laboratoy built coulometer based on operational amplifiers (assembled by
DCZ). A Cary IE UV-Visible Spectrophotometer (Varian Analytical
Instruments, Palo Alto, CA) was used to measure the concentration of
ferritin following size-exclusion chromatography. Atomic absorption
measurements used an AA-575 Atomic Absorption Spectrophotometer
(Varian Analytical Instruments, Palo Alto, CA).
27


Experimental Procedures
The concentration of ferritin was determined using UV-VIS
spectroscopy (lambda max=280 nm). The ferritin was initially eluted
through the G-25 Sephadex column to remove any excess ions. The
ferritin was prepared for analysis in pH 7 phosphate buffer. Bovine
albumin was used as the standard within the range of 20-200 mg / mL. For
the studies involving the comparison of 3-MPA, 6-MHA, and 8-MOA the
ferritin was further purified by size exclusion chromatography. The
protein was eluted through the gravity column, containing G-200
Sephadex, with G-200 buffer, then eluded through the G-25 Sephadex
column. Bovine albumin was used as the standard within the range of
0.5-10.0 mg/mL. The calibration curves were linear within the range of
standards.
The number of electrons transferred per ferritin molecule (n-value)
was determined by controlled potential electrolysis. A 3-MPA modified
gold electrode was immersed in a solution containing 2.0 X 1010 M
ferritin, pH 7 phosphate buffer, and 1 M Na2SC>4 at an applied potential of
0.00 V. With stirring the potential of the electrode was stepped to -0.32 V.
The reductive charge was integrated until no appreciable charge was
measured. The background charge was integrated using a 3-MPA
modified gold electrode immersed in a solution containing only the buffer
and electrolyte.
The average number of iron atoms per ferritin molecule was
determined using flame atomic absorption spectroscopy. The iron
standards were prepared in a range of 1.0 to 10.0 ppm using ferrous sulfate
heptahydrate in 2 M nitric acid. The measurements were conducted at
28


248.3 nm with 20 mA of applied current to the iron hollow cathode lamp.
The calibration curve was linear within the range of standards.
The heterogeneous rate constant was determined by mixing solid
sodium hydrosulfite with an equivalent amount of ferritin. The solution
of reduced ferritin was added to an equimolar quantity of oxidized ferritin
which resulted in equal concentrations of both oxidized and reduced
forms. A layer of 3-MPA was formed at the gold electrode, then immersed
in ferritin solution at 0.00 V. Under vigorous stirring the potential was
scanned from 0.00 V to -0.50 V. Using the mid-peak potential as an
estimate for the equilibrium potential, and assuming a transfer coefficient
of 0.5, a plot of log (i) vs. overpotential was constructed from which the
heterogeneous rate constant was determined. The same procedure was
employed when 6-MHA and 8-MOA were used as promoters.
The gold electrodes were prepared by initial annealing with a
natural gas/ air flame. Subsequently the electrode was cleaned with
sulfuric acid in the electrochemical cell by cycling the potential between 1.5
and -0.35 V at one minute intervals. After cycling three times a clean
electrode surface was confirmed by comparing its cyclic voltammagram
curve with that of a known clean gold electrode. In order to rinse the
electrode free of dissolved adsorbate compound the electrode cavity is
flushed 10 times in three second intervals. Hydroquinone, was used to
demonstrate the effectiveness of this rinsing method. A cyclic
voltammagram was obtained by immersing a clean gold electrode in a
solution of 0.010 M hydroquinone, and 1 M H2SO4, then scanning
anodically from 0.20 V to 0.70 V, then cathodically to 0.20 V. Next a clean
electrode was immersed in a new solution of 0.010 M hydroquinone, and 1
29


M H2SO4. Subsequently the electrode was transferred under nitrogen to
another cell containing pure electrolyte where the electrode cavity was
flushed 10 times. The rinsed electrode was then transferred under
nitrogen to an electrochemical cell containing a solution of 1 M H2SO4. A
cyclic voltammagram was obtained by scanning anodically from 0.20 V to
0.70 V, then cathodically to 0.20 V.
The area of the electrode was determined by initially obtaining a
clean gold electrode cyclic voltammagram. The clean gold electrode was
rinsed free of dissolved iodide then transferred to an electrochemical cell
containing a solution of sulfuric add with 0.010 M potassium iodide.
After a potential of 0.00 V was applied for five minutes the gold electrode
was transferred to another electrochemical cell containing sulfuric acid. A
cyclic voltammagram of the electrode with the adsorbed layer of iodine
was scanned within the range of -0.35 to 1.5 V. The difference between the
oxidation of adsorbed iodine and the surface oxide of the dean electrode
was used to determine the electrode area by obtaining the number of
moles of adsorbed iodine using the Faraday relationship.
The promoter layers were adsorbed to the dean electrode surface by
initially transferring the electrode to an eledrochemical cell containing a
solution of the adsorbate. The adsorbate was prepared in an eledrolyte
solution of pH 7 phosphate buffer and 1 M Na2SC>4. Once in the adsorbate
solution a potential of 0.00 V was applied for five minutes. During the
five minute interval the nitrogen flushing procedure was implemented at
two and four minutes to allow fresh solution to come in contad with the
electrode. Finally the electrode was removed under nitrogen, rinsed with
PDW, then transferred to another electrochemical cell containing fresh
30


electrolyte solution where further rinsing was carried out. The electrode
was then transferred under nitrogen to another electrochemical cell
containing fresh electrolyte solution where a cyclic voltammagram was
obtained of the adsorbed layer.
The coverage of the mercaptans on the gold electrode was measured
by determining the surface excess to deposited copper (Malem et al., 1993).
Initially a clean gold electrode was immersed into a solution of 1.0 X 10_5
M copper sulfate in 1 M H2SO4 with an applied potential of 0.4 V. Then
the potential was stepped to -0.1 V for 10 minutes while stirring. By
scanning anodically to 0.4 V a cyclic voltammagram was obtained from
which the copper oxidation UPD peak, occurring at 0.18 V, was integrated
by cutting and weighing. The electrode was cleaned, modified with a new
layer of mercaptan, transferred into the copper solution, and finally an
anodic stripping voltammagram was scanned. The difference in peak area
between the two cyclic voltammagrams was proportional to the surface
coverage of the mercaptan on the gold electrode.
The cyclic voltammagrams of proteins using a gold electrode
modified by adsorbed thiols were obtained by transferring the modified
gold electrode under nitrogen to an electrochemical cell containing a
solution of known protein concentration with pH 7 phosphate buffer and
1 M Na2SC>4. The electrode was exposed to fresh protein solution after
three minutes via flushing the electrode cavity. Generally, for the ferritin
protein cyclic voltammagrams the initial potential was 0.00 V. The
potential was scanned cathodically to -0.5 V, then anodically to 0.00 V.
The absence of any free iron in the ferritin samples was determined
voltammetrically. To determine the voltammetric peak morphology and
31


peak potentials of the iron-nitrilotriacetate complex a cyclic
voltammagram was obtained by immersing a clean gold electrode in a
solution of tris pH 7 buffer, 1 M Na2SC>4, 0.010 M nitrilotriacetate, and
0.001 M ferrous sulfate then scanning anodically from 0.00 V to 0.35 V
then cathodically to 0.00 V. Subsequently another cyclic voltammagram
was obtained by immersing a clean gold electrode in a solution of tris pH 7
buffer, 1 M Na2SC>4, 0.010 M nitrilotriacetate, and 4.9 mg/mL ferritin then
scanning using the same parameters.
Electrochemical studies of ferritin were performed using EDTA as a
complexing agent. Initially a clean electrode was modified with a 3-MPA
layer then immersed in a solution of pH 7 buffer, 1 M Na2S04, 0.01 M
EDTA, and 4.9 mg/mL ferritin. A cyclic voltammagram was obtained by
scanning cathodically from 0.00 V to -0.50 V, then anodically to 0.00 V. To
amplify the signal from the resulting EDTA-iron complex, the potential of
a 3-MPA modified gold was poised at -0.32 V with stirring for 120 minutes.
Then the cyclic voltammagram of the iron-EDTA complex was obtained by
stepping the potential to 0.00 V, scanning cathodically to -0.50 V and then
anodically to 0.30 V.
32


CHAPTER 3
RESULTS AND DISCUSSION
Instrumental Analysis of the Cellular Ferritin Protein
Using UV-Visible spectroscopy the concentration of post-Sephadex
G-25 column horse spleen ferritin was determined. The bovine albumin
protein standard curve was linear through the region of 20-200 mg/mL. A
value of 97.8 mg/mL was obtained for the commercially available ferritin
protein concentration.
Using UV-Visible spectroscopy in a similar fashion the
concentration of the post G-200 Sephadex column and post G-25 Sephadex
ferritin was determined. In this case the bovine albumin protein standard
curve was linear through the region of 0.5-10.0 mg/mL.
Voltammetry of Ferritin at a Clean Gold Surface
The cyclic voltammagram for the bare polycrystalline gold electrode
indicated the oxidation of water, the reduction of H+, and the UPD peaks
of the surface oxide; additional peaks were present confirming that the
electrode is clean (Figure 3.1). The cyclic voltammagram of ferritin at a
bare gold electrode indicated only a trace of electroactivity (Figure 3.2). It is
unclear from this data why ferritin does not exhibit electroactivity at a bare
gold electrode surface. Upon modifying the bare gold electrode with
certain adsorbates the electroactivity of ferritin was enhanced. The cyclic
voltammagrams indicated substantial cathodic and anodic currents due to
the redox reactions involving iron within the ferritin inner core.
33


Volts vs Ag/AgCl)
Figure 3.1-Cyclic voltammagram of a clean poly crystalline gold
electrode in 1M H2S04, electrode area 2.15 crn^, scan rate 100
mV/sec.
34


CURRENT (Microamps)
Figure 3.2-Cyclic voltammagram of horse spleen ferritin in pH 7
phosphate buffer, ferritin concentration 4.9 mg/mL, electrode
area 2.15 cm^, scan rate 100 mV/sec.
35


Modification of the Gold Electrode Surface
The electrode surface modifiers used in these studies for promoting
the direct electron transfer of ferritin were carboxylate terminated aliphatic
mercaptans of various chain lengths (C3, Cg, Cg). Initially the surface of
the bare gold was modified by immersing the electrode into a solution of
3-MPA and pH 7 phosphate buffer. A potential of 0.00 V was used to form
an adsorbed monolayer on the electrode surface. The electrode was
removed from the solution after five minutes under nitrogen and rinsed
in buffer using the nitrogen flushing process. The electrode with the
newly formed adsorbed monolayer was then transferred to another
electrochemical cell. The cyclic voltammagram was obtained by
immersing the modified electrode into the pH 7 phosphate buffer initially
at 0.00 V, scanning cathodically to -0.50V and anodically to 0.00 V (Figure
3.3). The cyclic voltammagram in this potential range is relatively
featureless. This cyclic voltammagram served as the background for the
subsequent cyclic voltammagrams of ferritin.
The surface coverage of the mercaptans on the gold electrode was
measured by determining the surface excess to deposited copper. The
surface coverage was determined with the following results: 3-MPA; 0.27,
6-MHA; 0.57, 8-MOA; 0.66.
36


-0.6
-0.
-0.2
0.0
POTENTIAL (Volts vs AO/AqCI)
Figure 3.3-Cyclic voltammagram of monolayer of 3-MPA adsorbed on
polycrystalline gold electrode in pH 7 buffer, 3-MPA
concentration 0.001 M, electrode area 2.15 cm^, scan rate 100
mV/sec.
37


Cyclic Voltammagrams of Ferritin
A cyclic voltanunagram of ferritin was obtained using a 3-MPA
modified electrode. The modified electrode was immersed into a solution
of pH 7 buffer and 4.9 mg/mL ferritin initially at 0.00V, scanning
cathodically to -0.50 V and anodically to 0.00 V (Figure 3.4). This cyclic
voltammagram of ferritin indicates that the monolayer of 3-MPA
promotes the direct electron transfer of ferritin at the gold electrode. The
cyclic voltammagram indicates a cathodic wave at -0.25 V, with the anodic
wave at -0.12 V, resulting a mid-peak potential of -0.19 V, and a peak
potential difference of 130 mV at 100 mV/s.
The sharp current increase beginning at about -0.37 V is the bulk
reduction of the hydrogen ion. To confirm that the bulk reduction of
hydrogen does not interfere with the ferritin analysis the cathodic limit
switching potential was varied from -0.4 V to -0.25 V (Figure 3.5). These
cyclic voltammagrams show that the anodic branch does not include a
significant amount of the current associated with the re-oxidation of
hydrogen. The anodic current is indeed related to the cathodic current, in
that they represent the anodic and cathodic branches of a common redox
couple.
Using controlled potential electrolysis to reduce the inner ferritin
iron, at an 3-MPA modified gold electrode, the n-value was obtained of
1910+190 electrons transferred per ferritin molecule. Atomic absorption
was used to determine the iron content in ferritin; the value of 2070+247
iron atoms per ferritin molecule was obtained. From this information it
was determined that one electron was transferred per iron atom. This
result is consistent with previously reported data using electrochemical
38


(Microamps)
Figure 3.4-Cyclic voltammagram of horse spleen ferritin at a
polycrystalline gold electrode with 3-MPA layer adsorbed in pH
7 buffer, 3-MPA concentration; 0.001 M, ferritin concentration
4.9 mg/ mL, electrode area 2.15 cm^, scan rate 100 mV/sec.
39


CURRENT (Microamps)
POTENTIAL (Volts vs Ag/AgCl) POTENTIAL (Volts vs Ag/AgCI)
-0.4 -0.2 0.0 0.4 -0.2 0.0
POTENTIAL (Volts vs Ag/AgCl) POTENTIAL (Volts vs Ag/AgCI)
Figure 3.5-Cyclic voltammagram of horse spleen ferritin at a
polycrystalline gold electrode with 3-MPA layer adsorbed in pH
7 buffer with varying cathodic limits (a) -0.40 V, (b) -0.35 V,
(c) -0.30 V, (d) -0.25 V 3-MPA concentration; 0.001 M, ferritin
concentration 4.9 mg/mL, electrode area 2.15 cm^, scan rate 100
mV/sec.
40


mediators (Watt et al., 1985). Using controlled potential electrolysis to
reduce the inner ferritin iron, at an 3-MPA modified gold electrode, the n-
value was obtained of 1910+190 electrons transferred per ferritin molecule.
Atomic absorption was used to determine the iron content in ferritin; the
value of 2070+247 iron atoms per ferritin molecule was obtained. From
this information it was determined that one electron was transferred per
iron atom. This result is consistent with previously reported data using
electrochemical mediators (Watt et al., 1985).
Using a ferritin concentration of 1.15 mg/mL (2.55 x 10 M/cm ) the
heterogeneous rate constants were obtained from Tafel plots (Figure 3.6)
Assuming a transfer coefficient of 0.5, the rate constants for ferritin with
the adsorbed layers were determined with the following results: 3-MPA;
1.24 x 10'6, 6-MHA; 3.37 x 10'7, 8-MOA; 2.63 xlO'7.
41


Log ip
Chartl
FERRITIN TAFEL PLOT 3-MPA
Figure 3.6-Tafel plot of horse spleen ferritin at a
polycrystalline gold electrode with 3-MPA layer adsorbed in pH
7 buffer, 3-MPA concentration; 0.001 M, ferritin concentration
4.9 mg/mL, electrode area 2.15 cm^, scan rate 100 mV/ sec.
42


Concentration Effects on the Electrochemistry of Ferritin
The cyclic voltammagrams of ferritin at concentrations of 4.9, 1.0,
and 0.49 mg/ mL using a 3-MPA gold modified electrode showed
corresponding peak currents (Figure 3.7). The concentration of 0.49
mg/ mL was the lowest level of ferritin where the peak potentials could be
measured. Using a scan rate of 100 mV/ s at each concentration the
cathodic peak potentials remained constant at -0.25 V, and the anodic peak
constant at -0.12 V.
Typically low protein concentrations are used in electrochemical
studies since in the low concentration range, explanations of
protein/ electrode interactions are simplified. The molecular weight of
horse spleen ferritin is approximately 450,000 Da. The concentration used
extensively in this work, 4.9 mg/mL, is equivalent to 1.1 X 10 mol / cm .
Potential step chromoamperometric measurements yielded a diffusion
coefficient of 5.7 X 10 cm /sec for ferritin. This measured result is
expected to be small in tight of the fact that ferritin is such a large
molecule. The magnitude of this value is three orders of magnitude
smaller than reported for horse heart myoglobin which is a protein 30
times smaller in mass (Taniguchi, 1992). Given this data, an n-value of
2
1910, an electrode area of 2.15 cm a scan rate of 100 mV / sec, and the
equation for an irreversible process (Eqn. 3.1),
ip = (2.99 X 105)n(ana)1' ' V''2 Eqn. 3.1
a peak current of 903 uA is calculated form theory (Bard et al., 1980). This
calculated value is two orders of magnitude larger than the experimental
43


(Mictotfflpt)
POTENTIAL (Vow A^A^O)
Figure 3.7-Cyclic voltammagrams of horse spleen ferritin at a
polyciystalline gold electrode with 3-MPA layer
adsorbed in pH 7 buffer with varying concentrations
(a) 0.49 mg/mL (b) 1.0 mg/mL (c) 4.9 mg/mL.
44


value of 7.7 uA under the same set of conditions. For a peak current of 7.7
-15
uA the diffusion coefficient would have to be on the order of 10
2,
cm /sec
Using the Stokes-Einstein equation, which allows for an
approximation of the diffusion coefficient for species moving through a
liquid medium, the calculated diffusion coefficient is 4.1 X 10" cm/sec.
The viscosity of the solvent has been estimated at 0.0089 poise which is the
viscosity of water at 25C. When this value is substituted into the peak
current equation the calculated peak current is 0.24 uA, which is much
closer to the results obtained experimentally.
Scan Rate Dependence on the Electrochemistry of Ferritin
The scan rate of the ferritin cyclic voltammagrams was from 50-400
mV/s using the 3-MPA modified gold electrode (Figure 3.8). The peak
potential difference increased with scan rate, varying within the range of
125-225 mV indicating that the electron transfer kinetics are relatively
slow which is consistent with the rate constant data. The cathodic peak
currents increased correspondingly with scan rate following an
approximate (scan rate) 1/2. The cathodic peak current/anodic peak
current ( ipc/ipa) ratio becomes larger as the scan rate is increased
indicating that adsorption by ferritin (oxidized form) is occurring. Both
diffusing ferritin (ox) and adsorbed ferritin (ox) are contributing to the
cathodic current.
To confirm whether ferritin adsorbs irreversibly to the 3-MPA layer,
a modified gold electrode with the 3-MPA was immersed in a solution of
ferritin at a potential of 0.00 V for five minutes. The electrode was then
45


POTENTIAL (Volts vs Ag/AgCI) POTENTIAL (Volts vs Ag/AgCI)
POTENTIAL (Volts vs Ag/AgCI) POTENTIAL (Volts vs Ag/AgCI)
Figure 3.8-Cyclic voltammagrams of horse spleen ferritin at a
polycrystalline gold electrode with 3-MPA layer
adsorbed in pH 7 buffer with varying scan rates
(a) 50 mV/sec (b) 100 mV/sec (c) 200 mV/sec
(d) 400 mV/sec.
46


removed under nitrogen then rinsed with pH 7 phosphate buffer and
subsequently immersed in pH 7 phosphate buffer/I M Na2SC>4 solution.
A cyclic voltammagram was obtained indicating only the presence of the
3-MPA background layer without any electroactivity of ferritin. This
result indicates that ferritin does not irreversibly adsorb to the 3-MPA
modified gold electrode layer. Previous studies with other proteins have
indicated that the adsorption of proteins occurs to some extent preceding
the electron transfer step (Sagara et al., 1991). Even though ferritin does
not undergo irreversible adsorption, the data suggests that a least weak
adsorption occurs.
Ferritin Iron Versus Free Iron
Initially the commercially available ferritin used in all of the experiments
is eluted through the G-25 sephadex column with pH 7 buffer which
replaces the transport buffer and removes any non-ferritin iron. It was
necessary to confirm that only inner ferritin iron was responsible for the
redox currents observed. Nitrilotriacetate (NTA) is a powerful
complexing agent which will serve as an iron sponge, especially in neutral
or basic conditions. The formal potential for the bis(NTA) iron (II) -
bis(NTA) iron (HI) complex is in the region of 0.05 V (vs. Ag / AgCl) which
is sufficiently separated from ferritin redox potentials. By immersing a
clean gold electrode in a solution of 4.9 mg/mL ferritin, 0.010 M
nitrilotriacetate, pH 7 tris buffer, and 1 M Na2SC>4 the cyclic
voltammagram was scanned. The resulting cyclic voltammagram
resulted in an absence of an iron-NTA complex, indicating that free iron
was not present in the ferritin sample (Figure 3.9a). The 3-MPA layer was
47


not used in this experiment since the layer might have inhibited the
electron transfer of the iron-NTA complex. Tris buffer was used instead of
phosphate buffer since the solubility products of iron(II) and iron(III)
phosphates are very small and added iron would quickly be removed from
solution by precipitation.
For comparison the cyclic voltammagram of the iron-NTA complex
was obtained by immersing the clean gold electrode into a solution of 0.010
M ferrous sulfate, 0.010 M nitrilotriacetate, pH 7 tris buffer, and 1 M
Na2SC>4 (Figure 3.9b). The cyclic voltammagram indicates a reversible
wave at an E of approximately 0.05 V. The iron-NTA complex forms in
even the presence of relatively high hydroxide concentrations at pH 7.0.
48


(Idujioisii'i) ueuuro
Figure 3.9-Cyclic voltammagrams of (a) ferritin and
nitrilotriacetate and (b) ferritin, nitrilotriacetate, and ferrous
sulfate at a polycrystalline gold electrode without the 3-MPA
layer in pH 7 tris buffer, ferritin concentration 4.9 mg/ mL,
nitrilotriacetate concentration; 0.010 M, ferrous sulfate
concentration 0.001 M, electrode area 2.15 cm.2, scan rate 100
mV/sec.
49


Current-Potential of Apoferritin
Apoferritin is the ferritin protein structure with the intra-protein
ferritin iron absent. The voltammetric currents observed with ferritin are
large compared with those obtained from apoferritin. The only
appreciable current in the scan of apoferritin is present due to the adsorbed
3-MPA layer (Figure 3.10). The absence of any additional currents in
apoferritin strongly suggests that the voltammetric waves present in
ferritin originate from the oxidation and reduction of iron in the
inorganic ferritin core.
Electrochemicallv Induced Iron Transport Studies Using EDTA
Even though iron is sequestered by ferritin in an organism, studies
to this point have not been able to show that iron exits the ferritin protein
core. Previous studies have proposed that ferritin must degrade before the
release of the inner core ferritin iron (Funk et al., 1985). The results of
previous studies suggest the oxidation of iron (II) to iron (HI) is involved
as iron enters the inner core of ferritin (Pape et al., 1968). Other in vitro
experiments have shown iron can be induced to exit the ferritin protein if
the iron is first reduced from iron (IE) to iron (II) (Sirivech et al., 1974).
The direct electron transfer of ferritin catalyzed by self-assembled short-
chain monolayer can be employed to study whether the electrochemical
reduction of ferritin iron can induce iron to be released from the ferritin
core.
50


Figure 3.10-Cyclic voltammagram of horse spleen apoferritin at a
polycrystalline gold electrode with 3-MPA layer adsorbed in pH
7 buffer, 3-MPA concentration 0.001 M, apoferritin
concentration 4.9 mg/mL, electrode area 2.15 cm^, scan rate 100
mV / sec.
51


If iron is released by ferritin upon reduction, then in the presence of
a strong complexing agent, the reduced iron should be induced to leave
the protein shell. EDTA, a strong complexing agent for many metal ions,
was used to complex ferritin iron once it was reduced. As discussed
previously ferritin iron is not removed by the presence of nitrilotriacetate,
therefore EDTA was not expected to directly remove ferritin iron (III). The
cyclic voltammagram of ferritin in the presence of EDTA using the 3-MPA
modified electrode indicates a reduction of the ferritin iron in the negative
potential direction at approximately -0.25 V. However upon switching the
scan direction at -0.5 V there is an absence of the anodic wave (Figure 3.11).
This result suggests that as the ferritin iron is reduced electrochemically
the iron is released as Fe2+ then is rapidly consumed by EDTA. Scanning
the potential anodically does not produce an anodic ferritin iron wave
since the iron is bound by EDTA and no longer in the ferritin inner core.
The scan rate dependence of the ferritin-EDTA solution shows an increase
in the cathodic wave as the scan rate is increased. However the anodic
peak region for ferritin iron is absent of any increase in electroactivity
which indicates there is not an increase in iron released as the scan rate is
increased (Figure 3.12).
A cyclic voltammagram was obtained with only EDTA and ferric
sulfate to determine the peak potentials of the iron-EDTA redox reaction
(Figure 3.13b). To verify an iron-EDTA complex had formed following the
reduction of ferritin iron, the potential of the 3-MPA modified gold
electrode was poised at -0.31 V for 120 minutes. During this period
sufficient iron was reduced to form a detectable iron-EDTA complex. After
the 120 minutes a cyclic voltammagram was obtained indicating the
52


presence of the iron-EDTA complex (Figure 3.12a). These results are
strong evidence that iron is released by ferritin when the inner core iron is
reduced to Fe^+. These results are in agreement with the in vitro studies
which suggest the reduction of iron precedes the release of iron from the
inner core of ferritin.
53


iMiciotmpi)
Figure 3.11-Cyclic voltammagrams of (a) horse spleen ferritin and EDTA
(b) horse spleen ferritin, at a polycrystalline gold electrode 3-
MPA layer adsorbed in pH 7 buffer, 3-MPA concentration
0.001M, ferritin concentration 4.9 mg/mL, EDTA
concentration 0.010 M, electrode area 2.15 cm^, scan rate 100
mV / sec.
54


(Micioimpi)
Figure 3.12-Cyclic voltammagrams of horse spleen ferritin and
0.010 M EDTA, at a polycrystalline gold electrode with
3-MPA layer adsorbed in pH 7 buffer, electrode area
2.15 cm2, scan rate; (a) 50 mV/sec, (b) 100 mV/sec, (c) 200
mV/sec.
55


(Micro tmpi)
POTENTIAL (Volts vs Ag/AgCI)
Figure 3.13-Cyclic voltammagrams of (a) horse spleen ferritin
and EDTA, taken after stirring at -0.31 V for 2 hr. at a
poly crystalline gold electrode with 3-MPA layer
adsorbed in pH 7 buffer and (b) iron-EDTA complex at a
polycrystalline gold electrode in 1 M H2SO4,. 3-MPA
concentration 0.001 M, ferritin concentration 4.9
mg/mL, EDTA concentration 0.010 M, ferric sulfate
concentration 0.010 M, electrode area 2.15 cm^, scan rate
100 mV/sec.
56


Effects of Varying Mercaptan Chain Lengths on the
Electrochemistry of Ferritin
While effecting the direct electron transfer of ferritin is essential in
understanding its electrochemistry, it is necessary to develop a molecular
level description of the proteins interaction with the electrode surface. As
described here, the electrochemical system will be perturbed by changing
one of the characteristics of the promoting layer: the carbon chain length.
The direct electron transfer between the ferritin iron core and the
modified gold electrode was probed using modified layers of 3-MPA, 6-
MHA, and 8-MOA. For these studies, the ferritin was purified further
using size exclusion chromatography. Obtaining a cyclic voltammagram
with an adsorbed layer of 3-MPA a comparison between the original
purified ferritin (G-25 Sephadex column only) and the further purified
ferritin (G-25 Sephadex column and G-200 sephadex column) was made to
investigate any difference in redox activity under different purity of
ferritin (Figure 3.14). Although the resulting cyclic voltammagrams
indicated the same peak potentials the further purified ferritin had slightly
higher cathodic peak. This enhanced redox activity of the further purified
ferritin was possibly due to the absence of oligomeric forms of ferritin.
The cyclic voltammagrams of each background adsorbed layer
showed similar characteristics (Figure 3.15). The peak current potential
behavior of ferritin (4.6 mg/ mL) at various promoter layers is shown
(Figure 3.16). Using a scan rate of 100 mV/s for each promoter the cathodic
peak remained relatively constant at -0.35 V, and the anodic peak at -0.21
V. However the peak currents decreased when the promoters chain
length increased. The definition of the cathodic and anodic peaks
57


CURRENf (Microamps)
POTENTIAL (Volts vs Ag/AgCI)
Figure 3.14-Cyclic voltammagrams of horse spleen ferritin (a) originally
purified and (b) further purified at a poly crystalline gold
electrode with 3-MPA layer adsorbed in pH 7 buffer, 3-MPA
concentration 0.001 M, ferritin concentration 4.9 mg/ mL,
electrode area 2.15 cm^, scan rate 100 mV/sec.
58


(Microamps)
POTENTIAL (Volts vs Ag/AgCI)
Figure 3.15-Cyclic voltammagrams of monolayers of (a) 3-MPA
(b) 6-MHA, (c) 8-MOA adsorbed on polycrystalline gold
electrode in pH 7 buffer, mercaptan concentration 0.001 M.
2
electrode area 2.15 cm scan rate 100 mV/sec.
59


(Microamps)
POTENTIAL (Volts vs Ag/AgCI)
POTENTIAL (Volts vs Ag/AgCI)
bigure 3.16-Cyclic voltammagrams of horse spleen ferritin at a
polycrystalline gold electrode with (a) 3-MPA, (b) 6-MHA,
(c) 8-MOA layer adsorbed in pH 7 buffer, mercaptan
2
concentration 0.001 M electrode area 2.15 cm scan rate 100
mV/sec.
60


decreased as the chain length was increased indicating decrease in the
effectiveness of the promoter.
The scan rate of the ferritin cyclic voltammagrams varying from 50-
400 mV/ s using a gold electrode modified with 3-MPA, 6-MHA, and 8-
MOA are shown in ((Figure 3.17), (Figure 3.18), and (Figure 3.19)
respectively. The peak potential difference increases with scan rate,
varying within the range of 125-225 mV indicating that the electron
transfer kinetics are relatively slow. The cathodic peak currents increased
correspondingly with scan rate as expected. The results of the varying
mercaptan chain lengths studies are summarized in Table 3.1.
61


UHItUJti (Micioaraps) CUlWB/r (Micioamps)
POTENTIAL (Volts vs Ag/AgCI)
POTENTIAL (Volts vs Ag/AgCI)
POTENTIAL (Volts vs Ag/AgCi)
Figure 3.17-Cyclic voltammagrams of horse spleen ferritin at a
poly crystalline gold electrode with 3-MPA layer adsorbed in
pH 7 buffer with varying scan rates, (a) 50 mV / sec, (b) 100
mV/ sec, (c) 200 mV/ sec, (d) 400 mV/sec.
62


(Microamps)
POTENTIAL (Volts vs Ag/AgCI)
mV/sec, (c) 200 mV/sec, (d) 400 mV/sec.
63


Figure 3.19-Cyclic voltammagrams of horse spleen ferritin at a
polycrystalline gold electrode with 8-MOA layer adsorbed in
pH 7 buffer with varying scan rates, (a) 50 mV/ sec, (b) 100
mV/ sec, (c) 200 mV/ sec, (d) 400 mV/ sec.
64


Table 3.1- Summary of varying mercaptan chain length studies*.
Adsorbed Layer
3-MPA 6-MHA 8-MOA
Rate constant 1.24X106 3.37X107 2.63X1 O7
Surface coverage 0.27 0.56 0.66
Equilibrium potential ECV 0.30 0.30 0.30
Cathodic peak current (ipC) uA 26.3 8.5 7.5
Anodic peak current (ipa) uA 5.2 5.1 3.8
*Scan rate 100 mV/sec.
65


As the chain length of adsorbed aliphatic molecules increases, the
coverage also increases With the increase in chain length comes a
concomitant increase in adsorbate packing efficiency. The decrease in the
electroactivity of ferritin as the order of self-assembly increases suggests
that the relative disorder of the 3-MPA layer is important in the direct
electron transfer process at the electrode. The relative randomness of the
3-MPA exposes more metal surface of the electrode. It can be inferred that
while the carboxylate moiety is responsible for certain electrochemically
favorable interactions with ferritin, that efficient electron transfer
depends, at least in part, on significant access to the metal electrode itself.
This inference can be extended to further conjecture that the mode of
electron transfer of ferritin with the surface
hopping" through the layer but rather fairly
does not involve "electron-
close contact with the surface.
66


CHAPTER 4
CONCLUSIONS
The primary goals of this research have been accomplished. An
effective promoter has been found to probe the redox activity of ferritin.
The level of redox activity of ferritin at a bare gold is low or non-existent.
However, when the gold electrode surface was modified with a monolayer
composed of short chain sulfhydryl terminated acids the direct electron
transfer of ferritin iron at the solid electrode surface was promoted. The
adsorbed layers project a negatively charged group into the solution which
apparently interacts in a manner which favors direct electron transfer at
the electrode. The resulting redox reaction was a transfer of 1910 electrons
per ferritin molecule which occurred at a relatively slow rate. The
resulting cyclic voltammagrams suggest weak adsorption of ferritin at the
mercaptan modified surface prior to the reduction of inner ferritin iron.
Cyclic voltammagrams obtained using ferritin purified with ion
separation (G-25 Sephadex) and size exclusion (G-200 sephadex)
chromatography indicated better peak resolution probably due to the
absence of oligomeric forms of ferritin. Obtaining cyclic voltammagrams
in the presence of nitrilotriacetate indicates that free iron is absent in the
ferritin sample. Additionally, no current was observed when using
apoferritin to obtain cyclic voltammagrams. These results indicate the
redox reactions involving the iron within the ferritin inorganic core were
responsible for the currents observed with ferritin. By reducing ferritin in
the presence of EDTA a cathodic current peak was present, however the
anodic current peak was not present suggesting upon the reduction of
67


iron, the iron is induced to exit the protein and complex with EDTA. The
studies involving the various chain lengths indicate that the
electroactivity is decreased when the chain lengths were increased. The
heterogeneous rate constants decreased as the adsorbate chain length was
increased, indicating slower kinetics as the chain lengths were increased.
Although the peak potentials remained constant anodic and cathodic peak
currents decreased as the chain lengths were increased. The
differentiation between the various chain lengths suggests that the
relatively greater randomness of the 3-MPA layer is partly responsible for
the increased redox activity of ferritin. These studies have shown short
chain carboxylates promote the direct electron transfer of ferritin at a gold
electrode and provide for an effective technique to probe electron transfer
kinetics.
68


BIBLIOGRAPHY
Bard, A.J. & Faulkner, L.R., (1980) Electrochemical Methods
John Wiley & Sons, New York
Caskey, J.H., Jones J.H., Miller, Y.E., & Seligman, P.A. (1983)
Proc. Nat'l. Acad. Sci. U.S.A. 80: 482-486
Chasteen, N.D. & Xu. B. (1991) J. Biol. Chem. 266: 1966-1970
Clegg, G.A., Fitton, J.E., Harrison, P.M. & Treffry, A. (1980)
Prog. Biophys. Mol. Biol. 36: 56-86
Collinson, M., Bowden, E.F., & Tarlov, M.J. (1992) Langmuir 8: 1247-1250
Gorun,S.M., Papaefthymiou, G.C., Frankel, R.B., & Lippard, S.J. (1987) J.
Am. Chem. Soc.109: 3337-3348
Gutteiidge, J.M.C., Halliwell, B., Treffry, A., Harrison, P.M., & Blake, D.
(1983) Biochem J. 209: 557-560
Harrison, P.M., White, J.L., Smith, A., Farrants, G.W., Ford, G.C., Rice,
D.W., Addison, J.M., & Treffry, A. (1985) "Proteins of Iron Storage and
Transport"; Spik, G., Montreuil, J., Crichton, R.R. & Mazurier, J., Eds.,
Elsevier Science Publishers, B.V., 67-79
Heusterspreute, M., & Crichton, R.R. (1981) Fedn. Eur. Biochem. Soc. Lett.
129: 322-327
Ivanovics, G. and Vargha, L.Z. (1944) Physiological Chemistry
281: 156-162
Jacobs, D., Watt, G.D., Frankel, R.B., & Papaefthymiou, G.C. (1989)
Biochem. J. 28: 9216-9221
Kadir, F.H., Al-Massad, F.K., Fatemi, S.J., Singh, H.K., Wilson, T.W., &
Moore, G.R. (1991) Biochem. J. 278: 817-820
Klausner, R.D., Rouault, T.A. & Harford, J.B. (1993)
Cell 72: 19-28
69


Lipschitz, D.A., Bothwell, T.H., & Seftel, H.C. (1971)
Br. J Haematol 20: 155
Mack, U.D., Cooksley, W.G.E., Feris, R.A., Powell, L.W. & Halliday, J.W.
(1981) Br. J. Haematol 47: 403
Malem, F., & Mandler, D. (1993) Anal. Chem. 65: 37-41
May P. & Williams D. (1977) FEBS Lett 78: 134
Monroe, M.C., & Munro, H.N. (1978) In Stave U, ed: Perinatal physiology
(2nd edition), New Tork: Plenum, 425
Price, D.J., & Joshi J.G. (1983) J. Biol. Chem 258: 10873
Sagara, T., Murakami, H., Iharashi, S., & Sato, H. (1991) Langmuir 7:3190-
3196
Salonen, J.T., Nyyssonen, K., Korpela, H., Toumilheto, J., Seppanen, R., &
Salonen, R. (1992) Circulation 86: 803-811
Siimes, M.A. & Dallman, P.R. (1974) Br. }. Haematol 28: 7
Tanihuchi, I., Wantanabe, K., Tominaga, M., & Hawkridge, F.M. (1992)
Denki Kagaku 60: 1043-1049
Theil, E.C. (1987) Ann. Rev. Bioiochem. 56: 289-315
Tominaga, M. Kumagai, Tv Takita, S., & Taniguchi, L (1992) J. Electroanal.
Chem. 333: 331-338
Treffry, A. & Harrison, P.M. (1982) "The Biochemistry and Physiology of
Iron"; Saltman, P. & Hegenauer, J., Eds., Elsevier North Holland, Inc., 459-
462
Treffry, A., Hirzmann, J., Yewdall, S.J., & Harrison, P.M. (1993) FEBS Lett.
302: 108-112
Watt, G.D. (1979) Analytical Biochemistry 99: 399-407
70


Watt, G.D., Frankel, R.B., & Papaefthymiou, G.C. (1985) Proc. Nat'l, Acad.
Sci. USA 82:3640-3643
Watt, G.D., Frankel, R.B., Jacobs, D., Huang, H. & Papaefthymiou, G.C.
(1992) Biochemistry 31: 5672-5679
Watt, G.D., Jacobs, D., & Frankel, R.B. (1988) Proc. Nat'l. Acad. Sci. USA 85:
7457-7461
Weinbrerg, E.D. (1974) Science184: 952
Worwood, M., Brook, J.D., Cragg, S., Hellkuhl, J. & Jones, B.M. (1985)
Hum. Genet. 69: 371-374
71