Studies of electrochemically induced iron release of horse spleen ferritin using long optical path length thin-layer spectroelectrochemistry

Material Information

Studies of electrochemically induced iron release of horse spleen ferritin using long optical path length thin-layer spectroelectrochemistry
Kreutzian, Todd B
Publication Date:
Physical Description:
62 leaves : illustrations ; 28 cm


Subjects / Keywords:
Ferritin ( lcsh )
Iron proteins ( lcsh )
Biological transport ( lcsh )
Electrochemistry ( lcsh )
Spectrum analysis ( lcsh )
Biological transport ( fast )
Electrochemistry ( fast )
Ferritin ( fast )
Iron proteins ( fast )
Spectrum analysis ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 59-62).
Department of Chemistry
Statement of Responsibility:
by Todd B. Kreutzian.

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:
49630827 ( OCLC )
LD1190.L46 2001m .K73 ( lcc )

Full Text
ToddB. Kreutzian
B.A. St. Olaf College, 1993
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Science

This thesis for the Master of Science
degree by
Todd B. Kreutzian
has been approved
Donald C. Zapien
John A. Lanning

Kreutzian, ToddB. (M.S., Chemistry)
Studies of Electrochemically Induced Iron Release of Horse Spleen Ferritin Using
Long Optical Path Length Thin-Layer Spectroelectrochemistry
Thesis directed by Associate Professor Donald C. Zapien
Iron released from the direct reduction of adsorbed ferritin was
quantitated. Horse spleen ferritin was adsorbed onto indium tin oxide (ITO)
electrodes. A long optical path-length electrochemical cell (LOPTLTE) was
designed and constructed to accommodate the ITO/adsorbed ferritin electrode. A
negative potential step in the presence of an iron-chelating agent initiated the
reduction and release of the core iron, and the absorbance of the complex was
measured using UV-visible spectroscopy. The LOPTLTE cell was also used to
obtain the current-potential curve; the integrated current was used to calculate the
amount of iron reduced using the Faraday law. The amount of iron released was
measured to be within 19% of the amount of iron reduced.
The kinetics of iron release show marked pH dependence with the rate
increasing with decreasing pH. The electron transfer coupled with the migration
rate of iron from the core to the protein surface was determined to be 2 4 sec'1.
This rate is significantly faster than previously reported in the literature. The

quantitation of the released iron confirms that ferritin forms a single molecular
layer on ITO, and that iron release is induced by applied negative potential. Iron
release rate curves are sigmodial in shape suggesting a non-uniform rate of iron
release from the protein core.
This abstract accurately represents the content of the candidates thesis. I
recommend its publication.

Thank you to my advisor Dr. Donald Zapien, for his willingness to teach a
chromatographer electrochemistry. Thank you to Dr. Larry Anderson for his
immeasurable help in the kinetics section of this thesis. I would also like to thank
Stephanie Portfolio and Paul Glenn of Geneva Pharmaceuticals, Inc. for their
understanding and support during my research.
I would like to recognize the National Science Foundation (Grant Number
CHE-0070875) for their grant that allowed this work to be done.

Figures ..........................................................
1. Introduction..................................................
1.1 Ferritin protein structure and function.......................
2. Experimental..................................................
2.1 Reagents......................................................
2.2. Equipment....................................................
3. Long optical path length thin layer electrochemical cell (LOPLTLE)
3.1 Cell fabrication..............................................
3.2 Effective path length determination..........................
3.3 Electrode area determination.................................
3.4 LOPLTLE volume determination.................................
3.5 LOPLTLE testing..............................................
4. Kinetics measurement apparatus................................
4.1 Fabrication of optical bench.................................
5. Procedures....................................................
5.1 F erritin purification.......................................
5.2 ITO electrode preparation....................................
. 1

5.3 Ferritin adsorption onto ITO electrode............................ 26
5.4 Solution preparation.............................................. 26
5.5 Cyclic voltammetery............................................... 27
5.5.1 Calculation of iron release...................................... 30
5.6 Iron release using LOPLTLE and visible spectrometry............... 31
5.7 Iron release kinetics using LOPLTLE............................... 34
6. Results and discussion............................................. 36
6.1 Quantitation of iron release..................................... 36
6.2 Iron release kinetics............................................. 38
7. Conclusion........................................................ 57
References............................................................ 59

1.1 Cartoon of ferritin showing arrangement of subunits........... 3
1.2 Cross sectional diagram of ferritin........................... 4
1.3 Ribbon representation of loaded ferritin...................... 5
3.1 Exploded front view of LOPLTLE cell........................... 15
3.2 Side view of LOPLTLE cell.................................... 16
3.3 Cyclic voltammogram of ferroin............................... 19
4.1 Optical Layout for kinetics study............................. 22
5.1 Absorbance vs. BSA concentration calibration curve............ 25
5.2 Cyclic voltammogram of adsorbed ferritin on ITO.............. 28
5.3 Cyclic voltammogram of adsorbed ferritin on ITO in the
presence of 1,10-Phenanthroline............................... 29
5.4 Absorbance vs. ferroin concentration calibration curve....... 33
6.2.1 Absorbance vs. time graph for pH dependent kinetics,
1 sec. time scale........................................... 42
6.2.2 Absorbance vs. time graph for pH dependent kinetics,
2.5 sec. time scale........................................ 43
6.2.3 Phen Iron complex concentration vs. time graph for pH
dependent kinetics. Modeled rates are overlaid. 1 sec. time scale.... 49
6.2.4 Phen Iron complex concentration vs. time graph for pH
dependent kinetics. Modeled rates are overlaid. 2.5 sec. time scale.. 50

6.2.5 Absorbance vs. time graph for pH dependent kinetics,
250 ms time scale.......................................... 53

Table 6.1 Error analysis of iron release kinetic data

1. Introduction
1.1 Ferritin protein structure and function
Ferritin is the primary storehouse of nonhemoglobin iron in most
organisms. Ferritins other functions are to sequester and release iron in response
to biological triggers thus regulating iron stores in the body. Problems in the
regulation if iron can cause anemia or hemochromatosis (iron deficiency or iron
overload). Ferritin is the only protein capable of controlling the phase transition
of a metal from bound or free ions to a crystalline structure, and in the reverse
process breaking the crystalline structure and releasing the metal back into
Ferritin is a self-assembling protein composed of 24 polypeptide helical
subunits held together through noncovalent interactions arranged in 4,3,2
symmetry (rhombic dodecahedron, cubic space group F432, a 18.5 nm). Each
subunit consists of four long a helices, forming a bundle with a fifth short a helix
arranged at roughly 60 to the bundle axis. Two of the long helices are connected
by a loop and two of these loops from neighboring subunits arrange to form a /?
sheet. Figure 1.1 shows the basic arrangement of the a-helices that form the
protein coat of ferritin1. This particular spatial arrangement allows for the
formation of channels along the threefold and fourfold symmetry axes. The

channels are thought to play an important part in the deposition and release of
iron. The six channels with fourfold symmetry are hydrophobic, whereas the
eight channels with threefold symmetry are hydrophilic It is through the
hydrophilic channels that iron passes into the protein shell3.
Ferritin has a spherical shape (~ 12 nm in diameter) with a hollow center
(~8 nm in diameter) in which the iron is mineralized in a crystalline form that
resembles the ferric oxide mineral ferrihydrite (Fe3+0(0H)) 4 Figure 1.2 shows
a cross sectional representation of ferritin5. See figure 1.3 for a cross sectional
ribbon representation of iron-loaded ferritin6.

Figure 1.1

Figure 1.2
Cross sectional representation of ferritin

Figure 1.3
Ribbon representation of loaded ferritin.

Iron in the +2 oxidation state is oxidized to the +3 state as it enters the
protein shell. The iron can be induced to exit the protein shell when the ferritin
iron is first reduced to the +2 oxidation state. The electron transfer steps involved
in this process are unclear .
Ferritin is an unusually robust protein. It can be heated to about 80C
without apparent damage. It can also tolerate changes in pH from 2 to 10. This
stability is unmatched by any other protein8.
For years our laboratory, and others, have studied ferritin iron reduction
and its subsequent release based on the assumption that one iron atom is reduced
per one electron transferred9. This assumption has been based upon
electrochemical measurements measuring the faradic charge required to reduce
iron in the protein core, or the spectroscopic measurement of iron released
following treatment with homogenous chemical reductants
It is known that cytochrome c irreversibly adsorbs at ITO electrodes as
shown by a well-defined current potential curve10. Electrochemical and
spectroscopic measurements were employed to determine the surface coverage of
cytochrome c. Reported results show agreement between the two measurements
within a factor of 2. The correlation between electrochemical and spectroscopic
measurements suggests that either cytochrome c does not adsorb as a reproducible
monolayer or the structure of cytochrome c is altered upon adsorption. The result

does, however, show the potential of combining electrochemical and
spectroscopic measurements in investigating surface coverages and adsorbed
protein structure vs. solution protein structure.
Previous work from our laboratory has suggested that ferritin can be
adsorbed at ITO, and that iron can be induced to exit ferritin when sufficient
negative potential is applied11. The adsorption of ferritin at ITO cannot be proven
by electrochemical methods alone. Current-potential curves that have been
interpreted as iron core reduction may be due to iron adsorbed directly at the ITO.
If ferritin spills its iron out onto the ITO surface during the adsorption process it
is possible that ferritin itself does not adsorb at the ITO. Spectroscopy combined
with electroanalytical methods can be used to confirm that ferritin does indeed
adsorb at ITO, and that iron is released from the protein core following an
electrochemically-induced reduction in the presence of a chelating agent.
Spectroelectrochemistry has developed into a powerful tool for chemists
since its earliest use reported by Kuwana, Darlington and Leedy in 196412 and
Hubbard and Anson in 196413. Hubbard and coworkers have developed
numerous thin layer electrochemical cells designs14. These cells are useful in
studying the electrochemical properties of compounds, however, they do not
allow the simultaneous collection of optical data. The desire to combine the
measurement of electrochemical with optical properties led to the development of

optically transparent thin-layer electrodes (OTTLE). Numerous OTTLE designs
have been reported including gold minigrids sandwiched between microscope
slides15, metallized-plastic electrodes16 and one using ITO as the electrode/optical
window placed into holes cut in a glass culture tube17. OTTLE generally have
path lengths on the order of 50 to 200 |im. While path lengths such as these are
sufficient for highly absorbing species or high concentrations of the species being
measured they are not adequate if the requirements of high absorptivity and
concentrations are not met.
Long optical path length thin-layer electrochemical cells (LOPLTLE)
were developed to overcome the problems associated with the short path length
OTTLE. Like the OTTLE before them, many different types of LOPLTLE have
been designed. Most LOPLTLE are constructed as cuvet type cells with path
lengths on the order of 0.1 to 1.0 cm18. These cuvet sized LOPLTLE were
designed to easily fit into a standard cuvet holder of a UV/Vis spectrophotometer
or inside the cuvet itself. The cuvet style LOPLTLE are easily fabricated and no
special mounting plates or aligning mounts are necessary making them ideally
suited for quick feasibility studies. Cuvet LOPLTLE provide adequate path
lengths for compounds with high extinction coefficients, however like their
OTTLE counterparts, they lack the necessary path length for compounds with
small molar absorptivities such as biomolecules. Their relatively short path

lengths necessitate the use of high sensitivity spectrophotometers. Kuwana and
coworkers designed an early LOPLTLE (1.6 cm path length, 100 pm solution
thickness) in 198319, this LOPTLTE became the inspiration for the cell described
in this thesis. Other more elaborate types of LOPLTLE have been fabricated
ranging from drilling a small diameter hole (500 pm) in glassy carbon to provide
the transparency20 to a circulating type LOPLTLE that allowed the solution
studied to be continuously stirred to overcome the problem of diffusion21.
LOPLTLE cells have been examined for use as detectors for high performance
liquid chromatography and flow injection analysis in order to combine
absorbance/fluorescence with electrochemical data22. Fiber optics have been used
to couple a thin-layer cell with a UV/Vis spectrophotometer, thus avoiding the
problems associated with the optical properties of transparent electrodes23.
A long optical path length thin-layer electrochemical cell (LOPLTLE) was
designed and used to measure the amount of iron released following core iron
reduction by direct electron transfer. The LOPLTLE was constructed to answer
the questions of (1) whether ferritin in fact adsorbs at ITO, and (2) The number
of iron atoms reduced per 1 electron transferred, and (3) whether ferritin adsorbs
in monolayer coverage at ITO. Another application of the LOPLTLE design is to

measure the iron release kinetics following direct the reduction of ferritin without
using homogenous chemical reductants.

2. Experimental
2.1 Reagents
Indium tin oxide on glass (100 ohms) was donated by Applied Films
Technology, (Longmont, CO). Ferroin (0.025 M stock), ferritin (>85 %), and
ferrous sulfate hexahydrate (99+% purity) were obtained from Sigma-Aldrich
Chemical Company (St. Louis MO). Sodium dihydrogen phosphate (ACS grade),
tris(hydroxymethyl)aminomethane (molecular biology grade),
1,10-phenanthroline (97%), sodium hydroxide (reagent grade), and hydrochloric
acid (37%, reagent grade) were purchased from VWR (San Francisco, CA) and
sodium chloride (reagent grade) from Fisher Scientific (Pittsburgh, PA). All
reagents except for ferritin were used without further purification. Water was
purified by distilling water vapor through a heated platinum catalyst in the
presence of oxygen24.
2.2 Equipment
All electrochemical experiments were performed using a Cypress model
Omni 90 analog potentiostat (Lawrence, KS) and a Bio Analytical Systems model
RXY recorder (West Lafayette, IN). A Perkin Elmer (Norwalk, CT) model 552A
UV/Vis spectrophotometer was used for determining ferritin concentration and
for the spectrophotometric quantitation of iron released by adsorbed ferritin. Size

exclusion purification was accomplished using a FLEX column (2.5 cm i.d. x 100
cm) obtained from Kontes (Vineland, NJ). For the kinetics study applied
electrode potential and photomultiplier signal data were simultaneously collected
on a dual channel digital oscilloscope (Tektronix Model 430A). Raw data from
the oscilloscope was imported into EXCEL for data manipulation and

3. Long Optical Path Length Thin Layer Electrochemical Cell
3.1 Cell Fabrication
The side and front views of the LOPLTLE cell are shown in Figures 3.1
and 3.2, respectively.
The cell body was machined out of a single block of Plexiglas. A pair of
polyethylene spacers (86 pm) were used to separate the ITO electrode from the
cell body defining the thickness of the cell solution; the length of the spacers
(2.02 cm) defined the optical path length. The thicknesses of the polyethylene
spacers were verified using a calibrated optical imaging system (IA32 imaging
system, LECO corp., St Joseph MI). A rectangular fine glass frit was imbedded
into the cell body in order to separate the thin layer from the reservoir containing
the reference and auxiliary electrodes. The glass frit and cell body were cemented
together using quick-set epoxy (Loctite Corp.) and made coplanar using a surface
grinder. The purpose of the frit was to provide sufficient cell conductivity and
solution separation.
A Vicor disk separated the reference electrode (Ag/AgCl) solution from
the electrolyte solution. Platinum wire (24 mm2 total immersed area) was used
for the auxiliary electrode. An aluminum reflector plate was mounted on top of
the ITO electrode to reduce reflection losses as the focused beam passed through
the cell. The reflector was also used to apply even pressure on the ITO electrode,

providing a uniform thin-layer solution thickness. Female luer fittings were
installed into the cell body to facilitate solution introduction and exit. Silicone
rubber gaskets (1/16 thick prior to compression, Accurate Conversion Resources,
Denver CO) were placed between Lucite-L windows (15 x 28 x 3 mm) and the
cell body. The windows were pressed against the gaskets using a spring clamp.
An acrylic cell was used to adjust the alignment of the thin-layer with respect to
the incident beam.

Figure 3.1
Front view of LOPLTLE cell. (1) ITO working electrode. (2) Lucite L window.
(3) Polished aluminum reflector plate. (4) Acrylic pressure plate. (5) 89 pm
polyethylene spacer. (6) Solution exit (polypropylene luer fitting). (7) Fine
porosity frit. (8) Auxiliary and reference reservoir. (9) Solution entrance
(polypropylene luer fitting). (10) Plexiglas cell body.

Figure 3.2
Side View of LOPLTLE cell. Pressure plate and aluminum reflector removed for
clarity. (1) ITO working electrode. (2) 89 p.m polyethylene spacer. (3) Lucite-L
window. (4) Silicon rubber seal. (5) Fine porosity frit. (6) Solution entrance
channel. (7) Solution exit (polypropylene luer fitting). (8) Solution exit channel.

3.2 Effective path length determination
The effective LOPLTLE path length was determined by UV/Vis
spectroscopy. The absorbance of a known concentration of ferroin in TRIS buffer
solution in a standard 1 cm cuvet was recorded and compared to the absorbance of
the same ferroin solution in the LOPLTLE. The total effective path length of the
LOPLTLE was than calculated from the data. The effective path length was
determined to be 2.08 cm compared to a physical path length of 2.01 cm.
3.3. Electrode area determination
Electrode area was the product of the width of the electrode between the
spacers and the length of the electrode between the silicone gaskets, and
determined to be 3.9 cm2 using a calibrated ruler (NIST traceable).
3.4 LOPLTLE volume determination
The thin-layer volume of the LOPLTLE cell was determined as follows.
The luer ports were capped and the cell was filled with triethylcitrate (density
1.137 g/mL, MW 276.3) through the silicon mbber gaskets. The weight
difference between the filled and empty cell yielded the weight of the
triethylcitrate within the thin-layer void. The thin-layer solution volume,
calculated using the density of triethylcitrate, was determined to be

130 p,L 1 |aL.
3.5 LOPLTLE testing
The suitability of the LOPLTLE as an electrochemical cell was tested by
scanning the current-potential curve of tris-l,10-phenanthroline iron (II) complex
(lx 10'4 M in 1 M KC1). The curve was well-defined and contained anodic and
cathodic branches. The tris-l,10-phenanthroline iron (II) complex system is
electro chemically quasi-reversible. The current-potential curve obtained using
the LOPLTLE compared well with the I-E curve using an H-cell that is known to
be a suitable electrochemical cell. See figure 3.3 below.

0.00 0.25 0.50 0.75 1.00 1.25
Figure 3.3
Current potential curve of Ferroin. 5 x 10'4 M ferroin in 1 M KC1. UO working
electrode. The LOPLTLE was assembled and the potential scanned in 5e4 M
ferroin in 1 M KC1. Scan Rate: 100 mV/sec. Electrode Area 3.9 cm2.

4. Kinetics measurement apparatus
4.1 Fabrication of optical bench
An optical bench was fabricated that allowed precise beam adjustment and
attenuation. The optical layout of the apparatus used in measuring the kinetics of
iron release is shown in Figure 4.1.
Visible radiation was provided by a 100 pm fiber optic cable with a
tungsten filament lamp (Ocean Optics, Dunedin FL). Radiation from the fiber
optic cable was collimated using a 1.5 inch focal length double convex lens, and
passed though a variable width exit slit. The optical fiber, lens and exit slit were
mounted on an X-Y-Z translator for alignment purposes. The LOTLTLE was
mounted on a rotating mounting bracket on a height adjustable plate. Collimated
radiation passed though the LOTLTLE to a concave mirror that focused the beam
into a Czemy-Turner monochromator. The wavelength setting of the
monochromator was set to 512 nm, Xmax of Fe(phen)32+ complex. The
combination of the monochromator entrance and exit slits adjustments provided
the necessary signal attenuation to avoid saturation of the photomultipler tube
(PMT) detector. The PMT supply voltage was set to -600 V; this voltage setting
gave the least amount of noise. Channel 1 on the digital oscilloscope (potential
step voltage) was set to 500 mV full scale; Channel 2 (PMT response) set to 1.00
mV full scale. Experimental time scales of 2.5 sec, 1.0 sec, and 250 ms were used

to folly investigate the kinetics. Raw photomultiplier signal vs. time data was
reprocessed to concentration vs. time prior to rate constant modeling. See Figures
6.2.3 and 6.2.4 for data converted to concentration vs. time format.

Figure 4.1
Optical layout for iron release kinetics study: (1) Source (2) 50 pm Optical fiber
(3) X-Y-Z Translator mount (4) Convex-convex lens (5) Exit slit (6) Cell
Mounting platform (7) LOPLTLE (8) Concave mirror (9) Monochromator
entrance slit (10) Monochromator (11) Photomultiplier tube detector (12) Power
supply (13) Digital oscilloscope. Note: Not to scale.

5. Procedures
5.1 Protein Purification
Ferritin (type I from horse spleen) was purified by size-exclusion
chromatography using a Sephadex G-200 stationary phase and eluted with a
buffer consisting of 0.02M sodium phosphate, 0.9% sodium chloride, 0.002M
phenylmethyl sulfonyl fluoride and 0.05% sodium azide. Ferritin concentration
was determined after purification using the Bradford method25.
The concentration of iron in the purified ferritin sample was determined
spectrophotometrically. A series of ferrous sulfate standards was prepared by the
following procedure. Aliquots of varying ferrous ion concentration were treated
with 5.00 ml of 1.6 M hydroxylamine hydrochloride, the pH adjusted to 5.0 with
3.3 M sodium acetate, treated with 5.00 ml of 0.01 M 1, 10-phenanthroline, and
brought to 100.0 ml volume. A 500 pL ferritin sample was digested in 250 pL of
hot (70C) sulfuric acid (18 M), and cooled to room temperature. Approximately
10 ml of water were added to the sample, and complexed with 1,10-
phenanthroline as described for the standards. The concentration of iron in the
ferritin sample was determined by projecting its absorbance (510 nm) onto the
concentration axis. Figure 5.1 shows an example calibration curve of absorbance
vs. concentration of BSA (bolvine serum albumin). The non-zero intercept is a
result of water being used as the reference in the UV/Vis system. The sample and

standard solutions are blue in color. A blank consisting of all reagents except for
B S A or ferritin was not prepared. A blank without BSA or ferritin is green in
color. It is this lack of a true blank that causes the intercept to be non-zero.

BSA Standard
Figure 5.1

5.2 ITO electrode preparation
The ITO electrodes (Applied Films Technology, Longmont, CO) were cut
into 20 x 30 mm rectangles. The leading and tailing edges of the electrodes were
beveled approximately 10 degrees to improve the seal between the electrode and
the silicone gasket. The electrodes were cleaned by sonication in a mixture of
saturated Alconox in ethanol, followed by repetitive rinsing in purified water.
The ITO working electrode was allowed to hydrate in purified water for twenty-
four hours at room temperature.
5.3 Ferritin adsorption onto ITO electrode
Adsorbed layers of ferritin were formed on the ITO electrodes as follows.
Clean ITO electrodes were immersed in 0.1 mg/ml of purified ferritin in 1.0 M
phosphate buffer (pt= 1.0) at pH 7.0, for twenty-four hours at room temperature.
The electrodes were then rinsed free of dissolved ferritin with pure 0.5 M TRIS /
1.0 M NaCl pH 7.0 solution
5.4 Solution preparation
All solutions were prepared in triply distilled water and the pH adjusted (if
necessary) with concentrated hydrochloric acid or 5 N sodium hydroxide. All
solutions were deaerated with nitrogen prior to use. All solutions were stored in

tightly sealed glass bottles for no more than one week. TRIS was used as the
buffer due to its buffering capacity in the pH ranges of interest, pKa 8.1, its high
purity, and its lack of metal binding.
5.5 Cyclic Voltammetry
Cyclic voltammetry of ITO/ferritin electrodes was accomplished in the
LOPLTLE cell. The 0.50 M TRIS/1.0 M NaCl buffer was deaerated with
pressurized nitrogen, and introduced into the cell via a plastic syringe. The
electrode lead was clipped to an overhanging portion of the ITO electrode and the
potential scanned between +0.050 and -1,300V, at a scan rate of 100 mV/sec. See
Figures 5.2 and 5.3.

Figure 5.2
Current potential curve of adsorbed ferritin on ITO. A clean ITO electrode was
immersed in a solution of 0.1 mg/mL purified ferritin and 1.0M pH 7 phosphate
buffer for 24 hours. After rinsing with 0.5M TRIS / 1.0M NaCl pH 7 buffer. The
LOPLTLE was assembled and the potential scanned in 0.5M TRIS / 1.0M NaCl
pH 7 buffer. Scan Rate: 100 mV/sec. Electrode Area 3.9cm2.

-1.40 -1.10 -0.80 -0.50 -0.20 +0.10
Figure 5.3
Current Potential curve of adsorbed ferritin on ITO in the presence of lOmM
1,10-phenanthroline. Electrode preparation and conditions same as figure 5.2

5.5.1 Calculation of iron release
Current-potential curves of ITO/ferritin electrodes were ran immediately
following UV/Vis experiments (described below), using the LOPLTLE cell. The
current-potential curves were run in order to correlate the moles of iron calculated
from the UV/Vis release experiment to that determined electrochemically. The
cathodic faradic current from -0.41 to -1.30 V was integrated, and the number of
moles of iron electrolyzed was calculated using the Faraday law. It was assumed
that one electron was transferred per iron atom in the ferritin core. Current-
potential scans were performed in triplicate.
The number of moles of iron released, determined electrochemically, was
calculated using the Faraday law.
Q-Qb =nF (number of moles of Fe)
Q Qb was calculated from the cathodic peak area by a simple cut and weigh
method. Calculated in coulombs,
n = 1 (one electron transfer) per iron atom
F = Faradays constant, 96485 coulombs per mole of e'
Number of moles of Fe=

5.6 Iron Release using LOPLTLE and visible spectrometry
Iron release measurements were conducted using the LOPLTLE cell. The
LOPLTLE cell fitted with the ITO/ferritin electrode and was rinsed with not less
than 20 mL of the appropriate blank solution, 0.5 M TRIS / 1.0 M NaCl buffer
solution or 0.5 M TRIS / 1.0 M NaCl / 1,10-Phenanthroline solution, before all
measurements to prevent cross-contamination. The assembled LOPLTLE cell
was aligned in the commercial spectrophotometer by monitoring transmittance
signal. The reference cell was a standard 1 cm quartz cell filled with a variable
concentration of ferroin and TRIS buffer solution to attenuate the reference signal.
The reference/auxiliary reservoir was filled with 0.5 M TRIS /1.0 M NaCl buffer
solution and the electrodes placed in the solution. The thin-layer was filled with
0.5 M TRIS / 1.0 MNaCl buffer solution containing lOmM 1,10-phenanthroline.
The initial potential applied was +0.050V. The potential was then stepped to
-1,300V, and the absorbance of the Fe(phen)32+ complex was recorded. The
absorbances of ferroin standard solutions ranging in concentration from 2.5 x 10'6
M to 1.5 x10'4M were measured. The calibration curve was plotted. The
absorbance of the phenathroline complex of the released iron was projected onto
the concentration axis.

Absorbance of the Fe(phen)32+ complex was projected onto the
concentration axis in order to determine the total amount of iron released. The
cell was then removed from the spectrophotometer and thoroughly cleaned using
methanol and triply distilled water. After cleaning the cell it was placed in fresh
0.5 M TRIS / 1.0 M NaCl buffer solution for a minimum of thirty minutes in
order to recondition the cell frit. Due to the significant variability of absorbance
caused by slight changes in cell alignment the ferroin linearity was repeated for
each individual sample. Figure 5.4 shows an example calibration curve of
absorbance vs. ferroin standard concentration.

Trial #3 of 0.1 mg/mL Ferritin o n ITO
(Step +0.05V ano die to -1.300V cathodic, at 512 nm)
Figure 5.4

5.7 Iron release kinetics using LOPLTLE
Measurements of iron release kinetics were performed using the
LOTLTLE cell. To examine the pH dependence of kinetics of iron release, four
solutions of 0.5 M TRIS / 1.0 M NaCl / 10 mM 1,10-Phenanthroline were
prepared at pHs 6.0, 7.0, 8.0 and 9.0. Purified ferritin was adsorbed onto clean
ITO electrodes as described in the Section 5.3. The LOTLTLE cell was
assembled and filled with 0.5 M TRIS / 1.0 M NaCl /10 mM 1,10-Phenanthroline
solution of the desired pH. Alignment of the beam through the cell and
monochrometer was a simple task due to the ruggedness of the optical bench
design. The photomultiplier tube power supply was set at 600 V, and the
monochromators entrance and exit slits were opened until a stable signal with
minimal noise was displayed on the digital oscilloscope. The potential was
poised at +0.05 V, then stepped to -1.300 V. The potential step triggered
simultaneous the collection of voltage vs. time and photomultiplier tube radiant
power vs. time data. After data collection the LOTLTLE was rinsed in place with
not less than 20 milliters of 0.5 M TRIS / 1.0 M NaCl buffer solution. The
LOTLTLE was than removed from the mount assembly and completely
disassembled. The LOTLTLE was rinsed with triply distilled water, then
methanol, followed by 0.5 M TRIS / 1.0 M NaCl buffer solution. The LOTLTLE
was soaked in 0.5 M TRIS /1.0 M NaCl buffer solution for thirty minutes to

recondition the cell frit. After soaking the LOTLTLE it was again rinsed with 0.5
M TRIS / 1.0 M NaCl buffer solution and reassembled with a new ITO / ferritin
electrode. Not less than 20 milliliters of the next pH solution was added to the
LOTLTLE and the process repeated.
Time, applied potential and absorbance data was imported from the
oscilloscope to Microsoft EXCEL for data manipulation and presentation.
Concentration vs. Time data was calculated using Beers law.
log£ be
I0 = Radiation in,
I = Radiation out
C = Concentration in Moles
b = Path length in cm
s = Molar absorptivity in M'1 cm'1 (11500 for Phen Iron complex)26

6. Results and Discussion of Iron Release
6.1 Quantitation of Iron Release
Phen was used as the indicator due to its fast reaction rate with iron (ki=
5.6 X 104 M'1 sec-1 (27)). The stability of the formed complex is indicated by its
formation constant, Kform at 25 C is (5 X 10'22)'1 (28). Stopped-flow
spectroscopy studies have shown that without the interference of apoferritin the
rate of formation of the Fe2+(phen)3 complex depends on the concentration of
phenanthroline. The stopped-flow spectroscopy studies also show that the rate-
limiting step of the complex is not the initial binding of Fe2+ to phenanthroline but
the later step in the formation of the complete Fe2+-(phen)3 complex29. The
relative inertness of phenanthroline towards chemical reaction other than salt
formation or chelation is a significant asset in its analytical applications. The
excellent stability of phenanthroline makes it a excellent indicator because it will
not break down and interfere with the analysis30. The orange-red phenanthroline
complex of iron forms quantitatively over the pH range of 2-930.
The total amount of Fe-phen complex measured from the UV/VIS
experiments was 1.14 x 10'4 molar. Using a total measured cell volume of 1.304
x 10' liters the total moles of iron released was 1.49 x 10' moles corresponding
to 8.97 x 1015 iron atoms.

The total moles of iron electrolyzed from the chronoamperometry
experiments was 1.25 x 10' moles of iron, which corresponds to 7.53 x 10 iron
The correlations between the two methods of iron measurement agree
within 19%. An agreement of 19% is quite good when one considers the
numerous possible sources of error inherent in each technique.
Possible sources of error include irreproducibilitys in electrode coverage
and thickness spacer placement limiting the total amount of adsorbed ferritin
available for electrolysis. Statistical analysis using the t-test shows that the two
methods do not significantly differ (texp=1.423, ttheor=4-303 at the 95% confidence
Based upon ferritins diameter of 12 to 13 nm the maximum theoretical
area occupied per molecule is 1.1 x 10'12 cm2 31. The theoretical maximum
packing density of ferritin on ITO is 1.5 pmol/cm2. Previous studies have
determined that ferritin probably forms a monolayer on ITO electrodes11. The
calculated packing density from chronoamperometry was 1.4 0.04 pmol/cm2.
Comparison of the experimental to theoretical packing densities shows that a
monolayer of ferritin was formed on the ITO. In addition this result indicates that
all the iron within ferritin is released upon reduction in the presence of

6.2 Iron release kinetics
Direct electron transfer overcomes many of the problems inherent to
classical methods measuring iron release kinetics of ferritin. Many different
chemical reductants have been used such as reduced riboflavin, ascorbate,
dithionite, glutathione and dihydroflavins. The technique of direct electron
transfer eliminates the problems of molecular orientation and diffusional
limitations of the chemical reductants. Direct electron transfer provides the
electrons without the complicating factors such as absorbance interference and
oxidation of the chemical reductants by atmospheric oxygen. Chemical
reductants when used to measure kinetic rates often requires reaction times on the
scale of minutes to hours even days in some cases. The long times needed for
kinetic data can be detrimental to the quality of the results due to possible slow
protein denaturing or conformational changes. Results using direct electron
transfer show that iron is completely released from the protein core in about one
second at the slowest rate of release. The speed of electron transfer minimizes the
potential errors resulting from protein denaturing and conformational changes.
The iron release kinetic curves exhibit a sigmodial shape. This sigmodial
curve shape (see figures 6.2.1 and 6.2.2) suggests a nonunifoim rate of iron
reduction in the core. I.G. Macara and co-workers while studying iron uptake of
ferritin using potassium iodate and sodium thiosulphate as the oxidizing mixture

noted that the sigmodial curve is reminiscent of those obtained for reactions
catalyzed by allosteric enzymes. They calculated rate constants of core formation
based only on the linear portion of the curve32. Separate studies indicate that
similar sigmodial curves were obtained for iron release33. The similarity of the
curve shape for uptake and release of iron suggests that the mechanisms of uptake
and release are closely related.
T. Jones et al. obtained similar sigmodial iron release curves when
investigating iron release from ferritin by dihydroflavins and dihydroflavin
analogues34. The work was conducted under anaerobic conditions with a constant
(a dihydroflavin generating system was used) and known concentration of
dihydroflavin. The anaerobic conditions eliminated kinetic artifacts due to the
oxidation of the dihydroflavin. In calculating the rate constants, only the rate of
maximum release was used. The slow initial portion of the curve was disregarded
for lack of an appropriate model to explain the slow region. They state that the
sigmodial curve is always associated with complete iron release. They suggest
the following possibilities for the slow portion of the sigmodial progress curve:
(1) It may be a function of the lesser chemical reactivity of Fe3+ at the
outer surface of the core when compared to that which is highly
ordered in the crystal.

Fe3+ at the outer surface of the core not incorporated into the crystal matrix may
be more difficult to reduce. Electron transfer within the crystal matrix can occur
much faster and with greater efficiency because the iron atoms are within van der
Waals radii of each other. Lack of an electron transport bridge to iron not in the
crystal matrix would slow its release rate compared to iron in the highly ordered
(2) It may be due to the slow access of the chemical reducing agent to the
A dihydroflavin molecule (M.W. of 788 g/mole) in solution must first pass within
close proximity to a ferritin molecule and orient itself correctly in order to engage
in electron transfer. Using modified dihydroflavins the T. Jones and colleagues
concluded that in the case of dihydroflavin the molecule must pass through the
four fold channels of the ferritin shell (9 12 A across at the narrowest point) to
the iron core in order to effect electron transfer. The large size of dihydroflavin
should prevent it from entering the four fold channel. Dihydroflavin must engage
in electron transfer with accepters on the surface to ferritin. The combination of
proximity and orientation may cause the slow portion of the sigmodial iron
release curve. However, the sigmodial curve is still apparent when the direct
electron transfer to immobilized ferritin is performed. This effectively eliminates

the possibility that the sigmodial curve shape is caused by the slow approach of
the reducing agent to an optimal position and orientiation.

Absorbance vs. Time
1 sec. Time Scale
S -0.003
| -0.004
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time (seconds)
pH 6
a pH 8
o pH 9
Figure 6.2.1
Absorbance vs. time graph for pH dependent kinetics, 1 sec. time scale. 2500
data points over 1 sec. collection rate.

Absorbance vs. Time
2.5 sec. Time Scale
o pH 6 AVG
pH 7 AVG
A pH 8 AVG
o pH 9 AVG
Figure 6.2.2
Absorbance vs. time graph for pH dependent kinetics, 1 sec. time scale. 2500
data points over 2.5 sec. collection rate.

D.L. Jacobs and co-workers suggest that the slow initial portion of the
sigmoidal progress curve involves very little if any proton uptake. As the core
becomes more reduced proton uptake increases to a maximum of 2 Kr per e' (2H+
per Fe3+ reduced) during the fast portion of iron release. They suggest that the
slow initial phase of reduction is accompanied by only small pH changes and it is
not until a significant portion (20-30%) of the core is reduced that a pH change is
seen35. The initial reduction of the core, during which very little iron is released,
is thought to be the cause of the sigmodial curve. The studies by Jacobs and co-
workers differed from previous studies in that no chelating agent was used in the
analysis of iron release. Proton prior to and during iron release was monitored by
optical titration in the 350 -800 nm range using S2O42' as the titrant. The lack of a
chelating agent combined with the fact that a sigmodial release curve was still
obtained eliminates the possibility that the sigmodial iron release curve found
using direct electron transfer is in any way caused by the chelating agent, 1-10-
F. Funk and colleagues reported sigmodial iron release curves from
ferritin when using dithionite as a chemical reductant. An increase in iron release
rate was noted for acidic pHs. The rate increase was attributed to the
participation of H+ in the activated dithionite complex36.

We designed a simplified mechanistic model to explain the data. In the
model the release of iron from the crystalline core of ferritin occurs in six steps.
Equations 6.2.1 through 6.2.6 below detail the model.
(6.2.1) Fe(III) core C Fe2+ core
(6.2.2) Fe2+ migration to the surface of the ferritin shell
(6.2.3) Fe2+Surface + Surface Fe2+ Surface + H*core
(6.2.4) Fe Surface Phen ^ FePhen in solution
(6.2.5) FePhen in solution Phen ;n solution F eP hen2 msoiution
(6.2.6) FePhen2 in solution ^ Phen jn solution ^ FePhen^ in solution
The experimental data fit the model above. We cannot determine the
overall rate-limiting step from the data. The migration of iron from the core to the
surface (eqn. 6.2.2) coupled with the electron transfer rate (eqn. 6.2.1) determines
the overall rate of the reaction. For the purpose of the kinetic analysis we treat ki
as the first order rate constant for:
ki = Fe3+COre > Fe2+SUrface (equation 6.2.2)
rate = kifFe3^]1
The first order rate constant ki was calculated to be from 2 to 4 sec'1. It is
logical to assume that the electron transfer rate using direct electron transfer to
induce core reduction of immobilized ferritin is faster than that of the
mobilization of iron out of the protein core. Because ferritin is in direct contact

with the continuous supply of electrons complicating factors found when using
chemical reductants such as proximity and molecular orientation are eliminated.
The literature value for the rate determining step in Fe(phen)32+ formation
is k= 5.6 X 1041VT1 sec"1 (27).
How the electrons reach the core of ferritin is unknown. Electrons must
either tunnel or be shuttled through the thick (up to 25 A) protein coat to reach the
protein core. It is unclear which mechanism is correct. Studies conducted in
which ferritin was modified to contain a heme group increase the rate of iron
release37. This has lead Moore et al. to suggest that the rate determining step in
the entire iron release process is the electron transfer across the protein coat.
When ferritin is modified with a heme group the chemical reductant used does not
change the measured release rate. Using unmodified ferritin iron release rate is
dependent upon the nature of the chemical reductant. In Moore and coworkers
production of heme modified ferritin the heme was incorporated into the ferritin
shell. This modification of ferritin did not lead to dissociation of the 24 subunits,
however, it was not made clear where the heme binds in relation to the iron exit
channels. Iron release of heme modified horse spleen ferritin using the chemical
reductant 4-Hydroxycinnamic acid (single electron donor) was reported to be 5.2
x 10'6 p.moles Fe2+ released sec'1. This is significantly slower than the initial
release rate calculated using direct electron transfer of unmodified horse spleen

ferritin that was calculated to be between 2.5 x 10'2 to 5.0 x 10'2 pmoles Fe2+
released sec' (based upon an initial bound iron concentration of 1.2 x 10' moles
calculated from the total amount of iron electrolyzed using voltammetry). The
amount of iron calculated from voltammetry is valid based upon the fact that one
iron atom is reduced for one electron transferred and all of the core iron has exited
the protein. The large disparity in results suggests that heme plays an important
role in electron transfer using chemical reductants in solution; however, the
results cannot be applied to direct electron transfer techniques. The possibility
that heme modification of ferritin modifies the iron exit channels cannot be
discounted. Modification of the iron exit channels by heme may increase the rate
of iron release leading to the assumption that one has increased the rate of
electron transfer. Through the use of direct electron transfer the entire release
process was greatly accelerated. Which step, electron transfer or iron migration,
was increased is unclear. Once iron in the core is reduced, the iron migration
from the core to the protein surface should be the same regardless of the reduction
method. If this assumption is correct, direct electron transfer accelerates the
transfer of electrons across the protein coat. With a large increase in electron
transport it is logical to assume that the overall rate-limiting step is the migration
of iron from the core to the surface.

The proton exchange (6.2.3) rate is coupled with the rate at which TRIS
buffer releases protons. This rate was calculated to be much faster than the iron
migration step (6.2.2) and is represented by k2.
k2 Fe surface H*"surface F6 surface core (equation 6.2.3)
=k[Fe2i surface ][/T surface ]
The rate constant k2 was calculated to be in the range of 5 x 108 to 7 x 108
M'1 s1. The kinetics of the iron release from the surface of the protein may be
limited by the rate of proton release from the TRIS buffer.

| 3.00E-05
1 second time scale kinetics data
Experimental Results with Model Results Overlaid
kj = 2 sec"1, k2 = 5 x 108 M"1 sec'1
pH 6
* pH 7
A pH 8
O pH 9
Model pH 6
- Model pH 7
- -Model pH 8
Model pH 9
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time (seconds)
Figure 6.2.3
Phen Iron complex concentration vs. time graph for pH dependent kinetics with
modeled rate curves, 1 sec. time scale. Data collection rate as figure 6.2.1.

Figure 6.2.4
Phen Iron complex concentration vs. time graph for pH dependent kinetics,
with modeled rate curves, 2.5 sec. time scale. Data collection rate as figure 6.2.1.

A modeling computer program known as ACUCHEM was used to fit the
experimental data38. For purposes of modeling the rates, the initial iron
concentration was determined to be 5.8 x 10'5 M and the proton varied between
1.0 x 10"6 M and 1.0 x 1 O9 M; the concentration of 1,10-phenanthroline was 0.01
M. The rate constants ki and k2 were adjusted in order to achieve a best fit to the
experimental data. The model used to calculate k2 assumed that the solution was
perfectly buffered, i.e. pH remains constant throughout the entire iron release
process. A first order rate constant for the coupled electron transfer and iron
migration step (ki) best fit the experimental data while a bimolecular rate constant
best described the pH dependent portion (k2) of the model.
There is a noticeable lag in the response immediately after the voltage is
stepped. This lag is independent of pH. The lag for each pH is about 50 ms (see
figure 6.2.5). This is most likely a combination of three distinct processes. The
first being the time needed to reach reducing voltage, the potentiostaf s RC time
constant. For the Omni 90 potentiostat used the RC time for a voltage step of
+0.050 V to 1.300 V was measured to be 182 ps. The RC time constant being
the primary cause of the measured lag is highly unlikely. The second explanation
of the observed lag is the electron rate transfer from the TTO electrode to the iron
core of ferritin. The electron rate transfer from ITO to the core cannot be

measured but is assumed to be fast compared to the initial time needed for the
first reduced iron atoms to migrate to the surface of the protein. This initial
migration time is the most likely cause of the lag in the data. It must be noted that
the first order rate constant (ki=2 4 sec'1) is a combination of these three
processes. H.F. Bienfait and Van Den Briel in their work with
monodehydroascorbate suggest that the lag period is caused by the necessary
formation of an active compound that is necessary to mobilize iron from the
core39. In their case the active compound was oxidized ascorbate.

Absorbance vs. Time
250 ms Time Scale
0.05 0.1 0.15 0.2
Time (seconds)
pH 6 Avg
pH 7 Avg
o pH 8 Avg
a pH 9 Avg
Figure 6.2.5
Absorbance vs. time graph for pH dependent kinetics, 250 ms time scale. 2500
data points over 250 ms collection rate.

The kinetic data collected at different time resolutions show moderate differences
in iron release rate. The source of these differences is attributed to experimental
error. Kinetic data for the different time scales was collected using freshly
prepared buffer solutions and freshly purified ferritin. It must be assumed that the
buffer solutions used for the two experiments were not identical, slight differences
in pH and ionic strength were expected. Another source of error in the
measurements was that of temperature changes that would have a significant
impact on the measured rates. The LOPLTLE cell was not thermostated which
may contribute to the temperature-induced error. The iron release rate
dependence with temperature has been documented using chemical reductants40.
As temperature is increased, the iron release rate from the protein core is also
increased. This temperature dependence is expected to hold true for this direct
electron transfer study.
The deviation of the absorbance value at each time point (triplicate
preparations, n=3) at each pH level was calculated. The sum of the deviations
was then averaged. The average deviation is shown in table 6.1 below. See
figures 6.2.3 and 6.2.4 for modeled kinetic curves overlaid on experimental data.

2.5 Sec. Time Scale Data 1.0 Sec. Time Scale Data
pH 6 3.2 x l pH 7 4.3 x 10' 1.0 x 10'5
pH 8 2.8 x 10 6.4 x 10'
pH 9 2.8 x 10 2.8 x 10'
Table 6.1
The large increase in the uncertainty of the calculated molarity for pH 6, 7,
and 8 from data collected over 2.5 sec. to that collected over 1.0 sec. can be
attributed to the total number of data points collected for each. 5000 data points
were collected for each experiment. Noise was more pronounced in the data
collected over 1.0 sec leading to errors on average approximately 2.2 times that of
data collected over 2.5 sec. The error increases roughly corresponds to the
decrease in total data collection time. Error was identical for data collected at
pH 9. Data collected at pH is not valid for reasons discussed below.
The calculated rate constants do correlate well to the experimental data at
pH 6, pH 7, pH 8 but not for pH 9 (See figures 6.2.3 and 6.2.4). The large
deviation between the kinetic models and pH 9 data can be explained in terms of
buffering capacity. As reduction is taking place in the protein core and iron is

exiting hydrogen ions must enter the core to balance the negatively charged core.
TRIS, pKa of 8.08 (Ka=8.41 x 10'9) is a weak acid and therefore works better as a
buffer at low pH values. At pH 9 the ratio of [H +] to Ka is 0.12 making it
ineffective at this pH. An effective buffer has a [H+] to Ka ratio of close to 1.
This is an indication that TRIS does not have the necessary if ions to donate to
maintain the solution pH as electrons are released from the core. As a result
solution pH increases and the protein core becomes charged. The increase in
solution pH and the core charging manifests itself in a reduction of the rate iron
release from the core. The rate of iron release at pH 9 cannot be accurately
described using a TRIS buffer system. The poor buffer capacity of TRIS at pH 9
results in kinetic curves that cannot be used in the explanation of iron release
kinetics. The rise in pH resulting from a poorly buffered system explains the
differences in modeled data (which assumes constant pH) and experimental data.
The dependence of iron release rate from ferritin protein to 1,10-Phenanthroline
with initial protein iron concentration has been previously examined. As the iron
content of the protein is increased the rate of release is decreased33. In these
studies the iron concentration, or loading, of the protein was 1954 iron atoms per
ferritin molecule. In all measurements the same lot of ferritin was used and
purified by the steps discussed in Section 5.1.

7. Conclusion
The close correlation between the total amount of iron released between
electrochemical and spectroscopic techniques has proven the following:
(1) One iron atom is reduced for each electron transferred.
(2) Iron does indeed adsorb at monolayer coverages at ITO.
(3) Iron is released from the ferritin core upon application of sufficient
negative potential in the presence of 1,10-phenanthroline.
The three points above were previously assumptions based upon electrochemical
results only. Through the use of the LOPLTLE cell the assumptions were proven
to be correct. If the iron reduced / electrons transferred ratio was not 1:1 the total
amount of iron measured from voltammetry results and spectroscopic results
would not have correlated. The total ferritin measured agrees well with a
theoretical monolayer amount. All iron is released from the protein core upon
electrochemical reduction, if it was not the total iron results measured using
voltammetry would have been significantly larger than the iron measured by
UV/Vis spectroscopy.
The LOPLTLE provided the means in which to measure kinetic iron
release rates. Kinetic rates measurements were greatly simplified through by
using direct electron transfer. No chemical reductant was used; therefore,
proximity, orientation and concentration effects did not influence the measured

rates. Previous rate determining methods measured the iron release rate at
equilibrium; direct electron transfer creates a nonequilibrium state in which the
iron release rates were measured. The sigmodial rate curve obtained using direct
electron transfer is consistent with the sigmodial rate curves using homogenous
reductants indicating that direct electron transfer is a viable method to investigate
iron release rates of ferritin.

1 L. Que, Jr., ed, Metal Clusters in Proteins, American Chemical Society, 1988,
Kaim, Schwederski, Biomorganic Chemistry: Inorganic Elements in the
Chemistry of Life, John Wiley & Sons 1994, 165-166
3 E.C. Thiel, Ann. Rev. Biochem. 1987, 56, 289
4 (a) P.M. Harrison, S.C. Andrews, P.J. Artymiuk, G.C. Ford, J.R. Guest, H.
Hirzmann, D.M. Livingstone, J.M.A. Smith, A. Treffry, S.J. Yewdall, Adv. Inorg.
Chem. 1991, 36, 449 (b) A.S. Pereia, P. Tavares, S.G. Lloyd, D. Danger, D.E.
Edmondson, E.C. Theil, B.H. Huynh, Biochem. 1997, 36, 7917-7927
5 E.C. Theil, H. Takagi, G.W. Small, L. He, A.R. Tipton, D. Danger, Inorganica
Chimica Acta 2000, 297, 242-251. The figure is from
http ://biochem. ncsu. edu/faculty/theil/theil. htm
6 http ://www. chemistry.wustl. edu/EduDev/LabTutorials/Ferritin/Ferritin.html
7 (a) B. Xu, N.D Chasteen, J. Biol. Chem. 1991, 266, 1996 (b) G.C. Ford, P.M.
Harrison, D.W. Rice, J.M.A. Smith, A. Treffery, J.L. White, J.J. Yariv, Phil.
Trans. R. Soc. Lond. 1984, B304, 551-565
J. Feder, I. Giaever, Journal of Colloid and Interface Science, 1980, 78 number
1, 144
9 (a) Alaa-Eldin F. Nassar, W.S. Willis, J.F. Rusling, Anal. Chem., 1995, 67,
2386-2392 (b) I. Taniguchi, K, Wantanabe, M. Tominaga, F.M. Hawkridge, J.
Electroanal. Chem., 1992, 333, 331-338
10 K.B. Koller, F.M. Hawkridge, H.N. Blount, J. Electroanal. Chem. Interfacial
Chem., 1984, 161, 355-376

11 (a) Moon-Son Pyon, R.J. Cherry, A.J. Bjomsen, D.C. Zapien, Langmuir, 1999,
15, 7040-7046 (b) RJ. Cherry, AJ. Bjornsen, D.C. Zapien, Langmuir 1998, 14,
1971-1973 (c) D.C. Zapien, M.A. Johnson, J. Electroanal. Anal. Chem., 2000, 0,
12 T. Kuwana, R.K. Darlington, D.W. Leedy, Anal. Chem. 1964, 36, 2023-2025
13 A.T. Hubbard, C.F. Anson, Anal. Chem. 1964, 36 vol. 4, 723
14 (a) M.P. Soriaga, P.H. Wilson, A.T. Hubbard, J. Electroanal. Chem. 1982, 142,
317 (b) M.P. Soriaga, A.T. Hubbard, J. Am. Chem. Soc. 1982, 104, 3937 (c) M.P.
Soriaga, E. Binamira-Soriaga, A.T. Hubbard, J.B. Benziger, K.-W.P. Pang, Inorg.
Chem. 1985, 24, 65
15 E.A. Blubaugh, AM. Yacynych, W.R Heineman, Anal. Chem. 1979, Vol. 51
no. 4, 561-565 (b) E.F. Bowden, F.M. Hawkridge, J. Electroanal. Chem. 1981,
125, 367-386
16 R. Cieslinski, N.R. Armstrong, Anal. Chem. 1979, Vol. 51 no. 4, 565-568
17 S.C. Paulson, C.M. Elliott, Anal. Chem. 1996, 68, 1711-1716
18 (a) Y. Zhangyu, G. Tiande, Q. Mei, Anal. Chem. 1994, 66, 497-502 (b) Y. Gui,
S.A. Soper, T. Kuwana, Anal. Chem. 1988, 60, 1645-1648 9 (c) N.J. Simmons,
M.D. Porter, Anal. Chem. 1997, 69, 2866-2869 (d) M.J. Simone, W.R.
Heineman, G.P. Kreishman, Anal. Chem. 1982, 54, 2382-2384 (e) K.A.
Rubinson, H.B. Mark, Jr., Anal. Chem. 1982, 54, 1204-1206
19 J. Zak, M.D. Porter, T. Kuwana, Anal. Chem. 1983, 55, 2219-2222
20 M.D. Porter, T. Kuwana, Anal. Chem. 1984, 56, 529-534
21 J.L. Anderson, Anal. Chem. 1979, 51, 2312
22 T.R. Nagy, J.L. Anderson, Anal. Chem. 1991, 63, 2668-2672
23 J.D. Brewster, J.L. Anderson, Anal. Chem. 1982, 54, 2560-2566

24 B.E. Conway, H. Angerstein-Kozlowska, W.B.A. Sharp, E.E. Criddle, Anal.
Chem. 1973, 45, 1331 '
25 Bradford, M.M. Anal. Biochem. 1976, 72, 248-54
26 T.G. Hoy, P.M. Harrison, M. Shabbir, I.G. Macara, Biochem. J. 1974, 137, 67-
27 Deb C.C, HazraD.K., Lahiri S.C., Z. Phys. Chem. (Leipzig) 1989, 270(4), 777-
28 (a) T.S. Lee, I.M. Kolthoff, D.L.Leussing, J. Am. Chem Soc. 1950, 70, 3596-
3600 (b) I.M. Kolthoff, D.L. Leussing, T.S. Lee, J. Am. Chem Soc. 1950, 72,
29 A. Trefffy, Z. Zhao, M.A. Quail, J.R. Guest, P.M. Harrison, Biochem. 1997, 36,
30 Alfred A. Schilt, Analytical Applications of 1,10-Phenanthroline and Related
Compounds, Pergamon Press, 1969, p. 1-4
31 Harrison P.M., Arosio, P., Biochemica Et Biophysica Acta 1996, 1275, 161-
32 Macara, I.C., Hoy, T.G., Harrison, P.M. Biochem. J. 1972, 126, 151-162
33 Harrison, P.M., Hoy, T.G., Macara, I.G., Hoare, R.J., Biochem. J. 1974, 143,
34 Jones, T., Spencer, R., Walsh, C. Biochemistry 1978, 17, 4011-4016
35 Jacobs, D.L., Watt, G.D., Frankel, R.B., Papafthymiou, G.C. Biochemistry
1989, 28, 1650-1655
36 F. Funk, J.P. Lenders, R.R. Crichton, W. Schneider, Eur. J. Biochem 1985, 152,
37 Moore, G.K,Kadir, F.H, Al-Massad F., J. Inorg. Bio. Chem., 1992, 47, 175-181

38 W. Braun, J.T. Herron, D.K. Kahaner, Int. J. Chem. Kinetics 1988, 20, 51-62
39 H.F. Bienfait, Van Den Briel, Biochimica et Biophysica Acta 1980, 631, 507-
40 M. WagstafF, A. Jacobs, The Biochemistry and Physiology of Iron, Elsevier
North Holland, 1982, 463-471