Use of Câ‚„ reversed-phase HPLC as a tool for peptide map analysis of proteins

Material Information

Use of Câ‚„ reversed-phase HPLC as a tool for peptide map analysis of proteins
Hammack, Barbara N
Publication Date:
Physical Description:
ix, 78 leaves : illustrations ; 29 cm


Subjects / Keywords:
Ferritin ( lcsh )
Peptides ( lcsh )
Ferritin ( fast )
Peptides ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 74-78).
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Science, Department of Chemistry
Statement of Responsibility:
by Barbara N. Hammack.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
31159851 ( OCLC )
LD1190.L46 1994m .H35 ( lcc )

Full Text
Barbara N. Hammack
B.S., Colorado State University, 1972
M.A., Oakland University, 1974
B.S., University of Colorado, 1986
A thesis submitted to the
Faculty of the Graduate School of 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
Barbara N. Hammack
has been approved for the
Department of
4T- V- U

Hammack, Barbara N. (M.S., Chemistry)
Use of C4 Reversed-Phase HPLC as a Tool for Peptide Map
Analysis of Proteins
Thesis directed by Assistant Professor Corinne Campbell
While the ubiquitous iron storage protein cellular
ferritin has been extensively studied and well
characterized, relatively little is known about
glycosylated heart ferritin and serum ferritin. It has
been hypothesized that the later two proteins have
similar or identical primary structures, which may or may
not be different from cellular ferritin. Peptide mapping
is a powerful tool for the analysis of proteins. One
goal of this project was to develop efficient conditions
for the peptide map analysis of native proteins and
peptides of varying hydrophobicity. A system was
developed utilizing a C4 reverse phase column with a
linear gradient from 0% to 50% acetonitrile for fifty
minutes. Peptide map analysis, using this system,
revealed that glycosylated heart ferritin and serum
ferritin were quite similar. However, the relationship

between these two proteins and cellular ferritin appears
This abstract accurately represents the content of the
candidate's thesis. I recommend its publication.
Corinne Campbell

I would like to thank Dr. Corinne Campbell for her
patience and continuous support and direction. In
addition, I am grateful to the other members of my
committee, Dr. Dyckes and Dr. Zapien, for their helpful
suggestions, guidance, and encouragement. I would like
to thank Linda Butcher, Peter Clemens, Hai Dinh, Meg
Langfur and Josh Saliman for their hard efforts. The
support offered by Denise Kent and Jimmy Yee from Waters
is very much appreciated. I am indebted to Rhondda
Wells, Jayne Graham, and Aubrey Robinson for their
support, encouragement, and tolerance during the
completion of this project.
This project was funded by a National Institute of
Health grant, GM 47658.

ABBREVIATIONS ...................................... viii
1. INTRODUCTION ...................................... 1
Cellular Ferritin .................... . 1
Structure of Cellular Ferritin .... 1
Function of Cellular Ferritin .... 6
Regulation of Cellular Ferritin
Synthesis .......................... 9
Serum Ferritin............................... 10
Structure of Serum Ferritin ............ 11
Clinical Significance of Serum
Ferritin.......................... 12
Function of Serum Ferritin ............. 15
Heart Ferritin............................... 15
Peptide Map Analysis of Proteins ............ 18
High Performance Liquid Chromatography . 21
2. RESEARCH OBJECTIVE ............................... 25
3. EXPERIMENTAL..................................... 26
Materials................................... 26
Instrumentation ............................ 27
Assay Methods............................... 2 8

HPLC Analysis of Native Holoproteins . 2 8
Preparation of and HPLC Analysis of
Denatured Proteins ............ 29
Trypsin and Chymotrypsin Digestion of
Native Proteins .................. 29
Trypsin and Chymotrypsin Digestion of
Denatured Proteins ............... 31
4. RESULTS........................................... 32
Native and Denatured Holoproteins ........... 32
Trypsin and Chymotrypsin Peptide
Maps.............................. 39
Cytochrome c...................... 39
Lysozyme.......................... 41
Bovine Serum Albumin.............. 45
Myoglobin......................... 48
Peptide Map Analysis of Ferritin ............ 48
Cellular Ferritin ..................... 48
L-subunit.............................. 52
Glycosylated Heart Ferritin ........... 55
Serum Ferritin......................... 58
5. DISCUSSION........................................ 63
6. FUTURE STUDIES.................................... 72
REFERENCES............................................. 74

Amino Acids:
Con A
Aspartic acid
Glutamic acid
Atrial natriuretic factor
Bovine serum albumin
complementary deoxyribonucleic acid
Concanavalin A
Endoplasmic reticulum

HoSplFt Horse spleen ferritin
HPLC High performance liquid chromatography-
mRNA messenger ribonucleic acid
RNA ribonucleic acid
TFA Trifluoroacetic acid
TPCK L-l-tosylamido-2-phenylethylchloromethyl

Cellular Ferritin
The cellular ferritins are a family of multi-subunit
proteins found in most mammalian cells as well as
invertebrates, higher plants, fungi, and bacteria.
Cellular ferritin was first crystallized by Luafberger in
1937, and six years later the first X-ray diffraction
photographs were reported by Fankuchen(1). The primary
functions of ferritin include the detoxification and
storage of iron. In different cell types, ferritin
serves specialized functions, such as short and long term
iron storage, as well as intracellular housekeeping
functions, including iron reserve for cytochromes,
nitrogenase, ribonucleotide reductase, and detoxification
in the presence of excess iron (2).
Structure of Cellular Ferritin
Cellular ferritin is composed of a spherical protein
coat (apoferritin) with a molecular weight approximately
450,000 which surrounds a core of hydrous ferric oxide

[Fe(III)0-OH]. The iron core can contain up to 4500 iron
atoms as well as variable amounts of phosphate(2).
Apoferritin is composed of 24 subunits (Fig. 1.1).
Two subunit types known as L (19,766 Da) and H (21,099
Da) comprise the protein shell in varying proportions
thus allowing for several different isoferritins(2).
Ferritin is found mainly in the liver, spleen, and bone
marrow, although it is also present in the mucosal cells
of the small intestine, reticulocytes, placenta, testis,
kidney, skeletal and heart muscle. On the basis of
different electrophoretic mobilities, immunodiffusion,
and amino acid composition, different ferritins have been
demonstrated in different organs within the same
species(3,4). Liver and spleen ferritin contain mainly L
subunit, whereas heart, placenta and brain ferritin may
contain over 50% H subunit(1).
Examination of human heart and liver ferritin
revealed subunits which differed in electrophoretic
mobility in denaturing gels. Initially, these subunits
were designated H(eart) and L(iver). As more ferritins
were characterized, the terminology changed to H(eavy)
and L(ight). The L subunit, sometimes referred to as the

Figure 1.1. Apoferritin consists of twenty-four subunits
with a 4-3-2 symmetry composed of varying ratios of L and
H chains (adapted from Pauline M. Harrison (1)).

basic isoferritin, is associated with a relatively higher,
pi and higher iron content. The H subunit is
characterized with a relatively lower pi and lower iron
content(2). Despite the differences between the amino
acid sequences of the two types of subunits, there is
experimental evidence that the secondary structures of El-
rich and L-rich ferritins are virtually identical(5).
Horse spleen ferritin (HoSplFt) crystallized from 40
mM CdS04 is characterized as an octahedral of cubic
symmetry, composed of approximately 85% L subunits (1).
Each subunit contains 174 amino acids which fold in such
a way as to form two pairs of alpha-helixes joined by a
long loop, with a tightly packed core of hydrophobic
residues (Fig. 1.2). Two of the long helices, A and C,.
appear to be on the outside of the protein, while helices
B and D are on the inside. Numerous inter-subunit salt
bridges, hydrogen.bonds, and hydrophobic interactions to
a large degree account for the thermostability of the
protein. Eight hydrophilic channels are lined with the
conserved residues Asp 127 and Glu 130. These
hydrophilic channels are formed by the interactions of
three subunits near the amino end of each subunit. It is

Figure 1.2. Secondary structure of ferritin L chain
(adapted from Pauline M. Harrison (1)).

thought that these channels are of prime importance in
the transport of metal into and out of the protein. Six
hydrophobic channels with a fourfold axis of symmetry are
present and are lined with residues such as Leu 165(1,2).
Function of Cellular Ferritin
Although one of the major functions of ferritin
relates to iron storage, it also binds several nonferrous
metal ions such as copper, zinc, vanadium, aluminum,
cadmium, beryllium, and terbium. Substantially lesser
amounts of these metal ions bind to apoferritin than
holoferritin, possibly owing to the high affinity of
ferritin for iron(6).
With respect to the detoxification of iron, it has
been reported the ferritin H-chain contains a ferroxidase
center (7). It would appear the ferritin H-chain contains
an iron oxidation site composed of seven conserved
residues responsible for the formation of Fe(III)-oxo-
bridged dimer (8).
Normal oxidative metabolism results in the
generation of free radicals which subsequently can alter
the integrity of macromolecules such as proteins, lipids,
and nucleic acids(6). These oxidation reactions require a

catalyst, such as iron. It has been found that the
superoxide anion reductively mobilizes ferritin iron,
resulting in oxidative damage (Fig. 1.3)(6). In addition
to the superoxide anion, there is a growing list of
compounds, such as paraquat, adriamycin and alloxan,
which have been reported to release iron from
At physiological pH, ferrous iron (Fell)is highly
toxic, as discussed above. However, in the presence of
dioxygen, ferritin is formed when apoferritin is
presented with Fe (II). It is thought Fe(II) enters
through the hydrophilic channels and is oxidized to
Fe(III) as it enters the ferric oxyhydroxide center
(2,10,11) .
It has been reported that the H subunit of ferritin
has an inhibitory effect on cell growth and proliferation
(12). Ferritin binding sites specific for the H-subunit
have been found present on cells expressing the
proliferation markers HLA-DR, MLR3, interleukin 2 (IL2),
and transferrin receptors. It has been postulated the H-
subunit binding sites found on proliferating lymphocytes
serve as proliferation markers with the function of

Figure 1.3. The Fenton reaction in which the oxidation
of iron results in the generation of a peroxide radical

downregulating proliferation (13,14). This regulatory-
effect would appear to be dependent on the ferroxidase
activity of the subunit. L subunits have not been shown
to suppress cell proliferation. It has been postulated
that by catalyzing the reaction of ferrous to ferric
iron, H subunits may prevent transferrin from
contributing its iron to cells. The iron in transferrin
is in the ferric form, and is donated to cells in the
ferrous form (15). Thus, H-subunit rich ferritin may be
considered a regulatory cytokine whose function is to
downregulate cell proliferation.
Regulation of Cellular Ferritin Synthesis
The regulation of cellular ferritin synthesis occurs
at both the translational and transcriptional level. The
51 untranslated regions of the L subunit and H subunit
mRNA contain a very similar 28 nucleotide sequence (iron-
responsive element, IRE) which regulates translation by
binding to a repressor identified as cytosolic aconitase
(16). The mechanism by which it senses the level of
cellular iron revolves around an iron-sulfur cluster
(16,17,18,19). When the level of cellular iron is low,

aconitase binds to a stem-loop structure in which six
nucleotides provide the single-stranded loop, thus
inactivating mRNA translation. When iron levels are
increased, aconitase has a reduced affinity for the IRE
and dissociates allowing the initiation of translation
(16,17,18,19). Additionally, Campbell et al. have
demonstrated iron-dependent ferritin synthesis in rat
liver and spleen in response to inflammation(20).
During iron uptake, transcriptional regulation of
ferritin synthesis is also evidenced. There appears to
be a preferential increase in L subunits which form
apoferritins with a high L subunit population. These L
subunit rich protein shells are associated with a greater
capacity to sequester iron (17).
Serum Ferritin
Ferritin is localized primarily intracellularly,
however small amounts are present in serum. While
cellular ferritin has been widely studied and well
characterized with respect to structure and synthesis,
little is known about serum ferritin. Serum ferritin is
immunologically similar to liver and spleen ferritin, and

detected with antibodies to L subunit rich ferritins
(21). The presence of serum ferritin was first reported
by Reissman and Dietrich in 1956. In a precipitation
assay using rabbit anti-human ferritin, ferritin was
found in serum from patients with hepatic necrosis.
Following the development of a more sensitive assay for
ferritin, it was also detected in normal serum (21).
Structure of Serum Ferritin
Serum.ferritin has been characterized with a lower
iron content than cellular ferritin. In patients with
iron overload, serum ferritin contained 0.02-0.07 fxg of
Fe/pig of protein compared, with >0.2 /xg of Fe//xg of
cellular ferritin in the liver and spleen(21). In
addition, serum ferritin has also been characterized with
a glycosylated subunit. Worwood has reported that
approximately sixty percent of ferritin from normal serum
binds to Con A, a plant lectin that binds glycoproteins
containing mannose and/or glucose residues (21). Cellular
ferritin from the liver, spleen or heart shows little or
no binding to Con A indicating a lack of
glycosylation(21). Furthermore, acidic isoferritins from

serum are converted to basic isoferritins when incubated
with neuraminidase ostensibly owing to the removal of
sialic acid residues, whereas the isoelectric point of
cellular ferritins is unchanged (22) Serum ferritin
from patients with idiopathic hemochromatosis, contains a
subunit, named G, which stains for carbohydrates. The G
subunit had a higher molecular weight of 23,000 (22).
Santambrogio et al. (23)found that the G subunit bound to
Concanavalin A and was recognized by anti ferritin L
subunit monoclonal antibodies, but not by anti ferritin H
subunit antibodies. Thus, it appears the G subunit is
immunologically similar to the L subunit (23).
Clinical Significance of Serum Ferritin
While many characteristics of serum ferritin remain
to be elucidated, its clinical importance may outweigh
that of the extensively studied and well characterized
cellular ferritin. Numerous methodologies have been
developed to assay serum ferritin including
radioimmunoassay, immunoradiometric, and enzymatic assays
(24). Serum ferritin concentrations have been used to
examine the level of iron stores in the body. Serum

ferritin concentrations typically average between 15 and
300 /ig/L in adults, although the values are somewhat
lower in women owing to iron loss as a result of
menstruation and childbirth(21). Serum ferritin
concentrations have also been utilized to reflect
pathological states. Clinical cases of idiopathic
hemochromatosis are typically characterized by elevated
serum ferritin concentrations ranging from 500 to 5000
/ig/L (21). Borecki et al. have demonstrated the efficacy
of using serum ferritin concentrations to differentiate
hemochromatosis homozygotes from other genotypes (25).
Patients with iron overload resulting from blood
transfusions are also characterized with elevated serum
ferritin concentrations(21). Iron overload is also
clinically significant in patients with homozygous beta
thalassemia. In one study by Olivieri et al.f reduction
of tissue iron stores and normalization of serum ferritin
(from 2174 /ig/L to 251 .jug/L) followed therapy with an
oral chelator (LI)(26).
In Hodgkin's disease, serum ferritin concentrations
have been correlated to the disease stages (21) In a
study following fifty children with newly diagnosed

Hodgkin's disease, a linear increase in serum ferritin
levels with advancing stages of the disease was
demonstrated, indicating relationship between prognosis
and serum ferritin concentrations.
Prediagnostic, value for serum ferritin levels have
also been demonstrated. Using stored serum samples from
a fixed population in Hiroshima and Nagasaki, Japan, an
elevated risk for stomach cancer was associated with low
antecedent serum ferritin levels (27).
Smith et al. have investigated serum ferritin as a
potential tumor marker for renal cell carcinoma (28). In
thirty-two patients with renal cell carcinoma, the serum
ferritin levels were evaluated pre and postoperatively.
A statistically significant increase in serum ferritin
levels with advancing disease stage, and well as a
significant decrease in serum ferritin levels after
nephrectomy for stages 1 and 2 disease was demonstrated
(29). Correlations have also been found between serum
ferritin levels and the stage of neuroblastoma (28).

Function of Serum Ferritin
An inverse relationship is typically demonstrated
between serum ferritin and transferrin levels. Poor
survival was associated with decreased transferrin and
increased serum ferritin. While all cells require iron
for growth, it appears malignant cells grow faster and
require iron more than normal cells. The major source of
iron to cells for growth in transferrin. Cells produce
membrane bound transferrin receptors and capture iron.
When malignant cells express an increased number of
transferrin receptors, large amounts of transferrin are
bound, increasing the concentration of transferrin in the
vicinity. This leads to an overall decrease in the serum
concentration. Thus, decreased transferrin and increased
ferritin may be associated with a poor prognosis owing to
more malignant tumors (30,31).
Heart Ferritin
The main function of the heart as a pump providing
oxygenated blood to the vascular system has been
understood for centuries. In 1628, William Harvey's
"Essay on the Motion of the Heart and the Blood in

Animals" was published describing this muscular organ.
However, more recently it has been discovered the heart
is also an endocrine gland which secretes a peptide
hormone called atrial natriuretic factor (ANF) (33). This
hormone plays an important role in the regulation of
blood pressure and blood volume as well as in the
excretion of water, sodium and potassium. Its effects
are exerted widely, including the blood vessels, kidneys,
adrenal glands and a large number of regulatory regions
in the brain. In 1956, Kisch noted the presence of dense
bodies in cardiocytes of the guinea pig atria. Later, it
was reported that these dense bodies found in muscle
cells were present in all mammalian atria, including
humans. In 1974, at the University of Montreal, it was
observed these granules were very similar to storage
granules seen in endocrine cells. It was later shown by
Pierre-Yves Hatt and his colleagues that the number of
granules in the atrial cardiocytes increased when the
amount of sodium in an animal diet was reduced, implying
the granules store some substance pertaining to sodium
balance(32). In 1981, homogenized rat atria was injected
into rats and a rapid and massive diuresis and

natriuresis was observed(32). In 1983, the peptide was
isolated and purified by Cantin and Genest (32,33).
In 1989, Campbell et al. demonstrated that ferritin
mRNA was present on both endoplasmic reticulum bound
polyribosomes and free polyribosomes in rat heart (34).
ER bound polyribosomes and cytoplasmic polyribosomes were
quantitatively separated and examined for ferritin mRNA
content. Ferritin mRNA was found on both free and bound
polyribosomes. Significant amounts of ferritin mRNA were
found on bound polyribosomes which could not be accounted
for by cross contamination studies. ANF cDNA hybridized
to ER bound ribosomes mRNA but not with liver RNA.
These data suggest that heart cells are competent to
synthesize a ferritin destined for secretion (34).
Recently, Campbell et al. have reported the
isolation and characterization of a glycosylated heart
ferritin from horse heart. It was found to specifically
bind to Con A, indicating a glycoprotein with high
mannose or glucosyl side chains. Based on electron
microscopy, the glycosylated heart ferritin was shown to
be smaller and similar in size to serum ferritin (3-5 nm
diameters) as opposed to cellular ferritin (10 nm). It

was immunologically cross reactive with antibodies raised
against, both horse cellular ferritin and horse serum
ferritin as shown by western blot analysis. Antibodies
raised against horse serum ferritin cross reacted with
the glycosylated heart ferritin, but not with cellular
ferritin, indicating antigenic determinants shared with
serum ferritin.and glycosylated heart ferritin, but not
with cellular ferritin. Unlike cellular ferritin, which
contain only 21 kDa subunits, the glycosylated heart
ferritin was shown to be composed of subunits of 66,000,
60,500, 53,500, 43,400, and 29,500 MW by SDS-PAGE and
western blot analysis. Serum ferritin was composed of
similar subunits. The glycosylated heart ferritin had an
iron content similar to serum ferritin, and far less than
that of cellular ferritin, as determined by flame atomic
absorption spectroscopy(35).
Peptide. Map- Analysis of Proteins
One of the most powerful and commonly used
techniques for protein analysis is a peptide map. A
peptide map refers to the separation and visualization of
peptide fragments. This analytical technique is widely

used for the identification, characterization, and
structural examination of proteins, and for the isolation
of peptide fragments in the enzymatic digests (36,37).
The first step of this process involves the specific
cleavage of a protein by enzymatic or chemical means.
Enzymatic methods utilize various enzymes such as
trypsin, chymotrypsin, and elastase, which differ in
specificities(Table 1.1). Once the digestion is
complete, the resulting fragments are separated by a
method such as liquid chromatography to yield a peptide
map. This separation is analogous to a fingerprint in
that it is unique for every protein (36). Thus, if two
proteins have identical primary sequences, then digestion
of each protein with a specific endopeptidase should
yield identical peptide fragments(38,39). Because of its
high resolution and selectivity, reversed-phase liquid
chromatography is the primary mode to accomplish this
separation (discussed below)(36,40,41). The standard
procedure employs gradients using aqueous TFA as the
starting solvent with increasing concentrations of an
organic solvent such as acetonitrile in the presence of

Table 1.1.
Endopeptidases and cleavage specificities.
Endooeotidase Peotide bonds cleaved
Trypsin Lys-X, Arg-X
Chymotrypsin Phe-X, Tyr-X, Trp-X
Elastase Ala-X, Gly-X
Peptide bonds are cleaved on the carbonyl side of each
amino acid listed.

trifluoroacetic acid (below). It is performed at a pH of
approximately 2, where the carboxyl functional groups are
protonated. Additionally, TFA serves as an ion pair to
the positively charged amino functions (42,43,44).
High Performance Liquid Chromatography
High performance liquid chromatography is a rapid
separation technique with wide sample applicability.
When molecules are injected to the top of a
chromatographic column, they are subjected to two
opposing forces. One force is the mobile phase which
carries the molecules through the column, while the
stationary phase tends to retain them. As the molecules
travel through the column, a dynamic equilibrium is
established as they pass between the mobile and
stationary phases, eventually emerging separately at the
end of the column. There are several different modes of
separation (45)
Adsorption chromatography typically utilizes a polar
stationary phase such as silica gel, and a non-polar
solvent such as hexane or dichioromethane. The surface
of the silica gel consists of a random distribution of

surface silanol (Si-OH) groups. Separation occurs as a
result of hydrogen bonding between the silanol groups on
the silica gel and solute molecules. Molecules capable
of a high degree of hydrogen bonding are more strongly
retained, thus solutes are eluted in order of increasing
polarity. This is also referred to as normal phase or
liquid-solid chromatography. It is particularly
applicable for the separation of compounds highly soluble
in organic solvents (45,46,47).
Reversed phase chromatography is by far the most
popular chromatographic technique used today, with at
least 60% of all analytical separations utilizing this
mode. The term reversed phase was originally coined by
Howard and Martin in 1950. They carried out liquid-
liquid chromatography of a stationary phase of paraffin
oil and n-octane with an aqueous mobile phase. This was
the reverse, of the conventional method which employed a
polar stationary phase and non polar or less polar mobile
Stainless steel, columns 50-300 mm in length with an
internal diameter of 4-5 mm are typically used to pack
the stationary phase. The stationary phase is typically

spherical silica particles chemically bonded to an
alkylsilyl compound (49). The most popular stationary
phase is the C18 type in which octadecylsilyl groups are
bonded to the silica surface, although silica with C22,
C8, C4, C3, C2, and Cl are also available (46).
An aqueous mobile phase is used, typically a
water/methanol or water/acetonitrile solution(45,50). In
general, solute retention takes place because of
hydrophobic interactions between the molecules and the
chemically bonded hydrocarbon groups on the stationary
phase. Very polar solvents favor retention while
nonpolar mobile phases such as acetonitrile diminish it
(50) .
Reversed phase chromatography tends to promote
protein denaturation. This stems from the use of
hydrophobic interactions on the stationary phase as well
as organic modifiers in the mobile phase. Biologically
active proteins exist as tightly coiled chains with most
of the polar groups oriented toward the outside where
they interact with the aqueous environment, and the non-
polar groups on the inside where they interact together.
This configuration tends to stabilize the protein and

minimize the free energy of the protein in an aqueous
solution. When the protein interacts with the stationary
phase, the minimum free energy would be achieved by
maximizing the hydrophobic interactions with the non-
polar groups, while orienting the polar groups toward the
polar mobile phase. This leads to an uncoiled,
biologically inactive conformer (50).

Reversed-phase chromatography is by far the most
popular method for analyzing peptides, where C18 columns
are typically utilized. Owing to the predominant alpha
helix structure of ferritin, one would expect the protein
to be characterized with a high degree of hydrophobicity
and therefore one would expect a great deal of difficulty
eluting such a hydrophobic molecule from this column. An
alternative is the less retaining C4 column. Therefore,
it was a goal of this project to determine the conditions
required for successful peptide mapping on a C4 reversed-
phase column in order to elucidate the structural
similarities between proteins of varying hydrophobicity.
The second goal of this research was to utilize this
system in order to determine the similarities or
differences between the primary structures of cellular
ferritin, glycosylated heart ferritin, and serum
ferritin. This structural study will further our
understanding of the functional relationship between
glycosylated heart ferritin and serum ferritin.

Horse heart Cytochrome C and myoglobin from horse
skeletal muscle were purchased from Calbiochem (San
Diego,CA). Bovine serum albumin, ferritin from horse
spleen, chymotrypsin, and trifluoroactic acid were
purchased from Sigma Chemical Company (St. Louis, MO).
Lysozyme, TCPK treated trypsin were purchased from
Worthington Biochemical Corporation (Freehold, NJ).
Ammonium bicarbonate was purchased from Fisher Scientific
(Pittsburgh, PA). HPLC grade water and 4mm syringe
filters (0.45 fi) were purchased from Baxter(McGaw
Park,IL). HPLC grade acetonitrile was purchased from
Fisher Scientific (Pittsburgh,PA). Microcon-10
concentrators were purchased from Amicon,Inc (Beverly,
MA). NAP-5 columns were purchased from Pharmacia Biotech
(Piscataway, NJ). Serum ferritin was purified in our
laboratory by Linda Butcher. Glycosylated heart ferritin
and L-chain ferritin were purified in our laboratory by
Hai Dinh. Purified ferritin L-chain subunit was

resuspended in 0.1 M ammonium bicarbonate buffer using a
dialysis column (Pharmacia NAP-5 column). Purified
glycosylated heart ferritin, serum ferritin, and ferritin
L-chain subunit were concentration to final
concentrations using microcon-10 concentrators.
Cytochrome c, myoglobin, bovine serum albumin, and
lysozyme were not further purified. According to
Calbiochem, the purity of cytochrome c was 98% by
spectrophotometry, and the purity of myoglobin was
greater than 98% by molecular exclusion chromatography.
The purity of bovine serum albumin was reported by Sigma
Chemical Company to be 96-99%. Sterile 0.45 /im
centrifuge filters, purchased from Fisher Scientific
(Pittsburgh, PA) were used to filter solutions prior to
assay on the HPLC.
The HPLC system was purchased from Waters. This
included Waters 501 HPLC pump, Waters U6K injector,
Waters 486 absorbance detector, and the Waters pump
control module. The initial software version 1.0 was
subsequently updated to version 1.1 and 2.0. The HPLC

system interfaced with a 386/33i NEC Powermate.
A Hi-Pore RP-304 column, 250 x 4.6 cm, particle size 5
urn, pore size 300 A was purchased from BioRad
Assay Methods
HPLC Analysis of Native Holoproteins
Native cytochrome C, bovine serum albumin, lysozyme
and myoglobin were suspended in 0.1 M ammonium
bicarbonate buffer, pH 8.0, at concentrations of 10
mg/ml, 5 mg/ml, lmg/ml, 0.5 mg/ml and and
frozen until chromatographed. Prior to HPLC analysis,
protein solutions were mixed with one volume of 0.1%
TFA/H20 (solvent A), and filtered using 4mm (0.45 /j)
filters. Ten microliter samples were applied to a C4
reverse phase column (Hi-Pore RP-304, BioRad) at a flow
rate of 1 ml/min and absorbance was detected at 214 nm.

The following gradient was used:
Time (minutes) %A solvent 0.i%tfa/h2o %B solvent 0.1%TFA/ acetonitrile
Initial 100 0
30 0 100
35 0 100
40 100 0
50 100 0
Preparation of and HPLC Analysis of Denatured Proteins
Cytochrome c, bovine serum albumin, lysozyme, and
myoglobin were suspended and diluted as above, then
boiled at 100 C, for ten minutes. The denatured proteins
were subsequently cooled to 25 and assayed as described
Trypsin and Chvmotrvpsin Digestion of Native Proteins
TCPK treated trypsin or chymotrypsin was dissolved
in 0.1 M ammonium bicarbonate buffer, pH 8.0, and diluted
to lmg/ml concentration. One volume of the trypsin or

chymotrypsin was added to 10 mg/ml concentrations of
cytochrome c, bovine serum albumin, myoglobin, and
lysozyme to produce a final trypsin/protein or
chymotrypsin/protein ratio of 10% (w/w). The solutions
were incubated for 2 hours at 25 C then trypsin or
chymotrypsin was inactivated by boiling at 100 c for 5
minutes. The digests were allowed to cool to room
temperature, then frozen until chromatographed. Each
digest was diluted with one volume of aqueous 0.1%TFA and
filtered using 4mm (0.45 pore size) syringe filters prior
to HPLC assay. The sample volume applied to the column
was 10 fih. The following linear gradient, absorbance at
214 nm, was the final gradient used to assay the digests:
Time (minutes) Flow (ml/min) %A solvent 0.1%TFA/H20 %B solvent 0.1%TFA/ acetonitrile
0.0 1.0 100 0.0
50 1.0 50 50
55 1.0 0.0 100
60 1.0 100 0.0
70 . 1.0 100 0.0

Trypsin and Chymotrypsin Digestion of Denatured Proteins
Cytochrome c, bovine serum albumin, myoglobin., and
lysozyme (10 mg/ml) were boiled at 100 C for ten
minutes, and allowed to cool to room temperature. Each
protein was mixed with one volume of trypsin or
chymotrypsin (1 mg/ml) and incubated for 2 hours.
Following incubation, trypsin or chymotrypsin was
inactivated by boiling each solution at 100 C for five
minutes. The solutions were allowed to cool to 25 C,
then frozen until chromatographed. Prior to analysis on
the HPLC, each digest was diluted with one volume of 0.1%
aqueous TFA and filtered using 4mm (0.45 pore size)
syringe filters. The same HPLC running conditions and
linear gradient were used as for trypsin or chymotrypsin
digests of native proteins (above).

Native and Denatured Holoproteins
Solute retention on a C4 reverse phase column is
primarily a function of hydrophobic interactions with the
hydrocarbonaceous stationary phase surface. Solutes are
retained in decreasing order of polarity. Figure 4.1
depicts the chromatograms and elution times in increasing
order of elution time of cytochrome c, lysozyme, BSA,
myoglobin, and cellular ferritin applied separately in
the native state to a C4 reverse phase column. L-subunit
ferritin was obtained already in the denatured state.
Cytochrome c consists of a single polypeptide chain of
104 residues(51). Based on X-ray analysis, the secondary
structure of cytochrome c is characterized with
approximately 40% alpha helix structure(52). Lysozyme is
an enzyme found in egg white which consists of a single
polypeptide chain and contains1 129 residues(53). Bovine
serum albumin consists of a single polypeptide chain of
581 amino acid residues folded in three similar

Figure 4.1. Chromatographs of native proteins on a C4
reverse phase column.

globular domains. Each domain contains three loops held
together by disulfide bonds (54). Horse myoglobin is a
heme containing proteins which consists of 153 residues.
Based on X-ray analysis, the secondary structure of
myoglobin is characterized with approximately 80% alpha
helix structure (52). Cellular ferritin was previously
described. An index of hydrophobicity for each of the
proteins was calculated based on the total number of
tryptophan, phenylalanine, valine, leucine, and
isoleucine amino acid residues per molecule (Table 4.1).
There is a direct correlation between the elution time of
native proteins and the percent hydrophobicity of each
protein. There would appear to be a correlation between
elution time of native proteins and molecular weight,
although this is less clear cut.
Figure 4.2 is an overlay chromatogram of
glycosylated heart ferritin.and the gradient blank. The
peak representing glycosylated heart ferritin eluted at
22.9 minutes.
It is difficult to determine the elution time of
serum ferritin because of the lack of a clearly
identifiable peak. The chromatograms of serum ferritin

Frotein Elution Time of Native Protein (minutes) Elution Time of Denatured Protein (minutes) %Hydrophobic Molecular Weight (daltons)
Cytochrome C 21.2 21.2 ' 19.2 12,300
Lysozyme 22.2 22.3 22.5 14,400
BSA 22.9 22.7 24.0 66,000
Myoglobin 24.0 24.1 27.4 18,800
Cellular Ft 25.8 26.1 27.7 450,000
| L-subunit Ft 25.9 28.2 19,800
% Hydrophobicity calculated as the total number Trp, Phe, Val, Leu, and
Table 4.1. Summai^r of elution times of native and
denatured proteins as a function of hydrophobicity on a
C4 reverse phase column.

0.00 5.00
io!ob 'is!oo' 20.00 25.00 30.00 35.00 40.00 45.00 50.o<
Figure 4.2 Overlay of chromatograms of glycosylated
heart ferritin (A) and the gradient blank (B).

are characterized with unexplained contaminants.
Figure 4.3 shows the chromatograms and elution times
of cytochrome c, lysozyme, BSA, myoglobin, cellular
ferritin, and L-subunit ferritin applied to a C4 reverse
phase column in the denatured state. The chromatograms
are presented in increasing order of calculated
hydrophobicity. The elution times for cytochrome c,
lysozyme, BSA, myoglobin, and cellular ferritin are
comparable to the elution times of these proteins applied
to the column in the native state. The peak height and
peak width of denatured cytochrome c, lysozyme, BSA, and
myoglobin are comparable to those in the native state.
The peak shape of denatured cellular ferritin is
asymmetrical, although the peak shape of native cellular
ferritin is symmetrical. Denatured cellular ferritin
eluted at 26.1 minutes, however L-subunit ferritin eluted
at 25.9 minutes. This would appear inconsistent based on
the calculated index of hydrophobicity.

Time (minutes)
Figure 4.3. Chromatographs of denatured proteins on a C4
reverse phase column.

Tryptic and Chymotryptic Peptide Maps
Cytochrome c
When proteins fold into their biological
conformation, the polar amino acids are located on the
outside of the protein, while the nonpolar amino acids
are sequested on the inside of the molecule. Trypsin
cleaves on the carboxyl side of the polar amino acids
lysine and arginine. Therefore, one would expect little
difference between the tryptic peptide maps in the native
and denatured state. The chromatograms of native and
denatured cytochrome c tryptic peptide maps (Fig. 4.4)
show fairly symmetrical peaks with reasonable resolution.
There is a difference between the two, however,
partcularly evident during the period of 30 to 45
minutes. The peaks are of different heights and differ
in resolution.
Chymotrypsin cleaves on the carboxyl side of the
aromatic amino acids phenylalanine, tryptophan, and
tyrosine as well as possibly large, hydrophobic residues.
Since hydrophobic residues tend to be sequestered on the
inside of a protein, denaturing the protein first may
make these hydrophobic residues more susceptible to

Figure 4.4. A comparison of native and denatured
cytochrome c tryptic peptide maps.

digestion by chymotrypsin. A comparison of native and
denatured cytochrome c chymotryptic peptide maps (Fig.
4.5) shows some similarity in the number of fragments and
peak shapes. Some of the peaks of the denatured
cytochrome c, primarily in the middle portion of the
gradient, appear to be of a higher absorbance.
Figure 4.6 shows a comparison of native and
denatured lysozyme tryptic peptide maps. There are
similarities with respect to the number of fragments and
peak heights in the first 25 minutes. However, during
the middle portion of the gradient, there are differences
in the number of fragments.
A comparison of native and denatured lysozyme
chymotryptic peptide maps is shown in figure 4.7. From 0
to 35 minutes, the fragments are similar in number and
shape, although differ somewhat in peak height. During
the portion of the gradient from 35 to 50 minutes, there
are peaks present in the chromatogram of the denatured
lysozyme which are not present in the chromatogram of the
native lysozyme.

Figure 4.5. A comparison of native and denatured
cytochrome c chymotrypsin peptide maps.

Figure 4.6. Comparison of native and denatured lysozyme
tryptic peptide maps.

Figure 4.7. Comparison of native and denatured lysozyme
chymotryptic peptide maps.

Bovine Serum Albumin
A comparison of native and denatured BSA tryptic
peptide maps is shown in Figure 4.8. Overall, there is
similarity between the two profiles with respect to a
comparable number of fragments of similar peak shape and
height. There is a fairly wide peak present in the
chromatogram of the native BSA tryptic peptide map
eluting at approximately 50 minutes which is not present
in the chromatogram of the denatured BSA tryptic peptide
map. Additionally, there are six peaks eluting between
35 and 40 minutes which appear more completely resolved
in the chromatogram of the native BSA tryptic peptide
map. Figure 4.9 shows a comparison of native and
denatured BSA chymotryptic peptide maps. Again, there is
a wide peak present in the chromatogram of the native BSA
chymotrypsin peptide map which is absent in the
chromatogram of the denatured BSA chymotryptic peptide
map. For the most part, the peaks in the denatured BSA
chymotryptic peptide map would appear more resolved from
each other than the peaks in the native BSA chymotryptic
peptide map. In addition, there would appear to be
increased peak height characterizing most of the peaks in

Figure 4.8. Comparison of native and denatured BSA
tryptic peptide maps.

Figure 4.9. Comparison of native and denatured BSA
chymotryptic peptide maps.

the denatured BSA chymotryptic peptide map relative to
the native BSA chymotryptic peptide map.
Figure 4.10 shows the tryptic peptide maps of native
and denatured myoglobin. There are striking differences
between the two chromatograms. These differences can be
seen in the number of fragments, peak heights, and the
symmetry of the peaks. Figure 4.11 shows a comparison of
native and denatured myoglobin chymotryptic peptide maps.
Again, there are striking differences between the
chymotrypsin digests of myoglobin in the native and
denatured state. The differences are again seen in the
number of peaks and peak heights.
Peptide Map Analysis of Ferritin
Cellular Ferritin
Figure 4.12 shows a comparison of native and
denatured cellular ferritin tryptic peptide maps. The
tryptic peptide map of native cellular ferritin appears
rather flat with only a few peaks, of very short height.
The tryptic map of denatured ferritin shows more evidence

Figure 4.10. Comparison of native and denatured
myoglobin tryptic peptide maps.

Time (minutes)
Figure 4.11. Comparison of native and denatured
myoglobin chymotryptic peptide maps.

Figure 4.12. Comparison of native and denatured cellular
ferritin tryptic peptide maps.

of fragments. Overall, the peaks appear somewhat
asymmetrical and incompletely resolved from each other.
A comparison of native and denatured cellular ferritin
chymotryptic peptide maps is shown in Figure 4.13. The
chymotryptic peptide map of native cellular ferritin is
remarkable for its lack of distinct peaks. The peaks
present are few in number and of very short peak height.
The chymotryptic peptide map of denatured cellular
ferritin evinces more fragments of somewhat asymmetrical
peak shape. The fairly flat tryptic and chymotryptic
peptide maps of.native cellular ferritin seem to
underscore the reported thermostability of this protein.
Figure 4.14 shows the chromatogram of the L-subunit
tryptic peptide map. During the first 5 minutes of the
gradient, two rather unresolved peaks are present
followed closely by a somewhat wide peak with an
asymmetrical shape. The majority of the peaks evidenced
are eluted between approximately 20 minutes and 50
minutes. There is a problem with many of the peaks in
that although they are distinct from one another, the

Figure 4.13. Comparison of native and denatured cellular
ferritin chymotryptic peptide maps.

Figure 4.14. Chromatogram of ferritin L-subunit tryptic
peptide map.

baseline resolution is not complete. Figure 4.15 shows
an overlay chromatogram of L-subunit and cellular
ferritin tryptic peptide maps as well as the gradient
blank. This overlay chromatogram focuses on the portion
of the gradient from 15 to 60 minutes. There appear to
be more peaks present, in the tryptic peptide map of
ferritin L-subunit than in the tryptic peptide map of
cellular ferritin. This is particularly evident in the
portion of the gradient from approximately 35 minutes to
60 minutes. One would expect the more hydrophobic
fragments to elute during this time, thus the ferritin L-
subunit would appear more susceptible to enzymatic
digestion. Since horse spleen ferritin is primarily
composed of L-subunits, one would expect a high degree of
similarity when comparing the tryptic peptide maps of the
two. While there appear to be more peaks present in the
tryptic peptide map of the ferritin L-subunit, there are
also similar peaks of comparable elution time, peak shape
and peak height present in the two chromatograms.
Glycosylated Heart Ferritin
Figure 4.16 shows the tryptic peptide map of

Figure 4.15. Comparison of ferritin L-subunit tryptic
peptide map (A), the gradient blank(B), and cellular
ferritin tryptic peptide map (C).

Figure 4.16. Chromatogram of tryptic digestion of
glycosylated heart ferritin from 15 minutes to 60 minutes
of a linear gradient.

glycoslylated heart ferritin. The data shown represent
the portion of the gradient from 15 minutes to 60
minutes. The peaks resulting from the enzymatic
digestion appear reasonably well resolved from one
another. The peaks are fairly symmetrical in shape.
There are several peaks present which seem to be
attibuted to the autolysis of trypsin. These are found
during the portion of the gradient from approximately 15
minutes to 35 minutes.
Serum Ferritin
Figure 4.17 shows the portion of the tryptic peptide
map of serum ferritin from 15 minutes to 60 minutes. As
mentioned above, some of the fragments, particularly
during the portion of the gradient from approximately 15
minutes to 35 minutes, seem to be attibutable to the
autolysis of trypsin. The peaks present seem reasonably
well resolved from one another. A portion of the peaks
are symmetrical, although there are peaks present
characterized with a shoulder.
Figure 4.18 shows an overlay of the tryptic peptide
maps of serum ferritin and glycosylated heart ferritin,

Figure 4.17. Chromatogram of trypsin digestion of serum
ferritin during the portion of a linear gradient from 15
minutes to 60 minutes.

Figure 4.18. Chromatograms of serum ferritin trypsin
digestion (A), glycosiylated heart ferritin trypsin
digestion (B), and the gradient blank (C) during the
portion of a linear gradient from 15 minutes to 60
minutes. Absorbance was limited to 0.04 absorbance

as well as the gradient blank. A comparison of the two
chromatograms indicates peaks of comparable elution time,
peak shape, and peak height present in both. There would
also appear to be unique peaks present in the two trypsin
digestions. The peaks of comparable elution time shape
and height seem to occur during the time span from 20
minutes to 30 minutes of the linear gradient. There also
seem to be similar peaks present in the time span from
approximately 35 to 40 minutes. The peaks present from
50 to 50 minutes are also present in the gradient blank.
Figure 4.19 shows an overlay of the tryptic peptide
maps of ferritin L-subunit, serum ferritin, and
glycoslylated heart ferritin, as well as the gradient
blank. The peaks present in the tryptic peptide map of
ferritin L-subunit appear to be of increased peak height,
and for the most part, of differing elution times
relative to the peptide maps of. glycosylated heart
ferritin and serum ferritin (Fig. 4.19).

Figure 4.19. Overlay of trypsin digestion of ferritin L-
subunit (L-Ft), trypsin digestion of serum ferritin
(sFt), trypsin digestion of glycoslylated heart ferritin
(GHtFt), and the gradient blank (Grad blk) during the
portion of a linear gradient from 15 minutes to 60

One goal of this project was to develop a procedure
for the adequate resolution of peptide maps. This was
accomplished using a C4 reverse phase column with a 0 -
50% acetonitrile linear gradient. The second goal of
this project was to compare the primary structure of
cellular ferritin, glycosylated heart ferritin, and serum
ferritin by examining peptide maps of each. In order to
achieve the first goal it was necessary to run a number
of proteins of known primary sequence in order to gain
knowledge with respect to their elution from the C4
column. There appeared to be a direct correlation
between the elution time of holoproteins and their degree
of hydrophobicity (Fig. 4.1).
The results of five separate trypsin digestions of
cytochrome c were reproducible (data not shown). There
are also similarities between the tryptic maps of native
and denatured cytochrome c, particularly during the
earlier portion of the linear gradient. There are two

peaks eluting near thirty five minutes which seem
inverted with respect to peak height, although the
explanation for this is not altogether clear.
Nonhydrolyzed cytochrome c eluted at 43 minutes. There
are no discernable peaks indicative of nonhydrolyzed
cytochrome c on either of the tryptic maps of cytochrome
c. Based on reproducible results from separate enzyme
digestions and the absence of nonhydrolyzed cytochrome c,
it was concluded that acceptable conditions for tryptic
peptide mapping were established.
The comparison of chymotryptic peptide maps of
native and denatured cytochrome c shows similarity
between the two with respect to the number of.fragments
with comparable elution times (Fig. 4.5). A peak
ostensibly corresponding to nonhydrolyzed cytochrome c is
evidenced in the chymotrypsin peptide map of native
cytochrome c, yet unremarkable in the chymotrypsin map of
denatured cytochrome c. This is consistent with the
substrate specificity of chymotrypsin where it cleaves
aromatic amino acid residues which one would expect to be
sequestered on the interior of the protein. By
denaturing the protein prior to proteolysis, these

hydrophobic residues would be more accessible and
therefore hydrolyzed.
An analysis of the tryptic maps of native and
denatured lysozyme immediately brings up two problems
associated with peptide map analysis in this project
(Figs. 4.6, 4.7). The first problem relates to the
failure to reduce disulfide bonds. Disulfide bonds
crosslink portions of the polypeptide backbone which are
spatially close. This widely occurring post-
translational process is particularly common among
secreted proteins. Proteins with intact disulfide bonds
tend to be somewhat resistant to proteolysis. This seems
logical in light of steric constraints on proteolysis if
the potential cleavage sites are near disulfide bonds or
within disulfide loops. This may lead to a partial
digestion, resulting in more peptide fragments than one
would expect based on known primary sequences. Hence, it
is difficult to establish the accuracy of an obtained
peptide map relative to published chromatograms or the
known primary sequence.
The presence of more peptide fragments than expected
owing to partial digestion is exacerbated by a second

problem encountered in this project. Additional
fragments ostensibly corresponding to the autolysis of
the endopeptidases trypsin and chymotrypsin were present
making the interpretation of peaks more difficult than
necessary. This problem seemed more pronounced with
respect to trypsin. The addition of 5mM CaCl2 may help
reduce the autolysis of trypsin.
While these problems were present with the tryptic
and chymotryptic maps of lysozyme, they were exacerbated
with respect to the digestions of BSA. Lysozyme is a
single polypeptide containing 129 residues with four
disulfide bonds. BSA is composed of 581 residues with 17
disulfide bonds, thus producing a complex map following
proteolysis. An analysis of the trypsin and chymotrypsin
maps for native and denatured BSA basically reveals a set
of several fragments reasonably well separated from each
other (Figs. 4.8, 4.9). The appearance of a relatively
broad peak eluting at 50 minutes in the maps of native
BSA corresponds to nonhydrolyzed BSA. This suggests that
longer incubation times are needed.
Myoglobin is a heme-containing protein composed of
153 residues. Interestingly, horse myoglobin is a rather

compact molecule which does not contain any cysteine
residues. A comparison of tryptic maps of native and
denatured myoglobin reveals quite different chromatograms
(Figs 4.10, 4.11). It seems evident changes stemming
from denaturation render the two trypsin maps of
myoglobin incomparable. Although the tryptic peptide map
of native myoglobin,reveals a set of separated fragments,
the peak eluting at 54 minutes corresponds to
nonhydrolyzed myoglobin, suggesting the incubation time
should probably be lengthened. The chymotryptic map of
native myoglobin reveals a large peak corresponding to
nonhydrolyzed myoglobin. This seems consistent with the
substrate specificity of chymotrypsin and the compact
nature of the protein. One would expect hydrophobic
residues to be rather protected from enzymatic digestion.
As previously stated, the second goal of this
project was to utilize the established conditions on a C4
reverse phase column to compare the primary structures of
cellular ferritin, glycosylated heart ferritin, and serum
ferritin. Consistent with the reported structural
differences, glycosylated heart ferritin eluted before
ferritin L-subunit. It was not possible to compare

these chromatograms with that of serum ferritin owing to
the lack of a clearly identifiable peak for serum
ferritin. These experiments were limited because of the
difficulty in purifying adequate quantities of serum
ferritin which is present in horse serum at a
concentration of only 10 ng/ml. The laboratory is
currently exploring means of improving yield and
obtaining greater volumes of horse serum.
Cellular ferritin (horse spleen ferritin) consists
primarily of L-subunits. The primary structure of L-
subunit from horse spleen contains two cysteine amino
acids, residue 48 and residue 126. It would.appear these
two cysteine residues are separated by too great a
spatial distance to.participate in the formation of a
disulfide bridge. An analysis of tryptic peptide maps of
cellular ferritin suggests substantially greater
digestion occurs with the denatured cellular ferritin
(Figs. 4.12, 4.13). This seems reasonable to expect
owing to the quaternary structure of cellular ferritin.
Subunit dissociation should render the protein more
susceptible to proteolysis. The chymotryptic maps are
consistent with greater proteolysis following

denaturation as well. The peaks eluting at 60 minutes
would appear to represent nonhydrolyzed protein,
suggesting improved peptide mapping may result following
increased incubation times.
During the purification of horse spleen L-subunit,
the specimen was boiled in an denaturing buffer
containing S-mercaptoethanol, which reduces cysteine
residues. Cellular ferritin was not treated with this
denaturing buffer. However, peptide mapping is rooted in
the fact the degree of similarity or dissimilarity
between peptide fragments reflects the degree of
similarity in the primary structure between proteins.
Horse spleen L-subunit contains only two cysteine
residues which would appear too far apart to generate a
disulfide bond. Hence, it seems reasonable that a
comparison of fragments stemming from the trypsin
digestion of L-subunit and cellular ferritin would not
result in differences owing to disulfide bonds. One
might expect a more partial digestion of the cellular
ferritin because of the hydrophobic nature of the protein
and thus perhaps incomplete subunit dissociation. A
comparison of the tryptic maps of ferritin L-subunit with

that of cellular ferritin indicates several peaks of
comparable shape and elution time suggesting similarities
in the primary structure of the two analyzed proteins
(Fig. 4.15). However, there would appear to be more
fragments present in the trypsin digestion of L-subunit,
consistent with the expected more complete digestion.
The tryptic maps of glycosylated heart ferritin and
serum ferritin were quite similar as evidenced by peaks
of comparable shape and elution times. This suggests
similarities in the primary sequences between the two
proteins (Fig. 4.18). The similarities seem most
pronounced during the earlier portion of the linear
gradient, corresponding to relatively more polar
fragments. This seems consistent with the existence of
glycosylated subunits.
The tryptic maps of L-subunit seemed to lack
similarity with the peptide maps of glycosylated heart
ferritin and serum ferritin. However, it has been
reported that glycosylated heart ferritin and serum
ferritin are cross-reactive with anti-L-subunit
antibodies, suggesting shared antigenic determinants. A
comparison of tryptic maps of glycosylated heart ferritin

and serum ferritin with H-subunit may shed some light.

The ambiguous relationship between ferritin L-
subunit and glycosylated heart ferritin and serum
ferritin could be further explored by sequencing the
peptide fragments resulting from tryptic digestion.
Additionally, it would be beneficial to compare the
tryptic digestions of glycosylated heart ferritin and
serum ferritin with ferritin H-subunit.
In addition, experiments are underway to utilize
immobilized trypsin and chymotrypsin on controlled pore
glass beads. This may provide a more expeditious method
of enzymatic digestion.
Glycosylation of subunits can be determined by the
removal of the carbohydrate side chains by
endoglycosidase H prior to analysis on the C4 column.
Specific fragments could also be evaluated for the
presence of glycosylation.
While the results of this research suggest a
similarity in the primary sequences of glycosylated heart
ferritin and serum ferritin as evidenced by similar

tryptic maps, many questions remain. Specifically,
separated peptide fragments could be analyzed for
molecular weight by thermospray mass spectrometry.
Finally, an on-line technique could be established
incorporating fragmentation by immobilized
endopeptidases, separation on a HPLC column, and
detection with thermospray mass spectrometry (55).

1. Harrison, P. M.; Ford, G. C.; Rice, D. W.; Smith, J
M. A.; Treffry, A.; White, J. L. Biochemical
Society Transactions 1987, 744-748.
2. Theil, E. C. Ann. Rev. Biochem 1987, 56., 289-315.
3. Passaniti, A.; Roth, T. F. Biochem. J. 1989, 258.
4. Crichton, R.; Millar, J.; Cumming, R.; Bryce, C.
Biochem. J. 1973, 131, 51-59.
5. Collawn, J. F., Jr.; Donato, H., Jr.; Fish, W. W.
Biochimica et Biophysica Acta 1986, 871. 235-242.
6. Joshi, J.; Clauberg, M. Biofactors 1988, 1 (3), 207
212 .
7. Levi, S.; Luzzago, A.; Cesareni, G.; Cozzi, A. ;
Franceschinelli, F.; Albertini, A.; Arosio, P. The
Journal of Biological Chemistry 1988, 263 (24),
December 5, 18086-18091.
8. Bauminger., E. R.; Harrison, P. M.; Hechel, D.;
Hodson, N. W.; Nowik, I.; Treffry, A.; Yewdall, S.
J. Biochemical Journal 1993, 296, December 15, 709-
9. Reif, D. W. Free Radical Biology & Medicine 1992,
12, 417-427.
10. Seligman, P. A.; Klausner, R. D.; Huebers, H. A. In
Molecular Basis of Blood Disease; Neinhuis A.;
Majerus G.; Stamatoyannopoulos G.; Leder,
G.,Ed.:Saunders: Philadelphia,PA,1987; pp 219-244.
11. Munro, H. N.; Linder, M. C. Physiological Reviews
1978, 58 (2), 317-396.

12. Konijn, A.; Meyron-Holtz, E. G.; Levy, R.; Ben-
Bassat, H.; Matzner, Y. FEBS Letters 1990, 263 (2),
April, 229-232.
13. Faragion, S.; Fracanzani, A. L.; Brando, B.; Arosio,
P.; Levi, S.; Fiorellil, G. Blood 1991, 78 (4), 15
August, 1056-1061.
14. Fargion, S.; Cappellini, M. D.; Fracanzani, A. L.;
De Feo, T. M.; Levi, S.; Arosio, P.; Fiorelli, G.
American Journal of Hematology 1992, 39., 264-268.
15. Broxmeyer, H. E. Journal of Laboratory Clinical
Medicine 1992, September, 367-370.
16. Klausner, R. D.; Stout, C. D.;Rouault, T. A.
Harford, J. B. Chemistry and Biology 1994, 15
April, 14-15.
17. Munro, H. N. Journal of Cellular Biochemistry 1990,
44. 107-115.
18. Theil, E. C. Journal of Biological Chemistry 1990,
265 (9), 25 March, 4771-4774.
19. Kaptain, S.; Downey, W. E.; Tang, C.; Philpott, C.;
Haile, D.; Orloff, D. G.; Harford, J. B.; Roulault,
T. A.; Klausner, R. D. Proc. Natl. Acad. Sci. USA
1.991, M, November, 10109-10113.
20. Campbell, C. H.; Solgonick, R. M.; Linder, M. C.
Biochemical and Biophysical Research Communications
1989, 160 (2), 28 April, 453-459.
21. Worwood, M. Clinical Science 1986, 70, 215-220.
22. Cragg, S. J.; Wagstaff, M.; Worwood, M. Biochem. J.
1981, 199, 565-571.
23. Santambrogio, P.; Cozzi, A.; Levi, S.; Arosio, P.
British Journal of Haematology 1987, 65, 235-237.

29 .
32 .
Dawson, D.; Fish, D.; Shackleton, P. Clin. Lab
Haemat. 1992, 14, 47-52.
Borecki, I.; Rao, D.; Yaouang, J.; Lalouel, J. Human
Heredity 1990. 40. 159-166.
Olivieri, N. F.; Koren, G.; Matsui, D.; Liu, P. P.;
Blendis, L.; Cameron, R.; McClelland, R. A.;
Templeton, D. M. Blood 1992, 79 (10), 15 May, 2741-
2748 .
Akiba, S.; Neriishi, K.; Blot, W. J.; Kabuto, M.;
Stevens, R. G.; Kato, H.; Land, C. E. Cancer 1991,
67, 15 March, 1707-1712.
Iancu, T. C.; Shiloh, H.; Kedar, A. Cancer 1988, 61
(15 June), 2497-2502.
Esen, A.; Ozen, H.; Ayhan, A.; Ergen, A.; Tasar, C.;
Remzi, F. Journal of Urology 1991, 145. June, 1134-
1137. '
Bomford, A.; Munro, H. Pathobiologv 1992, 60., 10-18.
Hann, H. L.; Lange, B.; Stahlhut, M. W.; McGlynn, K.
A. Cancer 1990, 66, 15 July, 313-316.
Cantin, M.; Genest, J. Scientific American 1986,
February, 76-81.
Rosenzweig, A.; Seidman, C. E. Annual Review
Biochemistry 1991, 60., 229-255.
Campbell, C. H.; Ismail, R.; Linder, M. C. Archives
of Biochemistry and Biophysics 1989, 273 (1), 89-98.
Campbell, C. H.; Crocker, D.; Gruntmeir, J. J.;
Head, M.; Kelly, T.; Langfur, M. I.; Leimer, A. H.
Journal of Cellular Biochemistry 1993, 53., 420-432.
Dong, M. W.; Tran, A. Journal of Chromatography
1990, 125-139.

37. Kalghatigi, K.; Horvath, C. Journal of
Chromatography 1988, 443. 343-354.
38. Judd, R. C. Methods in Enzymoloav 1990, 182. 613-
39. Hancock, W. ; Bishop, C.; Prestidge, R.; Hearn, M.
Analytical Biochemistry 1978, 89., 203-212.
40. Dong, M. W. Adv Chromatography 1992, 32., 21-51.
41. Stone, K. L.; Elliott, J. I.; Peterson, G.;
McMurray, W. ; Williams, K. R. Methods in Enzymology
1990, 193., 389-412.
42. Young, P. M.; Wheat, T. E. Journal of Chromatography
1990, 512, 273-281.
43. Young, P. M. ; Wheat, T. E. Peptide Research 199 0, 3.
(6), 287-292.
44. Hoeger, C.; Galyean, R.; Boublik, J.; McClintock,
R. ; Rivier, J. BioChromatographv 1987, 2. (3), 134-
142 .
45. Newton, R. In High-Performance Liquid Chromatography
in Endocrinology; Makin, H.; Newton, R., Ed.:
Springer-Verlag: Berlin ,1988; pp 1-31.
46. Lim, C. In HPLC of Small Molecules: A Practical
Approach: Lim, C., Ed.: IRL Press: Oxford,England,
1986; pp 1-12.
47. Scott, R. P. In Advances in Chromatography;
Giddings, J. C., Ed.: Marcel Dekker,Inc.: New York,
1982; pp 167-196.
48. Fallon, A.; Booth, R.; Bell, L. Laboratory
Techniques in Biochemistry and Molecular Biology:
Vol. 17: Applications of HPLC in Biochemistry;
Burdon, R., Ed.: Elsevier: Amsterdam, 1987.

49. Regnier, F. E.; Gooding, K. M. Analytical
Biochemistry 1980, 103. 1-25.
50. Huang, J.; Guiochon, G. Journal of Chromatography
1989, 492, 431-469.
51. DiBello, C.; Gozzini, L. International Journal of
Peptide Protein Research 1993, 41, 34-42.
52. Principles of Biochemistry. Second Edition;
Lehninger, A.; Nelson, D.; Cox, M., Ed.: Worth
Publishers,Inc: New York, NY, 1993.
53. Canfield, R. Journal of Biological Chemistry 1963,
238 (8), August, 2698-2706.
54. Spector, A. Methods in Enzymology 1986, 128. 320-
55. Stachowiak, K.; Dyckes, D. Peptide Research 1989,
2, 267-274.