Peptide mapping as a preliminary step in analysis of hydrophobic proteins

Material Information

Peptide mapping as a preliminary step in analysis of hydrophobic proteins
Millis, Sherri Z
Publication Date:
Physical Description:
vi, 54 leaves : illustrations ; 29 cm


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


Includes bibliographical references (leaves 51-54).
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Science, Department of Chemistry
Statement of Responsibility:
by Sherri Z. Millis.

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:
34366310 ( OCLC )
LD1190.L46 1995m .M55 ( lcc )

Full Text
Sheni Z. Millis
B.A., Cornell College, 1984
A tfiesis 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
Sherri Z. Millis
has been approved for the
Department of
5k /?5~
Corinne H. Campbell

Millis, Sherri Z. (M.S., Chemistry)
Peptide Mapping as a Preliminary Step in Analysis of Hydrophobic Proteins
Thesis directed by Professor Douglas Dyckes
Ferritins are a family of multisubunit hydrophobic proteins, resistant to heat
denaturation. Previous attempts at peptide mapping comparisons on a reversed-
phase HPLC column using heat denaturation proved uninformative, possibly due
to the percent hydrophobicity of the protein. The goal of this project was to devise
a method that would expose hydrophobic proteins more completely to enzymatic
digestion and thereby increase the fragments available for peptide mapping. Horse
spleen ferritin was degraded with cyanogen bromide prior to traditional enzymatic
digestion in order to expose more of the protein to the enzyme than had been
available by heat denaturation. The digested protein was then injected onto the C4
reversed-phase HPLC column. The resulting peptide map produced twice as many
peaks as previous maps. This procedure generates peptide maps which will be
useful standards for comparison of known ferritins to novel ferritins or comparisons
of other hydrophobic protein groups.
This abstract accurately represents the cont
recommend its publication.
ite's thesis.

Acknowledgments vi
Protein Structure 1
Protein Sequence Analysis 4
DNA Sequencing 5
Edman Degradation 6
Endo- and Exopeptidase Cleavage 7
Chemical Cleavage 10
Chromatography Methods 12
Mass Spectrometry and Combined Techniques 13
Peptide Mapping 15
Ferritins 18
Overview of Cellular Ferritins 18
Serum Ferritin 22
Materials 26

Instrumentation 26
Methods 27
Preparation of Standard Protein Solutions
for Analysis 27
Preparation of Heat Denatured Protein 27
Cyanogen Bromide Degradation 28
Formic Acid Standard Preparation 28
Enzymatic Digestion 29
HPLC Analysis 29
Native and Denatured Ferritin Analysis 31
Cyanogen Bromide Degraded Ferritin Analysis 37

I would like to thank Dr. Douglas Dyckes, my committee chair, for his
advice, support, and guidance. I am also grateful to the other members of my
committee for their time and suggestions.
I would like to send a special thanks to the entire staff at Reproductive
Genetics Center for their tolerance for the last two years, and to Dawn Cherico, a
colleague in the Chemistry program, for keeping me sane throughout this endeavor.
I would never have completed this thesis without the understanding and assistance
provided by these people.
I would also like to thank my husband, Michael, and my children, Courtney
and Zachary, for their continuous support, patience, and encouragement, without
which I might never have embarked upon this adventure and without which I surely
would never have completed this adventure.

Protein Structure
Proteins are a class of chemical compounds that consist mostly of long-
chain polymers of amino acids, which are often complex. These compounds are
ubiquitous throughout nature, and are die primary constituents of the cells of both
plants and animals. Proteins are involved in such diverse interactions as storage
deposition and transport, enzymatic catalysis, immunological defense, mechanical
support, and hormonal regulation. In attempting to understand proteins, chemists
often study the relationship of their structure to function.1
The amino acids in proteins are linked by peptide bonds. The specific order
of these amino acids is referred to as the protein's primary structure and is unique to
each protein. The location of disulfide bridges is also included in a primary
structure description, producing a covalent bond description of the protein. The
first complete sequence of a protein was determined in 1953 by Frederick Sanger.2
He sequenced the protein insulin, which has 51 amino acid residues.
The sequence of a given protein is determined by the gene that encodes it.
Often it is characteristic of the species and sometimes of the organ from which it is
derived. Many proteins have very similar structures which vary by only a few

amino adds from spedes to spedes. These proteins are thought to have evolved
from a single genetic protein precursor and reflect evolutionaiy change (divergent
evolution).3 By studying the amino add sequences of related proteins from the
same species or of equivalent proteins from different spedes, scientists can
postulate how far proteins have diverged in die course of evolution. Descriptions
of the changes are depicted with phylogenetic trees.4 The branches of each tree are
proportional in length to die number of differences in the amino acid sequences of
the protein being studied. Evolutionary relationships between the species can be
derived from this information. Species that have relatively few differences in the
primary structures of specific proteins are thought to be closely related. Often
similarities in protein sequences are used to relate structure to function. Positions in
proteins that are conserved even at great evolutionary distances between the species
may be essential for the proper function of the protein. On the other hand, if die
amino acid sequences of two proteins are very different, but die proteins have a
similar three-dimensional structure as defined below, an evolutionary convergence
may exist
The primary sequence of a protein is only one step in describing a protein.
The amino acids orient themselves in various ways to form a secondary structure as
seen by x-ray diffraction.5 This secondary structure is determined in part by the
angle of rotation of the peptide bond, hydrogen bonding, van der Waals forces, and
hydrophobic interactions of side chain constituents of the amino adds.6 Thus,
secondary structure is a description of the steric relationship of amino acid residues

close to one another in the linear sequence. The two most common secondary
structures are the alpha-helix and Ihe beta-pleated sheet.
The relationship of secondary structural units to each other is tertiary
structure. For example, the tertiary structure of myoglobin consists of eight alpha-
helical regions. Two non-helical regions separate two of the alpha-helices, and the
rest of the helical regions are separated by proline bends.
If a protein contains more than one polypeptide chain it exhibits quaternary
structure, a description of the way the subunits are packed together. The individual
chains in such a protein are called subunits, and specific areas in these quaternary
structures are called domains.
The amino acid sequence specifies the tertiary conformation of the protein
through stabilization of specific conformations. Folding of the polypeptide chain
into its biologically active form (also called the native conformation) appears to be a
cooperative, sequential process dependent in part upon the primary structure. The
folding into a biologically active molecule has long been considered a spontaneous
process, independent of cofactors and enzyme-catalyzed reactions.7 Primary
sequence alone seems sufficient for correct folding. Some larger proteins, though,
seem to need more than primary sequence information for proper folding. Recent
evidence shows that a class of proteins called chaperone proteins are probably
involved in the folding process of these larger proteins. The chaperone protein
sequesters the newly forming protein, preventing it from interacting with other
proteins, and holding it in a position that encourages specific interactions within

itself, such as hydrogen bonds or disulfide bridges. The protein then folds into its
functional three-dimensional configuration. Chaperone proteins may speed up
some stages of folding8 and aid in the refolding of heat-denatured proteins and
proteins that have unfolded to cross intracellular membranes.9
In unnatural conditions such as high temperatures or extreme pH, proteins
may lose their functional configuration through a major change from the original
native structure, without alteration of 1he amino acid sequence.10 The unfolded
molecule is termed denatured. Denaturation can cause dissociation of a polypeptide
protein into its respective subunits as well as loss of configuration of the individual
subunits. Protein denaturation may be reversible or irreversible, depending upon
the denaturing conditions.11 The stabilities of proteins to unfolding via different
denaturing conditions has been found to correlate with a number of factors such as
the number of hydrogen bonds and salt bridges and amount of surface area on the
protein. Mutant forms of proteins have shown changes in stability of the native
state due to single amino acid changes. Examples of these changes are loss of
hydrogen bond capability or distortion of regions on the protein.12 The ability of
proteins to resist denaturation techniques will thus be affected in part by their
structural qualities, such as the number of hydrophobic amino acid residues.
Protein Sequence Analysis
Analysis of amino acid sequences is an important tool in acquiring the
primaiy structure of proteins. This information can be used for many purposes,

such as correlation of structure with function or identification of structural features
common to an entire family of proteins. Although sequence determination cannot
yet serve for precise prediction of tertiary conformation, it is a fundamental step
toward the goal of understanding the protein as a whole. Being able to define
which amino acids are conserved throughout a family of proteins will ultimately
show which regions are important structurally or functionally.
DNA Sequencing
cDNA sequencing is currently the fastest and simplest method to determine
primary protein structure, but it has drawbacks. cDNA is complementary DNA,
synthesized in the laboratory by a process utilizing reverse-transcriptase catalysis of
any RNA template to produce DNA that is complementary in base sequence to the
RNA. The limit to cDNA sequencing is that RNA contains only the introns, or
expressed genes. Exons and unexpressed genes have been spliced before
transcription to RNA. Therefore, synthesis of DNA from RNA will not include
these genes. Also, mature protein products may have frame shift mutations and
post-translational modifications that cannot be determined by DNA sequencing. In
order to ascertain these variations in proteins and to confirm DNA sequence
information, many techniques can be utilized for protein sequence determination.13
The techniques include Edman degradation, exopeptidase and endopeptidase
digestion, chemical cleavage, chromatography, mass spectrometiic analysis of

peptide fragments, and combined liquid chromatography/mass spectrometry.
Nucleic acid and protein sequence data can thus complement each other.
Edman Degradation
Edman degradation is a method of chemically labeling the amino-terminal
residue of a protein with phenylisothiocyanate (PITC), and cleaving it. Edman's
method formed the foundation for all further sequencing studies.14 The study of
amino-terminal residues is almost a traditional beginning for the determination of
sequence.15 Once the initial amino-terminal amino acid has been identified, any
other cleavage methods will show at least one fragment with the identical amino-
terminal residue, forming a basis for laying out the entire final sequence. The
cleavage of the amino-terminal residue does not disturb the bonds between the
other peptides using the Edman method.16 The resulting phenylthiohydantoin of
the N-terminal amino acid (PTH-amino acid) can then be identified by various
chromatographic techniques. If the procedure is repeated on the peptide, the
second amino acid residue can be identified, then the next, and so on down the
backbone. Edman degradation has been automated since its inception in 195617
and is still the most widely used method for the primary structural characterization
of proteins and peptides when reasonable quantities (usually a few milligrams) are
available. It allows analysis of very complex proteins and comparison of species
differences.18 Its drawbacks include sample loss, side reactions, incomplete
cleavage, insufficient sensitivity in the analytical techniques used to scrutinize

cleaved amino acid derivatives, blocked N-terminus, and length of time involved in
analysis. All of these problems are magnified when analyzing low nanomole
quantities of polypeptides.
Endo- and Exopeptidase Cleavage
Terminal amino acids of proteins and peptides can also be removed by
exopeptidases. Table 1.1 shows the cleavage points of the common exopeptidases.
Only a few chemical methods are available for C-terminal determination, so it is
usually done enzymatically, using carboxypeptidases. Carboxypeptidase enzymes
selectively hydrolyze proteins by successive release of amino add residues from the
C-terminus. The different types of carboxypeptidases have individual drawbacks.
For example, Carboxypeptidase A does not hydrolyze bonds involving C-terminal
arginyl, tysl, or prolyl residues.19
Endopeptidases hydrolyze internal peptide bonds. This technique is used to
generate and purify size-appropriate fragments of the protein, because many
proteins are too large to be effectively sequenced using Edman degradation.
Examples include trypsin, chymotrypsin, thermofysin, and elastin. These
endopeptidases are often used as a first step in sequence determination by cleaving
the protein into large fragments. Secondary degradation is often performed using
subtilisin, papain, and pepsin, because they have broader specificity, leading to
smaller fragments. Table 1.1 shows the cleavage points of the common
endopeptidases. The smaller fragments produced combined with terminal end

Table 1.1
Exo- and Endopeptidases and Specific Cleavage Sites
Exooeotidase Peptide Bonds Cleaved
Carboxypeptidase A, C, Y Peptide-X
Carboxypeptidase B Peptide-Arg, Peptide-Lys
Leucine Aminopeptidase X-Peptide
Aminopeptidase M Endopeptidase X-Peptide
Trypsin Arg-X, Lys-X
Chymotrypsin Trp-X, Tyr-X, (Phe-X)
Elastase Ala-X, Gly-X
Staphylococcal Protease Asp-X, Glu-X
Pepsin A Phe-X, Leu-X
Thermolysin X-Ala, X-Ile, X-Leu, X-Met X-Phe, X-Tyr, X-Val
This is not a complete listing. Most other proteinases are too nonspecific or too
restricted to be of general use and are therefore not listed here.

identification or mass spectrometric techniques make reconstruction of the entire
sequence much simpler than removing one amino acid at a time as in Edman
degradation. As seen in Table 1.1, trypsin is one of the most often used
endopeptidases, because it hydrolyses the peptide bonds at the carboxyl group of
lysine and arginine, residues that are universally found in proteins.20 The drawback
to trypsin digestion is its rate dependence upon the adjacent residue. The problem
with rate dependence is that some sites are cleaved very early in the digestion
process while other sites are cleaved much later. Timeline studies generally need to
be conducted to clarify whether all sites have been cleaved in order to determine
amount of time needed for complete digestion of a specific protein. Timeline
studies can, of course, be used to one's advantage also, to show when specific sites
are cleaved. It also allows for exhibition of the chemistry of the reaction. Rates of
cleavage are influenced by the chemical nature of side chains in tiie vicinity; polar
groups decrease the rate of hydrolysis, and side chains with a net negative charge
adjacent to the lysine or arginine significantly reduce the rate of hydrolysis;
repetitive residues of lysine and/or arginine are also attacked by trypsin at a
considerably lower rate. Thus basic adjacent residues cleave the fastest and prolyl
residues are not cleaved, because of their imino bond.
Minor contaminants of the enzyme such as chymotrypsin can endanger the
specificity of the digestion. Thus, treatment of the purified trypsin with 1-(1-
tosylamido-2-phenyl-ethyl-chloro-methyl ketone (TPCK) is important, because
TPCK blocks tiie active center of chymotrypsin.21 A useful technique in sequence

determination is to digest separate batches of the protein with different
endopeptidases in order to produce overlapping sequences to narrow down the
possible sequence pattern.
Chemical Cleavage
Chemical methods of cleavage are also employed for protein sequence
determination.22 The most successful chemical method has been cleavage at
methionyl residues with cyanogen bromide.2324 The success of non-enzymatic
cleavage by cyanogen bromide of the chain at methionyl residues is due to the
specificity for methionine and rarity of methionine residues in proteins. This
increases the probability of a few large peptide fragments which provide valuable
structural information. The nucleophilic attack by the sulfur in the thioether side
chain of methionine on the carbon of CNBr results in displacement of the bromine
atom. The bond involving the carboxyl group of methionine is then cleaved25 to
yield a mix of homoserine and homoserine lactone plus methyl thiocyanate. The
cyanogen bromide degradation reaction is seen in Figure 1.1.
Cyanogen bromide degradation is generally performed in 0.1 N HC1, 70%
formic acid, or anhydrous TFA. Formic acid is generally considered the better
solvent, hi this highly acidic solution the protein can be effectively denatured and
broken into roughly two to four fragments which can be further digested by other
methods.25 CNBr has been employed with excellent results in studies related to

Figure 1.1
Cyanogen Bromide Degradation
Methionyl residue in peptide
Cyanogen Bromide
Br HjC-C= N ch2 Hz? V CHRNHC C H \ TvIHCRHCO- h2 H2C'^C'V'0 Br \ / NHCC NHCRHCO H Iminolactone Bromide
Cyanosulfonium Bromide Methyl thiocyanate
\ /
Homoserine lactone derivative
Peptide fragment

elucidation of primary structure of several proteins, including myoglobin.26 The
primary drawback is that adjacent seryl or threonyl residues can prevent cleavage of
methionine at those sites through participation of their hydroxyl groups.27
Comparisons of chemical cleavage versus enzymatic cleavage show that
chemical cleavage is oiten incomplete and the complex resulting mixture difficult to
separate. Enzymatic cleavage shows higher cleavage rates at specific cleavage sites
resulting in distinct sets of proteins. Enzymatic cleavage is advantageous with small
amounts. Many chemical modifications change the solubility pattern of the
peptides.27 Combining enzyme hydrolysates with chemical hydrolysates to get
structure information can eliminate or reduce some of their individual drawbacks.
Chromatography Methods
Another technique used either in conjunction with chemical or
enzymatic cleavage is chromatography. Chromatographic methods used in the
determination of amino add sequences apply the concept of exposing the analyte to
two phases, a stationary phase which can take many forms, and a mobile phase
which is dther gas or liquid. The protein is introduced to the system via the mobile
phase. For example, if the mobile phase is a liquid, die protein is dissolved in the
liquid and passed over or through the stationary phase. Separation is contingent
upon different degrees of physical interaction of the substances with the stationary
phase of the system. These possible physicochemical interactions are surface
adsorption, liquid partition, ion exchange, steric exclusion, and molecular affinity.

Other variable conditions include mode of detection, physical environment, and
condition of the sample.28
At the height of success of nucleic acid sequencing techniques, a new form
of chromatography, High Performance Liquid Chromatography (HPLC) became a
tool in protein chemistry. It had been available since the 1960s but its use for
peptide and protein analysis had not become widespread. Improvements in
microprocessor-controlled HPLC machinery and in the column-support materials
enabled these tools to make advancements in protein chemistry. Its greatest
successes are in the ability to analyze extremely small samples and separation of
biologically inactive or stable polypeptides.13 Reversed-phase chromatography
was developed in the 1970s and uses a non-polar or weakly polar stationary phase
and a polar eluent.
Thin-layer chromatography separation techniques have been widely applied
to peptide mixtures from ribosomal proteins. Ribosomal protein mixtures are rather
complex and consist of many similarly sized peptides of similar charge and are
poorly resolved by paper chromatography or electrophoretic techniques. Thin-layer
chromatographic methods are best for small peptide separation, although some
peptides up to 60 residues in length have been well resolved.29
Mass Spectrometry and Combined Techniques
Mass spectrometric techniques have been used in recent years for primaiy
sequence determination, even though it is based on entirely different principles

from the conventional approaches. It has the advantage of rapid analysis and
determination of blocked amino terminus substituents.30 as well as being the most
definitive method for identification of modified residues. Field desorption,31 ^^Cf
plasma desorption,32 and fast atom bombardment3334 mass spectrometric
techniques have all been used to characterize nonderivatized peptides.
Conventional mass spectrometry requires volatilization of the analyte;
therefore, thermally labile proteins have generally not been amenable to direct mass
spectrometric analysis. Chemical derivatization for GC/MS has been successful in
mass determination in these types of molecules.
Thermospray liquid chromatography/mass spectrometry has greatly
enhanced protein sequencing by permitting analysis of thermally labile molecules
without derivatization It is reliable, quick, sensitive and able to detect unusual
structures.35 Its major drawback is in its inability to analyze large proteins without
first cleaving them into smaller segments.
In 1988 electrospray ionization-mass spectrometry (ESI-MS) was
introduced as a tool for the ionization of large biomolecules. Large proteins are
ionized by spraying a very dilute solution from the tip of a needle across an
electrostatic field gradient of a few kV forming multiprotonated molecules. At
about the same time, matrix assisted laser desorption/ionization (MALDI) was
introduced and has been successful.3637 in molecular weight determinations of very
large proteins without the need to generate smaller fragments of the protein. This
is accomplished by placing the protein directly onto a chemical matrix, then exciting

it with a laser beam for analysis. Although ESI-MS has been made commercially
available, MALDI is not yet widely available.
Peptide Mapping
In 1980 H.D. Niall stated that more effort needed to be put into methods
for specific cleavage of proteins into a few large fragments that are isolatable to
decrease sequence time and to be able to more quickly compare mature proteins to
cDNA protein sequences.38 Although LC/MS techniques may someday become
routine laboratory procedures, current sequence techniques are either lengthy, some
taking several years to produce the amino acid sequence of a single protein, or are
not available to the general science populous. Peptide mapping is a procedure that
is available now and can determine similarities and differences in protein sequences
without having to obtain die entire sequence. Peptide mapping can utilize cleavage
of proteins into large fragments for quick comparisons of those fragments. Peptide
mapping also employs many techniques that generally are used in some part of the
sequencing process, anyway, so it is a useful preliminary analytical process.
The sample to be mapped can be a mixture of compounds that is separated
into its constituent components on die HPLC or a single compound that is
fragmented into smaller peptides through either enzymatic or chemical cleavage.
The resulting fragments are then separated by one of the various chromatographic
techniques. Small changes in amino acid composition can change chromatographic
retention. For example, if two proteins differ in sequence by a single amino acid

residue, one of which is hydrophobic and the other is hydrophilic, this difference
can potentially cause die peak representative of the fragment containing that amino
acid to shift. The resulting difference in peak retention between the two maps can
be used to draw conclusions regarding the difference in the two proteins. This
makes the HPLC an excellent technique for quick preliminary verifications of
variations in the amino acid sequence.39 Although the specific amino acid change
cannot be determined this way, scientists can draw general conclusions regarding
overall similarities between two proteins. The most common method utilizes
enzymatic cleavage (discussed above) followed by separation on a chromatographic
column to yield the peptide map. Peptide maps are unique for eveiy protein.40
Reversed phase HPLC is the method utilized most often in peptide map
construction, because of its resolution and selectivity. RP-HPLC allows protein and
peptide separations to be performed based on the basis of small changes in
polarity.41 Reversed phase chromatography differs from most other
chromatographic techniques where the attractive forces between the stationaiy
phase and mobile phase dominate. Reversed phase employs the above-mentioned
procedure of subjecting the enzymatically cleaved protein to a mobile phase which
consists of water, or water and an organic modifier. The mobile phase is forced at
high pressures through a nonpolar stationaiy phase of alkyl chains bonded to silica
particles. These alkyl chains vary in length from column to column. The primary
interactions between the two phases are hydrophobic. Selectivity is determined by
specific mobile phase effects. The pump which forces the mobile phase through the

column can operate in two different modes: isocratic, where the composition of the
mobile phase remains constant through the chromatographic run, or gradient mode,
in which the mobile phase composition is changed continuously and/or step-wise
throughout the run.42 Peptide mapping has been successful with as little as 50 ng
of protein.43
Clearly, many methods are available to help in the determination of
structure and amino acid sequence analysis. Peptide mapping, however, is simple,
fast, and can be used to make preliminaiy judgments about proteins. The
information obtained can be used to tailor further studies. Site-specific differences
can be very apparent with the change in a single peak on peptide maps. Peptide
mapping is thus an excellent candidate for the study of families of proteins. Novel
proteins that seem related to a specific family of proteins can be compared to well-
studied ones for preliminary relationship identification without determining the
entire sequence of the protein. Preliminary determinations can be made about the
absence or presence of post-translational modifications, since some families of
proteins differ in their specific sequences by these modifications. One such family
of proteins being studied that would be amenable to peptide mapping for the above
reasons is the ferritins.

Overview of Cellular Ferritins
The cellular fenitin protein sequesters iron in the cell as a response to iron
overload. This family of highly conserved, vety large, multi-subunit proteins is
found in most vertebrate and invertebrate cells, higher plants, fungi, and bacteria.
Although variations in function, structure, and amount of ferritin exist between
species and within different organs of a species, all known ferritins share some
structural properties which suggests the presence of cell-specific features of genetic
Ferritin is a large spherical protein coat surrounding a core of hydrous ferric
oxide. Ferritin has a mass of approximately 450,000 Daltons (Da) and can
sequester up to 4,500 iron atoms, along with variable amounts of phosphate which
are thought to enter through channels in the protein.44
Ferritin is comprised of 24 subunits packed together to form a compact
shell. The subunits are of two types which vaiy based on the organ of origin within
the same species,45 producing isofenitins. The L subunit is 19,766 Da and the H
subunit is 21,099 Da. Experimental evidence shows that the tertiary conformation
for both L and H subunits is virtually identical, despite differences in their amino
acid sequences.46 X-ray diffraction patterns show that structural homology
between ferritins' assembled subunits is retained in spite of differences in primary
structure of the individual subunits and fluctuation in the number of the respective

subunits per active ferritin.44 RP-HPLC analysis of ferritins has separated the
chemically distinct subunits.474*
The horse spleen fenitin is composed of approximately 90% L subunits.
These subunits cause die thermostability of the protein through numerous inter-
subunit salt bridges, hydrogen bonds, and hydrophobic interactions. Its L subunit
has 174 amino acids in a tertiary conformation of two pairs of alpha helixes joined
by a loop and one short helix at the end.49 Figure 1.2 shows the primaiy structure
of horse spleen ferritin L subunit. The primaiy sequence is labeled with the regions
of the five alpha helices. Figure 1.3 shows the tertiary and quaternary structure of
Sequence conservation in known ferritins is related to the three-dimensional
structure. Side chains on the outside surface of the shell have die most differences
from each other, with only about 33% of the amino acids being identical.
Approximately half die salt bridges and hydrogen bonds are conserved or replaced
with similar acting amino acids, continuing the fiinctionality of their position.
Conservation of structure is actually less inside ferritin on the iron storage surface.
Functional adaptation is a probable explanation. The highly conserved region is
around the 3-fold axes where metal binding ligands are found in all known
sequences. Residues in the outer funnel region are also conserved or conservatively

Figure 1.2
Primary Structure of Horse Spleen Ferritin L Subunit
alpha-helix A
alpha-helix B
alpha-helix C
alpha-helix D
alpha-helix £
All methionine residues (M) are shown in bold. Alpha-helix regions are underlined
and labeled. Arginine (R) and lysine (K) are shown in italics.

Figure 1.3
Tertiary and Quaternary Structure of Horse Spleen Ferritin L Subunit
(adapted from Pauline M. Harrison (49))

Serum Ferritin
Most ferritin is located intracellularly. Small amounts have been found in
the serum.50 While serum ferritin is immunologically similar to cellular ferritin, it
has a lower iron content Approximately 60% of serum ferritin binds to
concanavilin A (Con A), indicating a glycosylated subunit.51 Cellular ferritin does
not bind to Con A. Glycosylation is a post-translational modification of amino acid
side chains, which is used in cell signaling. Other differences in cellular and serum
ferritin that are indicative of the presence of a carbohydrate group on a serum
ferritin subunit and absence on cellular ferritin are 1) the isoelectric point of cellular
ferritins is unchanged when incubated with neuraminidase, but acidic isofenitins
from serum are converted to basic isofenitins (this indicates the removal of sialic
acid residues)52 and 2) the serum ferritin in patients with hemochromatosis, an iron
overload disease, has a subunit which "stains for carbohydrates on PAGE,
whereas cellular ferritin does not Serum ferritin is elevated in many pathological
conditions, while cellular ferritin levels remain steady.51 The complete amino acid
sequence and molecular weight of serum ferritin is not yet known.
Recently, a ferritin was isolated from horse heart that has similar
characteristics to serum ferritin such as binding to Con A and smaller size
compared to cellular ferritin.53 Horse heart ferritin and serum ferritin may be more
closely related to each other than to cellular ferritins, and they may in fact be the
same molecule. Utilizing peptide mapping techniques to manipulate characteristics
which might exploit any differences between die various ferritins would be valuable

in answering some of these questions.
Peptide maps of different ferritins can help in structural identification and
in identification of amino acid changes which would affect structure. Chemical and
enzymatic cleavage of ferritins would produce fragments of the molecules which
would elute based on hydrophobic and hydrophilic interactions with die column. A
single change from a hydrophobic to a hydrophilic amino acid, for example, could
change the elution time of the fragment, and the peak would shift on the
chromatogram, as has been seen with single amino acid changes in hemoglobin
mutants.54 While the specific change in amino acid residue(s) will not be
identifiable by this method, similarities and differences between two ferritin
molecules can be recognized using this procedure. Evolutionaiy relationships
within and between ferritins can be studied. Whether or not other iron-sequestering
molecules are from conveiging or diverging evolutionary precursors can be
postulated based on simple peptide maps showing peak shifts.
The concern with peptide mapping of ferritins to date has been their
hydrophobicity. An index for hydrophobicity can be calculated for a protein based
on the total number of hydrophobic amino acid residues per molecule.55 Adding
up the number of hydrophobic amino acid residues in cellular ferritin reveals it to
be one of the more hydrophobic of the known proteins, at 29% hydrophobicity
(compared to cytochrome c at 21% hydrophobicity). Using cellular fenitin as a
standard, fins study devised a method that works with the hydrophobic properties
of the ferritins to develop peptide maps which can be used for structural

comparisons. This method can be also extrapolated to other hydrophobic protein
groups for structural comparison analysis.

The purpose of this study was to develop techniques which could be used in
preliminary comparison studies of the serum, glycosylated horse heart, and cellular
ferritin proteins and in comparison studies of other hydrophobic proteins.
Distinguishing structural and amino acid differences of cellular and serum ferritin
is an important step in understanding their biological roles and how their differences
may play a role in disease.
Ordinary structural and sequence analysis techniques do not work well with
cellular ferritin, because of its extremely hydrophobic composition. Other
hydrophobic proteins may also need to be exploited for structural analysis.
Therefore, methods for denaturation and sequence analysis specific to these
characteristics needed to be developed for elucidation of structural and sequence
differences. Since peptide mapping is a relatively quick, very sensitive technique
for establishing preliminary data and fragments for further analysis, it was chosen
as the construct with which to develop a method that would enhance analysis of
ferritins and other hydrophobic proteins.

Ammonium bicarbonate, HPLC grade acetonitrile, and cyanogen bromide
were purchased from Fisher Scientific (Pittsburgh, PA). Horse spleen ferritin and
trifluoroacetic acid were purchased from Sigma Chemical Company (St. Louis,
MO). TPCK treated trypsin was purchased from Worthington Biochemical
Corporation (Freehold, NJ). Myoglobin from horse skeletal muscle was purchased
form Calbiochem (San Diego, CA). Nap-5 (G-25 gel exclusion) columns were
purchased form Pharmacia Biotech (Piscataway, NJ). 4 mm syringe filters were
purchased from Baxter (McGraw Park, IL). HPLC grade water was purified in
our laboratory.
The HPLC was an integrated Waters system and included a Waters 501
HPLC pump, U6K injector, 486 absorbance detector, and pump control module,
all interfaced with a 386/33i NEC poweimate, using Millipore Millenium 2010
software version 2.0 (Waters Chromatography Division, Millipore Corporation,
Milford, MA).

The HPLC column was a Hi-Pore RP-304 C4 column with dimensions of
250x4.6 mm., particle size 5 micrometer, and pore size 300 Angstroms, purchased
from BioRad (Hercules, CA).
Preparation of Standard Protein Solutions for Analysis
Cellular horse spleen ferritin was chosen as a standard in developing
methods which would work for ferritin structural comparisons, because the large
percentage of L subunit in the protein would increase die homogeneity of the
sample. Purchased protein was suspended at a concentration of 5 mg/'mL in 0.1 M
ammonium bicarbonate buffer, pH 8.0 by buffer exchange on a G-25 gel exclusion
chromatography column. The protein was then chromatographed as described
Preparation of Heat Denatured Protein
Native proteins in a concentration of 10 mg/mL in 0.1 M ammonium
bicarbonate were heat denatured by boiling for 10 minutes at 100 C. The protein
was cooled to 25 C and chromatographed as described below.

Cyanogen Bromide Degradation
Solid cyanogen bromide (a 100-fold molar excess with respect to the
methionines) was dissolved in 1 mL of a 70% formic acid solution. The solution
was added to 10 mg of protein dissolved in 0.1 mL of 0.1 M ammonium
bicarbonate. The vessel was stoppered, and suspension was maintained by using a
magnetic stirrer. The reaction was allowed to proceed for 24 hours at room
temperature. The solution was transferred to a polypropylene tube, and the by-
products were removed by cold-trap vacuum evaporation until approximately 0.3
mL of liquid remained. The residue was resuspended in 0.1M ammonium
bicarbonate, filtered using 4mm filters, and stored frozen until used (which was
generally within a week).
Formic Acid Standard Preparation
10 mg of protein dissolved in 0.1 mL of 0.1 M ammonium bicarbonate was
dissolved in 1 mL of a 70% formic acid solution. The vessel was stoppered and
stirred with a magnetic stirrer for 24 hours at room temperature. The solution was
transferred to a polypropylene tube, and the by-products were removed by cold-
trap vacuum evaporation until approximately 0.3 mL of liquid remained. The
residue was resuspended in 0.1M ammonium bicarbonate, filtered using 4mm
filters, and stored frozen until used.

Enzymatic Digestion
Tiyptic digestion was performed on either denatured or CNBr treated
proteins, prepared as described above. The TPCK trypsin was dissolved in 0.1 M
ammonium bicarbonate buffer (pH 8) and diluted to lmg/mL concentration. The
trypsin was then added to the protein or peptide mixture in a 1:50 ratio (v/v). The
solutions were allowed to react for 24 or 48 hours at room temperature, then frozen
(to stop digestion) until analyzed.
HPLC Analysis
Protein and peptide solutions were suspended in 0.1 M ammonium
bicarbonate buffer and filtered through 4 mm filters prior to injection on the C4
reversed phase HPLC. 10-100 microliter samples were applied to the column, and
absorbance was monitored at 214 nm.
The following linear gradient was used:
Table 3.1
Solvent Gradient
Time (min) Flow (ml/min) %A Solvent 0.1% TFA/ H20 %B Solvent 0.1% TFA/ Acetonitrile
0 1.0 100 0
50 1.0 . 50 50
55 1.0 0 100
60 1.0 100 0
70 1.0 100 0

The solvents were kept at a continuous flow rate of 1 mL/min with a linear
gradation from 100% solvent A, a solution of 0.1% TFA in water, to 50% solvent
B, a solution of 0.1% TFA in acetonitrile. Washing the column with 100%
acetonitrile at 55 minutes, reversing the linear gradient back to 100% solvent, and
reequilibrating for a ten minute period were essential for satisfactoiy
The solvents were filtered and degassed for 30 minutes before use. TFA
was added to the solvents as a mobile phase modifier and to facilitate the solubility
of the hydrophobic proteins used. TFA dissolves hydrophobic proteins and is
volatile,56 a property that would be useful for later LC/MS techniques.

Native and Denatured Ferritin Analysis
The reversed-phase HPLC analysis of both native and heat denatured
cellular horse spleen ferritin is shown respectively in Figures 3.1 and 3.2. Figure
3.1 shows the elution of a well resolved peak at 58 minutes for the native cellular
horse spleen ferritin. Figure 3.2 shows elution of the denatured cellular ferritin at
the same time as the native molecule. The only difference is a shoulder on the later
side of the peak. The shoulder elution later in the gradient is indicative of more
hydrophobic residues being exposed to the column, slowing the elution slightly.
Little else is indicative of any loss of the tertiary structure of the protein when
compared to the native chromatogram of Figure 3.1.
Subsequent tryptic digests of the native and heat denatured ferritin are
shown in Figures 3.3 and 3.4. These chromatograms indicate the need for some
other form of denaturation or degradation to be executed on the protein before
enzymatic digestion. The chromatogram of the HPLC analysis of 48 hour trypsin-
digested native cellular ferritin shown in Figure 3.3 shows only two to three peaks
from cleavage by the trypsin. Based on the previously identified amino acid
sequence seen in Figure 1.2, trypsin should cleave the molecule at 19 lysine and

Figure 3.1. A sample of native cellular horse spleen ferritin was injected onto a C4
reverse-phase HPLC column and eluted using a 0 to 50% acetonitrile gradient over
50 minutes, followed by a 5 minute gradient from 50 to 100% acetonitrile (shown
in Table 3.1.) with monitoring at 214 nm.

Figure 3.2. A sample of native cellular horse spleen ferritin was heat-treated at
100 C for 10 minutes, then cooled and injected onto a C4 reverse-phase HPLC
column and eluted using a 0 to 50% acetonitrile gradient over 50 minutes, followed
by a 5 minute gradient from 50 to 100% acetonitrile (shown in Table 3.1.) with
monitoring at 214 nra.

Figure 3.3. A sample of native cellular horse spleen ferritin was digested for 48
hours with TPCK-treated trypsin, then injected onto a C4 reverse-phase HPLC
column and eluted using a 0 to 50% acetonitrile gradient over SO minutes, followed
by a 5 minute gradient from 50 to 100% acetonitrile (shown in Table 3.1.) with
monitoring at 214 nm.

Figure 3.4. A sample of cellular horse spleen ferritin heat-treated for 10 minutes at
100 C then cooled, was digested for 48 hours with trypsin, then injected onto a
reverse-phase HPLC column and eluted using a 0 to 50% acetonitrile gradient over
50 minutes, followed by a 5 minute gradient from 50 to 100% acetonitrile (shown
in Table 3.1.) with monitoring at 214 nm.

arginine sites in the protein subunit sequence.47 Figure 3.4 supports the need for
some other form of denaturation of the protein. Despite 48 hours of digestion by
trypsin, the area of the nonhydrolyzed protein peak remains quite large. Fewer
than ten other peaks are visible above the baseline; only three are substantially large
in area.
Few of the 19 sites were accessible to die enzyme. Whether due to its
extreme hydrophobicity or due to the possibility that ferritin retains its structural
integrity and renatures upon cooling, heat denaturation alone does not make the
complete protein accessible to enzymatic digestion. L subunit is composed of
greater than 73% alpha-helical structure. The alpha-helix is held together by the
carboxyl group of each amino acid residue hydrogen bonding with the amino group
of the amino acid four residues ahead in the linear sequence. The turn created by
this hydrogen bonding holds the amino acids quite close together in a tightly coiled
chain. Also, when two alpha-helices come together, they form a folding unit which
further stabilizes them.57 Four of the five alpha-helical regions in L-subunit are
paired. With both the internal portion of each alpha-helix and the paired portion of
between the two alpha-helices, many of the amino acids are sheltered from the
external environment. When the enzyme reacts with the native subunit, it is only
able to access some of the amino acid residues. All the evidence seems to indicate
that harsher degradation methods would be necessary to truly denature ferritin in
order to increase the accessibility of the enzyme to the protein for digestion. Thus,
other methods were examined to degrade fire protein.

Cyanogen Bromide Degraded Ferritin Analysis
The major drawback to degradation of ferritin is its resistance to
denaturation. Unfortunately, the difficulty in using harsher methods is that most
anything that will denature ferritin is also going to denature the enzymes used to
digest ferritin. The best candidate for ferritin degradation or degradation of other
hydrophobic proteins is therefore cyanogen bromide cleavage, because the reaction
is executed in the mild conditions of 70% formic acid, and the byproducts can be
removed by volatilization before beginning enzymatic digestion.
As expected, cleavage of the methionine residues of the protein with
cyanogen bromide before protease digestion was an effective method of degrading
native ferritin. Figure 3.5 shows a much smaller peak for the nonhydrolyzed
protein remaining alter cleavage with cyanogen bromide. Approximately nine other
peaks are distinguishable. When compared to the heat denatured, 48 hour trypsin
digested chromatogram in Figure 3.4, cyanogen bromide cleavage appears to
produce more cleaved products than the method which used both heat denaturation
and trypsin digestion. With three methionines in a subunit of ferritin, ten peaks are
possible: the three fragments from complete fragmentation, the uncleaved protein
peak, and four peaks from incomplete cleavage of the protein at the various sites.
(The methionine residues are in bold face in the primary sequence portion of
Figure 1.2 to show their positions.) The protein is now much more accessible for
enzymatic cleavage. The disadvantage of chemical cleavage is this incomplete

Figure 3.S. A sample of cellular horse spleen ferritin was degraded with cyanogen
bromide for 24 hours in 70% formic acid, diluted with water, evaporated, then
injected onto a C4 reverse-phase HPLC column and eluted using a 0 to 50%
acetonitrile gradient over 50 minutes, followed by a 5 minute gradient from 50 to
100% acetonitrile (shown in Table 3.1.) with monitoring at 214 nm.

cleavage of all methionine residues, producing these partially cleaved peptide
fragments. Another drawback to the cyanogen bromide cleavage method is the
production of by-products as shown previously in Figure 1.1. These by-products
are volatilized during lyophilization. Since our method used cold-trap vacuum
vaporization, a minute amount of the by-products were left behind with the small
portion of liquid that remained in the flask. If the residue were evaporated to
diyness by this method, the protein may irreversibly adhere to the container.
Therefore, it was more beneficial to leave minute quantities of by-product, which
were determined to be eluting with the solvent front and not contributing to any of
the peaks seen on the chromatograms.
Figure 3.6 shows formic acid treated cellular ferritin without cyanogen
bromide. This chromatogram is useful as a standard to show that the peaks seen in
3.5 are due almost entirety to the cyanogen bromide degradation. Formic acid alone
does not degrade the protein. The chromatogram of the formic acid reaction
(Figure 3.6) is overlaid with the chromatogram of the cyanogen bromide reaction
(Figure 3.5) to contrast the two. The overlay is shown in Figure 3.7.
While cyanogen bromide peptide mapping shows resolved peaks, it is also
advantageous to combine the enzymatic digestion with the chemical cleavage to
produce a map with even more peaks for comparison. These secondary peaks can
also be collected and analyzed with a mass spectrometer system. A 24 hour tryptic
digestion of the cyanogen bromide cleaved ferritin is shown in Figure 3.8. Figure
3.9 overlays the 24 hour trypsin digestion of the cyanogen bromide degraded

Figure 3.6. A sample of 24 hour formic acid-treated cellular horse spleen ferritin
used as a control was injected onto a C4 reverse-phase HPLC column and eluted
using a 0 to 50% acetonitrile gradient over 50 minutes, followed by a 5 minute
gradient from 50 to 100% acetonitrile (shown in Table 3.1.) with monitoring at 214

Figure 3.7. Overlay of cyanogen bromide cleaved cellular horse spleen ferritin
(upper line) with the control of formic acid treated cellular horse spleen ferritin
(Figure 3.6, lower line). Both were injected onto a C4 reverse-phase HPLC
column and eluted using a 0 to 50% acetonitrile gradient over 50 minutes, followed
by a 5 minute gradient from 50 to 100% acetonitrile (shown in Table 3.1.) with
monitoring at 214 nm.

Figure 3.8. A sample of cyanogen bromide cleaved cellular horse spleen ferritin
was digested with trypsin for 24 hours, then injected onto a reverse-phase
HPLC column and eluted using a 0 to 50% acetonitrile gradient over 50 minutes,
followed by a 5 niite gradient from 50 to 100% acetonitrile (shown in Table 3.1.)
with monitoring at 214 nm.

Figure 3.9. Overlay of 24 hour trypsin-digested cyanogen bromide degraded
cellular horse spleen ferritin (Figure 3.8, upper line) with 48 hour trypsin-digested
heat-treated cellular horse spleen ferritin (Figure 3.4, lower line). Both were
injected onto a C4 reverse-phase HPLC column and eluted using a 0 to 50%
acetonitrile gradient over 50 minutes, followed by a 5 minute gradient from 50 to
100% acetonitrile (shown in Table 3.1.) with monitoring at 214 nm.

protein (Figure 3.7) to the 48 hour trypsin digestion of the heat denatured protein
(Figure 3.4). The overlaid chromatograms show the increase in number of peaks
obtained by degrading the protein with cyanogen bromide versus heat denaturation
before enzymatic digestion.

Changes in amino acid sequence from one protein to another in a family of
proteins by die process of evolutionary divergence can be used to make judgments
about structural and functional changes exhibited by those proteins. Comparison of
the amino acid changes of potentially related proteins by one of the many amino
acid sequence analysis techniques mentioned in the introduction is a primary step in
identifying some of these similarities and differences. Different methods must be
tried to establish which methods work best for any given group or type of proteins.
In this study, enzymatic digestion of native ferritin was shown to be
ineffective in producing informative peptide maps. Fewer than 5 peaks out of 20
possible total peaks including the nonhydrolyzed protein will not provide much
information to compare with other maps. The preparatory heat denaturation of the
protein prior to enzymatic digestion was also determined to be an ineffective
denaturation method for use on ferritins. While this method produced more peaks
than digestion of the native protein, fewer than 10 peaks were still observed. The
ineffectiveness of these methods may be related to the protein's structure: its
hydrophobicity may cause resistance to enzymatic digestion or temperature
changes. In addition, the characteristics which cause the hydrophobicity may also
promote renaturation of the protein upon cooling.

Because heat denaturation was not sufficient in degrading the protein, other
strategies were adopted to increase degradation of the protein and make it more
accessible to enzymatic digestion. This study shows that cyanogen bromide
degradation of the ferritin was an effective method of degrading the protein. Ten
fragments were produced by the cyanogen bromide method, which was as many
fragments as had been produced by both heat denaturation and enzymatic
digestion. The degradation allows for more complete enzyme-digested peptide
maps of the cellular horse spleen ferritin with an increased numbers of peaks that
can be used to establish preliminary identification and structural characteristics of
This study moved beyond the previous efforts by clearly delineating an
increased number of peaks available for comparison. The previous peptide maps
done in this laboratory comparing trypsin and chymotrypsin digestion of horse
spleen ferritin to serum ferritin and glycosylated heart ferritin were unable to
delineate clearly between the serum and glycosylated ferritins because of the
resistance to heat denaturation.51 The method used for those maps involved heat
denaturation of the ferritins prior to protease digestion and HPLC analysis. It was
determined that some other method needed to be used to more completely denature
and digest ferritins before a more informative peptide map analysis could be
Chemical degradation of cellular ferritin by cyanogen bromide cleavage
served this purpose by producing a few large fragments. HPLC analysis of these

fragments generated peaks which will be useful later to compare to maps of other
fenitins for preliminaiy analysis. Enzymatic digestion of the fragments produced
by cyanogen bromide degradation and injection on the HPLC also produced
chromatograms which provide enough information to compare to other ferritin
Peptide mapping of fenitins by using both a chemical cleavage to degrade
the protein and enzymatic digestion to further fractionate the peptides is an
excellent preliminaiy approach for comparison of ferritins and possibly other
hydrophobic proteins. It eliminates the denaturing difficulties associated with heat
resistance and allows for more complete enzymatic digestion of the protein, despite
a large hydrophobic composition. The resulting maps, which show many more
peaks, can be used for comparisons with serum and glycosylated heart ferritin when
they become available for analysis. Until more reliable methods are made available
for complete amino acid sequencing of the ferritins, this method will allow for some
basic comparisons to be made. This method will also allow for analysis of other
hydrophobic proteins which are resistant to common methods of denaturation and

The information gained from these maps could be valuable in eventually
answering many of the questions currently posed about different ferritins such as
determining glycosylation, or correlating the structural changes which occur in
disease states with elevated levels of the serum ferritin. While it may be years
before some of the questions to serum fenitin and glycosylated heart ferritin are
answered, some preliminary answers may be resolved through initial identification
and analysis by chemical and enzymatic digestion and subsequent peptide mapping.
The next step for peptide mapping would be to obtain serum ferritin and
glycosylated heart fenitin and compare their peptide maps obtained using the
methods developed in this study. The degradation method using cyanogen bromide
cleavage alone or cyanogen bromide followed by enzymatic digestion with trypsin
should produce maps that can be definitively compared to the maps obtained herein
of cellular horse spleen fenitin. The primary sequence shows that in the currently
known ferritins only one of the three methionines is completely conserved.
Utilizing the peptide map to compare to unknown ferritins, one would be able to
see whether one or both of the other two methionines have been conserved and
postulate evolutionary divergence.

An improvement over the current protocol would be use of a lyophilizer
instead of cold-trap evaporation to remove die cyanogen bromide by-products.
Contaminants would be further decreased by the use of a lyophilizer.
Another reasonable study would be to use a solvent on the HPLC with a
formic acid base, since formic acid is used in the chemical cleavage. This would
eliminate the concern of possible contamination with formic acid. Formic acid as a
solvent has been considered to be a good substitute for the TFA,58 and can be
volatilized later, which would also be excellent for future studies of ferritin using
combined LC/MS.
One of the disadvantages of cyanogen bromide degradation, the incomplete
cleavage of all methionine residues resulting in partially cleaved peptide fragments,
can actually be used to an advantage, both for continued peptide mapping and
when switching over to the mass spectrometric analysis. Analyzed fractions of the
three completely cleaved peptides will overlap with the six incompletely cleaved
fragments. On the RP-HPLC, the fractions of individual peaks could be collected
and concentrated down. The peptide could be further digested with another
enzyme to cleave other sites. Digestion could even be done with an on-line
technique, by immobilized endopeptidases, then separated on the HPLC column,
and analyzed by mass spectrometry, amino acid sequencing can be performed
utilizing die overlapping segments.59
Once preliminary similarities and differences between the known cellular
ferritins and serum ferritin or glycosylated ferritin are defined, continued analysis,

by LC/MS molecular weight determination and amino acid sequence determination
of the fractions obtained by these peptide mapping methods, will further answer the
questions surrounding the ferritins.
Structural analysis of other hydrophobic proteins can also be attempted
using this method, since it strives to increase accessibility of techniques to the
hydrophobic regions for more efficient and complete investigation into the protein
of interest.

1Iight, A., Proteins: Structure and Function (New Jersey: Prentice-Hall, Inc.,
1974) vii, 86.
2 Rfyle, A.P.; Sanger, F.; Smith, L. F.; Kitai, R., Biochemistry Journal 60 (1955):
3 Doolittle, R. F., Science 214 (1981): 149-159.
4 Dickerson, R.; Geis, L, The Structure and Action of Proteins (New York: Harper
6 Row, 1969) 60.
5 Pauling, L.; Corey, R.B.; Branson, H.R., Proceedings of the National Academy
of Science 37 (1951): 205-211.
^Creighton, T.E., Proteins Structure and Molecular Properties (New York: W. H.
Freeman and Company, 1983) 133-149.
7 Anfinsen, C.B., Science 181 (1973): 223.
8 Craig, E.A., Science 262 (19931:1902-1903.
9 Ezzell, C., Journal of NIH Research 6 (1994): 31-34.
10 Tanford, C., Protein Denaturation. (1981): 121-282.
11 Hermans, J., Jr., Methods of Biochemical Analysis. Vol. XIII: 81-111.
12 SteUwagen, E.; Wilgus, H., Nature 122 (1978): 33-42.
13Hughes, H., In Methods of Biochemical Analysis #29: D. Glick, Ed. (New York:
John Wiley & Sons, 1983) 59-135.
14 Chemical Synthesis and Sequencing of Peptides and Proteins. Liu, T.;

Schechter, A.; Heinrikson, R.; Condliffe, P., Ed. (New York: Elsevier North
Holland Inc., 1981)79.
15 Schroeder, W. A., The Primary Structure of Proteins (New York: Haiper &
Row, 1968)69.
16 Edman, P., Acta Chemica Scandinavica 10 (1956): 761.
17 Edman, P.; Begg, G., European Journal of Biochemistry 1 (1967): 80.
18 Chemical Synthesis and Sequencing of Peptides and Proteins. Liu, T.;
Schechter, A.; Heinrikson, R.; Condliffe, P., Ed. (New York: Elsevier North
Holland Inc., 1981)91.
1 ^Protein Sequence Determination. Needleman, S., Ed. (New York: Springer-
Verlag, 1970) 83-84.
20Methods in Enzvmologv XIX. Perlmann, G.E., Ed. (New York: Academic
Press, 1970) 57.
21 Gross, E.; Witkop, B., Journal of Biological Chemistry 237, 6 (1962): 1856
22 Witkop, B., Advances in Protein Chemistry 16(1961): 221-321.
23 Gross, E., Methods of Enzvmologv 11 (1967): 238-255.
24 Protein Sequence Determination. Needleman, S., Ed. (New York: Springer-
Verlag, 1970) 138-141.
25Edmundson, A., Nature 4. 27 (1963): 353-357.
26Chemical Synthesis and Sequencing of Peptides and Proteins. Liu, T.;
Schechter, A.; Heinrikson, R.; Condliffe, P., Ed. (New York: Elsevier North
Holland Inc., 1981)90.
27 Tietz, N., Fundamentals of Clinical Chemistry (Philadelphia: W.B. Saunders
Company, 1987) 105.

28Huang,J.; Guiochon, G., Journal of Chromatography 492 (1989): 431-469.
29Biemann, K., Biochemical Applications of Mass Spectrometry (New York:
Elsevier North Holland Inc., 1986) 30.
30Matsuo, S., Biochemical Mass Spectrometry (1981): 34.
31 Hakannson, P., Journal of the American Chemical Society (19821: 4878.
32Barber, T., Biomedical Mass Spectrometry (1982):72.
33 Rinehart, J., Science. 212 (1982): 933-937.
34 Kim, H.Y.; Pilosof, D.; Dyckes, D.F; Vestal, M.L, Journal of the American
Chemical Society 106, 24 (1984): 7304-7309.
35 Karas, Analytical Chemistry 60. 20 (1988): 2299-2301.
36Chait, B.T.; Kent, S.B.; Science 257 (1992): 1885-1894.
37 Niall, H. D., Methods in Peptide and Protein Sequence Analysis. Birr, C., Ed.
(New York: Elsevier/North-Holland Biomedical Press, 1980) 123-128.
38Janiss, L.J.; Regnier, F.E., Journal of Chromatography 444, (1988): 1-11.
39 Dong, M.W.; Tran, A., Journal of Chromatography 499 (1990): 125-139.
40Castagnola, I.; Cassiano, L.; De Cristofaro, R.; Landolfi, R., Journal of
Chromatography 440 (1988): 231-251.
41Tietz, N., Fundamentals of Clinical Chemistry (Philadelphia: W.B. Saunders
Company, 1987) 105.
42 Stone, K.L.; Williams, K.R., Journal of Chromatography 359 (1986): 203-212.
43Theil, E.C., Annual Review of Biochemistry 56 (1987): 289-315.

44 Passaniti, A.; Roth, T.F. Biochemical Journal 258 (1989): 413-419.
45 Collawn, J.F., Jr.; Donato, H., Jr.; Fish, W.W., Biochimica et Biophvsica Acta
871 (1986): 235-242.
46 Gowan, L.; Collawn, J., Jr.; Fish, W., Journal of Chromatography 264
(1983): 463-468.
47 Collawn, J., Jr.; Donato, H., Jr., Upshur, J. Comparative Biochemical
Physiology 81B (1985): 901-904.
48 Harrison, P.M.; Ford, F.G.; Rice, D.W.; Smith, M.A.; Treffry, A.; White, J.L.,
Biochemical Society Transactions 15 (1987): 744-748.
49Hammock, B. N., Master's Thesis, University of Colorado at Denver 1994.
50Worwood, M., Clinical Science 70 (1986): 215-220.
51Cragg, S.; Worwood, M., Biochemical Journal 199 (1981): 565-571.
52 Campbell, C.H., Journal of Cellular Biochemistry 53 (1993): 420-432.
53Schroeder, W.; Shelton, J.B.; Shelton, J.R., Journal of Chromatography 174
(1979): 385-392.
54 Parker, J.; Guo, D.; Hodges, R. Journal of Chromatography 359 (1986): 499-
55Hermodson, M.; Mahoney, W.C., In Chemical Synthesis and Sequencing of
Peptides and Proteins. Liu, T.; Schechter, A.; Heinrikson, R.; Condliffe, P., Ed.
(New York: Elsevier North Holland Inc., 1981): 122.
56 Karplus, M.; Weaver, D.L., Nature 260 (1976): 404-406.
57 Poll, D.J.; Harding, D.R.K.; Journal of Chromatography 469 (1989): 231-239.
58 Stachowiak, K.; Dyckes, D., Peptide Research 2 (1989): 267-274.