Note: Descriptions are shown in the official language in which they were submitted.
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
__ 1 -_
ENHANCED PROTEIN SEPARATION AND ANAL YSIS
STATEMENT OF GOVERNMENTAL INTEREST
This work was supported by funds from the National Institutes of Health (NIH),
under Heart
and Lung grant number 24526. The government has certain rights in this
invention.
FIELD
This disclosure relates to the field of proteomics, and particularly to
eWanced protein
separation techniques useful in the study of proteomes.
BACKGROUND
The mitochondrion is one of the most complex as well as one of the most
important
organelles in a eukaryotic cell. It consists of multiple comparhnents (Fret'
and Mannella, TIBS,
25:319-324, 2000; Perkins et al., J. Bioenerg. Bionaerubr., 30:431-442, 1998;
Perkins et al., J. Struct.
Bial., 119:260-272, 1997) containing a vast number of proteins which must
somehow be arranged to
cant' out a variety of processes fundamental to cell function. These processes
include heme
synthesis, the TCA cycle, (3-oxidation of fatty acids, the urea cycle,
electron transport, and oxidative
phosphorylation. Electron transport and oxidative phosphorylation alone
require the coordinated
action of five enzyme complexes, which together are comprised of an estimated
86 different
structural protev~s (Saraste, Science, 283:1488-1493, 1999). In addition,
there are non-structural
proteins which are required for the proper assembly and regulation of these
complexes (e.g., Surf I
and ScoII) (Sue et al., Ann. Neurol., 47:589-595, 2000; Papadopoulou et al.,
Nat. Genet., 23:333-337,
1999; Tiranti et al., Am. J. Hzrm. Genet., 63:1609-1621, 1998; Poyau et al.,
Hum. Genet., 106:194-
205, 2000). To add further complexity to this organelle, 13 of the structural
proteins are encoded by
mitochondrial DNA (Taanman, Biochinz. Biophys. Acta, 1410:103-123, 1999).
There is also increasing evidence that mitochondria play an unportant role in
cell death and
aging. The importance of mitochondria to apoptosis was first indicated when
bcl-2 was identified as
a mitochondrial protein that could prevent apoptosis (Hockenbery et al.,
NatZrre, 348:334-336, 1990).
Since this initial observation, it has further been noted that in a cell-free
system, DNA fragmentation
was dependent upon a mitochondria) fraction (Newmeyer, Cell, 79:353-364,
1994). In addition, it is
now known that during apoptosis, mitochondria release several pro-apoptotic
proteins including
c5rtochrome c (Liu et al., Cell, 86:147-157, 1996) and apoptosis inducing
factor (Susin et al., J. Exp.
~Lled., 184:13 31-1341, 1996). These facts have led to suggestions that
mitochondria) dysfunction, by
increasing the rate of apoptosis, is critically important in neurodegenerative
disorders including
Alzheimer's and Parkinson's diseases (Lemasters et al., J. Bioenerg.
Bionaernbr., 31:305-319, 1999;
Beal, Trends NeZrrosci., 23:298-304, 2000).
Due to the fundamental role mitochondria play in cell life and cell death,
interest in a
mitochondria) proteome map has grown significantly (Scharfe et al., NZrcleic
Acids Res., 28:155-158,
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
__ 2 __
2000; Rabilloud et al., Elect~~oplaoresis, 19:1006-1014, 1998). Such a map
would allow researchers
to compare the pattern obtained from an altered mitochondria) sample, such as
a cell line from a
patient with a mitochondria) disease, to a reference map and would provide
inforn~ation about
differences in protein expression. Up to now, most attempts to obtain a human
mitochondria) 2-D
map have involved solubilization of whole mitochondria or even whole cells
(Rabilloud et al.,
Electrophoresis, 19:1006-1014, 1995; Seow et al., Electrophoresis, 31:1787-
17813, 2000; Langen et
al., Electrophoresis, 20:907-916, 1999). This has led to elaborate two-
dimensional (2-D) patterns
containing more spots than can be optimally resolved for analysis,
particularly as many proteins
appear to be present in multiple forms due to post-translational and/or
preparative modifications (e.g.
deamidation). In addition, such maps provide little information about the
assembly-state or
functionality of individual protein complexes. Furthermore, a disproportionate
number of proteil~s in
the mitochondrion are membrane associated making them difficult to solubilize
for isoelectric
focusing.
There is an ongoing effort in several laboratories to obtain a mitochondria)
proteome for use
in diagnosis of diseases, to identify targets for drug therapy, and to screen
for unwanted drug side
effects. The most advanced human mitochondria) proteome has been reported by
Rabilloud and
colleagues (Electrophoresis, 19:1006-1014, 1998). Their approach has been to
resolve placental
mitochondria) proteins using the now classical 2-D-gel methodology of
isoelectric focusing in the
first dimension and SDS-PAGE in the second dimension. However, there are a
number of problems
with this most straightforward approach. First, the vast number of spots are
not optimally separated,
particularly as many components appear to be present in multiple f011115 due
to post-h~anslational
modification and/or modification occurring during sample preparation. In
addition, a considerable
number of mitochondria) proteins are small, i.e., MW below 10,000, and these
proteins are often
difficult to resolve by standard methods. Furthemnore, a surprisingly large
number of mitochondria)
proteins are highly basic (pKs > 9.0), and a majority of these proteins are
membrane bound. Of the
membrane-associated proteins, a high proportion is hydrophobic and difficult
to solubilize. Thus,
they are not well represented in the proteome of Rabilloud et al.
(Elecn~oplaoresis, 19:1006-1014,
1998), as these authors acknowledge.
In spite of recent advances, current 2-D-PAGE analysis is still inadequate for
separating all
of the proteins in a system, or even all of the proteins in an organelle. It
is to inadequacies in existing
separation and analysis techniques that this invention is directed.
SUMMARY OF THE DISCLOSURE
The inventors have surprisingly found that proteome analysis can be
dramatically improved
by including a preliminary separation of samples based on their interactions
with other proteins (their
tertiary structure). Examples of such preliminary separation are sucrose
gradients and non-denaturing
gel electrophoresis. Using a preliminary separation step that does not fully
disrupt the tertiary
structure of protein complexes, a third dimension can be added to traditional
proteomics analysis.
The three separations are based on (A) association of proteins in complexes,
(B) isoelectric point, and
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
__ 3 __
(C) size. Addition of the preliminary separation (e.g., separation through a
sucrose gradient) enables
detection of disturbances in protein-protein interactions in a system, such as
may be caused by
changes ui protein expression level, protein confirmation, or post-
translational protein modifications,
for example. In addition, this preluninary separation step provides the
surprising advantage of
S permitting a higher proportion of hydrophobic proteins to be separated and
identified in subsequent
analysis steps.
To address the above problems, the uiventors have developed a 3-dimensional (3-
D) system
for analysis of proteomes, such as the mitochondrial proteome. In a preferred
embodunent, the first
step involves reproducible, discontinuous sucrose gradient separation of
detergent-solubilized
proteins. The fractions obtained in this step contain protein complexes
differentiated by size. These
fractions then can be used to measure biologically relevant enzyme activities,
to separate proteins by
standard SDS-PAGE, and to resolve proteins by 2-D gel electrophoresis (e. g.,
using IEF in the first
dimension followed by SDS-PAGE in the second dimension). This approach greatly
enhances the
resolution of proteins and further provides functional information about
protein complexes withvi the
system.
Provided herein in specific embodiments are methods for creating three-
dvnensional
representations of the protein complement of a biological sample. Examples of
these methods
include at least three sequential separation phases, wherein the first is a
non-denaturuig separation
(such as a size or buoyant density gradient separation, e.g., sucrose gradient
separation, or aqueous 2-
phase partitioning, or a non-denaturing agarose gel electrophoresis
separation). The resultant
separated sample is then divided into a plurality of identifiable sub-
fractions, which occur in an
identifiable order based on fractionation or other criteria. One or more, or
all, of these fractions are
then subjected to second and third separation stages.
The second and third phases (which occur subsequent to the fwst phase but not
necessarily in
that order) can be separations based on net protein charge (e.g., isoelectric
focusing, capillary
electrophoresis, or isotachyphoresis) or protein size (e.g., SDS-PAGE, sizing
gel, or mass
spech~oscopy).
The data produced by the sequential separation of the proteins, or
representations of this
data, then can be assembled into a three-dimensional representation of the
proteins in the original
sample.
The foregoing and other features and advantages will become more apparent from
the
following detailed description of several embodiments, which proceeds with
reference to the
accompanying figures.
BRIEF DESCRIPTION OF THE FIGURES
FIG. I shows a comparison of sucrose gradient fractions, separated by SDS-
PAGE, from
three different sources of mitochondria: (FIG. IA) Bovine heart; (FIG. 1B)
Human brain; (FIG. 1C)
MRC-5 fibroblasts. The fractions were obtained from gradient B with the 35%
fraction omitted as
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
__ 4 __
described herein. Lanes 1-9 correspond to Fractions 1-9 respectively in each
gel. The samples were
applied to 10-20% SDS PAGE and stained with SyproRubyTM protein gel stain.
FIG. 2 shows TCA precipitated fractions from bovine heart mitochondria after
separation by
sucrose gradient centrifizgation. 500 p) of each fraction was TCA precipitated
and run on an 8 ~0%
polyacrylamide gel to optimize separation of proteuis in the molecular weight
range between 19-200
kDa. The gels were stained with Comassie Brilliant Blue. (FIG. ?A) Lanes 1-9
correspond to
fractions 1-9 respectively as obtained from gradient B. (FIG. 3B) Fraction 1
from gradient A.
FIG. 3 shows a Western blot analysis of mitochondria) proteins from three
different sources
after separation using sucrose gradient A. The complexes were identified by
subunit specific
monoclonal antibodies as described herein. (FIG. 3A) Western blot of subunits
from the five
respiratory chain complexes in sucrose gradient fractions from MRC-5
mitochondria. (FIG. 3B-3F)
Results of a densitometric scan of the gel of (FIG. 3A). Each respiratory
chain complex subunit was
plotted individually; the darkest intensity for each antibody was set to 100%.
Shown are gradients of
bovine heart (dotted), MRC-5 fibroblasts (solid) and MRC-5 rho0 (dashed).
FIG. 4 is a graph showilig the ATPase (solid line) and creatine kinase (dashed
line) activity
measurements measured in bovine heart mitochondria) fractions separated on
gradient A.
FIG. 5 shows 2-D gels of bovine heart mitochondria) proteins in (FIG. 5A)
fraction 3 and
(FIG. 5B) fraction 4 after sucrose gradient B. Proteins were separated on IPG
stt~ips (3-10 linear)
prior to separation on a 10% homogenous SDS-polyacrylamide gel. The gels were
stained with
SyproRubyT"1 protein gel stain and imaged using a Fuji FLA3000 scanner.
Proteins mainly present in
fraction 3 are highlighted with solid boxes, proteins mainly present in
fraction 4 are highlighted by
circles, and proteins unique in either fraction are highlighted by dashed
boxes.
FIG. 6 shows a pictorial model of an example of tln-ee dimensional protein
separation and
the information obtauied from each dimension.
DETAILED DESCRIPTION
I. A~~l'eVlCILTOJ7S
2-DE: 2-dimensional electrophoresis
COX: cytochrome oxidase
IEF: iso=electric focusing
IPG: immobilized pH gradients
LM: laurylmaltoside ((3-dodecyl maltopyranoside)
mAb: monoclonal antibody
PD: population doubling
PMSF: phenyl methylsulfonylfluoride
SDS: sodium dodecyl sulfate
II. Terms
Unless otherwise noted, technical terms are used according to conventional
usage.
Definitions of common terms in molecular biology may be found for instance in
Benjamin Lewin,
Genes b~; published by Oxford University Press, 1994 (ISBN 0-19-8542S7-9);
hendrew et al. (eds.),
The Encyclopedia of Molecular Biology, published by Blackwell Science Ltd.,
1994 (ISBN 0-632-
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
__ 5 __
02182-9); and Robert A. Meyers (ed.), Molecular Biology and BiotecJ7rzoloy: a
Compre7ensive Desk
Reference, published by VCH Publishers, Inc., 1995 (ISBN 1-56081-569-8).
In order to facilitate review of the various embodiments, the following
explanations of
specific teens are provided:
Isolated: An "isolated" biological component (such as a nucleic acid molecule,
protein or
organelle) has been substantially separated or purified away from other
biological components in the
cell of the organism in which the component naturally occurs, i.e., other
chromosomal and extra-
chromosomal DNA and RNA, proteins and organelles.
Pharmaceutical/therapeutic agent: Any agent, such as a protein, peptide (e.
g., hormone
peptide), other organic molecule, inorganic molecule, or combination thereof,
that has one or more
effects on a biological system.
Proteomics: Global, whole-cell analysis of gene expression at the protein
level, yielding a
protein profile for a given cell or tissue. The comparison of two protein
profiles (proteomes) from
cells that have been differently treated provides information on the effects
the treatment or condition
has on protein expression and modification. Subproteomics is analysis of the
protein profile of a
portion a cell, for instance of an organelle or a protein complex. Thus, a
mitochondria) proteome is
the profile of the protein expression content of a mitochondrion under certain
conditions.
Purified: The terns "purified" does not require absolute punt) ; rather, it is
intended as a
relative teen. Thus, for example, a purified protein preparation is one in
which the protein referred to
is more pure than the protein in its nah~ral environment within a cell or
within a production reaction
chamber (as appropriate). Likewise, a purified organelle preparation is one in
which the specified
organelle is more pure than in its naW ral environment within a cell, so that
only relatively
insubstantial amounts (e. g., less than 10% relative) of other organelles (or
markers for other
organelles) are present in the preparation.
Separate(d)/Separation: To spatially dissociate components, such as
biomolecules. The
components (for example, proteins or peptides) are usually separated based on
one or more specific
characteristics, such as molecular weight or mass, charge or isoelectric
point, confomnation,
association in a complex, and so forth. Separation may be accomplished by any
number of
techniques, such as sucrose gradient cent<~ifugation, aqueous or organic
partitioning (e. g., 2-phase
paaitionvig), non-denaturing gel electrophoresis, isoelectric focusing gel
electrophoresis, capillary
electrophoresis, isotachyphoresis, mass spectroscopy, chromatography (e.g.,
HPLC), polyacrylamide
gel electrophoresis (PAGE, such as SDS-PAGE), and so forth.
Once a sample is subjected to a separation, it can be divided into sub-samples
or fractions.
These fractions may be divided in an order, which may be correlated for
instance with a characteristic
that was used to separate the components. Thus, a sample subjected to sucrose
gradient separation
can logically be divided into fractions based on the final density. Proteins
or other biomolecules that
are separated by an isoelectric focusing gel can be fractionated (e.g., the
gel divided into strips) that
are correlated with their net charge. Likewise, molecules subjected to SDS-
PAGE separation can be
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
__ ( __
fractionated based on their molecular weight. The division of a separated
sample into fractions, in
some order based on that separation, is well known to those of ordinary skill
in the art.
As used herein, separation is not an absolute team (in that separation need
not be perfect or
"complete" for components to be "separated"). Thus, when a sample is subjected
to a separation
technique and the resultant separated sample is divided into fractions (e.g.,
fractions from a sucrose
Gradients, bands from a gel, and so forthj, components withui the sample can
still be referred to as
"separated" even though they occur in more than one of the fractions.
Subject: Living multi-cellular vertebrate organisms, a category that includes
both human
and non-human mammals.
Unless other,vise explained, all technical and scientific teams used herein
have the same
meaning as connnonly understood by one of ordinary skill in the au to which
the invention belongs.
Although methods and materials sunilar or equivalent to those described herein
can be used in the
practice or testing of the present invention, suitable methods and materials
are described below. All
publications, patent applications, patents, and other references mentioned
herein are incorporated by
reference in their entirety. In case of conflict, the present specification,
includuig explanations of
terms, will control. In addition, the materials, methods, and examples are
illustrative only and not
intended to be limiting.
III. Ovemiem of Sevennl Ejnbodiments
A first embodiment is a method for creating a three-dimensional representation
of a protevi
complement of a biological sample (e.g., a sample from an animal, a plant, a
microbe, a fungus, or so
forth). Examples of this method involve separating proteins contained in the
biological sample using
a non-denaturing separation process to produce a separated sample; dividing
the separated sample
into a plurality of identifiable sub-fractions having an order (for instance,
the order in which they are
removed from the separation); subjecting at least tlv0 of the sub-fractions to
at least one denaturuig
separation process (though the sub-fractions need not be subjected to the same
process) based on
protein size and at least one separation process based on protein charge
(though the sub-fractions
need not be subjected to the same process), to produce a two-dimensional
separation of the proteins
in the sub-fractions; producing a representation of each two-dimensional
separation of proteins; and
assembling (e.g., through or involving computer processing) the plurality of
representations in order
to produce a three-dimensional representation of the proteins in the sample.
In specific embodunents, the non-denaturing process comprises separation on an
osmotic
gradient, for instance a discontinuous or continuous gradient, such as a
sucrose gradient.
In specific embodiments, the denaturing separation process based on protein
size comprises
separation on a SDS-PAGE gel.
In certain embodiments, the denaturing separation process based on protein
charge
comprises separation on an isoelectric focusing gel.
It is particularly expected that in some embodiment, the denaturing separation
process based
on protein charge is carried out before the denaturing separation process
based on protein size.
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
__ 7 __
Also provided are methods for creating a three-dimensional representation of a
protein
complement of a biological sample, where the biological sample is an
organellar preparation. In
particular examples of such methods, the organellar preparation is a
preparation enriched (e.g., by
?0%, by 40%, by 60%, by 80%, or more compared to a starting sample) for
nuclei, endoplasmic
reticulum, mitochondria, plastids, lysosomes (vacuoles), peroxisomes, cytosol
components, or plasma
membranes. Specific envisioned methods are methods wherein the organellar
preparation is a
preparation enriched for mitochondria.
In fizrther methods, the biological sample is prefractionated into a plurality
of pre-fractions
prior to being separated into sub-fractions using a non-denaturing separation
process, and the three-
dunensional representation of the proteins is a three-dimensional
representation of a pre-fraction. In
examples of such methods, the method of claim 1 s, where assembling the
plurality of representations
in order to produce a three-dhnensional representation of the proteins in the
sample fizrther comprises
assembling a plurality of three-dimensional representations of individual pre-
fractions into a single
three-dunensional representation.
Particular embodiments provide methods for creating a three-dimensional
representation of a
protein complement of a biological sample, which methods involve subjecting
each of the sub-
fractions to at least one denaturing separation process based on protein size
(though not all sub-
fractions need be subjected to the same process) and at least one separation
process based on protein
charge (though not all sub-fractions need be subjected to the same process),
to produce a two-
?0 dimensional separation of the protevis in the sub-fi~actions.
The disclosure also provides three-dimensional representations of the protein
complement of
a sample, which representations are Generated by any one of the disclosed
methods.
A further embodiment includes a method for creating a three-dimensional
representation of a
protein complement of a mitochondrial sample, which method involves separating
protehis contained
?5 in the mitochondrial sample using a sucrose gradient to produce a separated
sample; dividing the
separated sample into a plurality of identifiable sub-fractions having an
order (such as the order in
which the samples are removed from the sucrose gradient); subjecting at least
some of the sub-
fractions to isoelech~ic gel electrophoresis, followed by SDS-PAGE, to produce
a two-dimensional
separation of the proteins in the sub-fractions; producing a representation of
each hvo-dimensional
30 separation of proteins having a plurality of individual features; and
assembling the plurality of
representations in order to produce a three-dunensional representation of the
proteins in the sample.
Examples of such methods fi~rther involve identifying at least one feature in
the three dimensional
representation.
Methods described herein are methods of generating a three-dimensional protein
profiles of
35 biological samples. In particular embodiments, such methods are methods of
generating a three-
dimensional protein profile for a disease or condition, wherein the biological
sample is a sample from
an organism known to be afflicted with the disease or condition. An example of
such an embodiment
method is a method where the disease or condition is linked to mitochondrial
fi~nction, the three-
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
__ g __
dimensional protein profile generated is a mitochondria) disease/condition-
linked profile, and the
biological sample comprises an organellar preparation enriched for
mitochondria.
A further embodiment is a method of screening for a compound useful in
treating, reducing,
or preventing a disease or condition linked to mitochondria) function, or
development or progression
of a disease or condition linked to mitochondria) function, which method
involves determining if
application of a test compound to a subject alters a mitochondria)
disease/condition-linked profile
produced from the subject, so that the profile less closely resembles a
mitochondria)
disease/condition-linked profile than it did prior to such h~eahnent, and
selecting a compound that so
alters the profile.
Yet another embodiment is a method of determining drug or treatment
effectiveness onside
effects, which method involves applying a drug or treatment to an organism or
a cell sample from the
organism; taking a biological sample from the organism or the cell sample from
the organism;
analyzing the biological sample to produce a test three dunensional protein
profile for the subject;
comparing the test three dunensional protein profile for the organism with a
control three dimensional
protein profile (for instance, a profile generated prior to the treatment or
drug application); and
drawing conclusions about the effectiveness or side effects of the drug or
treatment based on
differences or similarities between the test three dimensional protein profile
and the control three
dimensional protein profile. In specific examples of such methods, the drug or
treatment is a dmg or
treatment for a mitochondria)-linked disease or condition, and the test and
control three dimensional
protein profiles are three dimensional mitochondria) protein profiles.
11! Enhanced Protein Analysis
The methods described herein provide enhanced protein separation techniques,
which
facilitate proteomic analysis of living systems. In one embodiment, sucrose
gradient centrifugation is
combined with two-dimensional gel electrophoresis to produce a three-
dimensional representation of
the proteome. The resulting three-dimensional separation of proteins addresses
several of the
problems encountered during previous attempts to obtain complete proteome
maps, such as resolution
of proteins and solubility of hydrophobic proteins during isoelectric
focusing. In addition, the new
protein separation techniques described herein provide functional infornation
about protein
complexes within an organelle (or other subcellular division) that is not
obtained with two
dimensional gel electrophoresis of, for instance, whole mitochondria or a
whole cell.
As mitochondria play critical roles in both cell life and cell death, there is
great interest in
obtaining a human mitochondria) proteome map. Such a map is expected to be
useful in diagnosing
diseases, identifying targets for drug therapy, and in screening for drug
effects and side effects. The
mitochondrion was therefore used as an example system to explore the potential
of the described
protein separation techniques. Other protein systems can be analyzed with the
methods provided
herein in, including for instance other organelles, as well as specific tissue
or cell types. The methods
herein are therefore not intended to be limited to analysis of mitochondria)
protein analysis.
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
__ 9 __
In a representative example described in detail herein, mitochondrial proteins
are separated
based on their associations in the organelle (as shown in FIG. 6). A
significant number of
mitochondrial proteins exist ir7 viuo as multi-polypeptide complexes. Examples
include the five
complexes of the oxidative phosphorylation machinery (Saraste, Science,
?83:1488-1493, 1999), the
mitochondrial ribosome (Curgy, Biol. Cell, 54:1-38, 1985), the mitochondrial
nucleoid (a complex of
mtDNA and assorted nucleotide binding proteins; Newman et al., Nucleic Acids
Res., 24:386-393,
1996), as well as the TIM and TOM complexes (Pfanner and Meijer, CZnw. Biol.,
7:8100-8103,
1997) involved in protein import into the organelle, and the permeability
transition pore, which has
been lulled to apoptosis (Fontaine and Bernardi, J. Bioeneng. Biomembr.,
31:335-345, 1999).
The first step of this embodiment of the described separation analysis is a
discontinuous
sucrose gradient that separates the component polypeptides by the sizes of the
complexes in which
they participate. The effectiveness of this separation method is demonstt~ated
herein using the
OXPHOS component proteins, whose location could be readily followed using this
laboratory's set
of monoclonal antibodies specific to these complexes. During the course of
this study, the protein
patterns reported here have been obtained more than 50 times, confirming that
this method is highly
reproducible. In addition, a similar separation of complexes is obtained for
various tissue samples,
demonstrating the broad applicability of the described separation techniques.
An advantage of the sucrose gradient pre-fractionation step (or other primary
separation) for
proteome analysis is that it separates the total protein complement of the
starting sample into
"workable" fractions for subsequent electrophoretic separation, as well as
providing functionally
relevant information (i.e. assembly state, activit)~) about the various
proteins. This simplification
afforded b~~ separating total proteins into fractions allows some of the
problems encountered in
previous proteome attempts to be dealt with in a manageable manner. With fewer
proteins per
sample, the same fraction can be run on multiple Gels of varying isoelectric
point ranges, which helps
solve the "range of isoelectric point" problem while still producing svnple
enough patterns for
subsequent analysis.
In addition, during conventional 2-D electrophoresis, many hydrophobic
proteins are lost
during the initial isoelectric focusing step. Infornation about the
hydrophobic proteins, which would
have normally been lost during the isoelectric focusing step, now can be
obtained by identifying the
fragment sizes in the mass spectrometry analysis which are present in the one-
dimensional gel but are
absent in the 2-D-gel. With the herein-described separation methods, the same
sample can be
subjected both to one-dunensional SDS-PAGE and to 2-D gel electrophoresis
(consisting of
isoelectric focusing followed by SDS-PAGE). There are fewer proteins in each
of the fractions
produced using sucrose-gradient pre-fractionation than in whole mitochondrial
samples. Recent
advances in mass spectrometry allow for identification of individual
components in mixtures of
proteins, and therefore mass spectrometry can be carried out on both the
individual spots in the
described 3-D-gel and on the equivalent size band from the one-dunensional
gel.
Two other useful aspects of the described prefractionation are that activity
measurements
can be obtained from the same fractions being subjected to electrophoresis and
that proteins present
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
-- 10 --
in low copy number can be concentrated into one of the fractions. Activity
measurements provide the
added information of whether the complexes being studied are functional
(and/or to what extent they
are functional). By concentrating low copy number proteins into smaller
fractions, detection is made
easier. By way of example, results are discussed below that demonstrate that
the subunits of
Complex I are concentrated in fraction I . Samples produced using methods
described herein can and
are being examined to detect proteins of even lower abundance such as SURF-1,
a protein that
catalyzes cytochrome c oxidase assembly.
A. Types of Proteomes
The protein separation methods described herein can be used to provided
enhanced
separation and identification to any protein system, and are not limited to
the example systems
presented in detail herein. In essence, any proteome can be generated using
the described techniques;
the larger the number of component proteins, the more advantageous it is to
pre-fractionate the
protein sample prior to electrophoretic analysis as described herein. Thus,
the described methods can
be used to generate proteomes from various organisms, including microbes,
plants, animals (for
1 ~ instance, humans).
The described enhanced proteui separation methods can also be used to produce
proteomes
for sub-cellular fractions to produce subproteomes (e.g., on an organelle by
organelle basis, or system
by system within a cell). Sub-proteomes can be produced from any cell fraction
that can be reliably
produced. Representative examples of sub-proteomes that can be analyzed using
the described
30 enhanced protein separation methods include (but are not intended to be
limited to): nuclear,
mitochondrial, lysosomal/vacuolar, endoplasmic reticulum ER), secretoiy system
as a whole, plastid
(e.g., chloroplast), peroxisomal, and cytosol (not all of which will be found
in all cells).
Proteomes can be assembled for whole cells using the described techniques; the
proteins
from whole cells are advantageously sub-divided (for instance, by organelle)
prior to non-denaturing
25 (e.g., sucrose gradient] separation and subsequent electrophoretic
analysis. Thus, individual sub-
cellular proteomes such as those described above can be assembled (for
instance, using a computer
system) into the comprehensive proteome of a whole cell. However, such sub-
division of the cell is
not essential. Entire cell protein preparations can be separated on long
sucrose gradients, and
numerous fractions collected for subsequent denaturing analysis. It is,
however, advantageous in
30 some embodiments to take advantage of the compartmentalization of
eukaryotic cells to further
simplify the protein profile being examined.
The described techniques also permit enhanced detection of protein-interaction
perturbations
caused by protein co- and/or post-translational modification. Since such
modifications often
influence and/or control the ability of proteins to interact in complexes, the
non-denaturing separation
35 that is integral to the described protein separation methods perniits
separation of differentially
modified protein forms. This is believed to simplify the interpretation of
proteomic data, as well as
providing more inforniation on the functional forms of specific proteins. Co-
and post-translational
modifications are discussed, for instance, in Chapter 4 of Wilkins et al.,
(Proteorne Research: NeH
Frontiers in Fzn7ctional Genontics, Springer-Verlag, Berlin, 1997; ISBN 3-540-
62753-7).
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
-- 11 --
The pivotal importance of the mitochondrion makes it an excellent example
proteux system
in which to demonstrate the effectiveness and reliability of the described
enhanced protein separation
techniques. The mitochondrion is important in several cellular processes,
including the generation of
"energy" and programmed cell death (apoptosis). In addition, defects in this
organelle contribute to,
and are frequently a primary cause of, many human diseases. These defects are
often caused by
mutations in mitochondrial proteins such as enzymes uivolved in fatty acid
metabolism and oxidative
phosphoiylation (Eaton et al., Biocl7em. J., 320:345-357, 1996; Wallace,
Science, 283:1482-1488,
1999). Freidrich's Ataxia is just one example of a disorder that is known to
be caused by a mutated
mitochondrial protein (Lodi et al., Proc. Nat!. Aca~l. Sci. USA, 96:11492-
11495, 1999). Other
diseases linked to defective mitochondrial proteins are being reported with
increased frequency
(Scharfe et al., Nucleic Acids Res., 28:155-158, 2000). Mitochondria also play
a key role ui apoptosis
or progranmied cell death (Mignotte et al., Eur. J. Biochem., 252:1-15, 199S).
Enhanced rates of cell
death, due in part to mitochondrial dysfunction, are now considered to be an
important component of
Alzheimer's, Parkinson's, and Huntington's diseases (Lemasters et al., J.
Bioene~g. Biomentbr.,
31:305-319, 1999; Beal, Trends Neurosci., 23:298-304, 2000; Schapira,
BioclTinz Biopl?vs. Acta,
1410:159-170, 1999). Mitochondrial dysfunction can also lead to decreased
rates of cell death, and
roles for mitochondria in cancer have been described (Polyak et al., Nat.
Genet., 20:291-293, 1998;
Fliss et al., Scienee, 287:2017-2019, 2000). In addition, many drugs used in
treatment of diseases
such as cancer and AIDS have mitotoxic effects. For instance, AZT can be
problematic for patients
due to severe disruptive mitochondrial effects (Yerroum et al., Acta
Nea~ropathol., 100:82-86, 2000).
B. Separation Techniques
Certain enhancements arising from the separations described herein are
accomplished by the
combination of a non-denat<iring pre-separation of a protein sample into less-
complex sub-samples
(fractions), followed by subsequent denaturing separation. Various specific
separation techniques
can be used for each of these two portions of the separation, and indeed it is
contemplated that
different separation techniques can be employed to separate different sub-
fractions from the same
original biological sample. In general, however, the first separation
technique employed is one that
retains or substantially retains the functionality of at least one complex of
interest in the protein
sample.
One example of the first separation is a size or buoyant density gradient
separation method,
such as a discontinuous sucrose gradient, that separates the component
polypeptides of the sample by
the sizes of the complexes) in which they participate. Sucrose gradients for
the separation of
proteins are well known, and modifications to the disclosed sucrose gradient
method are
contemplated. Such modifications may include the use of a continuous rather
than discontinuous
gradient and different gradient conditions (for instance, different sucrose
concentrations, different
buffers, or different osmoticum). The length of the gradient can also be
varied, with longer gradients
expected to give better overall separation of proteins and protein complexes,
and to provide a larger
number of fractions that are then each individually analyzed using a
denaturing system.
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
-- 12 --
Other "mild" separation techniques that are suitable for the first separation
phase include
aqueous 2-phase partitioning and non-denaturing agarose gel electrophoresis
separation (such as
native blue gels).
Once the original protein sample is pre-fractionated into a few to several
fractions, one or
more usually two additional separations are performed; the order of these
subsequent separation
phases is not critical, but for ease of description they will be referred to
as the second and third
separation phases.
In specific embodiments, each of the fractions (or a select subset of them,
for instance a
cluster of fractions, every other fraction, every third, and so forth)
produced by the first separation is
further separated using net charge and size, usually in a denaturing system.
For instance, in the
second separation phase of the procedure, the individual proteins in a complex
are separated by net
charge. Typically, this occurs by separation in an isoelectric focusing (IEF)
gel. Other techniques for
separating and isolating the proteins include capillary electrophoresis or
isotachyphoresis. In many
ictstances, non-protein components in the sample are removed during
preparation of the samples) for
IEF. '
In the third separation, the individual proteins are separated by size (e. g.,
by SDS-PAGE or
sizing gel, or by mass spectroscopy). Mass spectroscopy may be performed after
separated proteins
are fragmented with an enzyme (such as trypsin) or a chemical cleaving agent
(such as cyanogen
bromide). The peptide mass profile (peptide fmge~print) obtained from mass
spech~ometry is
compared with theoretical fragmentation patterns derived from sequence date in
genomic databases in
order to aid in identifying the proteins. Additionally, Edman sequencing can
be used in identifying
peptides.
Representative examples of such separation techniques are presented below, in
Examples 3
and 4; representative results from an analysis of the mitochondrial proteome
are presented in the
accompanying figures. Other examples of two-dimensional electrophoretic
analysis are well known;
see, for instance, Chapter 2 of Wilkins et al. (Proteome Research: New
Frontiers in FZmctional
Geno~rrics, Springer-Verlag, Berlin, 1997; ISSN 3-540-62753-7).
Proteins can be visualized on denaturing gels using any of various known
stains. However,
some stains are more advantageous than are others. For instance, the use of
SyproRubyTM dye
(Molecular Probes, Oregon) allows seamless throughput fr0111 the gels to mass
spectrometry, as well
as providing the best sensitivity available to date in staining individual
proteins for identification.
Traditional buffering systems can be used for separating proteins in the
component
fractionations of the described systems. However, as is well known to those of
ordinary skill in the
art of protein separation, minor modifications to such buffer conditions can
be made to optimize the
buffers for individual raw protein preparations (see, for instance, the
discussion of two-dimensional
electrophoresis in Chapter 2 of Wilkins et al., Proteome Research: Nev~
Frontiers in FarrrctiorTal
Genonrics, Springer-Verlag, Berlin, 1997; ISBN 3-540-62753-7). Possible
modifications include, for
instances, changes in the pH, osmoticum (e.g., sucrose), salt content and/or
concentration of
individual solutions (e.g., the solutions used to make gradients, gels, and/or
the buffers used to run
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
-- 13 --
the gels). The temperature, voltage, and amperage at which individual gels are
run also can be
modified, as can the speed and duration of gradient equilibration and
centrifizgation. One of ordinary
skill in the relevant art will know not only how to vary these and other
relevant conditions, but will
also know the effects such variations are likely to have on the operation of
the system (e.g., the likely
effects on protein separation). All such minor variations of conditions that
are used to optimize
separation conditions are encompassed herein.
C. Identification of Individual Features
Proteins separated using the herein-described enhanced techniques can be
analyzed using
any of various well-known techniques. For instance, those protein
identification techniques currently
used to analyze individual protein features in 2-D proteomes can be used. Such
techniques are well
known, and examples can be found for instance in Wilkiils et al. (Proteome
Research: New Frontiers
in Famctional Genomics, Springer-Verlag, Berlil~, 1997; ISBN 3-540-62753-7).
In paaicular,
Chapter 3 of Wilkins et al. provides insights on such techniques. Examples of
applicable protein
identification techniques include (but are not intended to be limited to)
protein activity assays (see
Example 5); antibody recognition (Western mapping using, for instance, mAbs to
individual known
proteins; see Example 6); direct comparison to previous proteomic maps (on
which features have
been identified through any method); mass spectrophotometry; and database
screening for peptide
sequence matches, for instance using peptides) removed from a gel or blot.
D. Raw Data, Data Assembly, Automation, and Data Analysis
The fornl of data presentation from the described enhanced protein separation
methods is
largely a matter of individual preference. For instance, the individual gels
produced can be viewed
individually, as is discussed below in specific examples. Though the raw data
resulting from the
sequential protein separations can be read by an individual, it is
advantageous considering the vast
amount of information contained in each protein profile to process the data
using a computer.
In certain embodiments, therefore, it is advantageous to scan staaied gels
and/or blots
produced using the herein-described separation methods into a computer for the
processuig and/or
analysis of the raw data. Programs exist, and more are being developed, that
permit subtraction of gel
or stain artifacts, calculation of relative and/or apparent pI, molecular
weight, and amount of each
protein feature, and/or calculation of protein-protein interactions (for
instance by comparing different
pre-fractionated samples produced using the described methods).
The computer-assisted comparison of multiple gels can be used to detect
changes in
proteomes, which changes can be linked to (for instance) disease progression,
environmental or other
stimuli, clinical treatment, developmental changes, and so forth. Such
comparisons also pern~it
standardization of gel results, for instance by consistently identifying
features bet,veen different gels
(such as gels produced in different laboratories, or using proteins from
different samples).
Computer scanning of the described protein gel profiles also permits the
assembly of a set
(or sub-set) of pre-fractionated samples into a three-dimensional map, such as
is displayed in the
simplified model shown in FIG. 6. In this example, individual gels (the Z
dimension) represent
different fractions from a non-denaturing discontinuous sucrose gradient; each
fraction has
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
-- 14 --
subsequently been further separated using isoelectric focusing (the X
dimension) and denaturing
SDS-PAGE analysis (the Y dimension).
1! Applications
With the provision herein of systems for enhanced protein separation and the
generation of
fme-detail, three-dimensional representations of protein profiles of tissues,
cells, organelles and so
forth, the use of these representations to accurately compare protein profiles
under different
conditions is enabled. This excellent comparison system can be used, for
instance: (1) to identify
novel or previously unidentified proteins (for instance in an organelle, such
as the mitochondrion);
(2) in detection and diagnosis of disease, disease state, or prediction of
disease progression
(prognosis); (3) in development and testing of pharmaceutical agents; (4) for
tracking of drug efficacy
in a subject, and (5) for tracking of drug toxicity in a subject. Sample
comparison can be between
healthy and diseased tissues (e.g., biopsy) or cells (or cell cultures),
diseased tissue at different stages
(e.g., different cancer stages or the stages of other progressive diseases),
tissue before and after drug
or other treatment, and so forth.
Some clinical, biomedical, and biological applications of proteomics are
described, for
instance, in Chapters 8 and 9 of Wilkins et al. (Proteome Research: Nem
Frontiers in Fmzctional
Ger~orrrics, Springer-Verlag, Berlin, 1997; ISBN 3-540-62753-7).
Specific embodiments are illustrated by the following non-limiting Examples.
EXAMPLES
General Materials and Methods
Materials used for biochemistry were fi~om Sigma Chemical Company (St. Louis,
MO),
unless otherwise stated. Laurylmaltoside (LM) was purchased from Calbiochem
(La Jolla, CA). IPG
strips 3-10 (18cm) were purchased from Amersham Pharmacia Biotech
(Piscata~~~ay, NJ).
SyproRubyT"' protein gel stain was obtained from Molecular Probes Inc.
(Eugene, OR). All
chemicals used for 2-D electrophoresis were from Genomic Solutions (Ann Arbor,
MI).
Trichloroacetic acid (TCA) precipitation was done according to Petersen (Anal.
Biochenr.,
83:346-356, 1977). One-dunensional mini gels were nm essentially according to
Laemmli (Natarre,
227:680-685, 1970) using 10-20% gradient polyacrylamide gels. The gels were
stained with
Coomassie Brilliant Blue (Downer et al., Biochemistry, 15:2930-2936, 1976) or
SyproRubyT"'
protein gel stain using known procedures (Berggren et al., Electrophoresis,
21:2509-2521, 2000).
Example I: Preparation of a Biological Sample
This example provides descriptions of how one sample type, isolated
mitochondria, can be
prepared from various tissues for analysis using the separation systems
described herein. Other
tissue, cell, or subcellular preparations also can be examined; such samples
can be prepared using any
conventional means.
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
--15--
In certaui embodiments, it is beneficial that the final preparation is not
substantially
denatured (e.g., so that in vivo protein-protein interactions have been
substantially maintained). W
general, the more pure the target sample is, the better the results will be
from the proteomic analysis.
Preparation of mitochondria from bovine heart
All steps for purifying mitochondria were done at 4°C unless otherwise
stated.
The ventricles of a fresh bovine heart were cleaned of any connective tissue
and fat before
being minced into small pieces. About 600 ml of a Tris/sucrose buffer (0.2 mM
EDTA, 0.25 M
sucrose, 10 mM Tris/HCl pH 7.8) was added to 300 g of minced tissue and then
blended in a blaring
Blender for 30 seconds at high speed followed by 30 seconds at low speed. The
pH was checked and,
I O if necessary, adjusted to 7.8 with 2 M Tris before repeating the blending
and adjustment of pH. The
blended tissue was homogenized further with an Ultrahmex (Kinematica,
Switzerland) (3.5 seconds
at speed 9) followed by additional pH adjustment if needed. The homogenate was
centrifuged at
185 x g for 15 minutes in a KAJ-9 (Beckman, USA) rotor and the supernatant was
filtered tiwough
four layers of cheesecloth. The filtrate was homogenized in a glass
homogenizer with a tight fitting
15 Teflon pestle and centrifuged at 740 x g for 10 minutes in a ICAJ-9 rotor.
The pellet was discarded
and the resultuig supernatant was centrifuged at 20,600 x g for 15 minutes in
a GSA rotor to pellet the
mitochondria. The pellets were washed twice in the Tris/sucrose buffer
supplemented with 0.5 mM
PMSF before the final pellet was resuspended in a small amount of buffer.
After deterniining the
protein concentration the mitochondria were frozen at -80° C.
20 Preparation of mitochondria from NIRC-5 fibroblasts
MRC-5 fibroblasts were obtained from the American Type Culture Collection. The
population doubling (PD) of the cells was in the range of 35-45 before
harvesting to isolate
mitochondria. Rho°- MRCS fibroblasts were derived by culturing MRC-5
fibroblasts (PD=28-30)
continuously in media supplemented with 50 ng/ml ethidium bromide for a
further 16 PD's. All cells
25 were grown as described before (Marusich et al., Biochi»r. Biophvs. Acta,
1362:145-59, 1997) in high
glucose Dulbecco's modified Eagle's medium, supplemented with 10% bovine calf
serum, 50 pg/ml
uridine, 110 ~ghnl pyruvate, and 10 mM HEPES buffer to maximize growth rates.
For the preparation of mitochondria, 12-16 plates ( 150 mM diameter) of
confluent MRC-5
fibroblasts were harvested, and the cells were washed three times in Ca'+,
Mg'+ free phosphate
30 buffered saline (CMF-PBS). To improve the cell disruption, the cell pellets
were frozen at -80° C
for at least an hour. After thawing, 5 ml of homogenization buffer (0.5 pg/ml
leupeptin, 0.5 pg/ml
pepstatin, 1 mM PMSF, 350 mM sucrose, 1 mM EGTA, 1 mM EDTA, 10 mM HEPES/NaOH
pH 7.4) was added and the pellets were homogenized in a glass homogenizer with
a Teflon pestle.
The homogenate was centrifuged (1,500 x g, 10 minutes, 4° C) and the
supernatant was transferred
35 into a clean tube. The homogenization was repeated twice with the pellet,
and the supernatants ware
combined. The three combined supernatants were centrifuged (1,500 x g, 10
minutes, 4° C) and the
pellet discarded. The resulting supernatant was once more centrifuged to
pellet the mitochondria
(10,000 x g, 12 minutes, 4° C). The supernatant was discarded and the
pellet was resuspended in
ml wash buffer (0.5 pg/ml leupeptin, 0.5 pg/ml pepstatin, 1 mM PMSF, 250 mM
sucrose, 1 mM
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
-- 16 --
EGTA, I mM EDTA, 10 mM Tris/HCl pH 7.5). The centrifugation was repeated
(10,000 x g, 12
minutes, 4° C) and the final pellet was resuspended in 200-500 p) wash
buffer. After measuring the
protein concentration, the mitochondria were frozen at -80° C.
Preparation of mitochondria from human brain tissue
Human brain tissue was obtained from the Harvard Brain Tissue Resource Center,
which is
supported in part by PHS grant number MH/NS 31862. The mitochondria from human
brain tissue
were essentially prepared as described for MRC-5 fibroblasts and were a kind
gift of Dr. Leslie A.
Shinobu (Massachusetts General Hospital).
Further modifications in the described methodologies may improve the data that
can be
produced using the described system for protein separation and proteome
analysis. One such
modification would be to improve the purity of the mitochondria) preparation
used. Such
modification likely may be limited by recent evidence of interaction between
the mitochondrion and
other organelles, e.g., the ER (Rizzuto et al., Science, 280:1763-1766, 1998).
Example 2: Non-denaturing Separation of the Biological Sample
This is a representative example of a non-denaturing separation technique,
discontinuous
sucrose gradient analysis, which can be used to separate biological components
based on their
protein-proteui interactions.
Separation of mitochondria) proteins by sucrose gradient fractionation
Two slightly different sucrose gradients have been employed for the separation
of
mitochondria) complexes after extraction. The first gradient is optimized for
the purification of the
respiratory chain complex I, whereas the second is optvnized for the use in 2-
DE (two dunensional
electrophoresis). These gradients are referred to herein as gradient A and
gradient B, respectively.
Mitochondria prepared from three different sources (bovine heart, cultured MRC-
5
fibroblasts, and human brain) as described above, were solubilized for
analysis using 1 % LM.
Mitochondria (1-5 mg) were pelleted (TLA 100.2 Beckman rotor, 10,000 x g, 10
minutes, 4° C) and
resuspended at a protein concentration of 5 mg/ml in 100 mM Tris/HCI, 1 mM
EDTA, pH 7.5, 1
~g/ml pepstatin, 1 pg/ml leupeptin, I mM PMSF, 1 % LM. The mitochondria were
incubated in this
solution fur 20 minutes on ice with stirring, before the membranes were
pelleted again by
centrifugation (TLA100.2, 185,000 x g, 20 minutes, 4° C). The
supernatant (250 ~1, 500 p), or 1 ml)
was layered on top of a sucrose gradient. Composition of gradient A: 250 ~I
(35%), 500 ~1 (30%),
750 ~1 (27.5%), )ml (25%), 1 ml (20%), 1 ml (15%). Gradient B: 500 p) each of
the following
sucrose concentrations: 35%, 32.5%, 30 %, 27.5%, 25%, 22.5%, 20%, 17.5%, 15%.
The 35%
fraction was omitted, when 1 ml of supernatant was to be applied to the
gradient. Both gradients
were centrifuged overnight at 4° C (150,000 x g, 16.5 hours, SW 50.1,
acc. 7, dec.7). All sucrose
solutions contained 100 mM Tris/HCI, 0.05% LM, 1mM EDTA. The sucrose gradient
was
fractionated into nine fractions from the bottom of the W be into 500 p)
fractions, which were frozen
at - 80° C.
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
__ 17 __
Example 3: Denaturing Separation of the Biological Sample
This example provides one method for fizrther separating proteins in
fi~actions of a sucrose
gradient, using denaturing gel electrophoresis, specifically SDS-PAGE. In some
embodiments, this
separation step is perfornied immediately after separation of the sample using
a non-denaturing
system (e.g., sucrose gradient fractionation). In other embodiments,
fractionated samples are first
subjected to isoelectric focusing gel analysis, then applied to a denaturing
gel for final analysis.
For SDS-PAGE analysis, 10-20 u1 of each fraction was loaded per lane. The
composition of
fractions after SDS-PAGE and subsequent staining with SyproRubyT"' protein gel
stain is shown in
FIG. I .
Results
There is considerable difference in the overall staining pattern between the
three different
tissue samples, but this is to be expected for several reasons. First, heart
tissue is rich in
mitochondria and the mitochondria are easily purified essentially free of
other organellar membranes.
This is not the case for brain or cultured fibroblasts. Brain and fibroblast
mitochondria isolated by
I S differential centrifugation can contain many other vesicular membranes
that are closely connected to
the mitochondria, including endoplasmic reticulum (ER) and the Golgi
apparattis. Indeed, small
amounts of both ER and Golgi proteins have been shown to be present in our
purified brain and
fibroblast mitochondrial samples as indicated by Western blot analysis using
organelle specific
antibodies. Furthermore, based on quantitative Western blotting with mAbs to
each complex, the
levels of respiratory chain proteins per mg total mitochondria protein is
three times higher in heart
than in brain and 6-7 times higher in heart than u~ cultured fibroblasts.
The experiments shown in FIG. I use only a portion of each sucrose gradient
fraction,
leaving sufficient material for enzyme assays and additional gel
electrophoresis analysis. However,
to ensure better visualization of low copy number proteins, entire sucrose
gradient fractions may be
TCA precipitated and subjected to 10-20% gradient SDS-PAGE (FIG. 2a,b). FIG.
2b and lane 1 of
Figures 1 and 2a include proteins in complexes larger than 700,000 Da that are
still assembled after
1 % LM treatment. As shown in FIG. 2b, the gel pattern obtained from the first
fraction of Gradient A
matches previously published patterns for complex I (Walker et al., Methods
E~7~~mol., 260:14-34,
1995) for the molecular weight range analyzed. This complex contains at least
42 different
polypeptides (Skehel et aL, FEBS Lett., 438:301-305, 1998), approximately half
of which are larger
than 19 kDa, and qumtitative estimates of the levels of complex I in beef
heart mitochondria range
from 60 to 130 pmol/ mg mitochondrial protein (Smith et al., FEBSLett., I
10:297-282, 1980;
Albracht et al., FEBS Lett., 104:179-200, 1979). Thus, complex I is enriched
in fraction 1 after the
gradient and can be easily visualized after TCA precipitation.
Example 4: Second Denaturing Separation of the Biological Sample
This example provides one method for further separating proteins in fractions
of a sucrose
gradient, using isoelectric focusing gel electrophoresis.
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
--ls--
For the IEF dimension, 50-100 p1 sample as prepared in Examples I and 2 was
used fur each
strip. To each sample, 3 p1 10% SDS and 12.5 p1 10% LM were added before
adjusting the total
volume to 500 p1 with urea sample buffer (5 M urea, 2 M thiourea, 2 % CHAPS, 1
% zwittergent 3-
10, 0.8% carrier ampholytes, 65 mM DTT). Each IPG (immobilized pH gradient)
strip was
rehydrated in this solution overnight. The IEF dimension was run in the pHaser
isoelechic focusing
unit (Genomic Solutions) as suggested by the manufacturer.
After reaching equilibrium, each ship was incubated in 2 ml of 375 mM Tris, 50
mM DTT,
3% SDS pH S.6 for 10 minutes at room temperature with gentle shaking before
being transferred to
the second dimension.
For the second dimension, 10% polyacrylamide gels with a pH of 9.2 or 17.5%
standard
homogenous slab gels, both with 4.5% stacking gels, were used. The gels were
wn in the
Investigator 2-D gel tank (Genomic Solutions) according to the manufacturer's
protocol. After
completion of the run, the gels were fixed in 10% methanol, 7% acetic acid for
one hour and stained
with SyproRubyT"' protein gel stain as described in Berggren et al.
(Elecb~oplzoresis, 21:2509-2521,
2000).
Imaging of the gels was carried out with an FLA3000 fluorescent image analyzer
(Fuji
Photo Film, Tokyo, Japan) with a 473 nm excitation filter and a 5S0 mn long
pass emission filter.
Results
Representative results are shown in FIG. 5, which shows the profiles of two
fractions of the
sucrose gradient of bovine heart mitochondria. The protein profiles of the
prefractionated
mitochondria are greatly simplified compwed to typical 2-D profiles of whole
mitochondria.
Fraction 3 contains 56 identifiable spots at the levels of protein loaded
while fraction 4 contains about
90. v~ contrast, whole mitochondria preparations loaded at an equal protein
amount show 350 clearly
identifiable spots on a single gel. The number of spots on a single gel ca~~
be significantly uicreased
by loading more protein; the resolution of proteins, however, decreases as the
spot number increases.
Using the disclosed 3-D methods, there is some overlap of polypeptide content
between fractions,
which may be seen as a complication in that the same protein is identified
more than once. However,
it is also an advantage in that patterns may he aligned using common
"landmarks," thus facilitating
profile comparison and assembly of a three-dinnensional visualization system.
In FIG. 5, selected
spots with very different vitensities in the two fractions, or that are unique
to an individual fraction,
are circled or encased by squares.
Identification of the various spots can be carried out using mass
spectrometry, for instance.
Other approaches also can be used to identifying each spot, including complex-
specific purification
and immunologic (e.g., mAb based) identification.
Example 5: Identification of Individual Features in the Proteome: Activity
assays
This example provides illustrations of specific representative methods that
can be used to
identify individual proteins (features) within the multi-dimensional protein
profiles described herein.
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
-- 19 --
ATP hydrolysis and creative kinase activity measurements
To show that protein complexes are still functional after separation on the
sucrose gradient,
ATPase and creative kinase activity measurements were carried out on each
fraction of the sucrose
gradient of bovine heart mitochondria. For the activity measurements, a 20 p1
aliquot from each
sucrose gradient fraction was used. ATP hydrolysis was measured with a
regenerating system as
described by Lotscher et al. (Biochemistry, 23:4140-4143, 1984). The creative
kinase activity was
measured according to Bucher et al. (Handbuch den physiologisch- zrnd
pathalogisch-chemischen
.4nalyse, Hoppe-Seyler/Thierfelder, vol. VI/A, Springer, Berlin, Gottingen,
Heidelberg, 1964, pp.
292-339.).
The activity in each fraction is expressed as a relative percentage of the
maxunum activity in
the peak fraction (FIG. 4). The highest ATP hydrolysis activity was measured
in fraction 4, which is
in agreement with the position of complex V in the gradient, as indicated by
the Western blot for
complex V-a. Creative kinase has been reported to form an octameric complex
with an estimated
molecular weight of 400 kDa (Schlegel et al., Biol. Chenr., 263:16942-16953,
1988). Monomeric
bovine heart complex V is estvnated to be 550 kDa (Schagger et al., Anal.
Biochem., 217:220-230,
1994). The creative kinase activity peaks slightly after the ATP hydrolysis
activity, in agreement
with these estimated molecular weights.
Example 6: Identification of Individual Features in the Proteome: Western
Blotting
In addition to the above-described methods, monoclonal antibodies can be used
to identify
individual proteuis separated using the techniques described herein. At this
time, mAbs specific to
seven subunits of complex I, t<vo of complex II, three of complex III, ten of
complex IV, and three of
complex V are available for such identification, for instance. This example
provides descriptions of
Western blotting using some of these mAbs to identify specific protein spots
separated as described
above.
Western blotting of one dimensional gels
Western blotting was done essentially according to Mamsich et al. (Biochim.
Biophys. Acta,
1362:145-59, 1997) with the following modifications. Proteins were transferred
to 0.45 pm
polyvinylidine difluoride (PVDF) membrane (Millipore, Bedford, MA) using a
semi-dry transfer
system (Amersham Pharmacia Biotech) according to the manufacturer's
specifications. Reactive
bands were detected using the ECL PIusTM detection reagent (Amersham Pharmacia
Biotech) and
were imaged using the image analyzer Storm 860 (Molecular Dynamics, Sunnyvale,
CA).
Fluorescence was quantified using NIH Image. All antibodies used in this study
were prepared in the
monoclonal antibody facility at the University of Oregon. The antibodies were
used at the following
concentrations: Complex I 39 kDa (2 ~g/ml), complex II 30 kDa (5 pg/ml),
complex III Core 2 (0.4
pg/ml), complex IV COX II (2 pg/ml), complex IV COX Va (2 ~g/ml), complex V
alpha (4 pg/ml).
The antibodies were all mouse monoclonals.
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
__
Results
As shown in FIG. 3a, complex I runs the furthest in the gradient (based on the
39 kDa
polypeptide), followed by complex V (a subunit) and complex III duner (Core
?), complex IV (COX
Va and II), and finally complex II (30 kDa subunit). This ordering of the
complexes from highest
molecular weight to lowest molecular weight is in accordance with previous
estimates of their
molecular weights (Schagger et al., Anal. Bioche»a., 217:220-230, 1994).
Densitometric scans of the Western blots from either MRC-5 fibroblasts or
bovine heart
proteins were quantified and, for convenience, the relative expression levels
of each subunit in the
various fractions were expressed as a percentage of the highest intensity band
in the gradient. The
distribution of each subunit in the gradient was then plotted (FIG. 3b-f). The
broad distribution seen
for complex III likely arises in part because this complex can be a monomer as
well as a duner. The
ATP synthase also broadly distributes; this is likely due to the partial
disruption of the F,F° into F,
and F° components.
These plots are not representative of the absolute levels of the complexes,
because the blots
were developed to identify even small amounts of each complex. Therefore, the
levels of protein
present in the peak fractions are grossly under represented because of
antibody sat<iration effects. A
better measure of the levels of the complexes present in the fractions is the
staining intensity of the
bands in FIG. l, which show that complexes III and V are concentrated in
fraction 4. Nonetheless,
the plots do reveal that the distribution of each complex in the sucrose
gradient is nearly identical for
30 the bovine heart and MRC-5 mitochondria) extracts.
In order to show that the sucrose gradient is sensitive to molecular weight
changes and
complex assembly, mitochondria from MRC-5 fibroblast lacking mitochondria) DNA
(Rho°) were
used. Though these fibroblasts are respiration-deficient, they can be cultured
in a medium favoring
glycolysis. Western blotting revealed a considerable shift in the distribution
of subunits for each of
the complexes with mitochondrially encoded subunits (FIG. 3 b-d, f). As
expected, only complex II,
which does not contain any mitochondrially-encoded subunits, failed to shift
positions in the gradient
(see FIG. 3e).
Western blotting of two-dimensional gels
For Western blotting of t,vo-dimensional gels, 10-20 p) of each fraction was
loaded per lane.
Using appropriate antibodies, Fractions 4 and 5 were analyzed and spots
corresponding to
complex V a, complex III core 2, and complex V d (all ui fraction 4) could be
clearly identified, as
could the spot in fraction 5 corresponding to complex IV Va. Three of these
proteins (complex V a,
complex III core ?, and complex V d) were not identified in the human
placental mitochondria)
proteome of Rabilloud et al. (Elech~ophoresis, 19:1006-1014, 1998). This
example demonstrates that
the separation methods provided herein, coupled with monoclonal antibody
analysis, are powerful
tools in proteomics.
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
__ 21 __
Example 7: Processing, Assembly, and Analysis of Data
This example provides a representative system used for converting raw data
produced in the
described separation systems into data sets that can be used to compare the
protein profile of two (or
more) different samples.
By way of example only, such processing uicludes scanning of individual t,vo-
dimensional
gels, and assemblaig several scanned images into a three-dimensional
representation of the protein
profile. This can be accomplished, for instance, by sequentially stacking the
individual images in an
order that reflects the order of the corresponding sub-fractions in the non-
denaturing fractionation
(e.g., in the sucrose gradient). This is schematically illustrated in FIG. 6.
Example 8: Detection of Alterations in the Mitochondrial Proteome Caused by
Disease
With the provision herein of enhanced methods to separate proteins from
tissue, cell, or sub-
cellular (e.g., organelle) samples, methods are now enabled for using the
resulting proteomes to
identify, diagnose, prognose, and ti~ack diseases and other clinically
important conditions that alter
protein expression profiles. Such alterations in protein expression profiles
include changes in the
amounts of individual proteins, changes in the localization of protein
expression, changes in the
temporal regulation of protein regulation, and particularly changes in protein-
protein
interactions/associations (e. g., changes in the patterns of proteui complex
expression).
By way of example only, the mitochondrial proteomes described above can be
used to detect
protein expression and association changes associated with Alzheuner's.
Samples from known
Alzheimer sufferer and/or a known healthy control can be separated and
analyzed as described
herein, to provide standard protein fmgerprint(s). To determine if an
uidividual suffers from
Alzheimer's, a biological sample from that person is prepared and separated
under similar or
essentially identical conditions to a standard (e.g., a healthy and/or a known
diseased sample). The
resultant three-dimensional protein fingerprints are stained, for instance
with SyproRubyTM protein
gel stain as described herein, and compared to determine what proteins are
increased or decreased.
It is advantageous in some instances to use computer assisted scanning and
comparison
procedures to produce a difference map between the two protein fingerprints.
This difference map
can provide qualitative and/or quantitative inforniation regarding protein
levels in the controls) and
experimental samples. In certain embodunents, proteins are identified that
vary at least 20% in
protein level (or level in a particular fraction, or at a particular location
on a gel) between the t,vo
samples. Some proteins may vary considerably more than 20%, for instance by
more than 30%, more
than 40%, more than 50%, and so forth. In some instances, proteins that are
present in the healthy
control may be completely absent in the experimental or disease sample, and
vice versa.
This disclosure provides enhanced systems for protein separation and analysis,
which
systems can be augmented through the use of computers and automation.
Biological influences or
events can be correlated with alterations in a proteome or subproteome, thus
permitting disease
diagnosis, prognosis, pharmaceutical agent efficacy testing, and
pharmaceutical agent identification,
based on observations of such alterations. It will be apparent that the
precise details of the methods,
CA 02440323 2003-09-08
WO 02/082079 PCT/US02/08723
__ ?~ __
products, and devices described may be varied or modified without departing
from the spirit of the
invention. We claim all such modifications and variations that fall within the
scope and spirit of the
claims below.
