Proteomics in brain research: potentials and limitations.

Abstract

The advent of proteomics techniques has been enthusiastically accepted in most areas of biology and medicine. In neuroscience, a host of applications was proposed ranging from neurotoxicology, neurometabolism, determination of the proteome of the individual brain areas in health and disease, to name a few. Only recently, the limitations of the method have been shown, hampering the rapid spreading of the technology, which in principle consists of two-dimensional gel electrophoresis with in-gel protein digestion of protein spots and identification by mass-spectrometrical approaches or microsequencing. The identification, including quantification using specific software, of brain protein classes, like enzymes, cytoskeleton proteins, heat shock proteins/chaperones, proteins of the transcription and translation machinery, synaptosomal proteins, antioxidant proteins, is a clear domain of proteomics. Furthermore, the concomitant detection of several hundred proteins on a gel allows the demonstration of an expressional pattern, rather generated by a reliable, protein-chemical method than by immunoreactivity, proposed by protein-arrays. An additional advantage is that hitherto unknown proteins, so far only proposed from their nucleic acid structure, designated as hypothetical proteins, can be identified as brain proteins. As to shortcomings and disadvantages of the method we would point to the major problem, the failure to separate hydrophobic proteins. There is so far no way to analyse the vast majority of these proteins in gels. Several other analytical problems need to be overcome, but once the latter problem can be solved, there is nothing to stop the method for a large scale analysis of membrane proteins in neuroscience.

Full text

Page 1

UNCORRECTED PROOF
doi:10.1016/S0301-0082(03)00036-4
Progress in Neurobiology 611 (2003) 1–19
Review
3
Proteomics in brain research: potentials and limitations
4
Gert Lubeca,∗, Kurt Krapfenbauera,b, Michael Fountoulakisb
aDepartment of Pediatrics, University of Vienna, Währinger Gürtel 18, A 1090 Vienna, Austria
bF. Hoffmann-La Roche, Genomics and Proteomics Technologies, CH 4070 Basel, Switzerland
5
6
7
Received 21 November 2002; accepted 24 February 2003
8
9
Abstract
10
The advent of proteomics techniques has been enthusiastically accepted in most areas of biology and medicine. In neuroscience, a
host of applications was proposed ranging from neurotoxicology, neurometabolism, determination of the proteome of the individual brain
areas in health and disease, to name a few. Only recently, the limitations of the method have been shown, hampering the rapid spreading
of the technology, which in principle consists of two-dimensional gel electrophoresis with in-gel protein digestion of protein spots and
identificationbymass-spectrometricalapproachesormicrosequencing.Theidentification,includingquantificationusingspecificsoftware,
of brain protein classes, like enzymes, cytoskeleton proteins, heat shock proteins/chaperones, proteins of the transcription and translation
machinery,synaptosomalproteins,antioxidantproteins,isacleardomainofproteomics.Furthermore,theconcomitantdetectionofseveral
hundred proteins on a gel allows the demonstration of an expressional pattern, rather generated by a reliable, protein-chemical method than
by immunoreactivity, proposed by protein-arrays. An additional advantage is that hitherto unknown proteins, so far only proposed from
their nucleic acid structure, designated as hypothetical proteins, can be identified as brain proteins. As to shortcomings and disadvantages
of the method we would point to the major problem, the failure to separate hydrophobic proteins. There is so far no way to analyse the vast
majority of these proteins in gels. Several other analytical problems need to be overcome, but once the latter problem can be solved, there
is nothing to stop the method for a large scale analysis of membrane proteins in neuroscience.
© 2003 Published by Elsevier Science Ltd.
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Contents
1. Introduction ................................................................................
2. Typical analytical protocols for brain proteomics ..............................................
2.1. Sample preparation .....................................................................
2.2. Two-dimensional electrophoresis, first and second dimension...............................
2.3. Spot excision and in-gel digestion........................................................
2.4. Mass-spectrometrical procedures.........................................................
2.4.1. Matrix-assisted laser desorption ionization time-of-flight mass spectrometry
(MALDI-TOF-MS) ...............................................................
2.4.2. Tandem mass spectrometry........................................................
2.5. Quantification of protein spots...........................................................
3. Generation of a human brain protein map.....................................................
4. Classes of brain proteins identified and verified in brain so far (incomplete and gradually increasing)
4.1. Cytoskeleton proteins ...................................................................
4.2. Intermediary metabolism enzymes (IME) .................................................
4.3. Signalling proteins......................................................................
4.4. Antioxidant proteins ....................................................................
4.5. Ubiquitination and proteasome-related proteins............................................
4.6. RNA binding proteins and proteins serving in transcription, splicing and elongation..........
4.7. Heat shock proteins and chaperones......................................................
4.8. Neuronal proteins.......................................................................
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
∗Corresponding author. Tel.: +43-1-40400-3488;
fax: +43-1-40400-3194.
E-mail address: gert.lubec@akh-wien.ac.at (G. Lubec).
1
2
0301-0082/03/$ – see front matter © 2003 Published by Elsevier Science Ltd.

Page 2

UNCORRECTED PROOF
2
G. Lubec et al./Progress in Neurobiology 611 (2003) 1–19
4.9. Glial-associated proteins.................................................................
4.10. Hypothetical proteins..............................................................
5. Potential applications........................................................................
6. Limitations.................................................................................
6.1. Low abundance proteins.................................................................
6.2. Hydrophobicity of proteins ..............................................................
6.3. High and low molecular weight proteins..................................................
6.4. Post-translational modifications ..........................................................
6.5. Other limitations........................................................................
6.6. Verification of results ...................................................................
7. Conclusion and perspectives .................................................................
Acknowledgements ............................................................................
References ....................................................................................
00
00
00
00
00
00
00
00
00
00
00
00
00
1. Introduction
26
Proteomics is the science and methodology of the study
of the proteome, i.e. all proteins expressed in a cell or tissue,
rather than proteins one by one. The advent of proteomics
was a major step forward, comparable to the introduction of
molecular biological methods in the past. It fits within the
concept that the determination of protein rather than RNA
levels has major advantages as it is the proteins that carry
out functions. There is a long and unpredictable way from
RNA to protein, a fact known to all scientists working with
gene hunting methods, like differential display and subtrac-
tive hybridization. Huge subtractive libraries have been gen-
erated, but when the differentially expressed mRNAs were
studied at the protein level, only a small percentage of aber-
rant transcriptomes could be verified. This does not mean, of
course, that proteomics is fully replacing these techniques,
but proteomics is a very valuable tool for protein hunting,
i.e. comparing cell and tissue proteomes under physiologi-
cal and pathophysiological conditions. The trend, however,
seemstobethatoncetheproteinisdetermined,thetranscrip-
tional level is examined to find the underlying mechanism
for the increase or decrease of a certain gene product. More-
over, information about the presence of isoforms and post-
translational modifications can be obtained by proteomics.
In the following, we give only an outline, an example of
a proteomic method that has been used in many studies, and
we are aware that much valuable work is not mentioned and
many important studies are not cited. The review is written
to enable the neuroscientist to catch the spirit of proteomics.
After introducing the methodology, we present typical hu-
man brain protein maps, identify the protein classes, recom-
mend applications, discuss the shortcomings and limitations
of proteomics, and, finally, draw a conclusion.
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
2. Typical analytical protocols for brain proteomics
59
Proteomics consists mainly of two steps: (1) separation
of proteins usually by two-dimensional gel electrophoresis
60
61
and (2) protein analysis and identification, mainly by mass
spectrometry. Of course, other protein identification meth-
ods, like amino acid composition analysis, N-terminal se-
quencing, or immunochemistry, as well as column chro-
matography can be used (Fountoulakis, 2001), or other bio-
chemical techniques can be applied for protein enrichment
(Fountoulakis and Takács, 2002).
A critical step is the extraction of proteins from cells or
tissues. At this step, one must decide which brain areas,
cellular fractions or subfractions, etc. will be analysed. The
importance of separation science in proteomics research is
addressed in detail in a recent publication (Isaaq, 2001).
We here present a brain proteomics ready-to-use standard
protocol.
62
63
64
65
66
67
68
69
70
71
72
73
74
75
2.1. Sample preparation
76
Braintissue(0.1g)issuspendedin0.5mlofsamplebuffer
consisting of 20mM Tris, 7M urea (Merck, Darmstadt, Ger-
many), 2M thiourea (Sigma, St. Louis, MO, USA), 4%
CHAPS(Sigma),10mM1,4-dithioerythritol(Merck),1mM
EDTA (Merck), 1mM PMSF (Sigma) and 1?l/mg each of
pepstatin A, chymostatin, leupeptin and antipain. The sus-
pension is sonicated for approximately 30s and centrifuged
further at 150,000×g for 45min to obtain a cytosolic frac-
tion. The protein concentration of the supernatant is deter-
mined using the Coomassie blue method.
77
78
79
80
81
82
83
84
85
86
2.2. Two-dimensional electrophoresis, first and second
dimension
87
88
In the first dimension, separation of proteins on the basis
of differences in their net charge, isoelectric focusing (IEF),
takes place. For this purpose an immobilized pH gradient
(IPG) is used; this has been formed prior to the IEF run and
is generated by acrylamide derivatives, i.e. immobilines on a
plastic sheet. The IPG strips are commercially available and
can be stored frozen for longer periods. 1mg of protein is
applied on immobilized pH 3–10 non-linear gradient strips
(Amersham Biotechnologies, Uppsala, Sweden) in sample
89
90
91
92
93
94
95
96
97

Page 3

UNCORRECTED PROOF
G. Lubec et al./Progress in Neurobiology 611 (2003) 1–19
3
cups at their basic and acidic ends (Langen et al., 1997;
Fountoulakis, 2000). Focusing (Multiphor II Electrophoresis
Unit, Amersham Biotechnologies) starts at 200V, and the
voltage is gradually increased to 5000 at 3V/min and kept
constant for 24h (approximately 180,000V totally, due to
the specific matrix effects of brain).
In the second dimension, proteins are separated ac-
cording to their size in standard polyacrylamide gels.
During this step, proteins are negatively charged by the
addition of sodium dodecylsulphate (SDS). The second-
dimensional separation is usually performed on 9–
16% SDS linear-gradient polyacrylamide gels. The gels
(180mm × 200mm × 1.5mm) are run at 40mA per gel.
After protein fixation for 12h in 40% methanol, contain-
ing 5% phosphoric acid, gels are stained with colloidal
Coomassie blue (Novex, San Diego, CA, USA) for 48h.
Molecular masses are determined by running standard pro-
tein markers (Gibco, Basel, Switzerland), covering the range
10–200kDa. pI values are used as given by the supplier
of the IPG strips. Excess dye is washed out from the gels
with water, and gels are scanned in an Agfa DUOSCAN
densitometer (resolution 200). Electronic images of the gels
are recorded using Photoshop (Adobe) software.
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
2.3. Spot excision and in-gel digestion
121
Proteinspotsareexcisedwithaspotpickerandplacedinto
96-well microtiter plates. Each spot is destained with 100?l
of 30% acetonitrile in 50mM ammonium bicarbonate, dried
in a speedvac evaporator and processed in the presence of a
low salt buffer (Fountoulakis and Langen, 1997). Each dried
gel piece is then rehydrated with 4?l of 1mM Tris–HCl, pH
9.0containing50ngtrypsin(RocheDiagnostics,Mannheim,
Germany). After 16h at room temperature, 7?l of water
is added to each gel piece and the samples are shaken for
10min. 4?l of 50% acetonitrile containing 0.3% trifluo-
roacetic acid and the standard peptides des-Arg-bradykinin
(Sigma) and adrenocorticotropic hormone fragment 18–39
(Sigma) in water are added to each gel piece and shaken for
10min. The application of samples is performed with the
use of a sample processor. 1.5?l of the peptide mixture are
simultaneously applied with 1?l of matrix consisting of a
saturated solution of alpha-cyano-4-hydroxycinnamic acid
(Sigma) in 50% acetonitrile containing 0.1% trifluoroacetic
acid.
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
2.4. Mass-spectrometrical procedures
141
2.4.1. Matrix-assisted laser desorption ionization
time-of-flight mass spectrometry (MALDI-TOF-MS)
Theproteinsareidentifiedbymatrix-assistedlaserdesorp-
tion ionization time-of-flight mass spectrometry (MALDI-
TOF-MS) on the basis of peptide mass matching (Henzel
et al., 1993; Lahm and Langen, 2000), following in-gel di-
gestion with trypsin. About 1000 spots are excised from
each gel. An accelerating voltage of 20kV is used. Peptide
142
143
144
145
146
147
148
149
matching and protein searches are performed automatically.
The peptide masses are compared to the theoretical peptide
masses of all available proteins from all species. Monoiso-
topic masses are used and a mass tolerance of 0.0025% is
allowed (Berndt et al., 1999). Gevaert and Candekerckhove
(2000) have recently reviewed protein identification meth-
ods in proteomics highlighting the use of mass spectromet-
rical and other methods and describe the two principles of
peptide mass fingerprinting and post-source decay analysis.
150
151
152
153
154
155
156
157
158
2.4.2. Tandem mass spectrometry
Proteins that cannot be identified by MALDI-TOF-MS,
for example the low-molecular mass proteins that yield only
a few peptides (Fountoulakis et al., 1998a), are analysed
by tandem MS, using the Qq-TOF approach. Unseparated
peptide mixtures from the tryptic, in-gel digest of unknown
proteins are subjected to nano-electrospray ionization tan-
dem mass spectrometry. The tryptic peptides from the
wells of the microtiter plates are desalted and concentrated
on a pulled capillary containing approximately 100nl of
POROS R2 reverse phase material (Applied Biosystems,
Framingham, MA) and eluted with 1?l of 60% methanol
in 5% formic acid directly into the nano-electrospray nee-
dle. Electrospray ionization mass spectra are acquired on a
QSTAR Pulsar i quadrupole TOF tandem mass spectrom-
eter (AB/MDS-Sciex, Toronto, Canada) equipped with a
nano-electrospray ion source (MDS Protana, Odense, Den-
mark). Fragmentation of individual peptides by tandem
mass spectrometry yields a stretch of sequence together
with its location in the peptide (peptide sequence tag). With
these sequence tags, appropriate databases are searched
using the instrument’s software (Mann and Wilm, 1994).
If no protein is identified, de novo sequencing can be per-
formed, considering also isobaric residues, like Leu/Ile, as
well as possible modifications. With such deduced sequen-
ces, a search in protein databases of homologous species can
be performed with standard tools (FASTA, BLAST). Mod-
ern mass spectrometers can distinguish between residues of
similar mass, like Gln/Lys or Phe/methionine sulphoxide,
but they cannot distinguish between Leu/Ile, which have ex-
actly the same mass. In the latter case, a search repeat in the
databases with either residue is necessary. Each identifica-
tion is characterized by a score number, which is calculated
on the basis of the peptide matches. The researcher can ac-
cept or reject the identification proposed, considering addi-
tional data from the 2D gel analysis, such as molecular mass
and pI values. This method is labour intensive and cannot
be performed in a high-throughput mode today.
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
2.5. Quantification of protein spots
197
Proteinspotsareoutlined(firstautomaticallyandthenma-
nually) and quantified using the Image Master 2D Elite soft-
ware (Amersham Biotechnologies). Modern imaging and
quantification software provide automation and accuracy in
the analysis of a large number of gels. Background subtrac-
198
199
200
201
202

Page 4

UNCORRECTED PROOF
4
G. Lubec et al./Progress in Neurobiology 611 (2003) 1–19
tion is enabled, and gel averaging allows evaluation of dif-
ferences between runs. The protein level is expressed as the
percentage of the spot volume representing a certain protein
in comparison with the volume of all proteins in the gel.
203
204
205
206
3. Generation of a human brain protein map
207
To illustrate results of proteomics, we provide a practical
example of an application by showing typical brain protein
208
209
Fig. 1. Brain 2D protein map of the mitochondrial fraction from human frontal. The proteins were separated on a pH 3–10 non-linear IPG strip, followed
by a 9–16% SDS-gel. The gel was stained with colloidal Coomassie blue. Note the presence of mitochondrial marker proteins: aconitate hydratase,
mitochondrial precursor [Q99798]; ATP synthase alpha chain, mitochondrial precursor [P25705]; dihydrolipoamide succinyltransferase component (E2)
[P36957]; glycerol-3-phosphate dehydrogenase mitochondrial precursor [P35571].
maps from normal human frontal cortex after subfractiona-
tion into mitochondrial, microsomal and cytosolic fractions
(Figs. 1–3, respectively). In Table 1, we provide data for pro-
tein identification assignment, including peptide matches,
probability of assignment of a random identity and theoret-
ical pI and molecular weight values. Analysis of spots from
the three brain fractions was performed using the proteomics
methodology described (2D gels and MALDI-TOF-MS) and
resulted in the identification of 308 different gene products.
A part of the proteins presented here have been published in
210
211
212
213
214
215
216
217
218

Page 5

UNCORRECTED PROOF
G. Lubec et al./Progress in Neurobiology 611 (2003) 1–19
5
Fig. 2. Brain 2D protein map of the microsomal fraction from human frontal cortex. The gel was prepared as stated in Fig. 1.
a previous report (Langen et al., 1999), and this communi-
cation therefore represents the updating of identified brain
proteins.
219
220
221
4. Classes of brain proteins identified and verified in
brain so far (incomplete and gradually increasing)
222
223
4.1. Cytoskeleton proteins
224
Cytoskeleton proteins (CP) are not only serving as scaf-
folding structures but also play a major role in transport
and signalling (dos Remedios and Thomas, 2001). Reports
225
226
227
about the separation and identification of CP have been pub-
lished for several organ systems (Srinivasan et al., 2001;
Kovarova et al., 2000; Steiner et al., 2000), including the
brain (Weitzdoerfer et al., 2001, 2002; Lubec et al., 2001;
Gulesserian et al., 2002), and for several cellular compart-
ments. Table 1 includes 27 cytoskeletal brain proteins. They
can be seen in Figs. 1–3 mainly as high abundance proteins.
228
229
230
231
232
233
234
4.2. Intermediary metabolism enzymes (IME)
235
This group of enzyme proteins (IME) is the largest protein
classofTable1andisstronglyrepresentedinallmapsgener-
ated so far, mainly as high abundance proteins. The proteins
236
237
238