18 Proteomics in Fish Processing 405et al. 1999); nevertheless an approach that involves
several gels, each of a different pH range, from the
same sample is advocated for representative inclu-
sion of alkaline proteins when studying entire pro-
teomes (Cordwell et al. 2000). Indeed, Cordwell et
al. were able to significantly improve the representa-
tion of alkaline proteins in their study on the rela-
tively highly alkaline Helicobacter pyloriproteome
by using both pH 6–11 and pH 9–12 IPGs (Bae et al.
2003).
A second drawback of 2DE has to do with the
extreme difference in expression levels of the cell’s
various proteins, which can be as much as 10,000-
fold. This leads to swamping of low abundance pro-
teins by high abundance ones on the two-dimensional
map, rendering analysis of low abundance proteins
difficult or impossible. For applications such as spe-
cies identification or the study of major biochemical
pathways, where the proteins of interest are present
in relatively high abundance, this does not present a
problem. However, when investigating, for example,
regulatory cascades, the proteins of interest are likely
to be present in very low abundance and may at times
be undetectable because of the dominance of high
abundance proteins. Simply increasing the amount of
sample is usually not an option, as it will give rise to
overloading artifacts in the gels (O’Farrell 1975). In
transcriptomic studies, where a similar disparity can
be seen in the abundance of RNA transcripts present,
this problem can be overcome by amplifying the low
abundance transcripts using the polymerase chain
reaction (PCR), but no such technique is available for
proteins. The remaining option, then, is fractionation
of the protein sample in order to weed out the high
abundance proteins, allowing a larger sample of the
remaining proteins to be analyzed. A large number of
fractionation protocols, both specific and general, are
available. Thus, Østergaard et al. used acetone pre-
cipitation to reduce the abundance of hordeins pres-
ent in barley(Hordeum vulgare)extracts (Østergaard
et al. 2002), whereas Locke et al. used preparative
isoelectrofocusing to fractionate Chinese snow pea
(Pisum sativum macrocarpon)lysates into fractions
covering three pH regions (Locke et al. 2002). The
fractionation method of choice will depend on the
specific requirements of the study and on the tis-
sue being studied. Discussion of some fractionation
methods can be found in Butt et al. (2001), Corthals
et al. (1997), Dreger (2003), Issaq et al. (2002), Lo-
pez et al. (2000), Millea and Krull (2003), Pieper et
al. (2003), and Rothemund et al. (2003).
IDENTIFICATION BYPEPTIDEMASS
FINGERPRINTINGIdentification of proteins on 2DE gels is most com-
monly achieved via mass spectrometry of trypsin
digests. Briefly, the spot of interest is excised from
the gel and digested with trypsin (or another pro-
tease), and the resulting peptide mixture is analyzed
by mass spectrometry. The most popular mass spec-
trometry method is matrix-assisted laser desorp-
tion/ionization–time-of-flight (MALDI-TOF) mass
spectrometry (Courchesne and Patterson 1999),
where peptides are suspended in a matrix of small,
organic, UV-absorbing molecules (such as 2,5-
dihydroxybenzoic acid), followed by ionization by a
laser at the excitation wavelength of the matrix mol-
ecules and acceleration of the ionized peptides in an
electrostatic field into a flight tube where the time of
flight of each peptide is measured, giving its expect-
ed mass.
The resulting spectrum of peptide masses (Fig.
18.5) is then used for protein identification by search-
ing against expected peptide masses calculated from
data in protein sequence databases, such as Swiss
Prot or the National Center for Biotechnology In-
formation (NCBI) nonredundant protein sequences
data base, using the appropriate software. Several
programs are available, many with a web-based
open-access interface. The ExPASy Tools website
(http://www.expasy.org/tools) contains links to most
of the available software for protein identification
and several other tools. Attaining a high identifica-
tion rate is problematic in fish and seafood pro-
teomics due to the relative paucity of available pro-
tein sequence data for these animals. As can be seen
in Table 18.1, this problem is surprisingly acute for
species of commercial importance. To circumvent
this problem, it is possible to take advantage of the
available nucleotide sequences, which in many cas-
es are more extensive than the protein sequences
available, to obtain a tentative identity. How useful
this method is will depend on the length and quality
of the available nucleotide sequences. It is important
to realize, however, that an identity obtained in this
manner is less reliable than that obtained through
protein sequences and should be regarded only as
tentative in the absence of corroborating evidence
(such as two-dimensional immunoblots, correlated
activity measurements, or transcript abundance). In
their work on the rainbow trout (Oncorhynchus
mykiss)liver proteome, Martin et al. (2003b) and