Quantification of Proteins by iTRAQ - LC-MS/MS in Proteomics: Methods and Applications - page 205

reagent. The four-channel reagent has reporter ion masses at
114-117, with a small balance group. The eight-channel has
reporters at 113-119 and 121 (120 is left blank due to the pres-
ence of the phenylalanine immonium ion) and a larger balance
group to accommodate the extra isotopes. Practically, we have
noticed little difference between the two. Tandem mass tags are
available through Thermo Scientific.
iTRAQ has been successful in many experimental settings.
The primary advantages over an alternative labelling strategy such
as SILAC ( see Chapter 11 in this volume for a SILAC protocol)
are that it is applicable to primary samples, e.g. human bioflu-
ids ( 3 - 5 ) and disease tissues ( 6 - 8 ) or primary tissues from animal
models ( 9 ) . The multiplex nature of the reagent is also advan-
tageous over other chemical labelling methodologies such as
labelling with 18 O during digestion ( 10 ) or with 13 C-acrylamide
( 11 ) , making studies on time courses ( 12 ) possible, or allowing
an increased number of samples to be analysed against the same
control simultaneously, saving on sample preparation and analy-
sis time ( 13 ) . Since iTRAQ also labels all peptides in a sample,
it is ideally suited to the quantification of posttranslational mod-
ifications ( 14 - 16 ) and can also be used to label whole proteins
prior to subsequent protein fractionation steps, if required ( 17 ) .
An alternative to iTRAQ is quantification by label-free approaches
( see Chapters 9 , 10 ,and 13 for protocols on MS-based label-free
quantification of proteins and peptides).
As with many proteomics methods for relative quantifica-
tion, success in this technique is heavily dependent on sample
preparation. The peptide-reactive group on the iTRAQ tags is a
N -hydroxysuccinimide (NHS) moiety, which will label peptides
via free amine groups, namely those at the N-terminus and on
lysine side chains. As a result, it is important to ensure that all
buffers used to prepare protein prior to labelling should not con-
tain free amines, so common reagents such as Tris and ammo-
nium bicarbonate should be avoided. This chemistry will also
react with cysteine, so these should be blocked prior to labelling,
and at lower pH can also react with tyrosine residues. Since nei-
ther of these reactions goes to completion (i.e. 100% labelling)
they should be avoided or at least minimised at all costs with care-
ful design of sample buffer systems. Also, since no method is per-
fect, it is important to know the technical errors introduced via
the sample handling procedure. To test this, it is recommended
that each experiment uses all channels with at least two chan-
nels taken up by technical or biological replicates. If this is not
possible (i.e. a time course experiment), it is recommended that
a separate experiment is run containing such replicates to assess
inter-experimental variation.