Overview of Quantitative LC-MS Techniques for Proteomics and Activitomics - LC-MS/MS in Proteomics: Methods and Applications - page 21

mixture for downstream LC-MS/MS-based identification. Sub-
sequent work used the ICAT strategy to successfully identify and
determine the ratios of abundance of 491 proteins between the
microsomal fractions of naïve and in vitro-differentiated human
myeloid leukaemia cells ( 7 ) , whilst Griffin et al. combined ICAT
labelling with LC-MALDI-MS/MS for targeted protein identifi-
cation for better sample utilisation and reduced MS instrument
time ( 8 ) .
Several technical problems were reported with the original
deuterated ICAT reagents. Prominent of these was the differ-
ential elution of ICAT-labelled peptide pairs in reversed-phase
HPLC (RP-LC) leading to errors in the quantification ( 9 , 10 ) .
A further problem was the relatively low efficiencies of labelled
peptide collision-induced dissociation (CID), which was specu-
lated to be due to the relatively large size of the ICAT moiety
( 11 ) . Poor recovery of tagged peptides from avidin may also con-
tribute to reduced proteomic coverage. To improve these issues,
a new generation of ICAT reagents were developed where the
deuterium atoms were substituted for nine 12 C (light) or 13 C
(heavy) atoms and an acid-cleavable linker added. These new
generation cleavable ICAT (cICAT) reagents gave more precise
co-migration of the light- and heavy-tagged peptides in RP-LC,
whilst the cleavage strategy eliminated undesired residual frag-
mentation of biotinylated peptides and improved recovery ( 12 ) .
However, an additional clean-up step was required for removal
of the cleaved biotin moiety prior to LC-MS analysis. A subse-
quent large-scale investigation of the cICAT strategy, which mea-
sured protein changes in Escherichia coli treated with an inhibitor
of fatty acid biosynthesis, quantified more than 24,000 pep-
tides in four independent runs using an ion-trap mass spectrom-
eter ( 13 ) . Good reproducibility in quantification was reported
(median CV of ratios was 18.6%), and on average >450 unique
proteins were identified per experiment. However, the method
was biased towards the detection of acidic proteins and under-
represented small and hydrophobic proteins.
In another multiple-run study examining proteomic changes
in lymphoma cells following different ligand and drug treatments,
Vaughn et al. reported that the majority of ICAT-labelled pro-
teins were identified by single peptides with only 24-41% of these
containing cysteine residues ( 14 ) . Eighty-five percent of cysteine-
containing peptides yielded quantification data, but this repre-
sented only 28% of all identified proteins. This identification of
proteins from single peptides highlights a significant drawback of
cysteine-labelling strategies, where more false-positive identifica-
tions are likely to occur and quantification accuracy will be lower
than when data from multiple peptides derived from the same
protein is combined.