Overview of Quantitative LC-MS Techniques for Proteomics and Activitomics - LC-MS/MS in Proteomics: Methods and Applications - page 24

proteolytic digestion. Here, one or two 16 O atoms are replaced by
one or two 18 O atoms through enzyme-catalysed exchange in the
presence of H 2 18 O. Since any variability in labelling relies solely
on the digestion step, this method is expected to give smaller
technical variations than do the two-step chemical labelling meth-
ods, where both labelling and digestion are a source of introduced
variation. The method was first suggested as a protein quantifica-
tion tool when 18 O-labelled internal standards were generated
for absolute quantification by MALDI-MS ( 43 ) . This was fol-
lowed by the reporting of conditions for protein labelling ( 44 )
and then the first proteomic application of the method, where
trypsin was used to incorporate two 18 O atoms into the C ter-
mini of all tryptic peptides and was applied to compare proteins
from two serotypes of adenovirus ( 3 ) . Subsequent work by the
same group showed that Glu-C could also be used for labelling,
that 18 O-labelled and unlabelled ( 16 O) peptide pairs co-eluted
in RP-LC and measurements of isotope ratios by LC-MS were
accurate and precise ( 45 ) . Applications include combining with
2D-LC-FT-ICR and an accurate mass and time (AMT) tag strat-
egy to identify and quantify 429 distinct plasma proteins from
an individual prior to and after lipopolysaccharide administration
( 46 ) ; combining with 16 O/ 18 O-methanol esterification, immo-
bilised metal-ion affinity chromatography and RP-LC followed
by neutral loss-dependent MS 3 for phosphopeptide identification
in the study of lysophosphatidic acid-induced chemotaxis ( 47 ) ;
the differential analysis of NF-

κ B transcription factor complexes

following TNF- stimulation ( 48 ) ; the labelling of a 'univer-

sal' reference sample of pooled plasma for spiking into indi-
vidual unlabelled samples for quantitative analysis across clinical
samples ( 49 ) .
Several drawbacks of the proteolytic labelling strategy are
apparent. Only two samples can be compared simultaneously,
C-terminal peptides of proteins cannot be quantified and variable
incorporation of 18 O into peptides can occur ( 50 ) . There is also
a lack of computational tools for accurate quantification of pep-
tide differences and this is exacerbated by the overlap of isotopic
envelopes for 16 O- and 18 O 1 -and 18 O 2 -labelled peptides. How-
ever, some methods have been reported for correcting 16 O/ 18 O
ratios from overlapping isotopic multiplets ( 51 , 52 ) . Rao et al.
( 52 ) have also shown the potential of Lys-N labelling, where con-
ditions were established such that only a single 18 Oatomwas
incorporated into the C terminus of each peptide, compared to
incorporation of one or two 18 O atoms when using trypsin, Lys-C
or Glu-C. Finally, a significant degree of chemical back exchange
of the carboxyl 18 O can occur in H 2 16 O solvent, particularly at
extreme pH, although handling recommendations to limit this
have been reported ( 44 ) .

α