flavonoids were successfully assigned by their
elemental compositions. These results clearly
showed differences in flavonoid metabolism
among accessions and developmental stages
(Suzuki et al. 2008 ). At that time, peak picking
and MS analysis were mainly performed by
manual procedures. After this leading study, the
next metabolomics platforms incorporated auto-
mated procedures using newly developed soft-
ware and supported the comprehensive
metabolite annotation of tomato (Solanum lyco-
persicum) (Iijima et al. 2008 ).
MS/MS is a powerful technique for the struc-
tural elucidation of metabolites because it dis-
criminates fragment ions derived from related
compounds that belong to the same class of
metabolites but are partially modified by plant-
specific methyltransferases and glycosidases. MS/
MS spectra can provide an indication of putative
structures of metabolites via manual interpretation
or automated comparison with databases of frag-
mentation patterns. In conventional untargeted
analysis, MS/MS data are acquired for only a
limited number of peaks of interest, which are
selected by a data mining method, and additional
MS/MS analyses are required if other interesting
peaks are found by another data mining method.
Matsuda et al. ( 2009 ) developed a peak annotation
procedure on the basis of an MS/MS spectral tag
(MS2T) library for untargeted metabolic profiling
analysis. MS/MS spectra of most of the LC peaks
were acquired and stored in MS2T libraries. The
experiments for spectral acquisition were repeated
25 times, altering the mass ranges used to select
precursor ions. The libraries were created from
MS/MS spectra acquired using the automatic data
acquisition function of ultra-performance LC
quadrupole-time-of-flight-MS (UPLC-QTOF) in
experiments distinct from conventional metabolic
profiling analyses (Matsuda et al. 2009 ). UPLC
employing a sub-2-μm column dramatically
improves metabolite separation. The narrowed
peaks (3–6 s) are detected by high-sensitivity
scanning of QTOF-MS (Fig.16.1a, b). Using this
platform, untargeted MS2Ts were collected in
several developmental stages of Arabidopsis
(Matsuda et al. 2010 ), and the total of MS2Ts to
date have amounted to more than 1 million
derived from 10 plant species (A. thaliana,Oryza
sativa,Triticum aestivum,S. lycopersicum,Gly-
cine max,Charasp.,Pyrus communis,L. japo-
nicus,Allium cepaandBrassica napus). Together
with the above-mentioned analytical procedure,
an integrated searching system using a reference
database of MS2Ts has been established at RI-
KEN PSC (Akiyama et al. 2008 ; Lim et al. 2003 ;
Hank et al. 2003 ).
In untargeted analytical platforms with LC-
FTICR/MS and UPLC-QTOF/MS, thousands of
known and unknown peaks are detected and
characterized by elemental composition analysis
of HR/MS and MS/MS similarity searches of
MS2Ts (Iijima et al. 2008 ; Matsuda et al. 2009 ).
Given that these instruments are known as big
data generators (approximately 20 GB per
30 min of acquisition in the profile mode), and
the analytical samples for a project usually
number at least a hundred, a practical strategy of
research is needed for high-throughput metabolic
profiling.16.4 Widely Targeted AnalysisOn a tandem quadrupole mass spectrometer (TQ-
MS), targeted metabolites can be analysed using
a non-scanning MS technique of high sensitivity
and selectivity, termed selective reaction moni-
toring (SRM). In SRM, two mass analysers are
used as static massfilters to monitor a specified
fragment ion of a selected precursor ion
(Fig.16.1c). One of the essential parameters of
SRM is the specific pair of m/z values associated
with the precursor and fragment ions selected,
referred to as a “transition.” The instrument
continuously records the intensity of the ions that
match the selected transition, showing a chro-
matogram instead of MS spectra. SRM using
TQ-MS has been used for the analysis of phy-
tohormones in trace amounts (Kojima et al.
2009 ). A TQ-MS instrument can rapidly switch
the transitions monitored and can record the
intensity of each transition as a function of the
retention time, allowing the measurement of
multiple SRM transitions within the same
experiment. A recently introduced high-speed16 Metabolomics 175