encode the target enzyme. Identification of the
metabolites in the plant materials and reaction
products of the enzyme assays are essential for
metabolic studies including the functional char-
acterization of biosynthetic genes. The tech-
niques that have been used in thefield of plant
secondary metabolism since its pioneering stud-
ies are classical low-throughput techniques, such
as LC equipped with a UV or photodiode array
detector, infusion EI-MS and nuclear magnetic
resonance spectroscopy (Akashi et al.1998a,b,
1999a, b; Sawada et al. 2002 ). Until very
recently, studies in thisfield have not benefited
from the advances in technology that enabled
metabolomics.
Validation of gene functions using a model
plant provides reliable clues to the complex
metabolism of leguminous plants. In the late
1990s, studies aimed at the comprehensive clar-
ification of structural genes involved in the bio-
synthesis of isoflavonoid phytoalexins (Shimada
et al. 2003 , 2005 , 2006 , 2007 ), and triterpenes
(Sawai et al.2006a, b)ofL. japonicus were
begun with the aim of establishing a basis for
functional genomics and molecular genetic
approaches for the model legume. Candidate
genes and cDNAs were cloned with the powerful
aid of genomics data, such as expression
sequence tags and genome sequence information,
both of which have been accumulated in the
genome project of L. japonicus promoted by
Kazusa DNA Research Institute (KDRI). How-
ever, the analysis of metabolites and elucidation
of gene functions by biochemical assays depen-
ded on“classical”procedures, i.e., the isolation
of metabolites and their structural identification
by low-throughput techniques capable of han-
dling only limited numbers of targeted
metabolites.
16.3 Untargeted Analysis
for Metabolomics
Elucidation of the comprehensive metabolic
profile, i.e., untargeted metabolic profiling,
requires high-throughput and high-sensitivity/
high-selectivity analysis of a broad range of
phytochemicals. To meet these requirements,
untargeted metabolomics aims to make the best
use of MS-based technology combined with
chromatographic separation. In transcriptomics,
sequence-based gene annotation as well as
expression data is essential for biologicalfind-
ings. The analogy with transcriptomics suggests
the necessity of metabolite annotation, which can
also be implemented by MS. GC-MS has the
advantage that its standard ionization method, EI,
reproducibly generates fragment ions, and this
technologyfirst enabled the high-throughput and
unbiased analysis of thousands of metabolites
from plants and other organisms (Dixon and
Sumner 2003 ). Desbrosses et al. ( 2005 ) charac-
terized the metabolic profiles ofL. japonicusby
GC-MS, measuring relative levels of primary
metabolites in nodules, roots, leaves andflowers
of symbiotic plants. Discrete metabolic charac-
teristics were revealed by principal component
analysis and hierarchical cluster analysis. Prior to
measurement, they created mass spectral tag
libraries that contained both retention time of GC
and fragment ions representing metabolites
(Desbrosses et al. 2005 ).
The recent development of LC-MS and suffi-
ciently accurate high-resolution MS (HR/MS),
which allows us to define the elemental compo-
sitions of detected ions, has extended the poten-
tial of plant metabolomics. In this technique,
high-resolution mass data are provided as
metabolite annotation. HR/MS of LC-Fourier
transfer ion cyclotron resonance-MS (LC-
FTICR-MS) is an example of the most advanced
metabolomic techniques for untargeted analysis
(Marshall and Hendrickson 2008 ; Glauser et al.
2012 ).
KDRI is a pioneer in metabolomics with LC-
FTICR-MS, which was applied to characterize
the metabolic diversity between the two gener-
ally used accessions ofL. japonicus, B-129 Gifu
and MG-20 Miyakojima. These accessions differ
in stem colour and are accordingly thought to
differ in theirflavonoid metabolism (Suzuki et al.
2008 ). Natural metabolic variation among wild-
type accessions has important implications for
the mechanisms of environmental adaptation by
plants. A total of 61 known and unknown174 Y. Sawada and T. Aoki