Sample Preparation in Global Metabolomics of Biological Fluids and Tissues - Proteomic and Metabolomic Approaches to Biomarker Discovery - page 62

chloroform/methanol/water (also known as the
Bligh and Dyer method) extraction solvents,
currently predominate in the

for an evaporation/reconstitution step in such
approaches. Masson et al. recommended two-
step extraction, an aqueous extraction step with
methanol/water followed by a dichlorome-
thane/methanol (3/1, v/v) extraction step. 90,91
The resulting aqueous and organic extracts are
evaporated, reconstituted in methanol/water
(1/1), and injected individually inLC-MS. Overall,
this approach performed better than biphasic
mixture single-step extraction and was highly
reproducible with overall mean RSDs of 23.0%
for aqueous extracts and 33.1% for organic
extracts. A lyophilization/reconstitution step is
also useful for GC-MS analysis, whereby dried
extracts can be redissolved in an appropriate
organic solvent for derivatization. 58 In addition
to solvent extraction, Parab et al. also investigated
nine types of SPE cartridges for liver tissue metab-
olomics including both ion-exchange and
reversed-phase materials and found that the best
performance was for hydrophilic e lipophilic
balanced (HLB) polymeric sorbents such as
divinylbenzene/n-vinylpyrrolidone copolymer. 92
Considering the limited availability of clinical
tissues for biomarker discovery, a recent trend in
tissue metabolomics is to develop single-
extraction protocols directly compatible with
multiple analytical platforms. For

field of tissue metab-
olomics. 5,15,51,58,59,84,86 e 88 Keeping the chloroform
portion at less than or equal to 20% avoids phase
separation, and dissolved chloroform can serve
as a carrier to bring the phospholipids into the
methanolic phase in uniphasic mixtures. 88 For
instance, Lin et al. compared the performance of
perchloric acid, acetonitrile/water, methanol/
water, and methanol/chloroform/water for the
extraction of

fish muscle and liver tissues using
NMR and concluded that methanol/water/chlo-
roform extraction is the preferred method for
balanced recovery of both hydrophilic and hydro-
phobic metabolites and excellent reproduc-
ibility. 84 Sun et al . proposed a methanol/
chloroform/water extraction protocol for the anal-
ysis of mouse heart tissue in order to extract both
lipophilic and hydrophilic species, and eliminate
ionization suppression effects observed with
perchloric acid extraction. 17 Williams et al.
successfully used a similar extraction approach
for brain samples. 89 Beltran et al. also found that
the best performancewas obtained by usingmeth-
anol/water (1/1) and uniphasic methanol/
chloroform/water (7/2/1) mixtures, which out-
performed acetonitrile/water and acetonitrile/
chloroform/water mixtures for the extraction of
liver samples. 55 In addition, the extraction
enhanced the detection of aqueous compounds
in comparison to HR-MAS of the same liver tissue
but also slightlydecreasedmethod reproducibility
(13% RSD versus 7% RSD for extraction versus
HR-MAS). Using mixture design optimization,
chloroform/methanol/water (15/59/26) was the
optimal solvent composition for metabolite
extraction of hepatobiliary

instance,

a methanol/chloroform/water work ow was

rendered compatible for both NMR and LC-MS
analysis by the incorporation of an evaporation
step followed by reconstitution in deuterated
solvent (acetonitrile/water, 2/8, v/v). 55 The
authors clearly demonstrated that the use of
deuterated solvents had no impact on LC-MS
data acquired with no evidence of metabolite
deuteration, presumably due to quick back-
exchange reactions. Chloroform/methanol/
water extraction was also fully compatible with
UHPLC-MS and CE-MS analysis after reconsti-
tution in 50% and 20% methanol, respectively. 88
Dunn et al. proposed highly reproducible meth-
anol/water/chloroform extraction for placental
tissue with the methanol portion analyzed by
GC-MS while the chloroform portion is analyzed

uke Fasciola hepatica
tissue. 88 Geier et al. recently showed that the use
of a biphasic methanol/chloroform mixture for
C. elegans tissue resulted in signi cant loss of

nonpolar metabolites (100 e 300 metabolite
features), which preferentially partitioned into
a chloroform layer and were not injected if only
supernatant is analyzed, 85

supporting the need