ARTICLE

pubs.acs.org/ac

Expanding Coverage of the Metabolome for Global Metabolite

Profiling
Oscar Yanes,*,†,‡ Ralf Tautenhahn,† Gary J. Patti,† and Gary Siuzdak*,†
†
Departments of Chemistry and Molecular Biology, Center for Metabolomics, The Scripps Research Institute, La Jolla, California 92037,
United States
‡
Metabolomics Platform of the Spanish Biomedical Research Center in Diabetes and Associated Metabolic Disorders (CIBERDEM),
University Rovira i Virgili, Tarragona, 43007, Spain

ABSTRACT: Mass spectrometry-based metabolomics is the compre-

See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

hensive study of naturally occurring small molecules collectively

known as the metabolome. Given the vast structural diversity and
Downloaded via UNIV OF MANOUBA on December 14, 2018 at 10:06:43 (UTC).

chemical properties of endogenous metabolites, biological extraction

and chromatography methods bias the number, property, and con-
centration of metabolites detected by mass spectrometry and creates a
challenge for global untargeted studies. In this work, we used
Escherichia coli bacterial cells to explore the influence of solvent
polarity, temperature, and pH in extracting polar and nonpolar
metabolites simultaneously. In addition, we explored chromatographic
conditions involving different stationary and mobile phases that
optimize the separation and ionization of endogenous metabolite
extracts as well as a mixture of synthetic standards. Our results reveal
that hot polar solvents are the most efficient in extracting both
hydrophilic and hydrophobic metabolites simultaneously. In addition, ammonium fluoride in the mobile phase substantially
improved ionization efficiency in negative electrospray ionization mode by an average increase in signal intensity of 5.7 and over a
2-fold increase in the total number of features detected. The improvement in sensitivity with ammonium fluoride resulted in 3.5
times as many metabolite hits in databases compared to ammonium acetate or formic acid enriched mobile phases and allowed for
the identification of unique metabolites involved in fundamental cellular pathways.

Metabolomics is a rapidly growing field focused on the metabolomics study and thereby reflect overall coverage of the
profiling and quantification of small, naturally occurring com- metabolome. The intensity of each feature is also important in
pounds that collectively constitute the so-called metabolome. that a certain threshold is needed for accurate relative quantifica-
Electrospray ionization mass spectrometry coupled to liquid tion and further identification by tandem mass spectrometry
chromatography (LC-ESI MS) provides the most comprehen- (MS/MS). The intensity of the precursor ion dictates the
sive technology for metabolomics studies.1,2 Yet, two different signal-to-noise ratio of the tandem MS fragment ions, which is
mass spectrometry-based approaches have been described in the of critical importance in untargeted metabolomics studies because
field: targeted and untargeted metabolomics. In general, target- metabolites are identified by comparing the entire fragmentation
ing a specific metabolite or a small group of distinct metabolites is pattern of a naturally occurring compound to that of a pure
associated with hypothesis-driven studies3 and involves optimi- standard.
zation of chromatographic conditions (i.e., retention times) and Given the large number of molecules with chemical and
selected reaction monitoring (SRM) transitions with pure structural diversity constituting the metabolome, the method
standards.4 Recent studies showed that up to 50-100 metabo- used to biologically extract metabolites and separate them by
lites can be quantified using this approach.5,6 Untargeted meta- using liquid chromatography is fundamentally related to the
bolomics studies, in contrast, are designed to simultaneously number of features detected by MS with a sufficient signal-
profile the largest number of compounds possible and therefore intensity threshold. The basic philosophy of an untargeted
have the capacity to implicate previously unexplored biochemical metabolomics approach is to detect as many metabolites as
pathways.7,8 It is essential for the untargeted approach to possible to maximize the opportunity of identifying compounds
maximize ionization efficiency of metabolites over a broad mass that are dysregulated in a particular biological condition. There-
range (e.g., m/z 80-1000), since this determines the number fore, studies that comprehensively examine optimal extraction
and intensity (abundance) of the features to be analyzed. A
feature is defined as a molecular entity with a unique m/z and Received: November 11, 2010
retention time value. The number of features can be used as a Accepted: January 25, 2011
general metric for the comprehensiveness of a global Published: February 17, 2011

r 2011 American Chemical Society 2152 dx.doi.org/10.1021/ac102981k | Anal. Chem. 2011, 83, 2152–2161
Analytical Chemistry ARTICLE

methods and LC-MS conditions to detect the largest number of incubated 1-2 min at 80 °C (oven). A heating block or other
metabolites simultaneously are important. The isolation of methods can be used to control the temperature. The solution
metabolites from tissues, cells, or biofluids requires that proteins was centrifuged for 15 min at 13 000 rpm (4 °C), and the
are precipitated and that polar and nonpolar metabolites are supernatant was transferred to a new tube. The same process was
dissolved in solution without degradation. Further, the ideal repeated with the precipitate using 100 μL of hot methanol.
chromatographic conditions should retain and separate the Finally, the supernatants (∼300 μL total) were pooled in an
complex mixture of extracted metabolites using a mobile phase HPLC vial.
(solvent) that promotes ionization of the largest number of Hot EtOH/Ammonium Acetate. A volume of 400 μL of hot
analytes. (80 °C) 60% ethanol/40% water in 5 mM ammonium acetate, 1
In the present study, we used Escherichia coli as a model mM EDTA (pH 7.2) was added to the E. Coli pellet, vortexed for
organism to optimize metabolite extraction and chromatography 30 s, and incubated 1-2 min at 80 °C (oven). The solution was
conditions coupled to ESI-MS. Specifically, we applied seven centrifuged for 15 min at 13 000 rpm (4 °C), and the supernatant
extraction methods involving different aqueous/organic sol- was transferred to a new tube. The same process was repeated
vents, temperature, pH, and molecular weight filters to E. Coli twice, and the resulting supernatants were pooled in a 1.5 mL
cultures. Additionally, we explored the separation of 36 model tube. The solution was desiccated with a vacuum concen-
compounds of varying polarity from a standard mixture using trator (SpeedVac) at room temperature and redissolved in
different reverse-phase columns containing unique stationary 300 μL of 5% ethanol/water 5 mM ammonium acetate (pH
phases. Finally, we investigated the influence of the chromato- 7.2). Finally, the sample was centrifuged again for 10 min at
graphic mobile phase to enhance ionization efficiency of complex 13 000 rpm (4 °C), and the supernatant was transferred to an
mixtures of metabolites in negative ionization mode, which is HPLC vial.
generally associated with the detection of fewer features relative Cold EtOH/Ammonium Acetate. A volume of 300 μL of cold
to positive ionization mode. Overall, our results indicate that (4 °C) 60% ethanol/40% water in 5 mM ammonium acetate, 1
polar solvents (e.g., water, ethanol/water) in combination with mM EDTA (pH 7.2) was added to the E. Coli pellet, vortexed for
high temperature are more efficient in extracting both hydro- 30 s, and the sample was incubated 1 min in liquid nitrogen. The
phobic and hydrophilic metabolites compared to less polar sample was thawed at room temperature and incubated in liquid
solvents such as acetone or methanol. The choice of the nitrogen two more times. Next, the sample was incubated 1 h at
chromatographic mobile-phase conditions in negative ionization -20 °C followed by a 15 min centrifugation at 13 000 rpm. The
mode proved to be strikingly significant, and we report that the resultant supernatant (∼300 μL) was transferred to an
addition of ammonium fluoride substantially increased the HPLC vial.
absolute intensity of nearly all compounds analyzed up to 22- Boiling Water. A volume of 300 μL of LC-MS-grade water in
fold while also increasing the total number of features in E. Coli 1 mM HEPES and 1 mM EDTA (pH 7.2) was added to the
extracts by 2.50-fold compared to mobile phases enriched with E. Coli pellet, vortexed for 30 s, and the sample was incubated 1-
ammonium acetate or formic acid. 2 min in boiling water. Next, the sample was incubated for 1 min
in liquid nitrogen and thawed at room temperature. The
’ EXPERIMENTAL SECTION incubation in liquid nitrogen was repeated. Finally, the sample
was incubated 1 h at -20 °C, followed by 15 min centrifugation
Materials. All pure standards were purchased from Sigma at 13 000 rpm (4 °C). The resultant supernatant was transferred
Aldrich (St. Louis), except Peptide T (GenScript, Piscataway) to an HPLC vial.
and LysoPC (Cayman Chemical, Ann Arbor). Peptides Phe-Gly- Acetone/MeOH. A volume of 400 μL of cold (-20 °C)
Phe-Gly and Thymopoietin II fragment 32-36 were custom acetone was added to the E. Coli pellet, vortexed for 30 s, and
synthesized. Ammonium acetate, ammonium fluoride, formic the sample incubated 1 min in liquid nitrogen. The sample was
acid, and EDTA were purchased from Sigma Aldrich (St. Louis). thawed at room temperature and incubated in liquid nitrogen
LC-MS grade methanol, acetonitrile, and water were purchased two more times. After 1 h at -20 °C, the sample was centrifuged
from J.T. Baker (Phillipsburg). M9 minimal salts (5
) and at 13 000 rpm for 15 min. The resultant supernatant was
casamino acids were purchased from Difco (Franklin Lakes), transferred to a separate vial, and the precipitate was mixed with
and glucose, MgCl2, CaCl2, and meta-phosphoric acid (MPA) 200 μL of cold methanol/water/formic acid (86.5:12.5:1.0). The
were purchsed from Sigma Aldrich (St. Louis). The Microcon sample was vortexed for 30 s and then sonicated for 10 min
YM-3 (3,000 NMWL) devices and HPF Millex filters (4 °C) before leaving the sample 1 h at -20 °C. Next, the sample
(hydrophilic PTFE, 0.20 μm) were purchased from Millipore was centrifuged 15 min at 13 000 rpm (4 °C), and the super-
(Billerica, MA). natant was pooled with the previous. The solution was dried with
Growth of Escherichia coli. The E. Coli strain MC4100 (F-, a vacuum concentrator (SpeedVac) at room temperature and
araD139, Δ(arg F-lac)U169, ptsF25, relA1, flb5301, rpsL 150.λ-) redissolved in 300 μL of 95% acetonitrile/5% water. The final
was grown overnight at 37 °C in minimal media containing M9 solution was then centrifuged for 10 min at 13 000 rpm and the
salts (
1), 2% casamino acids, 0.2% glucose, 1 mM MgCl2, and supernatant transferred to an HPLC vial.
0.1 mM CaCl2. MPA. A volume of 300 μL of cold (4 °C) solution of 5% meta-
Metabolite Extraction Methods. An overnight batch culture phosphoric acid (MPA), 1 mM EDTA, 0.1% formic acid was
of E. Coli (25 mL) was divided into seven aliquots (1.5 mL each), added to the E. Coli pellet, vortexed for 30 s, and the sample was
and the supernatant was removed by centrifugation (5 min at incubated 1 min in liquid nitrogen. The sample was thawed at
3 000 rpm, 4 °C). Metabolites were extracted by using one of the room temperature and incubated in liquid nitrogen two more
following methods. times. Next, the sample was centrifuged 15 min at 13 000 rpm
Hot MeOH. A volume of 200 μL of hot (80 °C) methanol (4 °C), and the resultant supernatant (∼300 μL) was transferred
(100%) was added to the E. Coli pellet, vortexed for 30 s, and to an HPLC vial. It should be noted that the MPA buffer was
2153 dx.doi.org/10.1021/ac102981k |Anal. Chem. 2011, 83, 2152–2161
Analytical Chemistry ARTICLE

Figure 1. (A) Two-dimensional representation of the XCMS matrix of retention time, m/z, and feature intensity values using a multidimensional scaling
(MDS) plot. Data points for extraction methods producing similar features are closer to one another than data points for extraction methods producing
more dissimilar features. (B) Quantification of 31 identified metabolites extracted from E. Coli using seven different methods and analyzed with a C18
column (Cogent Bidentate) in positive ionization mode. Data points and error bars represent mean intensity values and standard deviation for three
replicates. The black solid vertical line indicates the intensity threshold required for tandem mass spectrometry, as established from our experimental
conditions. Parent ions below this threshold produced fragment ions with low signal-to-noise ratio (i.e., at background noise level). The intensity scale of
the X axis is log 10. LysoPC, lysophosphatidylcholine; LysoPE, lysophosphatidylethanolamine.

2154 dx.doi.org/10.1021/ac102981k |Anal. Chem. 2011, 83, 2152–2161

Analytical Chemistry ARTICLE

filtered with an HPF Millex filter (Millipore) before mixing with By using the statistical software R, each row of feature intensities
the E. Coli pellet. was normalized such that the highest value was 1. A two-
YM3. A volume of 300 μL of cold (4 °C) LC-MS-grade water dimensional representation of this matrix was calculated using
in 1 mM HEPES and 1 mM EDTA (pH 7.2) was added to the multidimensional scaling (MDS) as implemented within the
E. Coli pellet, vortexed for 30 s, and the sample was incubated for software R. In negative ionization mode, experimental blanks
1 min in liquid nitrogen. The sample was thawed at room were run in triplicate to remove “background” features arising
temperature and bath sonicated for 5 min in cold water. This from the mobile phases (e.g., ammonium fluoride and ammo-
process was repeated two more times. Next, the sample was nium acetate). “Background” features detected in each of the
centrifuged 10 min at 13 000 rpm (4 °C), and the resultant three blanks were removed from the E. Coli data run with the
supernatant was transferred to a microcon centrifugal filter unit same mobile phase. Standards were manually quantified by
with a YM-3 membrane (3 kDa nominal molecular weight cutoff extracting ion chromatograms and integrating peak intensities
(NMWCO)). The solution was spun down 1 h (4 °C), and the with Qualitative Analysis of MassHunter Workstation (Agilent
retentate was recovered in an HPLC vial. It should be noted that Technologies).
the membrane was spin-rinsed with deionized water 5 times to
remove trace amounts of glycerin before applying the sample.
Standard Mix 1 and 2. Each compound was dissolved in 50% ’ RESULTS
methanol/water and prepared at a final concentration of 10 μM
and 0.1 μM for LC-MS analysis. Influence of the Extraction Method on the Analysis of
LC-MS and MS/MS Analysis. Analyses were performed Polar and Nonpolar Metabolites. E. Coli pellets obtained from
using an HPLC system (1200 series, Agilent Technologies) the same batch culture were used to extract metabolites with
coupled to a 6538 UHD Accurate-Mass Q-TOF (Agilent Tech- seven different protocols (see the Experimental Section). The
nologies) operated in positive (ESIþ) or negative (ESI-) selected protocols represent examples in which different funda-
electrospray ionization mode. Vials containing extracted meta- mental conditions for metabolite solubility and stability are
bolites using one of the seven methods described above or the varied such as solvent polarity, temperature, pH, and molecular
standard mixture were kept at -20 °C prior to LC-MS analysis. weight cutoff filtering. In brief, in method Hot MeOH, metabo-
E. Coli extractions and standard mixtures were separated using a lites were extracted using hot 100% methanol. Method Hot
Cogent Bidentate C18: 4 μm, 100 Å, 150 mm
2.1 mm i.d. EtOH/Ammonium Acetate was modified from Buescher et al.,10
(catalog no. 40018-15P-2), a Waters XBridge C18, 3.5 μm, 135 with metabolites extracted using hot ethanol/water buffered at
Å, 150 mm
1.0 mm i.d. (part no. 186003128), or an Imtakt pH 7.2. In both methods, the solvents were preheated and mixed
Scherzo SM-C18, 3 μm, 13 nm, 150 mm
2 mm i.d. (product with the E. Coli pellet at 80 °C for a short period (∼1-2 min) to
no. SM025) column. When the instrument was operated in avoid thermal degradation or methyl/ethyl ester formation.
positive ionization mode, regardless of the column used, the Method Cold EtOH/Ammonium Acetate used the same solvent
solvent system was A = 0.1% formic acid in water, and B = 0.1% as method Hot EtOH/Ammonium Acetate, but the solvent was
formic acid in acetonitrile. When the instrument was operated in precooled in ice and the extraction was performed at low
negative ionization mode, we used one of the following pairs of temperature. Method BoilingWater was modified from Bhatta-
solvent systems: A = 0.1% formic acid in water, B = 0.1% formic charya et al.,11 with metabolites extracted using a polar solvent
acid in acetonitrile; A2 = 1 mM ammonium fluoride in water, B2 (100% H2O) buffered at pH 7.2 and incubated for a short time in
= acetonitrile; A3 = 5 mM ammonium acetate in water, B3 = 5 boiling water. In method Acetone/MeOH, metabolites were
mM ammonium acetate in 90% acetonitrile; A4 = 1 mM extracted at low temperature using a precooled nonpolar solvent
ammonium acetate in water, B4 = 200 mM ammonium acetate (100% acetone). Method MPA was modified from Rellan-
in 50% acetonitrile. The linear gradient elution used started at Alvarez et al.,12 with metabolites extracted using strong acidic/
100% A (time 0-5 min) and finished at 100% B (35-40 min). aqueous conditions (meta-phosphoric acid in water) at low
The injection volume was 8 μL. ESI conditions were gas temperature. Finally, in method YM3, metabolites were extracted
temperature 325 °C, drying gas 11 L/min, nebulizer 30 psig, using a cold aqueous solvent buffered at pH 7.2 and isolated
fragmentor 120 V, and skimmer 65 V. The instrument was set to using a centrifugal filter unit with a cutoff of 3 kDa. All extractions
acquire over the m/z range 80-1000 with an acquisition rate of were run as triplicates under the same LC-MS conditions using
1.3 spectra/s. MS/MS was performed in targeted mode, and the the Cogent Bidentate reverse-phase C18 column in positive
instrument was set to acquire over the m/z range 50-1000, with ionization mode (see the Experimental Section). Each data set
a default iso width (the width half-maximum of the quadrupole was visualized using a multidimensional scaling plot (Figure 1A)
mass bandpass used during MS/MS precursor isolation) of 4 m/ to show similarities in the results. With the use of feature
z. The collision energy was fixed at 20 V. intensities from the XCMS matrix, the data were scaled such
Data Processing. LC-MS data from the E. Coli extractions that similar methods are near each other and dissimilar methods
(ESIþ and ESI- modes) were processed by using XCMS are farther apart from each other. The multidimensional scaling
software9 (version 1.24.1) to detect and align features. Each plot shows short distances between the methods Hot EtOH/
metabolite extraction method was compared using the same Ammonium Acetate, Cold EtOH/Ammonium Acetate, and
column (Cogent Bidentate C18) and ionization mode (ESIþ). BoilingWater, highlighting that the number and intensities of
XCMS analysis of these data provided a matrix containing the features detected by XCMS in each method are similar. Hot
retention time, m/z value, and intensity of each feature for every MeOH is also relatively similar to the previous methods. Three
extraction method discussed above. Each row in the matrix extraction protocols, however, are strikingly different from the
represented a feature. It is important to note that while the others: Acetone/MeOH, MPA, and YM3. We deduced that
retention time and m/z values for each feature were consistent Acetone/MeOH is the most nonpolar of the seven extractions,
among extraction methods, the intensities of the features varied. and YM3 is the most polar.
2155 dx.doi.org/10.1021/ac102981k |Anal. Chem. 2011, 83, 2152–2161
Analytical Chemistry ARTICLE

Figure 2. Extracted ion chromatograms of 31 standard compounds separated by reverse-phase chromatography using a (A) XBridge C18, (B) Cogent
Bidentate C18, and (C) Scherzo SM-C18 column.
The mean intensity of 31 specific metabolites identified in signal-to-noise, produces ambiguous and unreliable fragment
E. Coli by using MS/MS data are shown in Figure 1B. Metabolites ions. On the basis of our experimental conditions, we established
characterized by different polarities and chemical functional a threshold for which the intensity of detected metabolites did
groups were extracted with different efficiencies based on the not allow for reliable MS/MS fragmentation (see black line in
method used. The intensity of a feature as determined by its Figure 1B). Our results indicate that the Acetone/MeOH
integrated LC-MS peak area has important implications in that protocol is the least efficient extraction to profile hydrophilic
metabolite identification requires MS/MS analysis. Tandem MS and hydrophobic metabolites simultaneously. Although this
in Q-TOF analyzers requires ion isolation, and without sufficient method provided increased detection of 4 phospholipids, only
2156 dx.doi.org/10.1021/ac102981k |Anal. Chem. 2011, 83, 2152–2161
Analytical Chemistry ARTICLE

Figure 3. Quantification of 36 standard compounds analyzed in negative ionization mode (ESI-) by using three different additives in the mobile phase:
1 mM ammonium fluoride, 5 mM ammonium acetate, or 0.1% formic acid. Compounds were separated by reverse-phase chromatography by using an
XBridge C18 column. Fold values indicate the difference in intensity between ammonium fluoride and the closest mobile phase. The intensity scale of the
X axis is log 10.

14 out of the 31 metabolites shown were detected. In contrast, methods is consistent with the clustering displayed in Figure 1A
method YM3 did not efficiently extract the 4 phospholipids or (see blue area). With the exception of succinyl-CoA, the rest of
other structurally unrelated metabolites such as flavin adenine the 31 metabolites were detected with high reproducibility in all
dinucleotide (FAD), flavin mononucleotide (FMN), and coen- three extractions. A total of 19 of the metabolites showed higher
zyme A (CoA). In addition, the intensity of metabolites such as intensity with Hot EtOH/Ammonium Acetate, as compared to 9
acetyl-CoA, succinyl-CoA, γ-glutamylcysteine, adenylsuccinic metabolites with BoilingWater, and only 2 with Cold EtOH/
acid, and ATP/ADP was below or slightly above the MS/MS Ammonium Acetate. Importantly, none of the 31 metabolites
threshold level. Using method Hot MeOH we detected 29 out of formed methyl or ethyl esters due to the short incubation time in
the 31 metabolites shown, with the exceptions being succinyl- boiling water or hot ethanol. Overall, we interpret our results in
CoA and ATP. It should be noted, however, that the intensity of 4 decreasing order of efficiency in extracting both polar and
metabolites (acetyl-CoA, FAD, CMP, and adenylsuccinic acid) nonpolar metabolites simultaneously as follows: Hot EtOH/
was below or slightly above the MS/MS threshold. With the use Ammonium Acetate < BoilingWater < Cold EtOH/Ammonium
of method MPA, all 31 metabolites were detected with intensities Acetate < MPA < Hot MeOH < Acetone/MeOH = YM3.
above the established MS/MS threshold, but we observed an Exploring Different C18 Stationary-Phases for Metabolite
unexpected high analytical variability for most of the compounds Profiling. Reliable quantification of thousands of chemically
(see error bars in Figure 1B). Methods Hot EtOH/Ammonium diverse features requires optimization of chromatographic se-
Acetate, Cold EtOH/Ammonium Acetate, and BoilingWater paration to reduce ion-suppression effects. Traditionally, untar-
appear to be the most efficient extraction protocols examined geted metabolomics analyses have been performed using reverse-
in this study, and the similarity of the results from these three phase (RP) C18 columns because they generally result in the
2157 dx.doi.org/10.1021/ac102981k |Anal. Chem. 2011, 83, 2152–2161
Analytical Chemistry ARTICLE

Figure 4. (A) Venn diagram representing the total number of features from LC-MS data of E. Coli samples extracted using the method Boiling Water
and analyzed using ammonium acetate or ammonium fluoride enriched mobile phases. (B) Quantification of 39 metabolites from E. Coli analyzed using
ammonium acetate and ammonium fluoride enriched mobile phases. The intensity scale of the X axis is log 2. Metabolites were separated by reverse-
phase chromatography by using an XBridge C18 column and detected in negative ionization mode (ESI-). Identification is based on accurate mass and
MS/MS data. Fold values indicate the difference in intensity between ammonium fluoride and ammonium acetate. Examples of unique metabolites
detected with ammonium fluoride are also represented. Data points and error bars represent mean intensity values and standard deviation for three
replicates.

detection of more features. RP C18 columns, however, are the full list of compounds in Standard Mix 1) using each column
limited in their capacity to retain hydrophilic compounds and with the same gradient and mobile phase (A = water, 0.1% FA; B
consequently result in ion suppression for polar metabolites in = acetonitrile, 0.1% FA). With the XBridge column, 14 out of the
the dead volume, thereby limiting MS coverage of the metabo- 31 compounds (45%) coeluted within the first 2 min
lome. Recent developments in RP C18 technology offer oppor- (Figure 2A). Similar results were obtained with traditional C18
tunities to improve retention of polar molecules and therefore columns (e.g., Zorbax, Atlantis T3) (data not shown). The
increase metabolite coverage in metabolomics analysis. We Cogent column retained more compounds, with only 7 of them
analyzed the ability of three different RP C18 columns with (22%) coeluting within the first 2 min (Figure 2B). With the
unique properties to retain polar model compounds: (i) the Scherzo column, we observed a remarkable improvement in the
XBridge C18 column characterized by broad pH range stability retention of polar compounds with only three compounds (10%)
(pH 1-11), (ii) the Cogent Bidentate C18 column characterized coeluting within the first 2 min (Figure 2C). Overall, all three
by silicon-hydride (Si-H) groups instead of the common silanol columns showed good performance in separating the rest of the
group (Si-OH),13,14 and (iii) the multimodal Scherzo SM-C18 compounds, although taurocholic acid, coenzyme A (CoA), and
column containing cation and anion ligands that allow for acetyl-CoA did not elute from the Scherzo column due to strong
reverse-phase separation in addition to anion and cation ex- ion exchange interactions with the stationary phase. Increasing
change. We analyzed a standard mixture of 31 model compounds the ionic strength of the mobile phase by adding 5 mM
characterized by different polarities (e.g., amino acids, tricar- ammonium acetate resulted in elution of taurocholic acid.
boxylic acids, vitamins, peptides, and xenobiotics, see Figure 2 for Further increasing the ionic strength of the mobile phase with
2158 dx.doi.org/10.1021/ac102981k |Anal. Chem. 2011, 83, 2152–2161
Analytical Chemistry ARTICLE

200 mM ammonium acetate resulted in the elution of CoA and using ammonium fluoride additive were not detected when using
acetyl-CoA. an ammonium acetate additive. A similar distribution of unique
Optimizing the Mobile Phase in Negative Ionization features was also found when the method BoilingWater and the
Mode for Global Metabolite Profiling. In electrospray ioniza- method Hot MeOH enriched with the same mobile phases were
tion, the composition of the solvent (i.e., mobile phase in LC- compared. A total of 1 008 features were detected using Hot
ESI MS) influences the gas-phase acid-base processes required MeOH with an ammonium acetate enriched mobile phase, and
for the ionization of the compounds to be analyzed. Global only 24% of these features overlapped with BoilingWater using
profiling of metabolites in positive ionization mode generally the same mobile phase (data not shown). The increased sensi-
produces more features compared to negative mode, most likely tivity achieved by using an ammonium fluoride enriched mobile
due to higher efficiency of protonation relative to deprotonation. phase was also reflected in the number of putative metabolite
Protonation is facilitated by the addition of an acid to the mobile identifications made on the basis of accurate mass from the
phase, such as formic acid, acetic acid, or TFA. In negative METLIN database. With the use of ammonium fluoride, 722
ionization mode, deprotonation in the gas phase is typically features out of 4 213 (17.1%) matched with known compounds
promoted by the addition of ammonium salts such as ammonium in METLIN (error < 5 ppm, [M - H]-). With the use of
formate or ammonium acetate. Given the strong basicity of the ammonium acetate, 211 out of the 1647 (12.8%) matched with
fluoride anion (F-) in the gas phase,15 we examined the effect of known compounds in METLIN. The difference reflects nearly a
adding ammonium fluoride to the mobile phase in negative- 3.5-fold increase in database hits. It is noteworthy that 81% of the
mode analysis. A previous report showed that fluoride anions in METLIN database hits using ammonium fluoride were not
the electrospray solvent resulted in increased deprotonation, present in the ammonium acetate analysis. Features with high
[M - H]-, of neutral steroids with higher abundances than other intensity, presumably corresponding to abundant metabolites,
anions tested.16 Using Standard Mix 2 containing 36 compounds were typically detected with both enriched mobile phases,
at 10 μM (see Figure 3 for full list of compounds), three different whereas less abundant metabolites were often uniquely detected
mobile phases were compared with the same reverse-phase C18 in ammonium fluoride enriched analyses. To confirm this
column (i.e., XBridge) and MS conditions in negative ionization observation, the mean intensity value was determined for various
mode: (i) 1 mM ammonium fluoride, (ii) 5 mM ammonium metabolites structurally identified based on accurate mass, reten-
acetate, and (iii) 0.1% formic acid. Mobile phases enriched with tion time, and MS/MS data (Figure 4B). The results were
ammonium fluoride and ammonium acetate were maintained at consistent with our previous data using Standard Mix 2 contain-
pH∼ 7, ruling out the possibility that differences in ionization ing 36 compounds. All metabolites identified showed higher
efficiency are related to the pH of the solution. intensity using an ammonium fluoride enriched mobile phase,
Figure 3 shows the intensity values and fold change of each with important cellular metabolites being detected only in the
compound, revealing that ammonium fluoride is a superior presence of ammonium fluoride.
additive to increase ionization efficiency in negative ionization
mode. On average, ammonium fluoride increased the intensity of
the compounds analyzed by 5.7-fold, including a 15-, 16-, and 22- ’ DISCUSSION
fold increase for quinidine, peptide T, and lysophosphatidylcho- Global metabolite profiling of biological samples is a challen-
line, respectively. Notably, when the same 36 compounds were ging task due to the chemical and structural diversity of naturally
prepared at 0.1 μM, 20 compounds (55%) were not detected occurring compounds ranging from polar metabolites such as
with 0.1% formic acid and 18 (50%) were not detected with amino acids and nucleotides to nonpolar molecules such as
ammonium acetate. Only 6 compounds (16%) were not detected steroids and membrane lipids. As a result of this diversity,
with ammonium fluoride, and all remaining compounds showed methods used for metabolite extraction and metabolite separa-
significantly higher intensity values with ammonium fluoride tion significantly influence the number and intensity of com-
compared to ammonium acetate and formic acid (data not pounds detected by ESI-MS analysis. The choice of extraction
shown). It is worth noting that lysophosphatidylcholine was and chromatography methods biases the chemical distribution of
detected as [M - H]- at 10 μM in negative ionization mode metabolites detected and is therefore problematic for untargeted
using ammonium fluoride; however, no signal of this compound metabolomics investigations aimed at accomplishing unbiased
was observed at 0.1 μM. Lysophosphatidylcholine, however, was profiling. The obvious complexity in performing untargeted
detected at 0.1 μM in positive ionization mode using 0.1% studies is that the metabolites of potential interest are unknown
formic acid. and therefore extraction and chromatography methods cannot
The total number of features and their absolute intensity were be tailored toward a specific chemical class of compounds. In this
also compared from E. Coli extractions analyzed by using context, it is unclear what evaluation criteria should be used to
ammonium fluoride and ammonium acetate enriched mobile assess the quality of untargeted metabolomics extraction and
phases. E. Coli samples in addition to blank samples were run as chromatography methods. The approach we have developed
triplicates with each mobile phase, and features consistently here involves using our metabolomics software XCMS to analyze
present in all three blanks were subtracted from the E. Coli data the number and intensity of features identified in each of different
as “background”. Then, only features present in the 3 E. Coli extraction protocols and LC-MS conditions. We interpret
replicates with intensity values above 5 000 ion counts were methods leading to the identification of more features of greater
considered for quantification purposes. The Venn diagram in intensity to be better suited for untargeted studies.
Figure 4A shows a total of 4 213 features obtained with the Decades of research has provided an extensive library of
method BoilingWater using an ammonium fluoride enriched detailed extraction and chromatography methods for analysis
mobile phase and 1 647 features from the method BoilingWater of unique classes of compounds, and it is not our intent to
using an ammonium acetate enriched mobile phase, which is a comprehensively survey all of them here. Rather, our study is
2.5-fold increase. Importantly, 77% of the features analyzed by aimed at providing an overview of easy and rapid extraction and
2159 dx.doi.org/10.1021/ac102981k |Anal. Chem. 2011, 83, 2152–2161
Analytical Chemistry ARTICLE

reverse phase (RP) LC-MS protocols for the metabolomics by using an ammonium fluoride enriched mobile phase. To the
scientist to use as a general guideline. Optimization of metabolite best of our knowledge, this approach has not been described
extraction has been recently pursued using a two-stage approach before for metabolomics studies. The use of an ammonium
or biphasic mixtures.17 Although these methods may reduce fluoride enriched mobile phase in negative ionization mode is not
complexity in the number of metabolites to be separated and intended to replace positive ionization mode analysis in meta-
ionized and may ultimately extend metabolite coverage, we have bolomics studies. Ammonium fluoride is stable in different
not explored them because they are generally more time- solvents (e.g., water, acetonitrile, methanol) and led to repro-
consuming and prone to analytical error. Other studies have ducible results. In addition, no contamination of the HPLC lines
shown that many metabolites can be extracted by a broad was observed after using ammonium fluoride. Importantly, the
spectrum of solvent mixtures,18-20 suggesting that there is no highest sensitivity was achieved when using 1 mM ammonium
specific extraction protocol that should be used in metabolomics. fluoride, and concentrations of 5 mM or higher introduced
Our data corroborate this point but highlight that a number of significant background in the mass spectra. The source of back-
considerations should be taken into account when carrying out ground was attributed to the impurities of the ammonium
global metabolite profiling studies. For example, highly polar fluoride stock solution and to the interaction at pH 7.0 of fluoride
solvents such as water or 100% nonpolar organic solvents such as anions with the silica of the stationary phase of most columns
acetone/methanol are not particularly suitable to extract both tested. The XBridge column, however, is stable within a wide pH
hydrophilic and hydrophobic metabolites simultaneously. In range and performed excellent with 1 mM ammonium fluoride,
particular, solvents with intermediate polarity such as a mixture generating mass spectra with low background. When a polymeric
of ethanol/water are more appropriate for this goal. We also (PSDVB) reverse-phase column was tested with ammonium
report that incubation of the sample with the solvent at high fluoride, such as the Hamilton PRP-1, the background was
temperatures (80-100 °C) for short periods of time (1-2 min) almost absent (similar to ammonium acetate) with no effect on
is more efficient than incubation of the sample at cold tempera- sensitivity. Overall, the consistency of our data with standard
tures (-20 °C). Water may have a polarity similar to some compounds and biological samples suggests that ammonium
organic solvents when boiled,21 which might explain the similar- fluoride should be used as a standard additive to mobile phases in
ity of the method Boiling Water with extraction protocols using negative ionization mode for global metabolomics studies.
high percentages of ethanol such as the methods Hot EtOH/
Ammonium Acetate and Cold EtOH/Ammonium Acetate. The ’ AUTHOR INFORMATION
general utility of the described extraction protocols to other cell
types, which have heavily weighted distributions of nonpolar Corresponding Author
metabolites (e.g., adipose tissue), possess increased tensile prop- *Oscar Yanes: address, Metabolomics Platform, Spanish Biome-
erties (e.g., muscle tissue), or contain functionalities that are not dical Research Center in Diabetes and Associated Metabolic
stable to the conditions employed, requires further investigation. Disorders (CIBERDEM), University Rovira i Virgili, Avda.
Much attention has been dedicated recently to different Països Catalans 26, 43007 Tarragona, Spain; phone, þ34 977-
chromatographic approaches to improve separation and analysis 297054; e-mail, oscar.yanes@urv.cat. Gary Siuzdak: address, The
of naturally occurring compounds. Hydrophilic interaction chro- Scripps Research Institute, Center for Metabolomics, 10550
matography (HILIC)-like stationary phases22 or the use of ion North Torrey Pines Road, La Jolla, CA 92037; phone, þ1 858-
pairing agents with reverse-phase columns23-25 improves the 784-9415; e-mail, siuzdak@scripps.edu.
retention of polar metabolites; however, each of these ap-
proaches has drawbacks for processing global metabolite profil- ’ ACKNOWLEDGMENT
ing data. HILIC columns do not retain hydrophobic compounds
well and are generally associated with broader peak shapes. The We thank Dr. Jesus Torres-Bacete for providing us with
use of ion pairing agents (e.g., tetrabutylammonium (TBA)) has Escherichia coli samples. We gratefully acknowledge financial
led to inconsistent results, lengthy equilibration, and general support from the California Institute of Regenerative Medicine
incompatibility with mass spectrometry and typically requires a (Grant TR1-01219), the National Institutes of Health (Grants
dedicated negative-mode LC stack as some ion pairing agents are R24 EY017540-04, P30 MH062261-10, and P01 DA026146-02),
difficult to remove from the instrument lines and cause contam- and NIH/NIA Grant L30 AG0 038036 (G.J.P.). Financial
ination affecting analysis in the positive-ionization mode. In support was also received from the Department of Energy
contrast, we demonstrate that the stationary phases utilized in (Grants FG02-07ER64325 and DE-AC0205CH11231).
our study show unique advantages for global metabolite profiling.
We found the flexibility of the Scherzo column to be particularly ’ REFERENCES
effective, as it combines a nonpolar stationary phase with anion (1) Nordstrom, A.; Want, E.; Northen, T.; Lehtio, J.; Siuzdak, G.
and cation ligands for ion exchange separation. Perhaps the major Anal. Chem. 2008, 80 (2), 421–429.
drawback is the strong ionic interaction of this stationary phase (2) Want, E. J.; Nordstrom, A.; Morita, H.; Siuzdak, G. J. Proteome
with some metabolites such as coenzyme A. This issue can be Res. 2007, 6 (2), 459–468.
resolved by increasing the ionic strength of the mobile phase with (3) Olszewski, K. L.; Mather, M. W.; Morrisey, J. M.; Garcia, B. A.;
200 mM ammonium acetate; however, it is important to mention Vaidya, A. B.; Rabinowitz, J. D.; Llinas, M. Nature 2010, 466 (7307),
774–778.
that percentages of acetonitrile higher than 50 led to precipita-
(4) Lu, W.; Bennett, B. D.; Rabinowitz, J. D. J Chromatogr., B: Anal.
tion of the ammonium salt, and low percentages of acetonitrile Technol. Biomed. Life Sci. 2008, 871 (2), 236–242.
likely affects the elution of highly hydrophobic compounds from (5) Bennett, B. D.; Kimball, E. H.; Gao, M.; Osterhout, R.; Van Dien,
the column. S. J.; Rabinowitz, J. D. Nat. Chem. Biol. 2009, 5 (8), 593–599.
Finally, a striking result from our work is the substantial (6) Kwon, Y. K.; Higgins, M. B.; Rabinowitz, J. D. ACS Chem. Biol.
improvement of ionization efficiency in negative ionization mode 2010, 5 (8), 787–795.

2160 dx.doi.org/10.1021/ac102981k |Anal. Chem. 2011, 83, 2152–2161

Analytical Chemistry ARTICLE

(7) Wikoff, W. R.; Anfora, A. T.; Liu, J.; Schultz, P. G.; Lesley, S. A.;
Peters, E. C.; Siuzdak, G. Proc. Natl. Acad. Sci. U.S.A. 2009, 106 (10),
3698–3703.
(8) Yanes, O.; Clark, J.; Wong, D. M.; Patti, G. J.; Sanchez-Ruiz, A.;
Benton, H. P.; Trauger, S. A.; Desponts, C.; Ding, S.; Siuzdak, G. Nat.
Chem. Biol. 2010, 6 (6), 411–417.
(9) Smith, C. A.; Want, E. J.; O’Maille, G.; Abagyan, R.; Siuzdak, G.
Anal. Chem. 2006, 78 (3), 779–787.
(10) Buescher, J. M.; Moco, S.; Sauer, U.; Zamboni, N. Anal. Chem.
2010, 82 (11), 4403–4412.
(11) Bhattacharya, M.; Fuhrman, L.; Ingram, A.; Nickerson, K. W.;
Conway, T. Anal. Biochem. 1995, 232 (1), 98–106.
(12) Rellan-Alvarez, R.; Hernandez, L. E.; Abadia, J.; Alvarez-Fer-
nandez, A. Anal. Biochem. 2006, 356 (2), 254–264.
(13) Pesek, J. J.; Matyska, M. T.; Fischer, S. M.; Sana, T. R.
J. Chromatogr., A 2008, 1204 (1), 48–55.
(14) Pesek, J. J.; Matyska, M. T.; Hearn, M. T.; Boysen, R. I.
J. Chromatogr., A 2009, 1216 (7), 1140–6.
(15) Wang, W. Q.; Cole, R. B. Anal. Chem. 2009, 81 (21),
8826–8838.
(16) Rannulu, N.; Cole, R. B. 58th American Society for Mass
Spectrometry Conference on Mass Spectrometry and Allied Topics, Salt
Lake City, UT, 2010.
(17) Masson, P.; Alves, A. C.; Ebbels, T. M.; Nicholson, J. K.; Want,
E. J. Anal. Chem. 2010, 82 (18), 7779–7786.
(18) Dietmair, S.; Timmins, N. E.; Gray, P. P.; Nielsen, L. K.;
Kromer, J. O. Anal. Biochem. 2010, 404 (2), 155–164.
(19) El Rammouz, R.; Letisse, F.; Durand, S.; Portais, J. C.; Moussa,
Z. W.; Fernandez, X. Anal. Biochem. 2010, 398 (2), 169–177.
(20) Rabinowitz, J. D.; Kimball, E. Anal. Chem. 2007, 79 (16),
6167–6173.
(21) Raynie, D. E. Anal. Chem. 2010, 82 (12), 4911–4916.
(22) Bajad, S. U.; Lu, W.; Kimball, E. H.; Yuan, J.; Peterson, C.;
Rabinowitz, J. D. J. Chromatogr., A 2006, 1125 (1), 76–88.
(23) Cordell, R. L.; Hill, S. J.; Ortori, C. A.; Barrett, D. A.
J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2008, 871 (1),
115–124.
(24) Lu, W.; Clasquin, M. F.; Melamud, E.; Amador-Noguez, D.;
Caudy, A. A.; Rabinowitz, J. D. Anal. Chem. 2010, 82 (8), 3212–3221.
(25) Luo, B.; Groenke, K.; Takors, R.; Wandrey, C.; Oldiges, M.
J. Chromatogr., A 2007, 1147 (2), 153–164.

2161 dx.doi.org/10.1021/ac102981k |Anal. Chem. 2011, 83, 2152–2161