massspectrometrystrategiesin metabolomicsaccepted as the gold standard in metabolite structural...

Mass Spectrometry Strategies inMetabolomics*Published, JBC Papers in Press, June 1, 2011, DOI 10.1074/jbc.R111.238691

Zhentian Lei, David V. Huhman, and Lloyd W. Sumner1

From the Plant Biology Division, The Samuel Roberts Noble Foundation,Ardmore, Oklahoma 73402

MS has evolved as a critical component in metabolomics,which seeks to answer biological questions through large-scalequalitative and quantitative analyses of the metabolome. MS-based metabolomics techniques offer an excellent combinationof sensitivity and selectivity, and they have become an indispen-sable platform in biology andmetabolomics. In thisminireview,various MS technologies used in metabolomics are briefly dis-cussed, and future needs are suggested.

Metabolomics is idealized as the large-scale, qualitative, andquantitative study of all metabolites in a given biological sys-tem. Unlike transcripts and proteins, the molecular identity ofmetabolites cannot be deduced from genomic information.Thus, the identification and quantification of metabolites mustrely on sophisticated instrumentation such as MS, NMR spec-troscopy, and laser-induced fluorescence detection. Each ofthese technologies has its own unique advantages and disad-vantages. Optimal selection of a particular technology dependson the goals of the study and is usually a compromise amongsensitivity, selectivity, and speed.NMR is highly selective and non-destructive and is generally

accepted as the gold standard in metabolite structural elucida-tion, but it suffers from relatively lower sensitivity. Laser-in-duced fluorescence is one of the most sensitive techniques, butit lacks the chemical selectivity that is critical in structural iden-tification. In contrast, MS offers a good combination of sensi-tivity and selectivity. Modern MS provides highly specificchemical information that is directly related to the chemicalstructure such as accurate mass, isotope distribution patternsfor elemental formula determination, and characteristic frag-ment ions for structural elucidation or identification via spec-tral matching to authentic compound data. Moreover, the highsensitivity of MS allows detection and measurement of pico-mole to femtomole levels of many primary and secondarymetabolites. These unique advantages make MS an importanttool in metabolomics (1, 2).Modern MS offers an array of technologies that differ in

operational principles and performance. Variations include

ionization technique, mass analyzer technology, resolvingpower, and mass accuracy. The most common ionization tech-niques in metabolomics include electron ionization, electro-spray ionization (ESI),2 and atmospheric pressure chemicalionization (APCI). Other ionization techniques such as chem-ical ionization, MALDI, and, more recently, desorption ESI(DESI) (3) and extractive ESI (EESI) (4) have also been used.Mass analyzers with different resolving powers have also beenused in metabolomics. These include ultrahigh and high reso-lution MS such as Fourier transform ion cyclotron resonanceMS (FT-ICR-MS), orbitrapMS, and multipass TOF-MS. How-ever, lower resolution instruments such as ion traps (both linearand three-dimensional quadrupoles) and single quadrupolesare utilized by many. Each of these mass analyzers has its ownadvantage and limitation. Selection of a specific MS platformfor metabolomics depends on the goal of the metabolomicsprojects, throughput, and instrumental costs. In this minire-view, we discuss MS strategies currently incorporated intometabolomics, including directMS analysis andMS coupled tochromatography for the analysis of the chemically complexmetabolome.

Direct MS Analysis

Direct MS analyses sample crude mixtures without chro-matographic separations. This approach is the least informativebut does provide a high throughput screening tool that is oftenthe only practical choice for large sample numbers such asthose encountered during clinical trials or screening of largemutant populations. For example, direct MS analyses enabledsuccessful screening of thousands of yeast mutants to help elu-cidate the functions ofmutated genes (5). DirectMS applicabil-ity inmetabolomics is broadened by advanced instrumentationcapable of high resolution, accurate mass measurements, andtandemMS (such as FT-ICR-MS and orbitrap MS).FT-ICR-MS is an important and powerful tool in direct MS

analyses due to its ultrahigh resolution (�1,000,000) and massaccuracy (�1 ppm). The high mass resolution is useful inempirical formula calculations and compound identification.For example, direct FT-ICR-MS analysis of a crude oil samplerevealed �111,000 features in a singular mass spectrum, fromwhich�8300 peaks could be assigned a unique elemental com-position (6). Similarly, �1000 unambiguous chemical formulaswere reportedly identified from the aerial parts of Arabidopsisusing direct FT-ICR-MS (7). The disadvantage of FT-ICR-MSis its formidable cost, which prohibits its widespread availabil-ity and routine use in many metabolomics laboratories.The orbitrap is a relatively newer mass analyzer that uses an

electrostatic field to trap ions (8). Orbitraps also have very high* This work was supported by Oklahoma Center for the Advancement of Sci-

ence and Technology Grant PSB10-027, National Science FoundationGrants MCB 1024976 and MCB 0520140, the LECO Corp., and Agilent Tech-nologies. Financial personnel support was provided by the Samuel RobertsNoble Foundation. This is the third article in the Thematic MinireviewSeries on Biological Applications of Mass Spectrometry. This minireviewwill be reprinted in the 2011 Minireview Compendium, which will be avail-able in January, 2012.

1 To whom correspondence should be addressed. E-mail: [email protected].

2 The abbreviations used are: ESI, electrospray ionization; APCI, atmosphericpressure chemical ionization; DESI, desorption ESI; EESI, extractive ESI; FT-ICR-MS, Fourier transform ion cyclotron resonance MS; CE, capillary elec-trophoresis; HS-SPME, headspace solid-phase microextraction; NP, nor-mal-phase; RP, reversed-phase; HILIC, hydrophilic interaction liquidchromatography; UHPLC, ultra-HPLC; CZE, capillary zone electrophoresis;CEC, capillary electrochromatography.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 286, NO. 29, pp. 25435–25442, July 22, 2011© 2011 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.

JULY 22, 2011 • VOLUME 286 • NUMBER 29 JOURNAL OF BIOLOGICAL CHEMISTRY 25435

MINIREVIEW This paper is available online at www.jbc.org

by guest on January 6, 2021http://w

ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

resolving power (typically 150,000) and excellentmass accuracy(1–5 ppm) and have been widely used in direct MS analyses ofbovine lipids (9), yeast sphingolipids (10), and plantmetabolites(11). Multiple pass TOF-MS instruments with high resolvingpower (40,000) and mass accuracy (�5 ppm) have also beenrecently developed and further facilitate the use of high resolu-tion MS in metabolomics. Three-dimensional and linear iontraps have been used in direct MS analysis of metabolites aswell. However, their relatively lower resolution and mass accu-racy limit their roles in direct analysis of complex samples.The development of several new ionization techniques such

as DESI, direct analysis in real time, and EESI facilitates the useof direct MS in metabolomics. DESI uses a charged ESI aerosolfocused on a separate surface to ionize the surface analytes,allowing not only liquid but also solid sample analyses directlybyMS (12). For example, DESI-MS revealed heterogeneous dis-tribution of antifungal compounds on the native surface of sea-weed (13) and spatial accumulation of different lipids on ratspinal cords (14). Other techniques such as direct analysis inreal-timeMS (15) andEESI-MS (4) have beenused also in directMS analyses, including analyses of insect hormones (16); met-abolic fingerprinting of human serum (17); screening of pesti-cides in produce (18); analyses of urine, milk, and pollutedwater (4); and fingerprinting of olive oil without sample prepa-ration (19). EESI has also been used to directly analyze gaseoussamples (20).Overall, direct MS analyses have been used for a diversity of

analytes; however, direct MS methods are very susceptible toion suppression or enhancement. In addition, direct MS datainterpretation can be challenging, as uniquemetabolite ions aredifficult to distinguish from adduct and product ions. Anotherdisadvantage is the inability to differentiate isomers. Themajority of these disadvantages can be surmounted by couplingMS to chromatographic separations.

Chromatography Coupled to MS

Coupling chromatography toMS offers an excellent solutionto complex mixture analyses and has been extensively used inmetabolomics. Chromatographic separation of metabolitesprior to MS analyses has several advantages: 1) reduces matrixeffects and ionization suppression, 2) separates isomers, 3) pro-vides additional and orthogonal data (i.e. retention time/factor/index) valuable for metabolite annotation, and 4) allows formore accurate quantification of individual metabolites. Cur-rently, three predominant chromatographic techniques havebeen incorporated inMS-basedmetabolomics, i.e.GC, LC, andcapillary electrophoresis (CE). Multidimensional separationtechniques such as two-dimensional GC (GC�GC) and LC(LC�LC) have further enabled the separation of even morecomplex biological mixtures but are less widely employed.Here, we review the applications of chromatography coupled toMS for metabolite profiling.GC-MS—GC-MS is ideally suited for the analyses of both

volatile and nonvolatile compounds following derivatization.The high resolution and reproducible chromatographic sepa-rations offered by modern capillary GC make it an excellenttool for complex metabolic mixture analyses. In addition, thestandardized MS electron ionization energy of 70 eV leads to

reproduciblemass spectra and highly transferable electron ion-ization MS spectral libraries that allow compound identifica-tion through mass spectral library matching. The highly repro-ducible retention indices can also be used for orthogonalconfirmation of compound identification such as in identifica-tion of stereoisomers that often produce similar mass spectrabut distinctly separate in the chromatographic domain. Severalspectral libraries have incorporated retention indices such asthe National Institute of Standards and Technologies (21) andMetabolomics FiehnLib (22).Volatiles are a specialized class ofmetabolites that contribute

to vegetable and fruit aromas and plant defense responses. Cur-rent metabolomics technologies used to study volatiles centeron GC-MS coupled to headspace solid-phase microextraction(HS-SPME) or other sorbent-based sampling techniques. HS-SPME is a sensitive and robust technique. The effectiveness ofvarious commercial HS-SPME fibers has recently been evalu-ated for the analysis of fruit volatiles and led to the identifica-tion of 14 novel volatiles (23). GC-MS-based analyses have alsobeen used to identify volatile tomato repellents against whitefly(24) and to profile volatiles from various plant species, includ-ing tomato (25) and grape (26).GC requires volatile and thermally stable analytes such as

those described above. However, relatively few compoundsmeet this requirement in their native state (e.g. short chain alco-hols, acid, esters, and hydrocarbons). Many other compoundscan be analyzed by GC only following derivatization, i.e. alkyla-tion and silylation (27, 28). An in-liner derivatization methodfor ultra-small sample volumes (2 �l down to 10 nl) wasrecently developed and used to profile the intracellular contentof frog oocytes (29). Although derivatization is often necessaryin GC analyses, it does introduce variability and producederivatization artifacts.GC has often been coupled to single quadrupole MS detec-

tors, which have the advantages of high sensitivity and gooddynamic range but suffer from slower scan rates and lowermassaccuracy relative to TOF-MS detectors. However, the availabil-ity, reliability, effectiveness, and affordable cost of GC-quadru-pole MS analyzers have made them a popular and robustmetabolomics platform.Othermass analyzers such asTOF-MSand triple quadrupoles have also been interfaced to GC. GC-triple quadrupole MS/MS is capable of multiple reaction mon-itoring of analytes, which can overcome the challenging identi-fication and quantification problems associatedwith co-elutinganalytes in complex matrices. It has been employed to detectmultiple pesticide classes in various fruits and vegetables (30–32); to profile sugar in olive fruits, leaves, and stems (33); toreveal responses of cell cultures to external stresses (27); and todetect fatty acid amides in human plasma (34).GC-TOF-MS technology offers high mass resolution, high

mass accuracy, and fast scan speeds. The relatively faster scanrates associated with TOF-MS are extremely useful for theaccurate deconvolution of overlapping high resolution or ultra-fast GC peaks such as those encountered during complex met-abolic mixture analyses. Recent application of GC-TOF-MS inmetabolomics includes large-scale metabolite profiling ofhuman serum (28) and plant samples (35–37).

MINIREVIEW: MS in Metabolomics

25436 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 286 • NUMBER 29 • JULY 22, 2011


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

A unique innovation was the development of GC�GC,which offers dramatically increased separation efficiencies andpeak capacities (38). In GC�GC, two capillary columns of dif-ferent stationary-phase selectivity are coupled in series througha flow modulator. Effluents from the first column (usually along non-polar column) are captured and transferred by themodulator onto the second column. The second column is nor-mally a short polar or semi-polar column that quickly separatesthe effluent within seconds before the next effluent enters thecolumn. The sharp and narrow peaks generated in fast GC orGC�GC require the use of fast scanning analyzers such asTOF-MS (i.e. approximately�100Hz) or “semi-fast” scan qua-drupoles (i.e.�20Hz) (39). Current GC�GC-TOF-MS analyz-ers can operate with very high acquisition rates, typically up to500 Hz, and offer higher resolution and sensitivity. Recentmetabolomics applications of GC�GC-MS include animals(40), plants (41–44), microorganisms (45), and other samplessuch as human serum and tissues (46–48). Fig. 1 (reprintedfrom Ref. 41 with permission) shows a recent application ofGC�GC-TOF-MS to resolve plant terpenoids.GC-MS is limited to volatile, thermally stable, and energeti-

cally stable compounds. Unfortunately, it is less amenable tolarge highly polarmetabolites due to their poor volatility. Chro-matographic analyses of these compounds usually rely on otherchromatographic techniques such as LC and CE.LC-MS—LC-MS is an important tool in metabolomics and

can be tailored for targeted or non-targeted metabolomics.

Although both normal-phase (NP) and reversed-phase (RP)columns have been employed in metabolomics, RP columnssuch as C18 and C8 are by far the most utilized. However, NPseparations provide complementary views of the metabolome,as demonstrated in metabolic profiling of urine using hydro-philic interaction liquid chromatography (HILIC)-MS and RP-HPLC-MS (49). HILIC is ideal for highly polar and ionic com-pounds and therefore suitable for samples that containpredominantly polar metabolites such as urine. LC-MS usingconventional C18 columns with particle sizes of 3–5 �m hasbeen widely used in metabolomics elucidation of plant second-ary metabolism (27, 50–52). Many established lipidomics pro-grams also rely upon LC-MS for the large-scale study of cellularlipids (53, 54), which has improved our understanding of lipidmetabolism, signaling, and neurodegenerative disorders suchas Alzheimer disease (55–57).The development of fast and more efficient ultra-HPLC

(UHPLC), which utilizes higher pressures (12,000–15,000 p.s.i.compared with �6000 p.s.i. for HPLC) and sub-2-�m packingparticles, has substantially increased chromatographic resolu-tion and peak capacity compared with HPLC. Fig. 2 shows aUHPLC-QTOF-MS base peak chromatogram of a highly com-plex metabolic plant mixture. The superiority of UHPLC hasalso been demonstrated by Nordstrom et al. (58), who reportedthat UPLC-MS resulted in a �20% increase in detectable com-ponents compared with a similar HPLC-MS-based approach.

FIGURE 1. Contour plot of two-dimensional GC-TOF-MS analysis of an n-hexane extract of transgenic Artemisia annua L. A, total ion chromatogram (TIC)plot (first dimension, x axis; second dimension, y axis; time unit, second). B, expanded view of the boxed area in A. Peaks that are numbered in both panels areterpenoids. This figure has been reprinted from Ref. 41 with permission.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

More recently, HILIC-UHPLC-MS has been introduced andused in urinary metabolic profiling (59).LC�LC utilizing two columns of different selectivity offers

an effective platform for separating both polar and non-polarcompounds simultaneously. The peak capacity of two-dimen-sional HPLC is much higher and is the product of the two inde-pendent dimensional peak capacities given that the first andsecond separations are truly orthogonal. The higher peakcapacities offer greater metabolome coverage. Fig. 3 (reprintedfrom Ref. 60 with permission) illustrates NP�RP and RP�RPtwo-dimensional HPLC separations, with the NP�RP systemachieving an overall peak capacity of 1095 when applied to theanalysis of a lemon oil extract (60). LC�LC is typically superiorto one-dimensional HPLC even if the columns used in two-dimensional LC are of similar chemistry (61). LC�LC-MS hasbeen employed for the analyses of carotenoids in differentorange juices (62), drugmetabolism (63), and triacylglycerols invegetable oil (64). LC�LC-MS has also been tailored for targetanalyses. For example, Aturki et al. (65) used an RP C18 columnin the first dimension to separate flavanone-7-O-glycosidesfrom a complex sample and then used a carboxymethylated�-cyclodextrin-based column in the second dimension toresolve the individual flavanone-7-O-glycoside stereoisomer.The disadvantages of LC�LC are its relatively complex setupand the loss of sensitivity due to a sample dilution effect in thesecond dimension (66).

The ionization technique selected for LC-MS-basedmetabo-lomics can also have a substantial impact on metabolite pro-files. Generally, ESI is ideal for semi-polar and polar com-pounds, whereas APCI is more suitable for neutral or less polarcompounds. These two ionization techniques provide comple-mentary data; both are desirable in large-scale non-targetedmetabolomics. For example, complementary ESI-APCI analy-ses resulted in an�20% increase in the number of detected ionsin a human blood serum extract (67). Multi-ionizationmode ordual ESI-APCI ion sources are now commercially availablefrom some instrument vendors, allowing for simultaneousacquisition of ESI and APCI data.ManymodernMS instruments are now capable of fast polar-

ity switching during data acquisition and have been exploited insimultaneous acquisition of both positive and negative ionmode data (68, 69). The use of both positive and negative ioni-zation LC-MS offers more comprehensive metabolome cover-age than the use of a single polarity (50, 67). Fig. 4 (reprintedfrom Ref. 50 with permission) shows the positive and nega-tive ion mode HPLC-MS chromatograms of Medicago trun-catula cell extracts. Several analytes were detected only inthe negative ion mode, whereas others were observed only inthe positive ion mode. Similarly, Nordstrom et al. (67) notedthat �90% of the human blood plasma ions observed in thepositive ion ESI mode were not found in the negative ionmode and vice versa.

FIGURE 2. UHPLC-quadrupole-TOF-MS base peak ion chromatogram of combined methanol extracts from soybean and M. truncatula (A17). Theseparation was performed on a Waters ACQUITY system using a Waters ACQUITY UPLC C18 column (2.1 � 100 mm) with 1.7-�m particles at a flow rate of 600�l/min and a linear gradient of 0.1% acetic acid/acetonitrile (5:95 to 30:70 over 30 min). Mass spectra were recorded on a Waters Q-Tof Premier massspectrometer.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

It is expected that LC-MS will continue to play an importantrole in MS-based metabolomics, and with the continuousadvancement of LC and MS technologies, both the sensitivityand depth of coverage of LC-MS-basedmetabolomics will con-tinue to improve. As demonstrated in a recent human serummetabolome study, higher metabolome coverage can best beachieved by using multiple metabolomics technologies (70).

Using a combination of platforms, including LC-MS/MS,GC-MS, TLC-GC-flame ionization detection, direct infusionMS/MS, and NMR, 4229 compounds were tentatively identi-fied fromhuman serum,with each platform identifying a subsetof unique compounds (70).CE-MS—CE separates analytes based on charge and size, and

it is particularly suitable for the analysis of highly polar and

FIGURE 3. Two-dimensional contour plots for a test mixture using NP�RP two-dimensional LC (A) and a steroid mixture using RP�RP two-dimen-sional LC (B). Compounds that were poorly separated in the first dimension NP separation (x axis) such as aromatic ethers (peaks 1–3), aromatic esters (peaks4 – 6), phenones (peaks 7–10), and steroids (peaks 19 –28) were clearly separated in the second dimension RP separation (y axis) (A). Despite its apparent lowerorthogonality, RP�RP two-dimensional LC still offers higher separation efficiencies relative to one-dimensional LC (B). When two RP columns (Cyano and C18)were used, steroids were separated, which could not be achieved using either Cyano or C18 one-dimensional LC (B). This figure has been reprinted from Ref. 60with permission.

FIGURE 4. HPLC-positive ion ESI-MS chromatogram (A), with inset showing a full positive ion MS spectrum of peak 15 (afrormosin �-D-glucoside), andHPLC-negative ion ESI-MS chromatogram (B) of M. truncatula cell extracts. Some peaks such as peaks 4, 30, 31, and 32 were detected only in negative ionmode (B) but not in positive ion mode (A). In contrast, peaks 3, 13, and 28 were observed in positive ion mode only (A). This figure has been reprinted from ourprevious work (50) with permission.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

ionic metabolites. Separation of neutral compounds can beachieved using micellar electrokinetic chromatography, whichemploys charged surfactants such as SDS to form chargedmicelles containing the analyte. CE is a fast, relatively inexpen-sive, and highly efficient separation technique. Capillary zoneelectrophoresis (CZE) is themost utilized separation techniquein CE-MS-based metabolomics because many compounds canbe separated readily by CZE. CE can be interfaced with variousMS analyzers; however, TOF-MS is the most commonly usedCE-MS analyzer due to its fast acquisition rates, which are nec-essary to statistically sample the narrow CZE peaks. ESI is theionization technique of choice for CE-MS.CE-MS has been used in both targeted and non-targeted

metabolomics. CE-TOF-MS was used for global profiling ofendogenous metabolites in tumor and normal tissues, and theresults revealed elevated glycolysis in tumor tissues evidencedby extremely low glucose, high lactate, and high glycolyticintermediates in tumor tissues (71). CE-TOF-MS metabolicprofiling of Illicium anisatum seed, pulp, stem, and leaf tissuesdetected �1000 tentative polar metabolites in 40 min andrevealed spatial distributions of numerous metabolites (72).Use of capillary electrochromatography (CEC) coupled to MSfor metabolomics has been reported as well. CEC achieves sep-arations using an electrostatic field imposed on a packed parti-cle or a monolithic column, which allows for the high separa-tion efficiency of CE with the high selectivity of the stationaryphase of LC columns. For example, CEC-ESI-MSmethods havebeen developed and used to analyze the metabolome of ahuman hepatocellular carcinoma cell line (73) and drug abusethrough urine analysis (74).Unfortunately, CE-MS has inherent limitations. These are

mainly low sensitivity, poor reproducibility, and electrochemi-cal reactions of metabolites. Recently, the performances ofGC-MS, LC-MS, and CE-MS were compared in quantitativemetabolomics, and it was concluded that CE lacked the neces-sary robustness and was the least suitable platform for analyz-ing complex biological samples (75).

Progressive MS-based Metabolomics Applications andQuantifications

The above noted MS technologies form the core analyticalplatforms used by most in metabolomics. However, metabolo-mics applications and techniques continue to expand. Severalhighlighted areas are discussed below, including spatiallyresolved metabolomics, fluxomics, integrated metabolomics,personalizedmedicine, and computationalmethods formetab-olite annotation.Advanced sampling and MS technologies have made it pos-

sible to perform spatially resolved metabolomics, which can beachieved through the profiling of laser-dissected tissues (76),single cell sampling using microcapillaries (77), or metaboliteimaging MS. Microdissection and single cell sampling usingmicrocapillaries provide information on differential metaboliteaccumulation in specific tissue or cell types, whereas imagingMS enables the spatial visualization of metabolites and theirrelative abundances across various tissues. Spatial metabolo-mics can be used to decipher the functional roles of the metab-

olites based upon their localization and/or co-localization withother metabolites/proteins/transcripts.Flux analyses or fluxomics is another important application

ofmetabolomics (78) and has been used formechanistic studies(79), integrated metabolomics for gene discovery (80) and per-sonalized medicine (81). For example, analysis of flux throughthe TCA cycle in a human cancer cell line confirmed glutami-nolysis and reductive carboxylation as the major cancer cellpathways that provide nitrogen for amino acid and nucleotide(adenine) syntheses (78). Differential regulation of the samemetabolic pathways in response to different elicitors in plantcells was identified through large-scalemetabolic profiling (79).Combined with transcriptomics, metabolomics was success-fully used to identify two ArabidopsisMyb transcription factorgenes that regulate aliphatic glucosinolate biosynthesis (80).The emerging MS-based omics technologies are providing asystems approach to disease and are transforming medicinefrom reactive to proactive (i.e. predictive, personalized, preven-tive, and participatory or “4P”) (82).In general, a significant proportion of profiled metabolites

remain unannotated, but several groups are creating largeMS/MS spectral libraries as part of an advanced scheme tocharacterize and cross-correlate both known and unknownmetabolites. For example,Matsuda et al. (83) have generated anMS2T library that contains nearly 1.6 million MS2T spectra tofacilitate peak annotation in non-targeted plant metabolomicsstudies. Annotation of metabolites through MS or MS/MSspectra typically involves spectral matching against spectrallibraries compiled with authentic standards. Identification ofmetabolites whose MS/MS spectra are not present in the spec-tral libraries remains very challenging and is being addressedthrough both empirical and computational MS. MetFrag, acomputer program for metabolite identification based upon insilico spectral fragmentation matching, was recently described(84). It is expected that continued development of computa-tional methods such as MetFrag and other fragment tree-re-lated software (85) will further assist metabolite identificationin metabolomics.Quantification in metabolomics is critical in understanding

biological processes, and it is generally performed in two man-ners, i.e. relative or absolute quantification. Relative quantifica-tion, which normalizes themetabolite signal intensity to that ofan internal standard or another relative metabolite, is typicallyused in large-scale non-targeted profiling. Absolute quantifica-tion uses external standards or internal isotopically labeledstandards to determine the absolute metabolite quantity and isused mostly in targeted metabolomics. The major obstacle inmetabolite quantification is that the metabolite’s signal inten-sity is dependent not only on its concentration but also on itschemical structure and matrix. Ion suppression and enhance-ment caused by matrix effects can result in inaccurate quanti-fication of the metabolites. Stable isotope dilution is one solu-tion for absolute quantification (i.e. stable isotope-labeledstandards are added to the samples to account for sample proc-essing variation and matrix effects encountered during MSanalysis). Absolute quantification in metabolic profiling ofobese and lean humans clearly linked branched chain aminoacid-related metabolites to the development of obesity-associ-




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

ated insulin resistance (86). Quantitative MS intermediarymetabolite profiling and 13C NMR-based flux analyses identi-fied a critical link between pyruvate and the TCA cycle medi-ated by pyruvate carboxylase, which plays an important role inregulating glucose-stimulated insulin secretion (87, 88). Quan-titative lipidomics revealed that abnormal lipid profiles occureven at the very earliest stages of diabetes (89). Stable isotopedilution GC-MS of cholesterol in human plasma revealed thatoxysterols are sensitive and specific biomarkers for Niemann-Pick type C1 disease (90).It is clear that stable isotope standards are critical for abso-

lute quantification in metabolomics. However, the availabilityof commercial isotope-labeled standards is limited, and thecosts can be prohibitive to large-scale use. Several groups havebegun to chemically synthesize a diverse range of stable iso-tope-labeled compounds for absolute quantification. Theseinclude fatty acids (91), human steroids (92), and plant hor-mones (93), but this approach is laborious and requires syn-thetic expertise not available in every metabolomics group. Inaddition, preparing multiple isotope-labeled standards forlarge-scale metabolomics is challenging given the diverse andcomplex structures of the metabolites. One way to circumventthis problem is to introduce a stable isotope tag to the metabo-lites through chemical labeling. A dimethylation method tolabel amine-containing metabolites using commercially avail-able and inexpensive 13C- or deuterium-labeled formaldehydehas been reported (94). Standards of known concentrations arelabeled in the same manner and used to generate calibrationcurves for absolute quantification.

Conclusions and Future Challenges

MS has become an indispensable productive tool in metabo-lomics due to its high sensitivity and selectivity. However, thereare still many metabolomics challenges, including limiteddynamic range, lack of comprehensive coverage, and limitedmetabolite annotations. Currently, the best dynamic rangeof modern MS is �106, which is significantly lower than theestimated concentration range of cellularmetabolites of 1012 ormore. In addition, the estimated number of metabolites withina given plant species can be 10,000 or more, and the currentmetabolome depth of coverage is approximately �20%. Thus,there is substantial need for improvement. Advancements andsolutions to the above limitations will continue to expand thescope and propel the utility of MS in metabolomics.

REFERENCES1. Sumner, L. W., Mendes, P., and Dixon, R. A. (2003) Phytochemistry 62,

817–8362. Bedair, M., and Sumner, L. W. (2008) Trends Anal. Chem. 27, 238–2503. Takats, Z., Wiseman, J. M., and Cooks, R. G. (2005) J. Mass Spectrom. 40,

1261–12754. Chen, H., Venter, A., and Cooks, R. G. (2006) Chem. Commun. 19,

2042–20445. Castrillo, J. I., Hayes, A., Mohammed, S., Gaskell, S. J., and Oliver, S. G.

(2003) Phytochemistry 62, 929–9376. Hughey, C. A., Rodgers, R. P., and Marshall, A. G. (2002) Anal. Chem. 74,

4145–41497. Giavalisco, P., Hummel, J., Lisec, J., Inostroza, A. C., Catchpole, G., and

Willmitzer, L. (2008) Anal. Chem. 80, 9417–94258. Makarov, A. (2000) Anal. Chem. 72, 1156–1162

9. Yang, K., Zhao, Z., Gross, R. W., and Han, X. (2007) PLoS ONE 2, e136810. Ejsing, C. S., Moehring, T., Bahr, U., Duchoslav, E., Karas, M., Simons, K.,

and Shevchenko, A. (2006) J. Mass Spectrom. 41, 372–38911. Allegrand, J., Touboul, D., Schmitz-Afonso, I., Guerineau, V., Giuliani, A.,

Le Ven, J., Champy, P., and Laprevote, O. (2010) Rapid Commun. MassSpectrom. 24, 3602–3608

12. Cooks, R. G., Ouyang, Z., Takats, Z., and Wiseman, J. M. (2006) Science311, 1566–1570

13. Lane, A. L., Nyadong, L., Galhena, A. S., Shearer, T. L., Stout, E. P., Parry,R. M., Kwasnik, M., Wang, M. D., Hay, M. E., Fernandez, F. M., andKubanek, J. (2009) Proc. Natl. Acad. Sci. U.S.A. 106, 7314–7319

14. Girod, M., Shi, Y., Cheng, J. X., and Cooks, R. G. (2010) J. Am. Soc. MassSpectrom. 21, 1177–1189

15. Cody, R. B., Laramee, J. A., and Durst, H. D. (2005) Anal. Chem. 77,2297–2302

16. Navare, A. T., Mayoral, J. G., Nouzova, M., Noriega, F. G., and Fernandez,F. M. (2010) Anal. Bioanal. Chem. 398, 3005–3013

17. Zhou, M., McDonald, J. F., and Fernandez, F. M. (2010) J. Am. Soc. MassSpectrom. 21, 68–75

18. Edison, S. E., Lin, L. A., Gamble, B. M., Wong, J., and Zhang, K. (2011)Rapid Commun. Mass Spectrom. 25, 127–139

19. Law,W. S., Chen, H.W., Balabin, R., Berchtold, C., Meier, L., and Zenobi,R. (2010) Analyst 135, 773–778

20. Zhu, L., Hu, Z., Gamez, G., Law, W. S., Chen, H., Yang, S., Chingin, K.,Balabin, R. M., Wang, R., Zhang, T., and Zenobi, R. (2010) Anal. Bioanal.Chem. 398, 405–413

21. Babushok, V. I., Linstrom, P. J., Reed, J. J., Zenkevich, I. G., Brown, R. L.,Mallard, W. G., and Stein, S. E. (2007) J. Chromatogr. A 1157, 414–421

22. Kind, T., Wohlgemuth, G., Lee do, Y., Lu, Y., Palazoglu, M., Shahbaz, S.,and Fiehn, O. (2009) Anal. Chem. 81, 10038–10048

23. Pereira, J., Pereira, J., and Camara, J. S. (2011) Talanta 83, 899–90624. Bleeker, P. M., Diergaarde, P. J., Ament, K., Guerra, J., Weidner, M.,

Schutz, S., de Both,M. T., Haring,M. A., and Schuurink, R. C. (2009) PlantPhysiol. 151, 925–935

25. Mayer, F., Takeoka, G. R., Buttery, R. G., Whitehand, L. C., Naim, M., andRabinowitch, H. D. (2008) J. Agric. Food Chem. 56, 3749–3757

26. Martin, D. M., Toub, O., Chiang, A., Lo, B. C., Ohse, S., Lund, S. T., andBohlmann, J. (2009) Proc. Natl. Acad. Sci. U.S.A. 106, 7245–7250

27. Broeckling, C. D., Huhman, D. V., Farag, M. A., Smith, J. T., May, G. D.,Mendes, P., Dixon, R. A., and Sumner, L. W. (2005) J. Exp. Bot. 56,323–336

28. Begley, P., Francis-McIntyre, S., Dunn,W. B., Broadhurst, D. I., Halsall, A.,Tseng, A., Knowles, J., Goodacre, R., andKell, D. B. (2009)Anal. Chem. 81,7038–7046

29. Koek, M. M., Bakels, F., Engel, W., van den Maagdenberg, A., Ferrari,M. D., Coulier, L., and Hankemeier, T. (2010) Anal. Chem. 82, 156–162

30. Bolanos, P. P., Moreno, J. L., Shtereva, D. D., Frenich, A. G., and Vidal, J. L.(2007) Rapid Commun. Mass Spectrom. 21, 2282–2294

31. Walorczyk, S. (2008) Rapid Commun. Mass Spectrom. 22, 3791–380132. Wong, J. W., Zhang, K., Tech, K., Hayward, D. G., Makovi, C. M.,

Krynitsky, A. J., Schenck, F. J., Banerjee, K., Dasgupta, S., and Brown, D.(2010) J. Agric. Food Chem. 58, 5868–5883

33. Gomez-Gozalez, S., Ruiz-Jimenez, J., Priego-Capote, F., and Luque deCastro, M. D. (2010) J. Agric. Food Chem. 58, 12292–12299

34. Zoerner, A. A., Gutzki, F. M., Suchy, M. T., Beckmann, B., Engeli, S.,Jordan, J., and Tsikas, D. (2009) J. Chromatogr. B 877, 2909–2923

35. Allwood, J. W., Erban, A., de Koning, S., Dunn, W. B., Luedemann, A.,Lommen, A., Kay, L., Loscher, R., Kopka, J., and Goodacre, R. (2009)Metabolomics 5, 479–496

36. Fukushima, A., Kusano, M., Redestig, H., Arita, M., and Saito, K. (2011)BMC Syst. Biol. 5, 1

37. Skogerson, K., Harrigan, G. G., Reynolds, T. L., Halls, S. C., Ruebelt, M.,Iandolino, A., Pandravada, A., Glenn, K. C., and Fiehn, O. (2010) J. Agric.Food Chem. 58, 3600–3610

38. Mondello, L., Tranchida, P. Q., Dugo, P., and Dugo, G. (2008)Mass Spec-trom. Rev. 27, 101–124

39. Adahchour, M., Brandt, M., Baier, H. U., Vreuls, R. J., Batenburg, A. M.,and Brinkman, U. A. (2005) J. Chromatogr. A 1067, 245–254




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

40. Welthagen,W., Shellie, R., Spranger, J., Ristow,M., Zimmermann, R., andFiehn, O. (2005)Metabolomics 1, 65–73

41. Ma, C., Wang, H., Lu, X., Wang, H., Xu, G., and Liu, B. (2009)Metabolo-mics 5, 497–506

42. Kusano, M., Fukushima, A., Kobayashi, M., Hayashi, N., Jonsson, P.,Moritz, T., Ebana, K., and Saito, K. (2007) J. Chromatogr. B 855, 71–79

43. Hope, J. L., Prazen, B. J., Nilsson, E. J., Lidstrom, M. E., and Synovec, R. E.(2005) Talanta 65, 380–388

44. Zhao, T., Krokene, P., Bjorklund, N., Långstrom, B., Solheim, H., Chris-tiansen, E., and Borg-Karlson, A. K. (2010) Phytochemistry 71, 1332–1341

45. Purcaro, G., Tranchida, P. Q., Dugo, P., La Camera, E., Bisignano, G.,Conte, L., and Mondello, L. (2010) J. Sep. Sci. 33, 2334–2340

46. Huang, X., and Regnier, F. E. (2007) Anal. Chem. 80, 107–11447. O’Hagan, S., Dunn, W. B., Knowles, J. D., Broadhurst, D., Williams, R.,

Ashworth, J. J., Cameron, M., and Kell, D. B. (2007) Anal. Chem. 79,464–476

48. Li, X., Xu, Z., Lu, X., Yang, X., Yin, P., Kong, H., Yu, Y., and Xu, G. (2009)Anal. Chim. Acta 633, 257–262

49. Cubbon, S., Bradbury, T., Wilson, J., and Thomas-Oates, J. (2007) Anal.Chem. 79, 8911–8918

50. Farag, M. A., Huhman, D. V., Lei, Z., and Sumner, L. W. (2007) Phyto-chemistry 68, 342–354

51. Huhman, D. V., Berhow, M. A., and Sumner, L. W. (2005) J. Agric. FoodChem. 53, 1914–1920

52. Huhman, D. V., and Sumner, L. W. (2002) Phytochemistry 59, 347–36053. Yang, K., Cheng, H., Gross, R. W., and Han, X. (2009) Anal. Chem. 81,

4356–436854. Gross, R. W., and Han, X. (2009) Am. J. Physiol. Endocrinol. Metab. 297,

E297–E30355. Zeng, Y., Cheng, H., Jiang, X., and Han, X. (2008) Biochem. J. 410, 81–9256. Sanchez-Mejia, R. O., Newman, J. W., Toh, S., Yu, G. Q., Zhou, Y., Hala-

bisky, B., Cisse, M., Scearce-Levie, K., Cheng, I. H., Gan, L., Palop, J. J.,Bonventre, J. V., and Mucke, L. (2008) Nat. Neurosci. 11, 1311–1318

57. Mattila, I., Seppanen-Laakso, T., Suortti, T., and Oresic, M. (2008)Meth-ods Mol. Biol. 456, 123–130

58. Nordstrom, A., O’Maille, G., Qin, C., and Siuzdak, G. (2006) Anal. Chem.78, 3289–3295

59. Spagou, K., Wilson, I. D., Masson, P., Theodoridis, G., Raikos, N., Coen,M., Holmes, E., Lindon, J. C., Plumb, R. S., Nicholson, J. K., andWant, E. J.(2011) Anal. Chem. 83, 382–390

60. Francois, I., de Villiers, A., Tienpont, B., David, F., and Sandra, P. (2008)J. Chromatogr. A 1178, 33–42

61. Stoll, D. R., Wang, X., and Carr, P. W. (2008) Anal. Chem. 80, 268–27862. Dugo, P., Giuffrida, D., Herrero, M., Donato, P., andMondello, L. (2009) J.

Sep. Sci. 32, 973–98063. Thomas, A., Deglon, J., Steimer, T., Mangin, P., Daali, Y., and Staub, C.

(2010) J. Sep. Sci. 33, 873–87964. Dugo, P., Kumm, T., Crupi, M. L., Cotroneo, A., and Mondello, L. (2006)

J. Chromatogr. A 1112, 269–27565. Aturki, Z., Brandi, V., and Sinibaldi, M. (2004) J. Agric. Food Chem. 52,

5303–530866. Mondello, L., Herrero, M., Kumm, T., Dugo, P., Cortes, H., and Dugo, G.

(2008) Anal. Chem. 80, 5418–542467. Nordstrom, A., Want, E., Northen, T., Lehtio, J., and Siuzdak, G. (2008)

Anal. Chem. 80, 421–42968. Tolonen, A., and Uusitalo, J. (2004) Rapid Commun. Mass Spectrom. 18,

3113–312269. Cai, F., Xu, W., Wei, H., Sun, L., Gao, S., Yang, Q., Feng, J., Zhang, F., and

Chen, W. (2010) J. Chromatogr. B 878, 1845–185470. Psychogios, N., Hau, D. D., Peng, J., Guo, A. C., Mandal, R., Bouatra, S.,

Sinelnikov, I., Krishnamurthy, R., Eisner, R., Gautam, B., Young, N., Xia, J.,Knox, C., Dong, E., Huang, P., Hollander, Z., Pedersen, T. L., Smith, S. R.,Bamforth, F., Greiner, R., McManus, B., Newman, J. W., Goodfriend, T.,and Wishart, D. S. (2011) PLoS ONE 6, e16957

71. Hirayama, A., Kami, K., Sugimoto, M., Sugawara, M., Toki, N., Onozuka,H., Kinoshita, T., Saito, N., Ochiai, A., Tomita, M., Esumi, H., and Soga, T.(2009) Cancer Res. 69, 4918–4925

72. Urakami, K., Zangiacomi, V., Yamaguchi, K., and Kusuhara, M. (2010)Biomed. Res. 31, 161–163

73. Kato, M., Onda, Y., Sekimoto, M., Degawa, M., and Toyo’oka, T. (2009)J. Chromatogr. A 1216, 8277–8282

74. Aturki, Z., D’Orazio, G., Rocco, A., Bortolotti, F., Gottardo, R., Tagliaro, F.,and Fanali, S. (2010) Electrophoresis 31, 1256–1263

75. Buscher, J. M., Czernik, D., Ewald, J. C., Sauer, U., and Zamboni, N. (2009)Anal. Chem. 81, 2135–2143

76. Schad, M., Mungur, R., Fiehn, O., and Kehr, J. (2005) Plant Methods 1, 277. Ebert, B., Zoller, D., Erban, A., Fehrle, I., Hartmann, J., Niehl, A., Kopka, J.,

and Fisahn, J. (2010) J. Exp. Bot. 61, 1321–133578. Hiller, K., Metallo, C. M., Kelleher, J. K., and Stephanopoulos, G. (2010)

Anal. Chem. 82, 6621–662879. Farag,M. A., Huhman, D. V., Dixon, R. A., and Sumner, L.W. (2008) Plant

Physiol. 146, 387–40280. Hirai, M. Y., Sugiyama, K., Sawada, Y., Tohge, T., Obayashi, T., Suzuki, A.,

Araki, R., Sakurai, N., Suzuki, H., Aoki, K., Goda, H., Nishizawa, O. I.,Shibata, D., and Saito, K. (2007) Proc. Natl. Acad. Sci. U.S.A. 104,6478–6483

81. Weston, A. D., and Hood, L. (2004) J. Proteome Res. 3, 179–19682. Hood, L., and Friend, S. H. (2011) Nat. Rev. Clin. Oncol. 8, 184–18783. Matsuda, F., Yonekura-Sakakibara, K., Niida, R., Kuromori, T., Shinozaki,

K., and Saito, K. (2009) Plant J. 57, 555–57784. Wolf, S., Schmidt, S., Muller-Hannemann, M., and Neumann, S. (2010)

BMC Bioinformatics 11, 14885. Mistrik, R. (October 14, 2004) U.S. Patent 10/967,01886. Newgard, C. B., An, J., Bain, J. R., Muehlbauer, M. J., Stevens, R. D., Lien,

L. F., Haqq, A.M., Shah, S. H., Arlotto,M., Slentz, C. A., Rochon, J., Gallup,D., Ilkayeva, O.,Wenner, B. R., Yancy,W. S., Jr., Eisenson, H.,Musante, G.,Surwit, R. S., Millington, D. S., Butler, M. D., and Svetkey, L. P. (2009) CellMetabolism 9, 311–326

87. Lu, D., Mulder, H., Zhao, P., Burgess, S. C., Jensen, M. V., Kamzolova, S.,Newgard, C. B., and Sherry, A. D. (2002) Proc. Natl. Acad. Sci. U.S.A. 99,2708–2713

88. Ronnebaum, S. M., Ilkayeva, O., Burgess, S. C., Joseph, J. W., Lu, D., Ste-vens, R. D., Becker, T. C., Sherry, A. D., Newgard, C. B., and Jensen, M. V.(2006) J. Biol. Chem. 281, 30593–30602

89. Han, X., Yang, J., Yang, K., Zhao, Z., Abendschein, D. R., and Gross, R. W.(2007) Biochemistry 46, 6417–6428

90. Porter, F. D., Scherrer, D. E., Lanier, M. H., Langmade, S. J., Molugu, V.,Gale, S. E., Olzeski, D., Sidhu, R., Dietzen, D. J., Fu, R., Wassif, C. A.,Yanjanin, N. M., Marso, S. P., House, J., Vite, C., Schaffer, J. E., and Ory,D. S. (2010) Sci. Transl. Med. 2, 56ra81

91. Mesaros, C., Lee, S. H., and Blair, I. A. (2010) Rapid Commun. Mass Spec-trom. 24, 3237–3247

92. Wudy, S. A., Hartmann, M., and Homoki, J. (2002) Steroids 67, 851–85793. Chiwocha, S. D., Abrams, S. R., Ambrose, S. J., Cutler, A. J., Loewen, M.,

Ross, A. R., and Kermode, A. R. (2003) Plant J. 35, 405–41794. Guo, K., Ji, C., and Li, L. (2007) Anal. Chem. 79, 8631–8638




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

Zhentian Lei, David V. Huhman and Lloyd W. SumnerMass Spectrometry Strategies in Metabolomics

doi: 10.1074/jbc.R111.238691 originally published online June 1, 20112011, 286:25435-25442.J. Biol. Chem.

10.1074/jbc.R111.238691Access the most updated version of this article at doi:

Alerts:

When a correction for this article is posted•

When this article is cited•

to choose from all of JBC's e-mail alertsClick here

http://www.jbc.org/content/286/29/25435.full.html#ref-list-1

This article cites 92 references, 10 of which can be accessed free at


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/lookup/doi/10.1074/jbc.R111.238691

http://www.jbc.org/cgi/alerts?alertType=citedby&addAlert=cited_by&cited_by_criteria_resid=jbc;286/29/25435&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/286/29/25435

http://www.jbc.org/cgi/alerts?alertType=correction&addAlert=correction&correction_criteria_value=286/29/25435&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/286/29/25435

http://www.jbc.org/cgi/alerts/etoc

http://www.jbc.org/content/286/29/25435.full.html#ref-list-1

http://www.jbc.org/

massspectrometrystrategiesin metabolomicsaccepted as the gold standard in metabolite structural...

Documents