B RIEFINGS IN FUNC TIONAL GENOMICS . VOL 9. NO 2.

139^148

doi:10.1093/bfgp/elp053

Plant metabolomicsmeeting the analytical challenges of comprehensive metabolite analysis

Adrian D. Hegeman
Advance Access publication date 11 January 2010

Abstract
Plant metabolomics has benefited from a rich array of pre-existing methodological approaches and bioanalytical knowledge for the characterization of the many chemically diverse classes of metabolites. While the field has pushed the implementation of unbiased and generally applicable strategies for metabolite extraction, fractionation and detection, significant challenges in fundamental activities such as compound identification and quantification still exist. This review provides an introduction to metabolomics terminology, methods and resources and discusses analytical limitations. Progress in the application of stable isotope dilution and stable isotopic metabolic labeling to metabolite identification and quantification are discussed. New strategies to address spatial distribution of metabolites via mass spectrometry-based imaging are also reviewed. Keywords: metabolomics; plant; stable isotope dilution; mass spectrometry; methodology

INTRODUCTION
Over the last decade a rich array of analytical strategies has been created for the high throughput analysis of plant metabolites. These strategies are typically referred to under the broader mantel of metabolomics and are often accompanied by terms that are misused or poorly defined to distinguish experimental approaches including: metabolite profiling, metabonomics, metabolic fingerprinting, or metabolic footprinting [1]. Metabolome is a term conceptualized by extrapolation from genome, and refers to the complete set of metabolites in a biological system often as they correspond to a complete set of genes [2]. This concept has emerged from systems biology, which attempts to measure and model various aspects of the molecular components (genes, transcripts, proteins and metabolites) of a biological system to generate information about the interactions between these molecules that can provide novel insight into the molecular basis for macroscopic properties of the system. Metabolomics is the characterization of the metabolome. While genomes have been sequenced

for many organisms, it is currently not possible to measure the complete metabolome for any but the simplest biological systems. It is difficult to estimate the numbers of metabolites that one might expect to observe even for model systems like Arabidopsis thaliana that have had complete and extensively annotated genomes for close to a decade. The number of compounds in the metabolome of a specific species has yet to be precisely determined, but estimates based on genome size vary from $600 in yeast to over 2000 for humans [3] and 500025 000 for different species of plants [4, 5]. Other approaches using high throughput metabolite analysis focus on a subset of useful information while avoiding the difficulties of comprehensive metabolite characterization. Metabolic fingerprinting uses signals from hundreds to thousands of metabolites for rapid sample classification via statistical analysis [6]. The term metabonomics [7] is sometimes used to refer to metabolic fingerprinting experiments thought its use has been discouraged due to the obvious potential for confusion with metabolomics [1]. Proponents of

Corresponding author. Adrian D. Hegeman, 305 Alderman Hall, 1970 Folwell Avenue, Saint Paul, MN 55108, USA. Tel: (612) 6263650; Fax: (612) 624-4041; Email: hegem007@umn.edu Adrian Hegeman is an Assistant Professor at the Departments of Horticultural Science and Plant Biology, Microbial and Plant Genomics Institute, University of Minnesota, Twin Cities campus. His research interests include the development of new methodology for high throughput plant metabolite analysis.
The Author 2010. Published by Oxford University Press. All rights reserved. For permissions, please email: journals.permissions@oxfordjournals.org

140

Hegeman
(MS) [26] but also infrared (IR) spectroscopy [27]. This review will primarily focus on NMR and MS strategies, which are by far the most commonly employed. MS approaches are coupled to a broad range of different separation/ionization strategies including liquid chromatography (LC-MS) [28, 29], direct infusion (DI-MS) [30], matrixassisted laser desorption ionization (MALDI-MS) [31], gas chromatography (GC-MS) [30, 32], capillary electrophoresis (CE-MS) [33] and multidimensional chromatographies (LC LC or GC GC MS) [34, 35]. While all of these approaches are general and can be applied to broad classes of metabolites, each one has different advantages and disadvantages with regard to the type of information provided, sensitivity and interferences. As a result, application of combinations of these approaches is advantageous [36, 37]. There are numerous examples of metabolomics studies that are successful within current analytical limitations and capabilities. These include the evaluation of substantial equivalence [38] for the safety assessment of genetically modified crops [39, 40]. Metabolomics has been used to assist in genome annotation [41]. New gene functions have been identified by analysis of genetic knockouts in phytoalexin biosynthesis [42], circadian clock function [43] and natural product biosynthesis [44]. Plant metabolic biotic [4547] and abiotic [4850] stress responses have examined [51, 52]. Metabolic engineering projects have used high throughput metabolite analysis to assess global consequences of their modifications [53]. Metabolic profiling of phytomedicinal species has proven a valuable tool for quality control [54] and optimization of yield [5557]. Metabolomics is also a particularly powerful tool for studying environmental and ecological interactions (reviewed in ref. [58]); the approach has been used to find molecular signatures for endophytes [59], for discovering unexpected bioactive compounds in plant herbivore interactions [45], and to discover metabolites induced by fungal infections [60]. Metabolite fingerprints are also being used to identify compounds associated with specific quality traits for rapid screening in plant breeding programs [61, 62]. Despite the fact that genomic information cannot be used as a constraint for metabolite analysis, metabolomics can be applied effectively to the analysis of species lacking sequenced genomes. Thus species that are not genome enabled may be examined using metabolomics approaches

metabonomics sometimes use the term broadly adding to the general confusion [8]. This approach does not typically attempt to identify the metabolites responsible for the signals but will use the signals to differentiate between populations based on some characteristic such as disease status or a quality trait. Often signals that are dominant classifiers will be identified to provide additional information about the molecular basis for the classified characteristic. Metabolite footprinting is a similar approach that examines the signals associated with an organisms growth in a substrate or medium by analysis of chemical species in the substrate [9, 10]. Lastly, metabolite profiling attempts to identify and quantify a specific class or classes of chemically related metabolites that often share chemical properties that facilitate simultaneous analysis [11]. Metabolite profiling experiments have been performed for many decades, but have benefited recently from the increased interest in metabolomics for more effective metabolite analysis. Plant metabolomics topics, especially its application to functional genomics have been extensively reviewed [1, 1214]. Other more specific topics have been reviewed including: metabolomics in systems biology [15]; and the application of metabolomics to crop plant improvement [16]; genetically modified crop plant evaluation [17]; human nutrition [18] and agriculture [19]. A robust set of data collection, curation and reporting standards have been (and are being) developed by the community [20, 21]. The analytical considerations of plant metabolomics have also been previously reviewed [22, 23]. High throughput approaches for the characterization of biomolecules including DNA sequencing, nucleic acid microarrays, and tandem mass spectrometry-based protein identification were required before genomic, transcriptomic and proteomic experiments could be conceived. While these omics disciplines were enabled by specific technical advances, metabolomics has exhibited far less dependence on any one technological breakthrough. This is partly because the vastly greater variety of chemical properties encompassed by metabolite pools compared to nucleic acids or proteins, but also because metabolites are not biopolymers like DNA, RNA and proteins linked by the genetic code and thus have not benefited from the increasing availability of genome sequence data. Instead, metabolomics has been approached using a variety of fairly mature analytical tools predominantly nuclear magnetic resonance (NMR) [24, 25] and mass spectrometry

Analytical challenges of comprehensive metabolite analysis

where transcriptomics likely fail. and proteomics would

141

ANALYTICAL CHALLENGES IN PLANT METABOLOMICS

The long-range goal of metabolomics might reasonably be defined as the characterization of an organisms complete complement of metabolites. In examining the analytical challenges involved in metabolite characterization (in contrast to genome sequencing) one begins to appreciate the present practical impossibility of comprehensive metabolomics: (i) the chemical diversity of a typical metabolome is enormous and requires multiple general approaches for extraction, fractionation and analysis to accommodate variation in solubility, reactivity and other chemico-physical properties [63]; (ii) dynamic range of metabolite concentrations can also be huge requiring efficient separation strategies so that minor components are not completely lost among those that are more abundant; (iii) spatial distribution includes specific organs, cell types, sub cellular domains, extracellular spaces and occasionally extends beyond the organism to compounds excreted into the environment [64]; (iv) temporal distribution is similarly large with variations ranging across an organisms lifespan, seasonal and circadian rhythms and faster chemical movements and oscillations [65]; and (v) unlike proteomics and transcriptional studies, genomic information cannot be used as a constraint for the identification of molecular species. Because of these challenges, state of the art metabolomics studies do not begin to approach the ultimate goal of comprehensive identification, quantification and localization of every metabolite. Subsequent sections explore the specifics of these challenges for metabolite extraction, identification, quantification, dynamic range and distribution examining various strategies for extending current methodological capabilities.

should be preserved; and (iii) metabolite solubilization should be reproducible and ideally complete. To stop metabolism extremes of temperature or pH are rapid and effective addition of acid or base or high temperatures result in significant degradation of many metabolite classes. Freezing helps to stabilize labile metabolites but tissue excision and freezing need to be accomplished quickly to avoid induction of wounding responses. Rapid freeze methods include immersion into liquid nitrogen, liquid nitrogen-cooled isopentane [66] and freeze clamping [67]. As freezing generally preserves protein structure steps need to be taken to avoid warming that may allow some enzyme activity to perturb metabolite pool composition. Freeze drying can help stabilize frozen samples, but may result in the loss of volatile metabolites and lyophilized samples can still reabsorb atmospheric water and regain some enzyme activity. One strategy uses polar organic solvent (alcohols or acetonitrile) extraction of frozen samples at low temperatures (72 C) to extract metabolites away from enzymes and cellular debris. Because of the low-temperature extraction is extremely inefficient and requires multiple repetitions. The initial extractions efficiently remove water from the enzymes making it possible to increase the extraction temperature (to 4 C) in subsequent rounds providing better extraction efficiency [68]. Several studies have suggested that 90% methanol/10% water is the best solvent for optimizing the number of extracted metabolites and the reproducibility among multiple extractions [37, 69, 70]. Although isopropanol water mixtures may be comparable in their efficiency while minimizing unwanted esterification side reactions. Other studies have examined optimal extraction protocols for combined LC-MS and GC-MS analysis [71], and for plant metabolites by NMR [72].

Detection, dynamic range and quantification

The analytical approaches commonly used for metabolomics studies are all capable of measuring broad ranges of molecule classes, but each technique also has limitations with regard to detection, sensitivity, dynamic range and quantification. NMR provides a great deal of structural information for virtually any biomolecule and signal intensity is linearly related to analyte abundance for facile quantification, but is not nearly as sensitive as any of the MS approaches [24]. Many different types of multidimensional NMR

Metabolite extraction
Most metabolomics studies (other than those that focus on excreted metabolite populations) must rely on cell/tissue homogenization followed by some form of extraction to remove compounds from cellular debris and place them into a context that is amenable to analysis. One strives to meet three (sometimes conflicting) goals: (i) metabolism must be quenched; (ii) chemical identity of metabolites

142

Hegeman
that two compounds, differing only in mass due to the substitution of one or more stable heavy atom isotopes, can be adequately co-purified and that the isotopic enrichment of the mixture can be determined following isolation [81]. Stable isotope dilution provides an ideal internal control where labeled compounds are added at the earliest possible stages to account for steps that are incomplete or partial, including extraction, derivatization and fractionation. Absolute quantification can be accomplished by addition of a known quantity of labeled standard prior to workup and analysis. Finding sources for labeled compounds can be problematic using this approach for measuring large numbers of compounds. One way to avoid this problem is to use an organisms capacity to incorporate inexpensive labeled nutrients (such as 13CO2 or 15NO3) to label every molecule. This approach, called metabolic labeling (reviewed in ref. [82]), has been used extensively for quantification in proteomics, to enhance NMR sensitivity [83], and is starting to be applied to metabolomics in plant suspension cells [84, 85] and in whole plants [30, 86, 87]. Relative quantification of multiple compounds can be easily accomplished with this approach either by mixing labeled and unlabeled organisms for every experimental sample, or by producing a large quantity of labeled plant material that can be spiked into large numbers of unlabeled experimental samples. Absolute quantification can be performed using a variant of this approach that first employs reverse isotope dilution [88, 89] to find the quantities of specific metabolites the large pool of labeled plant material. This can be accomplished using the more accessible unlabeled metabolite standards and then using that indexed labeled standard to quantify large numbers of unknown unlabeled experimental samples [77].

experiments can be used to resolve overlapping peaks in complex spectra [73]. GC-MS is limited to analytes that are thermally stable and volatile (or can be made volatile through derivatization) but is several orders of magnitude more sensitive than NMR and has a linear signal to concentration relationship. LC-MS is still more sensitive than GC-MS and more comprehensive as there are no volatility requirements and fewer issues with thermal lability. The electrospray ionization (ESI) process is competitive and may not always give a linear signal to concentration curve when multiple compounds are being ionized at the same time. This phenomenon is called ion suppression and it necessitates a high degree of internal control for consistent and accurate quantification of complex mixtures by LC-MS [74, 75]. Performing absolute quantification in MS metabolomics experiments limits the numbers of compounds that one can measure because of the requirement for calibration curves or internal standard curves using samples of standard reference compounds [76, 77]. As a compromise one often adopts semi-quantitative approaches like relative quantification where changes in abundance are measured in lieu of absolute quantities [78, 79]. It is important to note, however, that as purification and derivatization procedures become more complex, the error propagated through multiple steps cumulatively degrades the accuracy of quantitative measurements collected using these strategies. As a result, quantification becomes less accurate as attempts are made to overcome dynamic range limitations and to dig deeper into populations of low abundance metabolites. GC-MS is generally amenable to relative quantification by spectral comparison across carefully controlled and randomized sample sets. For LC-MS, because of ion suppression phenomena, relative quantification is much less reliable [80]. Both LC and GC-MS require a means of ubiquitous sample control through multistep manipulation for accurate quantification.

Metabolite identification
High throughput compound identification continues to be a significant challenge for metabolomics having multiple dimensions of ambiguity in the most commonly used approaches. High confidence identification typically relies on direct comparison to standard reference compounds to confirm the exact match to chromatographic and spectral data. Use of standard compounds is certainly neither high throughput nor in many cases feasible when reference compounds are not available. While publicly accessible data library resources are available for reference

Stable isotopes and metabolic labeling

Isotopic dilution analysis is one of the most effective approaches for bypassing the limitations imposed by the demands of multi-step purification and variations in ionization and detector efficiency. This approach can be used to avoid the confounding effects of ion suppression in ESI during LC-MS analysis of complex samples. Isotope dilution analysis uses the fact

Analytical challenges of comprehensive metabolite analysis

comparison, different techniques vary widely in how useful these comparisons are for compound identification depending on how well the methodology is standardized. GC-MS, for example, can be quite consistent between various instrument types and multiple laboratories for both electron ionization (EI) fragmentation patterns and indexed retention times making searchable library resources quite effective for GC-MS-based compound identification [26]. Similarly, NMR can make effective use of spectral library resources provided that proper sample standardization practices are employed [24]. LC-MS and MS/MS typically have a high degree of variation and a lack suitable standardization to provide spectral information or retention indexing in library data collections suitable for robust compound identification [90]. This is in part due to the wide variation in properties of LC media and the difficulty in selecting broadly optimized LC methods, but also because of the lack of consistent fragmentation across different MS/MS platforms [26]. Despite these obstacles to compound identification from LC-MS datasets the advantages of this approach, particularly with regard to sensitivity, have provided demand for the development of alternative compound identification tools. One such alternative uses very high mass accuracy measurements to allow the direct calculation of elemental composition for compounds <250350 amu. This approach typically requires mass errors <10 ppm and provides one or multiple elemental compositions calculated from the monoisotopic mass of each unknown compound using integer combinations of naturally occurring isotope masses [91]. Recently, stable isotopic labeling has been used to extend the mass range and precision of elemental composition assignment from accurate mass measurements by allowing the numbers of carbon and or nitrogen atoms to be directly measured thereby eliminating the carbon and nitrogen variables from the elemental composition calculation [30, 86, 87].

143

Spectral library resources

Web resources for MS-based metabolomics data analysis have been recently reviewed [92]. Mass spectral libraries are available from various commercial sources such as NIST 08 with 191 436 spectra from the National Institute for Standards and Technology and the Wiley Registry eighth ed. with $400 000 spectra. Many EI spectral libraries do not provide

adequate coverage of metabolites but instead contain large numbers of compounds from other classes including petrochemicals and pharmaceuticals. NIST libraries have been carefully curated though several cycles of revision and additions and also include GC retention indices for 44 008 compounds [93]. Metabolite specific GC-MS EI libraries are also available including plant specific collections in Oliver Fiehns lab containing 1162 spectra, and the Golm Metabolite Database with both GC-quadrupole (306 compounds) and GC-TOF (229 compounds). The Golm database also contains unidentified spectra from plant extracts for GC-quad (268 spectra) and GC-TOF (403 spectra) [94]. The Human Metabolome Database has a metabolite GC-MS EI library with 305 entries. All three of the metabolite specific libraries almost exclusively contain spectra of trimethylsilyl (TMS) derivatives and occasionally methyl oxime derivatives [95]. The range of reagents available and their ability to modify diverse functional groups (carboxylic acids, alcohols, amines, phosphoesters, sulpho esters), often without prior knowledge of the compound structures, has contributed to the wide adoption of TMS derivatization. Because of their general applicability, TMS techniques can be quite useful, however, they do have several drawbacks [96]. In LC-MS, MS/MS spectra can theoretically provide adequate fragmentation information to implement a spectral database matching strategy for identification of compounds analogous to that of GC-MS. In practice, however, MS/MS spectra vary from instrument to instrument, rendering spectral library matching strategies for MS/MS less effective. Still, efforts have been made to try to build MS/MS libraries for specific MS/MS instrumentation. Again, NIST has taken the lead with a collection of 14 802 spectra of 3898 positive and 1410 negative ions. The METLIN database has also made significant progress emphasizing metabolites with 573 positive ion and 587 negative ion spectra representing 881 total metabolites [97]. Perhaps the availability of these databases will help provide a basis set for MS/MS standardization that should make library searching of LC-MS/MS data a viable option. NMR and IR spectra can often be calculated although the results are not always accurate. The Madison Metabolomics Consortium Database has NMR spectral information for >19 700 metabolites and related compounds that includes many simulations in addition to spectra collected from standard compounds [98].

144

Hegeman
molecules and signals derived from the matrix. Four main strategies have been used to avoid this problem: the first uses an IR laser with an atmospheric pressure (AP) source, which excites water in the sample as the matrix [110]; the second employs higher molecular weight matrix molecules such as C60 or colloidal graphite, which do not produce as many interfering peaks [111]; the third uses tandem MS to monitor reaction products of a selected mass fragmentation event to dramatically reduce background noise [112]; the fourth uses matrix free laser desorption with a standard UV laser configuration and special surfaces such as porous silicon wafers [113]. Examples using MALDI imaging approaches plant samples include two studies looking at metabolites in fruit. Zhang and coworkers performed MALDI imaging on mature apple and strawberry section comparing standard MALDI matrix approaches to one using colloidal graphite [111]. This investigation demonstrated that colloidal graphite yielded superior results with less matrix interference. It should be pointed out that some MALDI instrumentation vendors discourage the use of graphite matrices due to short-circuiting of source electronics with excessive use of the highly conductive powder. Images were generated for apple and strawberry sections and compound distributions were generated from MS data only while MS/MS data were used to support compound assignments. MS/MS transitions for signal to noise enhancement was demonstrated for ellagic acid localization in the strawberry samples. This colloidal graphite approach was later used to examine cuticular waxes and flavonoids from various surfaces and cross-sections of Arabidopsis aerial tissues [114]. Li and coworkers have published three articles detailing using native water as matrix in samples prepared for analysis using an atmospheric pressure MALDI source with an IR laser [110, 115, 116]. One examined strawberry, banana and grape tissue and saw a host of sugars and organic acids as were seen by Zhang et al. except that fewer fatty acids were detected. This may be a consequence of using water as the matrix, which may favor hydrophilic analytes over lipids. Most of the oligosaccharides were observed as potassium adducts and some additional amino acids were observed that were not discussed by Zhang. This approach has an added advantage of bypassing the need for exogenous matrix application.

Spatial distribution and mass spectrometry-based imaging

Most experimental approaches for plant metabolomics employ homogenization and extraction procedures to remove metabolites from other plant components that interfere with analysis. Homogenization, of course, completely destroys potentially valuable information concerning metabolite distribution across a sample [64]. Brute force has been used to dissect, homogenize and analyze specific zones of larger plant organs such as melon [36], but this is not always possible for smaller tissues. Laser-assisted microdissection techniques are critical for sampling specific cell types but do not easily provide a holistic view of metabolite distribution (reviewed in ref. [99]). While many direct spatial imaging approaches exist these often fail to be either sufficiently comprehensive or sensitive to be useful for metabolomics. Several strategies using mass spectrometry (MS) provide measurements for multiple species with spatial resolution, although the resolution can be quite low [100, 101] (reviewed in ref. [102]). These include: Secondary Ion Mass Spectrometry (SIMS) imaging [103], Desorption Electrospray Ionization (DESI) imaging [104] and Matrix Assisted Laser Desorption Ionization (MALDI) imaging [105]. SIMS imaging offers the best resolution images ($1 nm) of the MS imaging options although molecular fragmentation resulting from high-energy ionization may limit this approach for metabolomics. SIMS has been used for imaging of lipids in frozen mammalian tissue sections using diagnostic product ions by TOF-SIMS [106], but has not been used extensively for plant tissues to date. Use of lower energy ion beams will improve the utility of this approach for biological samples but decrease resolution ($1 mm) [107, 108]. DESI is a newer approach that is significantly lower in resolution (>250 mm) but much lower energies allowing observation of intact molecular ions. MALDI Imaging has been the most widely used MS imaging approach for biological samples and provides resolutions from 20 to 200 mm [105, 109]. MALDI uses a laser instead of an ion beam (as in SIMS) and requires the physical application of a substance called a matrix to assist in the ionization process. MALDI like DESI is a soft ionization process that results primarily in intact molecular ions with a single charge. One of the main challenges for metabolomics applications of MALDI imaging concerns the high degree of spectral overlap between small

Analytical challenges of comprehensive metabolite analysis

145

CONCLUSION
Metabolomics is becoming an increasingly popular approach for identifying the molecular underpinnings of macroscopic biological phenomena. Its applicability to species with incomplete genomic data has made it quite useful for the analysis of properties of crop species prior to the complete sequencing of their genomes. In contrast to high throughput methodology for the analysis of DNA, RNA and proteins, current strategies for metabolite characterization still face significant obstacles. These challenges are largely caused by the high degree of chemical diversity among metabolite pools as well as the complexity of spatial and temporal distribution within living tissues. Plant metabolomics methodology and instrumentation are being developed at a rapid pace to address these analytical challenges.
8.

9.

10.

11.

12. 13. 14.

Key Points
 In the past decade plant metabolomics has experienced dramatic growth in both the technological aspects and in the types of biological questions being addressed.  Metabolomics is currently constrained by analytical obstacles affecting compound identification, quantification and localization that are largely brought about by the tremendous extent of chemical diversity across metabolite populations.  Stable isotope labeling can be used to provide additional control for extraction, fractionation, derivatization and detection to improve metabolite coverage quantification.  Emerging mass spectrometry-based imaging technologies will help provide a spatial dimension to metabolomics.

15. 16.

17.

18.

19.

20.

References
1. 2. Fiehn O. Metabolomics - the link between genotypes and phenotypes. Plant Mol Biol 2002;48:15571. Oliver SG, Winson MK, Kell DB, et al. Systematic functional analysis of the yeast genome. T rends Biotechnol 1998;16: 3738. Beecher CWW. The human metabolome. In: Harrigan GG, Goodacre R (eds). Metabolic Profiling: Its Role in Biomarker Discovery and Gene Function Analysis. Dordrecht: Kluwer Academic Publishers, 2003:3119. Trethewey RN. Metabolite profiling as an aid to metabolic engineering in plants. Curr Opin Plant Biol 2004;7:196201. Oksman-Caldentey KM, Saito K. Integrating genomics and metabolomics for engineering plant metabolic pathways. Curr Opin Biotechnol 2005;16:174179. Pauling L, Robinson AB, Teranishi R, et al. Quantitative analysis of urine vapor and breath by gas-liquid partition chromatography. Proc Natl Acad Sci USA 1971;68:23746. Nicholson JK, Lindon JC, Holmes E. Metabonomics: understanding the metabolic responses of living systems to pathophysiological stimuli via multivariate statistical analysis

21.

22. 23.

3.

24.

4. 5.

25.

26. 27.

6.

7.

of biological NMR spectroscopic data. Xenobiotica 1999;29: 11819. Lindon JC, Holmes E, Nicholson JK. Peer reviewed: so whats the deal with metabonomics? Anal Chem 2003;75: 384A91A. Allen J, Davey HM, Broadhurst D, et al. High-throughput classification of yeast mutants for functional genomics using metabolic footprinting. Nat Biotechnol 2003;21:6926. Kaderbhai Naheed N, Broadhurst David I, et al. Functional genomics via metabolic footprinting: monitoring metabolite secretion by Escherichia Coli tryptophan metabolism mutants using FT-IR and direct injection electrospray mass spectrometry. Comp Funct Genomics 2003;4:37691. Gates SC, Young ND, Holland JF, et al. Automated multicomponent analysis of biological mixtures by gas chromatography-mass spectrometry. Adv Mass Spectrom Biochem Med 1977;2:17181. Hall RD. Plant metabolomics: from holistic hope, to hype, to hot topic. New Phytol 2006;169:45368. Keurentjes JJB. Genetical metabolomics: closing in on phenotypes. Curr Opin Plant Biol 2009;12:22330. Schauer N, Fernie AR. Plant metabolomics: towards biological function and mechanism. T rendsPlantSci 2006;11:50816. Weckwerth W. Metabolomics in systems biology. Annu Rev Plant Biol 2003;54:66989. Fernie AR, Schauer N. Metabolomics-assisted breeding: a viable option for crop improvement? T rends Genet 2009; 25:3948. Harrigan GG, Martino-Catt S, Glenn KC. Metabolomics, metabolic diversity and genetic variation in crops. Metabolomics 2007;3:25972. Hall RD, Brouwer ID, Fitzgerald MA. Plant metabolomics and its potential application for human nutrition. Physiol Plant 2008;132:16275. Dixon RA, Gang DR, Charlton AJ, et al. Applications of metabolomics in agriculture. J Agric Food Chem 2006;54: 898494. Fiehn O, Sumner LW, Rhee SY, et al. Minimum reporting standards for plant biology context information in metabolomic studies. Metabolomics 2007;3:195201. Fiehn O, Wohlgemuth G, Scholz M, et al. Quality control for plant metabolomics: reporting MSI-compliant studies. PlantJ 2008;53:691704. Ryan D, Robards K. Analytical chemistry considerations in plant metabolomics. Sep Purif Rev 2006;35:31956. Sumner LW, Mendes P, Dixon RA. Plant metabolomics: large-scale phytochemistry in the functional genomics era. Phytochemistry 2003;62:81736. Eisenreich W, Bacher A. Advances of high-resolution NMR techniques in the structural and metabolic analysis of plant biochemistry. Phytochemistry 2007;68:2799815. Ward JL, Baker JM, Beale MH. Recent applications of NMR spectroscopy in plant metabolomics. Febs J 2007; 274:112631. Dettmer K, Aronov PA, Hammock BD. Mass spectrometrybased metabolomics. Mass Spectrom Rev 2007;26:5178. Mouille G, Johansen JN. Metabolic fingerprinting and profiling by MID-FT-IR (middle-Fourier transformedinfrared). In: Morot-Gaudry JF, Lea P, Briat JF (eds). Functional Plant Genomics. USA: Science Publishers Inc., 2007, 6648.

146

Hegeman
44. Dixon RA, Achnine L, Deavours BE, et al. Metabolomics and gene identification in plant natural product pathways. Biotechnol Agric For 2006;57:24359. 45. Jansen JJ, Allwood JW, Marsden-Edwards E, et al. Metabolomic analysis of the interaction between plants and herbivores. Metabolomics 2009;5:15061. 46. Leiss KA, Choi YH, Abdel-Farid IB, et al. NMR metabolomics of thrips (Frankliniella occidentalis) resistance in senecio hybrids. J Chem Ecol 2009;35:21929. 47. Leiss KA, Maltese F, Choi YH, et al. Identification of chlorogenic acid as a resistance factor for thrips in chrysanthemum. Plant Physiol 2009;150:156775. 48. Guy C, Kaplan F, Kopka J, et al. Metabolomics of temperature stress. Physiol Plant 2008;132:22035. 49. Ceoldo S, Toffali K, Mantovani S, et al. Metabolomics of daucus carota cultured cell lines under stressing conditions reveals interactions between phenolic compounds. Plant Sci 2009;176:55365. 50. Sanchez DH, Siahpoosh MR, Roessner U, etal. Plant metabolomics reveals conserved and divergent metabolic responses to salinity. Physiol Plant 2008;132:20919. 51. Shulaev V, Cortes D, Miller G, et al. Metabolomics for plant stress response. Physiol Plant 2008;132:199208. 52. Broeckling CD, Farag M, Huhman D, et al. Metabolomics and integrated functional genomics of Medicago truncatula responses to biotic and abiotic stress. Biol Plant-Microbe Interact 2006;5:1036. 53. Ishihara A, Matsuda F, Miyagawa H, et al. Metabolomics for metabolically manipulated plants: effects of tryptophan overproduction. Metabolomics 2007;3:31934. 54. van der Kooy F, Verpoorte R, Meyer JJM. Metabolomic quality control of claimed anti-malarial Artemisia afra herbal remedy and A. afra and A. annua plant extracts. S AfrJ Bot 2008;74:1869. 55. Mavituna F, Cenkci S, Walmsley R, et al. Modeling of the metabolism of Catharanthus roseus plant cell cultures for improved yields of pharmaceuticals. In: Proceedings of the 7th World Congress of Chemical Engineering, Glasgow, Scotland, 2005. pp. 857017. 56. Ma C, Wang H, Lu X, et al. Metabolic fingerprinting investigation of Artemisia annua L. in different stages of development by gas chromatography and gas chromatography-mass spectrometry. J Chromatogr A 2008; 1186:4129. 57. Shyur L-F, Yang N-S. Metabolomics for phytomedicine research and drug development. Curr Opin Chem Biol 2008;12:6671. 58. Viant MR. Recent developments in environmental metabolomics. Mol BioSyst 2008;4:9806. 59. Scherling C, Ulrich K, Ewald D, et al. A metabolic signature of the beneficial interaction of the endophyte Paenibacillus sp. isolate and in vitro-grown poplar plants revealed by metabolomics. Mol Plant-Microbe Interact 2009; 22:10327. 60. Abdel-Farid IB, Jahangir M, van den Hondel CAMJJ, et al. Fungal infection-induced metabolites in Brassica rapa. Plant Sci 2009;176:60815. 61. Hamzehzarghani H, Paranidharan V, Abu-Nada Y, et al. Metabolite profiling coupled with statistical analyses for potential high-throughput screening of quantitative resistance to fusarium head blight in wheat. Can J Plant Pathol 2008;30:2436.

28. De Vos RCH, Moco S, Lommen A, et al. Untargeted largescale plant metabolomics using liquid chromatography coupled to mass spectrometry. Nat Protoc 2007;2:77891. 29. Grata E, Boccard J, Guillarme D, et al. UPLC-TOF-MS for plant metabolomics: a sequential approach for wound marker analysis in Arabidopsis thaliana. J Chromatogr B: Anal Technol Biomed Life Sci 2008;871:26170. 30. Giavalisco P, Hummel J, Lisec J, etal. High-resolution direct infusion-based mass spectrometry in combination with whole 13C metabolome isotope labeling allows unambiguous assignment of chemical sum formulas. Anal Chem 2008; 80:941725. 31. Schroff R, Rulisek L, Doubsky J, et al. Acid-base-driven matrix-assisted mass spectrometry for targeted metabolomics. Proc Natl Acad Sci USA 2009;106:100926. 32. Kanani H, Chrysanthopoulos PK, Klapa MI. Standardizing GC-MS metabolomics. J Chromatogr B Anal Technol Biomed Life Sci 2008;871:191201. 33. Garcia-Perez I, Vallejo M, Garcia A, et al. Metabolic fingerprinting with capillary electrophoresis. J ChromatogrA 2008; 1204:1309. 34. Porter SEG, Stoll DR, Rutan SC, etal. Analysis of four-way two-dimensional liquid chromatography-diode array data: application to metabolomics. Anal Chem 2006;78:555969. 35. Kempa S, Hummel J, Schwemmer T, et al. An automated GCxGC-TOF-MS protocol for batch-wise extraction and alignment of mass isotopomer matrixes from differential 13 C-labelling experiments: a case study for photoautotrophic-mixotrophic grown Chlamydomonas reinhardtii cells. J Basic Microbiol 2009;49:8291. 36. Biais B, Allwood JW, Deborde C, et al. 1H NMR, GC-EITOFMS, and data set correlation for fruit metabolomics: application to spatial metabolite analysis in melon. Anal Chem 2009;81:288494. 37. tKindt R, Morreel K, Deforce D, et al. Joint GC-MS and LC-MS platforms for comprehensive plant metabolomics: repeatability and sample pre-treatment. J Chromatogr BAnalyt Technol Biomed Life Sci 2009;877:357280. 38. Kuiper HA, Kleter GA, Noteborn HPJM, et al. Substantial equivalence - an appropriate paradigm for the safety assessment of genetically modified foods? Toxicology 2002;181^ 182:42731. 39. Catchpole GS, Beckmann M, Enot DP, et al. Hierarchical metabolomics demonstrates substantial compositional similarity between genetically modified and conventional potato crops. Proc Natl Acad Sci USA 2005;102:1445862. 40. Beale MH, Ward JL, Baker JM. Establishing substantial equivalence: metabolomics. Methods Mol Biol 2008;478: 289303. 41. May P, Wienkoop S, Kempa S, et al. Metabolomics- and proteomics-assisted genome annotation and analysis of the draft metabolic network of Chlamydomonas reinhardtii. Genetics 2008;179:15766. 42. Boettcher C, Westphal L, Schmotz C, et al. The multifunctional enzyme CYP71B15 (phytoalexin deficient) converts cysteine-indole-3-acetonitrile to camalexin in the indole-3acetonitrile metabolic network of Arabidopsis thaliana. Plant Cell 2009;21:183045. 43. Fukushima A, Kusano M, Nakamichi N, et al. Impact of clock-associated Arabidopsis pseudo-response regulators in metabolic coordination. Proc Natl Acad Sc. USA 2009;106: 72516.

Analytical challenges of comprehensive metabolite analysis

62. Beckmann M, Enot DP, Overy DP, et al. Representation, comparison, and interpretation of metabolome fingerprint data for total composition analysis and quality trait investigation in potato cultivars. J Agric Food Chem 2007;55: 344451. 63. Roessner U. The chemical challenge of the metabolome. In: Villas-Boas SG, Roessner U, Hansen MAE, et al. (eds). Metabolome Analysis: An Introduction. Hoboken, NJ: WileyInterscience, 2007:1538. 64. Roessner U, Pettolino F. The importance of anatomy and physiology in plant metabolomics. Top Curr Genet 2007;18: 25378. 65. Sato S, Arita M, Soga T, et al. Time-resolved metabolomics reveals metabolic modulation in rice foliage. BMC Syst Biol 2008;2:51. 66. Tsai AL, Berka V, Kulmacz RJ, et al. An improved sample packing device for rapid freeze-trap electron paramagnetic resonance spectroscopy kinetic measurements. Anal Biochem 1998;264:16571. 67. Sharkey TD, Kobza J, Seemann JR, et al. Reduced cytosolic fructose-1,6-bisphosphatase activity leads to loss of oxygen sensitivity in a Flaveria linearis mutant. Plant Physiol 1988; 86:66771. 68. Kimball E, Rabinowitz JD. Identifying decomposition products in extracts of cellular metabolites. Anal Biochem 2006;358:27380. 69. Want EJ, OMaille G, Smith CA, et al. Solvent-dependent metabolite distribution, clustering, and protein extraction for serum profiling with mass spectrometry. Anal Chem 2006;78:74352. 70. Villas-Boas SG, Hojer-Pedersen J, Akesson M, et al. Global metabolite analysis of yeast: evaluation of sample preparation methods. Yeast 2005;22:115569. 71. tKindt R, De Veylder L, Storme M, et al. LC-MS metabolic profiling of Arabidopsis thaliana plant leaves and cell cultures: optimization of pre-LC-MS procedure parameters. J Chromatogr B AnalTechnol Biomed Life Sci 2008;871:3743. 72. Kruger NJ, Troncoso-Ponce MA, Ratcliffe RG. 1H NMR metabolite fingerprinting and metabolomic analysis of perchloric acid extracts from plant tissues. Nat Protoc 2008;3: 100112. 73. Lewis IA, Schommer SC, Hodis B, et al. Method for determining molar concentrations of metabolites in complex solutions from two-dimensional 1H-13C NMR spectra. Anal Chem 2007;79:938590. 74. Boettcher C, Roepenack-Lahaye Ev, Willscher E, et al. Evaluation of matrix effects in metabolite profiling based on capillary liquid cChromatography electrospray ionization quadrupole time-of-flight mass spectrometry. Anal Chem 2007;79:150713. 75. Annesley TM. Ion suppression in mass spectrometry. Clin Chem 2003;49:10414. 76. Kiefer P, Portais JC, Vorholt JA. Quantitative metabolome analysis using liquid chromatography-high-resolution mass spectrometry. Anal Biochem 2008;382:94100. 77. Bennett BD, Yuan J, Kimball EH, et al. Absolute quantitation of intracellular metabolite concentrations by an isotope ratio-based approach. Nat Protoc 2008;3:1299311. 78. Feldberg L, Venger I, Malitsky S, et al. Dual labeling of metabolites for metabolome analysis (DLEMMA): a new approach for the identification and relative quantification of metabolites by means of dual isotope labeling and

147

79.

80.

81. 82. 83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

liquid chromatography-mass spectrometry. Anal Chem 2009;81:925766. Huang X, Regnier FE. Differential metabolomics using stable isotope labeling and two-dimensional gas chromatography with time-of-flight mass spectrometry. Anal Chem 2008;80:10714. Bakhtiar R, Majumdar TK. Tracking problems and possible solutions in the quantitative determination of small molecule drugs and metabolites in biological fluids using liquid chromatography-mass spectrometry. J Pharmacol Toxicol Methods 2007;55:26278. Tolgyessy J, Braun T, Kyrs M. Isotope Dilution Analysis. Oxford: Pergamon Press, 1972. Beynon RJ, Pratt JM. Metabolic labeling of proteins for proteomics. Mol Cell Proteomics 2005;4:85772. Kikuchi J, Hirayama T. Practical aspects of uniform stable isotope labeling of higher plants for heteronuclear NMR-based metabolomics. Methods Mol Biol 2007;358: 27386. Engelsberger WR, Erban A, Kopka J, et al. Metabolic labeling of plant cell cultures with K15NO3 as a tool for quantitative analysis of proteins and metabolites. Plant Methods 2006;2:14. Harada K, Fukusaki E, Bamba T, et al. In vivo 15 N-enrichment of metabolites in suspension cultured cells and its application to metabolomics. Biotechnol Prog 2006;22: 100311. Giavalisco P, Kohl K, Hummel J, et al. 13C isotope-labeled metabolomes allowing for improved compound annotation and relative quantification in liquid chromatographymass spectrometry-based metabolomic research. Anal Chem 2009;81:654651. Hegeman AD, Schulte CF, Cui Q, et al. Stable isotope assisted assignment of elemental compositions for metabolomics. Anal Chem 2007;79:691221. Epstein E, Cohen JD, Bandurski RS. Concentration and metabolic turnover of indoles in germinating kernels of Zea mays L. Plant Physiol 1980;65:41521. Bialek K, Meudt W, Cohen J. Indole-3-acetic acid (IAA) and IAA conjugates applied to bean stem sections: IAA content and the growth response. Plant Physiol 1983;72: 6914. Kind T, Fiehn O. Metabolomic database annotations via query of elemental compositions: mass accuracy is insufficient even at less than 1 ppm. BMC Bioinformatics 2006;7: 234. Kind T, Fiehn O. Seven golden rules for heuristic filtering of molecular formulas obtained by accurate mass spectrometry. BMC Bioinformatics 2007;8:105. Tohge T, Fernie AR. Web-based resources for massspectrometry-based metabolomics: a users guide. Phytochemistry 2009;70:4506. Ausloos P, Clifton CL, Lias SG, et al. The critical evaluation of a comprehensive mass spectral library. J Am Soc Mass Spectrom 1999;10:28799. Schauer N, Steinhauser D, Strelkov S, etal. GC-MS libraries for the rapid identification of metabolites in complex biological samples. FEBS Lett 2005;579:13327. Fiehn O. Extending the breadth of metabolite profiling by gas chromatography coupled to mass spectrometry. T rends Analyt Chem 2008;27:2619.

148

Hegeman
107. Weibel D, Wong S, Lockyer N, et al. A C60 primary ion beam system for time of flight secondary ion mass spectrometry: its development and secondary ion yield characteristics. Anal Chem 2003;75:175464. 108. Sodhi RN. Time-of-flight secondary ion mass spectrometry (TOF-SIMS): versatility in chemical and imaging surface analysis. Analyst 2004;129:4837. 109. Cornett DS, Reyzer ML, Chaurand P, et al. MALDI imaging mass spectrometry: molecular snapshots of biochemical systems. Nat Methods 2007;4:82833. 110. Li Y, Shrestha B, Vertes A. Atmospheric pressure infrared MALDI imaging mass spectrometry for plant metabolomics. Anal Chem 2008;80:40720. 111. Zhang H, Cha S, Yeung ES. Colloidal Graphite-Assisted Laser Desorption/Ionization MS and MSn of Small Molecules. 2. Direct Profiling and MS Imaging of Small Metabolites from Fruits. Anal. Chem. (Washington, DC, US) 2007;79:65756584. 112. Khatib-Shahidi S, Andersson M, Herman JL, et al. Direct molecular analysis of whole-body animal tissue sections by imaging MALDI mass spectrometry. Anal Chem 2006;78: 6448456. 113. Liu Q, Guo Z, He L. Mass spectrometry imaging of small molecules using desorption/ionization on silicon. Anal Chem 2007;79:353541. 114. Cha S, Zhang H, Ilarslan HI, et al. Direct profiling and imaging of plant metabolites in intact tissues by using colloidal graphite-assisted laser desorption ionization mass spectrometry. PlantJ 2008;55:34860. 115. Li Y, Shrestha B, Vertes A. Atmospheric pressure molecular imaging by infrared MALDI mass spectrometry. Anal Chem 2007;79:52332. 116. Nemes P, Barton AA, Li Y, et al. Ambient molecular imaging and depth profiling of live tissue by infrared laser ablation electrospray ionization mass spectrometry. Anal Chem 2008;80:457582.

96. Little JL. Artifacts in trimethylsilyl derivatization reactions and ways to avoid them. J ChromatogrA 1999;844:122. 97. Sana TR, Roark JC, Li X, et al. Molecular formula and METLIN personal metabolite database matching applied to the identification of compounds generated by LC/TOF-MS. J BiomolTech 2008;19:25866. 98. Cui Q, Lewis IA, Hegeman AD, et al. Metabolite identification via the Madison Metabolomics Consortium Database. Nat Biotechnol 2008;26:1624. 99. Moco S, Schneider B, Vervoort J. Plant micrometabolomics: the analysis of endogenous metabolites present in a plant cell or tissue. J Proteome Res 2009;8: 1694703. 100. Stoeckli M, Chaurand P, Hallahan DE, et al. Imaging mass spectrometry: a new technology for the analysis of protein expression in mammalian tissues. Nat Med 2001;7:4936. 101. Todd PJ, Schaaff TG, Chaurand P, et al. Organic ion imaging of biological tissue with secondary ion mass spectrometry and matrix-assisted laser desorption/ionization. J Mass Spectrom 2001;36:35569. 102. McDonnell LA, Heeren RMA. Imaging mass spectrometry. Mass Spectrom Rev 2007;26:60643. 103. Guerquin-Kern JL, Wu TD, Quintana C, et al. Progress in analytical imaging of the cell by dynamic secondary ion mass spectrometry (SIMS microscopy). Biochim Biophys Acta 2005; 1724:22838. 104. Cooks RG, Ouyang Z, Takats Z, et al. Detection technologies. Ambient mass spectrometry. Science 2006;311: 156670. 105. Reyzer ML, Caprioli RM. MALDI-MS-based imaging of small molecules and proteins in tissues. Curr Opin Chem Biol 2007;11:2935. 106. Nygren H, Malmberg P. High resolution imaging by organic secondary ion mass spectrometry. T rends Biotechnol 2007;25:499504.

Copyright of Briefings in Functional Genomics & Proteomics is the property of Oxford University Press / UK and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.

Copyright of Briefings in Functional Genomics is the property of Oxford University Press / UK and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.