Novel Approach for High-Throughput Metabolic Screening of ... · ensure the global food supply (Khush, 2003) and to produce various chemicals and materials (Fischer and Emans, 2000;

Breakthrough Technologies

Novel Approach for High-Throughput MetabolicScreening of Whole Plants by Stable Isotopes

Lisa Maria Dersch, Veronique Beckers, Detlev Rasch, Guido Melzer1, Christoph Bolten2, Katina Kiep3,Horst Becker, Oliver Ernst Bläsing, Regine Fuchs, Thomas Ehrhardt, and Christoph Wittmann*

Institute of Systems Biotechnology, Saarland University, 66123 Saarbrücken, Germany (L.M.D., V.B., C.W.);Institute of Biochemical Engineering, University of Technology Braunschweig, 38106 Braunschweig, Germany(D.R., G.M., C.B., K.K.); BASF SE, 67117 Limburgerhof, Germany (H.B.); and Metanomics GmbH, 10589 Berlin,Germany (O.E.B., R.F., T.E.)

ORCID IDs: 0000-0001-5385-1739 (L.M.D.); 0000-0003-1928-5600 (D.R.); 0000-0003-0081-5050 (G.M.); 0000-0003-0418-8939 (C.B.);0000-0002-3294-3666 (T.E.); 0000-0002-7952-985X (C.W.).

Here, we demonstrate whole-plant metabolic profiling by stable isotope labeling and combustion isotope-ratio mass spectrometryfor precise quantification of assimilation, translocation, and molecular reallocation of 13CO2 and

15NH4NO3. The technology wasapplied to rice (Oryza sativa) plants at different growth stages. For adult plants, 13CO2 labeling revealed enhanced carbonassimilation of the flag leaf from flowering to late grain-filling stage, linked to efficient translocation into the panicle.Simultaneous 13CO2 and 15NH4NO3 labeling with hydroponically grown seedlings was used to quantify the relativedistribution of carbon and nitrogen. Two hours after labeling, assimilated carbon was mainly retained in the shoot (69%),whereas 7% entered the root and 24% was respired. Nitrogen, taken up via the root, was largely translocated into the shoot(85%). Salt-stressed seedlings showed decreased uptake and translocation of nitrogen (69%), whereas carbon metabolism wasunaffected. Coupled to a gas chromatograph, labeling analysis provided enrichment of proteinogenic amino acids. Thisrevealed significant protein synthesis in the panicle of adult plants, whereas protein biosynthesis in adult leaves was 8-foldlower than that in seedling shoots. Generally, amino acid enrichment was similar among biosynthetic families and allowed usto infer labeling dynamics of their precursors. On this basis, early and strong 13C enrichment of Embden-Meyerhof-Parnaspathway and pentose phosphate pathway intermediates indicated high activity of these routes. Applied to mode-of-actionanalysis of herbicides, the approach showed severe disturbance in the synthesis of branched-chain amino acids upontreatment with imazapyr. The established technology displays a breakthrough for quantitative high-throughput plantmetabolic phenotyping.

Metabolically engineered crops are highly desired toensure the global food supply (Khush, 2003) and toproduce various chemicals and materials (Fischerand Emans, 2000; Sharma and Sharma, 2009). Withoutdoubt, tailored plant metabolic engineering requires agood knowledge of the underlying metabolism. Thisexplains the strong interest in tools and technologies toanalyze plants on the systems level. Particularly, anal-ysis of plant metabolic pathways and intracellularfluxes bymeans of isotope experiments, coupled tomassspectrometry (MS) andNMR (Ratcliffe and Shachar-Hill,2006; Young et al., 2011), has predictive power formetabolic engineering (Kruger and Ratcliffe, 2009;Shachar-Hill, 2013). Such analyses, reported for cellsuspension cultures (Rontein et al., 2002; Kruger et al.,2007; Williams et al., 2008) and isolated plant organslike leaves (Schaefer et al., 1980; Cegelski and Schaefer,2005; Hasunuma et al., 2010), tubers (Roessner-Tunaliet al., 2004), and seeds (Schwender et al., 2003; Junkeret al., 2007), provide valuable insight into metabolismbut fail to describe the behavior of intact, whole plants(Allen et al., 2009), which is necessary to give a systemicpicture of metabolic functions under physiologicallyrelevant conditions (Cliquet et al., 1990; Römisch-Margl

1 Present address: Sandoz GmbH, 65929 Frankfurt am Main,Germany.

2 Present address: Nestlé Suisse S.A., 1814 La Tour-de-Peilz,Switzerland.

3 Present address: Evonik Industries AG, 45128 Essen, Germany.* Address correspondence to [email protected] author responsible for distribution of materials integral to

the findings presented in this article in accordance with the policydescribed in the Instructions for Authors (www.plantphysiol.org) is: Christoph Wittmann ([email protected]).

L.M.D. conducted the isotope labeling experiments, assisted in es-tablishment of the GC-IR-MS method, processed and evaluated mea-surement data, drew all figures, and wrote the article; V.B.contributed to isotope labeling studies; D.R. designed and con-structed the labeling reactors and provided technical assistance dur-ing isotope labeling studies; C.B., G.M., and K.K. assisted in isotopelabeling studies during preliminary layout and testing of the ap-proach; H.B. conducted C-IR-MS and GC-C-IRMS analytics; R.F. con-tributed to project layout and supervision, experimental design,isotope labeling studies, labeling analytics, data evaluation and pro-cessing, and writing of the manuscript; O.E.B. and T.E. provided ad-vice on experimental design and writing of the manuscript; and C.W.processed and evaluated measurement data, conceived and super-vised the project, and wrote the manuscript.

www.plantphysiol.org/cgi/doi/10.1104/pp.15.01217

Plant Physiology�, May 2016, Vol. 171, pp. 25–41, www.plantphysiol.org � 2016 American Society of Plant Biologists. All Rights Reserved. 25 www.plantphysiol.orgon May 21, 2020 - Published by Downloaded from

Copyright © 2016 American Society of Plant Biologists. All rights reserved.

http://orcid.org/0000-0001-5385-1739

http://orcid.org/0000-0003-1928-5600

http://orcid.org/0000-0003-0081-5050

http://orcid.org/0000-0003-0418-8939

http://orcid.org/0000-0002-3294-3666

http://orcid.org/0000-0002-7952-985X

mailto:[email protected]

http://www.plantphysiol.org




http://www.plantphysiol.org/cgi/doi/10.1104/pp.15.01217


et al., 2007). This is overcome by performing isotope-labeling experiments in entire plants, preferably using13CO2, a far safer tracer compound than its radiolabeledequivalent 14CO2 (Römisch-Margl et al., 2007), com-bined with labeling analysis by MS and NMR (Chenet al., 2011). Such experiments, however, suffer fromextended labeling time periods up to several hours oreven days (Hutchinson et al., 1976; Nouchi et al., 1995;Wu et al., 2009) and CO2 concentrations beyond naturalabundance to reach detectable amounts of enrichment(Römisch-Margl et al., 2007). As an example, the ap-plication of elevated CO2 levels can induce changes inplant metabolism (e.g. sink-source ratios, photosyn-thetic activity, and respiration), thereby affecting car-bon sequestration and growth (Arp, 1991; Zhu et al.,2014). Compared with conventional MS (0.05 atom %)and NMR, combustion-isotope ratio mass spectrome-try (C-IRMS) provides a far higher precision (0.0002atom %; Meier-Augenstein, 1999b), which makes thistechnique particularly attractive to analyze plants atlow enrichment. The high precision of C-IRMS evenallows for metabolic studies without preceding tracerapplication (Tcherkez et al., 2003; Yousfi et al., 2012,2013). Selected applications, which coupled C-IRMSwith elemental analysis (EA), have been used to traceturnover and incorporation processes in plants (Cliquetet al., 1990; Dyckmans et al., 2000; Nogués et al., 2004) aswell as to analyze plant interactions with the soil mi-crobial community (Griffiths et al., 2004; Leake et al.,2006; Wu et al., 2009). With the availability of gas chro-matography (GC)-coupled C-IRMS, enrichment analysiscould be targeted to individual metabolic compounds,like sugars, fatty acids, and amino acids (Derrien et al.,2004; Olsson et al., 2005; Molero et al., 2011; Lattanziet al., 2012).

Here, we describe a novel approach, which combines13C- and 15N-based whole-plant studies with subse-quent labeling analysis of entire plants, plant tissues,and individual metabolites using EA-C-IRMS andGC-C-IRMS for in vivo metabolic fingerprinting of themodel crop rice (Oryza sativa) under physiologicallyrelevant conditions. For the isotope experiments, spe-cific labeling reactors were designed and constructed,which allowed precise 13C-pulse labeling of plants aswell as simultaneous 13C and 15N tracing under well-defined environmental conditions. This offered a majorbenefit compared with previous studies: the applica-bility of physiological concentrations of 13CO2 and shortlabeling periods under highly controlled conditionsregarding light, temperature, and humidity. The de-veloped approach was applied to compare rice plantsat different developmental stages, concerning assimi-lation, translocation, and incorporation of label intometabolic intermediates after single or double labelingwith 13CO2 and 15NH4NO3. Particularly, studies com-bining 13C and 15N label detection in plants are rare sofar (Cliquet et al., 1990; Dyckmans et al., 2000). Thereby,the interaction of carbon/nitrogen metabolism andrelevant sink-source relations of individual plantorgans were studied. Furthermore, a stress-induced

phenotype was investigated by exposing rice seedlingsto high salinity, amajor abiotic stress. The examinationof the mode of action of herbicide treatment furtherunderlined the high potential of the technique to reli-ably quantify metabolic variation. The developed tech-nology provides plant metabolic engineers with asophisticated tool for fast screening and metabolicprofiling of distinct phenotypes, which perfectly com-plements more demanding isotopically nonstationarymetabolic flux analysis (Szecowka et al., 2013; Ma et al.,2014) to get a comprehensive picture of plant metabolicfunctions.

RESULTS

Construction of Labeling Reactors andExperimental Design

Three sizes and types of labeling reactor weredesigned, constructed, and validated to evaluate theirsuitability for in vivo 13C labeling studies (Fig. 1): twotube reactors for 13CO2 studies with soil-grown adultplants (250 L) and seedlings (100 L), respectively, and abox reactor for combined 13CO2 and

15NH4NO3 labelingstudies with hydroponic seedlings (125 L). For all re-actors, the wall material used allowed full transmissionof light (Supplemental Fig. S1), and temperature andhumidity could be maintained at desired values(Supplemental Fig. S2). Hence, the reactors allowedplants to be labeled under the same light, temperature,and humidity regimen that they were exposed to dur-ing growth in a research plant growth cabinet (phytochamber). Immediate replacement of ambient CO2 byan equimolar level of 13CO2was realized by an absorberunit connected to the reactor. A first set of experimentswith soil-grown rice seedlings was conducted to iden-tify the optimum conditions regarding the supplyof tracer (400 or 700 mL L21 13CO2) and the time pe-riod of incubation with 13CO2 (10, 60, and 180 min;Supplemental Fig. S3). Short incubation times of 10 minat ambient levels of 13CO2 (400 mL L21) were sufficientto allow precise estimation of the assimilated carbon,due to the high precision of the EA-C-IRMS measure-ment (Supplemental Fig. S3A). Labeled under theseconditions, the shoots of rice seedlings showed marked13C enrichment (i.e. a d value of 170‰ 6 6‰). The lowdeviation underlines that the assimilation was quantifiedwith excellent reproducibility, particularly consideringthe high complexity of the studied plant system. En-richment values were in an equal range after simulta-neous labeling of one, three, six, or 12 plants in the samereactor for 60 min, indicating that even larger sets ofplants and longer incubation times did not result in CO2limitation (Supplemental Fig. S3B). In order to examinediurnal effects on photosynthesis and CO2 assimilation,isotopic labeling studies with seedlings were conductedevery 1 h in the time frame of 2.5 to 9.5 h after sunrise.Labeling of plants at different times did not reveal sig-nificant differences from plants labeled at midday (6.5 hafter sunrise; Supplemental Fig. S3C). Accordingly, the

26 Plant Physiol. Vol. 171, 2016

Dersch et al.

www.plantphysiol.orgon May 21, 2020 - Published by Downloaded from Copyright © 2016 American Society of Plant Biologists. All rights reserved.

http://www.plantphysiol.org/cgi/content/full/pp.15.01217/DC1







Figure 1. Equipment designed and constructed for in vivo 13C and 15N labeling studies. A, Large tube reactor (0.5 m diameter,1.3 m height, and 255 L volume) for 13CO2 labeling of soil-grown adult rice plants. B, Small tube reactor (0.5 m diameter, 0.5 mheight, and 98 L volume) for 13CO2 labeling of soil-grown rice seedlings. C, Box reactor (0.5 3 0.5 3 0.5 m, 125 L volume) forsimultaneous 13CO2 and

15NH4NO3 labeling of hydroponic rice seedlings. All reactorswere equippedwith a temperature control,

Plant Physiol. Vol. 171, 2016 27

Screening Technology for Quantitative Isotope Experiments



routine work flow was as follows. Rice plants weregrown in a phyto chamber under ambient air, until theywere placed inside the enclosure shortly before the label-ing experiment. Ambient CO2 was then removed from thereactor within 30 s and replaced by an equimolar amountof 13CO2 (400 mL L21). After 10 min of incubation in thisatmosphere, plants were removed from the reactor andeither harvested directly for assessment of carbon assimi-lation or cultivated further in the phyto chamber for as-sessment of carbon translocation. For simultaneous tracingof 15N, roots of hydroponically grown seedlings weresupplied with 15NH4NO3. The isotopic enrichment of har-vested plant material and extracted amino acids was de-termined using EA-C-IRMS andGC-C-IRMS, respectively.

Combined Assessment of Carbon and NitrogenMetabolism in Rice Seedlings

The developed approachwas next used for combined13CO2 and

15NH4NO3 labeling of hydroponically grown

rice seedlings, in order to obtain an integrated picture ofcarbon and nitrogen metabolism (Fig. 2). Isotopicallylabeled 15NH4NO3 was provided to the roots, while atthe same time, 13CO2 was supplied to the shoot of a riceseedling, using the designed box reactor (Fig. 1C). Im-mediately after the pulse, maximum 13C enrichmentwas detected in the shoot (450‰6 45‰; Fig. 2A), while15N labeling was highest for the root (4,130‰ 6 25‰;Fig. 2C). However, the transport of carbon and nitrogenseemed fast, because significant 15N enrichment at thistime point was already found in the shoot (190‰ 625‰; Fig. 2D). Likewise, the root contained slightamounts of 13C (15‰6 5‰; Fig. 2B). Within 2 h afterassimilation, 15N and 13C levels were evenly equili-brated. During the ongoing chase period, 13C and 15Nenrichments decreased continuously in root and shoot.By integration of the measured 13C and 15N enrichmentdata, it was now possible to determine the percentagereallocation of label within 2 h after the labeling pulse(Fig. 3), thereby providing fast and quantitative access

Figure 1. (Continued.)comprising a water-cooled ventilator at the bottom plate of the reactor and an external cryostat. In addition, an external CO2 ad-sorption unit consisted of a high-power pump, an adsorber, and a fine dust filter. Prior to the experiments, the chosen plants wereplaced into the reactor, which were then closed gas tight by a rubber seal. Ambient CO2 was removed from the reactor within 30 s.The experiments were started by injecting the desired amounts of 13CO2 through an injection valve in the lid of each reactor.

Figure 2. Impact of salt stress on carbon and nitrogen assimilation and translocation of hydroponically grown rice seedlings. Thedata reflect 13C (A) and 15N (B) enrichment in the shoot and 13C (C) and 15N (D) enrichment in the root. The data shown aremeans6SD (n = 3) for stressed seedlings (100mMNaCl for 6 d; white bars) and untreated controls (gray bars). At the age of 12 d, rice seedlingswere simultaneously labeled with 13CO2 (400 mL L21) through the reactor gas phase andwith 15NH4NO3 (1.43 mM) supplied via thehydroponic growthmedium. The labeling pulses were applied for 10min, after which plants were either harvested directly to assessassimilation or cultivated further at ambient air and in nonlabeled medium up to 48 h to trace label translocation. The 13C and 15Nenrichment of freeze-dried plant material was analyzed by C-IRMS coupled to EA. Asterisks indicate significant differences betweenmean values (P # 0.05, Student’s t test). n.s., Not significant. The full data set is given in Supplemental Table S1.


Dersch et al.




to relative carbon and nitrogen fluxes in rice plants. Theroot-to-shoot ratio, calculated from the labeling data viaEquation 4 (see “Materials and Methods”), was 0.173.This seemed a proper estimate according to previouslyreported values for rice between 0.05 and 0.3 (Yoshida,1981). The majority of assimilated nitrogen (85%) wastransported to the shoot (Fig. 3A), whereas only 15%remained in the roots. Regarding assimilated carbon,the major fraction (69%) was retained in the shoot,whereas only 7%was translocated into the roots. A totalof 76% of assimilated carbon was thus recovered insidethe plant 2 h after assimilation, which indicated a lossof 24% via respiration. The retainment of carbon in theshoot was slightly higher than that for maize (Zea mays)plants, in which 53% of 13C is recovered inside the shootat elongation (Meng et al., 2013). The same calculation,using labeling data from samples taken 24 h after thepulse (i.e. including a dark period), revealed an in-creased loss of carbon through respiration by 52% (datanot shown), which agreed well with the correspondingdata obtained from the measurement of dry matterproduction, photosynthesis, and respiration (Tanakaand Yamaguchi, 1968).

Impact of Salt Stress on the Assimilation and Translocationof Carbon and Nitrogen in Rice Seedlings

Hydroponically grown rice seedlings were exposed to100 mM sodium chloride for 6 d prior to the labelingexperiment. Untreated plants served as a control. Im-mediately after the labeling pulse, root 15N enrichmentwas 4,190‰ 6 25‰ in the control plant and only2,490‰6 160‰ in the stressed plant, corresponding toa decreased ammonium uptake of 40% (Fig. 2). In addi-tion, the translocation of nitrogen from root to shoot was

strongly impaired, leading to significantly less enrich-ment (43%–62%) in the shoot of stressed plants com-pared with control plants at all sampling time points. Incontrast, carbon assimilation in the shoot was ratherunaltered (Fig. 2A). As described above for untreatedcontrol plants, the distribution of carbon and nitrogenwas inferred for a stressed plant by integrating 13C and15N labeling data. The root-to-shoot ratio estimated viaEquation 4 was only slightly increased under stressconditions (0.165 6 0.01 for control plants comparedwith 0.173 6 0.03 for stressed plants). Relative carbondistribution did not change in plants exposed to stress(Fig. 3). Nitrogen distribution, however, changed sig-nificantly, leading to the recovery of 30% of the assimi-lated nitrogen in roots compared with 15% in controlplants. The amount of labeled nitrogen transported tothe shoot was decreased accordingly, resulting in 69%compared with 86% in control plants (Fig. 3B).

Quantitative Assessment of Carbon Assimilation in AdultRice Plants

In the flowering and grain-filling stages, rice plantsexhibit different leaf types (i.e. normal leaves and flagleaves). The flag leaf is the uppermost leaf on a culmand the main carbon provider of the panicle when fullydeveloped (Rawson and Hofstra, 1969). Assimilationpatterns of plants in different developmental stages,sampled immediately after a 10-min labeling pulse,were compared (Fig. 4). Most of the label was found inthe leaves. Plants in the flowering stage showed similar13C enrichment in normal leaves (260‰ 6 40‰) andflag leaves (240‰ 6 50‰; Fig. 4B). This was also ob-served for the early grain-filling stage. During late grainfilling, however, the flag leaf was more important for

Figure 3. Relative fluxes of assimilatedcarbon and nitrogen in hydroponicallygrown rice seedlings under controlconditions (A) and exposed to salt stress(100mMNaCl for 6 d; B), calculated from13C and 15N enrichment data obtained 2hafter the labeling pulse (Fig. 2). The root-to-shoot ratio for seedlings under controlconditions (0.173) and seedlings exposedto high salinity (0.165) was calculatedfrom the 15N label distribution usingEquations 4 and 6. Based on this ratio, therelative reallocation of label betweenshoot and root was determined. The totalamount of assimilated carbon and nitro-gen at time zero was set to 100% carbonand nitrogen uptake, respectively, in or-der to provide relative data. The seed-lingswere analyzedat the ageof 12d. Thefull data sets are given in SupplementalTable S1.







the assimilation of 13CO2, indicated by significantlyhigher enrichment (170‰ 6 10‰ in normal leavescompared with 270‰ 6 30‰ in flag leaves), whichis similar to previous findings for wheat (Triticum aes-tivum; Rawson and Hofstra, 1969) and consistent withits function as amajor source organ for carbon transportinto the panicle. The stem did not exhibit any enrich-ment, while the panicle showed low 13C accumulationduring early grain filling (10‰ 6 5‰; Fig. 4B).

Quantitative Assessment of Carbon Translocation in AdultRice Plants

In order to assess carbon translocation, plants werepostcultivated in the phyto chamber after a labelingperiod of 10 min. Samples taken after distinct chaseperiods provided a time-resolved 13C sequestration

pattern. Immediately after the 13CO2 pulse, most of thelabel was recovered in the leaves (Fig. 4D).Within 2 h offurther growth, the 13C enrichment of flag leaves andnormal leaves declined by 45% and 41%, respectively,while the enrichment in the stem and panicle increased.This trend continued until a maximal enrichment wasreached after 24 h for the panicle (100‰ 6 20‰) andafter 48 h for the stem (50‰6 15‰; Fig. 4D). As for theseedling experiments, the data for the adult plantsshowed high precision and reproducibility.

Quantification of 13C and 15N Amino Acid EnrichmentUsing GC-C-IRMS

Using GC-C-IRMS, the measurement of labelingenrichment was extended to amino acids. For eachamino acid eluting from the GC column, the 13C and 15N

Figure 4. Carbon assimilation and translocation in adult rice plants, assessed by 13CO2 isotope experiments and labeling analysisby C-IRMS coupled to EA. A, Morphology of the studied plants with sampled leaf, flag leaf, stem, and panicle. B, Tissue-specific13C assimilation of rice at flowering stage (68 d), at early grain-filling stage (75 d), and at late grain-filling stage (88 d), for whichsoil-grown plants were labeled with 400 mL L21 13CO2 for 10 min and harvested directly for assessment of 13C enrichment. C,Morphology of the studied plants with identified carbon assimilation routes. D, Time-resolved carbon assimilation and trans-location of rice plants at late grain-filling stage (88 d), for which soil-grown plants were labeledwith 400 mL L21 13CO2 for 10 minand harvested directly or further cultivated at ambient air for 2, 4, 24, and 48 h prior to harvesting. In all cases, 13C enrichment offreeze-dried plant material is displayed as d13C (‰), corrected for natural labeling. Mean values6 SD (n = 3) are shown. Differentletters (a, b, or c) indicate significant differences betweenmeans (P# 0.05, one-way ANOVAwith Tukey’s test) of sampled organsat the individual time points. The full data sets of these experiments are given in Supplemental Tables S2 and S3.


Dersch et al.




enrichment was given by the ion intensity at a mass-to-charge ratio (m/z) of 45 and 29, reflecting 13CO2 and14N15N, respectively, formed by combustion in the in-strument. For 16 amino acids, satisfying signal qualitywas generally obtained and provided precise estimatesof their labeling status. Hereby, the amino acid pairsGlu/Gln andAsp/Asnwere each quantified as lumpedpools, due to conversion of the carboxamides Asn andGln into their corresponding acids during the proteinhydrolysis step of sample processing (Wittmann, 2007).Four amino acids (Cys, Met, Trp, and Arg) were notaccessible because they were degraded during thistreatment (Wittmann, 2007). Generally, 5 mg of lyophi-lized material was sufficient to provide high-qualitydata, independent of the type of tissue processed. Theconditions used for extraction, precipitation, and deriv-atization of protein from lyophilized plant material wascrucial with regard to the yield and purity of the aminoacids obtained. In contrast to extraction in Tris buffer(pH 8.8), extraction in hot water (100°C, 15 min) yieldedmuch less protein (data not shown). Similarly, precipi-tation of the extracted protein in ice-cold 10% (w/v) TCAwas less efficient than precipitation in ice-cold acetone(data not shown). Additional tests with alternativederivatization agents revealed that trimethylsilyl deriva-tives of the amino acids (e.g. obtainedwith trimethylsilyl-trifluoroacetamide)were not fully separated byGC. Evena set of variations in the temperature profile did not allowfull baseline separation for the amino acidsMet/Asp, Ile/Pro, and Glu/Phe. Full baseline separation was neces-sary, however, as the combustion of analytes into CO2and N2 prior to detection does not allow one to discrim-inate between overlapping analyte peaks. Derivatizationwith methyl-t-butyldimethylsilyl-trifluoroacetamide intot-butyldimethylsilyl derivatives finally led to full baselineseparation of the target analytes. Some of the seedlingsamples, however, did not provide an unambiguoussignal for Tyr, probably due to matrix overlay. Due to thefact that interference with background noise or matrixeffects leads to false results in isotope experiments(Wittmann, 2007), this amino acid was partly excludedfrom further interpretation. It also turned out that theaddition of dimethylformamide, commonly used toprocess microbial samples (Wittmann, 2007), was notcompatible with the GC-C-IRMS instrument used,because the solvent was largely transferred into thecombustion chamber. Accordingly, derivatization wasconducted without the addition of solvents. Tests withdifferent incubation times (30, 60, 90, and 120min) andtemperatures (80°C, 90°C, and 100°C) revealed thatthe combination of 60 min and 80°C provided theoptimum signal-to-noise ratio.

Isotope Distribution Patterns

Dynamic Incorporation of Carbon into Protein Amino Acids

The developed protocol was now used to quantify13C distribution patterns in rice related to amino acid

metabolism (Fig. 5). A first study with seedlingsrevealed that enrichment increased with time, as 13Cwas incorporated continuously into cell protein. Aminoacids differed strongly with regard to label incor-poration, which seemed to correlate to their biosyn-thetic origin. Particularly, Tyr and Phe, stemmingfrom intermediates of the Embden-Meyerhof-Parnas(EMP) pathway (phosphoenolpyruvate) and of the non-oxidative pentose phosphate (PP) pathway (erythrose4-phosphate), were enriched rather fast and to a greaterextent than all other amino acids and contributed up to9% and 15% of the entire enrichment, respectively (Fig.5A). Immediately after the pulse, significant 13C in-corporation also was observed for amino acids stem-ming from pyruvate, another intermediate of the EMPpathway (i.e. Ala, Ser, Gly, Val, and Leu). In contrast,labeling enrichment was delayed for amino acids syn-thesized from 2-oxoglutarate and oxaloacetate, in-termediates of the tricarboxylic acid cycle (Fig. 5A).Generally, highest enrichments were detected 24 h afterthe labeling pulse. After 48 h, the enrichment declinedfor most amino acids, probably due to a dilution with12CO2 taken up during the postlabeling incubation. The13C labeling profiles of Gly and Ser were similar, indi-cating that they originate from the same metabolicprecursor, 3-phosphoglycerate (Wittmann, 2007). Thesame trend also was found for other amino acid fami-lies, like the Asp family of amino acids, with Thr, Lys,Ile, and Asp itself. In comparison with rice seedlings,the general amino acid enrichment was, on average, 8times lower in the leaf of an adult plant during the lategrain-filling stage (Fig. 5A). This was interesting be-cause the total assimilation of carbon of both tissuetypes was in the same range (Fig. 4B; Supplemental Fig.S3) and matches perfectly with the changing role ofleaves from a strong sink during development to amajor source during reproduction, assimilating butnot incorporating carbon dioxide (Thrower, 1962;Turgeon, 1989). Low enrichment was equally observedfor normal leaves (60‰ 6 10‰ total amino acid en-richment), flag leaves (70‰ 6 10‰ total amino acidenrichment), and stem (40‰ 6 25‰ total amino acidenrichment) within 1 d of label translocation (Fig. 5B).In contrast, the panicle of adult rice plants was muchmore active regarding amino acid metabolic path-ways (280‰ 6 30‰ total amino acid enrichment).Here, highest enrichment was detected for Phe (39‰63‰) and Tyr (30‰ 6 2‰), followed by Ala (28‰ 64‰), Glu (23‰ 6 2‰), and Leu (22‰ 6 3‰), whichwas in accordance with the pattern detected for riceseedlings. The other amino acids showed lower enrich-ment (Fig. 5B).

Combined 13C and 15N Incorporation into ProteinAmino Acids

Amino acids of the shoot were analyzed for their13C and 15N enrichment from combined 13CO2 and15NH4NO3 labeling experiments with hydroponicallygrown rice seedlings (Fig. 6). For six selected amino







Figure 5. Incorporation of 13C into protein amino acids upon 13CO2 labeling. Enrichment of extracted amino acids was analyzed byGC-C-IRMS. A, Time-resolved pattern of a rice seedling (top, age of 12 d) and an adult rice plant at late grain filling (bottom, age of 88 d), forwhich soil-grown plantswere labeledwith 400mL L2113CO2 for 10min and then harvested directly or further cultivated at ambient air for2, 4, 24, and 48 h prior to harvesting. Statistical analysis was conducted using Student’s t test, whereby significant differences (P# 0.05)between seedling and adult leaves aremarkedwith asterisks. B, Tissue-resolved pattern of an adult rice plant at late grain filling, for whichplants were labeled with 400 mL L21 13CO2 for 10 min followed by 24 h of cultivation at ambient air prior to harvesting of leaf, flag leaf,stem, and panicle. Statistical analysis was done by one-way ANOVAwith Tukey’s test, whereby different letters (a, b, or c) indicate sig-nificant differences betweenmeans of the different tissues. C, Tissue-specific amino acidmetabolism, visualized as Venn diagrams, whichdisplay the relation of amino acids based on the statistical significance between measured 13C enrichments. Significant differences be-tweenmeans of amino acid 13C enrichments (P# 0.05) were determined by one-way ANOVAwith Tukey’s test. Amino acids that do notshowsignificantly different 13Cenrichment are located in equally colored ellipses. Theenrichment data areprovidedasmeanvalues (n=3)and reflect atomic percent excess corrected for natural isotopes. The full data sets are given in S4 to S6, respectively. To facilitate com-parison, the data are normalized for the highest enrichment of each data set, which was set to 100%, and visualized by a color codebetween yellow (0%) and red (100%). CIT, Citrate; E4P, erythrose 4-phosphate; F6P, Fru-6-P; FUM, fumarate; GAP, glyceraldehyde3-phosphate;G6P,Glc-6-P;MAL,malate;OAA, oxaloacetate; 2OG, 2-oxoglutarate; PEP, phosphoenolpyruvate; 3PG, 3-phosphoglycerate;PYR, pyruvate; R5P, ribose 5-phosphate; RU5P, ribulose 1,5-bisphosphate; SDL, seedling shoot; SUC, succinate.


Dersch et al.





acids (i.e. Ala, Gly, Pro, Ser, Asp, and Glu), root-baseddata with satisfying quality could be derived, whichallowed a tissue-specific examination, at least for thesemolecules (Fig. 6). Immediately after labeling, strong15N enrichment was detected for Ser, Asp, and Glu ofroot protein. At the same time, Ala, Gly, and Ser,extracted from shoot protein, revealed significant 13Cenrichment (Fig. 6). Within 2 h, labeled compoundswere distributed inside the plant, leading to combined13C- and 15N-enriched amino acids in shoot and root.Although all amino acids exhibited significantly dif-ferent 13C and 15N enrichment, the strongest differ-ences were found for shoot amino acids. Ala and Serexhibited higher 13C enrichment, whereas Pro, Asp,and Glu were more strongly enriched with 15N. Fourhours after the labeling pulse, similar amino acidenrichment patterns of shoot and root indicated theequilibration of label between these organs (Fig. 6).Amino acid enrichment patterns were similar for soil-grown and hydroponically grown rice seedlings(Supplemental Fig. S4), indicating that hydroponiccultivation did not significantly influence at least thispart of metabolism.

Mode-of-Action Analysis: Effect of Imazapyr Treatment onRice Seedlings

Herbicide treatment was used as a proof of concept todemonstrate the potential of the established technologyformode-of-action studies. Imazapyrwas chosen as awide-ly used herbicide, with a well-describedmechanism ofaction (i.e. the inhibition of acetolactate synthase inbranched-chain amino acid biosynthesis). The studieswere conducted using 12-d-old, soil-grown rice plants,which were exposed to imazapyr treatment or to acontrol treatment. Plants were pulse labeled with 13CO24 h after herbicide application (Fig. 7). At the time pointof the labeling experiment, no phenotypic alterations ofplantmorphology, as comparedwith the control plants,were observed (data not shown). The overall carbonassimilation at this time point was not disturbed, asindicated by equal 13C shoot enrichment of stressed andcontrol plants, namely 586‰ 6 26‰ and 595‰ 616‰, respectively (Fig. 7A). However, strongly dimin-ished label incorporation was detected in proteinogenicamino acids, which was particularly pronounced forthe branched-chain amino acids 2 h after the labelingpulse (Fig. 7B). Amino acid enrichment of stressedplants reached 19% to 75% of the enrichment detectedin control plants, except for the branched-chain aminoacids, which only reached 4% to 11%. On average, theenrichments were 2.7 times lower in imazapyr-treatedplants than in control plants, with the exception ofbranched-chain amino acids, which exhibited 7 to 26times less 13C enrichment. Observable phenotypic ef-fects became evident 7 d after imazapyr applicationand comprised inhibition of growth, chlorosis of above-ground plant parts, as well as a dieback of young leaves(Fig. 7C).

DISCUSSION

The quantitative analysis of pathway function andregulation is key to understanding and engineeringplant physiology (Maliga and Graham, 2004). Here, wedeveloped a cost-effective, high-throughput approachto assess whole-plant metabolic phenotypes in vivo. Asshown, we coupled 13CO2 and

15NH4NO3 pulse studiesunder precisely controlled physiological conditionswith ultra-precision GC-C-IRMS for parallel 13C and15N labeling analysis in different plant tissues and evenindividual molecular compounds to elucidate plantmetabolic traits. Different tube reactor layouts enabled13CO2 isotope experiments with soil-grown seedlingsand adult plants, up to a height of about 1 m (Fig. 1, Aand B), whereas parallel 13CO2 and

15NH4NO3 labelingwith hydroponic plant cultures could be conducted in aspecific box reactor (Fig. 1C). In contrast to previoustechniques (Tanaka and Osaki, 1983; Nouchi et al.,1995; Römisch-Margl et al., 2007), our methodologyminimizes potential alterations of the studied plantduring the experiment due to an atmosphere withambient CO2 levels, controlled temperature, humidity,and illumination, and incubation times of only 10 min.The careful experimental layout provided data withhigh precision and reproducibility among replicates,independent of the developmental stage of the ana-lyzed plants. Assimilation and translocation of carbonin different tissues (Figs. 3 and 4) and even individualmolecules (Fig. 5) could be quantified at deviationsbelow 20% for adult plants. Given the fact that plants inadvanced developmental stages are rather complex andsubject to a certain biological variation, due to differ-entiation into specific organs and cell types, the preci-sion and reproducibility can be regarded as excellent.For seedlings, they were partially even higher (Figs. 6and 7). This allowed us to accurately discriminate meta-bolic properties differing only slightly, which seems avaluable characteristic of our approach.

Assessment of Important Characteristics ofCarbon Metabolism

Beyond technical precision and accuracy, it was fur-ther important to validate the extent to which themethod could assess key properties of plant physiol-ogy. A catalog of experiments addressed this issue.Grain yield, the most prominent characteristic ofcrop plants, is strongly dependent on the source-sinkrelationship between plant parts (Kato et al., 2004),whereby the operational mode of a distinct plant organ,as source or sink, changes throughout development(Meng et al., 2013). Pioneering studies with rice plants,using radioactive 14CO2 at high dosage, had revealedthe translocation of assimilated carbon into the panicleduring grain filling (Cock and Yoshida, 1972). This wasquantitatively assessed here, however, by much saferand easier handling using stable 13CO2. Most of theassimilated carbon was recovered inside the rice pani-cle 24 h after the pulse (Figs. 4D and 5B). In line with






this, a 90% decrease of 13C in the assimilating tissue ofplants from the late grain-filling stage was detectedduring this chase period (Fig. 4D), which is in the rangeof values reported for other plants (Leake et al., 2006).During development from flowering to late grain fill-ing, the relative contribution of different leaves to car-bon assimilation probably corresponds to changingmetabolic requirements (Meng et al., 2013). During lategrain filling, the lower activity of carbon assimilationin mature leaves, relative to the flag leaf, indicatedmechanisms of senescence (Fig. 4). In senescing leaves,the Rubisco content and, as a consequence thereof,the photosynthetic activity decrease rapidly. Therefore,mature leaves cannot assimilate asmuchCO2 as youngerleaves (Makino et al., 1984). At night, 13C was furtherlost by respiration. Additional carbon loss might haveresulted from a combination of translocation to tissuesnot analyzed (e.g. the roots of plants grown on soil) anddilution of the label by overall plant biomass increaseduring the chase period (Leake et al., 2006).

Integrated Analysis of Carbon and Nitrogen Metabolism

The setup furthermore allowed combined labelingstudies with 13CO2 and

15NH4NO3, which are interest-ing due to the close connection of carbon and nitrogenmetabolism in plants but have been conducted only

rarely (Cliquet et al., 1990; Dyckmans et al., 2000). Thestudies highlighted fast interorgan distribution of car-bon and nitrogen in seedlings within 2 h after assimi-lation (Fig. 2), whereby the shoot operated as a majorsink and exhibited a 10- and 6-fold higher demand forcarbon and nitrogen compounds, respectively, com-pared with the root (Fig. 3A), which reflects sufficientsupply with water and nutrients, especially nitrogen(Wilson, 1988; Peuke et al., 1994). Obviously, the ex-amined seedlings received all essential macronutri-ents and micronutrients and were exposed to optimalbreeding conditions, concerning light, temperature,and humidity, so that carbon and nitrogen were usedmainly for shoot growth and development. The relativeflux maps for carbon and nitrogen, giving the relativedistribution among the different tissues on the basis ofisotope enrichment data, provide a straightforwardsnapshot of plant metabolism (Fig. 3). Given the avail-able instrumentation, such insights can be easily pro-vided within 1 d after an isotope experiment and thusallow for a fast evaluation (e.g. to study environmentalstresses as demonstrated for salt-stressed rice; Figs. 2and 3B). In light of breeding stress-tolerant plant lines,an important area of application is metabolic pheno-typing in order to identify protective mechanisms(Cramer et al., 2011) and changes in sink-source rela-tions (Roitsch, 1999; Albacete et al., 2014). Our datareflect these features for plants treatedwith salt, amajor

Figure 6. Integrated analysis of carbon and nitrogen metabolism by combined labeling of hydroponically grown rice seedlingswith 13CO2 and

15NH4NO3. Tissue-specific enrichment of extracted amino acids was quantified by GC-C-IRMS. Integrated viewson transport and biosynthetic routes of Ala, Gly, Ser, Glu/Gln, Asp/Asn, and Pro regarding assimilated 13C (light gray) and 15N(dark gray) labeling (A), assessed in shoot (B) and root (C), are represented by bar graphs over the first 4 h of tracing. The data reflectmeans6 SD (n = 3). At the age of 12 d, rice seedlingswere simultaneously labeledwith 400mL L2113CO2 and 1.43mM

15NH4NO3

for 10 min and then either harvested directly for assessment of label assimilation or cultivated further at ambient air for 2 to 4 h forassessment of label translocation. Asterisks indicate significant differences between mean values at P # 0.05 (Student’s t test).APE, Atomic percent excess; n.s., not significant. The full data set is given in Supplemental Table S7.


Dersch et al.




abiotic factor, particularly affecting the growth andproductivity of salt-sensitive crops (Roy et al., 2014).Salt stress resulted in a strong perturbation of nitrogenuptake by the root and transport to the shoot, whereascarbon assimilation and translocation were not signifi-cantly affected. Stressed plants kept relatively morenitrogen in the root as compared with nonstressedplants (Fig. 3), probably to support the supply withnitrogen-containing compatible solutes (Wang et al.,2012). Furthermore, enhanced retention of nitrogencompounds in the root might be due to adaptive mech-anisms to ensure adequate water and nutrient acquisi-tion from the rhizosphere (Sharp et al., 2004).

Labeling Dynamics of Amino Acids Provide QualitativeInsight into Metabolic Pathways

The coupling of C-IRMSwith GC separation, togetherwith the development of suitable protocols for samplingand sample processing, provided time-resolved enrich-ment data for protein-derived amino acids across differ-ent tissues of rice. Amino acids are the analytes of choicefor microbial pathway analysis, because they are muchmore abundant in cell extracts and protein than theirprecursors and provide extensive labeling information(Wittmann, 2007). On basis of the underlying biosyn-thetic precursor amino acid relationship, it is easy to de-duce the labeling patterns of the precursor metabolitesfrom labeling patterns of the corresponding amino acids.

The 13C labeling of amino acids from the same biosyn-thetic family was similar and obviously reflected theenrichment of their common precursor (Fig. 5A). In theseedling, amino acids stemming from precursors ofthe EMP pathway and the nonoxidative PP pathwayexhibited fast and strong 13C enrichment (Fig. 5A, top),examples being Phe, Tyr, Ser, Gly, and His. In contrast,Glu, Pro, Asp, and Thr, originating from tricarboxylicacid cycle-based precursors, were labeled much slowerand weaker. Although this does not directly providequantitative flux rates, the 13C enrichment data reveal afast and strong influx of 13C into the EMP pathway andthe nonoxidative PP pathway, indicating high activitiesof these routes. The rather low enrichment of tricarboxylicacid cycle-related amino acids most likely reflects adown-regulation of the cycle in the light (Tcherkez et al.,2005). In addition, the slow dynamics could be an indi-cation that Glu, Asp, and amino acids derived from themare not formed in the shoot but rather are synthesized inthe root, followed by transport into the shoot. This is in-deed taking place, which is discussed below on the basisof integrated 13C and 15N labeling data (Fig. 6). The 13Cenrichment in the shoot of a seedling was about 8- to10-fold higher than that in the leaf of an adult rice plant(Fig. 5A, bottom). Adult leaves assimilated the provided13CO2 to a high extent (Fig. 4) but incorporated only littleof it into protein. This matches with the changing role ofleaves from sink to source during reproduction, assimi-lating but not incorporating carbon dioxide (Thrower,1962; Turgeon, 1989).

Figure 7. A and B, Mode-of-action analysis of the effect of the herbicide imazapyr on rice seedlings using 13CO2 labeling incombination with EA-C-IRMS to assess carbon assimilation (A) and with GC-C-IRMS to assess 13C enrichment of extractedproteinogenic amino acids (B). The experimental setup contained seedlings at the age of 12 d subjected to imazapyr (62.5 g ha21

active ingredient) or control treatment (62.5 g ha21 control solution). Four hours after the treatment, seedlings were labeled for10 min with 400 mL L21 13CO2. The assimilation of 13C, determined immediately after the labeling pulse from freeze-dried plantmaterial, is expressed as d13C (‰), corrected for natural isotopes. The enrichment of extracted amino acids was determined after2 h of further cultivation at ambient air. The d13C values of the amino acids of imazapyr-treated rice seedlings were normalized tothose of the control seedlings. C, Phenotypes of rice seedlings 7 d after treatment with the control solution (left plant) and theimazapyr solution (right plant). Asterisks indicate significant differences between mean values of imazapyr-treated and controlplants at P # 0.05 (Student’s t test). n.s., Not significant. Mean values 6 SD (n = 3) are shown. The full data sets are given inSupplemental Tables S8 and S9.






Labeling Dynamics of Amino Acids IndicateTissue-Specific Metabolism

The 13C labeling of amino acids differed largelybetween individual tissues. For rice at late grain fill-ing, leaf, flag leaf, and stem incorporated only littlecarbon into protein (Fig. 5B). The panicle showed thehighest 13C enrichment among all tissues. Most as-similated carbon was hence translocated into thepanicle, where it was incorporated into proteinogenicamino acids. The panicle exhibited an active de novoprotein biosynthesis, reflecting that filling of thegrains results in a strong demand for protein precur-sors (Cock and Yoshida, 1972). A compound-orientedview, based on the statistical significance of labelingpatterns, groups amino acids into colored ellipsesaccording to their labeling similarity (Fig. 5C). Thisimmediately highlights specific metabolic finger-prints for all tissues of rice. Leaf and flag leaf differedonly in a few amino acids, whereas stem and paniclerevealed drastically altered patterns and evidencefor a tissue-specific carbon metabolism. This type ofvisualization further highlighted strong differencesbetween the carbon metabolism of an adult leaf and ayoung shoot.

Labeling Dynamics Provide Spatial Resolution of AminoAcid Metabolism

Combined 13C and 15N labeling of hydroponic riceseedlings showed an immediate occurrence of 15N-labeled Glu/Gln and Asp/Asn in the root, whereasthe 13C-labeled forms appeared much later (Fig. 6C).Obviously, these amino acids were formed fromcarbon precursors in the root through 15NH4 assimi-lation and transamination (Lam et al., 1995), whichexplains the exclusive enrichment with 15N (Fig. 6, Aand C). Subsequently, they were transported into theshoot (Funayama et al., 2013) to serve as amino groupdonors for the biosynthesis of other amino acids(Kiyomiya et al., 2001). Glu, with its rather high 15Nand low 13C enrichment, synthesized in the root andtransported to the shoot (Fig. 6B, 2-h time point), is theprecursor for the de novo synthesis of Pro, exhibiting asimilar labeling pattern.

The immediate 13C enrichment (Fig. 6B, 0-h timepoint) indicates a fast biosynthesis of Ala, Gly, andSer in the shoot. For Gly and Ser, this may be dueto photorespiration, which typically leads to highturnover rates for these amino acids (Gauthieret al., 2010). Overall, this provides a spatially re-solved picture of amino acid metabolism (Fig. 6),which explains the different dynamics of 13C and15N enrichment. From similar effects after simul-taneous labeling of rape (Brassica napus) plantswith 13C and 15N, it has been hypothesized that theincorporation of assimilated 13C into amino acidsis not tightly connected to nitrogen assimilation(Gauthier et al., 2010). This seems to also hold forrice.

The Response of Rice to the Herbicide ImazapyrComprises Impaired Biosynthesis of Branched-ChainAmino Acids and Enhanced Protein Turnover

Upon treatment with imazapyr, the biosynthesis ofVal, Leu, and Ile was strongly inhibited (Fig. 7B), ob-servable already 6 h after treatment. The overall carbonassimilation of the treated plant, however, was unaf-fected (Fig. 7A). This distinct phenotype perfectlymatches with the known mode of action of imazapyr.The herbicide is highly target specific. The enzymeacetohydroxyacid synthase, a key step in branched-chain amino acid biosynthesis, is the only site of ac-tion of this herbicide (Tan et al., 2005). In the longerterm, imazapyr causes a dysfunction of cell growth aswell as disruptions of DNA and protein biosynthesisand increased protein turnover to recycle branched-chain amino acids (Shaner and Reider, 1986; Scarponiet al., 1995; Royuela et al., 2000). As there was no morede novo synthesis, Val, Leu, and Ile exhibited lowestenrichment (3‰ 6 3‰ to 10‰6 3‰) among all aminoacids. In addition, our analysis demonstrates thatmetabolic changes at this early time point were not re-stricted to branched-chain amino acids as primary tar-gets. In fact, metabolism was affected much moreglobally. Newly assimilated 13C was incorporated intoprotein to a lesser extent in herbicide-treated plantsthan in control plants (Fig. 7B). Growth defects wereevident 7 d after imazapyr application (Fig. 7C), indi-cating a slow rate of plant death, which is related to theamount of intracellularly stored amino acids (Shanerand Singh, 1991). In this regard, our profiling toolboxappears valuable to study the mode-of-action of syn-thetic compounds at the initial level of plant pheno-typing, the first level of a three-tiered approach formode-of-action identification (Tresch, 2013).

CONCLUSION

Taken together, the developed approach is suitable toconduct stable isotope-labeling experiments for in vivowhole-plant metabolic profiling. The high technicaldata quality achievable under physiologically relevantconditions seems particularly promising to be appliedto (1) stress-response studies, (2) phenotypic screeningof plant lines, (3) mode-of-action studies, and (4) inte-grated analysis of carbon and nitrogen metabolism,among others. The underlying tracer studies requirelittle time and low effort compared with model-basedapproaches andare potent as a screening tool (Schwender,2008). Clearly, this type of analysis does not, per se,provide quantitative intracellular fluxes through in-dividual reactions and pathways, but it can suggestmetabolic activities (i.e. reallocation profiles or meta-bolic fingerprints). In this regard, our approach pro-vides a useful complementation to themore demandingapproaches of isotopically nonstationarymetabolic fluxanalysis (Szecowka et al., 2013; Ma et al., 2014). Thelatter offers the greatest potential for quantitative res-olution of metabolic fluxes in autotrophic plants but


Dersch et al.



involves extreme experimental and computational ef-fort, high cost, and low throughput to derive flux in-formation. In this regard, future combinations of bothtechnologies for two-tiered complementary analysiscould provide plant physiologists with a sophisti-cated toolbox for initial screening of several distinctphenotypes, thereby identifying the most promisingones for more comprehensive analysis. The high-precision GC-C-IRMS analysis, demonstrated herefor amino acids, might be easily widened to fattyacids (Meier-Augenstein, 2002; Richter et al., 2010;Panetta and Jahren, 2011), sterols (Jones et al., 1991),monosaccharides (Docherty et al., 2001; Derrien et al.,2003), and lignin (Goñi and Eglinton, 1996), whichare all accessible via such an instrumentation. Thispromises to extend the approach to other pathways ofprimary metabolism (e.g. lipid, sugar, and cell wallbiosynthesis) as well as to pathways of secondarymetabolism (e.g. involving phenol, isoprenoid, andalkaloid biosynthesis).

MATERIALS AND METHODS

Plants

Rice (Oryza sativa ssp. japonica ‘Nipponbare’) plants were obtained fromCropDesign.

Chemicals

As a tracer, 13CO2 (greater than 99 atom % 13C) was purchased from Eur-isotop. Labeled 15NH4NO3 (greater than 98 atom % 15N) was purchased fromSigma-Aldrich.

Media

Plants were grown either on soil (Einheitserde type GS90; 70% organic fiberpeat and 30% clay, pH 5.5–6; Einheitserde- und Humuswerke Gebr. Patzer) oron hydroponic medium (1.43 mM

15NH4NO3, 1 mM calcium chloride hexahy-drate, 0.18 mM magnesium sulfate heptahydrate, 1.32 mM potassium sulfate,0.32 mM monosodium phosphate, 1 mM ferric ethylendiaminetetraaceticacid, 8 mM manganese [II] chloride tetrahydrate, 0.15 mM zinc sulfate heptahy-drate, 0.15 mM copper [II] sulfate pentahydrate, 0.075 mM ammonium hepta-molybdate, and 1.39 mM boric acid; based on Yang et al., 1994). In isotopeexperiments, naturally labeled NH4NO3 was replaced by an equimolar amountof 15NH4NO3.

Plant Growth Conditions

Rice seedswere germinated onmoistfilter paper in a petri dish for 4 d at 26°Cin the dark. The seeds were transferred into light (500 mmol m22 s21 photo-synthetically active radiation) 1 d before they were transplanted either into0.7-dm3 pots used for further cultivation on soil or into hydroponic boxes. Priorto sowing, pots were soaked with deionized water containing 0.15% of thefungicide proplant (Stähler). Seeds were subjected to hot water treatment (60°Cfor 10 min) to prevent sheath rot. Rice plants were then grown under a 13/11-hday/night cycle at an average irradiance of 500 mmol m22 s21 photosyntheti-cally active radiation during the light phase (Powerstar HQI-BT 400W; Osram),a temperature cycle of 26°C/21°C, and a relative humidity of 60%. During thefirst 14 d of development, plants were irrigated two times per day withdeionized water. During the first 3 weeks, plants were watered via top irriga-tion and later via subirrigation. Between weeks 3 and 10, plants were fertilizedwith Hakaphos-Blau solution (0.3% [v/v] in deionized water; Compo) twiceper week, replacing one deionized water treatment, respectively. Plants indifferent developmental stages were used for 13CO2 labeling experiments.Rice seedlings were generally used for experimental purposes after 12 d of

cultivation. Flowering plants, as well as plants from the early and late grain-filling stages, were examined at 68, 75, and 88 d, respectively. For hydroponiccultures, plants were grown in hydroponic containers (27 3 17 3 12 cm) cov-ered with a perforated Styrofoam plate. Plastic meshes were placed inside theholes (2 cm diameter) on which pregerminated seedlings were cultivated withtheir roots freely suspended in the hydroponic medium. Rice seedlings weregenerally used for experimental purposes after 12 d of cultivation. The hy-droponic cultures were incubated under the same conditions regarding light,temperature, and humidity as soil-grown plants.

High-Salt Treatment

In salt stress experiments, hydroponically growing seedlings were subjectedto high-salt treatment for 6 d. For that purpose, 100 mM NaCl was added to thegrowth medium.

Imazapyr Treatment

Prior to the labeling experiment, soil-grown seedlings were subjected toimazapyr treatment, a nonselective herbicide of the imidazolinone group. Theimazapyr solution (0.3 mM imazapyr [BASF], 0.1% [v/v] dimethyl sulfoxide,and 0.1% [v/v] Dash E.C. [BASF]) was applied with an air brush to reflect adosage of 62.5 g ha21. Control plants were subjected to a solution containingonly the solvents used (0.1% [v/v] dimethyl sulfoxide and 0.1% [v/v] Dash E.C.). Labeling experiments were conducted 4 h after herbicide application.

Isotopic Labeling Experiments with 13CO2

The labeling experiments were conducted in specifically constructed label-ing reactors (Fig. 1). A large tube reactor (0.5mdiameter, 1.3mheight, and 255 Lvolume) was designed for 13CO2 isotope experiments with soil-grown adultplants (Fig. 1A), and a small tube reactor (0.5 m diameter, 0.5 m height, and 98 Lvolume) served for 13CO2 isotope experiments at the seedling stage (Fig. 1B).The tube reactors were built from Plexiglas acrylic sheets (5mm thickness; HansKeim Kunststoffe). The material allowed full spectral transmission of sunlightat wavelengths between 400 and 900 nm, containing the essential spectral in-terval for plant growth (Supplemental Fig. S1). Ethylene propylene dienemonomer rubber (Mercateo) was used as sealing material, because of its highflexibility and endurance. The rubber sealing was agglutinated to the Plexiglasor polycarbonate sheets with solvent-free glue (Loctite 406; Henkel). The tubewas conglutinated to the lid with a two-component adhesive (Pattex Stabilit;Henkel). For the maintenance of a constant temperature, a cooling system witha water-cooled ventilator (Peltier cooler/heater, 380 W, 24 V; Uwe Electronic)connected to an external cryostat (Lauda RMT20; Lauda Dr. R. Wobser) wasinstalled. The ventilator and the connectors were attached to the bottom plate ofthe reactor. To keep the humidity level at or above 60% of saturation during theexperiments, water-soaked cloth was placed in a glass beaker inside the en-closure. Temperature and relative humidity were monitored online by a hu-midity and temperature logger (Voltcraft DL-120 TH; Conrad Electronic).In addition, the CO2 concentration was measured (Voltcraft CM-100; ConradElectronic). All reactors allowed a precise adjustment of the 13CO2 level. For thispurpose, they contained two perforated pipes, one in the upper back part (ex-haust air) and one in the lower back part (supply air). These were connected to aCO2 adsorption unit, consisting of a high-power pump (6,000 L min21; Bravo2000, 220 V; Scoprega), a fine dust filter (Filter Cartridge A; Kärcher), and a CO2adsorber (5 L; Drägersorb 800+; Drägerwerk). All power supplies, except for theexternal absorber pump, had 24-V D.C. to allow safe handling. All labelingreactors were installed inside the phyto chamber (Svalöf), directly beside thebenches, where rice plants were grown at ambient air prior to and after thelabeling experiments. For the labeling studies, plants were placed insidethe enclosure. Hereby, the soil of the plant pots was covered with plastic wrapto avoid potential interference of CO2 respired from soil with the defined gasatmosphere. The ambient CO2was removed (less than 20mL L21) by adsorptionfor a short time period of about 30 s as described above. Immediately afterpurging, the desired amount of 13CO2 was injected through a valve (screw capwith silicone diaphragm), positioned in the lid of each enclosure, and the plantswere then incubated for a defined labeling period. Afterward, the plants or spe-cific parts were harvested directly or cultivated further in the phyto chamberunder ambient air prior to analysis. All harvested plants or plant tissues wereimmediately deep frozen in liquid nitrogen to stop metabolic activity. At eachsampling time point, at least three biological replicates were obtained. In thecase of seedling experiments, the whole shoot was harvested. Roots were






collected from hydroponically grown plantlets and treated equally. Of a full-grown plant, five tillers were harvested, whereby flag leaf, leaf, stem, and paniclewere separated, followed by immediate quenching in liquid nitrogen. The indi-vidual tissues originating from the five tillers were pooled. The outermost, pho-tosynthetically active leaves were removed from the stem prior to further treatment.Experiments were generally performed at 4 and 8 h after sunrise.

Combined 13C and 15N Labeling Experiment

Abox reactor (carbon/nitrogen reactor; 0.53 0.53 0.5m, 125 L volume)wasdeveloped for combined 13CO2 and

15NH4NO3 labeling with hydroponic ricecultures (Fig. 1C). Ammonium nitrate was selected because it generally servesas a nitrogen source and is taken up by the roots. Rice prefers ammonium overnitrate, as demonstrated by the higher uptake rate. Compared with nitrate, itsuptake is less dependent on light intensity and oxygen concentration at the rootsite (Sasakawa and Yamamoto, 1978). The reactor was built from polycarbonate(5 mm thickness; Hans Keim Kunststoffe), which allowed full spectral trans-mission of sunlight (Supplemental Fig. S1). Plants were pregrown in the phytochamber as hydroponic cultures. Prior to the experiment, cultures were trans-ferred from naturally labeled growth medium to a container with 15NH4NO3medium. The container was then placed inside the reactor on a perforated plate(polyvinylchloride, 5 mm; Hans Keim Kunststoffe), and the enclosure wasclosed tightly via fixation clamps. Immediately afterward, the 13CO2 pulse wasapplied, as described above. Monitoring and control of temperature, humidity,and 13CO2 level were done as described for the tube reactors (see above). At theend of the labeling incubation period, the box reactor was opened to ambient airand the plants were either harvested directly or retransferred into naturallylabeled growth medium after washing of the roots with naturally labeledgrowth medium. Sampling and sample processing were done as describedabove. Experiments were generally performed at 4 and 8 h after sunrise.

Quantification of Isotopic 13C and 15N Enrichment

Briefly,deep-frozenplantmaterialwas freezedried (ChristGamma2-16LSC;Martin Christ Gefriertrocknungsanlagen) and ground to a fine powder in a ballmill (3–5min, 30 Hz; Retsch MM300; Retsch). A sample (0.8–0.9 mg dry weight)was then transferred into a tin capsule (3.5 3 5 mm; HEKAtech). The isotopicenrichmentwas determined using an elemental analyzer (FLASH 2000; ThermoScientific) coupled to an isotope-ratio mass spectrometer (Delta V Plus; ThermoScientific). The gas flow was set to 300 mL min21, and the column temperaturewas kept at 41°C.

Quantification of Isotopic 13C and 15N Enrichment ofProteinogenic Amino Acids

Deep-frozen plant material was freeze dried (Christ Gamma 2-16 LSC;Martin Christ Gefriertrocknungsanlagen) as described previously (Hurkmanand Tanaka, 1986), with slight modifications as outlined below. Samples ofapproximately 5 mg dry weight were mixed with 400 mL of extraction buffer(0.175 M Tris-HCl, pH 8.8, 5% [w/v] SDS, 15% [v/v] glycerol, and 0.3 M dithi-othreitol), followed by centrifugation (13,000g, 10 min, and room temperature).The supernatant was mixed with 1.6 mL of ice-cold acetone and incubated forprecipitation of cell protein for 1 h at 220°C. After centrifugation (13,000g,10min, and 4°C), the protein pellet waswashed two timeswith 80% (v/v) acetone,air dried, and hydrolyzed into the amino acids for 12 h (125 mL, 6 M HCl, and100°C). The hydrolysate was purified (Millipore centrifugal filter units,Ultrafree-MC, Durapore polyvinylidene difluoride, 0.22 mm Merck KGaA).A volume of 50 mL of the hydrolysate was mixed with 50 mL of internal stan-dard (1.2 mM a-aminobutyric acid) and evaporated under a nitrogen stream.Amino acids were derivatized by the addition of 100 mL of N-methyl-N-t-butyldimethylsilyl-trifluoroacetamide (Macherey-Nagel) followed by incuba-tion for 1 h at 80°C. The isotopic composition of the amino acids was quantifiedby GC-C-IRMS (Trace GC Ultra gas chromatograph and Delta V Plus isotoperatio mass spectrometer; Thermo Scientific). A sample volume of 0.5 mL wasinjected via the PTV inlet (250°C) at a split ratio of 1:20. Helium was used as acarrier gas at a constant flow rate of 1 mL min21. Analytes were separated on afused silica capillary column (HP-5MS, 30 m 3 25 mm, 0.25 mm; Agilent) andthen transferred to an oxidizing combustion reactor (1,000°C). The initial oventemperature of 120°C was kept for 2 min. Afterward, the temperature was in-creased at a rate of 8°C min21 until 200°C. In a second gradient step, the tem-perature was raised to 300°C at 10°C min21 followed by a temperature hold for5 min. The GC-C-IRMS profile of the MBDSTFA-derivatized amino acids was

obtained in selective ionmonitoringmode via mass isomers of formed CO2 (m/z44, 45, and 46) and N2 (at m/z 28 and 29) at their corresponding retention times.Metabolite identification was achieved by gas chromatography-mass spec-trometry (7890A GC System, 7000 GC/MS Triple Quad, 7693 Autosampler;Agilent Technologies) measurements of single amino acids and their mixture(10 mM each) in full-scan acquisition mode, followed by a NIST mass spectralsearch (NIST MS search 2.0) and comparison of the chromatographic patternwith the one of the corresponding GC-C-IRMS measurement. Data acquisitionand evaluation were conducted using the software Isodat NT (Thermo Scien-tific). Analytical accurateness was a precondition for further data evaluation. Datapoints for which peak integrity was corrupted were omitted from graphical datarepresentation.

Calculation of 13C and 15N Enrichment

The d values (‰) are expressed relative to international standards as:

d ¼�Rsample 2Rstandard

Rstandard

�� 1000 ð1Þ

where R is the ratio of 13C/12C or 15N/14N. The 13C and 15Nmeasurements werecalibrated against Vienna Pee Dee Belemnite and atmospheric nitrogen, re-spectively, using the international reference materials cellulose (IAEA-CH-3)and ammonium sulfate (IAEA-N-1) as secondary standards. Typical values ford13C of unlabeled C3 plant material are between 234‰ and 222‰ (Meier-Augenstein, 1999a; Richter et al., 2010). Enrichment values also were ex-pressed as atomic percent (AP), which is the percentage of 13C or 15N atomsrelative to the total amount of carbon or nitrogen atoms in the sample. Atomicpercent enrichment was derived via Equation 2:

AP ¼ 100 � Rsample

1þ Rsampleð2Þ

Atomic percent excess (APE), the absolute value for isotopic enrichment, wascalculated by subtracting the enrichment of the control from the enrichment ofthe sample:

APE ¼ APsample 2APcontrol ð3Þ

Data Correction for Natural Isotopes

Naturally occurring isotopes of the analyte and of derivative groups wereconsidered (Wittmann, 2007), and data correction following GC-C-IRMSmeasurement was done as described previously (Metges and Daenzer, 2000;Docherty et al., 2001; Heinzle et al., 2008). Briefly, d values were normalized totheir respective unlabeled equivalents by subtracting the values of the unla-beled control from the corresponding value of the labeled sample:

dcorr ¼ dsample 2 dcontrol ð4Þwhere dcorr is the labeled compound, dsample is the derivatized labeled com-pound, and dcontrol is the derivatized unlabeled compound. For each experi-ment, unlabeled control plants that were harvested (at least three biologicalreplicates) were exposed to the exact same breeding conditions as the labeledplants (i.e. they were grown on the same batch of soil or hydroponic medium,respectively). Furthermore, unlabeled controls were subjected to the samesample-processing steps as labeled samples.

For the amino acid pathway illustrations (Fig. 5, A and B), the corrected d

values were normalized to 100%. Therefore, the highest dcorr value of the re-spective data set was set to 100%. The normalized value of any number x (Xnorm)of the original data set was calculated by the following equation:

Xnorm ¼ x � 100dmax

ð5Þ

where dmax was the highest dcorr value.

Determination of Root-to-Shoot Ratio andTranslocation Flux

To set up a fluxmap for a rice seedling, it was important to derive the root-to-shoot ratio of the analyzed plantlets. It was assumed that the decrease of


Dersch et al.




15N enrichment in the root within 2 h of tracing is only due to the translocationof label to the shoot. With this assumption, the root-to-shoot ratio was derivedvia the following formula:

Mshoot� �d15Nshoot;t2 2 d15Nshoot;t0

� ¼ Mroot� �d15Nroot;t0 2 d15Nroot;t2

� ð6ÞwhereM is the mass of shoot and root as indicated. A root:shoot ratio of 1:6 wasobtained for 12-d-old seedlings. This value was used to estimate the percentageenrichment of individual plant parts after 2 h of tracing. Considering that shootbiomass was 6 times as much as root biomass, it follows that 15N enrichment ofthe shoot was diluted by a factor of 6 upon translocation to the shoot, whereas13C enrichment accumulated 6-fold upon translocation to the root. Accordingly,d15N values of the shoot were multiplied by 6, whereas d13C values of the rootwere divided by 6. Shoot and root enrichment values of 13C as well as 15Nlabeling, at time point zero, were summed, displaying 100% uptake. On thisbasis, 13C and 15N percentage enrichments were estimated for root and shoot.The amount of 13C labeling that could not be recovered in plant tissue wasassumed to be lost by respiration.

Statistical Analysis

The statistical significance of differences between mean values was deter-mined using Student’s t test or one-wayANOVA followed by a posthoc Tukey’stest. Differences were considered significant when the P value was below 0.05.

Supplemental Data

The following supplemental materials are available.

Supplemental Figure S1. Absorption spectra of material used for the con-struction of labeling reactors.

Supplemental Figure S2. Profiles of temperature and relative humidityinside the different reactor housings.

Supplemental Figure S3. Influence of magnitude and duration of 13CO2labeling pulse, number of plants in the enclosure, and daytime on theenrichment of seedlings.

Supplemental Figure S4. Simplified pathway illustration of shoot aminoacid 13C enrichment of rice seedlings raised on soil compared with hy-droponic culture.

Supplemental Table S1. Raw data of Figure 2 and Figure 3, comparing 13Cwith 15N labeling of seedlings.

Supplemental Table S2. Raw data of Figure 4B, comparing 13C labeling ofplants in the flowering stage with plants in grain filling early and grainfilling late.

Supplemental Table S3. Raw data of Figure 4D, comparing 13C labeling ofthe different tissue types of plants in the late grain filling stage.

Supplemental Table S4. Raw data of Figure 5A, comparing 13C labeling ofa seedling shoot with the leaf of a plant in the late grain filling stage.

Supplemental Table S5. Raw data of Figure 5B, comparing 13C labeling ofthe different tissue types of a plant in the late grain filling stage.

Supplemental Table S6. Raw data of Figure 5C, statistical significancebetween measured amino acid enrichments.

Supplemental Table S7. Raw data of Figure 6, comparing shoot and root13C and 15N labeling.

Supplemental Table S8. Raw data of Figure 7A, displaying 13C enrichmentof imazapyr-treated rice seedlings, immediately after the labeling pulse.

Supplemental Table S9. Raw data of Figure 7B, displaying 13C enrichmentof imazapyr-treated rice seedlings after a two-hour chase period.

Supplemental Table S10. Raw data of Figure S2, displaying means (n 5 3)of temperature and humidity during the experiment inside the differentlabeling reactors.

Supplemental Table S11. Raw data of Figure S3, validation of labelingsystem.

Supplemental Table S12. Raw data of Figure S4, comparing 13C labeling ofseedlings grown on soil with seedlings grown on hydroponic medium.

ACKNOWLEDGMENTS

We thank Enrico Peter for input regarding experimental design and trou-bleshooting; Olaf Woiwode for cultivation of plants, maintenance of equip-ment, and assistance with the preparation and execution of experiments; as wellas Jürgen Kastler and René Müller for C-IRMS analyses and data evaluation.

ReceivedDecember 11, 2015; acceptedMarch 9, 2016; publishedMarch 10, 2016.

LITERATURE CITED

Albacete AA, Martínez-Andújar C, Pérez-Alfocea F (2014) Hormonaland metabolic regulation of source-sink relations under salinity anddrought: from plant survival to crop yield stability. Biotechnol Adv 32:12–30

Allen DK, Libourel IGL, Shachar-Hill Y (2009) Metabolic flux analysis inplants: coping with complexity. Plant Cell Environ 32: 1241–1257

Arp WJ (1991) Effects of source-sink relations on photosynthetic acclima-tion to elevated CO2. Plant Cell Environ 14: 869–875

Cegelski L, Schaefer J (2005) Glycine metabolism in intact leaves by in vivo13C and 15N labeling. J Biol Chem 280: 39238–39245

Chen WP, Yang XY, Harms GL, Gray WM, Hegeman AD, Cohen JD (2011)An automated growth enclosure for metabolic labeling of Arabidopsisthaliana with 13C-carbon dioxide: an in vivo labeling system for proteo-mics and metabolomics research. Proteome Sci 9: 9–23

Cliquet JB, Deléens E, Mariotti A (1990) C and N mobilization from stalkand leaves during kernel filling by C and N tracing in Zea mays L. PlantPhysiol 94: 1547–1553

Cock JH, Yoshida S (1972) Accumulation of 14C carbohydrate beforeflowering and its subsequent redistribution and respiration in the riceplant. Proc Crop Sci Soc Japan 41: 226–234

Cramer GR, Urano K, Delrot S, Pezzotti M, Shinozaki K (2011) Effects ofabiotic stress on plants: a systems biology perspective. BMC Plant Biol11: 163

Derrien D, Balesdent J, Marol C, Santaella C (2003) Measurement of the13C/12C ratio of soil-plant individual sugars by gas chromatography/combustion/isotope-ratio mass spectrometry of silylated derivatives.Rapid Commun Mass Spectrom 17: 2626–2631

Derrien D, Marol C, Balesdent J (2004) The dynamics of neutral sugars inthe rhizosphere of wheat: an approach by 13C pulse-labelling and GC/C/IRMS. Plant Soil 267: 243–253

Docherty G, Jones V, Evershed RP (2001) Practical and theoretical con-siderations in the gas chromatography/combustion/isotope ratio massspectrometry delta13C analysis of small polyfunctional compounds.Rapid Commun Mass Spectrom 15: 730–738

Dyckmans J, Flessa H, Shangguan Z, Beese F (2000) A dual 13C and 15Nlong-term labelling technique to investigate uptake and translocation ofC and N in beech (Fagus sylvatica L.). Isotopes Environ Health Stud 36:63–78

Fischer R, Emans N (2000) Molecular farming of pharmaceutical proteins.Transgenic Res 9: 279–299

Funayama K, Kojima S, Tabuchi-Kobayashi M, Sawa Y, Nakayama Y,Hayakawa T, Yamaya T (2013) Cytosolic glutamine synthetase1;2 isresponsible for the primary assimilation of ammonium in rice roots.Plant Cell Physiol 54: 934–943

Gauthier PPG, Bligny R, Gout E, Mahé A, Nogués S, Hodges M, TcherkezGGB (2010) In folio isotopic tracing demonstrates that nitrogen assimilationinto glutamate is mostly independent from current CO2 assimilation in illu-minated leaves of Brassica napus. New Phytol 185: 988–999

Goñi MA, Eglinton TI (1996) Stable carbon isotopic analyses of lignin-derived CuO oxidation products by isotope ratio monitoring-gas chromatography-mass-spectrometer (irm-GC-MS). Org Geochem24: 601–615

Griffiths RI, Manefield M, Ostle N, McNamara N, O’Donnell AG, BaileyMJ, Whiteley AS (2004) 13CO2 pulse labelling of plants in tandemwith stable isotope probing: methodological considerations for ex-amining microbial function in the rhizosphere. J Microbiol Methods58: 119–129

Hasunuma T, Harada K, Miyazawa S, Kondo A, Fukusaki E, Miyake C (2010)Metabolic turnover analysis by a combination of in vivo 13C-labelling from13CO2 and metabolic profiling with CE-MS/MS reveals rate-limiting stepsof the C3 photosynthetic pathway in Nicotiana tabacum leaves. J Exp Bot61: 1041–1051
























Heinzle E, Yuan Y, Kumar S, Wittmann C, Gehre M, Richnow HH,Wehrung P, Adam P, Albrecht P (2008) Analysis of 13C labeling en-richment in microbial culture applying metabolic tracer experimentsusing gas chromatography-combustion-isotope ratio mass spectrome-try. Anal Biochem 380: 202–210

Hurkman WJ, Tanaka CK (1986) Solubilization of plant membrane pro-teins for analysis by two-dimensional gel electrophoresis. Plant Physiol81: 802–806

Hutchinson CR, Stephen MT, Hsia S, Carver RA (1976) Biosyntheticstudies with 13CO2 of secondary plant metabolites: Nicotinana alkaloids.1. Initial experiments. J Am Chem Soc 98: 6006–6011

Jones D, Carter J, Eglinton G, Jumeau E, Fenwick C (1991) Determinationof d13C values of sedimentary straight chain and cyclic alcohols by gaschromatography/isotope ratio mass spectrometry. Biol Mass Spectrom20: 641–646

Junker BH, Lonien J, Heady LE, Rogers A, Schwender J (2007) Paralleldetermination of enzyme activities and in vivo fluxes in Brassica napusembryos grown on organic or inorganic nitrogen source. Phytochemis-try 68: 2232–2242

Kato M, Kobayashi K, Ogiso E, Yokoo M (2004) Photosynthesis and dry-matter production during ripening stage in a female-sterile line of rice.Plant Prod Sci 7: 184–188

Khush G (2003) Productivity improvements in rice. Nutr Rev 61: 114–116Kiyomiya S, Nakanishi H, Uchida H, Tsuji A, Nishiyama S, Futatsubashi

M, Tsukada H, Ishioka NS, Watanabe S, Ito T, et al (2001) Real timevisualization of 13N-translocation in rice under different environmentalconditions using positron emitting tracer imaging system. Plant Physiol125: 1743–1753

Kruger NJ, Huddleston JE, Le Lay P, Brown ND, Ratcliffe RG (2007)Network flux analysis: impact of 13C-substrates on metabolism inArabidopsis thaliana cell suspension cultures. Phytochemistry 68: 2176–2188

Kruger NJ, Ratcliffe RG (2009) Insights into plant metabolic networks fromsteady-state metabolic flux analysis. Biochimie 91: 697–702

Lam H, Coschigano K, Schultz C, Melo-Oliveira R, Tjaden G, Ngai N,Hsieh M, Coruni G (1995) Use of Arabidopsis mutants and genes tostudy amide amino acid biosynthesis. Plant Cell 7: 887–898

Lattanzi FA, Ostler U, Wild M, Morvan-Bertrand A, Decau ML, LehmeierCA, Meuriot F, Prud’homme MP, Schäufele R, Schnyder H (2012)Fluxes in central carbohydrate metabolism of source leaves in a fructan-storing C3 grass: rapid turnover and futile cycling of sucrose in con-tinuous light under contrasted nitrogen nutrition status. J Exp Bot 63:2363–2375

Leake JR, Ostle NJ, Rangel-Castro JI, Johnson D (2006) Carbon fluxesfrom plants through soil organisms determined by field 13CO2 pulse-labelling in an upland grassland. Appl Soil Ecol 33: 152–175

Ma F, Jazmin LJ, Young JD, Allen DK (2014) Isotopically nonstationary13C flux analysis of changes in Arabidopsis thaliana leaf metabolism dueto high light acclimation. Proc Natl Acad Sci USA 111: 16967–16972

Makino A, Mae T, Ohira K (1984) Relation between nitrogen and ribulose-1,5-bisphosphate carboxylase in rice leaves from emergence throughsenescence. Plant Cell Physiol 25: 429–437

Maliga P, Graham I (2004) Molecular farming and metabolic engineeringpromise a new generation of high-tech crops. Curr Opin Plant Biol 7:149–151

Meier-Augenstein W (1999a) Applied gas chromatography coupled toisotope ratio mass spectrometry. J Chromatogr A 842: 351–371

Meier-Augenstein W (1999b) Use of gas chromatography-combustion-isotope ratio mass spectrometry in nutrition and metabolic research.Curr Opin Clin Nutr Metab Care 2: 465–470

Meier-Augenstein W (2002) Stable isotope analysis of fatty acids by gaschromatography-isotope ratio mass spectrometry. Anal Chim Acta 465:63–79

Meng F, Dungait JAJ, Zhang X, He M, Guo Y, Wu W (2013) Investigationof photosynthate-C allocation 27 days after 13C-pulse labeling of Zeamays L. at different growth stages. Plant Soil 373: 755–764

Metges CC, Daenzer M (2000) 13C gas chromatography-combustion iso-tope ratio mass spectrometry analysis of N-pivaloyl amino acid esters oftissue and plasma samples. Anal Biochem 278: 156–164

Molero G, Aranjuelo I, Teixidor P, Araus JL, Nogués S (2011) Measure-ment of 13C and 15N isotope labeling by gas chromatography/combustion/isotope ratio mass spectrometry to study amino acid fluxes in a plant-microbe symbiotic association. Rapid Commun Mass Spectrom 25: 599–607

Nogués S, Tcherkez G, Cornic G, Ghashghaie J (2004) Respiratory carbonmetabolism following illumination in intact French bean leaves using13C/12C isotope labeling. Plant Physiol 136: 3245–3254

Nouchi I, Ito O, Harazono Y, Kouchi H (1995) Acceleration of 13C-labelledphotosynthate partitioning from leaves to panicles in rice plants ex-posed to chronic ozone at the reproductive stage. Environ Pollut 88:253–260

Olsson PA, van Aarle IM, Gavito ME, Bengtson P, Bengtsson G (2005) 13Cincorporation into signature fatty acids as an assay for carbon allocationin arbuscular mycorrhiza. Appl Environ Microbiol 71: 2592–2599

Panetta RJ, Jahren AH (2011) Single-step transesterification with simulta-neous concentration and stable isotope analysis of fatty acid methylesters by gas chromatography-combustion-isotope ratio mass spectrometry.Rapid Commun Mass Spectrom 25: 1373–1381

Peuke AD, Hartung W, Jeschke WD (1994) The uptake and flow of C, Nand ions between roots and shoots in Ricinus communis L. II. Grown withlow or high nitrate supply. J Exp Bot 45: 733–740

Ratcliffe RG, Shachar-Hill Y (2006) Measuring multiple fluxes throughplant metabolic networks. Plant J 45: 490–511

Rawson BHM, Hofstra G (1969) Translocation and remobilization of 14Cassimilated at different stages by each leaf of the wheat plant. Aust J BiolSci 22: 321–331

Richter EK, Spangenberg JE, Kreuzer M, Leiber F (2010) Characterizationof rapeseed (Brassica napus) oils by bulk C, O, H, and fatty acid C stableisotope analyses. J Agric Food Chem 58: 8048–8055

Roessner-Tunali U, Liu J, Leisse A, Balbo I, Perez-Melis A, Willmitzer L,Fernie AR (2004) Kinetics of labelling of organic and amino acids inpotato tubers by gas chromatography-mass spectrometry following in-cubation in 13C labelled isotopes. Plant J 39: 668–679

Roitsch T (1999) Source-sink regulation by sugar and stress. Curr OpinPlant Biol 2: 198–206

Römisch-Margl W, Schramek N, Radykewicz T, Ettenhuber C, Eylert E,Huber C, Römisch-Margl L, Schwarz C, Dobner M, Demmel N, et al(2007) 13CO2 as a universal metabolic tracer in isotopologue perturbationexperiments. Phytochemistry 68: 2273–2289

Rontein D, Dieuaide-Noubhani M, Dufourc EJ, Raymond P, Rolin D(2002) The metabolic architecture of plant cells: stability of central me-tabolism and flexibility of anabolic pathways during the growth cycle oftomato cells. J Biol Chem 277: 43948–43960

Roy SJ, Negrão S, Tester M (2014) Salt resistant crop plants. Curr OpinBiotechnol 26: 115–124

Royuela M, Gonzalez A, Gonzalez EM, Arrese-Igor C, Aparicio-TejoPM, Gonzalez-Murua C (2000) Physiological consequences of con-tinuous, sublethal imazethapyr supply to pea plants. J Plant Physiol157: 345–354

Sasakawa H, Yamamoto Y (1978) Comparison of the uptake of nitrateand ammonium by rice seedlings: influences of light, temperature,oxygen concentration, exogenous sucrose, and metabolic inhibitors.Plant Physiol 62: 665–669

Scarponi L, Alla MMN, Martinetti L (1995) Consequences on nitrogenmetabolism in soybean (Glycine max L.) as a result of Imazethapyr ac-tion on acetohydroxy acid synthase. J Agric Food Chem 43: 809–814

Schaefer J, Kier LD, Stejskal EO (1980) Characterization of photorespir-ation in intact leaves using carbon dioxide labeling. Plant Physiol 65:254–259

Schwender J (2008) Metabolic flux analysis as a tool in metabolic engi-neering of plants. Curr Opin Biotechnol 19: 131–137

Schwender J, Ohlrogge JB, Shachar-Hill Y (2003) A flux model of gly-colysis and the oxidative pentosephosphate pathway in developingBrassica napus embryos. J Biol Chem 278: 29442–29453

Shachar-Hill Y (2013) Metabolic network flux analysis for engineeringplant systems. Curr Opin Biotechnol 24: 247–255

Shaner DL, Reider ML (1986) Physiological responses of corn (Zea mays) toAC 243,997 in combination with valine, leucine, and isoleucine. PestBiochem Physiol 25: 248–257

Shaner DL, Singh B (1991) Imidazolinone-induced loss of acetohydrox-yacid synthase activity in maize is not due to the enzyme degradation.Plant Physiol 97: 1339–1341

Sharma AK, Sharma MK (2009) Plants as bioreactors: recent developmentsand emerging opportunities. Biotechnol Adv 27: 811–832

Sharp RE, Poroyko V, Hejlek LG, Spollen WG, Springer GK, Bohnert HJ,Nguyen HT (2004) Root growth maintenance during water deficits:physiology to functional genomics. J Exp Bot 55: 2343–2351


Dersch et al.



Szecowka M, Heise R, Tohge T, Nunes-Nesi A, Vosloh D, Huege J, Feil R,Lunn J, Nikoloski Z, Stitt M, et al (2013) Metabolic fluxes in an illu-minated Arabidopsis rosette. Plant Cell 25: 694–714

Tan S, Evans RR, Dahmer ML, Singh BK, Shaner DL (2005) Imidazolinone-tolerant crops: history, current status and future. Pest Manag Sci 61:246–257

Tanaka A, Osaki M (1983) Growth and behavior of photosynthesized 14Cin various crops in relation to productivity. Soil Sci Plant Nutr 29:147–158

Tanaka A, Yamaguchi J (1968) The growth efficiency in relation to thegrowth of the rice plant. Soil Sci Plant Nutr 14: 110–116

Tcherkez G, Cornic G, Bligny R, Gout E, Ghashghaie J (2005) In vivorespiratory metabolism of illuminated leaves. Plant Physiol 138: 1596–1606

Tcherkez G, Nogués S, Bleton J, Cornic G, Badeck F, Ghashghaie J (2003)Metabolic origin of carbon isotope composition of leaf dark-respiredCO2 in French bean. Plant Physiol 131: 237–244

Thrower SL (1962) Translocation of labelled assimilates in the soybean.Aust J Biol Sci 15: 629–649

Tresch S (2013) Strategies and future trends to identify the mode of actionof phytotoxic compounds. Plant Sci 212: 60–71

Turgeon R (1989) The sink-source transition in leaves. Annu Rev PlantPhysiol Plant Mol Biol 40: 119–138

Wang H, Zhang M, Guo R, Shi D, Liu B, Lin X, Yang C (2012) Effects ofsalt stress on ion balance and nitrogen metabolism of old and youngleaves in rice (Oryza sativa L.). BMC Plant Biol 12: 194

Williams TCR, Miguet L, Masakapalli SK, Kruger NJ, Sweetlove LJ,Ratcliffe RG (2008) Metabolic network fluxes in heterotrophic Arabidopsis

cells: stability of the flux distribution under different oxygenation conditions.Plant Physiol 148: 704–718

Wilson JB (1988) A review of evidence on the control of shoot:root ratio, inrelation to models. Ann Bot (Lond) 61: 433–449

Wittmann C (2007) Fluxome analysis using GC-MS. Microb Cell Fact 6: 6Wu WX, Liu W, Lu HH, Chen YX, Medha D, Janice T (2009) Use of 13C

labeling to assess carbon partitioning in transgenic and nontransgenic(parental) rice and their rhizosphere soil microbial communities. FEMSMicrobiol Ecol 67: 93–102

Yang X, Römheld V, Marschner H (1994) Effect of bicarbonate on root growthand accumulation of organic acids in Zn-inefficient and Zn-efficient ricecultivars (Oryza sativa L.). Plant and Soil 164: 1–7

Yoshida S (1981) Fundamentals of Rice Crop Science. International RiceResearch Institute. Los Baños, Philippines

Young JD, Shastri AA, Stephanopoulos G, Morgan JA (2011) Mappingphotoautotrophic metabolism with isotopically nonstationary 13C fluxanalysis. Metab Eng 13: 656–665

Yousfi S, Serret MD, Araus JL (2013) Comparative response of d13C, d18Oand d15N in durum wheat exposed to salinity at the vegetative andreproductive stages. Plant Cell Environ 36: 1214–1227

Yousfi S, Serret MD, Márquez AJ, Voltas J, Araus JL (2012) Combined useof d¹³C, d18O and d15N tracks nitrogen metabolism and genotypic ad-aptation of durum wheat to salinity and water deficit. New Phytol 194:230–244

Zhu C, Zhu J, Cao J, Jiang Q, Liu G, Ziska LH (2014) Biochemical andmolecular characteristics of leaf photosynthesis and relative seed yieldof two contrasting rice cultivars in response to elevated [CO₂]. J Exp Bot65: 6049–6056





Documents

Novel Approach for High-Throughput Metabolic Screening of ... · ensure the global food supply (Khush, 2003) and to produce various chemicals and materials (Fischer and Emans, 2000;