Computers & Electronics

A roadmap for interpreting 13C metabolite labeling patterns from cells

Description
A roadmap for interpreting 13C metabolite labeling patterns from cells
Published
of 13
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
   Aroadmapforinterpreting  13 C   metabolitelabelingpatternsfromcells Joerg   M   Buescher 1,2 ,Maciek   RAntoniewicz 3,* ,LaszloG   Boros 4,* ,Shawn   C   Burgess 5,* ,Henri   Brunengraber 6,* ,Clary   B   Clish 7,* ,Ralph   J   DeBerardinis 8,* ,Olivier   Feron 9,* ,   Christian   Frezza 10,* ,Bart   Ghesquiere 1,2,* ,EyalGottlieb 11,* ,Karsten   Hiller 12,* ,Russell   G   Jones 13,* ,   JurreJ   Kamphorst 14,* ,   Richard   GKibbey 15,* , Alec   C   Kimmelman 16,* ,Jason   WLocasale 17,* ,Sophia   Y   Lunt 18,* ,Oliver   DKMaddocks 11,* ,   CraigMalloy 19,* ,   Christian   MMetallo 20,* ,Emmanuelle   J   Meuillet 21,22,* ,Joshua   Munger 23,24,* ,Katharina   No ¨ h 25,* ,JoshuaD   Rabinowitz 26,* ,   Markus   Ralser 27,28,* ,Uwe   Sauer 29,* ,   GregoryStephanopoulos 30,* ,JulieSt-Pierre 31,* ,Daniel    A    Tennant 32,* ,   Christoph   Wittmann 33,* ,Matthew   GVanderHeiden 34,35,* ,Alexei   Vazquez 11,* ,Karen   Vousden 11,* ,   Jamey   DYoung 36,37,* ,   Nicola   Zamboni 29,* andSarah-Maria   Fendt 1,2 Measuring   intracellular   metabolism   has   increasingly   led   toimportant   insights   in   biomedical   research.  13 Ctracer   analysis,although   less   information-rich   than   quantitative  13 C   fluxanalysis   that   requires   computational   data   integration,   has   beenestablished   as   atime-efficient   method   to   unravel   relativepathway   activities,   qualitative   changes   in   pathwaycontributions,   and   nutrient   contributions.   Here,   we   reviewselected   key   issues   in   interpreting  13 C   metabolite   labelingpatterns,   with   the   goal   ofdrawing   accurate   conclusions   fromsteady   state   and   dynamic   stable   isotopic   tracer   experiments.  Addresses 1 VesaliusResearchCenter,VIB,Leuven,Belgium 2 DepartmentofOncology,KU   Leuven,Leuven,Belgium 3 DepartmentofChemicalandBiomolecularEngineering,Universityof Delaware,Newark,DE,USA  4 DepartmentofPediatrics,UCLASchool   ofMedicine,LosAngelesBiomedicalResearchInstituteat   the   Harbor-UCLAMedicalCenterandSidmap,LLC,LosAngeles,CA,USA  5  AdvancedImagingResearchCenter-DivisionofMetabolicMechanismsofDiseaseandDepartmentofPharmacology,TheUniversityofTexasSouthwesternMedicalCenter,Dallas,TX,USA  6 DepartmentofNutrition,   CaseWesternReserveUniversitySchoolof Medicine,Cleveland,OH,   USA  7 BroadInstituteofHarvardand   MIT,   Cambridge,MA,USA  8 Children’sMedicalCenterResearchInstitute,UT   SouthwesternMedicalCenter,Dallas,TX,USA  9 PoleofPharmacologyandTherapeutics(FATH),InstitutdeRechercheExpe´ rimentaleetClinique(IREC),Universite´  catholiquedeLouvain,Brussels,Belgium 10 MRCCancerUnit,   UniversityofCambridge,Hutchison/MRCResearchCentre,CambridgeBiomedicalCampus,Cambridge,UK 11 CancerResearchUK,BeatsonInstitute,Glasgow,UK 12 LuxembourgCentreforSystemsBiomedicine,Universityof Luxembourg,Esch-Belval,Luxembourg 13 GoodmanCancerResearchCentre,   DepartmentofPhysiology,McGillUniversity,Montreal,QC,Canada 14 InstituteofCancerSciences,UniversityofGlasgow,Glasgow,UK 15 InternalMedicine,Cellularand   MolecularPhysiology,YaleUniversitySchoolofMedicine,NewHaven,CT,USA  16 DivisionofGenomicStabilityandDNARepair,   Departmentof RadiationOncology,Dana-FarberCancer   Institute,Boston,MA,USA  17 DivisionofNutritionalSciences,CornellUniversity,Ithaca,NY,USA  18 DepartmentofBiochemistryandMolecularBiology,MichiganStateUniversity,EastLansing,MI,USA  19  AdvancedImagingResearchCenter-DivisionofMetabolicMechanismsofDiseaseandDepartmentof    Radiology,TheUniversityof TexasSouthwesternMedicalCenter,Dallas,TX,USA  20 DepartmentofBioengineering,UniversityofCalifornia,SanDiego,   LaJolla,CA,   USA  21 L’InstitutdesTechnologiesAvance´ esenSciencesduVivant(ITAV),ToulouseCedex1,France 22 TheUniversityof     ArizonaCancer   Center,and   Departmentof NutritionalSciences,TheUniversityofArizona,Tucson,AZ,   USA  23 DepartmentofBiochemistry,UniversityofRochesterMedicalCenter,Rochester,NY,USA  24 DepartmentofBiophysics,Universityof    RochesterMedicalCenter,Rochester,NY,USA  25 InstituteofBio-andGeosciences,IBG-1:Biotechnology,ForschungszentrumJu ¨ lichGmbH,Ju ¨ lich,Germany 26 DepartmentofChemistryandLewis – SiglerInstitutefor   IntegrativeGenomics,PrincetonUniversity,Princeton,NJ,USA  27 CambridgeSystemsBiologyCentreandDepartmentofBiochemistry,UniversityofCambridge,Cambridge,UK 28 DivisionofPhysiologyandMetabolism,MRC   NationalInstituteforMedicalResearch,London,UK 29 InstituteofMolecularSystemsBiology,ETHZurich,Zurich,Switzerland 30 DepartmentofChemicalEngineering,MassachusettsInstitute   of Technology,Cambridge,MA,USA  31 GoodmanCancerResearchCentre,andDepartmentofBiochemistry,McGillUniversity,Montreal,Quebec,Canada  Available   online   at   www.sciencedirect.com ScienceDirect  www.sciencedirect.com CurrentOpinioninBiotechnology  2015, 34 :189 – 201  32 SchoolofCancerSciences,CollegeofMedicalandDentalSciences,UniversityofBirmingham,Edgbaston,Birmingham,UK 33 InstituteofSystemsBiotechnology,SaarlandUniversity,Saarbru ¨ cken,Germany 34 KochInstitutefor   IntegrativeCancerResearchatMassachusettsInstituteofTechnology,BroadInstituteofHarvardandMIT,Cambridge,MA,USA  35 DepartmentofMedicalOncology,Dana-FarberCancerInstitute,Boston,MA,USA  36 DepartmentofChemicalandBiomolecularEngineering,VanderbiltUniversity,Nashville,TN,USA  37 DepartmentofMolecularPhysiologyandBiophysics,VanderbiltUniversity,Nashville,TN,USA Correspondingauthor:   Fendt,Sarah-Maria( sarah-maria.fendt@vib-kuleuven.be )*Authorsare   listed   inalphabeticorder. CurrentOpinionin   Biotechnology  2015, 34 :189 – 201Thisreviewcomesfromathemedissueon Systemsbiology  Editedby   SarahMariaFendt and   CostasDMaranashttp://dx.doi.org/10.1016/j.copbio.2015.02.003 0958-1669/  # 2015ElsevierLtd.Allrightsreserved. Introduction Investigating   cellular   metabolism   has   along-standinghistoryinvarious   research   areas   such   as   biochemistry,biotechnology   andcellular   physiology.   A   widelyapplica-bletoolbox   to   quantitatively   measure   intracellular   me-tabolism   has   beendeveloped   in   thecontext   of biochemicalengineering   [1].In   light   of    theemergingrealizationthat   altered   cellular   metabolism   contributes   tomanydiseases   includingcancer,   metabolic   syndromes,andneurodegenerative   disorders,   these   approaches   arebeingincreasingly   applied   to   address   biomedical   researchquestions   [2 – 8,9  ].Cellular   metabolism   can   be   characterized   bymany   pa-rametersincluding   nutrient   uptakeandmetabolite   secre-tionrates,   intracellular   metabolite   levels,   intracellularmetabolic   rates(fluxes),   nutrientcontributions   to   metab-oliteand   macromolecule   synthesis,   andpathway   activities[2,3,9  ,10 – 12].Metabolomics,   whichprovides   absolute   orrelative   intra-cellularorextracellular   metabolite   levels,   isabroad   andsensitive   method   to   detect   differences   in   metabolic   statesbetweenconditions   [13 – 16].   Changes   in   intracellularmetabolite   levels   indicate   an   altered   activityofthe   con-nectedconsuming   orproducing   reactions(e.g.   enzymatic,non-enzymatic,   or   transport   reactions)   [17  ,18  ,19 – 21].However,   concentration   changes   do   notreadily   allowconclusions   on   metabolic   rates(fluxes),   or   thedirectionof    thefluxchanges,   since   an   increase   in   metaboliteconcentration   can   both   be   indicative   of    increased   activityof    metabolite   producing   enzymes,butalso   decreasedactivityof    metabolite   consuming   enzymes.Incombination   with   growth   rates   (whichprovide   globalinformationon   metabolic   fluxes   to   biomassproduction),metabolite   uptake/secretion   rates   provide   amacroscopicpicture   of    overallmetabolism.   For   instance,   measuringthe   rate   of    glucose   depletion   from   themedia   reports   therate   of    glucose   used   by   cells   in   aculture   system.   However,thesedata   alone   areinsufficient   to   reveal   intracellularfluxesthroughout   thedifferent   metabolic   pathways.Toexamine   intracellular   fluxes   (metabolite   amount   con-verted/cell/time),   heavy   isotope   (most   frequently  13 C)labeled   nutrients   (tracers)are   commonly   utilized   [22 – 29].   In   formal  13 Cflux   analysis,   labeling   patterns   inintracellularmetabolites   resulting   from   metabolizing   a 13 C   labeled   nutrient,   cellular   uptake   andsecretion   rates,andprior   knowledge   of    thebiochemical   reaction   networkarecombined   to   computationally   estimate   metabolicfluxes[11,30  ,31 – 33,34  ].   In   practice,resolving   meta-bolicfluxes   from   measured   data   can   be   time   anddata-intensive.   In   many   cases,   however,   direct   interpretationof   13 C   labelingpatterns   (without   formal  13 C   fluxanalysis)issufficient   to   provideinformation   onrelative   pathwayactivities,   qualitative   changes   in   pathwaycontributionsviaalternative   metabolic   routes,   andnutrient   contribu-tionto   the   production   of    different   metabolites.   We   refertothis   direct   interpretation   of   13 Clabeling   patterns   as  13 Ctraceranalysis.   Here,   wediscuss   selectedimportantaspects   to   consider   when   performing  13 Ctraceranalysistoensurecorrect   data   interpretation   and   to   increase   theinsightobtained   by   stable   isotopic   tracer   experiments. Metabolicsteadystate   versusisotopicsteady state Metabolic   steady   state   requires   that   both,   intracellularmetabolitelevels   and   intracellularmetabolicfluxes   of    acellor   a   cell   populationare   constant   (Figure1a)[35]. Controlledculture   systems   thatensure   metabolic   steadystateare   continuous   cultures   (known   as   chemostats),   wherecellnumber   and   nutrient   concentrations   are   maintainedconstantthroughout   the   experiment   [36].More   commonly,experimentsare   performedat   pseudo-steady   state,   wherechangesin   metabolite   concentrations   and   fluxes   are   mini-malon   the   timescaleover   which   the   measurement   isbeingmade.   In   adherentmammalian   cellculture,   perfusionbioreactorsandnutrostats[37,38],wherenutrient   concen-trationsbut   notcell   number   are   constantover   time,areclosestto   a   chemostat.   In   conventional   monolayerculture,the   exponential   growth   phaseisoften   assumed   to   reflectmetabolicpseudo-steady   state,   because   cellsin   the   culturesteadilydivide   at   their   maximal   condition   specific   rate,giventhatnutrient   supply   doesnotbecome   limiting[39]. 190 Systems   biology  CurrentOpinioninBiotechnology  2015, 34 :189 – 201www.sciencedirect.com  So   longas   biologicalchanges   (e.g.differentiation)   occurslowly   relativetothe   timescaleof    metabolic   measurement,non-proliferatingcells   aregenerally   also   inmetabolic   pseu-do-steady   state.This   canbeverified   by   timeresolvedmeasurements   of    metabolic   parameters   of    interest[40].   Incasethe   biologicalsystemis   notinmetabolicpseudo-steadystate,for   example,   followingacutesignalingeventsor   nutri-ent   modulations,tracerexperiments   canstill   providequali-tativeand   quantitativeinformationon   metabolic   pathwayfluxes,but   interpretationof    non-steadystatedata   requiredifferent   approaches   [30  ,41 – 43]   than   thehere   discussed 13 C   traceranalysis   atmetabolic   pseudo-steady   state.Whilemetabolic   steady   state   characterizes   the   state   of metabolism,isotopicsteadystate   characterizes   the   en-richment   of    astable   isotopic   tracer   in   metabolites.   When   a 13 C   labeled   substrate   isadded   andsubsequently   metab-olized,the   metabolites   will   become   with   timeincreas-inglyenriched   for  13 Cuntil   thepoint   where   the  13 Cenrichment   isstable   over   time   (Figure   1b).From   apracticalperspective,   isotopicsteady   state   isreachedwhen  13 Cenrichment   into   a   given   metabolite   isstableovertime   relative   to   experimentalerror   and/or   the   desiredmeasurementaccuracy.   These   enrichment   dynamics   dif-ferdepending   ontheanalyzed   metabolite   and   the   traceremployed,since   the   timerequired   to   reach   isotopicsteadystate   dependson   both   the   fluxes   (i.e.   rate   of conversion)   from   thenutrient   tothatmetabolite,   andthepool   sizes   of    that   metabolite   and   allintermediatemetabolites.For   example,   upon   labeling   with  13 C-glu-cose,isotopic   steadystatein   glycolytic   intermediatestypicallyoccurswithin   minutes,   whereasfor   tricarboxylicacid(TCA)   cycle   intermediates   it   may   takeseveral   hours.For   many   amino   acids   that   areboth   produced   by   the   cellandaresupplemented   in   the   media   isotopicsteadystatemaynever   be   achieved   in   standard   monolayer   culture,dueto   constant   andrapidexchange   between   theintra-cellular   andtheextracellular   amino   acid   pools.In   such   a 13 C   tracer   analysis   Buescher   etal.   191 Figure1 mitochondrial malic enzyme    m  e   t  a   b  o   l   i  s  m    (   e .  g .   g   l  u  c  o  s  e  u  p   t  a   k  e  r  a   t  e   ) time [hours] metabolic steady state  glucose depleted glucose added     1   3    C  e  n  r   i  c   h  m  e  n   t   i  n  m  e   t  a   b  o   l   i   t  e  s time [minutes] steady state labeling  12 C 13   C M+0   10500   0.12M+1   22000   0.25M+2   15000   0.17 raw labeling data    m  e   t  a   b  o   l   i   t  e  w   i   t   h   t   h  r  e  e   C  a   t  o  m  s 13  C glucose added ion count (IC) (freely chosen example) total ion count (TIC) Σ  = 88000 MDV (IC/TIC) M+3   40500   0.46 (a)  metabolic changes  (b)  labeling changes 13 C-glutaminelactatealaninemalatepyruvate measured mitochondria TCA cycle  time [hours] dynamic labeling  time [minutes]time [minutes]alanine cytosol  (c)  mass distribution vector (MDV) metabolite A metabolite B  metabolite level metabolite level  pyruvatepyruvate (d)  cellular compartments glucose Current Opinion in Biotechnology Labelingbasics. (a) Time   dependentmetabolicchanges:Metabolismreachesametabolicsteadystatewhenthe   parametersofinterest(e.g.glucoseuptakerate)areconstantovertime. (b) Timedependentlabelingchanges:   Upon   additionof    anisotopicallylabeledcarbon   source,theisotopicenrichmentwillincreaseinthe   metabolitesuntil   thesteadystateenrichmentisreached. (c) Massdistributionvector(MDV)   (alsoknownasmassisotopomerdistribution(MID)vector):LabelingpatternsareMDVsthatconsistofthe   fractionalabundanceofeachisotopologue(alsoknownasmassisotopomer).Mdenotesmassofthe   unlabeledmetabolite. (d) Cellularcompartmentalization:Mostlabelingpatterndetectionmethodscannotresolvedifferentcellularcompartments,   thusthewholecellaveragelabelingpatternismeasured.www.sciencedirect.com CurrentOpinioninBiotechnology  2015, 34 :189 – 201  situation,   qualitative   tracer   analysis   can   easily   be   mislead-ing,and   quantitative,   formal   approachesarerequired(e.g.[44  ]).Key   aspects:    Proper   interpretation   of    labeling   data   depends   on   priorassessmentof    whetherthe   system   isatmetabolicpseudo-steady   state.   Ifso,   interpretation   of    tracer   dataismost   simpleif    labeling   isallowedto   proceed   also   toisotopicsteady-state.  The   time   to   reach   isotopicsteady   state   depends   bothonthe   tracer   being   employed   andthe   metabolitesbeinganalyzed.  Many   amino   acids   arefreely   exchanged   betweenintracellularandextracellular   pools.   This   can   preventlabelingfrom   reaching   isotopic   steadystate   andanyintracellular   metabolite   pool   that   isin   rapidexchangewith   alarger   extracellular   pool   issubject   to   thiscomplication. Labelingpatterns The   term   ‘labeling   pattern’   refers   to   amass   distributionvector(MDV)   (they   arealso   frequently   called   mass   iso-topomer   distribution   (MID)   vectors)   (Figure1c).Theshiftinmassof    a   metabolite   occurs   due   to   theincorpo-rationof    isotopes.   Metabolites   that   only   differ   in   theisotopecomposition   are   isotopologues   (theyarefrequent-lyalso   called   massisotopomers).   MDVsdescribe   thefractionalabundance   of    eachisotopologue   normalizedtothe   sum   of    allpossible   isotopologues.   Ametabolitewith n   carbonatoms   can   have0   to   n of    itscarbonatomslabeledwith  13 C,   resulting   in   isotopologues   that   increaseinmass(M)   fromM+0   (all   carbons   unlabeled   i.e.  12 C)   toM+ n (all   carbons   labeled   i.e.  13 C).   Hence,the   MDVrepresentstherelative   abundances   of    M+0   to   M+ n   iso-topologues   for   one   particular   metabolite   (Figure   1c).Consequently,   the   sum   of    allfractions   from   M+0   toM+ n is   100%   or   1.   Note   that   in   respect   to  13 Ceachisotopologuehas  n k   isotopomers   (same   isotope   com-position   but   differentposition   of    theisotope   within   themetabolite),   when   n denotes   thenumber   of    carbons   in   ametabolite   and    k thenumber   of    carbons   thatare  13 C(Figure1c).Isotopomers   can   onlybe   resolved   using   adetectionmethod   that   can   assign   aspecific   position   to   a 13 C   within   a   molecule   (e.g.   nuclear   magnetic   resonancespectroscopy   [45],mass   spectrometry   analysis   of    multiplefragments   [46]or   in   specificcasestandem   mass   spectrom-etry[47,48]).   Although   information   on   theposition   ofathe 13 C   labelcan   increase   the   information   content   of    labelingdata,theMDV   istypically   sufficient   to   draw   conclusions   onnutrientcontributions,   andalso   often   regardingpathwayactivities.Notably,while   wewill   discuss  13 Ctraceranaly-sis,the   above-described   MDVs   can   bealso   appliedto   otherstableisotopes   including  15 N   and  2 H.ToapplyMDVs   to   assess   nutrient   contributions   andpathwayactivities,   it   isimportant   to   first   correct   for   thepresenceof    naturally   occurring   isotopes,   for   example,  13 C(1.07%   natural   abundance   (na)),  15 N   (0.368%   na),  2 H(0.0115%   na),  17 O   (0.038%   na),  18 O(0.205%   na),  29 Si(4.6832%   na),or  30 Si(3.0872%   na)[49,50  ,51  ].   Forex-ample,   glutamate   and a -ketoglutarate,   whichare   normallyincomplete   exchange   andshare   thesame   carbon   back-bone,   should   accordinglyhavematchingMDVs.Yet,sincetheydifferin   theirmolecularformula,uncorrectedMDVsof    glutamate   and a -ketoglutarate   willnotmatchbecause   ofthe   natural   occurringisotopesin   N,   H,   andO.Foranalyticalmethodsthatrequiremetabolitederivati-zationto   enablechromatographic   separation(e.g.gaschromatography – massspectrometry),   thechemicalmod-ification   adds   additionalC,H,   N,O,andSiatomsto   themetabolites   [22,52].   Hence,   thenaturallabeling   of    allatomsinthemetaboliteandthederivatization   agentneedsto   betakeninto   account   when   performing   datacorrection.   For   analysisofunderivatized   metabolites(e.g.   byliquidchromatography – mass   spectrometry),   nat-urally   occurring  13 C   has   a   muchgreater   effectthan   othernaturalisotopes,   andit   is   minimally   imperativeto   correctforit.A   general   applicable   correction   matrixcan   be   formulatedbasedon   Eqn.   (1).  I  0  I  1  I  2 .   ..  I  n ...  I  n þ u 0BBBBBBBBBBB@1CCCCCCCCCCCA ¼  L  M  0 0  00 .   .. 0  L  M  0 1  L  M  1 0  0   .   .. 0  L  M  0 2  L  M  1 1  L  M  2 0  .   .. 0 ..   .   .   .   .   ...   .   ...   .   .  L  M  0 n  L  M  1 n  1  L  M  2 n  2  ....   .   ...   .   .   .   .   ...   .   ...   .   .  L  M  0 n þ u  L  M  1 n þ u  1  L  M  2 n þ u  2  ...  L  M  n u 0BBBBBBBBBBBB@1CCCCCCCCCCCCA   M  0  M  1  M  2 .   ..  M  n 0BBBBBB@1CCCCCCA (1)Here,thevector  I    denotes   the   fractionalabundances   of themeasured   metabolite   ions.  M    represents   the   MDVcorrectedfor   naturally   occurringisotopes.   n denotes   thenumber   of    carbonatomsthat   arepresent   in   the   analyzedmetabolite   ion   and   aresubject   to   isotope   labeling.   u denotesadditional   measured   ionabundancesbeyond   n srcinatingfrom   naturalisotopes   in   themetabolite   or   thederivatization.    L   denotes   the   correction   matrix   andthecolumns  L  M   k denote   the   theoreticalnatural   MDV   when    k (0to   n )carbonsare  13 C.   The   correction   matrix    L can   becalculated   based   on   thesum   formulaof    the   metabolite   ionunderconsideration   of    natural   isotope   abundances[49,53,54].To   solve   the   linear   equation   system   at   least n +1abundances   have   to   bemeasured.   Ifmorethan   n +1abundancesareconsidered,   this   results   in   an   overdeter-mined   system   and   provides   amore   robust   solution.   Tools 192   Systems   biology  CurrentOpinioninBiotechnology  2015, 34 :189 – 201www.sciencedirect.com  for   quickly   converting   raw   into   correctedMDVs   areavailable[55,56].Whenusing   analytical   approaches   involving   selected   ionmonitoring(SIM)   or   selected   reaction   monitoring   (SRM)massspectrometry,   it   isimportant   to   consider   upfront   thepotential   role   of    naturally   occurring   isotopes   when   settingtheselected   massrange   [50  ].   In   cases   involving   deriva-tization   with   Si-containing   reagents,   inclusionof    thesehighermass   rangesmay   be   important   and   therequiredmassrange   can   be   estimated   based   on   multinomial   ex-pansion(typically   a   shiftof    up   to   4amu   beyond   the   massofthe   fully   labeled   metabolite   shouldbe   considered).Comparison   between   labeled   andunlabeled   samples   issufficienttodetermine   whether   an   observed   mass   shifttruly   reflects   labeling   (as   opposed   to   merely   naturalisotope   abundance).   Itis   notappropriate,   however,   tosubtractthe   MDV   of    an   unlabeled   samplefrom   thelabeledsample.   Typically,   themain   natural   abundancepeak   in   theunlabeled   sample   willbe   M+1,   whereas   inlabeledsamples   natural   abundance   resultsin   peaks   athighermasses.Thenatural   occurring   isotopescan   be   also   used   to   validatetheappliedmass   spectrometry   method   forits   accuracy   tomeasureisotopologue   distributions   [22].Specifically,   me-tabolitescan   be   extracted   from   cellsfed   withnaturallylabelednutrients   (commonly   referred   toas   unlabelednutrients)   and   consequently   themeasured   MDV   of    thesemetabolites   should   accurately   (absolute   error   < 1.5%)   re-flectthe   theoretical   distribution   of    natural   occurring   iso-topes.With   this   validation   the   appliedmass   spectrometrymethod   can   be   improved   or   metabolites   for   which   theisotopologuedistribution   ismeasured   with   pooraccuracycan   be   excludes.   It   isimportant   tobe   aware   of    the   extent   of errorinMDV   measurements   andto   interpret   resultinglabeling   data   accordingly.   Randomerror   in   MDV   measure-mentis   oftensignificant   for   metabolites   thatare   lowabundance(i.e.   measurement   signalclose   to   noise).   Sys-tematic   errorin   MDV   measurement   ismore   serious   andcanreflect   metabolite   misannotations   or   overlaps   of    themeasuredmetabolite   ions   with   same   mass   ionsfrom   sam-plematrix   components.   Incase   the   accuracy   to   measureisotopologuedistributions   isvalidated,   datavariability   canbeasubject   oftheexperimental   procedure   (e.g.   inade-quatemetabolism   quenching)   or   thebiological   system   (e.g.rapid   metabolic   shifts   or   acontinuous   metabolicdrift).Keyaspects:  Correction   for   natural   abundance   facilitates   properinterpretation   of    labeling   data.  Subtracting   themeasured   MDV   of    anunlabeledmetabolitefrom   the   measured   MDV   of    thelabeledmetaboliteisnot   avalid   methodto   correct   fornaturalabundance.  Labeling   patterns   must   be   interpreted   in   light   of    theexperimentalerror   in   MDV   measurements   of    thechosenanalytical   approach.   Measurement   errorwilltypicallybehigherfor   low   abundance   compounds.  In   case   measurement   inaccuracy   can   beexcluded,   datavariability   can   result   from   theexperimental   procedureorthe   biologicalsystem. Cellularcompartments Eukaryotic   cellshave   organelles   such   as   mitochondria   andperoxisomes,andthese   organelles   result   in   intracellularcompartmentalization   ofmetabolitesandmetabolic   reac-tions.   Many   metabolites   are   present   in   multipleintracel-lularcompartments   and   even   spatialdistribution   within   acompartment   might   occur.This   adds   a   layer   ofcomplexi-tyto   understanding   metabolism.   Only   theaverage   label-ingpattern   and   metabolite   levels   from   allcompartmentswithinacell   can   be   measuredusing   most   current   tech-niques(Figure   1d)   [57,58].Dependingon   themetabolite   of    interest,   compartment-specificlabeling   patterns   insome   cases   can   be   inferredfromlabeling   of    metabolites   that   are   produced   exclusive-lyin   one   compartment   (Figure1d).For   example,pyru-vateis   found   both   in   the   cytosol   andin   the   mitochondria.Lactateandalanine   areboth   directly   produced   frompyruvate.Lactatedehydrogenase,   theenzymewhichinterconvertspyruvate   andlactate,   isastrictly   cytosolicenzyme[59],   an   assumption   inagreement   withthe   ob-servation   that   the   deletion   of    themitochondrial   pyruvatecarrier   does   not   affect   lactate   production   [60,61].   Thefindingthat   mitochondrial   pyruvate   carrier   deletion   dras-ticallyaffects   alanine   production   [60,61]   supports   thatalanine   isproduced   extensively   from   mitochondrial   py-ruvate[62].   Thus,   under   experimental   conditions   inwhich   neither   exogenous   alanine   nor   lactate   is   availabletocells,   lactate   labeling   likely   reflects   the   labeling   patternofcytosolic   pyruvate,   while   alanine   labeling   betterreflects   thelabeling   pattern   of    mitochondrial   pyruvate.Additionally,   engineered   compartment-specific   produc-tionof    metabolites   in   cellscan   also   be   used   to   providecompartment   specific   information.   For   example,   labelingofNADPH   inthemitochondria   and   thecytosol   wasdeterminedby   compartmentalized   transfer   of    deuteriumtothe   metabolite   2-hydroxyglutarate   (2-HG)   [63  ].Spe-cifically,   transient   expression   of    either   mutant   isocitratedehydrogenase   1   or   2   results   in   compartment   specificproductionof2-HG   that   utilizes   NADPH   availableinthat   location.   Thisapproach,and   asimilar   approach   butwithoutengineered   compartment   specific   production   of 2-HG   was   used   to   infercompartmentalized   serine   —glycineinterconversion   [63  ,64  ].Keyaspects:  In   most   casescellaveragelabelingpatternsaremeasured.Because   many   metabolites   are   present   in   more   thanone 13 C   tracer   analysis   Buescher   etal.   193 www.sciencedirect.com CurrentOpinioninBiotechnology  2015, 34 :189 – 201
Search
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks