Malate as a key carbon source of leaf dark-respired CO2 across different environmental conditions in potato plants

Jou nal o Expe imen al Bo any, Vol. 66, No. 19 pp. 5769–5781, 2015 doi:10.1093/jxb/e 279 Ad ance Access publica ion 2 July 2015 This pape is a ailable online ee o all access cha ges (see h p://jxb.ox o djou nals.o g/open_access.h ml o u he de ails) RESEARCH PAPER Mala e as a key ca bon sou ce o lea da k- espi ed CO2 ac oss di e en en i onmen al condi ions in po a oplan s Ma co M.Lehmann1,2,*, Ka ja T.Rinne1, Ca olaBlessing2, Rol T. W.Siegwol 1, NinaBuchmann2 and Roland A.We ne 2 1 Labo a o y o A mosphe ic Chemis y, Paul Sche e Ins i u e (PSI), CH-5232 Villigen, Swi ze land 2 Ins i u e o Ag icul u al Sciences, ETH Zu ich, Uni e si ae ss . 2, CH-8092 Zu ich, Swi ze land * To whom co espondence should be add essed. E-mail: ma [email p o ec ed] Recei ed 28 Janua y 2015; Re ised 30 Ap il 2015; Accep ed 6 May 2015 Edi o : Howa d G i i hs Abs ac Dissimila ion o ca bon sou ces du ing plan espi a ion in suppo o me abolic p ocesses esul s in he con inu- ous elease o CO2. The ca bon iso opic composi ion o lea da k- espi ed CO2 (i.e. δ13CR) shows daily en ichmen s up o 14.8‰ unde di e en en i onmen al condi ions. Howe e , he easons o his 13C en ichmen in lea da k- espi ed CO2 a e no ully unde s ood, since daily changes in δ13C o pu a i e lea espi a o y ca bon sou ces (δ13CRS) a e no ye clea . Thus, we exposed po a o plan s (Solanum ube osum) o di e en empe a u e and soil mois u e ea men s. We de e mined δ13CR wi h an in- ube incuba ion echnique and δ13CRS wi h compound-speci ic iso ope analysis du ing a daily cycle. The highes δ13CRS alues we e ound in he o ganic acid mala e unde di e en en i on- men al condi ions, showing less nega i e alues compa ed o δ13CR (up o 5.2‰) and compa ed o δ13CRS o soluble ca bohyd a es, ci a e and s a ch (up o 8.8‰). Mo eo e , linea ela ionships be ween δ13CR and δ13CRS among di - e en pu a i e ca bon sou ces we e s onges o mala e du ing day ime ( 2=0.69, P≤0.001) and nigh ime ( 2=0.36, P≤0.001) unde all en i onmen al condi ions. Amul iple linea eg ession analysis e ealed δ13CRS o mala e as he mos impo an ca bon sou ce in luencing δ13CR. Thus, ou esul s s ongly indica e mala e as a key ca bon sou ce o 13C en iched da k- espi ed CO2 in po a o plan s, p obably d i en by an anapleu o ic lux eplenishing in e media es o he K ebs cycle. Key wo ds: Compound-speci ic iso ope analysis (CSIA), d ough , o ganic acids, plan espi a ion, s able ca bon iso opes, suga s, empe a u e, ica boxylic acid (TCA) cycle. In oduc ion The in es iga ion o plan espi a ion as a majo p ocess in plan biochemis y has expanded ou unde s anding o ca - bon cycling in au o ophic o ganisms. Plan s dissimila e ca - bon sou ces o he p oduc ion o in e media es and educing equi alen s in suppo o me abolic p ocesses, he eby con in- uously eleasing CO2 ia plan espi a ion (Hopkins, 2006). Lea - espi ed CO2 is mainly de i ed om oxida i e deca box- yla ion eac ions ca alysed by enzymes om he K ebs cycle This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License (h p://c ea i ecommons.o g/licenses/by/3.0/), which pe mi s un es ic ed euse, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed. © The Au ho 2015. Published by Ox o d Uni e si y P ess on behal o he Socie y o Expe imen al Biology. Abb e ia ions: δ13CR, ca bon iso opic composi ion o lea da k- espi ed CO2; δ13CRS, ca bon iso opic composi ion o pu a i e lea espi a o y ca bon sou ces; An, ne assimila ion a e; Ci, in e cellula CO2 concen a ion; CSIA, compound-speci ic iso ope analysis; gs, s oma al conduc ance; HPLC, high pe o mance liquid ch oma og aphy; KC, K ebs cycle; LEDR, ligh -enhanced da k espi a ion; ME, malic enzyme; OAA, oxaloace a e; PDH, py u a e dehyd ogenase; PEPC, phosphoenolpy u a e ca boxylase; SPS, suc ose phospha e syn hase; SWC, olume ic soil wa e con en . a Na u al Resou ces Ins i u e Finland (Luke) on No embe 29, 2016h p://jxb.ox o djou nals.o g/Downloaded om 5770 | Lehmann e al. (KC) and om in e ac ing anabolic and ca abolic eac ions (Voe and Voe , 2011). Using s able iso opes, he pa hway o ca bon can be aced om pho osyn he ic ca bon ixa ion o espi a o y ca bon loss. On he one hand, C3 plan s disc imina e hea ily agains 13C due o pho osyn he ic iso ope ac iona ion, leading o gene al 13C deple ion in plan biomass o abou 20‰ in compa ison o a mosphe ic CO2 (Fa quha e al., 1989). The exac magni ude o pho osyn he ic ca bon iso ope disc imi- na ion depends on he in e cellula CO2 concen a ion (Ci) in he subs oma al ca i y, which is egula ed by o he physiolog- ical pa ame e s such as ne assimila ion a e (An) and s oma- al conduc ance (gs). En i onmen al condi ions such as ligh , empe a u e, soil mois u e, and ai humidi y will in luence hese pa ame e s and wi h hem he pho osyn he ic ca bon iso ope disc imina ion. On he o he hand, he ca bon iso- opic composi ion o lea da k- espi ed CO2 (i.e. δ13CR) has clea ly been shown o be less nega i e han lea me aboli es in se e al plan species (Ghashghaie e al., 2003; Bowling e al., 2008; We ne and Gessle , 2011; Ghashghaie and Badeck, 2014). In a daily cycle, lea da k- espi ed CO2 ollows a p o- g essi e 13C en ichmen du ing he day and a g adual 13C deple ion du ing he cou se o he nigh (Hymus e al., 2005; P a e e al., 2006), esul ing in a s ong empo al a iabili y o up o 14.8‰ (Ba bou e  al., 2007; We ne e  al., 2009; Wegene e al., 2010), which di e s among unc ional g oups (P iaul e al., 2009; We ne e al., 2009). δ13CR is he eby linked o he ca bon iso opic composi- ion o pu a i e lea espi a o y ca bon sou ces (i.e. δ13CRS) such as ca bohyd a es (soluble mono- and di-saccha ides, and s a ch) and o ganic acids. P e ious s udies showed ha en i onmen al d i e s such as empe a u e and soil mois u e in luence δ13CR and δ13CRS. Mo e nega i e δ13CR alues wi h inc easing empe a u e ha e been obse ed wi h sho - e m changes in lea empe a u e du ing da kness in Phaseolus ulga is (Tche kez e  al., 2003), while long- e m e ec s o highe empe a u es on δ13CR and δ13CRS ha e no ye been in es iga ed unde con olled condi ions. O he s udies ha e demons a ed less nega i e δ13CR and δ13CRS alues unde d y condi ions compa ed o hose unde we condi ions (Du anceau e  al., 1999; Ghashghaie e  al., 2001). Simila obse a ions we e made in ield expe imen s (Sun e al., 2009; Dubbe e al., 2012). Con e sely, mo e nega i e δ13CR al- ues ha e been ound unde d y condi ions o Medi e anean ees and he bs such as Que cus ilex and Tube a ia gu a a compa ed o hose unde we condi ions (Unge e al., 2010), which ha e been explained wi h accompanied inc eases in empe a u es and apou p essu e de ici . Ne e heless, he combined e ec s o empe a u e and soil mois u e on δ13CR and δ13CRS unde con olled condi ions ha e ye o be es ed. Mo eo e , δ13CR is de e mined by a ious pos -pho o- syn he ic ca bon iso ope ac iona ion p ocesses a pi o al b anching poin s in espi a o y pa hways, ca bon iso ope e ec s on enzyma ic eac ions, and changes in espi a o y subs a es ( o a de ailed e iew see We ne and Gessle , 2011). The 13C en ichmen in lea da k- espi ed CO2 i sel is hough o be a esul o agmen a ion ac iona ion p o- cesses based on he e ogeneous in amolecula ca bon iso ope dis ibu ion in espi a o y ca bon sou ces (Tche kez e  al., 2004). Fo ins ance, C-3 and C-4 posi ions o glucose a e known o be en iched in 13C compa ed o he o he mole- cule posi ions due o an iso ope e ec o he aldolase eac- ion (Rossmann e  al., 1991; Gleixne and Schmid , 1997). B eakdown o glucose du ing glycolysis p oduces py u a e wi h a 13C en iched C-1 posi ion ( o me C-3 and C-4 posi- ions o glucose). The ea e , he py u a e dehyd ogenase eac ion (PDH) eleases he C-1 posi ion as 13C en iched CO2, whe eas he mo e 13C deple ed ace yl-CoA esidue is used in he KC (P iaul e  al., 2009; We ne and Gessle , 2011). Thus, a PDH domina ed espi a o y pa hway may lead o 13C en ichmen in lea da k- espi ed CO2. Howe e , he knowledge abou δ13CR is o en based on ligh -acclima ed lea es, which ha e been ans e ed in o da kness o allow espi a o y measu emen s. This app oach holds an unp e en able bias known as ‘ligh -enhanced da k espi a ion’ (LEDR), which needs o be aken in o accoun when in e p e ing day ime δ13CR alues. LEDR is a sho - e m ligh -da k ansi ion pe iod, desc ibing an inc ease in he amoun o lea da k- espi ed CO2 sho ly upon da k- ening o abou 20 min, which depends on ligh in ensi y (A kin e al., 1998). On he one hand, LEDR may be in lu- enced by eassembly o he KC, which is hough o be only pa ially ac i e unde ligh condi ions (Tche kez e al., 2005; Swee lo e e  al., 2010; We ne and Gessle , 2011; We ne e al., 2011). On he o he hand, LEDR may be d i en by a b eakdown o a ligh -accumula ed mala e pool, causing 13C-en iched lea da k- espi ed CO2 (Ba bou e al., 2007; Gessle e al., 2009; We ne e al., 2009; Ba bou e al., 2011; We ne and Gessle , 2011). Mala e i sel is also known o be 13C en iched compa ed o o he ca bon sou ces (Gleixne e al., 1998; Ghashghaie e  al., 2001). The 13C en ichmen in mala e was a ibu ed o an anapleu o ic lux ia he phosphoenolpy u a e ca boxylase eac ion (PEPC), which ixes 13C-en iched hyd ogen ca bona e and eplenishes KC in e media es (Melze and O’Lea y, 1987; Sa idge and Blai , 2004). Thus, a possible b eakdown o mala e by he mi o- chond ial malic enzyme eac ion, o wi hin he KC, may in luence δ13CR (Ba bou e al., 2007; We ne e al., 2011). In addi ion, plan s may also use o a ce ain ex en mo e complex ca bon sou ces such as lipids and p o eins unde se e e en i onmen al condi ions o unde p olonged da k- ness (Tche kez e  al., 2003; Usadel e  al., 2008). Howe e , he d i ing p ocesses, he espi a o y ca bon sou ces, and he mechanisms causing changes in δ13CR du ing day and nigh a e no ully esol ed hus a . Hence, wi h his s udy we in end o assess wo majo esea ch ques ions. Wha causes he high daily a ia ions in δ13CR? How a e δ13CR and δ13CRS in luenced by empe a u e and soil mois u e condi ions? Ou main objec i es we e (i) o analyse he ela ionship be ween δ13CR and δ13CRS alues and (ii) o de e mine changes in δ13CR and δ13CRS alues, as well as in concen a ions o he pu a i e ca bon sou ces unde di e - en en i onmen al condi ions. The e o e, we exposed po a o plan s o di e en con olled empe a u e and soil mois u e condi ions and measu ed δ13CR wi h an in- ube incuba ion echnique, as well as δ13CRS and concen a ions o soluble a Na u al Resou ces Ins i u e Finland (Luke) on No embe 29, 2016h p://jxb.ox o djou nals.o g/Downloaded om Mala e as a key ca bon sou ce o lea da k- espi ed CO2 in po a o | 5771 ca bohyd a es, o ganic acids and s a ch om lea es wi h compound speci ic iso ope analysis (CSIA) on a daily basis. Ma e ials and me hods Plan ma e ial Po a o plan s (Solanum ube osum L. c . Annabell) we e g own om ube s o he same size in 5 l po s illed wi h ba k humus soil (Ökohum, He enho , Swi ze land) in a g eenhouse, wi h a e age empe a u es o 20/16°C and apou p essu e de ici s (VPD) o abou 0.9/0.4 kPa (day/nigh ). The plan s we e exposed o a 16 h dayligh pe iod supplemen ed by 400 W sodium-lamps (Powe one Son-T Plus, Philips, Ams e dam, Ne he lands). Fo y days a e plan ing, plan s we e ans e ed in o walk-in clima e chambe s o acclima i- za ion o 2 weeks. The 16 h dayligh in he clima e chambe s had an a e aged pho osyn he ic pho on lux densi y o ~400µmol m-2 s-1 a lea le el, hus plan s we e no ully ligh -sa u a ed. Be o e he ea - men pe iod, soil wa e s a us was op imal o a leas 3 d a e wa e - ing, while an indi idual plan consumed abou 300 ml wa e pe day. 50 ml o a 0.4% e ilize solu ion ( / , Gesal, Zü ich, Swi ze land) was applied wice o all plan s du ing he whole expe imen o 70d. T ea men s we e applied du ing he las 15 d o he expe imen . Plan s we e exposed o high empe a u e (Thigh) o 28/23°C (day/ nigh ) and low empe a u e condi ions (Tlow) o 22/17°C, a a VPD o abou 0.9/0.35 kPa o bo h empe a u e ea men s. Th ee cli- ma e chambe s we e used o eplica ion o each empe a u e ea - men . Wi hin each clima e chambe he e we e wo soil-mois u e ea men s wi h nine plan s each. D y soil mois u e condi ions we e kep cons an ly a 50–60% o he daily wa e consump ion o each indi idual plan , de e mined by weighing he en i e po s. Plan s unde we condi ions we e kep a 100%. The inal sampling pe iod las ed 32 h du ing he las 2 d o he expe imen , when d y soil condi ions we e es ablished o bo h em- pe a u e ea men s. Sampling was done on a daily basis e e y 2 h (nigh ime) o 4 h (day ime). Du ing sampling, indi idual plan s had 3–6 anks, wi h abou ou ully de eloped lea es pe ank. Always he hi d-las ully de eloped lea pe ank was sampled a all poin s in ime, bu wi hin 24 h only one sample was aken om each indi id- ual plan o a oid any s ess esponse induced by sampling. Sampled lea ma e ial was immedia ely ozen in liquid ni ogen and s o ed a −80°C. Subsequen ly, he lea ma e ial was eeze-d ied and milled o powde by a s eel ball mill (MM200, Re sch, Haan, Ge many) o all u he iso opic and biochemical analyses. In addi ion o lea sampling, ai CO2 samples om all six clima e chambe s we e col- lec ed a he same poin s in ime du ing he sampling pe iod, show- ing a mean δ13C alue o −12.2‰ and ypical daily a ia ions o SD ≤1.4‰; no di e ences be ween empe a u e ea men s (P≥0.05) and poin s in ime (P≥0.05; linea mixed e ec s model) we e obse ed du ing he daily cycle. Physiological measu emen s and biomass de e mina ion Se e al lea physiological pa ame e s we e de e mined wi h an in a ed gas analyse (LI-6400, LI-COR, Lincoln, Neb aska, USA), including ne assimila ion a e (An), in e cellula CO2 concen a ion (Ci), and s oma al conduc ance (gs). All measu emen s we e aken in he las 4 h o he dayligh phase. To moni o olume ic soil wa e con en (SWC), up o h ee soil mois u e senso s (EC-5 and log- ge Em5b, Decagon De ices, Pullman, USA) we e ins alled o each ea men . Sho ly a e he sampling pe iod, o al plan biomass was ha es ed, o en-d ied (a 60°C), and weighed. The esh ube weigh and ube coun (numbe o po a oes) we e de e mined. Ca bon iso ope and concen a ion analyses δ13C alues a e exp essed as desc ibed by C aig (1957) and modi ied by Coplen (2011): δ 13 sample s anda d C RR 1 ‰/ () =− whe e Rsample is he 13C/12C a io o he sample ma e ial and Rs anda d is ha o he in e na ional s anda d VPDB (Vienna Pee Dee Belemni e). De e mina ion o δ13CR The in- ube incuba ion echnique was used o he collec ion o lea da k- espi ed CO2 du ing day ime and nigh ime (We ne e  al., 2007). A lea was placed in a 12 ml gas- igh exe aine (Labco, Lampe e , UK), which was immedia ely da kened wi h a ligh p oo casing o igge lea da k espi a ion. The ube was hen lushed o 1 min wi h syn he ic ai un il a CO2- ee a mosphe e was es ab- lished, which was moni o ed wi h an in a ed gas analyse (LI-6262, LI-COR, Lincoln, Neb aska, USA). A e an incuba ion ime o 3 min in da kness, an aliquo o da k- espi ed CO2 was ans e ed wi h a gas- igh sy inge in o a new exe aine illed wi h d y N2. δ13CR alues we e de e mined wi h an IRMS, using a modi ied Gasbench II (The mo Fishe , B emen, Ge many) connec ed o a Del aplusXP- IRMS, simila o Zeeman e al. (2008). The ans e o he CO2 sam- ple in o a new exe aine , as well as he IRMS measu ing p ocedu e, we e bo h es ed wi h ai o known δ13C o CO2 o ensu e no iso- ope ac iona ion had occu ed. Measu emen p ecision o a qual- i y con ol s anda d ( h ee s anda ds pe 24 samples) was SD≤0.1‰. De e mina ion o δ13C in bulk lea es and lea s a ch Ex ac ion o lea s a ch was pe o med as desc ibed in p e ious s ud- ies (Wanek e al., 2001; Goe liche e al., 2006; Rich e e al., 2009). Lea s a ch was isola ed om 50 mg lea ma e ial wi h me hanol/ chlo o o m/wa e (MCW, 12:5:3, / / ) a 70°C o 30 min. Samples we e cen i uged (10 000 ×g, 2 min) and supe na an s emo ed, while he lea -s a ch-con aining pelle s we e washed wi h MCW and deionized wa e and d ied a oom empe a u e (RT). Pelle s we e hen e-suspended in wa e and boiled a 99°C o 15 min o acili- a e s a ch gela iniza ion. Subsequen ly, lea s a ch was enzyma i- cally diges ed wi h α-amylase (EC 3.2.1.1, Sigma-Ald ich, Buchs, Swi ze land) a 85°C o 2 h, and cleaned wi h cen i uga ion il e s o emo e enzymes (Vi aspin, Sa o ius, Gö ingen, Ge many). To de e mine δ13C o bulk lea es (δ13Clea ) and s a ch, an elemen al ana- lyse (Flash EA 1112 Se ies) coupled o a Del aplusXP-IRMS was used (bo h The mo Fishe , B emen, Ge many; We ne e al., 1999). Measu emen s o samples, blanks, and e e ence ma e ial ollowed he iden ical ea men p inciple desc ibed by We ne and B and (2001). The long- e m p ecision o a quali y con ol s anda d o all sequences was SD≤0.12‰. Iso opic and concen a ion analysis o soluble ca bohyd a es and o ganicacids Wa e -soluble compounds we e ex ac ed om 100 mg lea ma e- ial wi h wa e a 85°C o 30 min, simila o S ei e  al. (2013). Subsequen ly, soluble ca bohyd a es and o ganic acids we e sepa a ed by ion-exchange ch oma og aphy (Wanek e  al., 2001; Goe liche e al., 2006; Rich e e al., 2009), using Dowex 50WX8 in H+- o m and Dowex 1X8 in NaCOO-- o m (bo h 100–200 mesh, Sigma-Ald ich, Buchs, Swi ze land). To a oid clogging o he HPLC column by polyphenols, all samples designa ed o ca bohyd a e analyses we e il e ed wi h 100 mg Sep-Pak C18 Vac RC Ca idges (Wa e s AG, Mil o d, Massachuse s, USA). Finally, all ca bohy- d a e and o ganic acid samples we e cleaned wi h 0.45µm PTFE sy inge il e (In och oma AG, Zug, Swi ze land) p io o HPLC measu emen s. To de e mine δ13CRS alues and he concen a ions o soluble ca bo- hyd a es and o ganic acids, a HPLC-IRMS sys em consis ing o a high pe o mance liquid ch oma og aph coupled o a Del a V Ad an age IRMS by a LC IsoLink (all The mo Fishe , B emen, Ge many) a Na u al Resou ces Ins i u e Finland (Luke) on No embe 29, 2016h p://jxb.ox o djou nals.o g/Downloaded om 5772 | Lehmann e al. was used acco ding o K ummen e al. (2004). Ca bohyd a es we e sepa a ed on a 3 × 150 mm anion-exchange column Ca boPac PA20 (Dionex, Ol en, Swi ze land) using 2 mM NaOH as he mobile phase and a low speed o 250µl min-1 (Boschke e al., 2008; Rinne e al., 2012). Low column empe a u e o 20°C was used o p e en isome i- za ion o hexoses (Rinne e al., 2012). This enabled ch oma og aphic sepa a ion o suc ose and glucose, bu uc ose δ13CRS and concen- a ion measu emen s we e a ec ed by pa ial co-elu ion o uc ose wi h o he compounds. To co ec δ13CRS alues and o calcula e con- cen a ions om he peak a eas, in e spe sed s anda d solu ions in a concen a ion ange o 20–180 ng C µl-1 we e measu ed wi hin each sequence. The measu emen p ecision o δ13CRS alues in all ca bohy- d a e s anda ds was SD<0.5‰. Below a concen a ion o 60 ng C µl-1, he p ecision o uc ose s anda ds was lowe o ce ain ba ches, and he e o e hese esul s we e excluded. O ganic acids we e sepa a ed on a 4.6 × 300 mm Allu e O ganic Acids column (Res ek, Belle on e, USA) a 5–10°C. The mobile phase was a 100 mM monopo assium phospha e bu e (pH 3)wi h a low speed o 500μl min-1 (He mann e al., 2005). The measu emen p ecision o δ13C in o ganic acid s anda ds was SD<0.4‰. Low ci - a e concen a ions om Tlow samples (<45 ng C µl-1) impeded he analy ical accu acy o he δ13CRS alues, he e o e hese samples we e no aken in o accoun . All pu i ica ion s eps we e e i ied o each ba ch o 24 samples using 2.5 mg s anda d solu ions o known δ13C (by EA-IRMS) o all ca bohyd a es and o ganic acids measu ed in his s udy. Di e ences be ween δ13C alues be o e and a e pu i ica ion we e gene ally ≤0.2‰, indica ing no signi ican iso ope ac iona ion o any s and- a d. Mean eco e y was 101 ± 6% o uc ose, 96 ± 6% o glucose, 89 ± 3% o suc ose, 91 ± 3% o mala e, and 86 ± 3% o ci a e. De e mina ion o s a ch concen a ion Fo he ex ac ion o lea s a ch o concen a ion analyses we used a modi ied me hod o C i chley e al. (2001). Lea s a ch was iso- la ed wi h 1.12 M pe chlo ic acid om 50 mg lea ma e ial a RT o 15 min and cen i uged (10 min, 3000×g, 4°C). The supe na an was emo ed and he lea -s a ch-con aining pelle was washed ee om pigmen s wi h deionized wa e and e hanol. Pelle s we e hen d ied a RT, esuspended in wa e , and gela inized. Subsequen ly, s a ch samples we e enzyma ically hyd olysed o glucose o 2 h a 37°C wi h a solu ion mix o α-amylase (EC 3.2.1.1, Sigma-Ald ich, Buchs, Swi ze land) and α-amyloglucosidase (EC 3.2.1.3, Roche, Ro k euz, Swi ze land) in 220 mM sodium ace a e bu e (pH 4.8). The glucose concen a ion was de e mined a 340 nm wi h a 96-well mic opla e eade (EL×800, BioTek, Luze n, Swi ze land) using a coupled enzyma ic eac ion (Hoch e al., 2002). Po a o s a ch was used as a s anda d. Glucose concen a ions a e exp essed in mola - i y o s a ch monome s. Da a analysis R e sion 3.0.2 (R Co e Team, 2013) was used o (mul iple) lin- ea eg ession analyses and linea mixed e ec s models (R package nlme). Models included ixed e ec s ( empe a u e, soil mois u e, sampling ime) and andom e ec s (clima e chambe s, indi idual plan s). I applicable, δ13C alues and concen a ions we e loga i h- mically ans o med o ensu e no mal dis ibu ion. Fo he bes - i combina ion o he mul iple linea eg ession analysis, a iables we e excluded i P≥0.05. Resul s Physiological pa ame e s and biomass Physiological pa ame e s (An, Ci, gs, and SWC) o po a o plan s exposed o ou di e en ea men s we e moni o ed du ing he ea men pe iod o 15 d (Fig.1). The ne assimila ion a e declined du ing he ea men pe iod unde all ou ea men s (Fig.1A). Du ing he sampling pe iod (Fig.1A, day 15), An was signi ican ly in luenced by soil mois u e (P=0.02, Table1), wi h lowes alues (1.9µmol m-2 s-1) unde Thigh and d y condi- ions, and highes alues (5.4µmol m-2 s-1) unde Tlow and we condi ions, whe eas he empe a u e in luence on An was no signi ican (P=0.07, Table1) bu ended o cause lowe An al- ues unde Thigh han unde Tlow unde bo h soil mois u e con- di ions. The in e cellula CO2 concen a ion inc eased du ing he ea men pe iod o all ou ea men s (Fig.1B). Du ing he sampling pe iod (Fig.1B, day 15), Ci was independen ly in luenced by empe a u e (P=0.012, Table1) and soil mois u e (P=0.01, Table1), wi h lowes Ci (247.5µmol mol-1) unde Tlow and d y condi ions and highes Ci (332.8µmol mol-1) unde Thigh and we condi ions. S oma al conduc ance du ing he ea - men pe iod was lowe unde d y ea men s compa ed o hose unde we ea men s (Fig.1C). Du ing he sampling pe iod (Fig.1C, day 15), gs was signi ican ly in luenced by soil mois- u e (P≤0.001, Table1), wi h lowes gs (abou 0.06 mol m-2 s-1) in plan s o bo h d y ea men s and highes gs (0.22 mol m-2 s-1) in plan s unde Thigh and we condi ions, whe eas he em- pe a u e in luence unde we condi ions ended o cause highe gs alues unde Thigh han unde Tlow. The olume ic soil wa e con en was lowe unde d y condi ions (~7–14%) compa ed o we condi ions (23–27.5%) o he las 9 d o he ea men pe iod (Fig.1D), including he sampling pe iod (Fig.1D, day 15), whe e SWC was signi ican ly a ec ed only by soil mois u e ea men s (P=0.002, Table1). Gene ally, no signi ican in e - ac ions be ween empe a u e and soil mois u e we e obse ed o any pa ame e (Table1). In addi ion, only soil mois u e ea men s a ec ed plan biomass (P=0.008, Table1) and ube weigh (P=0.023, Table 1) aken sho ly a e he sampling pe iod, independen o empe a u e ea men s. Highes alues ended o be unde Tlow and we condi ions and lowes alues unde Thigh and d y condi ions (Tables 1, 2), indica ing di e en s ess le els c ea ed by he ou ea men s. Ca bon iso opes in po a olea es Daily cycles o δ13CR and δ13Clea δ13C alues o lea da k- espi ed CO2 (δ13CR) a ied signi i- can ly o e ime (P≤0.001, Table3) wi h alues in he ange o −21.9‰ and −32‰, declining s ongly du ing nigh ime and inc easing again du ing he day ime o all ou ea - men s (Fig.2A). An in e ac ion be ween empe a u e and ime showed ha he in luence o empe a u e di e ed wi h ime (P=0.014, Table3). Day ime δ13CR alues unde Thigh we e up o 4.7‰ mo e nega i e compa ed o hose unde Tlow, independ- en o soil mois u e condi ions, whe eas nigh ime δ13CR alues o bo h empe a u e ea men s we e e y simila , pa icula ly in he second nigh . D y soil mois u e condi ions caused less nega i e δ13CR alues compa ed o hose unde we condi ions du ing he daily cycle (P=0.013, Table3), wi h a maximum di - e ence o 2.7‰, independen o empe a u e ea men s. On a e age, he di e ence be ween day ime and nigh ime δ13CR alues was highes unde Tlow and we condi ions, a 5.7‰, and lowes unde Thigh and d y condi ions, a 2.5‰. a Na u al Resou ces Ins i u e Finland (Luke) on No embe 29, 2016h p://jxb.ox o djou nals.o g/Downloaded om Mala e as a key ca bon sou ce o lea da k- espi ed CO2 in po a o | 5773 The bulk lea ma e ial e lec s all en i onmen al condi ions expe ienced du ing he whole g ow h pe iod. δ13Clea o all ea men s showed no changes du ing he sampling pe iod and no in e ac ions be ween ea men s and ime (Fig.2B; Table3). Unde Thigh, δ13Clea alues we e up o 2.2‰ mo e nega i e compa ed o hose unde Tlow, esul ing in a signi i- can empe a u e e ec independen o soil mois u e condi- ions (P=0.022, Table3). Simila ly, soil mois u e showed a signi ican e ec on δ13Clea (P=0.005, Table 3), independ- en o empe a u e ea men s, wi h alues up o 1.1‰ less nega i e unde d y han unde we condi ions mainly du ing nigh ime. δ13CRS o soluble ca bohyd a es, o ganic acids, ands a ch Highes δ13C alues in pu a i e lea espi a o y ca bon sou ces (δ13CRS) we e ound in he o ganic acid mala e, while ● ● ● ● ● ● ● ● ● ● ● ● 0 2 4 6 8 10 (A) An (µmol m − 2 s − 1) ●● ● ● ●● ● ● ● ●● ● 150 200 250 300 350 (B) Ci (µmol mol − 1) ●●● ● ●●● ● ● ● ● ● 0.00 0.05 0.10 0.15 0.20 0.25 (C) gs (mol m − 2 s − 1) ● ●● ●● ● ● ● ●● ●●●● ● ● ● ●● ●● ● ● ● ●● ●●●● ● ● ● ● ●● ● ● ●●● ● ●● ● ● ●● 036912 15 0.0 0.1 0.2 0.3 ●Tlow We Tlow D y Thigh We Thigh D y (D) SWC (m3 m−3) T ea men pe iod (d) Fig.1. Physiological pa ame e s unde di e en en i onmen al condi ions du ing he ea men pe iod: (A) ne assimila ion a e (An, µmol m-2 s-1), (B) in e cellula CO2 concen a ion (Ci, µmol mol-1), (C) s oma al conduc ance (gs, mol m-2 s-1), (D) olume ic soil wa e con en (SWC, m3/m-3). Po a o plan s we e ea ed wi h a combina ion o Tlow (low empe a u e; closed symbols), Thigh (high empe a u e; open symbols), and we (ci cles) o d y ( iangles) condi ions. Boxed a eas indica e he sampling pe iod. Means ±SE a e gi en (n=3). Table1. En i onmen al in luences on physiological pa ame e s Resul s o linea mixed e ec s models es ing he e ec s o empe a u e (low, high) and soil mois u e (we , d y) on physiological pa ame e s (An, ne assimila ion a e; Ci, in e cellula CO2 concen a ion; gs, s oma al conduc ance; SWC, olume ic soil wa e con en ), o al plan biomass, ube weigh , and ube coun du ing he sampling pe iod. P- alues a e gi en o ea men s and hei in e ac ion. Signi ican di e ences a e gi en in bold (P≤0.05). Pa ame e AnCigsSWC Plan biomass Tube weigh Tube coun Tempe a u e 0.070 0.012 0.127 0.863 0.978 0.359 0.400 Soil mois u e 0.020 0.010 0.001 0.002 0.008 0.023 0.233 Temp.:mois u e 0.522 0.110 0.174 0.845 0.565 0.892 0.486 ● ● ● ● ● ●● ● ● ● ●●● ● ● ● ● ● ●● ● ● ● ●●● ● ● ●● ● ● ●● ● ●● ● ● −34 −32 −30 −28 −26 −24 −22 −20 (A) δ13CR (‰) ●● ●● ●●● ● ●●● ● ● ●● ●● ●●● ● ●●● ● ● ● ● ●● ●● ● ● ● ●●●● 048121620242 83 2 −34 −32 −30 −28 −26 −24 −22 −20 (B) Sampling pe iod (h) δ13Clea (‰) ●Tlow We Tlow D y Thigh We Thigh D y Fig.2. Daily cycles o he ca bon iso opic composi ion o (A) lea da k- espi ed CO2 (δ13CR) and (B) bulk lea es (δ13Clea ) unde di e en en i onmen al condi ions du ing he sampling pe iod. Po a o plan s we e ea ed wi h a combina ion o Tlow (low empe a u e; closed symbols), Thigh (high empe a u e; open symbols), and we (ci cles) o d y ( iangles) condi ions. G ey a eas indica e nigh ime. Means ±SE a e gi en (n=3). a Na u al Resou ces Ins i u e Finland (Luke) on No embe 29, 2016h p://jxb.ox o djou nals.o g/Downloaded om 5774 | Lehmann e al. soluble ca bohyd a es ( uc ose, glucose and suc ose) exhib- i ed gene ally lowes δ13CRS alues (Fig.3). δ13CRS o soluble ca bohyd a es o all ea men s we e in he ange o −27.2‰ and −36.6‰. Mo e nega i e δ13CRS alues o glucose and suc ose unde Thigh compa ed o hose unde Tlow we e ound, independen o soil mois u e condi ions, while less nega- i e δ13CRS alues unde d y condi ions compa ed o hose unde we condi ions we e obse ed, independen o em- pe a u e ea men s (Fig.3B, C; Table3). Signi ican in e - ac ions be ween empe a u e and ime o δ13CRS o glucose (P=0.008, Table3) and suc ose (P=0.003, Table3) showed ha daily cycles di e ed be ween empe a u es. Addi ionally, soil mois u e condi ions caused signi ican empo al a ia- ions du ing he daily cycle in δ13CRS o suc ose (P=0.002, Table3). We obse ed signi ican linea ela ionships be ween uc ose and glucose o δ13CRS ( 2=0.74, P≤0.001) and concen a ion alues ( 2=0.8, P≤0.001), while ela ionships be ween he o he δ13CRS alues and concen a ions o di e - en ca bon sou ces we e weake (da a no shown). Howe e , he de ian esul s o δ13CRS o uc ose in compa ison o he o he suga s a e assumed o e lec peak o e lap issues o his suga (Tables 3, 4). This is clea ly e lec ed also in he concen- a ion esul s (Fig.4A). Consequen ly, he uc ose esul s will no be discussed u he in de ail. δ13CRS o mala e (Fig. 3D) in he ange o −24‰ and −29.3‰ and δ13CRS o ci a e (Fig.3E) in he ange o −29.6‰ and −32.1‰ showed no empo al a ia ions (P=0.198 and P=0.052 o mala e and ci a e, espec i ely, Table 3). Signi ican in e ac ions be ween empe a u e and soil mois- u e ea men s we e obse ed o δ13CRS o mala e (P=0.017; Table3), esul ing in la ge di e ences be ween δ13CRS al- ues o soil mois u e condi ions unde Thigh han unde Tlow (Fig.3D). Ci a e showed less nega i e δ13CRS alues unde d y condi ions han unde we condi ions (P=0.009; Table3). δ13CRS o s a ch o all ea men s (Fig.3F), anging om −25.2‰ and −32.1‰, was in luenced by soil mois u e con- di ions (P=0.046, Table 3), independen o empe a u e ea men s, while empe a u e showed no signi ican e ec (P=0.107, Table 3). In addi ion, soil mois u e condi ions caused signi ican empo al a ia ions du ing he daily cycle in δ13CRS o s a ch (P=0.032, Table3). Concen a ions o soluble ca bohyd a es, o ganic acids, ands a ch Concen a ions o glucose o all ea men s (Fig.4B), ang- ing om 27 o 95µmol g DW-1, showed no empo al a ia- ions (P=0.927, Table3). In con as , concen a ions o suc ose Table2. Biomass and ube analyses a e sampling pe iod To al plan biomass (d y weigh ), ube weigh ( esh weigh ), and ube coun (numbe o po a oes) a e he sampling pe iod. Po a o plan s we e ea ed wi h a combina ion o Tlow (low empe a u e), Thigh (high empe a u e), and we o d y condi ions. Means ±SE a e gi en (n=3). Re e o Table 1 o s a is ical analysis. T ea men s Pa ame e Tlow we Tlow d y Thigh we Thigh d y To al biomass (g) 10.6 ± 1.4 7.8 ± 1.2 10.3 ± 1 8.1 ± 0.5 Tube weigh (g) 513.9 ± 18.4 458.3 ± 15 481.4 ± 15.9 430 ± 24.2 Tube coun (no.) 21.3 ± 2.7 20.2 ± 1.7 19.3 ± 1.2 19 ± 2.3 Table3. En i onmen al in luences on lea da k- espi ed CO2 and espi a o y ca bon sou ces Resul s o linea mixed e ec s models es ing he e ec s o empe a u e (low, high) and soil mois u e (we , d y) on δ13C alues in di e en pu a i e lea espi a o y ca bon sou ces, bulk lea es (δ13Clea ), and in lea da k- espi ed CO2 (δ13CR), as well as on concen a ions o di e en ca bon sou ces du ing he sampling pe iod. Resul s o uc ose a e a ec ed by co-elu ion wi h o he compounds. P- alues a e gi en o ea men s, ime, and hei in e ac ions. Signi ican di e ences a e gi en in bold (P≤0.05). δ13C Pa ame e F uc ose Glucose Suc ose Mala e Ci a e S a ch δ13CLea δ13CR Tempe a u e 0.019 0.004 0.028 0.015 n.a. 0.107 0.022 0.044 Soil mois u e 0.001 0.001 0.001 0.049 0.009 0.046 0.005 0.013 Time 0.001 0.195 0.081 0.198 0.052 0.001 0.066 0.001 Temp.:mois u e 0.035 0.063 0.543 0.017 n.a. 0.270 0.165 0.875 Temp.: ime 0.256 0.008 0.003 0.807 n.a. 0.113 0.812 0.014 Mois u e: ime 0.061 0.291 0.002 0.060 0.411 0.032 0.596 0.883 Concen a ion Pa ame e F uc ose Glucose Suc ose Mala e Ci a e S a ch Tempe a u e 0.663 0.352 0.142 0.011 n.a. 0.002 Soil mois u e 0.001 0.001 0.031 0.999 0.052 0.001 Time 0.016 0.927 0.001 0.035 0.110 0.001 Temp:mois u e 0.475 0.705 0.462 0.796 n.a. 0.001 Temp.: ime 0.901 0.847 0.113 0.387 n.a. 0.324 Mois u e: ime 0.831 0.629 0.063 0.889 0.895 0.071 n.a., no a ailable a Na u al Resou ces Ins i u e Finland (Luke) on No embe 29, 2016h p://jxb.ox o djou nals.o g/Downloaded om Mala e as a key ca bon sou ce o lea da k- espi ed CO2 in po a o | 5775 (Fig.4C) in he ange o 23 o 159µmol g DW-1 showed clea daily a ia ions (P≤0.001, Table3), wi h highes concen a ions o all ea men s by he end o he day, excep o Thigh and d y condi ions. Glucose concen a ions we e signi ican ly highe unde d y han unde we condi ions (P≤0.001, Table3), while con e se esul s we e obse ed o suc ose (P=0.031, Table3). Gene ally, no e ec o empe a u e on he concen a ion o any soluble ca bohyd a e was obse ed. Mala e concen a ions o all ea men s (Fig.4D), anging om 23 o 163µmol g DW-1, showed a daily pa e n wi h declining concen a ions in he beginning o he nigh and an inc ease a e 2–4 h in he da k (P=0.035, Table3). In con as o soluble ca bohyd a es, mala e concen a ions we e signi i- can ly highe unde Thigh han unde Tlow (P=0.011, Table3), bu we e no a ec ed by soil mois u e ea men s (P=0.999, Table3). Ci a e concen a ions unde Thigh o ~15µmol g DW-1 we e he lowes o all measu ed pu a i e ca bon sou ces a ailable o lea da k espi a ion and showed no changes due o soil mois u e ea men s and ime (Fig.4E; Table3). S a ch concen a ions (Fig. 4F), anging om 67 o 282 µmol g DW-1, showed signi ican empo al a ia ions (P≤0.001, Table3), independen o any ea men . The a e age s a ch concen a ion o 243µmol g DW-1 unde Tlow and we condi ions was clea ly highe (~2.5 imes) compa ed o hose unde o he ea men s. In addi ion, in e ac ions be ween empe a u e and soil mois u e ea men s led o smalle di - e ences be ween he alues o we and d y condi ions unde Thigh compa ed o hose unde Tlow (P≤ 0.001, Table3). Linea ela ionships be ween δ13CR and δ13CRS Linea eg ession analyses we e pe o med o unde s and he biochemical link be ween δ13CR and δ13CRS ac oss all ea men s (Table 4; Supplemen a y Fig. S1). δ13CRS o mala e explained mos o he daily a ia ion o δ13CR ( 2=0.26, P≤ 0.001), while he explana o y powe o uc- ose, glucose, and ci a e was lowe . The lowes linea ela- ionships du ing he daily cycle we e ound be ween δ13CR and δ13CRS o suc ose and s a ch. Due o he high daily a ia ions in δ13CR we ca ied ou he same analysis sepa- a ely o day ime and nigh ime. Day ime linea ela ion- ships we e gene ally s onge han du ing nigh ime, wi h δ13CR s ongly ela ed o δ13CRS o mala e, ci a e, and δ13Clea ( 2>0.6, P≤0.001), bu lowe ela ed o δ13CRS o sol- uble ca bohyd a es and s a ch. Du ing nigh ime, δ13CRS o mala e explained 36% o he a ia ion in δ13CR, bu δ13CRS ● ● ●● ●●●● ● ● ●● ●●●● ● ● ● ● ● ● ●● −37 −34 −31 −28 −25 −22 (A) F uc ose ● ● ● ● ●●●● ● ● ● ● ●●●● ● ● ● ● ● ● ● ● (B) Glucose ● ●● ● ● ●● ● ● ●● ● ● ●● ● ● ●● ● ● ● ● ● (C) Suc ose ●● ● ●● ●● ● ●● ● ●● ●● ● ● ● ●●● ●●● 0816 24 32 −37 −34 −31 −28 −25 −22 (D) Mala e ●● ● ●●●● ● 0816 24 32 (E) Ci a e ●Tlow We Tlow D y Thigh We Thigh D y ●●●●● ●●● ●●●●● ●●● ● ●● ● ● ● ● ● 0816 24 32 (F) S a ch δ 13 CRS (‰) Sampling pe iod (h) Fig.3. Daily cycles o he ca bon iso opic composi ion o di e en lea espi a o y ca bon sou ces (δ13CRS) unde di e en en i onmen al condi ions du ing he sampling pe iod: (A) uc ose, (B) glucose, (C) suc ose, (D) mala e, (E) ci a e, and (F) s a ch. Po a o plan s we e ea ed wi h a combina ion o Tlow (low empe a u e; closed symbols), Thigh (high empe a u e; open symbols), and we (ci cles) o d y ( iangles) condi ions. Resul s o uc ose a e a ec ed by co-elu ion wi h o he compounds. G ey a eas indica e nigh ime. Means ± SE a e gi en (n=2–3). Table4. Rela ionships be ween δ13C o lea da k- espi ed CO2 and δ13C o espi a o y ca bon sou ces Linea eg ession analyses ela ing δ13C o lea da k- espi ed CO2 o δ13C o pu a i e espi a o y ca bon sou ces and o δ13C o bulk lea es (δ13Clea ) ac oss all en i onmen al condi ions o day ime (0 h, 16 h, 24 h), o nigh ime (2 h, 4 h, 8 h, 26 h, 28 h, 32 h), and o he o al daily cycle (sampling pe iod o e 32 h). Resul s o uc ose a e a ec ed by co-elu ion wi h o he compounds. Gene ic eg ession equa ion y=mx+b was used. 2 alues a e gi en, s a s indica e P- alues. All co ela ion coe icien s we e posi i e. 2 Pu a i e ca bon sou ces Day ime Nigh ime Daily F uc ose 0.35*** 0.34*** 0.12*** Glucose 0.54*** 0.34*** 0.13*** Suc ose 0.59*** 0.20*** 0.04* Mala e 0.69*** 0.36*** 0.26*** Ci a e 0.67*** 0.28** 0.17** S a ch 0.48*** 0.16** 0.06* δ13CLea 0.63*** 0.33*** 0.20*** *, P≤0.05; **, P≤0.01; ***, P≤0.001 a Na u al Resou ces Ins i u e Finland (Luke) on No embe 29, 2016h p://jxb.ox o djou nals.o g/Downloaded om 5776 | Lehmann e al. o uc ose and glucose, as well as δ13Clea , showed simila ly high explana o ypowe . In luence o en i onmen al d i e s and ca bon sou ces on δ13CR Fu he mo e, a s epwise (backwa d) mul iple linea eg ession analysis was pe o med o iden i y en i onmen al d i e s and ca bon sou ces in luencing δ13CR (Table5). Day ime/nigh ime showed he s onges posi i e e ec on δ13CR (β=0.73, P≤0.001), while δ13CRS o mala e was he ca bon sou ce ha a ec ed δ13CR mos (β=0.4, P≤0.001). By compa ison, he in luence o δ13CRS o s a ch and soil mois u e condi ions on δ13CR alues wasmino . Discussion This s udy clea ly demons a es ha di e en empe a u e and soil mois u e condi ions in luence δ13C o lea da k- espi ed CO2 (δ13CR), δ13C o di e en pu a i e lea espi a o y ca - bon sou ces (δ13CRS), and concen a ions o ca bon sou ces du ing a daily cycle in po a o lea es. Fu he mo e, ou ind- ings s ongly indica e mala e as a key ca bon sou ce o day- ime and nigh ime δ13CR ac oss di e en en i onmen al condi ions. In luence o empe a u e and soil mois u e on iso opic composi ions A e 2 weeks o ea men , we al eady ound a clea empe a- u e e ec on δ13Clea , wi h up o 2.2‰ mo e nega i e δ13Clea alues unde Thigh condi ions compa ed o hose unde Tlow condi ions (Fig.2B). This is in ag eemen wi h a s udy show- ing mo e nega i e δ13C alues wi h inc easing empe a u e o bulk lea es o Xan hium species (Smi h e al., 1976). Simila o Tche kez e al. (2003) unde sho - e m empe a u e ea - men s, we obse ed mo e nega i e δ13CR alue wi h inc easing empe a u e (Fig.2A), bu due o ou long- e m ea men we ound also mo e nega i e δ13CRS alues (Fig.3). On he o he hand, d y condi ions in bo h o he empe a u e ea men s caused less nega i e δ13Clea , δ13CR, and δ13CRS alues com- pa ed o hose unde we condi ions, which is consis en wi h p e ious s udies unde con olled condi ions (Du anceau e al., 1999; Ghashghaie e al., 2001). The iso opic esul s unde he di e en en i onmen al con- di ions can be di ec ly linked o he lea gas exchange obse ed du ing he 32 h sampling pe iod (day 15 o he ea men pe iod). Inc easing empe a u e caused lowe An alues unde bo h soil mois u e condi ions (Fig.1A; Table1), indica ing ha plan s unde Thigh we e beyond he pho osyn he ic op imum. This esul is in ag eemen wi h ea lie s udies, showing ha cold-adap ed po a o plan s ha e educed a es o pho osyn he- sis wi h empe a u es abo e 20°C (Le y and Veilleux, 2007). Addi ionally, An migh be also in luenced by lea ageing, since Table5. En i onmen al d i e s and ca bon sou ces in luencing δ13C o lea da k- espi ed CO2 Resul o s epwise (backwa d) mul iple linea eg ession analysis showing he bes - i combina ion o independen en i onmen al d i e s ( empe a u e, soil mois u e, day ime/nigh ime), ime, and δ13C o glucose, suc ose, mala e, and s a ch as a iables in luencing δ13C o lea da k- espi ed CO2 (δ13CR) du ing he sampling pe iod in po a o lea es. S anda dized β-coe icien s and P- alues a e gi en. D i e s and ca bon sou ces in luencing δ13CR S anda dized β-coe icien P- alue Day ime/nigh ime 0.73 <0.001 Mala e 0.40 <0.001 S a ch 0.11 0.019 Soil mois u e 0.14 0.013 ● ● ●● ● ●● ● ● ● ●● ● ●● ● ●●● ● ● ● ●● 0 50 100 150 200 (A) F uc ose ● ●●● ● ● ● ● ● ●●● ● ● ● ● ●●●● ● ● ●● (B) Glucose ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● 0 50 100 150 200 (C) Suc ose ● ● ● ●● ● ●● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● 0816 24 32 0 50 100 150 200 (D) Mala e ● ● ●●●●●● 0816 24 32 (E) Ci a e ●Tlow We Tlow D y Thigh We Thigh D y ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ●●● ●●●● ● 0816 24 32 0 100 200 300 400 (F) S a ch Sampling pe iod (h) Concen a ion (µmol g DW − 1) Fig.4. Daily cycles o he concen a ion o di e en lea espi a o y ca bon sou ces unde di e en en i onmen al condi ions du ing he sampling pe iod: (A) uc ose, (B) glucose, (C) suc ose, (D) mala e, (E) ci a e, and (F) s a ch. Po a o plan s we e ea ed wi h a combina ion o Tlow (low empe a u e; closed symbols), Thigh (high empe a u e; open symbols), and we (ci cles) o d y ( iangles) condi ions. G ey a eas indica e nigh ime. Resul s o uc ose a e a ec ed by co-elu ion wi h o he compounds. To acili a e compa ison wi h o he me aboli es, suc ose concen a ions we e mul iplied by 2 o coun o hexose uni s, while s a ch concen a ions a e gi en in mola i y o s a ch monome s. No e di e en y-axis scale in (F). Means ±SE a e gi en (n=2–3). a Na u al Resou ces Ins i u e Finland (Luke) on No embe 29, 2016h p://jxb.ox o djou nals.o g/Downloaded om Mala e as a key ca bon sou ce o lea da k- espi ed CO2 in po a o | 5777 An dec eased unde all ea men s du ing he ea men pe iod. On he o he hand, gs ended o highe alues wi h inc easing empe a u e, bu only unde we condi ions (Fig.1C; Table1). An inc ease o gs unde Thigh migh be igge ed by inc easing anspi a ion a es, which could be a physiological esponse o compensa e educed a es o An by cooling he lea empe a- u e unde Thigh condi ions. Howe e , his was only obse ed in plan s unde Thigh and well-wa e ed condi ions, when SWC was high. Subsequen ly, lowe ca bon ixa ion and highe CO2 di usion in o he s oma al ca i ies unde Thigh, in compa ison o Tlow, caused an inc ease o Ci (Fig.1B) and mo e nega i e δ13CR and δ13CRS alues (Table6). Fu he mo e, d y soil mois- u e condi ions caused educed a es o An and gs compa ed o hose unde we condi ions (Fig.1A, C; Table1), independ- en o empe a u e ea men s. This can be explained wi h he se e e d ough s ess, e lec ing low SWC alues (Fig. 1D). Consequen ly, plan s unde d y condi ions expe ienced educed CO2 di usion in o he s oma al ca i ies, leading o lowe Ci and less nega i e δ13C alues (Table6). Plan s unde Thigh and d y condi ions showed he low- es pe o mance du ing he sampling pe iod compa ed o plan s unde o he ea men s, which is e lec ed in low An alues (Fig.1A), plan biomass, ube weigh and ube coun (Table2). δ13CR and δ13CRS in hese plan s we e expec ed o be he mos posi i e compa ed o o he ea men s due o a se e e d ough caused by he double e ec o high empe a- u e and d y soil mois u e. Ins ead, δ13CR and δ13CRS o he plan s unde he highes s ess le el (Thigh and d y condi ions) we e a he simila o hose unde lowes s ess le el (Tlow and we condi ions). This was pa icula ly obse ed o δ13CRS o soluble ca bohyd a es and s a ch (Fig.3). Again, his is an indica o o low An unde Thigh and d y condi ions, esul ing in a mode a e educ ion o Ci, while a he same ime gs s ongly educes CO2 di usion in o he s oma al ca i ies, causing an inc ease o Ci. Consequen ly, his led o in e media e δ13CR and δ13CRS alues unde Thigh and d y condi ions (Table6). In summa y ou indings indica e ha combined e ec s o em- pe a u e and soil mois u e condi ions on δ13CR and δ13CRS could cancel ou he indi idual e ec o each d i e . En i onmen al in luences on concen a ions o pu a i e ca bon sou ces Soil mois u e and empe a u e a ec ed concen a ions o pu a i e lea espi a o y ca bon sou ces di e en ly. Suc ose concen a ion dec eased unde d y condi ions (Fig. 4C; Table3), which is in con as o he ecen s udy by Lemoine e  al. (2013). This may be explained by educed a es o suc ose syn hesis due o lowe ing o he suc ose phospha e syn hase eac ion (SPS) (Vu e  al., 1998). The dec ease in he enzyme ac i i y is p obably igge ed by limi ed a es o phloem suga anspo obse ed unde d ough (Rueh e al., 2009). This in u n could be an explana ion o lowe plan biomass and ube weigh /coun in esponse o highe empe a u es and d y condi ions (Tables 1, 2). Subsequen ly, he inc ease o uc ose and glucose concen a ions unde d ough may also be a consequence o lowe SPS ac i i y (Fig. 4A, B; Table 3), since he demand o bo h hexoses o suc ose syn hesis was educed. Addi ionally, inc easing uc ose and glucose concen a ions unde d ough migh ha e osmo ic unc ionali y, main aining me abolic ac i i y (Lemoine e al., 2013). On he o he hand, mala e concen a ions inc eased wi h empe a u e (Fig.4D; Table3), which is mos likely a conse- quence o highe PEPC ac i i y (Chin hapalli e al., 2003). Highe mala e concen a ions may also suppo espi a- o y p ocesses in he KC o egula ion o s oma al opening (Finkemeie and Swee lo e, 2009). Mo eo e , dec eased s a ch concen a ions in lea es unde ea men s wi h highe en i onmen al s ess han Tlow and we condi ions (Fig.4F) we e simila o p e ious indings (Lemoine e al., 2013). The esul also suppo s he assump ion ha educed amoun s o assimila ed ca bon due o lowe An unde Thigh o d y con- di ions we e used o main enance o biochemical p ocesses a he han o ca bon s o age. Addi ionally, his indica es ha plan s unde Thigh o d y condi ions we e unde se e e en i onmen al s ess. Mala e as a key espi a o y ca bon sou ce o day ime and nigh ime δ13CR The daily cycle o δ13CR was highly a iable, showing less nega i e day ime and mo e nega i e nigh ime alues, while δ13CRS alues gene ally showed lowe changes du ing he same pe iod (Figs 2A, 3; Table3). δ13CRS alues o all ea - men s compa ed o δ13CR alues we e mo e nega i e o sol- uble ca bohyd a es (up o 9.3‰) and ci a e (up o 4.1‰), bu also less nega i e o s a ch (up o 4‰) and mala e (up o 5.2‰) du ing he daily cycle (Figs 2A, 3). In pa icula , mala e was s ongly en iched in 13C, by up o 8.8‰, compa ed o all o he pu a i e ca bon sou ces (Fig.3). This was simila o a p e ious s udy in es iga ing me aboli es in po a o lea es (Gleixne e al., 1998) and indica es a possible biochemical link be ween 13C en iched lea da k- espi ed CO2 and 13C en iched mala e. Fo a be e unde s anding o he o e all biochemical con- nec ions be ween δ13CR and di e en pu a i e ca bon sou ces, we ca ied ou linea eg ession analyses, independen o en i onmen al condi ions (Table4; Supplemen a y Fig. S1). Table6. Cohe ence be ween lea physiological pa ame e s and δ13C alues. Lea physiological pa ame e s and δ13C alues du ing he sampling pe iod in po a o plan s unde di e en ea men s compa ed o hose in po a o plan s g owing unde Tlow and we condi ions The ollowing a iables we e conside ed: An, ne assimila ion a e; Ci, in e cellula CO2 concen a ion; gs, s oma al conduc ance; δ13CR, δ13C o lea da k- espi ed CO2; δ13CRS, δ13C o di e en pu a i e espi a o y ca bon sou ces ( uc ose, glucose, suc ose, s a ch, and mala e). A ows indica e s ong (↑, ↓), in e media e (↗, ↘), o no changes (→) due o he in luence o ea men combina ions (Tlow, low empe a u e; Thigh, high empe a u e; and we o d y condi ions). T ea men s AngsCiδ13CRδ13CRS Tlow d y ↘ ↓ ↓ ↑ ↑ Thigh we ↘ ↑ ↑ ↓ ↓ Thigh d y ↓ ↓ ↗ → → a Na u al Resou ces Ins i u e Finland (Luke) on No embe 29, 2016h p://jxb.ox o djou nals.o g/Downloaded om