Intrapopulation genotypic variation of foliar secondary chemistry during leaf senescence and litter decomposition in silver birch (Betula pendula)

pls-08-01074 June 23, 2017 Time: 16:56 # 1 ORIGINAL RESEARCH published: 26 June 2017 doi: 10.3389/ pls.2017.01074 Edi ed by: José M. G ünzweig, Heb ew Uni e si y o Je usalem, Is ael Re iewed by: Ch is ophe Thaye Cole, Uni e si y o Wisconsin–Madison, Uni ed S a es Tanya Handa, Uni e si é du Québec à Mon éal, Canada *Co espondence: Juha Mikola [email p o ec ed] Special y sec ion: This a icle was submi ed o Func ional Plan Ecology, a sec ion o he jou nal F on ie s in Plan Science Recei ed: 03 Janua y 2017 Accep ed: 06 June 2017 Published: 26 June 2017 Ci a ion: Paaso U, Keski-Saa i S, Keinänen M, Ka inen H, Sil e T, Rousi M and Mikola J (2017) In apopula ion Geno ypic Va ia ion o Folia Seconda y Chemis y du ing Lea Senescence and Li e Decomposi ion in Sil e Bi ch (Be ula pendula). F on . Plan Sci. 8:1074. doi: 10.3389/ pls.2017.01074 In apopula ion Geno ypic Va ia ion o Folia Seconda y Chemis y du ing Lea Senescence and Li e Decomposi ion in Sil e Bi ch (Be ula pendula) Ulla Paaso1, Sa i a Keski-Saa i2, Ma kku Keinänen2, Heini Ka inen1, Ta ja Sil e 1, Ma i Rousi3and Juha Mikola1* 1Depa men o En i onmen al Sciences, Uni e si y o Helsinki, Lah i, Finland, 2Depa men o En i onmen al and Biological Sciences, Uni e si y o Eas e n Finland, Joensuu, Finland, 3Na u al Resou ces Ins i u e Finland (Luke), Helsinki, Finland Abundan seconda y me aboli es, such as condensed annins, and hei in e popula ion geno ypic a ia ion can emain h ough plan lea senescence and a ec li e decomposi ion. Whe he he in apopula ion geno ypic a ia ion o a mo e di e se asso men o seconda y me aboli es equally pe sis s h ough lea senescence and li e decomposi ion is no well unde s ood. We analyzed concen a ions o in acellula phenolics, epicu icula la onoid aglycones, epicu icula i e penoids, condensed annins, and lignin in g een lea es, senescen lea es and pa ly decomposed li e o sil e bi ch, Be ula pendula. B oad-sense he i abili y (H2) and coe icien o geno ypic a ia ion (CVG) we e es ima ed o me aboli es in senescen lea es and li e using 19 geno ypes selec ed om a B. pendula popula ion in sou he n Finland. We ound ha mos o he seconda y me aboli es emained h ough senescence and decomposi ion and ha hei pe sis ence was ela ed o hei chemical p ope ies. In apopula ion H2 and CVG o in acellula phenolics, epicu icula la onoid aglycones and condensed annins we e high and ema kably, inc eased om senescen lea es o decomposed li e . The ank o geno ypes in me aboli e concen a ions was pe sis en h ough li e decomposi ion. Lignin was an excep ion, howe e , wi h a diminishing geno ypic a ia ion du ing decomposi ion, and he concen a ions o lignin and condensed annins had a nega i e geno ypic co ela ion in he senescen lea es. Ou esul s show ha seconda y me aboli es and hei in apopula ion geno ypic a ia ion can o he mos pa emain h ough lea senescence and ea ly decomposi ion, which is a p e equisi e o ini ial li e quali y o p edic a ia ion in li e decomposi ion a es. Pe sis en geno ypic a ia ion also opens an a enue o selec ion o impac li e decomposi ion in B. pendula popula ions h ough ac ing on hei g een oliage seconda y chemis y. The nega i e geno ypic co ela ions and diminishing he i abili y o lignin concen a ions may, howe e , coun e ac his p ocess. Keywo ds: condensed annins, geno ypic a ia ion, he i abili y, lea li e decomposi ion, lignin, phenolic compounds, seconda y me aboli es, i e penoids F on ie s in Plan Science | www. on ie sin.o g 1June 2017 | Volume 8 | A icle 1074 pls-08-01074 June 23, 2017 Time: 16:56 # 2 Paaso e al. Seconda y Chemis y o Bi ch Li e INTRODUCTION Plan s p oduce an abundance o di e se seconda y me aboli es such as phenolics and e penoids. These compounds we e hough o be was e p oduc s un il F aenkel (1959) ecognized ha hey ha e an impo an ole in he bi o e de ense. Since his ea ly disco e y, unde s anding o he ole o seconda y me aboli es in plan ecology has g ea ly expanded (Theis and Le dau, 2003). Besides ac ing as an i-he bi o y agen s (Haukioja, 2003;Ma emyano e al., 2015), seconda y me aboli es can de end plan s agains mic obial a ack (Dixon, 2001;Ped as e al., 2003), egula e in e ac ions wi h bene icial mic obes such as myco hizal ungi (Akiyama e al., 2005) and p o ec plan s agains UV adia ion (Tegelbe g e al., 2001;Keski-Saa i e al., 2005). Seconda y me aboli es can also emain in lea li e , and by a ec ing soil mic obial ac i i y, li e decomposi ion, and nu ien cycling ha e e ec s on ecosys em unc ioning (No hup e al., 1998;Hä enschwile and Vi ousek, 2000;Schwei ze e al., 2004;Ko ilainen e al., 2009). While hese ‘a e li e’ e ec s a e widely ecognized, unde s anding o me aboli e dynamics in plan li e is o en based on one dominan g oup o compounds such as condensed annins (e.g., Schwei ze e al., 2004, 2008b) and s udies on he pe sis ence o a wide ange o indi idual compounds (e.g., Galle and Leb e on, 1995) ha e emained sca ce. Focusing mo e on he di e si y o less abundan me aboli es is encou aged by a ecen hypo hesis ha p oduc ion o low-abundance, low-molecula weigh seconda y compounds may se e as a mechanism o ees in opical o es s o en o ce ene gy s a a ion o soil decompose s and conse e he nu ien s in li e , whe e hey a e accessible by plan -associa ed myco hizal ungi (Hä enschwile e al., 2011). The dynamics o seconda y me aboli es du ing lea senescence and li e decomposi ion can be linked o hei chemical s uc u e (Galle and Leb e on, 1995), bu only a ely ha e such s udies been ca ied ou in su icien de ail o sc u inized in e ms o he chemical p ope ies o he compounds. Small di e ences in he chemical s uc u e o he me aboli es may s ongly a ec he a es o chemical p ocesses du ing decomposi ion. Fo example, in he senescen lea es and shed lea li e , dis up ion o cellula memb anes eleases phenolics om acuoles in o cy osol, whe e hey a e exposed o he enzyma ic ac i i y o polyphenol oxidases (PPOs). Compounds wi h ca echol moie y (o ho-diphenol) a e p e e en ial subs a es o PPOs (Rawel and Rohn, 2010), and a e he e o e oxidized mo e apidly han compounds ha lack icinal hyd oxyl g oups in hei phenolic ing s uc u e. Such di e ences may de e mine he a e o compounds du ing decomposi ion as was demons a ed o 14C-labeled pa a- and o ho-hyd oxybenzoic acids in aiga o es soils (Sugai and Schimel, 1993). Phenolic compounds a y widely in hei s uc u e and some o hem, such as la onoid glycosides and phenolic acids, dec ease in concen a ion al eady du ing lea senescence, while o he s, pa icula ly polyme s (such as lignin and condensed annins), wi hs and decomposi ion (Galle and Leb e on, 1995). S uc u al di e ences may also a ec o he chemical p ope ies ele an in decomposi ion, such as he hyd ophobici y and oxici y o he decomposing mic o- o ganisms. Th ee al e na i e hypo heses we e ecen ly o mula ed ega ding he ac o s ha con ol li e chemis y du ing decomposi ion (Wickings e al., 2012). The chemical con e gence hypo hesis s a es ha plan li e s s a esembling each o he o e he cou se o decomposi ion, while he ini ial li e quali y hypo hesis p oposes ha he ini ial chemis y can be used o simula e he quali y h oughou he decomposi ion. The decompose con ol hypo hesis unde lines he impo ance o dis inc decompose communi ies in luencing he li e chemis y du ing decomposi ion. Wickings e al. (2012) ound ha he chemis y o di e en li e ypes di e ged a he han con e ged, bu hei esul s also show ha he h ee hypo heses a e no mu ually exclusi e: he decompose communi ies ha e a key ole in egula ing changes in li e chemis y al hough he e ec s depend s ongly on he ini ial li e quali y. Recen s udies ha e also ound suppo o he chemical con e gence hypo hesis (Wallens ein e al., 2013;Pa sons e al., 2014), bu conclude ha his may be an o e simpli ica ion. La ge in e speci ic a ia ion in he composi ion and quan i y o seconda y me aboli es is known o lead o p o ound di e ences in soil o ganic ma e accumula ion and nu ien cycling in e es ial ecosys ems (Wa dle e al., 1997). In aspeci ic geno ypic a ia ion in seconda y me aboli es can be equally subs an ial as shown o ee species in Be ula (Keinänen e al., 1999;Lai inen e al., 2000, 2005), Populus (Schwei ze e al., 2008b), Alnus (Lece and Chau e , 2008), and Salix (Heiska e al., 2007). Gene ic a ia ion is a p e equisi e o na u al selec ion and e olu ion, bu ecen e idence sugges s ha i can also shape local communi ies and con ol ecosys em unc ioning, especially when ound in a dominan plan species (Whi ham e al., 2008; Genung e al., 2011;Pas o , 2017). T ee geno ypes a e known o di e in he composi ion o ungal and insec communi ies in hei canopies (Ba bou e al., 2009) and h ough lea li e all, o a ec he composi ion and unc ioning o soil mic obial communi ies (Schwei ze e al., 2008a;Mad i ch and Lind o h, 2011). Di e ences in li e quali y wi hin a popula ion can also lead o di e ences in ca bon and ni ogen luxes (Mad i ch and Hun e , 2005). In many o hese e ec s, seconda y me aboli es play a c ucial ole (Schwei ze e al., 2008a;Ba bou e al., 2009; Mad i ch and Lind o h, 2011), sugges ing ha hese compounds may be pa icula ly help ul in e ealing how na u al selec ion, ac ing on he gene ic s uc u e o a dominan plan popula ion, can d i e communi y composi ion and ecosys em unc ioning. In his s udy, we ocus on he a e o olia seconda y me aboli es and he pe sis ence o hei geno ypic a ia ion h ough lea senescence and li e decomposi ion in a Be ula pendula Ro h popula ion. Be ula pendula is a common, as - g owing deciduous ee in he no he n, and eas e n Eu ope (A kinson, 1992), whe e i o en domina es ea ly bo eal o es succession. Due o i s ecological and economic impo ance in he no he n a eas, he in a- and in e popula ion geno ypic a ia ion o B. pendula ai s ha e been a subjec o in ensi e esea ch. The s udies ha e co e ed ee g ow h (P i inen e al., 2003;Sil e e al., 2009;Mikola e al., 2014) and physiology (Sil e e al., 2008;Possen e al., 2014) as well as he bi o e suscep ibili y (Rousi e al., 1991, 1997;Pusenius e al., 2002;Sinkkonen e al., 2012) and decomposi ion o F on ie s in Plan Science | www. on ie sin.o g 2June 2017 | Volume 8 | A icle 1074 pls-08-01074 June 23, 2017 Time: 16:56 # 3 Paaso e al. Seconda y Chemis y o Bi ch Li e lea li e (Sil e e al., 2007, 2015). Folia seconda y me aboli es (Keinänen and Julkunen-Tii o, 1998;Lai inen e al., 2000), hei geno ypic a ia ion and ole in he bi o e and s ess esponses (Mu ikainen e al., 2000;Yamaji e al., 2003) a e equally well known o B. pendula. Fo example, he la onoid aglycones and i e penoids ound on B. pendula lea su ace can impai he g ow h and su i al o Lyman ia dispa la ae (Ma emyano e al., 2015). The abili y o seconda y me aboli es o explain he link be ween he bi o e esis ance and li e decomposi ion a e in B. pendula has also been es ed, wi h no ob ious ole ound (Sil e e al., 2015), bu he basic knowledge o al e a ions in he seconda y me aboli e p o iles and hei geno ypic a ia ion in lea senescence and li e decomposi ion is lacking. Ou s udy was designed o ill his gap o knowledge. We hypo hesized ha seconda y me aboli es, anging om he ample condensed annins and lignin o o he phenolics and i e penoids o lowe concen a ions (1) emain h ough lea senescence, (2) exhibi signi ican geno ypic a ia ion in he senescen B. pendula lea es, and (3) emain and p ese e hei geno ypic a ia ion h ough he ea ly phase o li e decomposi ion. These hypo heses, i suppo ed, would mani es he pe sis ence o geno ypic a ia ion o olia chemis y h ough senescence and decomposi ion: a p e equisi e o he ini ial li e quali y o p edic and he selec ion ac ing on olia chemis y o a ec li e decomposi ion. MATERIALS AND METHODS Field Si es, Plan Ma e ial, and Lea Sampling The lea ma e ial was collec ed om he Kuikannii y expe imen al si e, es ablished in 1999 on an abandoned ag icul u al ield in Punkaha ju, sou heas Finland (61◦470N, 29◦210E). The ees ha g ow in Kuikannii y consis o he mic op opaga ed p ogeny o 30 B. pendula ees (Lai inen e al., 2005), selec ed om a nea by 0.9-ha o es s and (Lai inen e al., 2000). The sou ce s and was na u ally egene a ed a e 1979 logging and he selec ed mo he ees g ow in six g oups, loca ed 10–60 m apa . The Kuikannii y si e consis s o six eplica e blocks, each including wo ees o each o he 30 geno ypes. Nine een o hese geno ypes we e used in his s udy (excep ha g een lea measu emen s we e es ic ed o eigh geno ypes a e quali y assessmen o he analyses, whe e accumula ion o polyme ic subs ances had caused inconsis encies in he de e mina ion o peak a eas), and o he wo ees wi h he same geno ype in each block, one was andomly chosen o he s udy. In 2008, when he lea es we e collec ed, he ees we e on a e age 11 m all. The he mal g owing season (i.e., he pe iod when he mean daily empe a u e emains abo e 5◦C) s a ed on Ap il 27, ended on Oc obe 29 and had a mean empe a u e o 10.8◦C (Finnish Me eo ological Ins i u e). To analyze he g een lea chemis y, a sample o 30 lea es (e e y second non-damaged lea om he ip o a sou h-side b anch, g owing a he minimum heigh o 150 cm) was collec ed om he ees in i e eplica e blocks (n=5) on June 26. The collec ed lea es we e immedia ely ozen in liquid ni ogen. Fo collec ing senescen lea es, wo sou h-side b anches (a he heigh o 140–300 cm) o each ee we e enclosed in mesh bags be o e lea all (Sep embe 8 o 10) in all six blocks (n=6, he numbe o eplica es was inc eased o i he numbe o blocks in he si e o decomposi ion; see below). The mesh bags we e collec ed a e lea all (Oc obe 28 o 30), hei con en s we e pooled wi hin a ee and andom subsamples o lea es we e aken o labo a o y analyses. Remaining lea es we e s o ed in plas ic bags in 4◦C un il No embe 5, when 10-g (d y mass equi alen ) samples we e used o es ablish li e pa ches on he g ound o a clea - cu , B. pendula-Pinus syl es is o es si e in Loppi, sou h Finland (60◦360N, 24◦240E). The soil in his si e is pos -glacial so ed ine sand wi h a pH o 5.0 and o al C and N concen a ions o 6 and 0.3%, espec i ely, in he uppe 0–5 cm laye (Mikola e al., 2014). The g ound laye ege a ion is domina ed by a e n P e idium aquilinum (L.) Kuhn, g asses Calamag os is a undinacea (L.) Ro h and Deschampsia lexuosa (L.) T in., and dwa sh ubs Vaccinium my illus L. and Vaccinium i is-idea L. (Mikola e al., 2014). Using a o es si e ins ead o he Kuikannii y si e (which was es ablished on an ag icul u al ield), we ensu ed ha he li e and me aboli es we e subjec ed o decomposi ion in a o es en i onmen , whe e he decompose s a e adap ed o ee li e decomposi ion. The li e pa ches we e alloca ed o six eplica e blocks ( ollowing he blocking ac o in he Kuikannii y si e) and we e co e ed, bu no enclosed, wi h a 1-mm mesh o p e en disappea ance and mixing o lea es. The senescen lea es used in he pa ches we e no d ied o d y mass measu emen s in o de o p ese e he mic obes, such as endophy es (Saikkonen e al., 2003, 2015), g owing on he lea es. The li e pa ches we e allowed o decompose in he ield un il June 24, 2009 (i.e., o 231 days) when 20 pa ly decomposed lea es we e andomly selec ed om each pa ch, s o ed in –76◦C and used o seconda y me aboli e analyses. The mean li e mass loss a his s age o decomposi ion was 9% (Sil e e al., unpublished da a). Analyses o Seconda y Me aboli es The subsamples o lea es and li e ha we e used o ex ac ing seconda y me aboli es we e g ound in liquid ni ogen and s o ed in –76◦C un il analysed. Upon analysis, he samples we e d ied o e nigh in a acuum cen i uge concen a o . Samples o 40 ±5 mg we e hen g ound using a s ainless s eel bead in a TissueLyse o 5 min, ex ac ed in 1 ml o 80% me hanol o 30 min, cen i uged (13000 pm, 2 min), and again ex ac ed wi h 1 ml o 100% o 10 min. The supe na an s we e d ied in a acuum concen a o a 45◦C and s o ed a 4◦C. Lignin was de e mined om he p ecipi an s (5 o 10 mg o lea o li e sample, espec i ely) using he me hod desc ibed in B inkmann e al. (2002) and he weigh o he ob ained biomass pelle was used as an es ima e o he lignin con en . The d ied supe na an was hen esuspended in 1.8 ml o 100% me hanol and he concen a ion o condensed annins (syn. p oan hocyanidins) was de e mined om a 100-µl aliquo o me hanol esuspension using he acid bu anol assay (Hage man, 2002). In he assay, 900 µl bu anol-HCl (5%) and 10 µl Fe3+- eagen we e added and he suspension was incuba ed a 90◦C o 50 min. A e being cooled wi h ice, he abso bance o he suspension was measu ed a 550 nm using cyanidin chlo ide (Ex asyn hese, Genay, F ance) as a quan i ica ion s anda d. F on ie s in Plan Science | www. on ie sin.o g 3June 2017 | Volume 8 | A icle 1074 pls-08-01074 June 23, 2017 Time: 16:56 # 4 Paaso e al. Seconda y Chemis y o Bi ch Li e Fo quan i ying he concen a ions o small-molecula phenolics and i e penoids, a 500-µl aliquo o he me hanol esuspension was d ied in a acuum cen i uge a 45◦C and s o ed a –20◦C. A e s o age, he samples we e dissol ed in 250 µl 100% me hanol and 250 µl dis illed wa e . High-pe o mance liquid ch oma og aphy-mass spec ome y (HPLC-MS) was hen pe o med using The mo Finnigan LC wi h he low spli in o wo be ween a The mo LTQ MS (The mo Finnigan, San Jose, CA, Uni ed S a es) wi h elec osp ay ioniza ion (ESI) and a Finnigan PDA de ec o wi h a subsequen Co ona Ul a cha ged ae osol de ec o . The column was C-18 Luna wi h an inne diame e o 2 mm, leng h o 150 mm and a pa icle size o 3 µm (Phenomenex, Denma k). The empe a u e o he ay was se o 18◦C and he column o 40◦C. The sol en s we e (A) 0.1% o mic acid (Sigma–Ald ich, S einheim, Ge many) in H2O and (B) 0.1 % o mic acid in ace oni ile (Ch omasol R g ade, Sigma–Ald ich). The low was 0.41 ml min−1and he elu ion was pe o med wi h a g adien as ollows: B s a ed wi h 5%, was inc eased o 50% by 15 min, o 60% by 35 min, o 85% by 45 min and o 98% by 50 min and hen kep a 98% o 5 min. The column was e u ned o i s s a ing condi ion wi h 15 min equilib a ion, gi ing a o al o 70 min o each un. The injec ion olume was 12 µl wi h a pa ial loop. The MS was un in a posi i e ion mode wi h a mass ange o 150–1500 m/z. The capilla y empe a u e was kep a 320◦C and he ol age a 5 V. The shea h gas low a e was kep a 20 ml min−1, auxilia y gas low a e a 5 ml min−1and sweep gas low a e a 5 ml min−1. The ube lens was se o 80 V. The a eas o he compounds we e in eg a ed using he Xcalibu so wa e. The compounds we e anno a ed using e en ion imes, UV spec a and HPLC-MS. Peak picking was done using Me Align so wa e (Lommen, 2009) based on he g een lea samples. All analyzed compounds wi h hei e en ion imes and quan i ica ion ions a e lis ed in Supplemen a y Table S1. The compounds we e coded in he o de o e en ion ime, which e lec s inc easing lipophilici y in he e e sed phase LC. The quan i ica ion was ca ied ou using comme cial s anda ds: i.e., chlo ogenic acid (Ald ich) o ca eoylquinic acids and hei de i a i es (CQAs), couma oylquinic acids (CouQAs) and 3,40-dihyd oxyp opiophenone-3-β-D-glucopy anoside (DHPPG); (+)-ca echin (Ald ich) o (+)-ca echin; que ce in 3-glucoside (Ex asyn hese) o my ice in, que ce in, and kaemp e ol de i a i es; and acace in (Ex asyn hese) o la onoid aglycones. T i e penoids a e epo ed as a bi a y uni s (peak a ea g−1d y mass). Fo he analyses o he i abili y and s a is ical signi icance o geno ypic a ia ion in he senescen lea es and decomposed li e , he small-molecula phenolics and i e penoids we e g ouped in o in acellula phenolics (including CQAs, CouQAs, DHPPG, (+)-ca echin, my ice in glycosides, que ce in glycosides, and kaemp e ol glycosides), epicu icula la onoid aglycones and epicu icula i e penoids. In addi ion, he in acellula phenolics we e es ed as subg oups; i.e., phenolic acids (CQAs and CouQAs), my ice in glycosides, que ce in glycosides, and kaemp e ol glycosides. S a is ical Analyses To a oid a mul i ude o mean es s and o allow easy s a is ical in e ence in he g aphs (Cumming, 2009), we in e p e ed he s a is ical signi icance o di e ences be ween g een lea , senescen lea and decomposing li e me aboli e concen a ions using 85% con idence in e als (CIs) o means. In his app oach, non-c ossing CIs o wo means deno e a s a is ically signi ican di e ence be ween he means. I is a common p ac ice o use 95% CIs, bu hey a e oo conse a i e o es ing mean di e ences and he bes app oxima ion o α=0.05 is achie ed using 85% CIs (Pay on e al., 2000). All concen a ion means we e calcula ed using da a om geno ypes 5, 6, 8, 12, 14, 15, 20, and 25 as hese we e a ailable o g een lea es. Since ou plan ma e ial consis ed o a cloned p ogeny o he selec ed B. pendula geno ypes, all ees wi hin a geno ype had an equal gene ic s uc u e. In such ma e ial, all a ia ion ha is ound wi hin geno ypes can be conside ed o be due o he a ia ion in en i onmen , o due o a measu emen e o , and all a ia ion ound be ween he geno ypes o be gene ic (Falcone , 1989). This geno ypic a ia ion includes bo h he addi i e and non-addi i e componen s, which canno be sepa a ed in cloned ma e ial, and only he deg ee o gene ic de e mina ion, i.e., he b oad-sense he i abili y (H2) can be calcula ed (Falcone , 1989). In ou s udy, he b oad-sense he i abili ies o he concen a ions o he h ee me aboli e g oups (in acellula phenolics, epicu icula la onoid aglycones, and epicu icula i e penoids), in acellula phenolic subg oups, soluble condensed annins and lignin we e calcula ed on indi idual plan basis acco ding o he Eq. 1, whe e σ2 Gand σ2 Ea e a iance componen s o geno ypes and e o , espec i ely (calcula ed using he SPSS GLM Va iance componen s p ocedu e). Following common p ac ice in o es b eeding, he eplica e block was included in he calcula ion model as a ixed ac o (which emo es he block-scale a ia ion om e o a iance). This di e s om ou ea lie B. pendula s udies in a na u al o es si e (Mikola e al., 2014;Sil e e al., 2015), whe e we we e also in e es ed in he size o he block-scale en i onmen al a ia ion and ea ed block as a andom ac o . H2=σ2 G/(σ2 G+σ2 E)(1) Coe icien s o geno ypic a ia ion (CVG) we e calcula ed acco ding o he Eq. 2, whe e ¯xis he pheno ypic mean. CVG=qσ2 G/¯x (2) The s a is ical signi icance o geno ypic a ia ion in he concen a ions o he h ee g oups o small-molecula me aboli es, in acellula phenolic subg oups, condensed annins, and lignin was es ed using he analysis o a iance. Following he he i abili y calcula ions, he geno ype was ea ed as a andom ac o and he ield eplica e block as a ixed ac o . The homogenei y o esidual a iance among geno ypes was es ed using he Le ene’s es and he no mal dis ibu ion o model esiduals using he Shapi o– Wilk es . To ul ill he a iance and no mali y assump ions, he da a o condensed annins was log10 ans o med and he da a o o he me aboli es, excluding lignin, squa e- oo ans o med. Lignin da a ul illed he assump ions wi hou a ans o ma ion. F on ie s in Plan Science | www. on ie sin.o g 4June 2017 | Volume 8 | A icle 1074 pls-08-01074 June 23, 2017 Time: 16:56 # 5 Paaso e al. Seconda y Chemis y o Bi ch Li e Rela ions among geno ypes and indi idual seconda y me aboli es we e u he examined using he p incipal componen analysis (PCA). Compounds wi h quali a i e a ia ion we e excluded om hese analyses and he senescen lea es and decomposed li e we e analyzed sepa a ely as no all he compounds o senescen lea es we e p esen in li e . Be o e PCA, bo h columns and ows o he da a ma ix we e ans o med o ha e a mean o ze o and a s anda d de ia ion o 1. This was done o educe he quan i a i e di e ences among he compounds and he samples and hus, ocus mo e on chemical p o iles. Compounds wi h skewed dis ibu ion we e log10 ans o med. The signi icance o geno ypic a ia ion along he i s wo PC axes was analyzed using axis sco es and he same ANOVA models as used o compound concen a ions. Geno ypic co ela ions among he compound g oups and be ween he senescen lea es and decomposed li e we e es ed using he Spea man ank co ela ion analysis. The pe sis ence o geno ypic a ia ion in he chemical p o iles e ealed by PCA be ween he senescen lea es and li e was es ed as ank co ela ions o he geno ype means o PC axis sco es. All s a is ical analyses we e pe o med using he SPSS 15.0.1 and SPSS 18 s a is ical packages (SPSS, Chicago, IL, Uni ed S a es) excep o he PCA, which was pe o med using he SIMCA-P+ so wa e (Ume ics AB, Umeå, Sweden). RESULTS Changes in Me aboli e Concen a ions du ing Lea Senescence and Li e Decomposi ion All hose seconda y me aboli es ha we e ound in g een summe lea es we e also de ec ed in senescen lea es, excep o CQAs (Figu e 1A). The concen a ions o CouQAs we e on a e age 97% lowe in he senescen han g een lea es and dec eased below he de ec ion limi du ing li e decomposi ion (Figu e 1A). DHPPG concen a ion was 99% lowe in he senescen han g een lea es and also dec eased below de ec ion du ing decomposi ion (Figu e 1A). The dec ease o la onol glycoside concen a ions du ing lea senescence and li e decomposi ion a ied among he la onoid subg oups (Figu es 1B–D). Concen a ions o my ice in glycosides dec eased on a e age by 93% du ing lea senescence and none o he six compounds was de ec ed in he decomposed li e (Figu e 1B). In con as , concen a ions o que ce in and kaemp e ol glycosides dec eased on a e age by 79 and 76% du ing lea senescence and all compounds, excep o kaemp e ol-3-glucoside, we e also de ec ed in he decomposed li e (Figu es 1C,D). Based on he compa ison o con idence in e als, he educ ion in concen a ion du ing lea senescence was s a is ically signi ican o all la onol glycosides, excep o kaemp e ol 3-a abino u anoside (Figu es 1B–D). Du ing li e decomposi ion, he concen a ion o que ce in and kaemp e ol glycosides dec eased on a e age by 86 and 52%, and excep o kaemp e ol 3-glucu onide, he dec ease was s a is ically signi ican in all compounds (Figu es 1C,D). The concen a ions o epicu icula la onoid aglycones a ied a lo in he g een lea es, bu displayed ela i ely simila dynamics du ing lea senescence and li e decomposi ion (Figu e 1E). The concen a ions we e on a e age 27% lowe in he senescen han g een lea es and he dec ease was s a is ically signi ican o nine o he 15 compounds (Figu e 1E). Fo one o he compounds (F15), he concen a ion inc eased by 70% (Figu e 1E). Du ing li e decomposi ion, he concen a ions o la onoid aglycones dec eased on a e age by 51% and he dec ease was s a is ically signi ican o all compounds (Figu e 1E). The mean concen a ion o i e penoids dec eased du ing lea senescence by 25% (Figu e 1F), bu his dec ease was d i en by one abundan compound T8 ha was anno a ed as 12- O-ace yl-3-O-malonylbe ula olien iol. When T8 was excluded om calcula ions, he mean concen a ion o i e penoids inc eased by 4%, and o he wo oco illol- ype i e penoids, papy i e ic acid (T7) and i s de i a i e (T6), he inc ease was s a is ically signi ican (Figu e 1F). Du ing li e decomposi ion, all i e penoids had pa allel dynamics, he mean concen a ion dec eased by 55% and he dec ease was s a is ically signi ican o all compounds (Figu e 1F). O he polyme s, he concen a ion o lignin inc eased by 51%, while he concen a ion o condensed annins did no change du ing lea senescence (Figu e 1G). Du ing li e decomposi ion, lignin concen a ion inc eased u he by 23%, bu annin concen a ion dec eased by 87% (Figu e 1G). The (+)-ca echin concen a ion dec eased by 97% du ing senescence, bu did no change du ing decomposi ion (Figu e 1G). Geno ypic Va ia ion in Me aboli e Concen a ions The seconda y me aboli es displayed bo h quali a i e (absence o p esence in only ce ain geno ypes) and quan i a i e ( ound in all geno ypes, bu in a ying quan i y) geno ypic a ia ion. Quali a i e a ia ion was ound among la onol glycosides: he 3-glucu onides we e lacking in ou o he 19 geno ypes (16, 24, 25 and 30) and he 3-a abino u anosides we e ound in ou geno ypes only (2, 8, 22, and 23). The quali a i e a ia ion emained h ough he senescence and decomposi ion as hese compounds we e also ound in he decomposed li e (Figu es 1B–D). Quan i a i e geno ypic a ia ion was ound in all compound g oups (Figu e 2 and Table 1). O he main me aboli e g oups, he epicu icula i e penoids (Figu e 2C) had he highes b oad-sense he i abili y, H2(0.281) and coe icien o geno ypic a ia ion, CVG(0.138) in he senescen lea es (Table 1). The o he g oups had e y simila he i abili ies (0.113–0.121), whe eas he CVG a ied mo e, wi h lignin and epicu icula la onoid aglycones ha ing lowe CVG han in acellula phenolics and condensed annins (Table 1). The geno ypic a ia ion was s a is ically no highly signi ican in he senescen lea es, excep o i e penoids (Table 1). Among he in acellula phenolic subg oups, my ice in glycosides and kaemp e ol glycosides had e y high alues o H2(0.398 and 0.327, espec i ely) and CVG(0.331 and 0.219), while hose o phenolic acids and que ce in glycosides esembled he alues o F on ie s in Plan Science | www. on ie sin.o g 5June 2017 | Volume 8 | A icle 1074 pls-08-01074 June 23, 2017 Time: 16:56 # 6 Paaso e al. Seconda y Chemis y o Bi ch Li e FIGURE 1 | Mean concen a ions (±85% CI, n= 5 o a abino u anosides and 33–38 o o he compounds, based on geno ypes 5, 6, 8, 12, 14, 15, 20, and 25) o phenolic compound g oups: (A) phenolic acids and DHPPG, (B) my ice in glycosides, (C) que ce in glycosides, (D) kaemp e ol glycosides, (E) epicu icula la onoid aglycones, (F) epicu icula i e penoids and (G) he polyme s in Be ula pendula g een lea es, senescen lea es and decomposed lea li e ( he in e al be ween g een and senescen lea es is 4 mon hs and be ween senescen lea es and decomposed li e 8 mon hs; i e penoids a e epo ed as peak a ea; CQA, Ca eoylquinic acid; CouQA, Couma oylquinic acid; DHPPG, 3,40-dihyd oxyp opiophenone-3-glucoside; M, My ice in; Q, Que ce in; K, Kaemp e ol; a , a abino u anoside; ap, a abinopy anoside; gal, galac oside; glc, glucoside; gl , glucu onide; h, hamnoside). F on ie s in Plan Science | www. on ie sin.o g 6June 2017 | Volume 8 | A icle 1074 pls-08-01074 June 23, 2017 Time: 16:56 # 7 Paaso e al. Seconda y Chemis y o Bi ch Li e FIGURE 2 | Concen a ions o (A) in acellula phenolics, (B) epicu icula la onoid aglycones, (C) epicu icula i e penoids, (D) condensed annins, and (E) lignin (mean + SE, n= 5–6) in he senescen lea es and decomposed li e o 19 Be ula pendula geno ypes (a anged in a dec easing o de o senescen lea annin concen a ions). in acellula phenolics in gene al (Table 1 and Supplemen a y Figu e S1). Du ing decomposi ion, he H2and CVGinc eased o in acellula phenolics, epicu icula la onoid aglycones, and condensed annins, emained he same o epicu icula i e penoids and dec eased o lignin (Figu e 2 and Table 1). As a esul , he decomposed li e had s a is ically highly signi ican geno ypic a ia ion in all compounds excep o lignin ha had los geno ypic a ia ion du ing decomposi ion (Table 1). The geno ypic a ia ion in he small-molecula compounds o senescen lea es was also clea ly isible in he PCA o indi idual compounds, whe e PC1 ep esen s en i onmen al a ia ion (P=0.172 o geno ype, P<0.001 o eplica e block) and PC2 F on ie s in Plan Science | www. on ie sin.o g 7June 2017 | Volume 8 | A icle 1074 pls-08-01074 June 23, 2017 Time: 16:56 # 8 Paaso e al. Seconda y Chemis y o Bi ch Li e TABLE 1 | The a iance componen s (σ2), b oad-sense he i abili y (H2), pheno ypic mean (¯ x), coe icien o a ia ion (CVG), and Fand Ps a is ics o ANOVA o he geno ypic a ia ion o seconda y me aboli es in he senescen lea es and decomposed li e o Be ula pendula (G =geno ype, E =e o ; means a e mg g−1d y mass, excep o epicu icula i e penoids peak a ea g−1d y mass; lignin no ans o med, condensed annins log10- ans o med, o he g oups squa e oo - ans o med; bold alues deno e s a is ically signi ican geno ype e ec s). σ2 Gσ2 EH2¯ xCVGF P Senescen lea es In acellula phenolics 0.080 0.617 0.115 2.58 0.110 1.69 0.060 Phenolic acids 4.7E-4 0.003 0.120 0.22 0.098 1.61 0.086 My ice in glycosides 0.037 0.056 0.398 0.58 0.331 4.76 <0.001 Que ce in glycosides 0.077 0.539 0.125 2.43 0.114 1.76 0.047 Kaemp e ol glycosides 0.015 0.031 0.327 0.56 0.219 3.58 <0.001 Epicu icula la onoid aglycones 0.005 0.038 0.116 0.92 0.077 1.67 0.062 Epicu icula i e penoids 247 632 0.281 114 0.138 3.08 <0.001 Condensed annins 0.018 0.142 0.113 1.19 0.113 1.95 0.021 Lignin 496 3613 0.121 491 0.045 1.79 0.040 Decomposed li e In acellula phenolics 0.021 0.057 0.269 1.13 0.128 2.99 <0.001 Que ce in glycosides 0.019 0.058 0.248 1.00 0.139 2.76 0.001 Kaemp e ol glycosides 0.006 0.005 0.528 0.49 0.158 6.96 <0.001 Epicu icula la onoid aglycones 0.005 0.020 0.200 0.66 0.107 2.40 0.004 Epicu icula i e penoids 115 352 0.246 79 0.136 2.75 0.001 Condensed annins 0.011 0.032 0.263 0.47 0.223 3.01 <0.001 Lignin 29 874 0.032 614 0.009 1.19 0.289 mos ly geno ypic a ia ion (P<0.001 o geno ype, P=0.042 o block) (Figu e 3A). In he decomposed li e , geno ypic a ia ion was signi ican along bo h he PC1 (P=0.018 o geno ype, P<0.001 o block) and he PC2 (P<0.001 o geno ype, P=0.003 o block) (Figu e 3B). The anks o geno ype mean sco es co ela ed posi i ely be ween he senescen lea es and decomposed li e o PC2 (ρ=0.87, P<0.001), bu no o PC1 (ρ=–0.43, P=0.064). The geno ype 16 was mos dis inc om o he s in bo h senescen lea es and decomposed li e (Figu es 3A,B). T i e penoids and he mos lipophilic la onoid aglycones (F11-F15) we e he compounds ha bes explained he geno ypic a ia ion along he PC axes (Figu es 3C,D). In he decomposed li e , ou i e penoids, including papy i e ic acid (T7) and i s de i a i e (T6), o med a igh clus e sepa a ed om he es o he compounds (Figu e 3D). Geno ypic Co ela ions among Me aboli es The geno ype mean concen a ions o he wo epicu icula compound g oups – la onoid aglycones and i e penoids – we e posi i ely co ela ed in he senescen lea es, whe eas he mean concen a ions o condensed annins co ela ed nega i ely wi h he concen a ions o la onoid aglycones and lignin (Table 2 and Supplemen a y Figu e S2). The posi i e co ela ion be ween he la onoid aglycones and i e penoids also emained in he decomposed li e (Table 2). Among he in acellula phenolic subg oups o he senescen lea es, concen a ions o phenolic acids co ela ed posi i ely wi h he concen a ions o que ce in glycosides (ρ=0.63, P=0.004, n=19) and kaemp e ol glycosides (ρ=0.52, P=0.023), which also co ela ed wi h each o he (ρ=0.63, P=0.004). The anks o geno ype mean concen a ions co ela ed posi i ely be ween he senescen lea es and decomposed li e o in acellula phenolics (ρ=0.70, P=0.001, n=19), la onoid aglycones (ρ=0.46, P=0.048), i e penoids (ρ=0.75, P<0.001) and condensed annins (ρ=0.62, P=0.004), bu no o lignin (ρ=0.16, P=0.514) (Figu e 2). DISCUSSION Ou esul s show ha al hough he concen a ions o many seconda y me aboli es dec eased signi ican ly du ing B. pendula lea senescence, all me aboli es excep o ca eoylquinic acids (CQAs) emained in he senescen lea es. As we hypo hesized, he emaining me aboli es also exhibi ed signi ican geno ypic a ia ion wi h high b oad-sense he i abili ies and coe icien s o geno ypic a ia ion. Du ing decomposi ion, mos me aboli es dec eased in concen a ion, sugges ing ha hey we e decomposed as e han he li e ma e ial on a e age, bu he geno ypic a ia ion was pe sis en . This was mani es ed by he inc easing he i abili ies and coe icien s o geno ypic a ia ion o he in acellula phenolics, su ace la onoid aglycones and condensed annins du ing li e decomposi ion. Con i ming he pe sis ence o geno ypic a ia ion, he geno ype anks in me aboli e concen a ions emained s able unde ield condi ions and mic obial deg ada ion. Conside ing ha seconda y me aboli es can a ec li e decomposi ion and nu ien cycling (Hä enschwile and Vi ousek, 2000;Schwei ze e al., 2004), hese esul s sugges ha by ac ing on he olia seconda y me aboli e p o iles o B. pendula popula ions, selec ion can be a signi ican d i e o li e decomposi ion. Lignin was an impo an excep ion among he compounds, howe e , as F on ie s in Plan Science | www. on ie sin.o g 8June 2017 | Volume 8 | A icle 1074 pls-08-01074 June 23, 2017 Time: 16:56 # 9 Paaso e al. Seconda y Chemis y o Bi ch Li e FIGURE 3 | P incipal componen analysis (PCA) g aphs o he seconda y me aboli e da a in (A) senescen lea es and (B) decomposed li e and he loadings o he compounds esponsible o he a ia ion on he PC axes as p(co ) alues in (C) senescen lea es and (D) decomposed li e . In (A,B), he 19 geno ypes a e shown as mean axes sco es wi h he e ical and ho izon al e o ba s depic ing ±1 SE (n= 4–6). CouQAs, Couma oylquinic acids; F, Fla onoid aglycone; FG, Fla onol glycosides; Kgal, Kaemp e ol 3-galac oside; K h, Kaemp e ol 3- hamnoside; Mgal, My ice in 3-galac oside; Mglc, My ice in 3-glucoside; Qap, Que ce in 3-a abinopy anoside; Qgal, Que ce in 3-galac oside; Q h, Que ce in 3- hamnoside; T, T i e penoid. TABLE 2 | Rank co ela ion coe icien s (Spea man’s ho) o geno ype mean concen a ions o seconda y me aboli es in he senescen lea es and decomposed li e o Be ula pendula (n=19; ∗P<0.05, ∗∗P<0.01, ∗∗∗P<0.001; bold alues deno e s a is ically signi ican co ela ions). Epicu icula la onoid aglycones Epicu icula i e penoids Condensed annins Lignin Senescen lea es In acellula phenolics 0.05 0.14 0.17 <0.01 Epicu icula la onoid aglycones 0.75∗∗∗ −0.47∗0.25 Epicu icula i e penoids −0.43 0.15 Condensed annins −0.51∗ Decomposed li e In acellula phenolics −0.02 −0.05 0.15 −0.37 Epicu icula la onoid aglycones 0.71∗∗ 0.29 −0.13 Epicu icula i e penoids 0.20 −0.25 Condensed annins −0.37 lignin concen a ions inc eased du ing lea senescence and li e decomposi ion and he geno ypic a ia ion disappea ed du ing decomposi ion. Mo eo e , ou esul s show ha he geno ype means o seconda y me aboli e concen a ions, like hose o condensed annins and lignin, can be nega i ely co ela ed in he senescen lea es. As lignin (Melillo e al., 1982; Hobbie e al., 2006;Talbo and T esede , 2012) and condensed annins (Schwei ze e al., 2004, 2008a) can bo h es ic li e decomposi ion, an associa ion be ween li e decomposi ion a e and a concen a ion g adien o one compound could be canceled ou by an in e se g adien o he o he . In such case, selec ion ac ing on he concen a ion o ei he compound in he g een lea es migh no lead o a signi ican change in li e decomposi ion a e in he popula ion. Also, equally impo an is o ecognize he ole o o he cha ac e is ics o li e chemis y, such as concen a ions o ni ogen (Sil e e al., 2007) and mic onu ien s (Makkonen e al., 2012; Ga cía-Palacios e al., 2016a), when weighing he oppo uni y o selec ion o F on ie s in Plan Science | www. on ie sin.o g 9June 2017 | Volume 8 | A icle 1074