Spatial variations in the molecular diversity of dissolved organic matter in water moving through a boreal forest in eastern Finland

1 Scien i ic RepoR s | 7:42102 | DOI: 10.1038/s ep42102 www.na u e.com/scien i ic epo s Spa ial a ia ions in he molecula di e si y o dissol ed o ganic ma e in wa e mo ing h ough a bo eal o es in eas e n Finland Jun’ichi o Ide1, Mizue Ohashi2, Ka su oshi Takahashi3, Yuko Sugiyama2, Si pa Pii ainen4, Pi kko Ko elainen5, Nobuhide Fuji ake6, Kei a o Yamase7, Nobuhi o Oh e8, Mina Mo i ani2, Miyako Ha a2 & Leena Finé 4 Dissol ed o ganic ma e (DOM) s ongly a ec s wa e quali y wi hin bo eal o es ecosys ems. Howe e , how he quali y o DOM i sel changes spa ially is no well unde s ood. In his s udy, o examine how he di e si y o DOM molecules a ies in wa e mo ing h ough a bo eal o es , he numbe o DOM molecules in di e en wa e samples, i.e., ainwa e , h ough all, soil wa e , g oundwa e , and s eam wa e was de e mined using Fou ie ans o m ion cyclo on esonance mass spec ome y (FT-ICR MS) in No way sp uce and Sco s pine s ands in eas e n Finland du ing May and June 2010. The numbe o molecula compounds iden i ied by FT-ICR MS (molecula di e si y) anged om 865 o 2,194, e ealing la ge DOM molecula di e si y in he wa e samples. Addi ionally, some o he molecula compounds we e sha ed be ween di e en wa e samples. The DOM molecula di e si y linea ly co ela ed wi h he numbe o low-biodeg adable molecules, such as, lignin-like molecules (lignins), bu no wi h dissol ed o ganic ca bon concen a ion. The numbe o lignins sha ed be ween di e en sampling loca ions was la ge han ha o any o he biomolecula class. Ou esul s sugges ha low-biodeg adable molecules, especially lignins, egula e spa ial a ia ions in DOM molecula di e si y in bo eal o es s. Dissol ed o ganic ma e (DOM) plays a cen al ole in changes in wa e chemis y wi hin bo eal o es eco- sys ems. I a ec s he acidi y and colo o he wa e ha passes h ough he ee canopy o o es loo 1–5. I o ms complexes wi h many nu ien s and me als and anspo s hem om o es ecosys ems o wa e cou ses6–10. Addi ionally, changes in he quali y and quan i y o DOM in luence ecological p ocesses, because DOM is a sou ce o ene gy and is easily modi ied by mic oo ganisms11,12. Bo eal o es s con ain mo e han one hi d o he ca bon e ained in o es ege a ion and soils globally13. Acco ding o a ecen in en o y o o ganic ca bon pools in bo eal o es soils in Fennoscandia, 70–80% o he o ganic ca bon is ound in he uppe 100 cm o mine al soil14. P e ious s udies showed ha dissol ed o ganic ca bon (DOC) concen a ion was i e imes highe in h ough all han in bulk deposi ion and also i e imes highe in soil wa e in he O-ho izon han in h ough all15,16. This is because DOM leaches in o he wa e om plan issues and dead o ganic ma e , and is also mic obially decomposed in soils. Addi ionally, DOM can be e ained in mine al soils since humic subs ances a e associa ed wi h amo phous i on and aluminum17. The elease and e en ion o DOM by biogeochemical p ocesses should al e no only he amoun o DOM bu also i s com- posi ion. Hyd ophobic DOM is p e e en ially adso bed by o es soil pa icles, which could inc ease he ela- i e abundance o hyd ophilic compounds in soil wa e 18. An inc ease in DOC concen a ion in soil wa e a e 1Ins i u e o Decision Science o a Sus ainable Socie y, Kyushu Uni e si y, Fukuoka 811-2415, Japan. 2G adua e School o Human Science and En i onmen , Uni e si y o Hyogo, Hyogo 670-0092, Japan. 3Na ional Ins i u e o Ad anced Indus ial Science and Technology, Tsukuba 305-8568, Japan. 4Na u al Resou ces Ins i u e Finland, P.O. Box 68, FIN-80101 Joensuu, Finland. 5Finnish En i onmen Ins i u e, P.O. Box 140, FIN-00251 Helsinki, Finland. 6Facul y o Ag icul u e, Kobe Uni e si y, Kobe 657-0013, Japan. 7Hyogo P e ec u al Technology Cen e o Ag icul u e, Fo es y and Fishe ies, Fo es y and Fo es P oduc s Resea ch Ins i u e, Hyogo 671-2515, Japan. 8Depa men o Social In o ma ics, G adua e School o In o ma ics, Kyo o Uni e si y, Kyo o 606-8501, Japan. Co espondence and eques s o ma e ials should be add essed o J.I. (email: [email p o ec ed]) Recei ed: 04 Oc obe 2016 accep ed: 04 Janua y 2017 Published: 10 Feb ua y 2017 OPEN www.na u e.com/scien i ic epo s/ 2 Scien i ic RepoR s | 7:42102 | DOI: 10.1038/s ep42102 clea -cu ing was also epo ed, sugges ing ha o es y ope a ions a ec he chemical composi ion o DOM19. Conside ing he esul s epo ed by Bou bonnie e and C eed20 ha eshe and younge o es - loo ma e ials eleased mo e biodeg adable DOM han olde ones, i is possible ha DOM eleased om logging esidues o li e in newly-plan ed o es s ands has a labile p ope y. Howe e , ew s udies ha e ocused on how he chemical composi ion o DOM changes in wa e as i mo es h ough o es ecosys ems21. Recen p og ess in he chemical cha ac e iza ion o DOM has enabled us o ob ain an o e all pic u e o molec- ula al e a ions/ ansi ions o DOM in wa e collec ed om o es s. Fou ie ans o m ion cyclo on esonance mass spec ome y (FT-ICR MS) is one o he mos inno a i e me hods and allows us o de ec di e en ypes o molecules in DOM based on he cyclo on equency o cha ged ions in a magne ic ield. This me hod has p o- ided ep oducible and eliable da a conce ning he mass spec um o DOM in i e and ma ine wa e samples, which consis o mo e han a hund ed di e en ypes o DOM molecules22–25. In addi ion o in o ma ion on he numbe and ypes o DOM molecules, he an K e elen diag am26 enables us o di ide each molecule de ec ed by FT-ICR MS in o di e en biomolecula classes, such as lignin-like molecules, p o eins, and annin-like mole- cules, o desc ip i e pu poses27,28. De ailed in o ma ion on DOM migh help us unde s and he p ocesses d i - ing changes in he chemical composi ion o DOM and he ac o s con olling hem. Kalbi z e al.29 in es iga ed changes in he chemical composi ion o DOM samples ex ac ed om di e en ma e ials, including maize s aw, o es loo s, pea , and ag icul u al soils, du ing a 90-day incuba ion using se e al me hods, such as ul a iole abso bance, luo escence emission spec oscopy, and Fou ie ans o m in a ed spec oscopy. Thei esul s indi- ca ed ha he pa ial deg ada ion o highe molecula weigh compounds, such as lignin dime s and alkyla oma - ics, inc eases he p opo ion o lowe molecula weigh compounds, such as lignin monome s and phenols, which a e biodeg ada ion p oduc s. They also indica e ha DOM biodeg ada ion can esul in ela i e en ichmen o a oma ic compounds, which a e ela i ely s able and slowly mine alized. These esul s sugges ha chemical com- posi ion o DOM a ies mainly ia bio ic p ocesses, whe eas he e a e molecula compounds ha a e mo e s able and mo e along wi h he wa e h ough he ecosys em. I is expec ed ha he FT-ICR MS analysis and subsequen biomolecua classi ica ion enhance ou unde s anding no only o chemical aspec s bu also o bio ic p ocesses in he molecula al e a ion/ ansi ion o DOM associa ed wi h he wa e mo emen . The objec i es o his s udy we e (1) o e alua e a ia ions in molecula di e si y and he molecula com- pounds di ided in o speci ic biomolecula classes in ainwa e passing h ough a bo eal o es using FT-ICR MS and an K e elen diag ams, and (2) o examine how sampling loca ion a ec s DOM molecula compounds in a bo eal o es ca chmen in eas e n Finland. Ou hypo heses we e as ollows: 1. The numbe o molecula compounds de ec able by FT-ICR MS (i.e., molecula di e si y) and subsequen - ly he numbe o each biomolecula class a y wi h wa e mo emen (i.e., ainwa e o s eam wa e ia h ough all, soil wa e , and g oundwa e ). 2. The molecula compounds iden i ied as he same DOM compound exis a di e en sampling loca ions. 3. DOM molecula di e si y inc eases wi h inc easing DOC concen a ion. Resul s Molecula di e si y and biomolecule classi ica ion o DOM. The numbe o DOM molecula com- pounds de ec ed as m/z peaks (molecula di e si y) was 1,293 ± 307 SD on a e age ac oss all o he wa e samples and anged om 865 o 2,194 (Fig.1a). The coe icien o a ia ion (CV) o he molecula di e si y was 23%. The molecula di e si y did no di e signi ican ly among sampling loca ions (p > 0.05). Each DOM molecula compound was assigned o one o se en biomolecula classes (i.e., lipids, p o eins, ami- nosuga s/ca bohyd a es (As/Ch), unsa u a ed hyd oca bons (UH), lignin-like molecules (lignins), annin-like molecules ( annins), o condensed a oma ic s uc u es (CAS)), o i was no ed ha i could no be assigned o any class. Ou esul s e ealed ha he numbe o lignins was la ge han ha o any o he class and only a ew molecules o As/Ch we e ound in e e y sample excep he bulk deposi ion sample ob ained in May (Fig.2). The numbe o lignins was especially la ge in h ough all, g oundwa e -PT, and s eam wa e samples in June. The numbe s o lipids o p o eins we e he second la ges a e lignins in he majo i y o he wa e samples; howe e , in some samples, he numbe o CAS was second o lignins. Rela ionships o biomolecula classes and DOC concen a ion o molecula di e si y. DOM molecula di e si y inc eased signi ican ly wi h inc easing numbe s o lignins, annins, and CAS in bo h May and June (Fig.3a: = 0.83, p < 0.001 in May, = 0.77, p < 0.001 in June; 3b: = 0.86, p < 0.001 in May, = 0.77, p < 0.001 in June; 3c: = 0.77, p < 0.01 in May, = 0.76, p < 0.001 in June). The p opo ion o he sum o lignins, annins, and CAS in he molecula di e si y in soil wa e was gene ally high (28–56%) compa ed wi h hose in bulk deposi ion and h ough all (20–31%), al hough no in compa ison o he h ough all in June (56%). The DOC concen a ion ac oss all o he wa e samples was 11.2 ± 10.4 mgC l–1 on a e age, anging om 0.7 o 36.4 mgC l–1 (Fig.1b). The CV o DOC concen a ion was 93%. The DOC concen a ion di e ed signi ican ly among he sampling loca ions (Fig.1b: p < 0.01). The e was no co ela ion be ween DOC concen a ion and DOM molecula di e si y (Fig.4: = 0.09, p = 0.742 in May, = 0.09, p = 0.722 in June). Spa ial a ia ion in DOM molecula composi ion. The numbe o molecula compounds de ec ed as he same m/z peak be ween di e en wa e samples anged om 153 o 1010. The p opo ion o hose compounds in he molecula di e si y anged om 10% o 60% depending on sampling loca ion. Addi ionally, he numbe o lignins sha ed be ween di e en wa e samples was la ge han ha o any o he biomolecula class and anged om 52 o 445. www.na u e.com/scien i ic epo s/ 3 Scien i ic RepoR s | 7:42102 | DOI: 10.1038/s ep42102 The dend og am including all species o molecula compounds de ec ed by FT-ICR MS in May (Fig.5a) showed ha bulk deposi ion and h ough all o med one small clus e , indica ing ha hese samples in May had many molecula compounds in common. The dend og ams o he se en biomolecula classes in May (Fig.5b–h) also showed ha lignins de ec ed in he bulk deposi ion and h ough all o med one small clus e . O-ho izon soil wa e (SWO(O1), SWY(O1) and SWY(O2) in Fig.5), g oundwa e -PT and s eam wa e we e g ouped in o he same clus e in he lignins, annins and CAS dend og ams in May (Fig.5 –h). The bulk deposi ion and h ough- all did no o m one small clus e in he dend og am including all molecula compounds iden i ied in June (Fig.6a); howe e , hey we e g ouped in o he same clus e when he dend og am was cu a he dissimila i y le el o 0.95. On he o he hand, he dend og ams o he se en biomolecula classes in June (Fig.6b–h) showed ha liginins de ec ed in he bulk deposi ion and h ough all we e sepa a ed in o he di e en clus e s. Th ough all, g oundwa e -PT and s eam wa e o med one clus e in he lignins, annins and CAS dend og ams (Fig.6 –h), which exhibi ed simila pa e ns o he dend og am including all molecula compounds iden i ied in June. S ong Numbe o m/z peaks 01000 20003000 (a) Bulk deposi ion Th ough all Soil wa e in F O (O-ho izon) Soil wa e in F O (E-ho izon) Soil wa e in F O (B-ho izon) So il wa e in F Y (O-ho izon) Soil wa e in F Y (E-ho izon) Soil wa e in F Y (B-ho izon) G oundwa e -PT G oundwa e -MO G oundwa e -MI S eam wa e DOC concen a ion [mgC l –1 ] 01020304 050 a a ab ab ac a a a a bc b b (b) Figu e 1. (a) A e age alues o he numbe o m/z peaks and (b) DOC concen a ion in bulk deposi ion, h ough all, soil wa e , g oundwa e , and s eam wa e du ing he s udy pe iod. E o ba s ep esen s anda d de ia ions. Values wi h he same le e indica e no signi ican di e ence (p > 0.05). FO and FY ep esen old- g ow h o es and young plan a ion s ands, espec i ely. See Me hods o he abb e ia ions o g oundwa e . 0500 1000150 02000 0500 1000 Numbe o m/z peaks Bulk deposi ion Th ough all Soil wa e in F O (O-ho izon) Soil wa e in F O (E-ho izon) Soil wa e in F O (B-ho izon) So il wa e in F Y (O-ho izon) Soil wa e in F Y (E-ho izon) Soil wa e in F Y (B-ho izon) G oundwa e -PT G oundwa e -MO G oundwa e -MI S eam wa e Ma yJ une n = 3 n = 1 n = 1 n = 2 n = 1 n = 2 n = 3 n = 3 n = 3 Lipids P o eins As / Ch UH LigninsTanninsCASUniden i ied Figu e 2. Classi ica ion o he m/z peaks in o speci ic biomolecula classes using a an K e elen diag am. FO and FY ep esen old-g ow h o es and young plan a ion s ands, espec i ely. n alues ep esen he numbe o lysime e s which collec ed soil wa e . The a e age alues a e shown o soil wa e samples wi h n = 2 o n = 3. See Me hods o he abb e ia ions o g oundwa e and biomolecula classes and see Figu esS1 and S2 o a ia ions in he numbe o m/z peaks o each biomolecula class in soil wa e samples. www.na u e.com/scien i ic epo s/ 4 Scien i ic RepoR s | 7:42102 | DOI: 10.1038/s ep42102 posi i e co ela ions we e ound be ween he cophene ic ma ices o Jacca d’s dis ance o all molecula com- pounds and o lignins bo h in May and June (Fig.7: = 0.84, p < 0.001 in May, = 0.87, p < 0.001 in June). Discussion The esul s o he FT-ICR MS analysis e ealed ha housands o di e en DOM molecules exis ed in each wa e sample collec ed om ou s udy si e, i.e., Kangas aa a ca chmen (Fig.1a). FT-ICR MS is one o he mos ad anced ins umen s a ailable o he de ec ion o ionizable DOM because i has ul ahigh mass esolu ion 0 200 400 600 800 1000 1200 0500 1000 1500 2000 2500 0 10 20 30 40 50 60 70 80 0500 1000 1500 2000 2500 0 20 40 60 80 100 120 140 160 0500 1000 1500 2000 2500 To al numbe o m/z peaks (a) (b) (c) Numbe o ligninsNumbe o anninsNumbe o CAS May: R2 = 0.70 p < 0.001 June: R2 = 0.60 p < 0.001 May: R2 = 0.74 p < 0.001 June: R2 = 0.59 p < 0.001 May: R2 = 0.59 p < 0.01 June: R2 = 0.58 p < 0.001 MayJune Figu e 3. (a) Rela ionship be ween he o al numbe o m/z peaks (molecula di e si y) and he numbe o peaks ep esen ing lignin-like molecules (lignins), (b) annin-like molecules ( annins), and (c) condensed a oma ic s uc u es (CAS). Solid and b oken lines ep esen eg ession lines o da a ob ained in May and June, espec i ely. www.na u e.com/scien i ic epo s/ 5 Scien i ic RepoR s | 7:42102 | DOI: 10.1038/s ep42102 and is o en coupled wi h a non-des uc i e ion sou ce, i.e., elec osp ay ioniza ion (ESI)30. Kim e al.31 analyzed he molecula composi ion o DOM in a black-wa e s eam loca ed in he Pinelands o New Je sey, USA using FT-ICR MS and de ec ed 18 peaks ha can be assigned molecula o mulas in he ange o m/z 469.0–469.3. Addi ionally, Wi e al.32 in es iga ed he molecula composi ion o ul ic acid in a wa e sample collec ed om he Suwannee Ri e , USA, using FT-ICR MS and de ec ed 21 peaks app oxima ely a he m/z 411 wi hin 0.35 in he mass spec um. These p e ious s udies indica ed ha FT-ICR MS allows o he de ec ion o DOM molec- ula compounds wi hin a na ow ange o m/z alues. Kim e al.33 de ec ed o e 3,000 molecula compounds in he mass spec um o he wa e samples ob ained in opical and empe a e o es ed s eams. Fu he mo e, Wozniak e al.34 de ec ed o e 3,000 molecula compounds in ae osol-de i ed wa e -soluble o ganic ca bon col- lec ed om u al a eas in Vi ginia and New Yo k, USA. Mazzoleni e al.35 de ec ed 1,368 molecula compounds in wa e -soluble a mosphe ic o ganic ma e con ained in og wa e . The numbe o molecula compounds de ec ed as m/z peaks (molecula di e si y) in his s udy was compa able o hose obse ed in he p e ious s udies. On he o he hand, no clea end in molecula di e si y acco ding o sampling loca ion was obse ed, al hough he molecula compounds could be di ided in o biomolecula classes, in which he numbe o lignins was la ges (Fig.2) (hypo hesis 1). The linea ela ionships be ween DOM molecula di e si y and he numbe s o lignins, annins, and CAS (Fig.3) sugges ha low-biodeg adable molecules egula e changes in DOM molecula di e si y36,37. Biodeg adable molecules, such as hyd ophilic o ganic ma e leached om ees, a e p e e en ially u ilized by mic oo ganisms, esul ing in an abundance o low-biodeg adable molecules, such as hyd ophobic a oma ic com- pounds emaining in soils29,38. Hu e al.39 sugges ed ha mic obial u iliza ion o labile componen s, such as sim- ple ca bohyd a es and amino acids, p oduced humic-like a oma ic componen s in he DOM ex ac ed om lea li e . These would also be e lec ed in he gene ally highe p opo ion o he sum o lignins, annins, and CAS in he molecula di e si y in soil wa e han in bulk deposi ion and h ough all. Howe e , some low-biodeg adable molecules can be p e e en ially adso bed by soil pa icles18,40. This migh be e lec ed in he a ia ions in he numbe o low-biodeg adable molecules among eplica es o soil wa e samples (see FigsS1 and S2). Simila pa e ns we e ound among he lignins, annins and CAS dend og ams (Figs6 and 7), implying ha dynamics o low-biodeg adable molecules we e simila wi hin bo eal o es ecosys ems. Compa isons o he m/z alue be ween di e en wa e samples e ealed ha common molecula compounds exis ed a di e en sampling loca ions (hypo hesis 2). We ound ha he numbe o lignins sha ed be ween di - e en sampling loca ions was la ge han ha o any o he biomolecula class. Addi ionally, co ela ion analyses be ween he cophene ic ma ices o Jacca d’s dis ance (Fig.7) showed ha he pa e n o he deg ee o simila i y among sampling loca ions o lignins e lec s ha o all molecula compounds de ec ed by FT-ICR MS. Indeed, simila pa e ns we e ound be ween he dend og ams o lignins and all molecula compounds in May and June (Figs5 and 6). These esul s indica e ha lignins could egula e spa ial a ia ions in he numbe o he common molecula compounds. Lignins a e phenolic polyme s ha o igina e mainly om ascula plan s in e es ial ecosys ems and a e conside ed o be ecalci an o ganic ma e s because hey a e deg aded a a lowe a e han o he subs ances o igina ing om plan s, such as, cellulosic and non-cellulosic polysaccha ides and p o eins41,42. This p esumably allows lignins o exis as he molecula compounds common o di e en sampling loca ions in ou s udy si e. On he o he hand, a numbe o di e en molecula compounds exis ed be ween sampling loca ions since common molecula compounds accoun ed o less han 50% o he molecula di e si y in he majo i y o he wa e samples. This may be ela ed o he ac ha he quali y o DOM is al e ed by se e al p ocesses. Fo ins ance, ee canopies and o es soils elease DOM molecula compounds in o ainwa e ; whe eas in ainwa e , hey a e mic obially con e ed o lowe molecula -weigh compounds and/o some a e e ained in mine al soils 0 500 1000 1500 2000 2500 010203 040 DOC concen a ion [mgC l –1 ] Numbe o m/z peaks MayJune Figu e 4. Rela ionship be ween he numbe o m/z peaks (molecula di e si y) and DOC concen a ion. www.na u e.com/scien i ic epo s/ 6 Scien i ic RepoR s | 7:42102 | DOI: 10.1038/s ep42102 G-MI G-MI G-MI G-MI Dissimila i y SW O (O2) G-PT SW Y (O1) ST TF SW O (B) SW O (E) BD SW Y (O2) SW O (O1) SW O (O3) G-MI SW Y (B) G-MO 0.85 0.90 0.95 1.00 1.05 TF ST SW O (O2) G-PT SW Y (O2) SWP_B SW O (O1) SW Y (O1) SWF_E BD G-MO G-MI SW O (O3) SW O (B) 0.80 0.90 1.00 SW O (O1) SW Y (O1) SW Y (O2) SW O (O3) SW Y (B) G-MO ST SW O (E) SW O (B) SW O (O2) G-PT BD TF 0.70 0.80 0.90 1.00 (a) May (b) Lipids G-MO BD TF SW O (B) SW O (E) G-PT SW O (O2) ST SW O (O3) SW Y (O1) SW O (O1) SW Y (O2) SW Y (B) 0.60.7 0.80.9 1.01.1 (c) P o eins G-MI BD G-PT TF SW O (O1) ST SW O (E) SW Y (O1) SW O (B) SW Y (O2) SW O (O2) SW Y (B) SW Y (O3) G-MO 0.60.7 0.80.9 1.01.1 1.21.3 (d) As / Ch (e) UH SW O (O2) G-PT ST SW O (O1) SW Y (O1) SW Y (O2) G-MO SW O (O3) SW O (E) SW O (B) SW Y (B) BD TF 0.50.6 0.70.8 0.9 ( ) Lignins SW O (O1) SW Y (O2) G-PT SW Y (O1) ST TF SW O (O3) SW O (O2) SW O (E) SW Y (B) SW O (B) BD G-MO 0.60.8 1.01.2 1.4 (g) Tannins (h) CAS ST SW O (O1) G-PT SW Y (O1) SW Y (O2) SW O (E) SW O (B) G-MI SW O (O2) SW O (O3) TF BD SW Y (B) G-MO 0.80.9 1.01.1 1.2 Dissimila i y G-MI BD = bulk deposi ion, TF = h ough all, SWO(O) = O-ho izon soil wa e in FO, SWO(E) = E-ho izon soil wa e in FO, SWO(B) = B-ho izon soil wa e in FO, SWY(O) = O-ho izon soil wa e in FY, SWY(B) = B-ho izon soil wa e in FY, G-PT = g oundwa e wi hin he pea laye , G-MO = g oundwa e wi hin he mo aine, G-MI = g oundwa e wi hin mine al soils unde he pea laye , ST = s eam wa e Figu e 5. (a) Clus e dend og ams o di e en sampling loca ions o all molecula compounds iden i ied and (b–h) o he se en biomolecula classes in May. The igu e in he pa en hesis ep esen s he iden i ica ion numbe o he h ee lysime e s (L1–L3). See Me hods o he abb e ia ions o biomolecula classes. www.na u e.com/scien i ic epo s/ 7 Scien i ic RepoR s | 7:42102 | DOI: 10.1038/s ep42102 Dissimila i y Dissimila i y ST TF G-PT SW Y (B2) SW Y (E3) G-MO SW O (E1) SW O (E3) SW O (B1) SW Y (B1) SW O (B2) SW O (B3) G-MI SW O (E2) SW Y (E1) BD SW Y (E2) 0.60.7 0.80.9 1.01.1 1.21.3 SW O (E1) SW Y (E3) BD SW O (E3) SW Y (B2) G-MO G-MI SW O (E3) SW O (B1) TF G-PT ST SW O (B2) SW Y (B1) SW O (E2) SW Y (E1) SW Y (E2) 0.60.8 1.01.2 1.4 TF G-PT ST SW O (E3) SW Y (E1) SW O (E2) SW Y (E2) SW Y (B2) G-MO SW O (E1) SW Y (E3) SW O (B1) SW O (B3) SW Y (B1) G-MI BD SW O (E2) 0.50.6 0.70.8 0.91.0 SW O (E1) SW Y (E3) SW O (E3) ST TF G-PT BD G-MO SW Y (B2) SW Y (E1) SW Y (B1) SW O (B1) SW O (B2) SW Y (E2) SW O (E2) SW O (B3) G-MI 0.70.8 0.91.0 1.11.2 SW Y (E2) SW Y (E3) SW Y (B1) SW Y (B2) SW O (E3) SW Y (E1) TF SW O (E1) G-PT SW O (E2) SW O (B1) BD ST G-MO G-MI SW O (B2) SW O (B3) 0.85 0.90 0.95 1.00 1.05 SW O (B3) SW O (E2) SW O (B2) SW O (B1) SW O (E1) SW Y (E3) SW Y (E2) SW Y (E2) G-MI G-MO BD SW Y (B1) SW O (E3) G-PT SW Y (E1) TF ST 0.80.9 1.01.1 1.2 SW Y (E2) SW Y (E3) SW Y (E1) SW Y (B1) SW Y (B2) SW O (E2) SW O (B1) SW O (B2) SW O (E3) SW O (B3) TF G-PT G-MO BD G-MI SW O (E1) ST 0.70.8 0.91.0 1.11.2 1.3 SW O (E2) SW O (B1) SW O (B3) SW O (B2) SW Y (B2) SW Y (E3) SW Y (B1) G-MI SW O (E1) SW O (E3) SW Y (E2) G-MO BD SW Y (E1) TF G-PT ST 0.60.7 0.80.9 1.01.1 (a) June (b) Lipids (c) P o eins (d) As / Ch (e) UH ( ) Lignins (g) Tannins (h) CAS BD = bulk deposi ion, TF = h ough all, SWO(E) = E-ho izon soil wa e in FO, SWO(B) = B-ho izon soil wa e in FO, SWY(E) = E-ho izon soil wa e in FY, SWY(B) = B-ho izon soil wa e in FY, G-PT = g oundwa e wi hin he pea laye , G-MO = g oundwa e wi hin he mo aine, G-MI = g oundwa e wi hin mine al soils unde he pea laye , ST = s eam wa e Figu e 6. (a) Clus e dend og ams o di e en sampling loca ions o all molecula compounds iden i ied and (b–h) o he se en biomolecula classes in June. The igu e in he pa en hesis ep esen s he iden i ica ion numbe o he h ee lysime e s (L1–L3). See Me hods o he abb e ia ions o biomolecula classes. www.na u e.com/scien i ic epo s/ 8 Scien i ic RepoR s | 7:42102 | DOI: 10.1038/s ep42102 ia co-p ecipi a ion40,43–45. Sleigh e and Ha che 46 no ed ha chemical s uc u es could di e e en hough he assigned molecula o mulas a e he same be ween di e en wa e samples o he FT-ICR MS analysis. They also showed di e en chemical s uc u es o C26H32O11 as an example o isome s, which ep esen wo majo e ac o y subs ances, i.e., lignin and ca boxyl ich alicyclic molecule, and explained ha hese subs ances ha e di e en eac i i ies and sou ces hough hey ha e he same molecula o mula. The e o e, i is possible ha e en molecula compounds conside ed lignins in he an K e elen diag am we e misiden i ied and consequen ly a la ge numbe o di e en molecula compounds exis ed be ween sampling loca ions han obse ed. This also would be suppo ed by Ree sma47 who s a es ha he molecule plo ed in o he egion o one o he biomolecula classes in he diag am does no necessa ily belong o ha class because he O/C and H/C a ios a e insu icien c i e ia o asc ibe a molecule o a ce ain biomolecua class. Howe e , se e al s udies desc ibed ha lignins a e majo componen s o e ac o y DOM in he e es ial eshwa e 12,30,48. In o es s, li e is conside ed o be an impo an sou ce o soil DOM49 and especially coni e ous li e , such as sp uce and pine needle li e could elease lignin-de i ed DOM componen s wi h a high a oma ici y in o he soils ia mic obial decomposi ion50. These suppo ou conclusion ha lignins could egula e spa ial a ia ions in he numbe o common molecula compounds and consequen ly molecula di e si y in bo eal o es s. Ve ical p o iles in DOM molecula di e si y and DOC concen a ion wi hin he Kangas aa a ca chmen (Fig.1) e ealed ha a ia ions in DOM molecula di e si y we e much smalle han hose in DOC concen- a ion. P esumably e lec ing his esul , molecula di e si y did no change in associa ion wi h DOC concen- a ion (Fig.4) (hypo hesis 3). These esul s demons a e ha DOC concen a ion alone does no su icien ly explain spa ial a ia ions in molecula di e si y. This could be a ibu ed o he ac ha molecula composi ions de ec ed by FT-ICR MS do no p o ide quan i a i e in o ma ion on he molecula compounds p esen in DOM51. Al e na i ely, i could be a ibu ed o he ac ha no all DOM molecula compounds in wa e samples a e de ec able because he en ichmen p ocedu e o DOM om wa e samples, such as he C18 solid phase ex ac- ion me hod used, is incomple e47 and also because non-ionizable o ganic compounds a e no cha ac e ized by FT-ICR MS30. Reem sma and These52 sugges ed he possibili y ha high molecula -weigh componen s o humic and ul ic acids we e less e ec i ely ionized by ESI han low molecula -weigh componen s. Addi ionally, Piccolo and Spi elle 53 s a ed ha when u ilizing ESI o de ec DOM molecula compounds, he dominance o hyd ophobic compounds in humic subs ances may inhibi ESI o hyd ophilic compounds due o sup amolecula associa ions o low-molecula o ganic compounds. Howe e , S enson e al.54 used ESI coupled o FT-ICR MS o de ec molecula compounds in Suwannee Ri e ul ic acids and epo ed ha ions gene a ed by ESI we e pa - ially ep esen a i e o he en i e humic subs ances s udied. In any case, ou esul s sugges ha DOM molecula di e si y does no inc ease o dec ease wi h DOC concen a ion in wa e samples collec ed om bo eal o es s. This s udy p esen s p elimina y bu unp eceden ed esul s on he changes in he molecula di e si y and composi ion o DOM ac oss wa e mo emen in a bo eal o es . A his s age, we could no examine empo al a ia ions in he quali y o DOM. Howe e , common DOM molecula compounds we e de ec ed in he same loca ion be ween May and June and he numbe o he lignins common o bo h May and June was la ge han ha o any o he biomolecula class o any wa e sample (Fig.S3). The e o e, i is possible ha lignins egula e no only spa ial bu also empo al a ia ions in molecula di e si y. Fu he esea ch is needed on he empo al a ia ions and on he ac o s d i ing he spa ial and empo al a ia ions in molecula di e si y and he mecha- nisms o p ese a ion and al e a ion o DOM molecula compounds ac oss wa e mo emen h oughou he yea in bo eal ecosys ems. Me hods Si e desc ip ion. This s udy was conduc ed in he Kangas aa a ca chmen (Fig.S4) loca ed in eas e n Finland (63°51′N, 28°58′E). The a ea and mean ele a ion o he ca chmen a e 56 ha and 187 m a.s.l., espec i ely. The ca chmen has a pe ennial b ook. The mean slope g adien o he ca chmen is 7%. F om 1981 o 2000, he -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 CAS Tannin s Lignins UH As/Ch P o ein s Lipids -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 CAS Tannin s Lignins UH As/Ch P o ein s Lipids Co ela ion coe icien ( ) Ma yJ une Figu e 7. Co ela ion coe icien s be ween cophene ic ma ices o Jacca d’s dis ance o all molecula compounds iden i ied and o each o he se en biomolecula classes in May and June. See Me hods o he abb e ia ions o biomolecula classes. www.na u e.com/scien i ic epo s/ 9 Scien i ic RepoR s | 7:42102 | DOI: 10.1038/s ep42102 mean annual ai empe a u e was 2.3 °C and p ecipi a ion was 527 mm, 35% o which was snow all55. The soils in he ca chmen mainly consis o i on podzols (haplic podzols) and ib ic his osols (pea ), which ha e de eloped on s ony ill ma e ial. The unde lying bed ock consis s o gneiss g ani e and g anodio i e. The p opo ion o pea land is 8% o he ca chmen a ea56. Mos o he o es s on he upland mine al soils (97%) we e old-g ow h o es s domina ed by No way sp uce (Picea abies L. Ka s .), Sco s pine (Pinus syl es is L.), and whi e and sil e bi ch (Be ula pubescens Eh h. and Be ula pendula Ro h., espec i ely). Eu opean aspen (Populus emula L.) we e also ound in he ca chmen 56. The expe imen al s ands in he ca chmen we e old-g ow h mixed o es (FO) domina ed by No way sp uce and 12-yea -old Sco s pine plan a ion (FY). The plan a ion was es ablished a e ha es ing a po ion o he No way sp uce s ands (Fig.S4). A 50 × 50 m plo was es ablished in each s and. No way sp uce, Sco s pine, and deciduous ees accoun ed o 53%, 33%, and 14%, espec i ely, o he o al s and olume o 260 m3 ha–1 in he FO. The ield laye ege a ion was domina ed by dwa sh ubs (Vaccinium i is-idaea L. and V. my illus L.) and he bo om laye by ea he mosses (Pleu ozium sch ebe i B id. and Hylocomium splendens (Hedw.) B. S. & G.). The o es loo consis ed o li e and mo humus. In 1996, s em-only ha es ing was ca ied ou in 35% o he ca chmen a ea and Sco s pine seedlings we e plan ed he ea e . B anches, lea es, and ee ops, he diame e s o which we e less han 8 cm, we e le on si e. The ield laye ege a ion was domina ed by g asses (Deschampsia lexuosa (L.) T in.) and he bs (Epilobium angus i olium L.), and he bo om laye by pionee mosses (Dic anum sp. and Poly ichum sp.) in he FY. The a e age hickness o he o e laying mo humus laye (O-ho izon), he elu ial (E) ho izon and he uppe illu ial (B) ho izon was 5.7, 6.8 and 13.9 cm, espec i ely. Wa e sampling. Bulk deposi ion, h ough all, soil wa e , g oundwa e , and s eam wa e we e sampled in May a e snowmel and in June be o e he d y summe mon hs o 2010. Bulk deposi ion was sampled using i e open collec o s (each 130.7 cm2) in an open a ea wi hin 500 m o he expe imen al plo s (Fig.S4). Th ough all was sampled using 20 collec o s o he same ype used o bulk deposi ion, placed sys ema ically a ound he FO plo s. The bulk deposi ion and h ough all samples we e collec ed o a week in each mon h and emp ied a e he week. The soil wa e was collec ed unde he O-ho izon, a a dep h o 15 cm ha was compa able o he E-ho izon and a a dep h o 35 cm ha was compa able o he B-ho izon by ension (60 kPa) lysime e s ha consis ed o 67-mm-long po ous cups wi h an ou e diame e o 12 mm (P80, Hoechs Ce amTec AG, Ge many). Th ee lysim- e e s (L1, L2, and L3) we e placed 1–2 m apa in e e y ho izon and each s and and soil wa e om each loca ion was collec ed o he same week as bulk deposi ion and h ough all. In May, no wa e samples we e ob ained om wo lysime e s in he E- and B-ho izons in FO, one lysime e in he O-ho izon, h ee lysime e s in he E-ho izon, and wo lysime e s in he B-ho izon in FY because o small ain all. In June, no wa e samples we e ob ained om h ee lysime e s in he O-ho izon in FO, h ee lysime e s in he O-ho izon and one lysime e in he B-ho izon in FY. The g oundwa e was ob ained om a g oundwa e ube ins alled a a dep h o 0.7 m in he pea laye (g oundwa e -PT), 2.5 m in he mo aine (g oundwa e -MO), and 2.4 m in he mine al soil unde he pea laye (g oundwa e -MI). The dep h o he pea laye a ound he g oundwa e -PT anged om 1.5 o 2.1 m. The g ound- wa e ube was made o polyamide and had an inne diame e o 30 mm and an ou e diame e o 40 mm. The s eam wa e samples we e collec ed a he ou le o he Kangas aa a ca chmen (Fig.S4). The g oundwa e and s eam wa e samples we e aken a he same day as o he wa e samples, whe e empe a u e, ain all, and 10-day an eceden ain all we e 8.1 °C in May and 13.5 °C in June, 0.8 mm in May and 0 mm in June, and 27.3 mm in May and 57.3 mm in June, espec i ely. All samples we e il e ed using p e-combus ed glass ibe il e s wi h a nominal po e size o 0.7 μ m (GF/F, Wha man, Japan) and ozen un il analysis. Chemical and da a analyses. DOC concen a ions we e measu ed using he combus ion ca aly ic oxida- ion me hod (TOC-5000; SHIMADZU Co p., Japan). DOM molecula compounds we e analyzed using elec o- sp ay ioniza ion (ESI) coupled o high- esolu ion FT-ICR MS. In his s udy, an FT-ICR mass spec ome e wi h a 9.4-T magne ic ield was used (APEX-Q-94e, B uke Dal onics Inc., MA, USA). Fil e ed samples we e ea ed wi h he C18 solid phase ex ac ion (SPE) me hod o emo e ino ganic sal s57. Wa e samples ex ac ed by his me hod we e dilu ed wi h deionized wa e and me hanol o yield a inal sample composi ion o 50/50 ( / ) o me hanol o wa e . The ex ac ion e iciency anged om 32% o 64%. The ioniza ion e iciency was enhanced by adding ammonium hyd oxide be o e ESI23. Samples we e injec ed in o he FT-ICR mass spec ome e using a sy inge pump a an in usion a e o 100 μ l h–1. All samples we e analyzed in he nega i e ion mode. The elec os- p ay ol age was op imized o each sample. Ions we e accumula ed in a hexapole o 2 s be o e hey we e ans- e ed o he ICR cell, and he 100 ansien s collec ed using a 2 M Wo d ime domain we e co-added. All spec a (Fig.S5) we e ex e nally calib a ed using an a ginine s anda d and in e nally calib a ed using a y acids. Mass lis s we e p oduced using a signal- o-noise a io (S/N) cu -o o 4. Iso ope peaks we e emo ed om he lis . The molecula o mula calcula o (Molecula Fo mula Calcula o e . 1.0; ©NHMFL, 1998) was used o assign an expec ed molecula o mula o each m/z alue wi h a mass accu acy ≤ 1 ppm 23,27. Only m/z alues in he ange o 180–500 we e inse ed in o he molecula o mula calcula o . The ollowing condi ions we e used o o mula assignmen : C = 0–∞ ; H = 0–∞ ; O = 0–∞ ; N = 0–5; S = 0–3; P = 0–3; DBE ≥ 027. A e he o mula assignmen by he molecula o mula calcula o , some o mulas no likely o be obse ed in na u al wa e we e elimina ed based on ules desc ibed in Kujawinski and Behn58 and Wozniak e al.34. To iden i y which biomolecula class each molecula compound belonged o, a an K e elen diag am was used based on he elemen al a ios o he expec ed molecula o mulas, i.e., he oxygen- o-ca bon (O/C) and hyd ogen- o-ca bon a ios (H/C)26 (Fig.S6). He e, we excluded he molecules ha had mul i-candida e o mulas assigned and hus did no ha e a pa icula ange o O/C and H/C alues ha alls in o a biomolecula class. The excluded molecules accoun ed o 40% o he o al m/z peaks on a e age, wi h a iabili y anging om 19% o 55%, and ended o inc ease wi h inc easing molecula weigh . Following he p o ocols p oposed by G annas