Contrasting complexing capacity of dissolved organic matter produced during the onset, development and decay of a simulated bloom of the marine diatom Skeletonema costatum

Con as ing complexing capaci y o dissol ed o ganic ma e p oduced du ing he onse , de elopmen and decay o a simula ed bloom o he ma ine dia om Skele onema cos a um J.I. Lo enzo a,b, ⁎, M. Nie o-Cid c , X.A. Ál a ez-Salgado c , P. Pé ez b , R. Bei as b a Depa men o Biology, Campus Middelheim, Uni e si y o An we p, G oenenbo ge laan 171/U7, BE-2020, An we p, Belgium b Uni e sidad de Vigo, Facul ad de Ciencias del Ma , Dep . de Ecología y Biología Animal, Lagoas-Ma cosende, E-36200 Vigo, Spain c CSIC, Ins i u o de In es igacións Ma iñas, Dep . de Oceanog a ía, Edua do Cabello 6, E-36208 Vigo, Spain Recei ed 2 Feb ua y 2005; ecei ed in e ised o m 28 Feb ua y 2006; accep ed 30 May 2006 A ailable online 25 July 2006 Abs ac The capaci y o na u al dissol ed o ganic ma e (DOM) p oduced du ing he onse , de elopmen and decay o a simula ed bloom o he ma ine dia om Skele onema cos a um o complex ee coppe has been ollowed o a 2 week pe iod. Coppe binding capaci y o he cul u e was measu ed by anodic s ipping ol amme y (ASV) wi h a hanging me cu y d op elec ode (HMDE). The concen a ion o dissol ed o ganic ca bon (DOC) and wo luo opho es, M (humic-like, Ex/Em: 320 nm/410 nm) and T (p o ein- like, Ex/Em: 280 nm/350 nm), we e ollowed du ing he cou se o he incuba ion. Models using DOC concen a ions alone could no accu a ely p edic he complexing capaci y o he cul u e, especially a he end o he bloom, and be e p edic ions we e ob ained when luo escence measu emen s we e conside ed. They we e help ul in cha ac e ising wo ypes o coppe ligands p oduced in he cul u e. The i s ype, aced by he luo escence o peak T, was ela ed o labile DOC di ec ly exuded by phy oplank on. The second ype, aced by he luo escence o peak M, was he e ac o y humic-like ma e ial p esumably p oduced in si u as a by-p oduc o he bac e ial deg ada ion o phy ogenic ma e ials. Du ing he onse and de elopmen o he bloom (days 0 o 7), he luo escence o peak T explains 60–80% o he o al complexing capaci y o he cul u e, sugges ing ha exuded “p o ein-like”compounds among o he exuded complexing agen s e icien ly complexed ee coppe . On he con a y, du ing he decay (days 8 o 13), hese ligands we e eplaced by humic subs ances as he complexing agen o coppe . © 2006 Else ie B.V. All igh s ese ed. Keywo ds: Dissol ed o ganic ma e ; Phy oplank on cul u e; Ligands; Fluo escence; Vol amme y 1. In oduc ion Ma ine dissol ed o ganic ma e (DOM) is he la ges pool o educed ca bon on he Ea h's su ace (Hedges, 2002). I is composed o a my iad o di e en compounds, co e ing a wide ange o molecula weigh s, chemical s uc u es and unc ions. F om he en i onmen al poin o iew, ma ine DOM plays se e al ele an oles (Sunda, 1995), some o hem being connec ed o he a e o me als in he en i onmen : (1) i con ibu es o he ecycling, accumula ion, and expo o biogenic ma e in ocean biogeochemical cycles, (2) i is an elec on dono in he pho o educ ion and solubilisa ion o i on and manganese and (3) i educes he oxici y o ee me al ions by he o ma ion o s able me al/o ganic complexes. Ma ine Chemis y 103 (2007) 61–75 www.else ie .com/loca e/ma chem ⁎Co esponding au ho . Depa men o Biology, Campus Middel- heim, Uni e si y o An we p, G oenenbo ge laan 171/U7, BE-2020, An we p, Belgium. Tel.: +32 32653478; ax: +32 32653497. E-mail add ess: [email p o ec ed] (J.I. Lo enzo). 0304-4203/$ - see on ma e © 2006 Else ie B.V. All igh s ese ed. doi:10.1016/j.ma chem.2006.05.009 Mos o he ma ine DOM is p oduced in si u om he accumula ion and subsequen deg ada ion o he p oduc s o syn hesis o ma ine phy oplank on (Biddanda and Benne , 1997; Ca lson, 2002), and only 10% o igina es on he con inen s (D u ele al.,1992). Mechanisms o DOM p oduc ion include di ec exuda ion by ma ine phy oplank on, cell au olysis in esponse o nu ien s ess condi ions o senescence, and zooplank on media ed libe- a ion in connec ion wi h an ine icien g azing p ocess (Naga a, 2000). Among he p oduc s libe a ed du ing he p oduc ion o ma ine phy oplank on a e pho osyn ha es, essen ially ca bohyd a es and amino acids (Mykles ad, 1995; Biddanda and Benne , 1997; G anum e al., 2002), mainly he la e ha e high capaci y o complex ee me al ions. In addi ion, pho ochemical and bac e ial espi a ion p ocesses gi e ise o a complex se ies o poo ly unde s ood ans o ma ions expe ienced by ma ine DOM, globally known as humi ica ion (Hedges, 1988; Kiebe e al., 1997). Me al complexa ion p ope ies o amino acids ha e been desc ibed a long ime ago and hey ha e been col- lec ed and c i ically e iewedin se e al da abases (Ma ell and Smi h, 1974; IUPAC se ies: Kiss e al., 1991; Pe i and Powell, 1993; Be hon, 1995; Yamauchi and Odani, 1996). Mos amino acids can ac as biden a e me al bind- ing ligands h ough hei α-amino and α-ca boxyla e g oups, and hei s oichiome ic s abili y cons an s a e hus e y simila o he one om glycine (logK Cu–L ∼8.1), because hei side chains a e usually no in ol ed in he me al binding p ocess. These s abili y cons an s a e low o e ec i ely complex me al ions in na u al sys ems because o hei low concen a ions, and specially in sea- wa e , due o he compe i ion wi h o he ca ions and he high ionic s eng h. Howe e , his idine and cys eine p e- sen s ong dono g oups in hei side chains, imidazole and hiol g oups, espec i ely, being much mo e e icien complexing amino acids. Mo eo e , p o eins wi h high me al binding p ope ies (e.g.,me allo hionein, zinc inge pep ides, me allop o eins) a e also ich in hose amino acids (Fa kas and Só ágó, 2002). Humic and ul ic acids, ei he o con inen al o in si u o igin, a e complex high molecula weigh sub- s ances, wi h a capaci y o complex me al ions ha de- pends on he abundance o ca boxylic acid g oups pe uni weigh (Mido ikawa and Tanoue, 1998). In seawa e , hey usually cons i u e 5–15% o he o al DOM concen a ion in ca bon uni s, ca. 8–25 μMC(Benne e al., 1992), al hough in es ua ies and coas al wa e s hey can each concen a ions as high as 125 μMC(Ishiwa a i, 1992; Bu ney, 1994). Coppe is an essen ial elemen o li ing o ganisms, a na u al componen o a a ie y o mine al sal s, and also an elemen used in many indus ial applica ions and in wa e piping. The e o e, na u al and an h opogenic inpu s o coppe o he en i onmen a e bo h ele an (Scheinbe g, 1991). Al hough an essen ial elemen , coppe is one o he mos oxicme als o many o ganismsabo e ce ainle els, especially o aqua ic in e eb a e la ae (His e al., 1999) and phy oplank on species (Sunda and Lewis, 1978). Inc easing a en ion has been paid o he DOM pool in en i onmen al s udies due o i s ole in bu e ing he oxici y o me als o aqua ic o ganisms (Campbell, 1995). The inclusion o DOM in wa e quali y c i e ia has also been ecommended (No , 1987; Allen and Hansen, 1996). The e o e, unde s anding how DOM a ec s me al bio- a ailabili y has been a key issue in ecen yea s. Se e al models ha e been de eloped o p edic he biological e ec o me als in he en i onmen . The bio ic ligand model, BLM (Di To o e al., 2001), combined he ee ion ac i i y model wi h he gill su ace in e ac ion model, and hus conside s me al specia ion in aqua ic sys ems oge- he wi h he e ec o compe ing ca ions a he memb ane su ace o heo ganisms. Ne e heless,disc epanciesha e been obse ed be ween he p edic ions made in e ms o dissol ed o ganic ca bon concen a ions (DOC) and he obse ed bioa ailabili y da a, sugges ing ha u he DOM cha ac e isa ion is needed (Playle, 1998; Ryan e al., 2004). Clea di e ences in me al complexa ion p ope ies ha e been also ound in na u al o ganic ma e om di e en sou ces (Benede i e al., 1996; Abb -B aun and F immel, 1999) ha canno be explained in e ms o hei DOC con en . The aim o his wo k is o s udy he p oduc ion and ans o ma ion o he DOM o igina ed du ing he onse , de elopmen and decay o a simula ed bloom o he ma ine dia om Skele onema cos a um and i s implica- ions o hei coppe binding p ope ies. The luo es- cence o DOM, in addi ion o adi ional case a iables such as DOC and chlo ophyll a, is in es iga ed o ge a be e cha ac e isa ion o DOM in he cul u e. 2. Ma e ial and me hods 2.1. The cul u e The ma ine dia om S. cos a um was cul u ed in a i icial seawa e (Lo enzo e al., 2002) in acid-washed 4- l polyca bona e bo les, wi h he aim o p oducing na u al DOM. Two eplica es, named A and B, we e pe o med. Thecul u eswe ekep a 18°Cunde a14hligh :10hda k pho ope iod (70 μmol quan a/m 2 s) wi h bubbling il e ed ai (0.45 μm PTFE, poly e a luo oe hylene), and gen le magne ic s i ing was p o ided when cell densi y inc eased. On he ini ial day, he a i icial seawa e was en iched wi h hal o he nu ien s o a s anda d /2 media (Guilla d, 1975) 62 J.I. Lo enzo e al. / Ma ine Chemis y 103 (2007) 61–75 wi hou added EDTA (e hylenediamine e aace ic acid). The ini ial concen a ions o N and P we e 700 μMand 18 μM, espec i ely. Bo h bo les we e inocula ed wi h a non-axenic cul u e o S. cos a um in exponen ial g ow h phase o each a inal densi y o 10 4 cells/ml. The pH o he medium was kep be ween 8.05 and 8.25 by adding dilu ed HCl when necessa y. All eagen s used we e analy ical- g ade quali y, o minimize he me al con en o he cul u ing media. The cul u es we e moni o ed along he expe imen ollowing he ime cou se o wo s a e a iables: (1) Chlo ophyll a(Chl a) concen a ion: 20 ml o he cul u e we e il e ed h ough Millipo e APFF glass- ib e p e il e s o o al Chl ade e mina ion. Sam- ples we e kep a −20 °C un il analyses. Chl awas ex ac ed in 90% ace one a −20 °C o ca. 24 h. Samples we e analysed using a Tu ne TD-700 luo ome e calib a ed agains pu e Chlo ophyll a (Sigma). (2) Cell densi y was assessed daily using a Neubaue haemocy ome e . 2.2. Fil a ion p ocedu es Abou 300 ml o each cul u e we e gen ly il e ed (∼150 mm Hg acuum) o minimize cell lysis, h ough acid cleaned 0.22 μm Millipo e PES (polye he sulphone) il e s. Acid cleaned polyca bona e il e holde s we e used o a oid me al con amina ion. The il a e was aken e e y 2 days du ing he cul u e and i was used o dissol ed o ganic ca bon (DOC) and luo escence o dissol ed o ganic ma e (FDOM) analyses and coppe i a ions. 2.3. Dissol ed o ganic ca bon (DOC) The il a e was collec ed in o 10 ml p ecombus ed (450 °C, 12 h) glass ampoules. A e acidi ica ion wi h H 3 PO 4 o pH <2, he ampoules we e hea -sealed and s o ed in he da k a 4 °C un il analysis. DOC was measu ed wi h a Shimadzu TOC-5000 o ganic ca bon analyse . The sys em was s anda dized wi h solu ions o po assium hyd ogen ph hala e. The concen a ion o DOC was de e mined by sub ac ing he a e age peak a ea om he ins umen blank a ea and di iding by he slope o he s anda d cu e. The CV was ∼1% and he accu acy o he measu emen s was success ully es ed wi h he TOC e e ence ma e ials p o ided by P o . D. Hansell (Uni e si y o Miami). 2.4. Fluo escence o dissol ed o ganic ma e (FDOM) Aliquo s o he DOC il a e we e collec ed be o e Cu addi ion in o 20 ml acid cleaned polyca bona e lasks and ozen a −20 °C un il analysis. FDOM o hese samples was measu ed wi h a Pe kin Elme LS 55 Luminescence spec ome e . The ins umen was equipped wi h a xenon discha ge lamp, equi alen o 20 kW o 8 μs du a ion, and a 1 cm qua z luo escence cell. Milli-Q wa e was used as a e e ence o luo escence analysis, and he in ensi y o he Raman peak was checked egula ly. Disc e e exci a- ion/emission pai measu emen s we e pe o med a peaks M (ma ine humic subs ances, Ex/Em: 320 nm/410 nm) and T (“p o ein-like subs ances”, Ex/Em: 280 nm/350 nm), speci ically yp ophan (Coble, 1996; Nie o-Cid e al., 2005; S edmon and Ma kage , 2005). Fou eplica e mea- su emen s we e pe o med o each Ex/Em wa eleng h pai s. A ou poin s anda d cu e was p epa ed daily wi h a mixed s anda d o quinine sulpha e (QS) and yp ophan (T p) in 0.05 mol/l sulphu ic acid (concen a ions om +0 o +150 ppb and om +0 o +100 ppb o QS and T p, espec i ely; Nie o-Cid e al., 2005). The equi alen con- cen a ion o e e y peak was de e mined by sub ac ing he a e age peak heigh om he blank heigh , and di iding by he slope o he s anda d cu e. Fluo escence uni s we e exp essed in ppb equi alen s o QS (ppb eq QS) o FDOM M and ppb equi alen s o T p (ppb eq T p) o FDOM T . No e ha ppb eq QS is iden ical o he quinine sulpha e uni s (QSU) epo ed p e iously in o he s udies. The p ecision o he luo escence measu emen s was ±0.1 ppb eq QS and ±0.6 ppb eq T p, espec i ely. In addi ion o hese measu emen s, a ial was pe - o med wi h he il a e om he six h day wi h di e en addi ions o Cu ( anging om 0 o 5 μM) o check he e ec o Cu complexa ion in he di e en luo opho es. Twen y-millili e samples we e spiked wi h inc easing Cu concen a ions and allowed o equilib a e o 24 h in he da k. A e ha ime, FDOM M and FDOM T we e measu ed. 2.5. Cu measu ing condi ions Squa e wa e anodic s ipping ol amme y (ASV) analyses we e ca ied ou wi h a hanging me cu y d op elec ode, a Ag/AgCl e e ence and a P - od auxilia y elec ode held in a Me ohm 663 VA pola og aphic s and coupled o an Eco-Chemie Au oLab PGSTAT10 po en- ios a . Samples we e placed in a 20 °C he mos a ic glass cell. A e solu ions had been pu ged o 250 s wi h N 2 , coppe was accumula ed on a me cu y d op o 0.52 mm 2 a −0.55 V o 15 s a he maximum s i ing speed (3000 pm) and 10 s o equilib a ion we e allowed be o e he ol age scan. The condi ions o he squa e- wa e (SW) scan we e an ini ial po en ial o −0.55 V, an SWampli ude o 25 mV, an SW equency o 25 Hz and a scan inc emen o 2 mV. Th ee ol ammog ams we e 63J.I. Lo enzo e al. / Ma ine Chemis y 103 (2007) 61–75 eco ded o each solu ion. To minimize po en ial e ec s o adso p ion o Cu o cell walls, cells we e only washed wi h sample solu ion du ing each i a ion. Samples wi h inc easing Cu addi ions we e measu ed in o de , s a ing om he cul u e il a e wi h no Cu addi ion. 2.6. Coppe i a ions Be ween 13 and 16 samples o 16 ml cul u e il a es we e dosed in 20 ml polyp opylene ials. Solu ions we e s o ed a −20 °C un il analyses. Be o e analyses, Fig. 1. Time e olu ion o he Skele onema cos a um cul u e and he dissol ed o ganic ma e concen a ion and composi ion. (a) Numbe o cells measu ed in cul u es A and B. E o ba s ep esen he 95% con iden ial in e als o he mean o h ee eplica es. (b) DOC concen a ion in cul u es A and B compa ed wi h cell numbe s. (c) Fluo escence cha ac e isa ion o he DOM pool: FDOM M is he luo escence o humic subs ances in quinine sulpha e uni s (ppb eq QS), and FDOM T is he luo escence o he a oma ic amino acids in yp ophan uni s (ppb eq T p). 64 J.I. Lo enzo e al. / Ma ine Chemis y 103 (2007) 61–75 inc easing Cu addi ions ( om 0 o 5 μmol/l) we e made o each se o samples om a dilu ed spec opho ome ic Cu s anda d solu ion (1000 ppm, Pan eac). Solu ions we e kep 24 h in he da k o allow he equilib a ion o he complexa ion eac ion. All addi ions we e p epa ed and analyses we e pe o med on a lamina low, class 100 il e ed ai , clean bench. All plas ic lab wa e was kep in 5% HNO 3 a leas 24 h and insed 5 imes wi h deionised wa e pu i iedbyion exchange( esis i i y≥18.2 MΩ/cm; Milli-Q®) be o e use. The peak heigh o each ol ammog am was mea- su ed, and i was ans o med in o labile Cu concen a- ions ([Cu′]) di iding by he slope o he linea segmen o each i a ion (Dona e al., 1994). Labile coppe is he equi alen concen a ion o coppe ha con ibu es o he ol amme ic signal o he same ex en as he ee me al ion (Mulle e al., 2003). The i a ion cu es we e explained assuming he simples complexa ion model; ha is, only one ype o ligand and a eac ion s oichi- ome y o 1:1. Ti a ion plo s we e i ed o Eq. (1), ob ained om he heo e ical complexa ion model p e iously explained (Lo enzo e al., 2002): CuV¼aþffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi a2þ4CuT=KV p2ð1Þ whe e a=(−Cu T +L+1/K′), and Cu T and Cu′a e he o al and labile coppe concen a ions (mol/l), espec i ely, L is he o al ligand concen a ion (mol/l) and K′is he condi ional s abili y cons an o he Cu–L complexes, alid o he condi ions o he expe imen al medium. This Fig. 2. E olu ion o he measu ed luo opho es du ing he cul u e. (a) Compa ison o FDOM T and DOC; (b) compa ison o ne p oduc ion a es o FDOM T and DOC along he cul u e; (c) compa ison o FDOM M and DOC; and (d) FDOM T ne p oduc ion a e e sus ime. Reg ession lines we e i ed o model II in (a) and (b), and o model I in (c) and (d), espec i ely. The eg ession equa ions (±s anda d e o ), he co ela ion coe icien s, and he signi icance o he ela ionships a e also p esen ed. 65J.I. Lo enzo e al. / Ma ine Chemis y 103 (2007) 61–75 K′is e e ed o labile coppe concen a ions, no o ee coppe ion concen a ions. I is necessa y o conside ino ganic coppe specia ion in he medium o con e ing K′ o K′ Cu 2+ (Dona e al., 1994). KV Cu2þ¼KVdaCu2þð2Þ whe e α Cu 2+ is he ino ganic eac ion side coe icien , which accoun s o he ino ganic coppe specia ion. A alueo 36wasassumed(Leal e al., 1999) o calcula ions. 2.7. S a is ical analysis The bes - i be ween any couple o a iables (X,Y) was ob ained minimizing he unc ion: X i ðXi−XiÞwXðYi−YiÞwY  2ð3Þ whe e w X and w Y a e weigh s o Xand Y espec i ely, wi h w X ,w Y ≥0 and w X +w Y =1. Any eg ession model can be exp essed as one o wo ex eme cases: (a) model I, which should be applied when w X =0, w Y =1 and (b) model II when w X =w Y =0.5 (Sokal and Rohl , 1995). S a is ical analyseswe epe o medwi hSTATISTICA e sion 6.0 (S a So ). The pa ame e s o he complexa- ion model in Cu i a ions we e es ima ed by means o non-linea i ing p ocedu es wi h Sigma Plo 2001 e - sion 7.0 (SPSS), which uses i e a i e p ocesses o minimizing he sum o he squa es o he esiduals by means o he Ma qua d -Le enbe g algo i hm. 3. Resul s 3.1. Onse , de elopmen and decay o he cul u ed S. cos a um bloom Fig. 1 shows he ime e olu ion o he cul u e. Simila pa e ns we e obse ed in bo h lasks (A and B) o he di e en measu emen s. Fig. 1a shows he loga i hmic inc ease in he numbe o cells and chlo ophyll a(Chl a) concen a ion du ing he ini ial 4 days. No lag ime was obse ed in he cul u e g ow h. F om days 4 o 6, he numbe o cells emained s a iona y, and a e day 8 he e was a con inuous dec ease in cell densi y. The maximum Chl aconcen a ion (700–1000 μg/l) was achie ed a day 7, and a e ha , a con inuous dec ease was also obse ed o Chl ale els. Fig. 1b shows he high co ela ion be ween DOC accumula ion in he cul u e and cell densi y du ing days 0 o 7 ( he anabolic phase o he bloom). The e was a con inuous inc ease in he DOC con en o he cul u e om he ini ial day, bu his inc ease was much highe om days 4 o 5. A e day 5, a clea DOC dec ease was obse ed. A la ge disc epancy was obse ed a e day 8 (named he ca abolic phase o he bloom), when cell densi y dec eased whe eas he DOC con en inc eased, especially a he end o he cul u e (no e he change o scale). High bac e ial densi ies we e obse ed a e day 4 Fig. 3. Fluo escence quenching in he p esence o coppe . Fil e ed samples om he six h day o he cul u e we e spiked wi h inc easing Cu concen a ions and luo escence o “humic-like”subs ances (FDOM M ) and o “p o ein-like”subs ances (FDOM T ) was measu ed a e 24 h. Fi ings o obse ed da a o exponen ial decay equa ions a e also p esen ed. 66 J.I. Lo enzo e al. / Ma ine Chemis y 103 (2007) 61–75 ( e ained in 0.22 μm il e s), and one sample was incu- ba ed o es bac e ial consump ion o he DOC in he cul u e. One aliquo o he cul u e in day 5 was il e ed h ough a 5 μm po e size GF/F il e , which emo es phy oplank on bu no bac e ia, and incuba ed o 24 h in he same cul u e condi ions. A DOC dec ease om 0.50 o 0.14 mM C was obse ed. Fig. 1c shows he e olu ion o he wo luo opho es measu ed du ing he cul u e. FDOM T ep oduced he same empo al pa e n han he DOC concen a ion (Fig. 2a). A high co ela ion was obse ed be ween bo h a iables, bu he co ela ion was e en highe when p o- duc ion/deg ada ion a es o FDOM T and DOC we e compa ed up o day 11 (Fig. 2b). FDOM M showed a comple ely di e en pa e n. A mono onic inc ease was obse ed in he cul u e and FDOM M was no signi ican ly co ela ed wi h DOC (Fig. 2c). The same esul was obse ed when compa ing FDOM M and DOC p oduc ion a es. Howe e , FDOM M p oduc ion a es we e highly co ela ed wi h ime (Fig. 2d). A signi ican dec ease in luo escence was obse ed when Cu addi ions we e made o samples con aining bo h luo opho es (Fig. 3). FDOM T was mo e a ec ed by Cu addi ions han FDOM M , and a 32% e sus a 20% luo escence dec ease we e measu ed, espec i ely. 3.2. E ol ing coppe binding capaci y o DOM p oduced by he cul u e Fig. 4a shows he e olu ion o Cu i a ion plo s in he cul u e o e ime. Some speci ic days we e selec ed o clea ly show he shi in he plo s o he igh , due o he inc ease in Cu complexing capaci y. The complex- a ion model p e iously desc ibed (Eq. (1)) accu a ely i ed he da a, and he ob ained 2 anged be ween 0.997 and 0.9996, e en when indi idual measu emen s and no mean alues whe e used o i ing pu poses. Coppe i a ion i ing pa ame e s (±95% con iden- ial in e al, CI) a e shown in Table 1. The e was a good ag eemen be ween he pa ame e s ob ained in bo h cul u es, as p e iously obse ed wi h DOC measu e- men s (Fig. 2b). The e was a ma ked inc ease in ligand concen a ion o e ime (see also Fig. 4b). The ligand inc ease showed an in e media e end o hose obse ed o bo h luo opho es: an ini ial and sha p inc ease a he end o he exponen ial g ow h (days 4–5) coinciding wi h FDOM T and no dec ease be ween days 5 and 8, as Table 1 Complexing capaci y (L±95% CI, μmol Cu/l) and condi ional s abili y cons an (logK′ Cu 2+ ±95% CI) in cul u es A and B a di e en ime in e als; n, numbe o coppe addi ions in each i a ion Time (days) L A (μmol Cu/l) LogK′ Cu 2+ nL B (μmol Cu/l) LogK′ Cu 2+ n 0 0.04±0.01 9.5±2 12 0.08±0.01 9.4±0.9 14 2 0.12±0.01 8.4±0.2 15 0.13±0.02 8.6±0.5 15 4 0.40±0.03 8.2±0.2 12 0.24±0.02 8.4±0.2 15 5 0.64±0.02 8.60±0.02 15 0.57±0.03 8.2±0.1 13 6 0.64±0.04 8.64±0.04 16 0.58±0.02 8.5±0.1 16 8 0.79±0.02 9.0±0.2 14 0.84±0.02 8.82±0.09 15 11 0.92±0.06 8.3±0.2 14 1.00±0.05 8.3±0.1 14 13 1.12±0.03 8.41±0.09 12 Fig. 4. (a) Time e olu ion o Cu i a ions o he il e ed cul u e. Solid lines ep esen he bes non-linea squa es i ings o he da a o Eq. (1). (b) Time e olu ion o me al ligand concen a ion in he cul u e. E o ba s ep esen he 95% con iden ial in e als ob ained om he s anda d e o s o he i ings. 67J.I. Lo enzo e al. / Ma ine Chemis y 103 (2007) 61–75 FDOM M . Day 8 showed s a is ically highe condi ional s abili y cons an s, coinciding wi h he ini ia ion o he ca abolic phase o he cul u e. The ligand capaci y a he ini ial day o he cul u e was e y low, which was also e lec ed in he low con idence o he ob ained K′ alue. The sample om day 5 incuba ed o bac e ial con- sump ion was also i a ed wi h Cu. The ligand concen- a ion in heincuba ed sample dec eased om 0.64±0.02 o 0.38±0.04 μM, i.e. a 40% dec ease. This esul con- as s wi h he 72% dec ease o DOC concen a ion in he same sample. 3.3. DOM composi ion e sus Cu binding capaci y Fig. 5a shows he ligand concen a ion measu ed du ing he cul u e e sus he DOC con en in he il a es. In he anabolic phase, a signi ican co ela ion was obse ed be ween he DOC concen a ion and he me al binding capaci y (ligand concen a ion) o he cul u e. Up o 70% o he ligand a iabili y could be explained by he DOC concen a ion. Howe e , when he ca abolic phase o he cul u e was conside ed, he eg ession did no p edic he ligand concen a ion. A sha p inc ease in ligand concen a ion was obse ed be ween days 8 and 11 coinciding wi h a clea dec ease in DOC. The opposi e end was obse ed on day 13, when a 10- old inc ease in DOC was p oduced, bu jus a small inc ease in ligand concen a ion was obse ed. The wo luo opho es we e used o p edic he ligand concen a ion in he cul u es (see Fig. 5b and c). FDOM T was he bes p edic i e a iable: 88% o he a iabili y o he ligand concen a ion in he anabolic Fig. 5. Rela ionships be ween he concen a ion o ligand in he cul u e and di e en measu emen s o he DOM con en in he medium. (a–c) Ligand concen a ions exp essed as a unc ion o he DOC, FDOM T and FDOM M o he cul u e, espec i ely. (d) Obse ed e sus expec ed ligand concen a ions ob ained om hemul iple model II eg ession. The mul iple eg ession (see ex , Eq. (4)) was calcula ed wi h da a om he anabolic phase.Whi e andg ey do s ep esen da a om he ca abolic phase i ed o he p e ious eg ession be o e and a e FDOM T co ec ions. Black do s and long dashed lines co espond o da a and i ings ob ained in he anabolic phase o he cul u e. Open do s and do ed lines co espond o da a and i ings om he ca abolic phase. The eg ession equa ions (±s anda d e o ), he co ela ion coe icien s and he signi icance o he model II ela ionships a e also p esen ed. 68 J.I. Lo enzo e al. / Ma ine Chemis y 103 (2007) 61–75 phase can be explained by FDOM T . The eg ession model wi h FDOM M was also signi ican , bu i did no imp o e he p edic i e powe o he model ob ained wi h he DOC concen a ion. None o hese h ee eg ession models es ed in he anabolic phase can be success ully applied o he ca abolic phase o he cul u es, and clea disc epancies wi h he p edic ed alues o he ligand concen a ions we e obse ed al eady o he measu ed alues on day 8 (open ci cles in Fig. 5a, b and c). A mul iple eg ession model conside ing he con i- bu ion o bo h luo opho es o he ligand concen a ion was also es ed in he anabolic phase o he cul u es. The ollowing equa ion was ob ained: ½L¼0:05ðF0:03Þþ0:007ðF0:001ÞFDOMT þ0:008ðF0:002ÞFDOMMð4Þ whe e L is he ligand concen a ion (μmol Cu/l), and he s anda d e o o p edic ed pa ame e s a e shown in b acke s (n=11). The eg ession model explains 96% o he obse ed ligand a iabili y in he cul u e (see black do s o Fig. 4d). A signi ican con ibu ion o bo h FDOM T and FDOM M o he ligand concen a ion was obse ed (p<0.001 and p=0.005, espec i ely). When he same model was applied o he ca abolic phase o he cul u e, good p edic ions we e ob ained o day 8. None o he p e ious models conside ing only DOC, FDOM T o FDOM M could p edic ligand concen a ions a his day. Howe e , expec ed ligand concen a ions (open do s in Fig. 4d) o he las days o he cul u e (days 11 and 13) depa ed signi ican ly om measu ed alues. Model p edic ions o e es ima e mea- su ed ligand concen a ions, coinciding wi h he clea decay o he cul u e. P obably, FDOM T a his pe iod is no ela ed o labile exuded DOC by algae bu o DOC libe a ed du ing cell lysis o he dead cul u e. The e o e, a new pa ame e was in oduced o Eq. (4) o accoun o a possible dec ease in ligand capaci y o measu ed FDOM T a his pe iod: ½L¼0:05 þ0:007ð1−aÞFDOMT þ0:008FDOMMð5Þ whe e αis he % o FDOM T wi hou me al binding p ope ies. Good i ings wi h he obse ed ligand con- cen a ions we e ob ained conside ing ha 85% and 100% o FDOM T had no me al complexing p ope ies in days 11 and 13, espec i ely (g ey do s in Fig. 4d; Table 2). The pa ame e s ob ained om Eq. (4) a e no di ec ly compa able, because bo h luo opho es a e no ex- p essed in he same uni s. Howe e , i is possible o calcula e he con ibu ion o bo h ac ions o he ligand concen a ion by conside ing he measu ed FDOM T and FDOM M a any ime (see Table 2). Du ing he ini ial phase o he cul u e, “p o ein-like subs ances” oge he wi h o he labile DOC compounds ( aced by FDOM T ) di ec ly exuded om phy oplank on li ing cells made up mos o he me al complexing capaci y o he cul u e (be ween 60% and 80%). Howe e , a e day 6, humic subs ance con ibu ion o he ligand concen a ions s a ed o be dominan . 4. Discussion 4.1. DOC p oduc ion du ing he cul u e Two phases ha e been dis inguished in he e olu ion o he cul u e, named anabolic and ca abolic. The anabolic phase coincided wi h he g owing and s a iona y phases (days 0 o 7), cha ac e ised by he exuda ion o DOM by S. cos a um li ing cells. The ca abolic phase coincided wi h he decay o he bloom (days 8 o 13), p obably caused by P nu ien limi a ion, cha ac e ised by inc eased cell lysis. DOC p oduc ion by S. cos a um has been desc ibed p e iously in ield (Igna iades, 1973) and in cul u e (Igna iades and Fogg, 1973) condi ions. The amoun and composi ion o DOC eleased by ma inedia oms has been p o en o be dependen on he physiological s a us o he cul u e (Chen and Wage sky, 1996a; Te zićand Ahel, 1998; G anum e al., 2002). Whe eas low molecula weigh ma e ials a e eleased in he exuda ion p ocesses du ing he g owing phases o cul u es, high molecula weigh ma e ials a e dominan in he senescen phases (Chen and Wage sky, 1996b). Table 2 Pe cen age o he measu ed ligand concen a ion explained by he FDOM T and FDOM M luo opho es (±s anda d de ia ion) du ing he cou se o he Skele onema cos a um cul u e acco ding o Eq. (4) Time(days) %L FDOM M %L FDOM T % Non-complexing FDOM T 0 28±1 72±1 0 2 22±2 78±5 0 4 37±13 63±23 0 5 28±2 72±6 0 6 44±1 56±1 0 8 58±2 42±2 0 11 92±6 8±1 85 13 100 0 100 The es ima ed pe cen age o FDOM T wi hou me al binding p ope ies acco ding o Eq. (5) in he anabolic phase is also shown in he ou h column. 69J.I. Lo enzo e al. / Ma ine Chemis y 103 (2007) 61–75