Molecular composition of dissolved organic matter from a wetland plant (Juncus effusus) after photochemical and microbial decomposition (1.25 yr): Common features with deep sea dissolved

Page 1

Organic Geochemistry 60 (2013) 62–71

Contents lists available at SciVerse ScienceDirect

Organic Geochemistry

journal homepage: www.elsevier .com/locate /orggeochem

Molecular composition of dissolved organic matter from a wetland plant(Juncus effusus) after photochemical and microbial decomposition(1.25 yr): Common features with deep sea dissolved organic matter

0146-6380/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.orggeochem.2013.04.013

⇑ Corresponding author. Tel.: +49 441 798 3348; fax: +49 441 798 3404.E-mail address: [email protected] (P.E. Rossel).

Pamela E. Rossel a,⇑, Anssi V. Vähätalo b,c, Matthias Witt d, Thorsten Dittmar a

a Max Planck Research Group for Marine Geochemistry, University of Oldenburg, ICBM, D-26111 Oldenburg, Germanyb Department of Biological and Environmental Sciences, University of Helsinki, FIN-00014 Helsinki, Finlandc Department of Biological and Environmental Science, University of Jyväskylä, Survontie 9, FI-40014 Jyväskylä, Finlandd Bruker Daltonics, Fahrenheitst. 4, 28359 Bremen, Germany

a r t i c l e i n f o a b s t r a c t

Article history:Received 21 December 2012Received in revised form 20 April 2013Accepted 30 April 2013Available online 9 May 2013

We hypothesized that microbial and photochemical processing of dissolved organic matter (DOM) deter-mines its molecular formula composition in aquatic systems to a greater degree than does the originalsource of the DOM. To test this hypothesis, we exposed DOM from a leachate of a wetland plant (Juncuseffusus) to solar radiation or incubated it in the dark for 1.25 yr. Analysis of the extracted DOM of theleachates via Fourier-transform ion cyclotron resonance mass spectrometry (FT-ICR-MS) identified2800 molecular formulae. Of the formulae in the initial DOM, 11% were lost during microbial decompo-sition in the dark and 54% under solar radiation. Solar radiation also produced a large number of formulaecontaining N, that were preferentially degraded by microorganisms (47% loss). We compared the ‘‘recal-citrant formulae’’, i.e. those not degraded in the experiment, with those of DOM from the deep North Paci-fic Ocean. Of the deep sea DOM formulae, 18% were present in the recalcitrant fraction of the initial DOM.An additional 18% of the formulae in marine DOM were photoproduced and recalcitrant, and 8% wereproduced by microbes in the experiment. Consequently, 44% of the deep sea DOM shares identical molec-ular formulae with the recalcitrant DOM from the experiment, most of which were produced by the com-bined action of sunlight and microbes. This indicates that processes in the water column may be moreimportant than the original source in determining the composition of bulk DOM.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

During the transport of dissolved organic matter (DOM) fromterrestrial, wetland and littoral environments to the ocean, selec-tive degradation takes place. Microbial decomposition, startingwith a labile part of the DOM, leaves a major recalcitrant fractionthat is transported to the open ocean (Wetzel, 1992). Part of thisterrestrial material, especially the light-absorbing aromatic frac-tion, is photoreactive and susceptible to degradation during solarexposure (Kouassi and Zika, 1992; Moran et al., 2000). This light-absorbing material (chromophoric DOM, CDOM), has an impacton ocean color (Coble, 2007) and influences the carbon cycle dueto its role in the biogeochemical processes for which solar radia-tion is important, such as primary production (Blough and DelVecchio, 2002). The role of photolysis on DOM alteration has beenwidely discussed and extensively studied in short term experi-ments of days to 1 or 2 months (Hernes and Benner, 2003;Kujawinski et al., 2004; Lou and Xie, 2006; Dittmar et al., 2007;

Gonsior et al., 2009; Stubbins et al., 2010). The effect of long termirradiation under natural conditions on timescales of months toyears has, however, been poorly studied (Vähätalo and Wetzel,2008). Combined effects of both photochemical and microbial pro-cesses are also considered important for the removal of DOM fromthe ocean (Mopper et al., 1991; Amon and Benner, 1996; Miller andMoran, 1997; Opsahl and Benner, 1998; Hernes and Benner, 2003).In addition to DOM decomposition due to solar radiation, produc-tion of non-bioavailable DOM as a consequence of photochemicaltransformation has been observed (Benner and Biddanda, 1998;Obernosterer et al., 1999). Studies have suggested that recalcitrantDOM may derive from plant constituents introduced to the oceanafter extensive degradation (Thurman, 1985) or from condensationreactions after solar exposure (Harvey et al., 1983). Furthermore, ifrecalcitrant OM results from freshly produced DOM in surfacewater due to solar exposure (Benner and Biddanda, 1998), it is alsopossible that a fraction of this material may accumulate over timeand persist in the deep sea DOM pool. The material could returnfrom the deep ocean to the surface, after being highly degraded,which may, at least in part, make it more photoreactive and ableto release bioavailable compounds after solar exposure, as

Page 2

P.E. Rossel et al. / Organic Geochemistry 60 (2013) 62–71 63

suggested by Benner and Biddanda (1998). Thus, both solar radia-tion and microbial decomposition are important processes alteringthe DOM pool in the water column due to both the removal of or-ganic compounds as well as the production of recalcitrant material.

We hypothesized that long term microbial and photochemicalprocessing of DOM would determine the molecular compositionof DOM in aquatic systems to a larger degree than the originalsource of DOM. To test this hypothesis, we exposed DOM from aleachate of a wetland plant (Juncus effusus) to solar radiationand/or incubated it in darkness for a total of 1.25 yr. The timescaleused for solar exposure was considered long enough to completethe photodegradation of CDOM. J. effusus, and vascular plant leach-ates in general, are most likely not the prime source of DOM in theocean (Hedges et al., 1997; Hernes and Benner, 2003, 2006). If ourhypothesis was correct, DOM from the experiments should havelittle molecular similarity with deep sea DOM at the beginning ofthe treatment, but should share major common molecular featureswith deep sea DOM at the end of the treatments. As a reference fordeep sea DOM, we used DOM from the O2 minimum zone of theNorth Pacific off Big Island, Hawaii. The molecular composition ofDOM was assessed via Fourier-transform ion cyclotron resonancemass spectrometry (FT-ICR-MS). The high resolution and high massaccuracy of FT-ICR-MS resolved the exact mass of thousands ofindividual compounds and allowed calculation of molecular for-mulae. This holistic molecular approach allowed us to follow thou-sands formulae in the course of the dark and light incubations andto compare experimental DOM with deep sea DOM at a molecularformula level.

2. Material and methods

2.1. Sample description and experimental setup

The experiment was performed with DOM leached from a com-mon vascular wetland plant, J. effusus, which is widely distributedin the northern hemisphere (Vähätalo and Wetzel, 2008). Wetlandsand their vascular macrophytes are the most important sources ofallochthonous DOM in surface water on the continents (Wetzel,1992). For details of the experimental setup, see Vähätalo andWetzel (2008). In brief, the plant leachate was prepared by inun-dating 30 l of a 1:1 mixture of senescent and fresh leaves from J.effusus, collected from Talladega (Alabama) wetland ecosystem,in 38 l deionized water. The mixture was degraded in a containerwhere dissolved O2 was rapidly consumed. After 19 weeks ofmostly anaerobic degradation, the water was gravity filteredthrough Whatman 50 low ash filter paper, followed by vacuum fil-tration through Whatman GF/F (0.7 lm) filters. The collected 8 l offiltrate were then lyophilized to obtain the ‘‘initial DOM’’ (6.7 g dry

Fig. 1. Absorption coefficient by (a) CDOM at 350 nm and concentration of (b) TOC initi(dark control), under solar radiation in the presence of low (light and low microbes) ordeviation of replicated determinations (n = 3–5; Vähätalo and Wetzel, 2008). The error bawith the mean for initial leachates.

wt.). Decomposition for 19 weeks before the start of the actualexperiment was performed to mimic the process occurring alsoin O2-limited environments, such as wetland soils. Furthermore,this anoxic degradation, in combination with the following aerobicdecomposition during the experiment, could account for additionalmicrobial capabilities during DOM transformation. The lyophilizedDOM was dissolved in GF/F-filtered water (3 l with 2.3 mg l�1 ofdissolved organic carbon, DOC) from Lake Tuscaloosa (Alabama,USA) to obtain a final concentration of total organic carbon (TOC)of 847 mg l�1 (Vähätalo and Wetzel, 2008). This DOM solutionwas distributed into seven 700 ml quartz flasks and one glass flaskand autoclaved before the experimental treatments. The headspaceof the flasks was filled with O2 to ensure sufficient oxygen formicrobial and photochemical oxidation. O2 was sequentially intro-duced over the course of the experiment to prevent anoxia (Vähät-alo and Wetzel, 2008). The quartz flasks were exposed to theambient solar radiation, while the glass flask was treated in thesame way except that it was wrapped in Al foil to serve as a darkcontrol. The flasks were placed horizontally on the roof top ofthe University of Alabama (Tuscaloosa, 33�1205900N; 87�2303600W)from 23 June to 29 July 2001 and from 8 August 2001 to 25 Sep-tember 2002 at the University of North Carolina (Chapel Hill,35�5402100N; 79�0301600W) (Vähätalo and Wetzel, 2008). Of thequartz flasks exposed to light, which received an identical amountof solar radiation, four became turbid from microbial growth dur-ing exposure (further referred to as ‘‘high microbes’’), while twodid not exhibit visible microbial turbidity (‘‘low microbes’’). It isnecessary to keep in mind that, despite the absence of visiblemicrobial turbidity, all treatments contained viable bacteria(Vähätalo and Wetzel, 2008).

According to Vähätalo and Wetzel (2008), TOC concentrationwas 847 mg l�1 and the absorption coefficient at 350 nm(aCDOM,350) was 2720 m�1 at the beginning of the experiment. Mi-crobes mineralized 38% of TOC and 41% of aCDOM,350 was lost duringthe 1.25 yr dark treatment (Fig. 1). The exposure to solar radiationbleached 99.9% of aCDOM,350 and decomposed 90% and 72% of TOC,in the presence of high or low microbial activity, respectively.

We compared J. effusus leachates on a molecular level with theSuwannee River fulvic acid reference material obtained from theInternational Humic Substances Society (IHSS). The initial andresulting DOM mixtures from the experiment were also comparedwith marine DOM collected at 700 m water depth from a stationin the North Pacific Ocean off Big Island (Hawaii) operated by theNational Energy Laboratory of Hawaiian Authorities (NELHA). ThisDOM was obtained from the O2 minimum zone (20–30 lM O2),which is also the NO�3 maximum [43 lM; P16, US Climate Variabilityand Predictability (CLIVAR) hydrographic program, http://www.cli-var.org/resources/data/clivar-carbon-and-hydrographic-sections].

ally in the Juncus effusus leachate (initial DOM), after 459 day treatment in the darkhigh (light and high microbes) microbial activity. The bars show mean ± standard

rs and the CDOM in the solar radiation treated leachates are barely visible compared

Page 3

64 P.E. Rossel et al. / Organic Geochemistry 60 (2013) 62–71

The absence of chlorofluorocarbons (CFC-11) and tritium indicatethat this water mass was old and had not been recently in contactwith the atmosphere (CLIVAR hydrographic program). Thus, DOMfrom this water mass was considered a good representation of deepsea DOM.

2.2. Sample extraction and FT-ICR-MS analysis

Immediately after completion of the respective treatments, thewater samples were lyophilized and stored in darkness at roomtemperature. All samples were initially analyzed using FT-ICR-MSafter dissolving the lyophilized samples in water:MeOH (1:1) with-out further purification. However, due to matrix interference, onlyfew molecular formulae (< 100) could be detected. To avoid matrixinterference, solid phase extraction (SPE) with a styrene divinylbenzene polymer column (Varian PPL) was performed prior tomolecular analysis. This procedure is one of the most efficientmethods for extracting DOM for MS analysis (Dittmar et al.,2008). PPL has a pore size of 150 Å and retains a wide spectrumof highly polar to non-polar compounds (Dittmar et al., 2008). Col-loidal material and small ionic compounds might be selectivelylost during the procedure. For SPE, lyophilized material from eachtreatment was dissolved in ultrapure water to a final concentrationof 42 mg C l�1. Before extraction, this dissolved material and thedeep seawater were acidified with HCl (analytical grade) to pH 2and passed via gravity through a PPL cartridge previously rinsedwith MeOH (ultrapure MS grade). Before elution, the cartridgeswere rinsed with several cartridge volumes of ultrapure water atpH 2 to wash out the salt, and dried under a stream of ultrapureN2. The sorbed DOM was then eluted with MeOH. Aliquots of theMeOH extracts were evaporated overnight and redissolved inultrapure water at pH 2 for DOC analysis. DOC concentration wasobtained using catalytic oxidation at high temperature on a TOC-V Shimadzu instrument. Based on comparison between the DOCconcentration of the original samples and their SPE-DOM, calcu-lated extraction efficiency was between 50% and 65%. Beforemolecular analysis, an aliquot of SPE-DOM and Suwannee Riverfulvic acid reference material were dissolved in a solution of 1:1MeOH:water to a final DOC concentration of ca. 12 mg C l�1. Thisstandardized sample matrix and DOC concentration enabled max-imum and reliable comparability of the FT-ICR-MS results from thedifferent samples analyzed the same day.

Molecular analysis was performed in negative ion mode with a12 Tesla FT-ICR-MS instrument (Bruker Solarix) equipped with anelectrospray ionization (ESI) source. ESI is a soft ionization tech-nique, which keeps the molecular ions intact and is widely usedin the analysis of natural organic mixtures (Koch et al., 2005,2008; Dittmar and Koch, 2006; Kujawinski and Behn, 2006; Tremb-lay et al., 2007; Gonsior et al., 2009; D’Andrilli et al., 2010). Thesamples were infused at 120 ll h�1 using a capillary voltage of4500 V and an ion accumulation of 0.1 s. Additionally, a skimmervoltage of the second ion funnel of 40 V was applied to avoid clus-ter formation. External calibration was performed based on argi-nine cluster, and internal linear calibration was performed with alist of > 50 known molecular formula mass peaks from m/z 250to 500 in each sample and Suwannee River fulvic acid referencematerial. To calculate formulae, the data were analyzed with DataAnalysis software (ESI Compass 1.3, Bruker Daltonics). Formula cal-culation was performed after the data were filtered via a signal tonoise (S/N) > 4 using the following criteria: C P O; O > (2P + S);H 6 2C + 2; N 6 3; S 6 2 and P 6 2. Additionally, formula assign-ments were isotopically confirmed for the most intense peaksand, based on this information, CH2 homologous series and CH2Ohomologous series were extended over the entire mass range.The CH2 homologous series differ only by multiples of CH2, whilethe CH2O homologous series are additionally separated by O atoms

(oxygen series). Formulae belonging to a homologous series withmore than two formulae were retained. In case of double assign-ments, CH2 and CH2O homologous series extended to lower massrange were considered correct. All assigned formulae, when themeasured and calculated masses were compared, had an er-ror < 0.5 ppm. The same criteria were used in the case of formulaeassignments for the deep sea DOM. In addition, we compared theexperimental DOM with deep sea DOM without applying the S/Nvalue to the deep sea DOM. This additional test provided a mostconservative limit for testing the presence of specific compoundsin marine DOM (details in Table 2 footnote ‘‘a’’).

After formulae were assigned, they were grouped according tothe following defined groups of compounds (note that some ofthe groups overlap): (i) ‘‘recalcitrant’’ is defined as the formulaein the initial material and which remained at the end of all treat-ments, (ii) ‘‘degraded’’ formulae are those that disappeared dueto solar radiation (absent from the light and low microbes treat-ment) or microbial decomposition (absent from the dark controlor light and high microbes treatments), (iii) ‘‘photoproduced’’, de-fined as formulae which appeared after solar exposure in the lightand low microbes treatment. They could occur in the light and highmicrobes but were absent from the initial material or the dark con-trol, (iv) ‘‘photoproduced recalcitrant’’, photoproduced compoundsin both the low and high microbes treatments, and (v) ‘‘microbiallyproduced’’, formulae in the dark control, but not at the beginning ofthe experiment. These compounds could also occur in the light andhigh microbes treatment.

A molecular formula is considered to be present or absent in theabove groups if its intensity is higher or lower than the detectionlimit (S/N 4). The relative intensities of each formula were not con-sidered in further calculations. We also calculated the average for-mula for each of these different groups based on the averagecontribution of C, H, O, N, P and S atoms. To evaluate the molecularvariability within each group, formulae were displayed accordingto their H/C and O/C ratios in van Krevelen diagrams. Contributionsof heteroatoms (N, S and P) were highlighted in the diagrams.

To assess the aromaticity of a compound based on its molecularformula, we calculated the double bond equivalent minus oxygen(DBE-O) and the modified aromaticity index (AImod, Koch and Ditt-mar, 2006) using Eqs. (1) and (2), respectively.

DBE-O ¼ ð1þ 0:5ð2C�Hþ Nþ PÞÞ � O ð1Þ

AImod ¼ ð1þ C� 0:5O� S� 0:5HÞ=ðC� 0:5O� S� N� PÞ ð2Þ

DBE-O may underestimate the unsaturation degree in the car-bon skeleton by assuming that all oxygens in DOM are in C@Ogroups, but it is a good measure of double bonds without the effectof oxygen. Values for the AImod > 0.5 and P 0.67 are unambiguousindicators for the presence of aromatic and condensed aromaticcompounds, respectively (Koch and Dittmar, 2006).

Carboxylic-rich alicyclic molecules (CRAMs) are assumed to bea major refractory component of freshwater and marine DOM(Hertkorn et al., 2006; Lam et al., 2007). Molecules belonging toCRAMs were defined as DBE/C = 0.30–0.68, DBE/H = 0.20–0.95and DBE/O 0.77–1.75, according to Hertkorn et al. (2006). ForCRAM determination, DBEs were obtained from Eq. (1) withoutthe subtraction of oxygen atoms.

3. Results

3.1. Molecular modifications induced by exposure to light andmicrobes

We assigned molecular formulae to 75–85% of the FT-ICR-MSmasses in DOM extracted from each treatment, not considering

Page 4

P.E. Rossel et al. / Organic Geochemistry 60 (2013) 62–71 65

13C isotopic peaks. Altogether, ca. 2800 different formulae were as-signed, excluding the 13C isotopic peaks. The initial DOM shared70% of the formulae found in the Suwannee River fulvic acid refer-ence material, indicating its similarity to aqueous freshwater hu-mic substances (Fig. 2a and b).

The initial sample and the dark treatment shared 1082 formu-lae, representing 88.9% of the formulae identified in the initialmaterial (Fig. 2b and c). These two treatments had also a similarlylow contribution of heteroatoms (Fig. 3). The initial and dark treat-ments were also similar in the number of CH2 and CH2O homolo-gous series (ca. 150) detected in both samples (Table 1). Of theformulae in the initial DOM, 11% were not found in the dark treat-ment, indicating microbial decomposition of these compounds.These decomposed compounds had an average molecular compo-sition of C18.7H16.5O8.9 (Fig. 4a).

Comparison between the initial DOM and both solar exposuretreatments (low and high microbes) indicated decomposition ofhigh molecular mass compounds (Fig. 2d and e, Table 1). Solar radi-ation treatments decreased the number of CHO compounds but in-creased the number of compounds with heteroatoms (Fig. 3). Indetail, 58% of the formulae in the initial material, representedmainly by CHO components (average composition C22.2H21.7O8.8),were degraded by the combined effect of light and microbes. Thelow and high microbe treatments showed only slight differences:

Fig. 2. Mass spectra (m/z 150–650) and an exemplary nominal mass for all samples. (a)light and low microbes, (e) light and high microbes and (f) deep sea DOM. One exemplaryfull spectra. The assigned formulae are provided only once for the measured exeminterpretation of the references to color in this figure legend, the reader is referred to th

54% and 58% of the initial compounds were degraded, respectively(Fig. 4b).

DBE-O values ranged broadly from �5 to 9 in the initial DOMand the dark treatment, with a unimodal distribution (Fig. 5).The DBE-O indices for irradiated samples had a bimodal distribu-tion, with a general shift towards saturated compounds betweenDBE-O of �6 and 5, but also including new formulae with DBE-Obetween 11 and 14 (Fig. 5). The decrease in DBE-O values in the so-lar exposed treatments was also accompanied by a decrease in theAImod (average AImod values in Fig. 5). These changes were accom-panied by a decrease in the number of aromatic homologous seriesand an increase in oxygen (CH2O-homologous series) series in bothirradiation treatments (Table 1).

3.2. New molecular formulae in the solar exposed samples

In addition to the loss of formulae due to photochemical andmicrobial decomposition, new formulae not observed in the initialDOM or the dark control were detected in both irradiation treat-ments (Fig. 4c). They accounted for 74% of the 1835 formulae as-signed in the light and low microbes treatment and increased thenumber of formulae with respect to those detected in the initialDOM and the dark control (Table 1). These new formulae predom-inantly contained heteroatoms (N, S and P containing compounds

Suwannee River fulvic acid reference material, (b) initial DOM, (c) dark control, (d)nominal mass is provided for the area highlighted within the red dashed lines in the

plary nominal masses throughout the consecutive samples from (a) to (f). (Fore web version of this article.)

Page 5

Fig. 3. Heteroatom composition of the initial and the treated DOM. Samples are displayed with the same color code as Fig. 2. The letters indicate the elements found. (Forinterpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Table 1Molecular information for SPE-DOM based on FT-ICR-MS analysis.

Sample No. offormulae

Average molecular mass ± standarddeviation

No. of CH2

seriesaNo. of oxygenseriesa

No. of aromaticseriesb

Average molecularcomposition

Initial DOM 1217 398 ± 104 158 159 57 C20.1H20.9O8.5

Dark control 1261 412 ± 105 153 154 51 C20.9H22.7O8.7

Light and lowmicrobes

1835c 359 ± 89 316 317 21 C16.5H21.4N1.0O7.4

Light and highmicrobes

1323 358 ± 77 223 217 14 C16.3H20.6N0.9O7.9

a CH2– and oxygen (CH2O–) series are homologous series separated by CH2 groups and additionally by oxygen, respectively.b Number of aromatic series based on a AImod > 0.5, assuming that if the first element of the series is aromatic the whole series is aromatic.c 1358 formulae (74%) of the light and low microbes treatment were photoproduced (could occur in the light and high microbes but were absent in the initial DOM or the

dark control).

66 P.E. Rossel et al. / Organic Geochemistry 60 (2013) 62–71

up to 90% of the new formulae), especially N (average molecularcomposition, considering all photoproduced formulae, wasC16.4H20.5N1.4O7.3, in which P and S contributed < 0.2; Figs. 3 and4c).

N alone or in combination with other heteroatoms was associ-ated with 70% of the new formulae detected in the low microbestreatment. The most intense mass peaks attributed to these formu-lae were confirmed by the presence of their respective 13C isotopicpeaks in the spectrum. From the new formulae detected in the lowmicrobes treatment, 44% were not detected in the high microbestreatment (Fig. 4d). The loss of these formulae, in the presence ofenhanced microbial activity, indicates biodegradation generatedby photochemical transformation. The average molecular composi-tion of the compounds that appeared in the low microbes treat-ment, but were not detected in the high microbes treatment, wasC17.4H23.3N1.4O6.8 with an average AImod of 0.45 and a standarddeviation of 0.23 (Fig. 3d).

A group of new compounds characterized by higher DBE-O val-ues (> 10) was also detected in both the low and high microbestreatments (Fig. 5). This type of aromatic compound (average AImod

0.5 with standard deviation of 0.05; Fig. 5) was distributed in 26CH2 homologous series in the low microbes treatment, 14 of whichwere common to the light and high microbes treatment. The com-parison of irradiated high and low microbes treatments indicatesthat compounds with DBE-O > 10 were partially degraded by mi-crobes. These aromatic compounds contained N, P or both.

3.3. Molecular composition of experimental DOM compared with deepsea DOM

Although the initial DOM shared similarities with SuwanneeRiver fulvic acid, the extensive photochemical and microbial trans-formation during the treatments altered its composition awayfrom the characteristics of freshwater humic substances and to-wards highly reworked marine DOM (Figs. 2, 4 and 6). Comparisonof the molecular mixtures resulting from this experiment and deepsea DOM indicated that the removal of compounds from our initialDOM, which were not present in the deep sea DOM material, wasmainly due to the effect of light (Fig. 4e). Aromatic compounds,which were preferentially removed during solar exposure, werealso absent from the deep sea DOM mixture. The initial DOM usedin this experiment contained between 493 and 571 formulae incommon with deep sea DOM, depending whether or not the datawere filtered via S/N 4, respectively (Table 2). From these commonformulae, only between 57% and 58% were recalcitrant (survivedall treatments), representing up to 18% of the deep sea DOM for-mulae (Table 2, Fig. 6). These recalcitrant formulae were character-ized by an average composition of C17.2H21.6O8.0 and average AImod

of 0.26 with standard deviation of 0.13 and DBE-O of �0.58 withstandard deviation of 1.9.

Of the formulae produced exclusively after exposure to solarradiation, 22–27% were the same as those in the deep sea DOM(Fig. 6). Not all these formulae met the criteria of photoproduced

Page 6

Fig. 4. Van Krevelen diagrams, including all formulae, for sample comparison. (a) Suwannee River fulvic acid reference material, initial DOM and dark control, (b) initial DOM,formulae not detected in the light and low microbes treatment (degraded in the light and low microbes) and additionally not detected in the light and high microbestreatment (degraded in the light and high microbes), (c) photoproduced formulae (present in the light and low microbes but absent from the initial and dark control material),highlighting the high contribution of those with heteroatoms (particularly N), (d) photoproduced formulae according to their bioavailability. ‘‘Photoproduced and degradedby microbes’’ represent the formulae in the light and low microbes but degraded in the light and high microbes treatment. ‘‘Photoproduced and recalcitrant’’ represents thephotoproduced formulae in the light and low microbes not degraded in the high microbes-treatment (present in both light treatments), (e) initial DOM and photodegradedformulae in comparison with deep sea DOM and (f) deep sea DOM (non-filtered via S/N) highlighting identical formulae with Juncus leachate. Initial DOM and recalcitrant(present in all treatments), photoproduced and recalcitrant (present in both solar exposed treatments) and microbially produced. The blue dashed line in (f) shows the areawith the predominance of CRAMs (see CRAM definition in Section 2). (For interpretation of the references to color in this figure legend, the reader is referred to the webversion of this article.)

P.E. Rossel et al. / Organic Geochemistry 60 (2013) 62–71 67

Page 7

Fig. 5. Double bond equivalents minus oxygen (DBE-O) and modified aromaticity index (AImod) for all samples. DBE-O and AImod values were obtained based on Eqs. (1) and(2), respectively. Mean values for AImod and O content for three different regions along the DBE-O axis (indicated by black lines) are presented in the same color as thesamples. Tentative molecular structures discussed in this work for some observed molecular formulae are presented as examples for DBE-O values.

68 P.E. Rossel et al. / Organic Geochemistry 60 (2013) 62–71

and ‘‘recalcitrant’’, i.e. not all were present in the low and high mi-crobes treatments. The photoproduced compounds present in boththe low and high microbial treatments represented about 18% ofthe deep sea DOM (Fig. 4f, Table 2) and were characterized by anaverage AImod of 0.29 with standard deviation of 0.14, DBE-O of�0.45 with standard deviation of 2.09 and an average compositionof C16.3H19.8N1.2O8.5.

In contrast to photoproduced compounds, formulae producedexclusively by microbes (dark control and light and high microbes)represented only 8% of those in the deep sea DOM (Fig. 4f, Table 2).Thus, considering the total number of formulae at the end of theexperiment, which were also present in the deep sea DOM, ca.44% of the deep sea DOM formulae (18% from the initial material,18% photoproduced and recalcitrant and 8% microbially derived)could be simulated in the experiment.

4. Discussion

Loss of high mass compounds in both solar exposure treatmentsis in agreement with previous reports (Opsahl and Benner, 1998;Dittmar et al., 2007; Helms et al., 2008; Stubbins et al., 2010). Pref-erential photodegradation of more unsaturated compounds, withhigher DBE-O values, observed in the experiment is also in agree-ment with FT-ICR-MS studies evaluating photolysis of terrigenousDOM on a molecular level for timescales of < 57 days (Kujawinskiet al., 2004; Gonsior et al., 2009; Stubbins et al., 2010). The shiftfrom positive to negative DBE-O values (higher number of oxygensthan double bonds in the molecule), was likely due to an increasein carboxylic and hydroxyl groups and/or decomposition of

aromatic moieties rich in C@C bonds in the compounds in the solarexposure samples, in agreement with previous reports (Kujawinskiet al., 2004; Gonsior et al., 2009; Vähätalo, 2009; Stubbins et al.,2010).

The decrease in the number of CHO compounds, accompaniedby the increase in heteroatom-containing compounds in our solarexposure treatments, has been observed in a transect from a riverto the ocean (Sleighter and Hatcher, 2008). In the transect from thelower Chesapeake Bay towards offshore, molecular diversity andthe compounds with heteroatoms increased in association with adecrease in CHO. Similarly, Aarnos et al. (2012) reported preferen-tial degradation of DOC over dissolved organic nitrogen (DON) in acoastal gradient.

Contrary to previous studies (Gonsior et al., 2009; Stubbinset al., 2010), we detected, however, a larger number of photopro-duced compounds. Among these were also aromatic and con-densed aromatics compounds, in agreement with their presencein a photodegradation experiment performed by Stubbins et al.(2010). These authors indicated the presence of five condensedaromatic compounds after DOM from Congo River was irradiated.The compounds were shown to contain N and some of them P, inagreement with the composition of the aromatic photoproductsin our experiment. Stubbins et al. (2010) had already suggestedthat, based on their aromatic structures, it is likely that they maybe resistant to biodegradation, although microbial decompositionwas not evaluated in their experiment. Nevertheless, the idea issupported by the only partial degradation of these compoundsafter more than a year of microbial degradation in our high mi-crobes treatment. Gonsior et al. (2009) also reported photopro-duced molecules from light exposed riverine DOM, although

Page 8

Fig. 6. Venn diagrams comparing deep sea DOM, initial DOM and produced and recalcitrant organic mixtures derived from the experiment. Formulae in the initial DOM arepresented in green. Formulae produced by microbes (present in the dark control and light and high microbes treatments, but absent from the initial DOM) and by solarradiation (present in the light and low microbes, but absent from the initial and dark control treatments) are displayed in red and orange circles, respectively. Of theseproduced formulae, those common with deep sea DOM are indicated in white color, highlighting in the dashed area and gray text the recalcitrant formulae (surviving at theend of the experiment). Recalcitrant formulae in the initial DOM or produced during the 1.25 yr degradation (blue circle) correspond to 44% of the deep sea DOM molecularfingerprint based on qualitative molecular data obtained from FT-ICR-MS. Notice that several other produced or initially recalcitrant formulae were not present in the deepsea DOM. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

P.E. Rossel et al. / Organic Geochemistry 60 (2013) 62–71 69

none were aromatic or condensed aromatics, in agreement withthe preferential loss of aromatics reported in previous studies(Tremblay et al., 2007). Although in our experiment and the Stub-bins et al. (2010) work, photoproduced aromatic molecules (in thisstudy with DBE-O > 10 and AImod > 0.5) were detected and seem to

Table 2Comparison of formulae in initial DOM and recalcitrant mixtures from the experiment wi

Group of compounds No. of formulae non-degraded (%)

Nocom

Deep ocean DOMa non-filtered and (filtered) byS/N

– 21

Initial DOM – 57Initial DOM and recalcitrantb 378 (31%) 32Microbially producedc – 14Photoproduced and recalcitrantd 758 (41%) 32Total of recalcitrant formulae after 1.25 yr

degradation– 79

a For comparison with the experiment samples, the deep sea DOM was filtered and nonparentheses, respectively. The difference between the filtered and non-filtered deep smolecular formulae without considering S/N, therefore increasing the number of detect

b The initial DOM and recalcitrant group of compounds are represented by all formulc The microbially produced group of compounds is represented by the formulae produ

These formulae were not present in the initial DOM.d The photoproduced and recalcitrant group of compounds is represented by formula

condition or dark control. They are defined here as recalcitrant because they survived the The group of compounds that remained at the end of the experiment, independent

present in the initial DOM and survived the treatments, represented between 37% to 44

be resistant to microbial degradation, these compounds seem notto be preserved during ocean residence time since we did not de-tect them in our deep sea DOM.

The production of aromatic and condensed aromaticcompounds with heteroatoms during solar exposure is a surprising

th deep ocean DOM.a

. of formulae inmon

No. of CH2-series incommon

No. of common carboxylic-rich alicyclic molecules(CRAMs)

Series Formulae

00 (1580) 294 (240) 193 (174) 1457 (1169)

1 (493) 79 (75) 58 (55) 435 (388)3 (288) 55 (53) 39 (38) 232 (210)2 (126) 60 (52) 43 (39) 92 (85)9 (290) 103 (88) 71 (64) 254 (230)4 (704)e 218 (193) 153 (141) 578 (525)

-filtered via S/N. Values derived from the comparison are provided with and withoutea DOM data is that the latter was only filtered via other criteria used to assigned peaks in the spectra.ae that survived all treatments. Thus molecular formulae present in all treatments.ced by microbes in the dark control and in the light and high microbes treatments.

e produced in the light and low microbes treatment but absent from the originale microbial degradation in the light and high microbes treatment.

if they were produced during the experiment and survived the treatments or were% of the molecular formulae present in deep sea DOM.

Page 9

70 P.E. Rossel et al. / Organic Geochemistry 60 (2013) 62–71

result. Possible explanations include: (i) removal of aliphatic sidechains or hydroxyl groups from the compounds increased AImod,making the compounds apparently more aromatic, (ii) their signalwas very low compared with the predominant CHO compounds inthe initial DOM and their relative abundance increased during theexperiment, and (iii) the compounds were only ionized after theywere partly decomposed by sunlight. Introduction of carboxylfunctional groups during exposure to sunlight could have in-creased ionization efficiency in ESI.

Preferential microbial decomposition of N compounds after so-lar exposure of our initial DOM suggests that solar radiation canproduce N-containing, bioavailable molecules from a relativelyN-poor source. Thus, photolysis of wetland-derived DOM may bean important mechanism for producing bioavailable N. The photo-chemical production of NHþ4 and amino acids from biologically re-calcitrant DOM has been reported (Bushaw-Newton and Moran,1999; Vähätalo et al., 2003; Vähätalo and Zepp, 2005; Vähätaloand Järvinen, 2007), but the production of such a molecularly di-verse group of N-containing compounds in our experiment hasnot been reported. Contrary to previous reports, the molecularmass (average 347 g mol�1) and C/N ratio (average 7.5) of theseN compounds later consumed by microbes were high in our study.Secondary introduction of NHþ4 that could have been photochemi-cally produced from a few abundant organic compounds into theDON pool (Vähätalo et al., 2003) could be an explanation for ourobservation. Not all these compounds were bioavailable on a yeartimescale. This is in general agreement with studies that reportedthat photodegradation of DOM produces both bioavailable (Kieberet al., 1989, 1990; Wetzel et al., 1995) and non-bioavailable com-pounds (Benner and Biddanda, 1998; Tranvik and Kokalj, 1998).Studies of photodegradation of marine and freshwater derivedDOM also indicated that light exposure of freshly produced algalDOM caused a decrease in biological degradation (Benner andBiddanda, 1998; Tranvik and Kokalj, 1998), while in terrigenousDOM this process generated biologically available compounds(Moran et al., 2000). Therefore, positive or negative effects in theproduction of bioavailable compounds have been related to thetype of material exposed to solar radiation. Our results suggest thatphotochemical transformation of wetland-derived DOM includescomplex reactions and can result in production of compounds witha range of reactivity, especially for N-containing compounds. Inter-estingly, the molecular diversity (i.e. number of formulae) in-creased during solar radiation in our experiments, especially forN compounds. This trend is surprising, because one would expectreduced molecular diversity after such an extensive DOC degrada-tion by light and microbes. It is likely that this aspect has beenoverlooked in previous photodegradation studies in which analyt-ical techniques other than FT-ICR-MS were used.

Identical formulae in the experimental and deep sea DOM do notnecessarily imply that the same structural isomers were present.Nevertheless, it is interesting that photochemical and microbialdegradation shaped a vascular plant leachate in a way that it be-came similar to deep sea DOM on a very broad and extensive molec-ular formula level. Similar photochemical and microbialdecomposition of freshwater DOM in natural waters would changeits chemical composition in such a way that it would resemble mar-ine DOM, and its source in the open ocean may not be recognizableusing traditional biomarker techniques. Along the same lines,Sleighter and Hatcher (2008) suggested that lignin phenols couldbe structurally modified during photochemical reactions in such away that lignin would not be recognized anymore with the CuO oxi-dation method. Additionally, the stable carbon isotopic approach,commonly used to identify carbon sources in the ocean, may alsobe confounded by isotopic shifts resulting from photodegradation(Vähätalo and Wetzel, 2008). Similarly, the optical properties of ter-restrial carbon and lignin normalized yields are also changed to-

ward marine values during photodegradation (Opsahl and Benner,1998; Vähätalo et al., 1999; Moran et al., 2000; Osburn et al.,2001; Spencer et al., 2009). These observations suggest that longterm photochemical and microbial reactions of DOM that occur inall surface waters can result in similar recalcitrant DOM mixtures.

Our results support our initial hypothesis that processes in theenvironment shape DOM in a way that the resulting recalcitrantDOM mixtures share common molecular features and become sim-ilar, despite sharp differences in source material. This is consistentwith the apparently ubiquitous presence of CRAM compounds infreshwater and marine systems. First described by Hertkorn et al.(2006) for marine DOM from the Pacific Ocean, CRAMs also seemto be a dominant component in freshwater DOM from Lake Ontario(Lam et al., 2007). The occurrence of CRAMs in freshwater and mar-ine DOM is possibly due to the similarity in biogeochemical pro-cesses in both environments. This is in agreement with theoverlap between our deep sea DOM and the recalcitrant mixturesderived from our experiment, which were predominantly repre-sented by CRAMs.

Although our initial hypothesis was largely supported by ourstudy, major molecular differences remained between the marineDOM and the DOM produced in our experiment. Ca. 55% of thedeep sea DOM formulae were not produced in our experiment.Limitations of our experiment are the use of unique samples, theJuncus leachate and the deep sea DOM from the North PacificOcean. At the moment it is difficult to say to which extent our re-sults may vary if a variety of vascular plants leachates and deep seaDOM collected from different locations were considered. Althougha single sample from deep sea was evaluated, the stability and longturnover times of deep sea DOM could imply that this unique sam-ple may well represent the typical DOM composition in the deepsea. However, it is hard to give a quantitative measure of its repre-sentation of whole deep sea DOM.

The representativeness of a single species of leachate (Juncus) ofthe typical vascular plant derived DOM entering from freshwater tothe ocean may only be possible to address if this material were di-rectly compared with the DOM composition from large rivers. Nev-ertheless, the similarity of the Juncus leachate with the SuwanneeRiver fulvic acid indicates that this leachate is similar to fulvic acidtype material, often the dominating class of compounds in freshwa-ter DOM. Although fulvic acid type DOM generally dominates theDOC concentration in many freshwaters transported to the ocean,also non-humic material is transported to the ocean. Additionalprocesses not included in our experiment must be also consideredin the transformation of DOM. Aspects like microbial communitycomposition and abiotic process other than photochemistry, suchas thermal alteration or interaction with particles, may also play amajor role in the production of recalcitrant DOM in the deep ocean.

5. Conclusions

This study evaluated the effect of light and microbial decomposi-tion over an initial DOM material represented by a leachate of a vas-cular plant, J. effusus. Together with a general decrease in aromaticcompounds due to photochemical reactions of the initial DOM, wealso detected the unexpected production of aromatic compoundsdue to solar radiation. Despite intensive decomposition of the initialDOM, molecular diversity increased during solar exposure, espe-cially for N-containing compounds. Regardless of the rather specificorigin of the initial DOM, a single species of wetland vascular plant,the study allowed us to provide preliminary evidence that photo-chemical and microbial alteration over the 1.25 yr timescale pro-duce molecular formulae identical to 44% of the deep sea DOMformulae. Most of the compounds were represented by CRAMs.Our results illustrate that intensively degraded DOM can share com-mon molecular features independent of its origin.

Page 10

P.E. Rossel et al. / Organic Geochemistry 60 (2013) 62–71 71

Acknowledgments

We thank J. Fuchser for the measurement of one of the powdersamples using FT-ICR-MS at Bruker Daltonics (Bremen) and K.Klaproth for checking missing formulae. We also thank J. Letelierfor helping with MATLAB programming and colleagues from theMarine Geochemistry group, J. Niggemann and M. Friebe, for pre-paring the deep ocean DOM at the National Energy Laboratory ofHawaian Authorities (NELHA) station. We thank the two anony-mous reviewers for constructive comments.

Associate Editor—E.A. Canuel

References

Amon, R.M.W., Benner, R., 1996. Photochemical and microbial consumption ofdissolved organic carbon and dissolved oxygen in the Amazon River system.Geochimica et Cosmochimica Acta 60, 1783–1792.

Aarnos, H., Ylöstalo, P., Vähätalo, A.V., 2012. Seasonal phototransformation ofdissolved organic matter to ammonium, dissolved inorganic carbon, and labilesubstrates supporting bacterial biomass across the Baltic Sea. Journal ofGeophysical Research 117. http://dx.doi.org/10.1029/2010JG001633.

Benner, R., Biddanda, B., 1998. Photochemical transformations of surface and deepmarine dissolved organic matter: effects on bacterial growth. Limnology andOceanography 43, 1373–1378.

Blough, N.V., Del Vecchio, R., 2002. Chromophoric DOM in the coastal environment.In: Hansell, D.A., Carlson, C.A. (Eds.), Biogeochemistry of Marine DissolvedOrganic Matter. Academic Press, Amsterdam, pp. 509–540.

Bushaw-Newton, K.L., Moran, M.A., 1999. Photochemical formation of biologicallyavailable nitrogen from dissolved humic substances in coastal marine systems.Aquatic Microbial Ecology 18, 285–292.

Coble, P.G., 2007. Marine optical biogeochemistry: the chemistry of ocean color.Chemical Reviews 107, 402–418.

D’Andrilli, J., Dittmar, T., Koch, B.P., Purcell, J.M., Marshall, A.G., Cooper, W.T., 2010.Comprehensive characterization of marine dissolved organic matter by Fouriertransform ion cyclotron resonance mass spectrometry with electrospray andatmospheric pressure photoionization. Rapid Communications in MassSpectrometry 24, 643–650.

Dittmar, T., Koch, B.P., 2006. Thermogenic organic matter dissolved in the abyssalocean. Marine Chemistry 102, 208–217.

Dittmar, T., Whitehead, K., Minor, C.E., Koch, B.P., 2007. Tracing terrigenousdissolved organic matter and its photochemical decay in the ocean by usingliquid chromatography/mass spectrometry. Marine Chemistry 107, 378–387.

Dittmar, T., Koch, B.P., Hertkon, N., Kattner, G., 2008. A simple and efficient methodfor solid phase extraction of dissolved organic matter (SPE-DOM) fromseawater. Limnology and Oceanography: Methods 6, 230–235.

Gonsior, M., Peake, B.M., Cooper, W.T., Podgorski, D., D’Andrilli, J., Cooper, W.J.,2009. Photochemically induced changes in dissolved organic matter identifiedby ultrahigh resolution Fourier transform ion cyclotron resonance massspectrometry. Environmental Science and Technology 43, 698–703.

Harvey, G.R., Boran, D.A., Chesal, L.A., Tokar, J.M., 1983. The structure of marinefulvic and humic acids. Marine Chemistry 12, 119–132.

Helms, J.R., Stubbins, A., Richtie, J.D., Minor, E.C., Kieber, D.J., Mopper, K., 2008.Absorption spectral slopes and slope ratios as indicators of molecular weight,source, and photobleaching of chromophoric dissolved organic matter.Limnology and Oceanography 53, 955–969.

Hedges, J.I., Keil, R.G., Benner, R., 1997. What happens to terrestrial organic matterin the ocean? Organic Geochemistry 27, 195–212.

Hernes, P.J., Benner, R., 2003. Photochemical and microbial degradation of dissolvedlignin phenols: implications for the fate of terrigenous dissolved organic matterin marine environments. Journal of Geophysical Research 108, 3291. http://dx.doi.org/10.1029/2002JC001421.

Hernes, P.J., Benner, R., 2006. Terrigenous organic matter sources and reactivity inthe North Atlantic Ocean and a comparison to the Arctic and Pacific oceans.Marine Chemistry 100, 66–79.

Hertkorn, N., Benner, R., Frommberger, M., Schmitt-Kopplin, P., Witt, M., Kaiser, K.,Kettrup, A., Hedges, J.I., 2006. Characterization of a major refractory componentof marine dissolved organic matter. Geochimica et Cosmochimica Acta 70,2990–3010.

Kieber, D.J., McDaniel, J., Mopper, K., 1989. Photochemical source of biologicalsubstrates in sea water: implications for carbon cycling. Nature 341, 637–639.

Kieber, R.J., Zhou, X., Mopper, K., 1990. Formation of carbonyl compounds from UV-Induced photodegradation of humic substances in natural waters: fate ofriverine carbon in the sea. Limnology and Oceanography 35, 1503–1515.

Koch, B.P., Dittmar, T., 2006. From mass to structure: an aromaticity index for high-resolution mass data of natural organic matter. Rapid Communications in MassSpectrometry 20, 926–932.

Koch, B.P., Witt, M., Engbrodt, R., Dittmar, T., Kattner, G., 2005. Molecular formulaeof marine and terrigenous dissolved organic matter detected by electrosprayionization Fourier transform ion cyclotron resonance mass spectrometry.Geochimica et Cosmochimica Acta 69, 3299–3308.

Koch, B.P., Ludwichowski, K.-U., Kattner, G., Dittmar, T., Witt, M., 2008. Advancedcharacterization of marine dissolved organic matter by combining reversed-phase liquid chromatography and FT-ICR-MS. Marine Chemistry 111, 233–241.

Kouassi, A.M., Zika, R.G., 1992. Light induced destruction of the absorbency propertyof dissolved organic matter in seawater. Toxicological and EnvironmentalChemistry 35, 195–211.

Kujawinski, E.B., Behn, M.D., 2006. Automated analysis of electrospray ionizationFourier transform ion cyclotron resonance mass spectra of natural organicmatter. Analytical Chemistry 78, 4363–4373.

Kujawinski, E.B., Del Vecchio, R., Blough, N.V., Klein, G.C., Marshall, A.G., 2004.Probing molecular-level transformations of dissolved organic matter: insightson photochemical degradation and protozoan modification of DOM fromelectrospray ionization Fourier transform ion cyclotron resonance massspectrometry. Marine Chemistry 92, 23–37.

Lam, B., Baer, A., Alaee, M., Lefebvre, B., Moser, A., Williams, A., Simpson, A.J., 2007.Major structural components in freshwater dissolved organic matter.Environmental Science and Technology 41, 8240–8247.

Lou, T., Xie, H., 2006. Photochemical alteration of the molecular weight of dissolvedorganic matter. Chemosphere 65, 2333–2342.

Miller, W.J., Moran, M.A., 1997. Interaction of photochemical and microbialprocesses in the degradation of refractory dissolved organic matter from acoastal marine environment. Limnology and Oceanography 4, 1317–1324.

Mopper, K., Zhou, X., Kieber, R.J., Kieber, D.J., Sikorski, R.J., Jones, R.D., 1991.Photochemical degradation of dissolved organic carbon and its impact on theoceanic carbon cycle. Nature 353, 60–62.

Moran Jr., M.A., Sheldon, W.M., Zepp, R.G., 2000. Carbon loss and optical propertychanges during long-term photochemical and microbial degradation ofestuarine dissolved organic matter. Limnology and Oceanography 45, 1254–1264.

Obernosterer, I., Reitner, B., Herndl, G.J., 1999. Contrasting effects of solar radiationon dissolved organic matter and its bioavailability to marine bacterioplankton.Limnology and Oceanography 44, 1645–1654.

Opsahl, S., Benner, R., 1998. Photochemical reactivity of dissolved lignin in river andocean waters. Limnology and Oceanography 43, 1297–1304.

Osburn, C.L., Morris, D.P., Thorn, K.A., Moeller, R.E., 2001. Chemical and opticalchanges in freshwater dissolved organic matter exposed to solar radiation.Biogeochemistry 54, 251–278.

Sleighter, R., Hatcher, P.G., 2008. Molecular characterization of dissolved organicmatter (DOM) along a river to ocean transect of the lower Chesapeake Bay byultrahigh resolution electrospray ionization Fourier transform ion cyclotronresonance mass spectrometry. Marine Chemistry 110, 140–152.

Spencer, R.G.M., Stubbins, A., Hernes, P.J., Baker, A., Mopper, K., Aufdenkampe, A.K.,Dyda, R.Y., Mwamba, V.L., Mangangu, A.M., Wabakanghanzi, J.N., Six, J., 2009.Photochemical degradation of dissolved organic matter and dissolved ligninphenols from the Congo River. Journal of Geophysical Research 114. http://dx.doi.org/10.1029/2009JG000 968.

Stubbins, A., Spencer, R.G.M., Chen, H., Hatcher, P.G., Mopper, K., Hernes, P.J.,Mwamba, V.L., Mangangu, A.M., Wabakanghanzi, J.N., Six, J., 2010. Illuminateddarkness: molecular signatures of Congo River dissolved organic matter and itsphotochemical alteration as revealed by ultrahigh precision mass spectrometry.Limnology and Oceanography 55, 1467–1477.

Thurman, E.M., 1985. Organic Geochemistry of Natural Waters. Kluwer AcademicPublishers Group, The Netherlands.

Tranvik, L., Kokalj, K., 1998. Decreased biodegradability of algal DOC due tointeractive effects of UV radiation and humic matter. Aquatic Microbial Ecology14, 301–307.

Tremblay, L.B., Dittmar, T., Marshall, A.G., Cooper, W.J., Cooper, W.T., 2007.Molecular characterization of dissolved organic matter in a North Brazilianmangrove porewater and mangrove-fringed estuaries by ultrahigh resolutionFourier transform-ion cyclotron resonance mass spectrometry and excitation/emission spectroscopy. Marine Chemistry 105, 15–29.

Vähätalo, A.V., 2009. Light, photolytic reactivity and chemical products. In: Likens,G.E. (Ed.), Encyclopedia of Inland Waters. Elsevier, Oxford, pp. 761–773.

Vähätalo, A.V., Järvinen, M., 2007. Photochemically produced bioavailable nitrogenfrom biological recalcitrant dissolved organic matter stimulates production of anitrogen-limited microbial food web in the Baltic Sea. Limnology andOceanography 52, 132–143.

Vähätalo, A.V., Wetzel, R.G., 2008. Long-term photochemical and microbialdecomposition of wetland-derived dissolved organic matter with alteration of13C:12C mass ratio. Limnology and Oceanography 53, 1387–1392.

Vähätalo, A.V., Zepp, R.G., 2005. Photochemical mineralization of dissolved organicnitrogen to ammonium in the Baltic Sea. Environmental Science andTechnology 39, 6985–6992.

Vähätalo, A.V., Salonen, K., Münster, U., Järvinen, M., Wetzel, R.G., 2003.Photochemical transformation of allochthonous organic matter providesbioavailable nutrients in a humic lake. Archiv für Hydrobiologie 156, 287–314.

Vähätalo, A.V., Salonen, K., Salonen, M.S., Hatakka, A., 1999. Photochemicalmineralization if synthetic lignin in lake water indicates enhanced turnoverof aromatic organic matter under solar radiation. Biodegradation 10, 415–420.

Wetzel, R.G., 1992. Gradient dominated ecosystems: sources and regulatoryfunctions of dissolved organic matter in freshwater ecosystems.Hydrobiologia 229, 1369–1380.

Wetzel, R.G., Hatcher, P.G., Bianchi, T.S., 1995. Natural photolysis by ultravioletirradiance of recalcitrant dissolved organic matter to simple substrates for rapidbacterial metabolism. Limnology and Oceanography 40, 1369–1380.