Mercury methylation rates of biofilm and plankton microorganisms from a hydroelectric reservoir in French Guiana

Science of the Total Environment 408 (2010) 1338–1348

Contents lists available at ScienceDirect

Science of the Total Environment j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / s c i t o t e n v

Mercury methylation rates of biofilm and plankton microorganisms from a hydroelectric reservoir in French Guiana L. Huguet a, S. Castelle b, J. Schäfer b, G. Blanc b, R. Maury-Brachet b, C. Reynouard c, F. Jorand a,⁎ a b c

Laboratoire de Chimie Physique et Microbiologie pour l'Environnement (LCPME), UMR 7564, CNRS-Nancy-Université, 405, rue de Vandœuvre, 54600 Villers-les-Nancy, France Université de Bordeaux UMR 5805 EPOC, Avenue des Facultés, 33405 Talence, France Laboratoire de recherche en environnement barrage de Petit-Saut (HYDRECO), BP 823, 97388 Kourou, French Guiana

a r t i c l e

i n f o

Article history: Received 5 May 2009 Received in revised form 18 October 2009 Accepted 20 October 2009 Available online 14 November 2009 Keywords: Plankton microorganisms Biofilm Isotope dilution Mercury methylation Tropical reservoir Estuary

a b s t r a c t The Petit-Saut ecosystem is a hydroelectric reservoir covering 365 km2 of flooded tropical forest. This reservoir and the Sinnamary Estuary downstream of the dam are subject to significant mercury methylation. The mercury methylation potential of plankton and biofilm microorganisms/components from different depths in the anoxic reservoir water column and from two different sites along the estuary was assessed. For this, reservoir water and samples of epiphytic biofilms from the trunk of a submerged tree in the anoxic water column and from submerged branches in the estuary were batch-incubated from 1 h to 3 months with a nominal 1000 ng/L spike of Hg(II) chloride enriched in 199Hg. Methylation rates were determined for different reservoir and estuarine communities under natural nutrient (reservoir water, estuary freshwater) and artificial nutrient (culture medium) conditions. Methylation rates in reservoir water incubations were the highest with plankton microorganisms sampled at − 9.5 m depth (0.5%/d) without addition of biofilm components. Mercury methylation rates of incubated biofilm components were strongly enhanced by nutrient addition. The results suggested that plankton microorganisms strongly contribute to the total Hg methylation in the Petit-Saut reservoir and in the Sinnamary Estuary. Moreover, specific methylation efficiencies (%Me199Hgnet/cell) suggested that plankton microorganisms could be more efficient methylating actors than biofilm consortia and that their methylation efficiency may be reduced in the presence of biofilm components. Extrapolation to the reservoir scale of the experimentally determined preliminary methylation efficiencies suggested that plankton microorganisms in the anoxic water column could produce up to 27 mol MeHg/year. Taking into account that (i) demethylation probably occurs in the reservoir and (ii) that the presence of biofilm components may limit the methylation efficiency of plankton microorganisms, this result is highly consistent with the annual net MeHg production estimated from mass balances (8.1 mol MeHg/year, Muresan et al., 2008a). © 2009 Elsevier B.V. All rights reserved.

1. Introduction Mercury (Hg) is considered a major pollutant since exposures to this heavy metal have environmental and human health consequences from femto- to nanomolar concentrations (Boening, 2000; Gochfeld, 2003; Zahir et al., 2005). Its environmental cycle is highly dynamic and complex, involving reduction/oxidation or methylation/demethylation with biotic or abiotic pathways (Rocha et al., 2000; Barkay and Wagner-Döbler, 2005; Garcia et al., 2005; Celo et al., 2006; Desrosiers et al., 2006; Poulain et al., 2007). Monomethylmercury CH3Hg+ (MeHg) is easily bioaccumulated and biomagnified in aquatic food chains, mainly due to fish consumption, and has serious negative effects on human health (Fréry et al., 2001; Baeyens et al., 2003; Li et al., 2008). ⁎ Corresponding author. LCPME UMR 7564 CNRS-Nancy-University, 405 rue de Vandœuvre, F-54600 VILLERS-LÈS-NANCY, France. Tel.: +33 3 83 68 52 48; fax: +33 3 83 27 54 44. E-mail address: [email protected] (F. Jorand). 0048-9697/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2009.10.058

In French Guiana, Hg levels in hair samples of more than half of a native Amerindian people were higher than the maximum value recommended by the World Health Organization (10 μg g− 1), i.e. at a level that may cause neurological disorder, especially for children (Fréry et al., 2001). In Amazonia, the main mercury sources originate from goldmining practices, including erosion of soils naturally enriched mercury (Boudou et al., 2005; Roulet et al., 1999). Great lakes and many hydroelectric reservoirs are preferential sites of MeHg production because of stratification and development of anoxic hypolimnion (Rosenberg et al., 1997; Pak and Bartha, 1998; Bellanger et al., 2004; Eckley and Hintelmann, 2006). The Petit-Saut Reservoir, on the Sinnamary River in French Guiana (South America), is a model system for studies on Hg behavior in tropical reservoirs (Coquery et al., 2003; Boudou et al., 2005; Peretyazhko et al., 2005; Peretyazhko et al., 2006; Muresan et al., 2008a). It was constructed on ancient gold-mining sites and still receives large amounts of Hg from its watershed due to gold mining (increasing soil erosion) and -extraction by amalgamation. About 15 kgHgy− 1 are provided to the reservoir by the influent rivers,

L. Huguet et al. / Science of the Total Environment 408 (2010) 1338–1348

and especially by the Leblond River affected by gold-mining activities (Coquery et al., 2003). The filling of this reservoir has started in January 1994 resulting in the immersion of 365 km² of tropical rainforest and naturally Hg-rich soils. Biomass immersion has led to strong disturbances of biogeochemical cycles in both the reservoir and the Sinnamary Estuary (Richard et al., 1997; De Mérona et al., 2005). Mercury methylation in the anoxic waters of the reservoir and in the Sinnamary Estuary downstream of the dam induces high Hg concentrations in fish representing a health risk for regional human populations depending on fish-diet (Durrieu et al., 2005; Dominique et al., 2007; Muresan et al., 2007, 2008b). In 2004/2005, the mean Hg concentrations in dorsal skeletal muscle of Curimata cyprinoides, a detritivorous/benthivorous fish species, were 10-fold higher in fish sampled downstream (3400 ± 240 ng g− 1) compared to fish sampled in the reservoir (320 ± 50 ng g− 1) and the MeHg fraction was close to 100% (Dominique et al., 2007). This fish species feeds by grazing immersed surfaces (rocks, tree branches and trunks, roots, etc.), recovered by an abundant slime assumed to be a biofilm. Biofilms are complex associations of multi-species microorganisms embedded in a polymeric matrix and adhering to a surface, usually in aqueous environments. This is assumed to be the preferential microbial way of life in the environment (De Beer and Stoodley, 2006). Previous work on biofilms sampled on glass sheets in the oxic water column (0–6 m depth) of the reservoir and in the Sinnamary Estuary reported high MeHg contributions to total Hg content (28 ± 6% and 40 ± 4%, respectively), suggesting that these biofilms represent a MeHg entry into the local aquatic food chain (Dominique et al., 2007). Relatively high amounts of MeHg (up to 200 ngg− 1 dw) (Dominique, 2006), suggested that biofilms in the anoxic part of the reservoir play together with the benthic communities (Coquery et al., 2003; Peretyazhko et al., 2005) a major role in Hg methylation at the reservoir scale. This is consistent with the generally observed vertical distributions of MeHg concentrations, i.e. maxima around −9 m and near the benthic interface, suggesting that these are main depth ranges of water column MeHg mobilization and/or production (Muresan et al., 2008a). For these reasons and because several recent studies have suggested involvement of biofilms in Hg methylation/demethylation processes (Desrosiers et al., 2006; Lin and Jay, 2007), one may suppose the biofilm on the flooded dead tree trunks still standing in the reservoir lake to be a key location for Hg transformations in the Petit-Saut reservoir. Furthermore, plankton microorganisms are abundant in both

1339

the epi-and hypolimnion of the reservoir (≈5–14× 106 cells mL− 1; Dumestre et al., 2001) and there is evidence that plankton may contribute to Hg methylation in the oxic layer of the reservoir water column (Dominique et al., 2007). Nevertheless, to date there is no study on the role of plankton microorganisms and/or the biofilm attached to the dead tree trunks along the water column on Hg methylation in the anoxic water of the reservoir. The main objective of the present work was to assess the Hg methylation potential of plankton microorganisms and the influence of biofilm components (microorganisms, organic matter, minerals) from the anoxic part of the reservoir and from the Sinnamary Estuary. For this, we used stable Hg isotope spikes followed by incubation at different time-scales in reservoir and estuary water (natural nutrient conditions) and in culture medium (artificial nutrient conditions). The use of different nutrient conditions was explorative and aimed at covering a potentially wider range of bacterial activity and/or Hg methylation. The specific methylation efficiencies of plankton microorganisms in the presence or not of biofilm components were compared and their role in Hg cycling at the whole reservoir scale is discussed. 2. Materials and methods 2.1. Study area and sampling stations The Petit-Saut dam (5 °04′ North, 53 °03′ West) was constructed on the Sinnamary River in French Guiana ~ 60 km upstream from its mouth to the Atlantic Ocean (Fig. 1). The Sinnamary River has an average discharge of 250 m3 s− 1 with important seasonal and interannual variability (170–340 m3 s− 1) (Sissakian, 1997). The reservoir lake has a maximum depth of up to 35 m (depending on the season) and is ~ 80 km long, covering ~365 km2 of uncleared tropical forest. A rough estimate suggests that the tree density in the submerged area is ~ 25 trees 100 m− 2, i.e. ~91 million trees at the reservoir scale and an available apparent surface for biofilm of ~ 3 × 109 m2 (Dominique et al., 2007). The reservoir water body is highly stratified with an oxic epilimnion and an anoxic hypolimnion separated by a quasi permanent oxycline located around 5–7 m depth. Average water residence time in the reservoir is ~5 months (Richard et al., 1997). Water temperature reaches ~30 °C at the surface and decreases with depth to ~ 25 °C at ~30 m. Downstream from the dam, the Sinnamary Estuary has an

Fig. 1. Location of the Petit-Saut hydroelectric reservoir in French Guiana, location of the sampling sites in the reservoir (Bagatelle station) and in the Sinnamary Estuary (Passerelle and Venus stations), and schematic drawing illustrating the biofilm sampling procedure.

1340

L. Huguet et al. / Science of the Total Environment 408 (2010) 1338–1348

average depth of ~4 m and is ~50 m wide (Fig. 1). Water temperature is ~26 °C and water residence time was estimated between 24 and 48 h (Gosse and Grégoire, 1997). A 3 m-high two-falls aerating weir was constructed ~400 m downstream of the dam in order to increase dissolved oxygen concentration to 2 mg L− 1 (Gosse and Grégoire, 1997). The biofilm were collected in March 2008 at three sampling stations (Fig. 1): i) the Bagatelle station (N5°01′, W53°02′) located in the reservoir at ~5 km upstream from the dam. This site has been chosen because it is not under the direct influence of any influents or the river channel. Furthermore, at this site the water column depth is relatively high with submerged trees of 30–40 m high, allowing sampling throughout the maximum depth range of the water column; ii) the Passerelle station (N5°03′54″, W53°03′08″) ~ 800 m downstream of the dam, where the thick red biofilms were observed and iii) the Venus station (N5°11′04″, 52°58′42″) ~ 20 km downstream from the dam. The Venus station was chosen because of its central position in the Sinnamary Estuary. 2.2. Biofilm and water sampling and preconditioning Professional divers sampled biofilm components mixed with reservoir water from size-calibrated (1500 cm2) water-resistant tropical wooden planks (attached to a dead tree trunk 8 months before sampling) by aspiration with a 8 mm inner diam. polyethylene tubing and a 12 V peristaltic pump (Masterflex, easy-load, Bioblock 7518-12) set on a pirogue and working at 1 L min− 1 (Fig. 1). The untreated wooden planks were made of an abundant local forest tree called Wapa (Fam. Caesalpinioideae, Eperua sp.), representing around 20% of the total wood volume in the tropical forest of French Guiana (CTFT, 1984). The wooden planks served as a support for biofilm development as a part of a larger research strategy implying biofilm sampling during different seasons. Recovering biofilms from surface-calibrated planks rather than rotting tree trunks, we expected to (i) avoid/limit contamination by the support (e.g. debris from the trunk) and (ii) provide reliable estimates on the biofilm density per surface unit. In fact, randomly introduced amounts of trunk debris may induce additional sample heterogeneity (from one depth to another, from one campaign to another) in terms of biofilm purity, microorganism density and -activity, isotopic composition, etc. Furthermore, using planks, the precise sampling point on the submerged tree trunk is the same from one campaign to another. One may consider that the biofilms developing on a local wood support would be more representative of the trunk biofilm than the biofilms developing on the classically used glass supports. However, these biofilm assemblages were obviously younger than the raw trunk biofilm (i.e.: the biofilm growing on the natural substrata). The biofilm samples were taken along a vertical profile (− 20 m, −15 m and −6.5 m) from bottom to top (Fig. 1). Similarly, water samples were retrieved from different depths (− 24.5 m, − 15 m, and −9.5 m). At every depth, the sampling device was thoroughly rinsed with ambient water before sampling to prevent sampling artefacts (e.g. cross-contamination). All samples were directly transferred via the pump into 3 L polyethylene suction bags (Serres Oy, Kauhajoki, FIN) which were put in vacuum conditions in order to preserve samples from atmosphere and stored in an isotherm container (Chest 36QT, Coleman®, Wichita, KS). In the Sinnamary Estuary (Passerelle (P) and Venus (V) stations) abundant biofilm was collected from immersed branches (≤0.5 m depth) in 125 mL sterile polypropylene sample containers (A1764E, Bioblock) and stored in an isotherm box until return to the on-site lab. Estuary water was also collected at each station in acid cleaned and autoclaved 250 mL glass bottles. Removing biofilms from the dead tree trunk partially destroyed their macroscopic structure (aggregates were still observed). Accordingly, we incubated clusters of biofilm microorganisms and associated material, rather than structurally intact biofilms, in the literal meaning.

The term “biofilm-components” is therefore used for biofilm material in the experimental batches. All further sample handling was performed in glove bags under N2 atmosphere (X37-37, I2R, Cheltenham, PA). Samples were treated within few hours after collection. Biofilm components in bioprocess bags were decanted for 1 h and the major part of the supernatant water was separated and conserved. A final volume of 250 mL of biofilm components suspended in the remaining water was divided and conditioned for the incubations. All the labware (glass, Teflon) were acid cleaned (HCl + HNO3 analytical grade; 10% v/v, 72 h and HCl analytical grade; 10% v/v, 72 h), thoroughly rinsed with ultrapure water (MilliQ/MilliRO quality, Millipore®), dried under a laminar flow hood and autoclaved at 120 °C during 15 min prior to being in contact with the samples. 2.3. Main characteristics of the raw biofilm and the experimental cultures The biofilms in the Petit-Saut Reservoir and in the Sinnamary Estuary have not been studied yet in terms of microbial and physico-chemical characterization. However, macroscopic observation suggested a strong difference between reservoir biofilms and those downstream from the dam. The first appeared relatively thin (approx. 1 mm thick) compared to the second (up to 1 cm thick). The estuarine biofilm was red–brown coloured, probably due to the oxidation of the ferrous species present in the anoxic part of the reservoir water. 2.3.1. Cell density measurements Total numbers of microorganisms were determined according to a protocol described in detail elsewhere (Lunau et al., 2005). Briefly, microorganisms in 1 mL sample aliquots were dispersed by 10% (v/v) methanol and sonication during 15 min at 35 °C (ultrasonic bath 120 W, RK31, 35 kHz, Sonoclean, Paris, France). Cells were then fixed with glutardialdehyde and stained directly on a black Nucleopore filter with SybrGreen I (Invitrogen S7567) in a mounting medium (buffered moviol and ascorbic acid as antioxidant). Bacteria were observed and counted by epifluorescence microscopy after several minutes, i.e. after installation of an equal dispersion of the mounting medium. For each sample, 30 view fields were enumerated. 2.3.2. Total organic carbon concentration and dry weight Biofilm samples were dried at 60 °C (72 h) immediately after sampling and stored in sealed containers at ambient temperature, then ground and homogenized. No carbonate removal was done because of the acidic pH (5–6) and the low carbonate concentrations at the sampling site (Abril G., pers. com.). Measurements were done with a 2500 CARLO-ERBA CN Elemental Analyzer. TOC was expressed as a percentage of dry weight. Dry weight of the suspended solid was measured by desiccating (105 °C) the filtrate (glass microfiber filter, GF/C grade, Whatman) or the centrifuged pellet (5000 ×g, 5 min) of 10 to 25 mL samples from the reservoir or the estuary, respectively. 2.3.3. Iron, sulphur and manganese total concentrations Exactly 3 mL of biofilm sample were diluted in MilliQ water® to 30 mL final volume and then 4 mL of HNO3 (65% suprapur) and 4 mL of H2O2 (35% suprapur) were added followed by ultrasonic treatment (ultrasonic bath 120 W, RK31, 35 kHz, Sonoclean, Paris, France) during 3 h. After extraction, the samples were filtered through 0.22 μm filters (Millipore, SLGP033RS) in order to prevent introduction of residual particles into the analyzer. Total Fe, S and Mn concentrations were determined by ICP-AES (Ultima, Jobin Yvon, Horiba Group) under standard conditions. 2.3.4. Total mercury and methylmercury concentrations Biofilm components and suspended matter in sample aliquots (10 to 15 mL) were recovered by filtration through 0.45 μm Teflon membranes (LCR, Millipore). The filters were immediately stored at

L. Huguet et al. / Science of the Total Environment 408 (2010) 1338–1348

−18 °C until analysis. Total mercury (HgT) concentration was measured by flameless atomic absorption spectrometry (AAS, AMA 254) after thermal sample decomposition at 750 °C under oxygen flow and amalgamation on a gold trap. Protocols and quality control are detailed in Dominique et al. (2007). Methylmercury concentrations in the suspensions containing biofilm components were measured by species-specific isotope dilution mass spectrometry analysis (SID-ICP-MS). The applied protocol including extraction (acid leaching), derivation (propylation; NaBPr4) and the analytical setup (gas chromatography coupled to ICP-MS:GC-ICP-MS) is based on the method described in detail elsewhere (Monperrus et al., 2004), that was, however, adjusted to biofilm samples as follows. After adding 202Hg-enriched MeHg to 10 ng L− 1 (CH202 3 HgCl; ERMAE670, isotopic purity 98%) and equilibration during 4 h (rotary shaking; 5 rpm Heidolph instruments; Reax 20), MeHg was extracted from plankton and biofilm slurries and certified reference materials by acidification to 1% final concentration using ultrapure HCl (J.T.Baker, INSTRA analyzed) combined with ultra-sonication for 1 h (150 W, 28 kHz, Ultrasonic bath 220, Deltasonic, Meaux, France). The suspension was then decanted for 15 min and 10 mL of the supernatant were recovered in a 20 mL borosilicate glass bottle with Teflon®-lined cap (RESTEK). The extract was then buffered with 500 μL of acetic acid-sodium acetate buffer (1 M; Sigma-Aldrich/ Merck Suprapur) and adjusted to pH 3.9 by addition of 100 μL of NH4OH (25% m/v, Suprapur, Merck). Then, 250 μL of isooctane (J.T. Baker, HPLC-analyzed) and 300 μL of sodium tetrapropylborate (Galab, 1% w/v in KOH 2% w/v) were added in order to derivatize (propylate) the MeHg (Monperrus et al., 2004; Castelle et al. 2007). After vigorous manual shaking (5 min) and centrifugation (2000 rpm; 20 min), the organic layer was transferred into 2 mL glass vials with Teflon®-lined-caps (Thermo Fisher Scientific) and stored at −18 °C until measurement by GC-ICP-MS. Quantification of MeHg in the organic extract was done by gas chromatography (Thermo Fisher Scientific; Focus GC) coupled to ICPMS (Thermo Fisher Scientific; X7) using a thermostatic interface as described elsewhere (Rodriguez Martin-Doimeadios et al., 2003). The detection limit of the method (3 sigma of the blank values including reagent blanks and sample handling; nominal volume 30 mL) was 0.05 pmol L− 1. The analytical results were continuously quality checked by analyzing different international certified reference materials (IAEA 356 polluted marine sediment, IAEA 405 estuarine sediment; IAEA 407 fish homogenate), as there exists no biofilm reference material and the algae reference material IAEA 140 (Fucus) certified for MeHg is no longer commercially available. For the contrasting reference materials used, accuracy was 94 ± 4% (IAEA 356, IAEA 405) and 102 ± 5% (IAEA 407) and precision (RSD) was generally better than 5%, supporting the idea that the applied extraction protocol produced reliable results for the biofilm microorganism suspensions. Methylmercury concentrations in the samples were calculated using the isotope dilution equation (Hintelmann and Evans, 1997). 2.4. Incubation conditions Assuming that microbial activity greatly influences Hg methylation, different sets of batch experiments were conducted. The methylation potential of plankton microorganisms contained in unfiltered reservoir water from the anoxic zone without addition of biofilm components was quantified in batches without nutrient supply. Additionally, biofilm components were incubated in ambient unfiltered water (including plankton microorganisms) sampled at the same time and depth/ location as the biofilms in order to study methylation in nutrient conditions close to those in the natural system. On the other hand, biofilm components were incubated in liquid culture medium (Table 1) to study methylation under high nutrient conditions allowing the growth of a large part of the biomass. Prior to incubation, 50 mL of biofilm components from three different depths or from the estuary

1341

Table 1 Global characteristics of the reservoir and estuarine waters: suspended particles matter (SPM), nitrogen Kjeldahl (NK), total phosphorus (PT), dissolved organic carbon (DOC), total organic carbon (TOC), mercury (HgT), methylmercury (MeHgT) and MeHgT/HgT ratio and composition of the culture medium.

SPM (mg L− 1)a, c NK (10− 6 mol L− 1)a, c PT (10− 6 mol L− 1)a, c DOC (mg L− 1)a, d TOC (mg L− 1)a, d HgT (ng L− 1)b, d MeHgT (ng L− 1)b, d MeHgT/HgTb, d pHa, d

Reservoir water

Estuarine water

Culture medium

3.2 ± 1.4 49.1 ± 29 1.3 ± 0.6 3.8 ± 0.9 4.3 ± 1.1 2.6 ± 1.2 0.4 ± 0.2 15% 5.5 ± 0.5

b5 25 ± 15 1.2 ± 0.2 4.9 ± 0.5 5.6 ± 0.2 2.2 ± 0.6 0.7 ± 0.2 32% 5.7 ± 0.3

Yeast extract Peptone Casamino acids Glucose Soluble starch K2PO4 MgSO4 Sodium pyruvate pH

0.5 g L− 1 0.75 g L− 1 0.25 g L− 1 0.5 g L− 1 0.5 g L− 1 0.3 g L− 1 0.3 g L− 1 0.3 g L− 1 7.2

The culture medium was N2-purged and autoclaved 15 min at 121 °C. Reservoir water: amean values from 2006 provided by Hydreco laboratory, b hypolimnion values (Muresan et al., 2008b). Estuarine water: cAbril et al. (2008), dMuresan et al. (2008a).

(Table 2) were transferred into 500 mL Teflon bottles and mixed with 300 mL of either unfiltered reservoir/estuary water or the culture medium. The incubations of plankton microorganisms were conducted in 30 mL Teflon bottles, containing 30 mL of unfiltered reservoir or estuarine water. As the dry weight of suspended solids in the reservoir is typically b5 mg L− 1 between −3 m and −25 m (De Junet et al., 2009), one may consider that the contribution of the water reservoir to the dry weight of the suspended matter in the batches is less than ~10–15%. All incubations were stopped by acidification (HCl 1%) after different time intervals between 1 h and 90 d (Table 3).

2.5. Enriched stable Hg isotope incubation Mercury methylation potential was evaluated from spiking precise amounts of 199Hg(II) chloride to the samples, followed by incubation (different time-series). The evolution over time of Me199Hg in the samples was then followed by species-specific isotope dilution GCICP-MS (Monperrus et al., 2003; Castelle et al., 2007). Preliminary tests with biofilm components from the Passerelle site, incubated with 3.5, 35 and 350 ng of 199Hg(II), suggested that under these experimental conditions (unfiltered natural waters, presence of biofilm components containing natural MeHg, potential adsorption to the batch container walls, analytical precision, etc.) only incubations with 350 ng of 199Hg(II) provided reproducible and significant results (data not shown). The resulting nominal Hg(II) concentration in the biofilm microorganism slurries (~1000 ng/L) was (i) similar in magnitude to that of other recent incubation studies (e.g. Mauro et al., 2002; Acha et al., 2005; Coelho-Souza et al., 2006; Table 4) and (ii) clearly below the minimum inhibitory concentrations previously described for different bacterial biofilms (ranging from 4 to 80 mg L− 1, depending on the bacterial species) (Harrison et al., 2004). Accordingly, in this work the samples were spiked with 350 ng of 199Hg(II) (Oak Ridge National Laboratory, 199Hg(II) chloride; isotopic purity 92%). Plankton microorganism and biofilm component incubations were performed under anaerobic conditions in the dark without shaking in a thermostated incubator at 30 °C. In spite of the aerobic environment in the Sinnamary Estuary (pO2 ~ 5 mg L− 1), the biofilms from the estuary were also incubated in anaerobic conditions. Due to the thickness of the biofilm (up to 1 cm), the presence of anaerobic zones could be expected within the biofilm. Moreover, as it is widely recognized that methylating activity involves anaerobic bacteria (sulfate reducing and/or iron reducing bacteria), incubation under anaerobic conditions could favor Hg methylation. Aliquots of 30 mL were sampled after 1 h, 18 h, 2.5 d, 4.5 d, 7 d and 90 d for MeHg analysis. The microorganism densities were determined in 1 mL aliquots (see Section 2.3.1) for 0 h and 7 d. The unfiltered aliquots were stabilized with 300 μL of

1342

L. Huguet et al. / Science of the Total Environment 408 (2010) 1338–1348

Table 2 Main characteristics of the reservoir and estuarine raw biofilms in terms of dry weight (DW) (per m− 2 of the wooden plank substrata), total concentrations (DW) of organic carbon (TOC), iron (FeT), sulphur (ST), manganese (MnT), mercury (HgT), methylmercury (MeHgT) and MeHgT/HgT ratio and dry weight and cell densities in the batches.

Reservoir biofilms Estuarine biofilms

Depth (m)/ location

DW (g m− 2)

TOC (% DW)

FeT (mg g− 1)

ST (mg g− 1)

MnT (mg g− 1)

HgT (μg g− 1)

MeHgT (ng g− 1)

MeHgT/HgT (%)

DW (mg L− 1)

Cell densities (×109 cells g− 1)

− 6.5 − 15 − 20 P V

0.41 ± 0.08 0.37 ± 0.07 0.96 ± 0.02 n.d. n.d.

34 31 27 10 13

87 ± 4 46 ± 2 20 ± 1 438 ± 22 144 ± 7

1.7 ± 0.1 18 ± 1 26 ± 1 6.4 ± 0.3 2.2 ± 0.1

7.9 ± 0.4 10.3 ± 0.5 1.4 ± 0.1 3.3 ± 0.2 3.2 ± 0.2

9.2 ± 1.5 9.2 ± 1.5 4.6 ± 0.8 4.1 ± 1.8 0.3 ± 0.1

20 ± 1 118 ± 6 100 ± 5 54 ± 3 15 ± 1

0.2 1.3 2.2 1.3 5.6

36 ± 7 32 ± 6 83 ± 1 344 ± 39 594 ± 122

58 ± 8 59 ± 11 16 ± 8 10 ± 2 4.4 ± 1.2

P = Passerelle, V = Venus. n.d. = not determined.

ultrapure HCl and stored at 4 °C until MeHg analysis (see Section 2.3.4). Incubations using plankton microorganisms were analyzed at the start and after 7 d (end of incubation). Due to the limited quantity of biofilm recovered, no replicate incubations were performed, because priority was given to the time-resolved observation of methylation rates. A first series of control (‘abiotic’) experiments was conducted by heating 30 mL aliquots of biofilm/water mixtures during 30 min at 100 °C, as no autoclave was available in the on-site laboratory. However, this thermal treatment did not inactivate the entire bacterial population and growth was observed in the culture medium incubation controls. This cannot be attributed to external contamination because the bottles were kept closed all the time between heating and incubation steps. However, methylation was retarded and clearly (3-fold) lower in those controls than in batches with non-heated samples (data not shown). For both culture medium and reservoir water, a second series of control (abiotic) experiments was re-conducted 3 weeks later in the metropolitan laboratory. No microorganism growth and no Hg methylation were observed after 7 d of incubation (30 °C, in the dark) in culture medium and heat-sterilized reservoir waters (121 °C, 30 min). 2.6. Calculation of methylation rates and specific methylation efficiencies Methylmercury concentrations in the samples were calculated using the isotope dilution equation (Hintelmann and Evans, 1997). The excess Me199Hg (i.e. net Me199Hg produced; Me199Hgnet) in the samples after incubation was evaluated from initial values and from natural MeHg isotopes in the sample (e.g. Me200Hg) and was attributed to methylation of the 199Hg(II) spike. The results were expressed as percent of Me199Hg produced during incubation

(Me199Hg/199Hg(II) × 100). Since the samples were incubated in the dark, abiotic photochemical degradation of Me199Hg was ruled out. Methylation rates were calculated by dividing the gains of MeHg by the incubation time (7 d) (% of Me199Hgnet d− 1). Specific methylation efficiencies were calculated for the 7 d incubations by dividing the methylation rates by the number of microorganisms counted (% of Me199Hgnet d− 1 cell− 1 mL− 1). 3. Results 3.1. Main characteristics of the raw biofilm The main characteristics of the biofilm samples are given in Table 2. Dry weight and cell densities were similar for the biofilm of the reservoir at − 6.5 m and −15 m with means of 0.41 ± 0.08 g m− 2 and 0.37 ± 0.07 g m− 2 giving 34 ± 6 mg L− 1 and 58 × 109 cells g− 1 in the corresponding batch experiments (Table 2). At − 20 m the dry weight was 2.5-fold higher and the cell density 3.5-fold lower than for the upper biofilms. In term of cells g− 1, estuarine biofilms from Passerelle and Venus showed the lowest cell densities (10 ± 0.2 × 109 and 4.4 ± 1.2 × 109 cells g− 1, respectively) and the highest dry weight in the batches (344 ± 39 and 594 ± 122 mg L− 1). Since the surface area of the substrata of these biofilms was not calibrated, the amount of biofilm in g m− 2 was not calculated. However, a rough estimation of the substrata surface indicates a value between 30 and 40 g m− 2 for Passerelle biofilms (not shown). Raw biofilms contained between 27 and 34% of TOC in the reservoir and 10 and 13% in the estuary (Table 2). Iron concentrations were the highest compared to sulphur and manganese concentrations except for the biofilm sampled at −20 m. Maximum Fe values occurred near the oxycline in the reservoir (87 ± 4 mg g− 1) and at the

Table 3 Net methylation potential (Me199Hgnet (%)), methylation rates and cell densities of batches containing planktonic microorganisms with or without addition of biofilm components incubated in the presence of 199Hg(II) during 1 h up to 90 d. Depth (m)/ location

Reservoir planktonic microorganisms with biofilm components

RW

CM

Reservoir planktonic microorganisms

RW

Estuarine planktonic microorganisms with biofilm components

EW CM

− 6.5 − 15 − 20 − 6.5 − 15 − 20 − 9.5 − 15 - 24.5 P V P V

Me199Hgnet (%)

Methylation rates (% d− 1)

Cell densities (106 × cells mL− 1)

1h

18 h

2.5 d

4.5 d

7d

90 d

Over 7 d

1h

7d

0.02 ± 0.01 0.09 ± 0.01 0.16 ± 0.03 0.01 ± 0.01 0.15 ± 0.02 0.07 ± 0.04 0.02 ± 0.01 0.06 ± 0.01 0.02 ± 0.01 0.02 ± 0.02 0.09 ± 0.02 bd.l. bd.l.

0.02 ± 0.01 0.06 ± 0.01 0.05 ± 0.01 0.03 ± 0.01 0.13 ± 0.02 0.07 ± 0.02 n.d. n.d. n.d. 0.44 ± 0.09 0.15 ± 0.02 0.04 ± 0.02 0.21 ± 0.04

0.02 ± 0.01 0.10 ± 0.04 0.08 ± 0.06 0.43 ± 0.08 0.09 ± 0.01 0.22 ± 0.03 n.d. n.d. n.d. 0.99 ± 0.2 0.25 ± 0.04 0.77 ± 0.09 0.71 ± 0.09

0.02 ± 0.01 0.05 ± 0.01 0.08 ± 0.02 0.02 ± 0.01 0.24 ± 0.03 1.11 ± 0.17 n.d. n.d. n.d. 1.1 ± 0.2 0.17 ± 0.02 4.3 ± 0.6 14 ± 3

0.03 ± 0.02 0.05 ± 0.01 0.18 ± 0.04 0.34 ± 0.13 0.32 ± 0.04 2.8 ± 0.5 3.5 ± 1.8 0.6 ± 0.1 0.05 ± 0.01 1.0 ± 0.1 0.26 ± 0.04 15 ± 2 44 ± 6

0.15 ± 0.02 bd.l. 1.9 ± 0.3 9.8 ± 1.2 10.3 ± 1.3 1.0 ± 0.15 n.d. n.d. n.d. 1.3 ± 0.2 0.23 ± 0.03 81 ± 19 85 ± 15

0.005 ± 0.003 0.007 ± 0.001 0.026 ± 0.006 0.053 ± 0.019 0.046 ± 0.006 0.40 ± 0.07 0.5 ± 0.3 0.09 ± 0.02 0.008 ± 0.001 0.15 ± 0.03 0.04 ± 0.01 2.1 ± 0.2 6.3 ± 0.9

2.1 ± 0.7 1.9 ± 0.7 1.3 ± 0.7 1.4 ± 0.4 1.9 ± 0.7 2.0 ± 0.6 2.1 ± 0.8 2.3 ± 0.6 1.7 ± 0.5 3.6 ± 1.1 2.6 ± 1.2 3.6 ± 1.1 2.6 ± 1.2

1.8 ± 1.3 1.9 ± 0.7 1.1 ± 0.3 424 ± 43 332 ± 31 660 ± 69 2.1 ± 0.5 2.4 ± 0.7 4.1 ± 1.1 20 ± 6 9.2 ± 1.7 261 ± 76 900 ± 92

The incubation medium was the reservoir water (RW), the estuarine water (EW) or the culture medium (CM). P = Passerelle, V = Venus. n.d. = not determined. d.l. detection limit. Standard deviation (± SD) represents uncertainty considering a 5% error on isotopic ratio measurement.

L. Huguet et al. / Science of the Total Environment 408 (2010) 1338–1348

1343

Table 4 Mean methylation rates of varied aquatic ecosystems. Location − ecosystem type

Dry weight/cell density

Hg spike

Bolivia — floodplain lake: macrophyte roots-associated periphyton Argentina — ultraoligothophic lake Moreno: photosynthetic plankton

17 g L− 1

203

103 phytoplanktonic cells mL− 1

197

Pure cultures of Desulfovibrio desulfuricans Canada — lakes: water column Brazil — Ribeirão das Lajes reservoir: phytoplankton Brazil — Fazenda Ipiranga Lake: floating macrophytes Florida — Everglades artificial wetland Brazil — eutrophic tropical Lagoinha lake Macrophytes-associated periphyton

Incubation time (d)

Methylation rate (% d— 1)

Ref.

HgCl2 670–13,300 ng L− 1

1

27.5–36.1

Acha et al. (2005)

Hg2+ 10 ng L− 1

3

Sterilized+darkness: ≤6 Sterilized + light:≤10 Light: ≤ 17

Ribeiro Guevara et al. (2008)

107 cells mL− 1

Hg(II) 50 ng L− 1 sulfide 5 μmol L− 1

2

Biofilms: 17.3 Plankton: 6.1

Lin and Jay (2007)

n.d.

199

1

Eckley and Hintelmann (2006)

Guimaraes et al. (1998)

HgCl2 2.4–7.5 ng L− 1

1.5–3.4 · 10 cells mL

203

−1

HgCl2 1750 ng L

4

0.56–15.8 Max: oxycline 1.5

33 g L− 1

203

HgCl2 1534 ng L− 1

3

2.2–3.5

7

−1

Mauro et al. (2002) 8–17 g L

−1

203

200

HgCl2 or HgCl2 456–913 ng L− 1

4–12

Cultivated periphyton Water (oxycline) Mediterranean Sea: Coastal waters Marine waters French Guiana–Petit-Saut reservoir: Reservoir biofilm microorganisms Water column planktonic bacteria Estuarine biofilm microorganisms Canada — oligotrophic boreal shield lake: epilithon

Coelho-Souza et al. (2006)

17 or 1.5–7.7 1.6 0.9 or 0.8

n.d.

199

HgCl2 1–2 ng L− 1

1

Monperrus et al. (2007) 0.8–6.3 0–0.5

36–83 mg L− 1/ 2 · 106 cells mL− 1 350–600 mg L− 1/ 3 · 106 cell mL− 1 10 g L− 1

199

Hg Cl2 1000 ng L− 1

7

0.04 0.5 0.2

This study

203

HgCl2 2 ng L− 1

48

0.4

Desrosiers et al. (2006)

n.d. = not determined.

Passerelle station in the estuary (438 ± 22 mg g− 1). Sulphur concentrations ranged from 1.7 ± 0.1 to 26 ± 1 mg g− 1 (maximum in the biofilm sampled at −20 m). Estuarine biofilm samples had similar Mn concentrations (3.3 ± 0.2 mg g− 1) at the 2 sampling stations and were clearly higher for reservoir biofilms with maximum value (10.3 ± 0.5 mg g− 1) at − 15 m. Total Hg concentrations were 9.2 ± 1.5 μg g− 1 in both biofilms sampled at − 5.5 and −15 m but methylmercury concentration was 6-fold higher at −15 m (118 ± 6 ng g− 1, the maximum observed). For the biofilm sampled at − 20 m, total Hg was 2-fold lower than for upper biofilms with MeHg being 100 ± 5 ngg− 1. Accordingly, the MeHgT/HgT ratio was ≤2.2% for the reservoir biofilms (Table 2). Downstream, the biofilm sample from the Passerelle station contained more Hg and MeHg (4.1± 1.8 μg g− 1, 54± 3 ngg− 1, respectively) than the Venus biofilm (0.3 ± 0.1 μgg− 1, 15± 1 ngg− 1, respectively). Interestingly, the MeHgT/HgT ratio was higher for the Venus biofilm than for the Passerelle biofilm (5.6 vs. 1.3%). 3.2. Mercury methylation potential 3.2.1. Methylation potential of unfiltered reservoir water with biofilm components The Hg methylation potential of reservoir water was assessed along incubation time with biofilm components in the presence of 199 Hg. In addition, the incubation experiments were performed in a bacterial culture medium to assess if the Hg methylation rate can be related to the cell growth and/or cell density. During the batch incubation of unfiltered reservoir water with biofilm components the cell densities were rather similar whatever the depth (average: 1.8 ± 0.4 × 106 cells mL− 1) and did not significantly change between 1 h and 7 d (Table 3). In contrast, cell densities increased by a factor of 2

(i.e. from 1.8 ± 0.3 × 106 to 472 ± 169 × 106 cells mL− 1; during 7 d in all batches using the artificial culture medium (Table 3). The percentage of net Me199Hg produced (Me199Hgnet) differed from one depth to another and evolved with time, depending on the incubation conditions (Table 3). Methylation of the spiked 199Hg occurred in all incubation batches whatever the depth or the nutrient conditions during the first week. The data suggest a rapid increase in Me199Hg within the first hour (up to 0.16% after 1 h), then Me199Hg remained relatively low (b0.5%) during the first week, except for the batch from −20 m incubated in culture medium (2.8 ± 0.3%). For most of the batches, Me199Hgnet levels tended to increase during the 90 d of incubation, reaching values close to ~ 10% in batches of biofilm components from −6.5 m and −15 m incubated in culture medium (Table 3). In contrast, the batch containing biofilm components from − 15 m incubated in reservoir water and the batch of biofilm components retrieved at −20 m and incubated in culture medium showed low % Me199Hgnet values at 90 d (Table 3). Methylation rates, determined after 7 d of incubation, appeared to increase with depth for biofilm components incubated in reservoir water from − 6.5 m (0.005 ± 0.003% d− 1) to −20 m depth (0.026 ± 0.006% d− 1) (Table 3). Incubation in the culture medium enhanced the methylation rates of nearly all the batches by up to one order of magnitude probably reflecting the cell density increase (×200 to 300 after one week, Table 3). Like methylation rates, the specific methylation efficiency expressed as % of Me199Hgnet d− 1 cell− 1 (Fig. 2), was higher for the batch of biofilm components from −20 m (27 ± 2 × 10− 9) than for the two upper depths −15 and −6.5 m (4.7× 10− 9 and 2.8 × 10− 9). Interestingly, the reservoir biofilm components incubated in the culture medium showed significantly lower specific mercury methylation efficiencies

1344

L. Huguet et al. / Science of the Total Environment 408 (2010) 1338–1348

Fig. 2. Specific methylation efficiencies (% Me199Hgnet d− 1 cell− 1 mL− 1) calculated over the first week of incubation for planktonic microorganisms with biofilm components (− 6.5 m, − 15 m, − 20 m, Passerelle and Venus) and for planktonic microorganisms without biofilm components (− 9.5 m, − 15 m and − 24.5 m). The incubation medium was the reservoir water (RW), the estuarine water (EW) or the culture medium (CM).

than those incubated in the reservoir water (ranging from 0.12 ± 0.03 × 10− 9 to 6.1 ± 0.4× 10− 9% Me199Hgnet d− 1 cell− 1 mL− 1, Fig. 2). 3.2.2. Methylation potential of unfiltered estuary water with biofilm components At t=1 h, cell densities in batches from the Sinnamary Estuary were similar for the Passerelle and Venus stations (4 and 3±1×106 cellsmL− 1 respectively; Table 3). During the experiment, growth was observed in both estuarine water and culture medium incubations. In estuarine water, the increase in cell density was two fold higher for Passerelle (20×106 and 9.2×106 cellsmL− 1 reached at 7d, respectively). In contrast, the opposite occurred in the culture medium, where cell densities reached 261± 76×106 and 900±92×106 cellsmL− 1 in Passerelle and Venus incubations, respectively (Table 3). Incubation of planktonic microorganisms with biofilm components from both the Passerelle and the Venus stations in unfiltered estuarine water resulted in significant Me199Hgnet increases over time, with methylation being up to 5-fold higher for Passerelle than for Venus (1.3 ± 0.2% vs. 0.23 ± 0.03%, respectively, after 90 d of incubation time, Table 3). Me199Hgnet level reached 0.99 ± 0.2 and 0.25 ±0.04% after 2.5 d respectively and remained almost constant until the end of the incubation period. In contrast, when incubated in culture medium, the planktonic microorganisms with biofilm components from the Venus site produced 3-fold more Me199Hgnet during the first 7 d (44 ±6%) than those from the Passerelle batch (15 ± 2%). At the end of the experiment (90 d), however, the Passerelle and the Venus batches showed very high and rather similar Me199Hgnet levels of up to 85 ± 15% and 81 ± 19% respectively (Table 3). The resulting Hg methylation rates were 0.15 ± 0.03% d− 1 (Passerelle site) and 0.04 ± 0.01% d− 1 (Venus site) (Table 3). As for Me199Hgnet levels, Hg methylation rates in culture medium were ~3-fold higher for Venus (6.3 ± 0.9% d− 1) than for the Passerelle site (2.1 ± 0.2% d− 1). In estuarine water incubations, the % of Me199Hgnet d− 1 cell− 1 mL− 1 was ~ 2-fold higher in the batch containing biofilm-components from the Passerelle site (7.7 ± 0.8 × 10− 9) than for that from the Venus site (4.0 ± 0.2 × 10− 9). In culture medium, the specific methylation efficiencies of estuarine plankton microorganisms and biofilm from both sites were similar (~7.5 ± 0.5 × 10− 9% Me199Hgnet d− 1 cell− 1 mL− 1; Fig. 2). 3.2.3. Methylation potential of plankton microorganisms from the reservoir Cell densities in incubations of plankton microorganisms in reservoir water sampled at −9.5 m and −15 m were rather similar

(2.2 ± 0.7 × 106 cells mL−1) and did not evolve over time, whereas in bottom water (−24.5 m) incubations cell density slightly increased from 1.7 ± 0.5×106 cells mL−1 to 4.1 ± 1.1 × 106 cells mL−1 (Table 3). Maximum Me199Hgnet (3.5 ± 1.8%, 7d) occurred for water sampled at −9.5 m (Table 3). This percentage was respectively ~100-fold and ~10-fold higher than those obtained for biofilm components in reservoir water from the nearest depth ranges (−6.5 m and −15 m), and in culture medium (Table 3). In the batch with water sampled at −15 m, plankton microorganisms produced 0.6 ± 0.1% of Me199Hgnet, i.e., 6-fold and 2-fold more Me199Hgnet, respectively, than the biofilmcontaining batches from the same depth with and without nutrient addition. The lowest Me199Hgnet production after 7 d occurred for the bottom water (−24.5 m) incubation (0.05 ± 0.01%). The respective Hg methylation rates decreased with depth from 0.5 ± 0.3%d− 1, to 0.008 ± 0.001%d− 1 for plankton microorganisms (Table 3). The same trend was observed for the specific methylation efficiencies with the highest value obtained for plankton microorganisms from −9.5 m (2.4 ± 0.6 × 10− 7% of Me199Hgnet d− 1 cell− 1 mL− 1). At −15 and −24.5 m, specific mercury methylation efficiencies were 1– 2 orders of magnitude lower (3.8± 0.6 × 10− 8 and 1.8 ± 0.3 × 10− 9% of Me199Hgnet cell− 1, Fig. 2). Cell densities in incubations of plankton microorganisms in reservoir water sampled at −9.5 m and −15 m were rather similar (2.2±0.7×106 cellsmL− 1) and did not evolve over time, whereas in bottom water (−24.5 m) incubations cell density slightly increased from 1.7±0.5 ×106 cellsmL− 1 to 4.1 ±1.1 ×106 cells mL− 1 (Table 3). 4. Discussion 4.1. Influence of nutrient supply on Hg methylation Damming the Sinnamary River immerged ~ 10 Mt of organic carbon (Galy-Lacaux et al., 1997) and nutrients easily available to different biogeochemical processes. In the first years, decomposition of the flooded terrestrial organic matter and the related nutrient recycling were intense and decreased over time to the recent, clearly lower level (Delmas et al., 2004; Abril et al., 2005). Accordingly, one may assume that reactive organic carbon and nutrients available for heterotrophic microorganisms were much more abundant in the early period after flooding. The colonisation of the diverse flooded substrata probably started at the same time as the increase of the water level, and over the years, the microbial communities adapted to the

L. Huguet et al. / Science of the Total Environment 408 (2010) 1338–1348

changing environmental conditions (available nutrients, Eh/pH, temperature, light intensity,…) (Dumestre et al., 2001). The incubation of biofilm components in batch experiments cannot reproduce strictosenso the environment of this microbial community and probably gives only a rough estimate of their in-situ methylation performances. Conditions in the present incubation experiments, with a culture medium on the one hand and real reservoir/estuarine water on the other, may probably reflect two contrasting situations: one with ‘unlimited’ optimum offer of highly available organic matter and nutrients and the other with less favorable conditions. Although the culture medium does probably not reproduce nutrient conditions in the reservoir during the first years after flooding, this setup may serve as a first approach to test the methylation potential of microorganism communities from Petit-Saut under contrasting nutrient conditions. Furthermore, one cannot exclude nutrient depletion in the batch reactors during batch incubations over days or weeks. Assuming that such depletion probably did not occur in the reservoir, the results obtained at the end of the incubation period (90 d) may roughly reflect the behavior of autochthonous microorganisms when they are submitted to strong nutrient limitation. The rapid onset of Me199Hgnet production in all batches with natural reservoir water suggests that the microorganisms and the respective waters from different depth ranges in the reservoir allow significant net methylation in the respective zones and under the current nutrient conditions in the Petit-Saut reservoir. Enhanced Hg methylation in batches using the culture medium could simply signify an increase among the methylating actors due to the cell growth. However, it could also be provided by the experimentally modified nutrient conditions (i) increased bacterial activity and/or (ii) favored Hg(II) uptake by bacteria. The presence of yeast extract and casamino acid components contained in the culture medium has been reported to favor bacterial Hg(II) uptake (Golding et al., 2002). The fact that Me199Hgnet production was more efficient in batches containing estuarine biofilm microorganisms (Table 3) compared to batches with reservoir biofilm microorganisms, may be attributed to the observed higher cell numbers rather than specific methylation efficiencies of the bacterial communities involved (Fig. 2). In fact, specific methylation efficiencies of batches containing plankton microorganisms and biofilm components from the planks were clearly higher (approx. 10×) for incubations in reservoir water than for those in culture medium (Fig. 2). Accordingly, the culture medium did not specifically support development, activity or efficiency of methylating bacteria, despite the strong increase in total bacteria number. This may suggest that, based on bacteria communities initially present in reservoir water and biofilms, the anaerobe culture medium favored growth of less- or non-methylating and/or demethylating microorganisms. For instance, the low amount of sulfate (2.5 mM, Table 1) and the presence of highly fermentable organic matter (glucose, starch,…) could limit the growth/activity of the sulfate reducing bacteria, a metabolic group known to be implied in Hg methylation. And thereafter the abundant labile organic carbon in the culture medium would be used by other bacteria. On the other hand, one cannot exclude Hg complexation by organic ligands or by additional sulfides produced in the culture medium, which would also decrease the Hg bioavailability to methylating agents (Fitzgerald et al. 2007). Contrary to results obtained with reservoir plankton microorganisms with biofilm components, the specific methylation efficiencies of estuarine plankton microorganisms with biofilm components were similar in estuarine water and in the culture medium incubations (Fig. 2), suggesting that both nutrient conditions were similarly favorable to methylating microorganism communities. 4.2. Mercury methylation by plankton microorganisms with and without biofilm components The unfiltered water from the anoxic zone of the reservoir was expected to contain plankton microorganisms including sulfate- and

1345

Fe-reducing bacteria (Dumestre et al., 1999), suggesting that these plankton communities would contribute to Hg methylation during the incubations (e.g. Compeau and Bartha, 1985; Kerin et al., 2006). The cell densities in the batch experiments containing plankton microorganisms in unfiltered reservoir water were not significantly different from those with biofilm components added (Table 3). The batches with biofilm components had lower methylation rates than those without. Accordingly, specific methylation efficiencies of plankton microorganisms alone from the anoxic reservoir water column were 1–2 orders of magnitude higher than those of plankton microorganisms mixed with biofilm components (Fig. 2), strongly suggesting that plankton microorganism communities in the water column may play a major role in Hg methylation at the reservoir scale and that the addition of biofilm-microorganisms and associated material suppressed the Hg methylation by the plankton microorganisms (− 15 m (Fig. 2)). The maximum of specific methylation efficiency observed at −9.5 m is consistent with (i) the generally observed maximum dissolved MeHg concentrations in the reservoir water column around − 9 m (Muresan et al., 2008a) and (ii) maximum relative abundances of sulfate reducing plankton bacteria in the reservoir water column around − 9 m to − 10 m depth (Dumestre et al., 2001). These results suggest (i) a major contribution of this metabolic group to Hg methylation in the anoxic reservoir water column and (ii) that the presence of biofilm components in the water column modifies Hg methylation by plankton microorganisms. 4.3. Reduced methylation efficiency of plankton microorganisms in the presence of biofilm components Disturbing biofilm structures by natural processes or by the applied experimental procedure may alter their methylating capacity, i.e. the experimentally determined Hg methylation potential of natural intact biofilms could be underestimated as well as overestimated. However, adding biofilm microorganisms to unfiltered reservoir water already containing methylating plankton microorganisms, we expected increased Hg methylation rates compared to batches without biofilm microorganisms. Surprisingly, the opposite occurred for batches with similar cell densities (Fig. 2), suggesting that, under the experimental conditions, the presence of biofilm components impaired Hg methylation. Mercury bioavailability, a key parameter for methylation (Fitzgerald et al., 2007), may account for the different results obtained for incubations with and without biofilm microorganisms. Aqueous HgS0, presumably one of the most available Hg species to bacteria (Drott et al., 2007), is commonly considered a major substratum for methylation, although this may imply its transformation to less stable compounds inside the cells prior to methylation. Recent work also evidenced that the formation of Hg-cysteine complexes may increase Hg uptake and methylation (Schaefer and Morel, 2009). Geochemical modelling (VisualMintek) using the environmental data of the reservoir presented in Table 1 (Nica Donnan model was used for DOM), data from Muresan et al. (2008b) (e.g. pH = 5; HS− = 2 μM; etc.) and data from pluriannual observation in the reservoir (Hydreco Laboratory; data not shown), obtained for similar seasonal and hydrological conditions (short dry season) suggests that aqueous HgS0, presumably one of the most available Hg species to bacteria (Drott et al., 2007), concentrations were in the 10− 15 to 10− 13 molar range in reservoir water from all depths beneath the oxycline. Although picomolar concentrations of neutral Hg-sulfide complexes are thought to exist in the reservoir hypolimnion, reduced methylation efficiency in batches containing plankton microorganisms and biofilm components may have resulted from their limited availability. For example, DOM may complex Hg and Fe at multiple acid sites corresponding to functional groups of carboxylic acids, phenols, ammonium ions, alcohols, and thiols (Chadwick et al., 2005; Liu et al., 2008). Mercury is expected to preferentially bind with thiol and other sulphur-containing groups (Ravichandran, 2004) and its complexation with DOM generally

1346

L. Huguet et al. / Science of the Total Environment 408 (2010) 1338–1348

limits the amount of inorganic Hg available to methylating bacteria (Barkay et al., 1997, Haitzer et al. 2002), because DOM molecules are too large to cross the cell membrane. Given that DOC represents about 80– 90% of the TOC in the reservoir water (Table 1), humic substances released from resuspended biofilms could additionally compete for Hg and thus limit Hg bioavailability (Haitzer et al. 2002). Furthermore, the physiological properties of biofilms are determined by bacterial cells, whereas their chemical and physical properties particularly depend on the extracellular polymeric matrix which is the main structural component of biofilm (De Beer and Stoodley, 2006). Biofilm extracellular matrix could lead to Hg and MeHg immobilizations (Hintelmann et al., 1993; Cheng et al., 2008) and thus, at least partly, explain the decreased methylation efficiency after adding biofilm components to unfiltered reservoir water. Finally, high total Fe concentrations in the biofilms from the PetitSaut system (Table 2) probably limit Hg availability to biofilm microorganisms due to sulfide precipitation and/or adsorption (Mehrotra and Sedlak, 2005). Interestingly, the highest Fe concentration in biofilmcomponents from the reservoir matched well with the lowest methylation rate (Tables 2 and 3). The biofilms in the reservoir contain much less sulfate reducing bacteria (~104 cells g− 1; d.w.) than iron reducing bacteria (~108 cells g− 1; d.w., unpublished data) which is another metabolic group implied in the Hg methylation (Kerin et al., 2006). Assuming that Fe concentration in biofilm reflects the balance between Fe-oxidizing and -reducing bacteria (i.e. between precipitation and reductive dissolution, respectively), Hg methylation in biofilms would be related to iron reducing activity, i.e. the Fe concentration in biofilm would be expected to be inversely related to MeHg/Hg. The design of the present study (e.g. sample numbers, number of depth ranges, etc.) does not allow establishing such relationships, indicating that further work should be focused on the microbiological control factors of the system. Further experiments with culture medium specifically favoring distinct (e.g. SRB and iron reducing bacteria) microorganism groups are necessary to improve our comprehension of their specific contribution to Hg methylation in the reservoir water column and in undisturbed biofilms. 4.4. Hg methylation at the Petit-Saut reservoir scale The methylation rates obtained from incubations of tropical plankton and biofilm microorganisms in natural waters are within the range of values reported for other aquatic ecosystems (Table 4), including marine waters from the Mediterranean Sea and of freshwater from Canadian or tropical lakes (Mauro et al., 2002; Eckley and Hintelmann, 2006; Monperrus et al., 2007). In tropical ecosystems, periphyton communities associated to macrophyte roots were reported to be key locations of intense Hg methylation and the role of photosynthetic microorganisms is significant inside those systems (Guimaraes et al., 1998; Mauro et al., 2002; Acha et al., 2005; Coelho-Souza et al., 2006; Ribeiro Guevara et al.,, 2008). Periphyton compositions are generally rather similar to that of biofilms, but the existing literature data suggest higher methylation activity for periphyton than that observed for the incubations of mixed plankton/biofilm incubations from the Petit-Saut reservoir (Tables 2 and 4; Mauro et al., 2002). The observed methylation rates were also clearly lower than those reported for mono-species pure culture biofilms (similar bacterial densities) and incubated in a medium which promotes sulfate reducing bacteria growth (Lin and Jay, 2007). Interestingly, the incubations in natural water over 3 months of planktonic microorganisms with biofilm components from the Petit-Saut reservoir clearly suggest that methylation continued after 7 d, because Me199Hgnet values after 90 d were ~5 to 10-fold higher than those after 7 d (Table 3). Accordingly, one may assume that the Me199Hgnet values observed after 7 d for reservoir planktonic microorganisms with biofilm components may not correspond to a final state, whereas in incubations with estuarine plankton microorganisms and biofilm components

Me199Hgnet did not evolve after 2.5 d, suggesting that methylation stopped after few days. This may eventually be attributed to nutrient depletion suspected in the batches, because in culture medium estuarine plankton/biofilm communities continued methylating during 3 months, using almost the total amount of 199Hg(II) added (Table 3). Given that the different types of microorganisms showed rather contrasting responses (e.g. specific methylation efficiencies) for incubations in natural water and in nutrient-rich culture medium, one may suppose that variations in the environmental nutrient status of the studied system would probably impact in-situ MeHg production. During the 2003–2004 period, maximum MeHg inputs into the Petit-Saut reservoir via rainfall and tributaries were estimated to 5.4 mol yr− 1 (i.e.: 1080 g MeHg yr− 1), whereas exportations to the downstream estuary averaged 13.5 mol yr− 1, suggesting that in-situ net MeHg production in the reservoir accounts for ~ 8 mol yr− 1, corresponding to an average methylation rate of 0.06% d– 1 (Muresan et al., 2008a). This value is rather close to the methylation rate obtained from reservoir plankton-microorganisms with biofilm components incubation in natural water (average ~ 0.01% d− 1) but is up to 10-fold lower than that of plankton microorganism incubations in natural water alone (average ~0.2% d− 1). Assuming that the hypolymnion volume of the reservoir is ~109 m3 (during dry season, surface area = 230 km2; average depth = 13 m; average depth of the oxycline = 5−7 m; Muresan et al., 2008a), an average methylation rate of plankton microorganisms in reservoir water of 0.2% d− 1, and an average dissolved Hg concentration of ~2 ng L− 1 (10− 11 M) in the reservoir (Muresan et al., 2008a), one may roughly extrapolate the experimentally obtained results to the reservoir scale. Using higher water volumes and lower MeHg concentrations as reported for the wet seasons (e.g. Muresan et al., 2008a) produced similar estimates at the daily timescale. This suggests that seasonal variations in MeHg production may be low and that these results may also be extrapolated to the annual timescale. This rough estimation suggests that plankton microorganisms could account for an in-situ MeHg production of ~27 mol yr− 1, which is three times higher than the estimated net in-situ MeHg production (~8 mol yr− 1; Muresan et al., 2008a) derived from flux estimates in the system. Comparatively, extrapolation of Hg methylation rates observed in the presence of biofilm components (~0.01% d− 1) gives 1.4 mol yr− 1, which appears rather low, compared to the existing mass balance. However, Hg methylation was supposed also for the oxic layer of the reservoir (Muresan et al., 2008a) and one cannot exclude Hg methylation by structurally intact biofilms. Accordingly, plankton microorganisms may strongly contribute to gross MeHg production in the reservoir that is partly compensated by antagonistic processes such as (i) adsorption of the MeHg on suspended particulate matter and following sedimentation, (ii) adsorption to biofilms on the immersed trunks, and/or (iii) demethylation. The MeHg/Hg ratio, commonly accepted as a proxy for Hg methylation (Fitzgerald et al., 2007; Drott et al., 2008), in the reservoir and estuarine biofilms was 10-fold lower than that of the unfiltered water (Table 2). This supports the idea that plankton microorganisms strongly contribute to MeHg production in tropical aquatic systems and that biofilms are not necessarily sites of intense net Hg methylation. Assuming that biofilms receive both MeHg and Hg from the water column, the low MeHg/Hg ratio would suggest (i) preponderant inorganic Hg assimilation due to adsorption and/or (co-) precipitation or (ii) active demethylation inside the biofilms. Given that biofilms are an important entry of MeHg into the aquatic food chain, further work is needed to assess the methylating potential of structurally intact tropical biofilms. 5. Conclusion This study provides the first lines of evidence showing the high Hg methylation potential of plankton microorganisms in tropical

L. Huguet et al. / Science of the Total Environment 408 (2010) 1338–1348

reservoirs and their downstream waters. This could be useful for assessing the impact of reservoirs on Hg cycle, especially in tropical regions that undergo gold-mining activities. Decreased specific methylation efficiencies in the water samples containing plankton microorganisms (from the anoxic reservoir water column) in the presence of biofilm components (under the experimental conditions and at constant cell densities) were attributed to limitation in bioavailable Hg (e.g. due to adsorption, (co-precipitation), changes in sulfide concentrations, etc.) in the batches. This and the low MeHg/ Hg ratios in biofilms may reflect predominant accumulation of inorganic Hg and/or demethylation supporting the idea that biofilms are not necessarily the predominant sites of net methylation in tropical aquatic systems. Adding nutrients promoted both microorganism growth and methylation rates, but the resulting lower specific methylation efficiencies suggest that, under the experimental conditions, the culture medium (i) impaired the growth of methylating bacteria, or (ii) enabled the growth of non-methylating or demethylating microorganisms. The surprisingly high methylation potentials of estuarine biofilm communities in culture medium, however, clearly suggest that additional nutrient transport to the estuary may modify the MeHg budget in this vulnerable ecosystem. Further work is needed to (i) quantify the methylation potential of structurally intact biofilms, (ii) identify the bacterial communities involved and quantify their specific methylation efficiencies and (iii) confirm the present results for other aquatic systems. Acknowledgments We would like to thank the divers Y. Godart and P. Motreff (Sub Cayman, Kourou, French Guiana) for sampling biofilms on the trunk and Dr. F. Redero, L. Lanceleur and F. Mathieu for technical support with GC-ICP-MS and TOC analyses. We greatly acknowledge the thoughtful and useful comments of both anonymous reviewers and the contributions of Professor A. Boudou, who initiated the project and of Drs S. Richard and O. Merdaci for fruitful discussion and the help during the first part of the project. This work was supported by the EDF Company and by the French National Scientific Research Center (CNRS/INSU, ACI-ECCO research programme). L. Huguet benefited from a Ph.D. Grant by EDF. References Abril G, Guérin F, Richard S, Delmas R, Galy-Lacaux C, Gosse P, et al. Carbon dioxide and methane emissions and the carbon budget of a 10-year old tropical reservoir (Petit Saut, French Guiana). Glob Biogeochem Cycles 2005;19, doi:10.1029/2005GB002457. Abril G, Boudou A, Cerdan P, Delmas R, Erard C, Forget PM, et al. Petit Saut : Bilan environnemental après 12 années de fonctionnement. Scientific Committe of the Petit-Saut Dam Report, Toulouse, France: Bucerep; 2008. Acha D, Iniguez V, Roulet M, Guimaraes JRD, Luna R, Alanoca L, et al. Sulfate-reducing bacteria in floating macrophyte rhizospheres from an amazonian floodplain lake in Bolivia and their association with Hg methylation. Appl Environ Microbiol 2005;71: 7531–5. Baeyens W, Leermarkers M, Papina T, Saprykin A, Brion N, Noyen J, et al. Bioconcentration and biomagnification of mercury and methylmercury in North Sea and Scheldt Estuary fish. Arch Environ Contam Toxicol 2003;45:498–508. Barkay T, Gillman M, Turner RR. Effects of dissolved organic carbon and salinity on bioavailability of mercury. Appl Environ Microbiol 1997;63:4267–71. Barkay T, Wagner-Döbler I. Microbial transformations of mercury: potentials, challenges, and achievements in controlling mercury toxicity in the environment. Adv Appl Microbiol 2005;57:1-52. Bellanger B, Huon S, Steinmann P, Chabaux F, Velasquez F, Vallès V, et al. Oxic–anoxic conditions in the water column of a tropical freshwater reservoir (Pena-Larga dam, NW Venezuela). Appl Geochem 2004;19:1295–314. Boening DW. Ecological effects, transport, and fate of mercury: a general review. Chemosphere 2000;40:1335–51. Boudou A, Maury-Brachet R, Coquery M, Durrieu G, Cossa D. Synergic effect of gold mining and damming on Hg contamination in fish. Environ Sci Technol 2005;39: 2448–54. Castelle S, Schäfer J, Blanc G, Audry S, Etcheber H, Lissalde J-P. 50-year record and solid state speciation of mercury in natural and contaminated reservoir sediment. Appl Geochem 2007;22:1359–70. Celo V, Lean DRS, Scott SL. Abiotic methylation of mercury in the aquatic environment. Sci Total Environ 2006;368:126–37.

1347

CTFT (Centre Technique Forestier Tropical) Fiches techniques sur les bois guyanais : Wapa (Technical sheets on French Guyanan woods: Wapa). Bois et Forêts des Tropiques. 1984; 204: 69–72. Chadwick SP, Babiarz CL, Hurley JP, Armstrong DE. Influences of iron, manganese, and dissolved organic carbon on the hypolimnetic cycling of amended mercury. Sci Total Environ 2005;368:177–88. Cheng J, Zhao W, Liu Y, Wu C, Liu C, Wang W. Adsorption properties and gaseous mercury transformation rate of natural biofilm. Bull Environ Contam Toxicol 2008;81: 516–20. Coelho-Souza SA, Guimaraes JR, Mauro JB, Miranda MR, Azevedo SM. Mercury methylation and bacterial activity associated to tropical phytoplankton. Sci Total Environ 2006;364:188–99. Compeau GC, Bartha R. Sulfate reducing bacteria: principal methylators of mercury in anoxic estuarine sediment. Appl Environ Microbiol 1985;50:498–502. Coquery M, Cossa D, Azemard S, Peretyazhko T, Charlet L. Methylmercury formation in the anoxic waters of the Petit-Saut reservoir (French Guiana) and its spreading in the adjacent Sinnamary River. J Phys IV 2003;107:327–31. De Beer D, Stoodley P. Microbial biofilms. Prokaryotes 2006;1:904–37 chapter 3.10. De Junet A, Abril G, Guérin F, Billy I, De Wit R. A multi-tracers analysis of sources and transfers of particulate organic matter in a tropical reservoir (Petit Saut, French Guiana). River Res Appl 2009;25:253–71. De Mérona B, Vigouroux R, Tejerina-Garro FL. Alteration of fish diversity downstream from Petit-Saut Dam in French Guiana. Implication of ecological strategies of fish species. Hydrobiologia 2005;551:33–47. Delmas R, Richard S, Guérin F, Abril G, Galy-Lacaux C, Delon C, Grégoire A. Long Term Greenhouse Gas Emissions from the Hydroelectric Reservoir of Petit Saut (French Guiana) and Potential Impacts. Vol 12: Tremblay, A.,Varfalvy, L., Roehm, C., Garneau, M, 2004. Desrosiers M, Planas D, Mucci A. Mercury methylation in the epilithon of boreal shield aquatic ecosystems. Environ Sci Technol 2006;40:1540–6. Dominique Y. Contamination par les différentes formes chimiques du mercure de la composante biologique du barrage hydroélectrique de Petit-Saut et des zones amont/ aval du fleuve Sinnamary, en Guyane francaise (études in situ et approches expérimentales). 2006. PhD Thesis, University Bordeaux 1, Bordeaux, France, 366 pages. Dominique Y, Maury-Brachet R, Muresan B, Vigouroux R, Richard S, Cossa D, et al. Biofilm and mercury availability as key factors for mercury accumulation in fish (Curimata cyprinoides) from a disturbed amazonian freshwater system. Environ Toxicol Chem 2007;26:45–52. Drott A, Lambertsson L, Björn E, Skyllberg U. Do potential methylation rates reflect accumulated methylmercury in contaminated sediments? Environ Sci Technol 2008;42:153–8. Drott A, Lambertsson L, Björn E, Skyllberg U. Importance of dissolved neutral mercury sulfides for methyl mercury production in contaminated sediments. Environ Sci Technol 2007;41:2270–6. Dumestre J-F, Casamayor EO, Massana R, Pedros-Alio C. Changes in bacterial and archaeal assemblages in an equatorial river induced by the water eutrophication of Petit Saut dam reservoir (French Guiana). Aquat Microb Ecol 2001;26:209–21. Dumestre J-F, Vaquer A, Gosse P, Richard S, Labroue L. Bacterial ecology of a young equatorial hydroelectric reservoir (Petit Saut, French Guiana). Hydrobiologia 1999;400: 75–83. Durrieu G, Maury-Brachet R, Boudou A. Goldmining and mercury contamination of the piscivorous fish Hoplias aimara in French Guiana (Amazon basin). Ecotoxicol Environ Saf 2005;60:315–23. Eckley CS, Hintelmann H. Determination of mercury methylation potentials in the water column of lakes across Canada. Sci Total Environ 2006;368:111–25. Fitzgerald WF, Lamborg CH, Hammerschmidt CR. Marine biogeochemical cycling of mercury. Chem Rev 2007;107:641–62. Fréry N, Maury-Brachet R, Maillot E, Deheeger M, de Mérona B, Boudou A. Gold-mining activities and mercury contamination of native amerindian communities in French Guiana: key role of fish in dietary uptake. Environ Health Perspect 2001;109: 449–56. Galy-Lacaux C, Delmas R, Jambert C, Dumestre J-F, Labroue L, Richard S. Gaseous emissions and oxygen consumption in hydroelectric dams: a case study in French Guyana. Glob Biogeochem Cycles 1997;11:471. Garcia E, Poulain AJ, Amyot M, Ariya PA. Diel variations in photoinduced oxidation of Hg0 in freshwater. Chemosphere 2005;59:977–81. Gochfeld M. Cases of mercury exposure, bioavailability, and absorption. Ecotoxicol Environ Saf 2003;56:174–9. Golding GR, Kelly CA, Sparling R, Loewen C, Rudd JWM, Barkay T. Evidence for facilitated uptake of Hg(II) by Vibrio anguillarum and Escherichia coli under anaerobic and aerobic conditions. Limnol Oceanogr 2002;47:967–75. Gosse P, Grégoire A. Artificial re-oxygenation of the Sinnamary, downstream of PetitSaut dam (French guiana). Hydroécol appl 1997;1–2:23–56 Tome 9. Guimaraes JRD, Meili M, Malm O, Maria de Souza Brito E. Hg methylation in sediments and floating meadows of a tropical lake in the Pantanal floodplain, Brazil. Sci Total Environ 1998;213:165–75. Haitzer M, Aiken GR, Ryan JN. Binding of mercury to dissolved organic matter. Environ Sci Technol 2002;36:3564–70. Harrison JJ, Ceri H, Stremick CA, Turner RJ. Biofilm susceptibility to metal toxicity. Environ Microbiol 2004;6:1220–7. Hintelmann H, Ebinghaus R, Wilken R-D. Accumulation of mercury(II) and methylmercury by microbial biofilms. Water Res 1993;27:237–42. Hintelmann H, Evans RD. Application of stable isotopes in environmental tracer studies — measurement of monomethylmercury (CH3Hg+) by isotope dilution ICP-MS and detection of species transformation. Fresenius' J Anal Chem 1997;358:378–85. Kerin EJ, Gilmour CC, Roden E, Suzuki MT, Coates JD, Mason RP. Mercury methylation by dissimilatory iron-reducing bacteria. Appl Environ Microbiol 2006;72:7919–21.

1348

L. Huguet et al. / Science of the Total Environment 408 (2010) 1338–1348

Li P, Feng X, Qiu G, Shang L, Wang S. Mercury exposure in the population from Wuchuan mercury mining area, Guizhou, China. Sci Total Environ 2008;395:72–9. Lin C-C, Jay JA. Mercury methylation by planktonic and biofilm cultures of Desulfovibrio desulfuricans. Environ Sci Technol 2007;41:6697-2007. Liu J, Valsaraj KT, Devai I, Delaune RD. Immobilization of aqueous Hg(II) by mackinawite (FeS). J Hazard Mater 2008;157:432–40. Lunau M, Lemke A, Walther K, Martens-habbena W, Simon M. An improved method for counting bacteria from sediments and turbid environments by epifluorescence microscopy. Environ Microbiol 2005;7:961–8. Mauro JBN, Guimaraes JRD, Hintelmann H, Watras CJ, Haack EA, Coelho-Souza SA. Mercury methylation in macrophytes, periphyton, and water — comparative studies with stable and radio-mercury additions. Anal Bioanal Chem 2002;374: 983–9. Mehrotra AS, Sedlak DL. Decrease in net mercury methylation rates following iron amendment to anoxic wetland sediment slurries. Environ Sci Technol 2005;39: 2564–70. Monperrus M, Krupp E, Amouroux D, Donard OFX, Rodriguez Martin-Doimeadios RC. Potential and limits of speciated isotope-dilution analysis for metrology and assessing environmental reactivity. Trends Anal Chem 2004;23:261–72. Monperrus M, Rodriguez Martin-Doimeadios RC, Scancar J, Amouroux D, Donard OFX. Simultaneous sample preparation and species-specific isotope dilution mass spectrometry analysis of monomethylmercury and tributyltin in a certified oyster tissue. Anal Chem 2003;75:4095–102. Monperrus M, Tessier E, Amouroux D, Leynaert A, Huonnic P, Donard OFX. Mercury methylation, demethylation and reduction rates in coastal and marine surface waters of the Mediterranean Sea. Mar Chem 2007;107:49–63. Muresan B, Cossa D, Richard S, Burban B. Mercury speciation and exchanges at the air–water interface of a tropical artificial reservoir, French Guiana. Sci Total Environ 2007;385: 132–45. Muresan B, Cossa D, Richard S, Dominique Y. Monomethylmercury sources in a tropical artificial reservoir. Appl Geochem 2008a;23:1101–26. Muresan B, Cossa D, Coquery M, Richard S. Mercury sources and transformations in a man-perturbed tidal estuary: the Sinnamary Estuary, French Guiana. Geochim Cosmochim Acta 2008b;72:5416–30. Pak K-R, Bartha R. Mercury methylation and demethylation in anoxic lake sediments and by strictly anaerobic bacteria. Appl Environ Microbiol 1998;64:1013–7.

Peretyazhko T, Van Cappellen P, Meile C, Coquery M, Musso M, Regnier P, et al. Biogeochemistry of major redox elements and mercury in a tropical reservoir lake (Petit Saut — French Guiana). Aquat Geochem 2005;11:33–55. Peretyazhko T, Charlet L, Muresan B, Kazimirov V, Cossa D. Formation of dissolved gaseous mercury in a tropical lake (Petit-Saut reservoir, French Guiana). Sci Total Environ 2006;364:260–71. Poulain AJ, Garcia E, Amyot M, Campbell PGC, Raofie F, Ariya PA. Biological and chemical redox transformations of mercury in fresh and salt waters of the high arctic during spring and summer. Environ Sci Technol 2007;41:1883–8. Ravichandran M. Interactions between mercury and dissolved organic matter — a review. Chemosphere 2004;55:319–31. Ribeiro Guevara SR, Queimalinos CP, Dieguez MdC, Arribére M. Methylmercury production in the water column of an ultraoligotrophic lake of Northern Patagonia, Argentina. Chemosphere 2008;72:578–85. Richard S, Arnoux A, Cerdan P. Evolution in physico-chemical water quality in the reservoir and downstream following the filling of Petit Saut dam (French Guiana). Hydroécol appl 1997;9:57–83. Rocha JC, Sargentini EJ, Zara LF, Rosa AH, Dos Santos A, Burba P. Reduction of mercury (II) by tropical river humic substances (Rio Negro) — a possible process of the mercury cycle in Brazil. Talanta 2000;53:551–9. Rodriguez Martin-Doimeadios RC, Monperrus M, Amouroux D, Donard OFX. Using speciated isotope dilution with GC-inductively coupled plasma MS to determine and unravel the artificial formation of monomethylmercury in certified reference sediments. Anal Chem 2003;75:3202–11. Rosenberg DM, Berkes F, Bodaly RA, Hecky RE, Kelly CA, Rudd JWM. Large-scale impacts of hydroelectric development. Environ rev 1997;5:27–54. Roulet M, Lucotte M, Farella N, Serique G, Coelho H, Sousa Passos CJ, et al. Effects of recent human colonization on the presence of mercury in amazonian ecosystems. Water Air Soil Pollut 1999;112:297–313. Schaefer JK, Morel FMM. High methylation rates of mercury bound to cysteine by Geobacter sulfurreducens. Nature Geoscience (2009). Sissakian C. General presentation of Petit Saut hydroelectric installation (French Guiana) and ecological monitoring program. Hydroécol appl 1997;1–2:1-21 Tome 9. Zahir F, Rizwi SJ, Haq SK, Khan RH. Low dose mercury toxicity and human health. Environ Toxicol Pharmacol 2005;20:351–60.