Environmental Research 125 (2013) 30–40

Contents lists available at ScienceDirect

Environmental Research

0013-93http://d

n CorrE-m

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

In-situ subaqueous capping of mercury-contaminated sedimentsin a fresh-water aquatic system, Part I—Bench-scale microcosm studyto assess methylmercury production

Paul M. Randall a,n, Ryan Fimmen b, Vivek Lal c, Ramona Darlington c

a U.S. Environmental Protection Agency, Office of Research and Development, National Risk Management ResearchLaboratory, 26 W. Martin Luther King Drive, Cincinnati, OH 45268, USAb Geosyntec Consultants, 150 E. Wilson Bridge Road, Suite 232, Worthington, OH 43085, USAc Battelle, 505 King Ave., Columbus, OH 43201, USA

a r t i c l e i n f o

Article history:Received 14 July 2012Received in revised form9 February 2013Accepted 2 March 2013Available online 13 June 2013

Keywords:MethylmercuryMercuryMethylationMicrocosmSediment

51/$ - see front matter Published by Elsevierx.doi.org/10.1016/j.envres.2013.03.012

esponding author.ail addresses: randall.paul@epa.gov, pran1053

a b s t r a c t

Bench-scale microcosm experiments were designed to provide a better understanding of the potentialfor Hg methylation in sediments from an aquatic environment. Experiments were conducted to examinethe function of sulfate concentration, lactate concentration, the presence/absence of an aqueousinorganic Hg spike, and the presence/absence of inoculums of Desulfovibrio desulfuricans, a strain ofsulfate-reducing bacteria (SRB) commonly found in the natural sediments of aquatic environments.Incubations were analyzed for both the rate and extent of (methylmercury) MeHg production.Methylation rates were estimated by analyzing MeHg and Hg after 2, 7, 14, 28, and 42 days.The production of metabolic byproducts, including dissolved gases as a proxy for metabolic utilizationof carbon substrate, was also monitored. In all treatments amended with lactate, sulfate, Hg, and SRB,MeHg was produced (37 ng/g-sediment dry weight) after only 48 h of incubation and reached amaximum sediment concentration of 127 ng/g-sediment dry weight after the 42 day incubation period.Aqueous phase production of MeHg was observed to be 10 ng/L after 2 day, reaching a maximumobserved concentration of 32.8 ng/L after 14 days, and declining to 10.8 ng/L at the end of the incubationperiod (42 day). The results of this study further demonstrates that, in the presence of an organic carbonsubstrate, sulfate, and the appropriate consortia of microorganisms, sedimentary Hg will be transformedinto MeHg through bacterial metabolism. Further, this study provided the basis for evaluation of anin-situ subaqueous capping strategy that may limit (or potentially enhance) MeHg production.

Published by Elsevier Inc.

1. Introduction

Mercury (Hg) is a heavy metal neurotoxin that is present in theenvironment from both natural and anthropogenic sources.It exists in three main forms in aquatic and marine environments:elemental (Hg0), inorganic (e.g., mercuric chloride [HgCl2],Hg [OH]2, Hg2+ metal-ligand complexes) and organic methylmer-cury (MeHg). Of these three forms, organic Hg is by far the mosttoxic; this has been most notably illustrated in the Minamata Bayincident (Hosokawa, 1993). Once Hg has been introduced into theaquatic environment, a complex set of biologically mediatedchemical reactions within the anaerobic region of the aquaticsediments leads to the conversion of Hg to MeHg. It is important todetermine conditions under which MeHg production is enhanced or,

Inc.

34@gmail.com (P.M. Randall).

more constructively, can be reduced by the implementation ofspecific remedial activities.

MeHg is formed mainly at the sediment-water interface, at thetransition from oxic to anoxic conditions, by the complex interac-tion of inorganic Hg and microorganisms (Whalin et al., 2007).The conversion of Hg to MeHg is widely accepted to be governedby the action of sulfate-reducing bacteria (SRB) (although not allSRB methylate Hg) which enzymatically catalyze the methylationof inorganic Hg (King et al., 2000, 2002; Warner et al., 2003; Drottet al., 2007). Also, iron-reducing bacteria have the ability tomethylate Hg; however, this was not explicitly considered here(Fleming et al., 2006). Despite the lack of information on thekinetics of individual mechanisms, it is generally accepted thatbiomethylation of Hg during metabolic sulfate reduction is theprimary pathway of Hg methylation in the natural aquatic envir-onment. Although the principal SRBs are responsible for Hgmethylation, the kinetics of Hg uptake through the cell membraneof SRB has not been reported in the literature. The best evidencefor the link between Hg methylation and sulfate reduction is likely

P.M. Randall et al. / Environmental Research 125 (2013) 30–40 31

found in work concluding that Hg methylation by SRB in pureculture is almost completely quenched with the addition ofmolybdate (a metabolic inhibitor to SRB)(Compeau and Bartha,1985; Gilmour and Henry, 1992; Gilmour et al., 1992, 1998; Kinget al., 1999). Similar work found that SRB did not methylate Hg toany appreciable extent in the absence of sulfate (Pak and Bartha,1998b). Although the precise mechanism of Hg methylation hasnot been fully elucidated, specific enzymatic mechanisms havebeen proposed (Gilmour and Henry, 1991; Choi et al., 1994; Pakand Bartha, 1998a). Previous microcosm studies have been con-ducted to investigate the transformation of Hg to MeHg and thefactors that affect MeHg production. In all studies, there wastransformation of Hg to MeHg in the presence of an appropriateorganic substrate (electron donor), sulfate (electron acceptor) andone or more strains of SRB (in either pure culture or sedimentslurries). The goal of this study is to determine the specificcombinations of electron donor (organic substrate) and electronacceptor (sulfate) which cause the greatest production of MeHgwhen in the presence of inorganic Hg and one specific strain ofSRB under sediment conditions with a mercury concentrationgreater than 100 ppm. Data from this study may assist in theinvestigation of a remedial strategy (i.e., the implementation of“active” vs. “passive” in-situ remedial cap).

In total, 19 different genera of SRB have been described in theliterature (Rooney-Varga et al., 1998), although more recent workspurred by the development of more sophisticated microbialmethods have led to the identification of others (Miletto et al.,2007). All of these SRB genera are hypothesized to have thepotential to methylate Hg in the aquatic environment under theappropriate conditions of electron donor, electron acceptor(sulfate) and in the presence of Hg. Research reported onHg methylation in the natural environment (not pure-culturestudies but those that use sediment slurries in artificial micro-cosms) have identified six different genera of SRB that, to date, areconsidered the main putative Hg methylators in the naturalaquatic environment: Desulfovibrio, Desulfobulbus, Desulfococcus,Desulfobacter, and Desulfobacterium (King et al., 2000). Studiescited in the previous reference are concerned primarily with theidentification of SRB strains in natural sediment slurries followedby measurement of Hg methylation rates (MMR, measured asnmol-MeHg mL−1 h−1) and sulfate reduction rates (SRR, measuredas nmol SO4 mL−1 h−1). King's study (King et al., 2000) describesthe putative Hg methylator in a consortium of different SRBisolated from natural environments and illustrates that all strainsof SRB have the same MMR/SRR ratio, thus providing evidence fora singular Hg methylation pathway shared across SRB genera.Three important observations were made in this study: (1) amongthose studied, the genera displaying the highest growth rate afterintroduction of Hg during the log-phase of growth was Desulfovibrio;(2) among those studied, the genera displaying the largest SRR wasDesulfovibrio; (3) amongst those studied, the genera demonstratingthe highest MMR amongst those investigated was Desulfobacteriumfollowed by Desulfovibrio. Thus, Desulfovibrio was chosen as the SRBfor use in the current study. The fact that previous pure-culturestudies have extensively used Desulfovibrio as a model SRB organism(Compeau and Bartha, 1985; Gilmour and Henry, 1991; Choi et al.,1994; Pak and Bartha, 1998a,b) supports this decision.

In order to ensure the necessary growth conditions, D. desulfuricansrequires the presence of Hg and sulfate (as an electron acceptor), aswell as the addition of an electron donor, typically as a carbon source.In choosing the appropriate electron donor for the current study, itwas important to ensure high SRR as well as a high growth rate of D.desulfuricans. Aquatic environments are not single colonies, but rathera complex and vast consortium of many different microorganisms.Thus, it is important that for on-going sulfate reduction and optimalconditions for Hg methylation, SRB are not outcompeted by other

microorganisms naturally present within the sediments. In particular,it has been shown under certain temperature regimes and dominantelectron donors, methanogenic organisms and acetogenic bacteria canoutcompete SRB leading to the reduction and, in some cases, theeventual removal of SRB from the system (Liamleam and Annachhatre,2007). Thus, suitable electron donors, such as lactate, which wouldensure a high growth rate of D. desulfuricans, were considered for usein the present study.

Previous studies found that high sulfate concentrations (over30 mg/L) inhibited methylation of Hg (Gilmour et al., 1992). It washypothesized that this was a geochemical effect rather thana biological one, since sulfate is reduced to sulfide during SRBheterotrophic activity, which reacts readily with inorganic Hg toprecipitate HgS (cinnabar) and renders Hg unavailable for methy-lation (Warner et al., 2003; Deonarine and Hsu-Kim, 2009).Another study supporting this hypothesis found that increasingsulfate concentration by ten-fold over natural concentrations didincrease MeHg production, but not proportionally. This non-stoichiometric relationship may be due to an increase in MeHgdemethylation or a decline in Hg bioavailability due to theformation of HgS. Another possible explanation is that sulfatemay no longer be the limiting agent, but rather the bioavailableorganic matter may be limiting. It should also be noted that incases where SO4 becomes the limiting reactant, SRB will revert tosyntrophic fermentation. However, this is not expected to occur inthe present study due to the excess of SO4 in the microcosmsystems. Microcosms were treated with variable sulfate amend-ments (no added sulfate, 50 mg/L, 200 mg/L), variable lactateconcentrations (no lactate, 250 mg/L, 1000 mg/L), with and with-out a Hg spike (150 mg/L as HgCl2), and with and without theaddition of D. desulfuricans. Levels of sulfate were chosen based onthe stoichiometry of the sulfate reducing equation when lactate isused as an electron donor in which 0.5 mol of SO4 are consumedfor every 1 mol of lactate. Here, there was an attempt to inten-tionally force a limiting SO4 condition:

CH3CHOHCOOH+0.5SO42−+H+-CH3COOH+CO2+0.5HS−+H2O

In order to follow the biomethylation of inorganic Hg from theperspective of the C-substrate, metabolites and known metabolicbyproducts were also measured in the microcosms at each timepoint. SRB metabolize lactate as part of their metabolic process,transforming it to acetate and then CO2. Methanogens can meta-bolize the products of SRB metabolism (namely acetate) fromwhich they produce CH4. Thus, it is important to measure the CH4

concentration in microcosms to determine the consumption ofcarbon resources by methanogens. Upon depletion of lactate, SRBcompete with methanogens for the limiting acetate. Lactate mayalso be fermented by certain organisms to form propionate. Thus,propionate is another important product that was analyzed in thisexperiment. The rates of lactate metabolism and formation ofacetate by SRB and propionate during fermentation (Gottschalk,1986; Freedman et al., 2002) were measured, in addition to theCO2 and CH4 end-products. The quantification of the C-substrateallows the evaluation of methanogenic activity relative to SRBactivity; the latter does not produce CH4.

In this study, bench-scale microcosm experiments were designedto provide a better understanding of the potential for Hg methylationin anthropogenically contaminated natural sediments as a function ofsulfate concentration, carbon substrate (lactate) concentration, thepresence/absence of an aqueous inorganic Hg spike (as HgCl2), andthe presence/absence of artificially inoculated SRB. The specific goalof this experiment was to characterize the extent of Hg methylationunder specific parameters in an aquatic system. This work will befollowed-up with a series of sorption, incubation and column studiesevaluating potential capping materials. In contrast to previous

Table 1Lake sediment and porewater chemistry.

Sediment, mg/kg dry wgt.Mercury 39.9Arsenic 7.46Barium 147Cadmium 0.57Chromium 41.1Lead 21.1Selenium 9.05Silver o0.3Porewater, μg/LMercury o0.2Arsenic o6.0Barium 64.2Cadmium o1.0Chromium o6.0Lead o1.0Selenium o6.0Silver o2.0Porewater, mg/LChloride 105Nitrate-N o0.5Sulfate 202Alkalinity, Bicarbonate 50.0

P.M. Randall et al. / Environmental Research 125 (2013) 30–4032

Hg methylation studies where lower level Hg concentrations wereconsidered, factors governing Hg methylation were considered in aseverely contaminated environment, with total initial Hg concentra-tions of 150 mg/L in the spiked microcosms. Worldwide, there aremany lakes with high mercury concentrations in the surface sedi-ments. For example, Ullrich reported mean mercury concentrations4150 mg/kg in surface sediments in Lake Balkyldak, Kazakhstanwith levels reaching up to 1500 mg/kg in surface sediments near awastewater outfall pipe (Ullrich et al., 2007). Another example wasJapan's Minamata Bay, where Hg concentrations as high as 600 mg/kg were detected in settled sediment (Hosokawa, 1993). The resultsof benchscale experiments will enable the determination of optimalconditions for biomethylation of inorganic Hg, which will beemployed in subsequent batch-sorption, incubation and column-transport experiments evaluating different capping materials. Theresults of this study attempts to answer the question of whatconditions of carbon substrate (lactate), electron donor (lactate),and electron acceptor (sulfate) cause the greatest potential formethylation of Hg under specific conditions.

2. Materials and methods

2.1. Preparation of materials

All 200 mL serum bottles used as microcosms for the incubation study wereinverted in glass trays and autoclaved at 121 1C for 1 h then wrapped in aluminumfoil to prevent dust or debris from entering the open bottles before their use in theincubation experiments. Tap water was added to approximately 75% of the heightof 1-L Pyrexs bottles with ceramic screw caps and then autoclaved at 121 1C for 1 hfor sterilization and to remove chlorine. Chlorine removal was confirmed byanalyzing autoclaved water via the Hach N,N-diethyl-p-phenylene-diamine (DPD)method using a DR/2000 colorimeter. All other materials used in the study, such asspatulas, beakers, funnels and graduated cylinders, were first cleaned withAlconoxs soap, rinsed once with tap water, and then rinsed three times withdeionized (DI) water before being autoclaved at 250 1F.

2.2. Sediment collection and preparation

Sediment for microcosm experiments was collected from a lake in the South-eastern U.S.A. Two 5-gallon buckets were used for sample collection and contain-ment. The sediment samples were collected beneath at least 1 ft depth of water toensure anoxic conditions typical of contaminated lake sediments. A Ponar dredgewas used to collect the top 4 in. of sediment. During sediment collection someexposure of sediment to air was inevitable, although this was avoided whenpossible. All buckets were sealed with no headspace to the extent possible foranalysis and stored in a cold room at 472 1C. Sediment samples were brought toroom temperature before use.

Sediment and porewater (obtained by decanting the overlaying water in the5-gal bucket) were analyzed for eight Resource Conservation and Recovery Act(RCRA) metals per standard methods SW1311/7470A for Hg and SW1311/6020 forother metals and major anions by USEPA Method E300 (results are presented inTable 1). These results are from discrete samples of homogenized sediment. Resultsfrom metals analysis confirmed elevated sediment Hg concentrations (39.9 mg-Hg/kg dry weight [DW] sediment). Sediment results also revealed the presence ofbarium (147 mg-Ba/kg DW sediment), chromium (41.1 mg-Cr/kg DW sediment)and lead (21.1 mg-Pb/kg DW sediment). Porewater chemistry results showedelevated levels of sulfate (202 mg/L as SO4) and chloride (105 mg/L as Cl).This level of sulfate in sediment porewater indicates that native and/or inoculatedSRB should have the appropriate electron acceptor for metabolic requirements.The sediments were also shown to have high moisture content (70%).

Prior to microcosm experiments, one of the 5 gallon buckets was placed insidean anaerobic glove box (Terra Universal, Fullerton, CA) containing an inert, oxygen-free atmosphere (95%/5% N2/H2, v/v). The overlaying water was decanted and thesediments were homogenized inside the glove box using a clean stainless steelspoon. Mixing was performed as quickly as possible to minimize drying and anyimpact to particle size. Twigs, shells, leaves and/or small stones were removed fromthe sediment samples. Visual observations, including sediment color and consis-tency, were recorded. Sediment was collected from the bucket and placed into1-L wide mouth amber-glass bottles for later use.

2.3. Stock solution preparation

A 4% Hg solution was prepared by dissolving 5.41 g of HgCl2 in a 100 mLvolumetric flask from solid HgCl2 (Sigma Aldrich, 499.5%). After complete

dissolution, the Hg stock solution was transferred to a 4 oz glass jar. A 1% sulfatesolution was prepared by dissolving 14.85 g of Na2SO4 (J.T. Baker,499.0%) in a 1 Lvolumetric flask (10 g-SO4/L). After dissolution, the sulfate stock solution wastransferred to a 1-L high density polyethylene (HDPE) container. A 10% sodiumlactate solution was prepared by adding 32.1 mL of sodium lactate syrup (FisherScientific, 60.0%) in a 250 mL volumetric container. After complete mixing, thesolution was transferred to a 250 mL HDPE container.

2.4. SRB inoculum preparation

D. desulfuricans were grown in a four component medium (a modification of(Alico and Liegey, 1966)): (1) 400 mL aqueous solution of 2.0 g magnesium sulfate(Sigma Aldrich, 499.5%), 5.0 g sodium citrate (Sigma Aldrich, 499.5%), 1.0 gcalcium sulfate (Sigma-Aldrich, 499.9%), and 1.0 g ammonium chloride (FisherScientific, 499.0%); (2) 200 mL aqueous solution of 500 mg dibasic potassiumphosphate (Sigma Aldrich, 499.0%); (3) 400 mL of an aqueous solution of 3.5 gsodium lactate syrup (Fisher Scientific, 60%), and 1.0 g of yeast extract (AcrosOrganics); and (4) 50 mL aqueous solution of ferrous ammonium sulfate (J.T. Baker,5% v/v). After complete mixing/dissolution, all solutions were separately vacuum-filtered through a 0.45 mm filter and transferred to separate glass jars. The pH ofeach component was separately adjusted to 7.5 in their respective glass jars:components 1 and 3 required the addition of approximately 1 mL of 0.1 M NaOH(Fisher Scientific) while component 2 required approximately 1 mL of 0.5 M HCl(Fisher Scientific). After pH adjustment, the solutions were autoclaved at 121 1C.After autoclaving, components 1, 2 and 3 were mixed in the anaerobic glove boxafter which 40 mL of component 4 was added to the mixture. After the brothsolution had been thoroughly mixed, 5 mL of the composite broth solution wastransferred to a 40 mL vial which would serve as an incubation chamber.An ampoule of freeze-dried D. desulfuricans (American Type Culture Collection,Manassas, VA) was opened and 0.5 mL of broth was added into the ampoule towash the bacterial culture into the incubation chamber. Each incubation chamberwas covered with aluminum foil and allowed to incubate for 48 h at 38 1C. Afterincubation, the broth/SRB mixture had attained a black hue indicating successfulgrowth of D. desulfuricans. The incubation vial was transferred to the anaerobicglove box and added to 50 mL of broth solution in a 1-L glass bottle after which thebottle was covered with aluminum foil and parafilm and the composite broth/SRBmixture was allowed to incubate for 48 h at 38 1C. After the second incubation theSRB solution was ready for use.

2.5. Microcosm bottle preparation

The microcosm bottles were prepared in an anaerobic environment in theanaerobic glove box. Thirty-three (33) grams of sediment from the previouslyprepared 1-L amber jars was added into the bottles using a thin-stemmed spatula(yielding 10 g sediment on a dry weight basis). Following the addition of sediment,90 mL of N2 (Praxair, 99.9% purity) sparged tap water was added to the microcosmbottles. Using a micro-pipette the bottles were then spiked with the requiredamounts of sodium lactate (final aqueous concentrations of either 0, 250, or1,000 mg/L as sodium lactate), Hg (final concentrations of either 0 or 150 mg/L asHg), SRB (either 0 or 3 mL of incubated inoculums) and sulfate (either 0, 48, or192 mg/L added sulfate) from the pre-prepared stock solutions. It is important

Table 2Experimental matrix for microcosm experiments.

MicrocosmID

Description SRBaddition

HgCl2spike(mg/L)

Lactateaddition(mg/L)

SO4

addition(mg/L)

1 No Hg spike, noSRB, no Lac, noSO4

No No None None

2 Hg spike, no SRB,no Lac, no SO4

No Yes None None

3 Hg spike, SRB, noLac, no SO4

Yes Yes None None

4 Hg spike, SRB,low Lac, no SO4

Yes Yes Low None

5 Hg spike, SRB,high Lac, no SO4

Yes Yes High None

6 Hg spike, SRB, noLac, low SO4 yes

Yes None Low

7 Hg spike, SRB, noLac, low SO4

Yes Yes None Low

8 Hg spike, SRB, noLac, high SO4

Yes Yes None High

9 Hg spike, SRB,low Lac, low SO4

Yes Yes Low Low

10 Hg spike, SRB,low Lac, high SO4

Yes Yes Low High

11 Hg spike, SRB,high Lac, low SO4

Yes Yes High Low

12 Hg spike, SRB,high Lac, high SO4

Yes Yes High High

Test bottles were prepared in duplicate and at seven time points (0, 2, 7, 10, 14, 28and 42 days) Microcosm 1 was a control with no amendments added.

P.M. Randall et al. / Environmental Research 125 (2013) 30–40 33

to point out that the 0/48/192 mg/L sulfate conditions refer to added sulfate, andthat the higher total sulfate concentrations are expected because of sulfate in thesediment porewater, sulfate on the lake sediment which may slowly desorb intothe aqueous phase, and sulfate present in the tap water used to prepare theincubations. The actual final concentrations of sulfate were determined by analyz-ing aliquots of the sediment slurry, and it was determined that the 0/48/192 mg/Ladded sulfate solutions were actually 42 mg/L, 149 mg/L, and 396 mg/L sulfate,respectively. However, throughout this paper, reference to the sulfate incubationswill be made according to the quantity of added sulfate (i.e., no added sulfate,48 mg/L, or 192 mg/L). Table 2 shows the experimental matrix. The glass incubationchamber was occasionally vortexed throughout the SRB addition to ensurea bacterial suspension in the glass incubation chamber. After addition of all ofthe required constituents, a glass pipette was used to fill the remaining volume upto 100 mL and the serum bottles were capped with a rubber stopper, sealed witha crimp cap and gently shaken by hand to mix. The bottles were kept at roomtemperature in the absence of light until they were sacrificed at thepre-determined times.

All Hg and mercury-containing waste was properly disposed of as hazardous waste.

2.6. pH/ORP/DO measurements

After the serum bottle was incubated for the appropriate time, the pH,oxidation reduction potential (ORP) and dissolved oxygen (DO) were measured.All measurements were conducted in the anaerobic glove box. Approximately 2 mLto 3 mL of water was extracted from the microcosm bottles with a 10 mL syringeand added into a clean 40 mL vial before electrochemical probes were used tomeasure pH, ORP and DO.

2.7. qPCR measurements

Two microcosm bottles (those sacrificed on day 0 and 42) were sub-sampledfor qPCR analysis by Microbial Insights (Rockford, TN). The samples were analyzedwithin 24 h of shipment, making the total incubation time 96 h to 120 h beforeqPCR. Samples were extracted using Power Soil deoxyribonucleic acid (DNA) kits(Mobio Laboratories, Solana Beach, CA) following the manufacturer's recommenda-tions. Real time PCR was then performed on each sample using target oligonucleo-tide probes (Suzuki et al., 2000). The gene that was amplified was the 16 S RNA ofthe δ-proteobacteria. For Taqman-based assays, one of the probes contained6-carboxy-fluorescein (6-FAM) as a reporter fluorochrome on the 5′ end, and N,N,N′,N′-tetramethyl-6-carboxy-rhodamine (TAMRA) as a quencher on the 3′ end. Eachreaction contained 1X TaqMan Universal PCR Master Mix (Applied Biosystems),

forward primer, reverse primer, Taqman probe and a DNA template of the M13gene specific for the extracted samples. The PCR conditions were as follows: 2 minat 50 1C and 10 min at 95 1C, followed by 50 cycles at 15 s/cycle at 95 1C and 1 minat 58 1C. The PCR reaction was carried out in an ABI Prism 7300 Sequence DetectionSystem (Applied Biosystems, Foster City, CA) (for details see (Stults et al., 2001) and(Harms et al., 2003)). For sybr green-based assays, real-time PCR was performed onan ABI 7000 Sequence Detection System (Applied Biosystems, Foster City, CA). PCRmixtures contained 1X Cloned Pfu Buffer (Stratagene, LaJolla, CA), 0.2 mM of eachof the four deoxynucleoside triphosphates, sybr green (diluted 1:30,000; MolecularProbes, Eugene, OR), and 1 U of Pfu Turbo HotStart DNA polymerase (Stratagene,LaJolla, CA). Annealing temperatures, primer concentrations, and MgCl2 concentra-tions varied depending on the primer set used (Hales et al., 1996; Greene et al.,2003). A calibration curve was obtained by using a serial dilution of a knownconcentration of positive control DNA.

2.8. Organic acids

Next, a sub-sample of the overlying water from each of the microcosm bottlesthat were spiked with lactate was extracted, filtered through a 0.25 mm Teflons

filter, and transferred to a 1.5 mL centrifuge tube. The centrifuge tube and samplewere then sent to Clemson University Department of Environmental Engineeringand Earth Sciences at the L.G. Rich Environmental Research Laboratory for organicacid analysis by liquid chromatography. Lactate, propionate, and acetate wereanalyzed on a Waters 600E high-performance liquid chromatography (HPLC)system composed of an autosampler (Waters 717plus), pumping system (Waters600), and a UV/Vis detector (Waters Model 490E) targeting 210 nm. Samples wereinjected through a sample loop (50 mL) onto an Aminexs HPX-87H ion exclusioncolumn (300 mm�7.8 mm; BioRad). Eluent (0.01 N H2SO4) was delivered at 0.6 mL/min. Five calibration standards were prepared ranging from 0 to 1, 500 mg/L. Ifsample concentrations were above the calibration curve, samples were diluted via a2X dilution and analyzed again.

2.9. Total mercury and methylmercury

After sub-samples were taken for pH/ORP/DO, qPCR and organic acid analysis,the entire microcosm bottles with remaining samples were re-sealed and sent toBattelle Sequim for both solid and aqueous Hg and MeHg analysis. Sedimentsamples were analyzed for total Hg as per U.S. EPA Method SW7471A (reportinglimit 0.04 mg/kg); aqueous samples were analyzed for Hg as per U.S. EPA MethodSW7470A (reporting limit: 0.200 mg/L).

Preparation of sediments for MeHg analysis was based on a modification of U.S.EPA Method 1630 (distillation, aqueous ethylation, purge and trap, desorption, andcold vapor atomic fluorescence spectrometry). Sediment samples were preparedbased on Bloom (1989). Aqueous samples were analyzed for MeHg as per U.S. EPAMethod 1630 without modification (reporting limits: 0.320 ng/g and 0.154 ng/L).

2.10. Sulfate

A sub-sample of the aqueous phase was filtered through a 0.45 mm Teflons

filter, placed in a glass vial and sent to DHL Analytical (Round Rock, TX) for sulfateanalysis. The sample was analyzed directly as per U.S. EPA Method 300 via ionchromatography using a Dionex Model DX-120 chromatograph using an IonPaccolumn with an ion-suppressed conductivity detector.

2.11. Dissolved organic carbon

A sub-sample of the aqueous phase from each microcosm bottle was filteredthrough a 0.45 mm Teflons

filter, placed in a glass vial and sent to DHL Analytical(Round Rock, TX) for dissolved organic carbon (DOC) analysis. The sample wasprepared by preserving a portion of sample with phosphoric acid and thenanalyzed as per U.S. EPA Method 415.1 via a Tekmar/Dorhmann Phoenix 8000instrument using air stripping of inorganic carbon followed by ultraviolet-catalyzedpersulfate oxidation of organic carbon and then followed by infrared detection/quantification of carbon.

2.12. Dissolved gases (methane and carbon dioxide)

A sub-sample of each microcosm bottle was placed in a glass vial and sent toDHL Analytical Laboratory (Round Rock, TX). The sample was then analyzedfollowing DHL's DGAS standard operating procedure by the RSKERR method.

3. Results and discussion

Bench-scale microcosm experiments were designed to providea better understanding of the Hg methylation potential

P.M. Randall et al. / Environmental Research 125 (2013) 30–4034

in anthropogenically-contaminated natural sediment environ-ments as a function of sulfate concentration, carbon substrateconcentration (lactate), the presence/absence of an Hg spike, andthe presence/absence of SRB. Levels of Hg (150 mg/L as Hg) for thisstudy were chosen to represent severely contaminated environ-ments. This analysis builds upon the emerging body of literaturedescribing Hg fate and transformation in sedimentary environ-ments, and expands upon the present scientific knowledge byextending the experimental conditions to higher levels of Hgcontamination than areas impacted exclusively through atmo-spheric deposition. All analyses presented in the figures thatfollow are based upon duplicate microcosm incubations, and theerror bars represent the standard deviation of the duplicates.

3.1. Aqueous Hg concentrations

The aqueous and sediment phases of each microcosm wereanalyzed for soluble and sediment-bound Hg at each of the timepoints during the experiment (0, 2, 7, 10, 14, 28 and 42 days).Results shown in Fig. 1 are from incubations amended with SRBand Hg (150 mg/L as HgCl2). The effects of experimental sulfateconcentrations on aqueous Hg concentrations are presented withno added lactate and 250 mg/L lactate. The low concentration ofaqueous Hg is most likely due to the fact that the lake sedimentsmay behave as a sorbent for inorganic Hg, and reducing conditionscause the reduction of sulfate to sulfide which will precipitate Hgas HgS. These are two explanations for the small amounts ofaqueous Hg compared to the Hg spike (mg/L levels after incubationcompared to mg/L levels before incubation). These figures presentaqueous phase Hg concentrations and do not reflect the total Hg

Fig. 1. Temporal evolution of total aqueous Hg in microcosms with zero addedlactate, 250 mg/L added lactate and varying amounts of added sulfate (“LowSulfate”¼48 mg/L added sulfate; “High Sulfate”¼192 mg/L added sulfate) demethy-lation and sorption of MeHg to sediments attained within 42 days.

concentration in the microcosms. The aqueous Hg concentrationsreported in Fig. 1(0.5 mg/L to 2 mg/L) are four orders of magnitudeless than the initial Hg spike of 150 mg/L. Because this discrepancyfar exceeds the MeHg produced in these systems, it is inferred thatthe majority of the Hg removed from the aqueous phase wasadsorbed onto the sediment or precipitated (e.g., as HgS).

Regardless of lactate addition, aqueous Hg concentrationdecreases rapidly (within two weeks) in the presence of addedsulfate (48 mg/L and 192 mg/L). Fig. 1 also presents the results ofaqueous phase Hg concentration in the presence of SRB and250 mg/L lactate at different experimental levels of sulfate.The lake sediment used for this study is moderately organic andcontains sufficient organic substrate to encourage the growth andmetabolism of SRB along with concurrent Hg methylation. There-fore, these microcosm incubations represent an environment thatis conducive to SRB growth (i.e., anoxic conditions with availableelectron sources). All systems displayed similar total aqueous Hgconcentrations regardless of sulfate addition. This is most likelydue to the precipitation of HgS under the reducing conditions hereor is a consequence of the high levels of sulfate already present inthe lake sediment porewater (202 mg/L); given the expectedheterogeneity of the sediments, it is difficult to speculate as tothe role of sediments as a source of sulfate.

3.2. Aqueous MeHg production

Fig. 2 presents aqueous MeHg concentrations at each timepoint for three microcosm experiments. In the background experi-ment, lake sediment was allowed to incubate without any amend-ments. The “Hg Spike” microcosm received a spike of Hg (150 mg/Las HgCl2) and was allowed to incubate in the absence of addedSRB. This microcosm was designed to evaluate the contribution ofthe native bacteria in Hg methylation. The third microcosm,“Hg Spike & SRB,” contained lake sediment, a Hg spike (150 mg/Las HgCl2) and added SRB. None of the microcosms presented inFig. 2 contained added sulfate; thus, the sulfate level in theseincubations is the amount in the porewaters of the lake sediment.Comparing the extent of Hg methylation in these three incuba-tions indicates that much of the native Hg is present in anoccluded form and not readily bioavailable, and most likelyassociated with the organic fraction of the sediment. These resultsindicate that the native consortium of bacteria and/or abioticmethylation processes are capable of methylating Hg, but not tothe extent of the microcosm with addition of both Hg and SRB(o12 ng/L MeHg observed over the 42-day experiment for the“Hg Spike” microcosm compared to 430 ng/L MeHg for the “Hg &

Fig. 2. Temporal evolution of aqueous MeHg production in lake sediments, withaddition of only HgCl2, and with addition of HgCl2 and SRB (without lactate).

Fig. 4. Temporal evolution of aqueous and sediment-associated MeHg in a systemwith added Hg spike (150 mg/L as Hg) and added SRB.

Table 3Kd values calculated at each time point from the“Background” system.

Time point (day) Kd (L/g)

2 5.157 5.12

10 11.6714 5.2528 5.4642 8.02

(For comparison, Rothenberg et al. (2008) reportedKd values of 1.3 L/g).

P.M. Randall et al. / Environmental Research 125 (2013) 30–40 35

SRB” microcosm). Abiotic methylation processes, for example byreaction of Hg(II) with acetate, humic or fulvic acid complexes,methylcobalt, methyltin, or reaction of Hg(0) with methyl iodide,are generally insignificant compared to biotic methylation pro-cesses, especially in freshwaters where conditions for SRB growthand metabolism exist. Previous studies provide a review of abioticmethylation processes and conditions in which they may besignificant compared to biological Hg methylation (Gardfeldtet al., 2003; Celo et al., 2006). These results provide significanceof microbial Hg methylation, and the fundamental importance ofboth a bioavailable source of Hg and active SRB for MeHgproduction.

At 14 days, significant differences in aqueous MeHg concentra-tions were observed between the system that received SRB andthe system that did not. The presence of high sulfate concentra-tions in the lake sediment along with an excess of aqueous Hg andadded SRB contributed to a favorable environment for the methy-lation of Hg. After Day 14, the aqueous MeHg concentrationdecreased in the “Hg Spike & SRB” microcosm to a level similarto that of the “Hg Spike” microcosm at the end of the experiment.In these microcosms, this may have occurred for the followingreasons: (1) physical/chemical adsorption of MeHg to sedimentsfollowing methylation; (2) bacterial demethylation of MeHg(although this route is generally thought to be insignificantcompared to abiotic demethylation (Celo et al., 2006); and (3)abiotic demethylation processes. This decline is important as theremay be an unidentified sink for MeHg which out paces MeHgproduction after 14 days, while sorption kinetics to sedimentaccounts for the steady increase in sediment associated MeHg.It is possible that continuously precipitated HgS is responsible fora Hg (or potentially, MeHg) sink that eventually reaches a criticalpoint at 14 days, beyond which Hg becomes the limiting reagentand MeHg production can no longer keep pace with sorption ofMeHg to HgS (Xiong et al., 2009). Both demethylation andadsorption may have contributed to the observation that both“Hg & SRB” and “Hg Spike” microcosms contained the sameaqueous MeHg concentration at the end of the experiment. Thismay indicate an equilibrium condition for methylation, demethy-lation and sorption of MeHg to sediments attained within 42 days.

3.3. Sediment MeHg production

Sediment-associated MeHg was also measured for the “Back-ground”, “Hg Spike” and “Hg Spike & SRB” microcosms to deter-mine the relative contribution of sediment-associated MeHg to thetotal MeHg in these experiments (Figs. 3 and 4). Both aqueous andsediment-associated MeHg concentrations increased during the42 day incubation in the “Background” incubation (Fig. 3), with

Fig. 3. Time series of MeHg concentration on sediments and in the aqueous phasefor the “Background” condition.

similar methylation dynamics in both aqueous and sedimentphases. Fig. 3 illustrates that MeHg equilibrium between aqueousphase and sediment is fast and Kd values calculated at each timepoint are given in Table 3. These data suggest two plausiblescenarios describing Hg dynamics during the microcosm study.In the first scenario, Hg is preferentially methylated in the aqueousphase and subsequently adsorbed to the organic components ofthe sediment. Assuming MeHg adsorption is relatively rapid, eachtime point can then be estimated as an equilibrium condition.A partition coefficient (Kd) for this sediment can be calculatedusing the aqueous and sediment-associated MeHg concentrationsat each time point during the period of increasing concentrationsusing the ratio of the concentration of MeHg in solution to theconcentration of MeHg measured on sediments. This approachproduces Kd values ranging from 3.7 L/g to 4.7 L/g. These values aresimilar to recently determined Kd values measured by Rothenberget al. (2008) for MeHg sorption onto a California freshwater lakesediment of 1.3 L/g. The second plausible scenario to describe thedata in Fig. 3 is that Hg is first being adsorbed to sediments andthen methylated on the sediment surface. Under this scenario,equilibrium between aqueous and sediment-bound Hg is estab-lished first, followed by methylation on the sediment surface. This,however, is a diffusion/bacterial motility limited phenomena andis slower than methylation in the aqueous phase.

Fig. 4 shows the temporal evolution of sediment-associatedMeHg concentrations in the systems receiving an Hg spike(Hg Spike) and those receiving both an Hg spike and addition ofSRB (Hg Spike & SRB) in the absence of added lactate or sulfate. Boththe “Hg Spike” and “Hg Spike & SRB” microcosms show a steadyincrease in sediment-associated MeHg from Day 0 to Day 14, afterwhich the sediment MeHg concentration approaches an asymptoticvalue. Both of these systems contain an Hg spike; however, thesystem that did not receive SRB displays an equivalent increase in

Fig. 5. Temporal evolution of sediment-associated MeHg in three systems:background, Hg spike, and Hg spike with SRB.

-5

-4

-3

-2

-1

0

1

2

5 5.5 6 6.5 7 7.5 8 8.5 9pH

pe

Day 0Day 2Day 7Day 10Day 14Day 28Day 42

SO 42-

H2SHS -

Fe 3+

Fe 2+

Fig. 6. pe-pH predominance area diagram for all microcosm incubations.

Fig. 7. Temporal evolution of aqueous sulfate concentration under differingexperimental conditions (No lactate or sulfate was added to these systems).

Fig. 8. Temporal evolution of aqueous sulfate concentration under different lactateconcentrations at an initial sulfate concentration of 192 mg/L added sulfate.

Table 4Stoichiometric relationship for lactate oxidation, sulfate reduction and methyl-mercury production in the high lactate, high sulfate with Hg and SRB spikenormalized to methylmercury production.

SO4 MeHg Lactate

1.7E+07 1 2.7E+081.8E+07 1 7.0E+061.1E+06 1 3.0E+055.2E+04 1 −2.6E+06a

−2.9E+04 1 −9.8E+053.9E+04 1 2.4E+05

a Negative values for sulfate reduction and lactate oxidation indicate anincrease in sulfate or lactate concentration at that time point.

P.M. Randall et al. / Environmental Research 125 (2013) 30–4036

sediment-associated MeHg. This indicates that native bacteria arecapable of methylating the experimentally amended Hg, but thenbecomes limited by available C-substrate or sulfate after 14 days.Fig. 5 illustrates the effect of added Hg (Hg Spike) and added Hg andSRB (Hg Spike & SRB) on the production of sediment associatedMeHg. Clearly the indigenous Hg within the lake sediments is notbioavailable as addition of inorganic Hg leads to production ofMeHg, with a maximum being reached after only 14 days.

3.4. Sulfate reduction

It is assumed that SRB methylation of inorganic Hg is the resultof the passive diffusion of Hg through the cell membrane (Benoitet al., 2002) during sulfate reduction. Fig. 6, depicting the pE–pHrelationship of the microcosm incubations, illustrates that Fe(II)/Fe(III) is the dominant redox couple for these sediments. Althoughexperimentally measured ORP values indicate that the backgroundredox status of these sediments is dominated by iron redoxdynamics, subsequent figures demonstrate that sulfate reductionis rapid in these incubations.

Fig. 7 shows the temporal dynamics of native sulfate within the“Background,” “Hg Spike” and “Hg Spike & SRB” microcosms (noneof the microcosms had added sulfate or lactate). There is littledifference between the “Background” and “Hg Spike” microcosms;both display a slight decrease in sulfate concentration over the42-day experiment. However, this reduction is less than 30% of theinitial concentration and is likely a result of sulfate reduction bynative SRB, given that abiotic reduction is unlikely under the redoxpotentials measured in these incubations. In contrast to the“Background” and “Hg Spike” microcosms in Fig. 7, the “Hg Spike& SRB” microcosm displayed a drastic reduction in sulfate overthe first 7 days and then a gradual reduction to negligible

concentrations at Day 14. This trend is further illustrated inFig. 8, in which three systems with added Hg and SRB and192 mg/L added sulfate were evaluated for sulfate reduction withdifferent C-substrate additions (0, 250 mg/L and 1000 mg/L lac-tate). These microcosms indicate that sulfate reduction and con-current Hg methylation is limited by the availability of labileC-substrate. Contrasting the “Hg Spike & SRB” system of Fig. 7(where sulfate reduction was completed after 14 days) to the “HgSpike & SRB High Lactate” system of Fig. 8 (where sulfate reductionwas completed after less than 2 days), it is clear that the presence ofa C-substrate (electron donor) enhances sulfate reduction in thesesystems and therefore likely also enhances Hg methylation. Thisfinding is consistent with the hypothesis that, in the presence oflabile Hg and excess sulfate, Hg methylation is C-substrate limited.

Fig. 10. Dissolved methane production during the 42-day incubation.

P.M. Randall et al. / Environmental Research 125 (2013) 30–40 37

The stoichiometry of carbon oxidation, sulfate reduction, andMeHg production in the aqueous phase has been calculated foreach time point for the high lactate, high sulfate, Hg and SRB spikesystem (Table 4). The results of the stoichiometric calculation arenormalized to MeHg production. The results illustrate that severalorders of magnitude greater amounts of sulfate and lactate mustbe processed compared to MeHg production. Partitioning of MeHgto the sediments from the aqueous phase at least partiallyaccounts for these large stoichiometric ratios for MeHg productionand sulfate. As such, these stoichiometric relationships do notrepresent the entire system (sediment and water) nor do theyinclude other sinks for MeHg (such as precipitated HgS or organicmatter). Nevertheless, these ratios indicate that large quantities ofsulfate and lactate must be processed by the microbial communityfor a relatively small production of MeHg.

3.5. Metabolic byproducts and potential competitionby methanogens

3.5.1. Dissolved gas metabolic byproductsDissolved CO2 and CH4 were measured as proxy for SRB and

methanogenic metabolic activity. CO2 concentrations are stronglycontrolled by lactate addition (Fig. 9), indicating that SRB meta-bolism is stimulated by the addition of a C-substrate in thesesystems. The native bacteria in the “Background” groups producednegligible quantities of CO2, while bacteria within SRB-amendedmicrocosms increased dissolved CO2 concentrations by a factor ofthree throughout the incubation period (average of the “no lactate”incubations). As lactate concentrations are increased, dissolvedCO2 concentrations increase regardless of sulfate addition, under-scoring the importance of available C-substrate for SRB metabo-lism in this system.

Dissolved CH4 was measured at low levels throughout the 42-dayincubation in the “Background” groups and SRB-amended microcosmswith no added lactate (Fig. 10). Methanogens are unable to competewith SRB in the incubations without excess labile C-substrate. CH4

increased steadily throughout the incubation period in the “LowLactate” microcosms beginning at Day 10 after methanogens hadconverted the byproducts of lactate metabolism by SRB directly to CH4.Dissolved CH4 production increased substantially in the “High Lactate”microcosms as the systems became sulfate-limited. This observation issupported by the results of the “High Lactate” system with no addedsulfate, where methanogens were observed to compete with SRB andproduced very high levels of CH4 (�3000mg/L). Adding more sulfateto the microcosms decreased CH4 production, presumably fromincreased SRB growth, thereby decreasing effective competition frommethanogens.

Fig. 9. Carbon dioxide production during the 42-day incubation.

3.5.2. High lactate microcosmsIt was shown that sulfate was completely reduced within 14 days

in microcosms containing SRB and adequate C-substrate (Figs. 7and 8). After SRB exhausted the supply of available sulfate in themicrocosms, they began to slow their metabolism. During this periodof slow growth and metabolism, SRB may be out-competed by otherbacteria for the remaining carbon metabolic substrate (especially inhigh lactate concentrations, representing an excess of electron donor)(Gottschalk, 1986). SRB metabolize lactate, transforming it to acetateand eventually CO2. It is important to note here that not all SRB canmetabolize lactate, however D. desulfuricans is known to be able tometabolize lactate (Liamleam and Annachhatre, 2007). However,during periods of metabolic distress for SRB, such as depleted sulfateconditions, slower-growing methanogens may successfully competewith SRB for available lactate. Propionate may also be used as anelectron donor for SRB, albeit not as efficiently as lactate or acetate;however, propionate is only formed after the depletion of sulfate,leading to the conclusion that depletion of propionate is due to the factthat SRB have switched to syntrophic fermentation to metabolizepropionate. Fig. 11 shows the temporal evolution of lactate, acetate andpropionate in microcosms with an added Hg spike, SRB and thehighest experimental concentration of lactate (1000 mg/L) at threedifferent levels of added sulfate (0 mg/L, 48 mg/L and 192mg/L).Within the first 7 days of the microcosm experiment, lactate wasdepleted concurrently with sulfate (see Fig. 8), suggesting a metabolicrelationship between the electron donor (lactate) and electron accep-tor (sulfate). In all systems, lactate was oxidized, sulfate was reducedand MeHg was produced.

Acetate was formed in all three incubations, an indicator of thecollective respiration of the native microbial consortia present inthe sediment. Acetate is rapidly metabolized by microbes within30 days at all sulfate levels. Differing sulfate amendments had nosignificant effect on the metabolism of acetate, indicating possiblecompetition from methanogens, which can quickly and easilymetabolize acetate in the absence of sulfate. After the initialreduction of sulfate (after Day 7, Fig. 8), SRB slowed theirmetabolism as they switched to fermentation and methanogensbecame the dominant species within the microcosm, evidenced byescalating CH4 production after Day 14 (Fig. 10).

3.5.3. Low lactate microcosmsFig. 12 shows the temporal evolution of lactate, acetate and

propionate, respectively, in microcosms with a Hg spike, SRB, anda low concentration of added lactate (250 mg/L) at three differentsulfate treatments (0 mg/L, 48 mg/L and 192mg/L). These microcosmsare identical to those in Fig. 11, but were amended with lower initiallactate concentrations (250 mg/L). In these microcosms, all lactate was

PROPIONATE

050

100150200

250300350400450

0 10 20 30 40Time (d)

[Pro

] mg/

L

No SulfateLow SulfateHigh Sulfate

LACTATE

0

100

200

300

400

500

600

700

800

0 10 20 30 40time (d)

[Lac

] mg/

L

No SulfateLow SulfateHigh Sulfate

Fig. 11. Temporal evolution of lactate, acetate and propionate in microcosms with anadded Hg spike, SRB and the highest experimental concentration of lactate (1000 mg/L)at three different levels of added sulfate (0 mg/L, 48 mg/L and 192mg/L).

Fig. 12. Temporal evolution of lactate, acetate and propionate in microcosms witha Hg spike, SRB, and a low concentration of added lactate (250 mg/L) at threedifferent sulfate treatments (0 mg/L, 48 mg/L and 192 mg/L).

P.M. Randall et al. / Environmental Research 125 (2013) 30–4038

consumed within 7 days in the high sulfate and No Added Sulfatemicrocosms, and within 2 days in the low sulfate microcosm. Thisillustrates that the SRB are able to more quickly metabolize lactate assulfate concentrations are increased. As expected, acetate was formedfrom the metabolism of lactate and is most prominent in the highsulfate microcosm where SRB metabolism is assumed to be highest.Propionate is formed most likely due to fermentative conditionsformed in the presence of high lactate (Gottschalk, 1986).

3.5.4. Evidence for methanogenic competitionPropionate is produced in greater quantities in the high lactate

microcosms, which may be attributed to the presence of compe-titive methanogens that thrive when C-substrate is in excess andconverted to acetate. Since SRB passively methylate Hg duringrespiration, the availability of C-substrate for SRB metabolism hasa relevant impact on MeHg production in aquatic sediments.

Any change in chemical factors promoting the metabolicactivity of SRB will typically lead to greater MeHg productionand thus greater measured aquatic MeHg levels. Figs. 13 and 14show the temporal evolution of aqueous MeHg in microcosmsreceiving a Hg spike, SRB, and differing sulfate concentrations atboth high and low lactate levels. Initial sulfate levels are shown tohave an impact on aqueous MeHg concentrations under highlactate conditions (1000 mg/L) (Fig. 13). The aqueous MeHg con-centrations were similar under the low lactate condition at allexperimental sulfate concentrations (0 mg/L, 48 mg/L and 192 mg/L)under the low lactate condition (Fig. 14). This indicates that SRBmetabolism depended on available C-substrate and not the level ofsulfate present. In these microcosms, SRB were provided with anexcess of C-substrate and, hence, metabolism became sulfate-limitedand more MeHg was produced in the microcosms with highersulfate concentrations.

Increasing the lactate level four-fold only increased the aqu-eous MeHg concentration by about 30% (Figs. 13 and 14), which

HIGH LACTATE

0

5

10

15

20

25

30

35

40

45

50

0 10 20 30 40time (d)

[MeH

g] (n

g/L)

No SulfateLow SulfateHigh Sulfate

Fig. 13. Aqueous MeHg concentrations under different experimental sulfateconcentrations with high (1000 mg/L) added lactate.

LOW LACTATE

0

5

10

15

20

25

30

35

40

45

50

0 10 20 30 40time (d)

[MeH

g] (n

g/L)

No SulfateLow SulfateHigh Sulfate

Fig. 14. Aqueous MeHg concentrations under different experimental sulfateconcentrations with low (250 mg/L) added lactate.

P.M. Randall et al. / Environmental Research 125 (2013) 30–40 39

is consistent with the non-stoichiometric relationship betweensulfate and MeHg reported by Gilmour et al. (1992). One explana-tion for this relationship is that SRB metabolism cannot utilizethe additional lactate due to competition from other microbes.Alternatively, aqueous MeHg concentrations may be limited bypartitioning to the sediment phase as the system approachesequilibrium and demethylation occurs. This latter explanation ismore likely, as CO2 respiration still increases under high lactateconditions compared to low lactate conditions. It has also beenrecently observed that spontaneous formation of HgS particles canlead to reduced bioavailability of Hg for methylation (Deonarineand Hsu-Kim, 2009). In order to properly evaluate the potential fornet aqueous MeHg production, it is important to determine thesorption characteristics of MeHg to site-specific sediments undersite-specific water chemistry conditions.

4. Summary and conclusions

The results of this study attempt to answer the question ofwhat conditions of carbon substrate (lactate) and electron accep-tor (sulfate) cause the greatest potential for methylation of Hg.To do so, 11 experimental microcosms were incubated for 42 daysto better elucidate factors controlling aquatic Hg methylation incontaminated sediments. The methylation process was examined asa function of sulfate concentration, carbon substrate concentration

(lactate), the presence/absence of an aqueous inorganic Hg spike,and the presence/absence of SRB. The presence of Hg in sedimentsis not an adequate indicator of active Hg methylation. MeHg isformed through a process of bio-methylation whereby inorganic Hgis methylated as a reaction ancillary to normal enzymatic processes.In this study, the methylation process is predominately controlledby the carbon substrate and sulfate available to SRB and bycompetition from methanogens for that carbon substrate.

Aqueous Hg concentrations were found to decrease quicklywith the addition of low or high sulfate additions, presumably as aresult of the sorption of Hg to the lake sediments and precipitationof HgS in these sulfate-rich sediments. Additionally, higher overallSRB metabolic activity (added lactate and added sulfate) wasobserved to decrease aqueous phase Hg concentration. AqueousMeHg was not readily produced by native microbes with orwithout the Hg spike. However, in the SRB-amended microcosmscontaining an Hg spike, up to 32 ng/L MeHg (compared to 1.2 ng/Lin the “Background” and 11 ng/L in the Background+Hg spike) wasproduced by Day 14 of the incubation without the addition ofsulfate or carbon substrate. This maximum rate of MeHg produc-tion is approximately 2.13�10−5% of the total Hg added to thesystem in the aqueous phase. This is far less than the resultsreported by Duran et al. (2008) who reported a Hg conversion of0.3% and less still to the findings of Monperrus et al. (2007) andco-workers reported a Hg conversion in the aqueous phase of 6.3%.The sediments used in this study may account for a large sink ofMeHg with fast sorption kinetics and, therefore, MeHg producedin the aqueous phase is expected to adsorb to the sediments andreduce the MeHg accumulation in the aqueous phase. Oh et al.(2010) and co-workers reported a MeHg production of 0.67 ng/g-DW on sediments after 2 years while studying surface sedimentsin a freshwater lake in Korea. Compared to the results presented inthis report MeHg production on sediments after 42 days (140 ng/g-DW) this number is far smaller. However, the fact that Oh et al.(2010) studied surface sediments is significant: biotic MeHgproduction is much greater in anaerobic environments than oxicenvironments which explain this discrepancy. Taking into accountthe sediment/water surface area in these experiments(2.58�10−3 m2), MeHg fluxes to sediment within this 42 dayexperiment were 59 ng/m2/d1. This value is larger than thatreported by Holmes and Lean (2006)(reporting a flux of 10.2 ng/m2/d in wetland sediments) and much larger than that reported byHollweg et al. (2009) and co-workers (reporting a flux of 0.215 ng/m2/d in Chesapeake Bay sediments). Again, this flux can beexplained by the strong affinity of MeHg for the sediments inthe current study compared to those in the studies by Holmes andLean (2006) and Hollweg et al. (2009) which had much lowerorganic carbon fractions. This MeHg then decreased through theduration of the incubation as the lactate was utilized by themicrobes. In the absence of adequate metabolic fuel (sulfate,carbon), MeHg levels are observed to decrease on the time scaleof days to weeks.

Hg methylation by SRB is limited by the depletion of sulfate andcarbon, and appears to be sensitive to competition by methano-gens for carbon substrate. SRB utilize lactate, transforming it toacetate, and eventually CO2. However, in the presence of excesslactate fermentation results in propionate production, and even-tually conversion to CO2 and CH4. It was found that in a highlactate environment, all lactate was utilized in the microcosmswithin seven days, despite the amount of sulfate added, and allacetate was consumed by Day 42. However, propionate producedby competing methanogens persists through the incubation.Microcosms with SRB addition demonstrated drastic sulfate reduc-tions after only 7 days. Once there is a shortage of sulfate in themicrocosm and SRB begin to slow their metabolism, methanogensmay begin to effectively compete for remaining labile C-substrate.

P.M. Randall et al. / Environmental Research 125 (2013) 30–4040

These experiments show that MeHg concentration peaks at Day 14within the microcosms, presumably as methanogens begin tobetter compete with SRB for available carbon substrate.The depletion of sulfate and carbon substrate and the competitionfor carbon by methanogens most likely limit overall Hg methyla-tion by SRB. Significant MeHg production is possible in bothsediments and the aqueous phase when conditions of adequatecarbon substrate, electron donor, electron acceptor and SRB arecollectively present. The implications of this from a sedimentremediation standpoint are that capping materials should: (1) beable to sequester MeHg produced in sediments and (2) notcontribute to the production of MeHg. Further research underwayin the laboratory will attempt to answer whether a specificcapping material will meet these two goals.

Acknowledgments

We would like to thank the anonymous reviewers who reviewedthe manuscript and made suggestions for its improvement. The U.S.Environmental Protection Agency through its Office of Research andDevelopment funded the research described here. This research hasnot been subjected to Agency review and therefore does not necessa-rily reflect the views of the Agency. Mention of trade names andproducts should not be interpreted as conveying official EPA approval,endorsement, or recommendation.

References

Alico, R.K., Liegey, F.W., 1966. Growth of Desulfovibrio desulfuricans under hetero-trophic and anaerobic conditions. J. Bacteriol. 91 1122-1114.

Benoit, J.M., Gilmour, C.C., Heyes, A., Mason, R.P., Miller, C.L., 2002. Geochemical andBiological Controls over Methylmercury Production and Degradation in AquaticEcosystems. Biogeochemistry of Environmentally Important Trace Elements.American Chemical Society pp. 262–297.

Bloom, N., 1989. Determination of picogram levels of methylmercury by aqueousphase ethylation, followed by cryogenic gas-chromatography with cold vaporatomic fluorescence detection. Can. J. Fish. Aquat. Sci. 46, 1131–1140.

Celo, V., Lean, D.R.S., Scott, S.L., 2006. Abiotic methylation of mercury in the aquaticenvironment. Sci. Total Environ. 368, 126–137.

Choi, S.C., Chase, T., Bartha, R., 1994. Metabolic pathways leading to mercurymethylation in Desulfovibrio-desulfuricans LS. Appl. Environ. Microbiol. 60,4072–4077.

Compeau, G.C., Bartha, R., 1985. Sulfate-reducing bacteria—principal methylators ofmercury in anoxic estuarine sediment. Appl. Environ. Microbiol. 50, 498–502.

Deonarine, A., Hsu-Kim, H., 2009. Precipitation of mercuric sulfide nanoparticles inNOM-containing water: implications for the natural environment. Environ. Sci.Technol. 43, 2368–2373.

Drott, A., Lambertsson, L., Bjorn, E., Skyllberg, U., 2007. Importance of dissolvedneutral mercury sulfides for methyl mercury production in contaminatedsediments. Environ. Sci. Technol. 41, 2270–2276.

Duran, R., Ranchou-Peyruse, M., Menuet, V., Monperrus, M., Bareille, G., Goni, M.S.,Salvado, J.C., Amouroux, D., Guyoneaud, R., Donard, O.F.X., Caumette, P., 2008.Mercury methylation by a microbial community from sediments of the AdourEstuary (Bay of Biscay, France). Environ. Pollut. 156, 951–958.

Fleming, E.J., Mack, E.E., Green, P.G., Nelson, D.C., 2006. Mercury methylation fromunexpected sources: molybdate-inhibited freshwater sediments and an iron-reducing bacterium. Appl. Environ. Microbiol. 72, 457–464.

Freedman, D.L., Cashwell, J.M., Kim, B.J., 2002. Biotransformation of explosive-gradenitrocellulose under denitrifying and sulfidogenic conditions. Waste Manage.22, 283–292.

Gardfeldt, K., Munthe, J., Stromberg, D., Lindqvist, O., 2003. A kinetic study on theabiotic methylation of divalent mercury in the aqueous phase. Sci. TotalEnviron. 304, 127–136.

Gilmour, C., Riedel, G.S., Ederington, M.C., Bell, J.T., Gill, G.A., Stordal, M.C., 1998.Methylmercury concentrations and production rates across a trophic gradientin the northern Everglades. Biogeochemistry 40, 327–345.

Gilmour, C.C., Henry, E.A., 1991. Mercury methylation in aquatic systems affected byacid deposition. Environ. Pollut. 71, 131–169.

Gilmour, C.C., Henry, E.A., 1992. Mercury Methylation By Sulfate-Reducing Bacteria—Biogeochemical and Pure Culture Studies. Abstracts of Papers of the Amer-ican Chemical Society. 203, 140-GEOC.

Gilmour, C.C., Henry, E.A., Mitchell, R., 1992. Sulfate stimulation of mercurymethylation in dresh-water sediments. Environ. Sci. Technol. 26, 2281–2287.

Gottschalk, G., 1986. Bacterial Metabolism, second ed. Springer-Verlag New YorkInc. New York, New York.

Greene, E.A., Hubert, C., Nemati, M., Jenneman, G.E., Voordouw, G., 2003. Nitritereductase activity of sulphate-reducing bacteria prevents their inhibition bynitrate-reducing, sulphide-oxidizing bacteria. Environ. Microbiol. 5, 607–617.

Hales, B.A., Edwards, C., Ritchie, D.A., Hall, G., Pickup, R.W., Saunders, J.R., 1996.Isolation and identification of methanogen-specific DNA from blanket bog featby PCR amplification and sequence analysis. Appl. Environ. Microbiol. 62,668–675.

Harms, G., Layton, A.C., Dionisi, H.M., Gregory, I.R., Garrett, V.M., Hawkins, S.A.,Robinson, K.G., Sayler, G.S., 2003. Real-time PCR quantification of nitrifyingbacteria in a municipal wastewater treatment plant. Environ. Sci. Technol. 37,343–351.

Hollweg, T.A., Gilmour, C.C., Mason, R.P., 2009. Methylmercury production insediments of Chesapeake Bay and the mid-Atlantic continental margin. Mar.Chem. 114, 86–101.

Holmes, J., Lean, D., 2006. Factors that influence methylmercury flux rates fromwetland sediments. Sci. Total Environ. 368, 306–319.

Hosokawa, Y., 1993. Remediation work for mercury Contaminated Bay—experi-ences of Minamata Bay Project, Japan. Water Sci. Technol. 28, 339–348.

King, J.K., Saunders, F.M., Lee, R.F., Jahnke, R.A., 1999. Coupling mercury methylationrates to sulfate reduction rates in marine sediments. Environ. Toxicol. Chem. 18,1362–1369.

King, J.K., Kostka, J.E., Frischer, M.E., Saunders, F.M., 2000. Sulfate-reducing bacteriamethylate mercury at variable rates in pure culture and in marine sediments.Appl. Environ. Microbiol. 66, 2430–2437.

King, J.K., Harmon, S.M., Fu, T.T., Gladden, J.B., 2002. Mercury removal, methylmer-cury formation, and sulfate-reducing bacteria profiles in wetland mesocosms.Chemosphere 46, 859–870.

Liamleam, W., Annachhatre, A.P., 2007. Electron donors for biological sulfatereduction. Biotechnol. Adv. 25, 452–463.

Miletto, M., Bodelier, P.L.E., Laanbroek, H.J., 2007. Improved PCR-DGGE for highresolution diversity screening of complex sulfate-reducing prokaryoticcommunities in soils and sediments. J. Microbiol. Methods 70, 103–111.

Monperrus, M., Tessier, E., Amouroux, D., Leynaert, A., Huonnic, P., Donard, O.F.X.,2007. Mercury methylation, demethylation and reduction rates in coastal andmarine surface waters of the Mediterranean Sea. Mar. Chem. 107, 49–63.

Oh, S., Kim, M.-K., Yi, S.-M., Zoh, K.-D., 2010. Distributions of total mercury andmethylmercury in surface sediments and fishes in Lake Shihwa, Korea. Sci. TotalEnviron. 408, 1059–1068.

Pak, K.R., Bartha, R., 1998a. Mercury methylation and demethylation in anoxic lakesediments and by strictly anaerobic bacteria. Appl. Environ. Microbiol. 64,1013–1017.

Pak, K.R., Bartha, R., 1998b. Mercury methylation by interspecies hydrogen andacetate transfer between sulfidogens and methanogens. Appl. Environ. Micro-biol. 64, 1987–1990.

Rooney-Varga, J.N., Genthner, B.R.S., Devereux, R., Willis, S.G., Friedman, S.D., Hines, M.E.,1998. Phylogenetic and physiological diversity of sulphate-reducing bacteria isolatedfrom a salt marsh sediment. Syst. Appl. Microbiol. 21, 557–568.

Rothenberg, S.E., Ambrose, R.F., Jay, J.A., 2008. Mercury cycling in surface water,pore water and sediments of Mugu Lagoon, CA, USA. Environ. Pollut. 154,32–45.

Stults, J.R., Snoeyenbos-West, O., Methe, B., Lovley, D.R., Chandler, D.P., 2001.Application of the 5′ fluorogenic exonuclease assay (TaqMan) for quantitativeribosomal DNA and rRNA analysis in sediments. Appl. Environ. Microbiol. 67,2781–2789.

Suzuki, M.T., Taylor, L.T., DeLong, E.F., 2000. Quantitative analysis of small-subunitrRNA genes in mixed microbial populations via 5′-nuclease assays. Appl.Environ. Microbiol. 66, 4605–4614.

Ullrich, S.M., Ilyushchenko, M.A., Kamberov, I.M., Tanton, T.W., 2007. Mercurycontamination in the vicinity of a derelict chlor-alkali plant. Part I: Sedimentand water contamination of Lake Balkyldak and the River Irtysh. Sci. TotalEnviron. 381, 1–16.

Warner, K.A., Roden, E.E., Bonzongo, J.-C., 2003. Microbial mercury transformationin anoxic freshwater sediments under iron-reducing and other electron-accepting conditions. Environ. Sci. Technol. 37, 2159–2165.

Whalin, L., Kim, E.-H., Mason, R., 2007. Factors influencing the oxidation, reduction,methylation and demethylation of mercury species in coastal waters. Mar.Chem. 107, 278–294.

Xiong, Z., He, F., Zhao, D., Barnett, M.O., 2009. Immobilization of mercury insediment using stabilized iron sulfide nanoparticles. Water Res. 43, 5171–5179.