Article pubs.acs.org/est
Temperature and the Sulfur Cycle Control Monomethylmercury Cycling in High Arctic Coastal Marine Sediments from Allen Bay, Nunavut, Canada K. A. St. Pierre,† J. Chétélat,‡ E. Yumvihoze,† and A. J. Poulain†,* †
Department of Biology, University of Ottawa, 30 Marie-Curie, Ottawa, Ontario, K1N 6N5, Canada Environment Canada, National Wildlife Research Centre, Ottawa, Ontario, Canada
‡
S Supporting Information *
ABSTRACT: Monomethylmercury (MMHg) is a neurotoxin of concern in the Canadian Arctic due to its tendency to bioaccumulate and the importance of fish and wildlife in the Inuit diet. In lakes and wetlands, microbial sediment communities are integral to the cycling of MMHg; however, the role of Arctic marine sediments is poorly understood. With projected warming, the effect of temperature on the production and degradation of MMHg in Arctic environments also remains unclear. We examined MMHg dynamics across a temperature gradient (4, 12, 24 °C) in marine sediments collected in Allen Bay, Nunavut. Slurries were spiked with stable mercury isotopes and amended with specific microbial stimulants and inhibitors, and subsampled over 12 days. Maximal methylation and demethylation potentials were low, ranging from below detection to 1.13 pmol g−1 h−1 and 0.02 pmol g−1 h−1, respectively, suggesting that sediments are likely not an important source of MMHg to overlying water. Our results suggest that warming may result in an increase in Hg methylation - controlled by temperature-dependent sulfate reduction, without a compensatory increase in demethylation. This study highlights the need for further research into the role of high Arctic marine sediments and climate on the Arctic marine MMHg budget.
■
responsible for methylation,12 recent work has shown that iron-reducing microbes (FeRM) and methanogens contribute to these processes and to the net accumulation of MMHg in the environment.13−15 Although both methylation and demethylation have been shown to occur quickly in freshwaters, in marine environments the rate of MMHg degradation is much reduced (10−6 s−1 in freshwater) as a result of increased stability conferred by elevated chloride concentrations.5 In freshwater environments, sediments play an important role in the cycling of MMHg, particularly in MMHg production, which necessitates reducing conditions. It has been hypothesized6,8 that the sedimentary contribution to the marine Arctic MMHg pool is negligible, with the majority of methylation activity occurring in the water column; however, little work has been done to determine whether this is the case. Indeed, there exists no known estimate of marine sediment Hg levels or methylation rates for the Arctic Archipelago.16
INTRODUCTION Monomethylmercury (MMHg) is a potent neurotoxin found in aquatic ecosystems worldwide with a tendency to bioaccumulate and biomagnify.1 In the Canadian Arctic, high MMHg levels pose a risk to the health of Inuit people whose traditional local diet is mostly comprised of marine mammals and fish.2 Despite the predominance of marine species in the Inuit diet, knowledge of MMHg dynamics in Arctic coastal systems is sorely lacking. Our current understanding has been derived from previous work focused on freshwater systems in the Arctic and/or coastal systems at temperate latitudes.3−5 Sources of MMHg in marine Arctic environments include both in situ production and external loading via atmospheric deposition, snowmelt following atmospheric depletion events and river and ocean current transport.6−8 Analysis of lake sediment cores suggests that mercury (Hg) deposition in the Arctic has increased during the last century and is being retained within the environment.9 Both MMHg production and degradation can occur via abiotic (mostly photochemical) and/or biotic pathways.10 Parks et. al (2013)11 recently identified two genes involved in Hg methylation, proposing a corrinoid protein-mediated mechanism common to all methylating bacteria. While sulfatereducing bacteria (SRB) are thought to be principally © 2014 American Chemical Society
Received: Revised: Accepted: Published: 2680
August 30, 2013 January 14, 2014 January 21, 2014 January 21, 2014 dx.doi.org/10.1021/es405253g | Environ. Sci. Technol. 2014, 48, 2680−2687
Environmental Science & Technology
Article
Table 1. Description of Microcosms and Experimental Set-Upa Hg Spikes, Stimulants and Inhibitors ID N1 C1 A1 T1 T2 T3 T4
description natural, unspiked natural, spiked abiotic control, spiked SRB inhibited MPA inhibited SRB and MPA inhibited, FeRM stimulated MPA inhibited, SRB stimulated
Hg spikes yes yes yes yes yes yes
Na2MoO4
BES
Fe(III)
yes yes yes
yes
Na2SO4
yes yes
yes
All microcosms were inoculated in duplicate. Na2MoO4, sodium molybdate; BES, sodium-2-bromoethane sulfonate; Fe(III), ferrihydrite − Fe(OH)3; Na2SO4, sodium sulfate; FeRM, iron-reducing prokaryotes; MPA, methanogens; SRB, sulfate-reducing bacteria. a
Microcosm Preparations. In the laboratory, all sediment manipulations and incubations were conducted under a nitrogen atmosphere in a Shel Lab Bactron anaerobic chamber. The sediments (0−2 cm depth) were thawed at 4 °C and homogenized. Approximately 75 g of sediments were transferred to 250 mL dark glass jars with Teflon-lined lids. 100 mL anoxic artificial seawater21 was then added and the resulting slurry shaken before spike, inhibitor and stimulant additions. Abiotic controls (A1) were prepared 48 h prior to the other microcosms by adding formaldehyde to a final concentration of 1%.22 Being volatile, formaldehyde is easily removed from the slurry before the start of the incubations. Formaldehyde incubations did not alter the pH of the slurry. All treatments (Table 1) were conducted in duplicate, reflecting the high cost of Arctic sample collection and the need to ensure that enough sediments be used in preparing each microcosm for MMHg extraction from each subsample. The agreement between duplicates in treatments where the trends observed were significant (e.g., C1 and T2, see Results and Discussion below) lends credibility to the use of duplicates in this case. The microcosms were incubated at 4, 12, and 24 °C in parallel, but staggered by 3 days to allow for immediate analysis of slurry chemistry following each subsampling. Temperatures were specifically chosen to represent current temperatures (4 °C), a possible future water temperature scenario given current climate projections (12 °C) and a much higher temperature (24 °C) to ensure that a temperature effect (if present) was detectable. Specific inhibitors and stimulants were used to assess the relative contributions of the microbial groups shown to play a major role in mercury methylation and demethylation: sulfatereducing bacteria (SRB), iron-reducing microbes (FeRM), and methanogens (MPA).12,13,15,23 Sodium molybdate (Na2MoO4) was used as an inhibitor and sodium sulfate (Na2SO4) as a stimulant of SRB.12 Although Na2MoO4 has been shown to inhibit some methanogens,24 its use is still widely accepted in SRB inhibition experiments.25,26 MPA were inhibited using sodium-2-bromoethane sulfonate (BES). Both MPA and SRB require hydrogen (H2), so inhibitors are required to definitively attribute methylation and demethylation activity in the system to either group. Despite there being no known inhibitor of microbial iron reduction, the simultaneous use of Na2MoO4, BES, and iron(III) oxyhydroxide theoretically promotes the iron reduction pathway.12,13 In freshwater sediments, amorphous iron(III) oxyhydroxide has been shown to be the most bioavailable form of iron for FeRM.27 Stock solutions of Na2MoO4 (1 M), Na2SO4 (1 M), and BES (0.5 M) were prepared in Milli-Q water. Amorphous ferrihydrite (Fe(OH)3) was synthesized according to Schwert-
Potentially compounding the effect of increases in Hg deposition is the effect of climate change, expected to be particularly pronounced in these regions with even conservative models projecting continued warming for at least the next century.17 While an increase in the rate of Hg methylation in pure cultures has been correlated with increased temperatures,18 few studies have examined the direct effect, if any, of temperature on the rates of Hg methylation and MMHg degradation in marine coastal Arctic environments. We investigrated the effects of temperature on mercury methylation and demethylation potentials in sediments collected from Allen Bay, Nunavut, Canada. Sediment slurries were incubated for two weeks with isotope-enriched mercury tracers and specific microbial inhibitors and stimulants. The objectives of the study were (a) to determine Hg methylation and demethylation potentials in marine Arctic sediments; (b) to assess the relative contribution of abiotic processes, SRB, FeRM, and methanogens to Hg methylation and demethylation; and (c) to establish the effect of temperature on overall potentials of de/ methylation in the sediments and the relative contributions of the microbial guilds of interest under different temperature regimes.
■
MATERIALS AND METHODS Site Description and Sample Collection. Sediment samples were collected from a tidal pool in Allen Bay (Cornwallis Island), Nunavut, Canada (74°47′N, 95°18′W) on August 12, 2011. The maximum depth of the pool sampled at low tide was around 1.0 m. At this depth, sediments would be susceptible to natural seasonal changes in air temperatures. This study site is representative of organic-matter poor, icecompacted coastal sediments of the Arctic.19,20 Over the last 10 years, average summer (July−August) temperatures recorded at the weather station in Resolute ranged, from 1 to 12.7 °C with extreme temperatures ranging from −3 to 20.1 °C. Sediments were collected using a long-bladed shovel, transferred to a sterile (autoclaved) HDPE bottle in the field without a headspace, and stored in a cooler. The same day, upon returning to the research facilities in Resolute Bay, the sediments were frozen until further analysis. Frozen sediments were transported to facilities in Ottawa where they were kept at −20 °C until the incubation experiments were setup. Freezing the sediments may alter the chemistry and the microbial communities present in the sediments; however, because of the shallow nature of the overlying water at this site, the sediments likely freeze naturally during the winter. Ambient dissolved total mercury and MMHG in the overlying water were 0.43 ng L−1 and 0.05 ng L−1, respectively (Supporting Information (SI) Table S1). 2681
dx.doi.org/10.1021/es405253g | Environ. Sci. Technol. 2014, 48, 2680−2687
Environmental Science & Technology
Article
mann and Cornell (2000).28 Inhibitors and stimulants were added prior to the Hg-tracers.25 With the exception of Na 2 MoO4 and Na 2 SO 4 additions (final concentrations following Compeau and Bartha 12), amendments were made based on ambient concentrations of the compounds of interest within the sediments (Table 2).
using the ferrozine32 and Cline’s methods,33,34 respectively. All analyses were conducted using a Varian Cary 300 spectrophotometer. Due to the time sensitive nature of the analyses, the time required for probe stabilization, and in an effort to limit microcosm exposure to light, pH and Eh were only measured regularly in the microcosms incubated at 24 °C. pH was measured in all microcosms with a Double Junction pHTestr 30 (Oakton Instruments), while Eh was measured using a VWR (Eh) SP21 m in treatments N1, T1 and T3. pH was not measured during the final subsamplings (t = 192 and 288 h) due to insufficient slurry volume in which to immerse the probe. MMHg Isotope Analyses. Subsamples for MMHg analysis were prepared according to the protocol outlined by Avramescu et al. (2010).35 Samples were acidified with 0.5 M nitric acid during thawing and the solution was separated from the sediment by centrifugation at 3000 rpm for 15 min. Between 1.2 and 4.5 g of sediments were used for analysis, depending on the quantity of sediments remaining after slurry removal. MM201HgCl was added as an internal standard to correct for procedural recovery during isotope dilution quantification.25 Concentrations of each of the isotope-enriched tracers were determined following gas chromatograph separation using inductively coupled plasma mass spectrometry (GC-ICPMS; GC, Agilent Technologies 7890x; ICPMS, Agilent Technologies 7700x).29 Four isotopes were measured: 198Hg (demethylation tracer), 199Hg (methylation tracer), 201Hg (internal standard), and 202Hg (ambient). The limit of detection for quantifying excess MM199Hg and MM198Hg, calculated according to Hintelmann and Evans,36 were 0.008 pmol.g−1 and 0.005 pmol.g−1, respectively. Statistical Analyses. All statistical analyses were completed in R.37 Owing to the length of the study, concentrations of all compounds of interest (S2−‑, Fe(II), MMHg) oscillated over time, potentially reflecting the effect of opposing processes (i.e., production/degradation; oxidation/reduction; etc.). In consequence, first order kinetics could not be assumed in all cases30 or corrected for6 and the rate constant approach was not representative of the dynamics of the system (e.g., Figure 1 and SI Figures S2 and S3). Data were visually assessed for linearity and divided into linear segments based on the dynamics
Table 2. Ambient Sediment Characteristics and Changes to 199 Hg2+ and CH3198HgCl Spikes and Specific Inhibitor Additions for Incubations parameter
Allen Bay
organic content water content ambient total mercury ambient methylmercury 199 HgII CH3198HgCl Na2MoO4 amendment (SRB inhibitor) Na2SO4 addition (SRB stimulant) Na-BES amendment (MPA inhibitor) Fe(OH)3 amendment (FeRM stimulant)
2.56 ± 0.31% 22.62 ± 1.68% 1.04 ± 0.10 ng/gwet (1.34 ng/gdry) 0.12 ± 0.01 ng/gwet (0.15 ng/gdry) 50 ng/gdry 7.5 ng/gdry 20 mMa 62.04 mMb 30 mMa 0.41 mMb
a
Compeau and Bartha 1985. bYu et al. 2012.
MMHg enriched in Hg isotope 198 (CH3198HgCl) was prepared by the methylcobalamin method29 and purified using anhydrous sodium sulfate. Fresh inorganic Hg (199HgCl2) solutions were prepared at the beginning of each incubation. Hintelmann et al. (2000)30 suggested Hg additions between 13 and 70% of ambient total mercury (THg) and MMHg; however, because ambient Hg concentrations were so low in Allen Bay sediments (1.34 ng THg g−1dry (6.7 pmol g−1) and 0.15 ng MMHg g−1dry (0.6 pmol.g−1); Table 2), isotopeenriched MMHg yields obtained during preliminary experiments (results not shown) using the proposed upper range of amendments were too low to be reliably detected by GCICPMS. Consequently, THg and MMHg concentrations were increased to 50 ng g−1 (240 pmol·g−1) and 7.5 ng g−1 (33.5 pmol g−1), respectively, yielding Hg concentrations still well within the natural levels found in the Arctic Ocean basin.31 Following Hg spikes, inhibitor and stimulant additions, approximately 1−2 mL of slurry was removed from each jar and transferred to a 125 mL glass serum bottle to measure methane production by methanogens in each treatment. Enough anoxic artificial seawater was then added to each serum bottle to fill to two-thirds of its volume and the bottle was crimp-sealed. Serum bottles were incubated for the duration of the incubations. At the end of the experiment, methane concentrations were subsequently determined using a SRI 8610C gas chromatograph as a way to test for the efficacy of the methanogenesis inhibitor used. Subsamples for both slurry geochemistry and mercury isotope analyses were removed at various times (0, 24, 48, 96, 144, (170 at 12 °C), 192, 288 h), transferred to sterile centrifuge tubes and frozen at −80 °C until further analysis. Sediment water content was measured by drying the sediments for 24 h at 105 °C, while organic carbon content (%) was determined using loss on ignition (LOI) (400 °C for 8 h).25 Aqueous Analyses. Subsamples were centrifuged, and the overlying water was transferred to clean syringes fitted with 0.2 μm filters (Sarstedt) under a nitrogen atmosphere. Ferrous iron (Fe(II)) and sulphide (S2‑) concentrations were determined
Figure 1. Production of MM199Hg+ (pmol g−1) over time in the methanogenesis-inhibited treatment (T2) at 4, 12, 24 °C. 2682
dx.doi.org/10.1021/es405253g | Environ. Sci. Technol. 2014, 48, 2680−2687
Environmental Science & Technology
Article
12 °C between 170 and 192 h (Figure 1). This finding is in contrast to previous studies, during which the majority of methylation activities occurred within the first 48 h of isotope additions.30 These systems were not monitored for as long as was done here, so it is unclear whether this difference in the timing of maximum methylation is a characteristic of the Allen Bay system, simply a reflection of the differences in monitoring times between studies, or the time needed for the systems to equilibrate following stimulant and inhibitor additions. The delay in methylation relative to previous studies may further reflect a dependence on other biogeochemical processes, such as sulfate reduction (discussed below). Maximum demethylation potentials were not affected by temperature (F = 1.49, p≫0.05), but were dependent on treatment (F = 4.66, p ≪ 0.05) (Figure 3). Tukey multiple
(increasing or decreasing concentrations) observed. Simple linear regression was used to calculate maximum sulphide production, iron reduction, Hg methylation and demethylation potentials for each microcosm on each linear segment. Additional information on linear segment division can be found in the SI. The maximum slope among the segments for a given microcosm was then chosen to represent the potential of the microcosm with respect to each of the biogeochemical processes of interest. Two-way analyses of variance (ANOVA) with Tukey multiple comparisons were used to evaluate differences in sulphide and iron production rates, and Hg methylation and demethylation potentials between microcosms and across temperature treatments.
■
RESULTS AND DISCUSSION Hg Methylation and MMHg Demethylation Potentials. Both methylation and demethylation activities were low within the sediments (SI Table S2). Methylation potentials ranged from the detection limit (0.008 pmol.g−1) to 1.12 pmol MM199Hg formed g−1 h−1. Demethylation potentials ranged from the detection limit (0.005 pmol.g−1) to 26.3 fmol MM198Hg lost g−1 h−1, 2 orders of magnitude less than the upper range of observed methylation potentials. The overall low methylation potentials from sediment slurries support previous studies6,38 which suggested that water column methylation was responsible for the greatest proportion of MMHg in polar marine waters. Concentrations of MMHg were variable over time, likely reflecting the length of the study and presumably time for both MMHg production and degradation. In the absence of significant sulfate reduction (i.e., all treatments except C1 and T2), variability in the potentials measured was associated with the uncertainty of the method at concentrations close to or below our detection limit. Note that all experiments performed at 4 °C yielded de/methylation potentials below our detection limits (0.008 pmol.g−1 for methylation and 0.005 pmol.g−1 for demethylation) (SI Table S2). Maximum methylation potentials differed significantly by temperature (F = 54.13, p ≪ 0.05) and treatment (F = 22.48, p ≪ 0.05) (Figure 2). Methylation potentials were highest in C1 (natural control) and T2 (MPA inhibited; SRB dominated) at
Figure 3. Maximum demethylation potentials in pmol g−1 h−1. Means (n = 2 replicated experiments) and error bars (±1 SD) are shown. Treatments are described in Table 1
comparisons revealed that the only significant difference was between the unspiked control (N1) and the iron-amended microcosm (T3), the microcosm with the greatest demethylation. Most demethylation occurred between 24 and 48 h after spike additions. Controls on MMHg in Arctic Marine Sediments. In the absence of significant organic matter (Table 2), sulfur chemistry appears to be the primary chemical control on MMHg concentrations in the Allen Bay sediments. Maximum sulphide production rates (SPR) ranged between 0.076 and 26.74 μM.h−1. SPR varied by temperature (F = 43.90; p < 0.05), treatment (F = 4.95; p < 0.05) and by treatment over temperature (interaction; F = 5.58; p < 0.05) (Figure 4). There was no significant difference between SPR at 4 and 12 °C but both differed significantly from 24 °C (p < 0.05), the temperature at which highest sulphide concentrations were observed. At 24 °C in C1 and T2 (where SRB dominated and MM199Hg production was highest at 12 °C), MM199Hg concentrations were highly correlated (R2 = 0.97 and 0.87, respectively; p ≪ 0.05) with sulphide concentrations (SI Figure S2, panels A and B). Based purely on metabolism, we would expect methylation rates to increase at higher temperatures, mirroring the increase in sulfate reduction; however, the highest methylation potentials were observed at 12 °C, not 24 °C. The correlation between sulphides and MMHg was not significant
Figure 2. Maximum methylation potentials in pmol g−1 h−1. Means (n = 2 replicated experiments) and error bars (±1 SD) are shown. Treatments are described in Table 1 2683
dx.doi.org/10.1021/es405253g | Environ. Sci. Technol. 2014, 48, 2680−2687
Environmental Science & Technology
Article
through the production of sulphides which chemically reduce Fe(III), sequestering Fe(II) in the form of iron sulphide minerals, or directly through enzymatic processes.43 This process can also affect Hg availability through the dissolution of adsorbed Hg(II) or Hg-sulphide complexes, thereby increasing the dissolved Hg(II) available for methylation.44 Although our results suggest the FeRM are not involved in the direct methylation of Hg, highest methylation potentials were measured in those treatments where both SRB and FeRM were active (C1, T2). FeRM may indirectly influence methylation by SRB as Fe(III) reduction can oxidize excess S2− to sulfate, thereby prolonging sulfate reduction and methylation.45 pH remained circumneutral over the course of the incubations but increased over time from 6.6 ± 0.1 to 7.4 ± 0.1 (not including the abiotic control, A1). The increase in pH likely reflects the consumption of H+ during microbial metabolism.46 The entire incubations were carried out under reducing conditions (−163.7 mV at t = 0 h). Eh decreased over time in the three treatments monitored. The change was however not uniform, with a much greater difference in the natural unspiked control (Δ = 253.1 mV) when compared to T1 and T3 (mean Δ = 47.4 mV). This potentially reflects the fact that sulfate reduction, a proton consuming process, was inhibited in both treatments, but not in the natural control. Both Eh and pH were well within the range over which sulfatereducing bacteria are known to occur.40 Factors controlling biotic methylation and demethylation ultimately reflect (1) the bioavailability of Hg, and/or (2) the activity of the microbes involved.47 Recent work showed that the genetic basis for methylation is found sporadically among a diversity of microbial genera.11 Likewise, demethylation activity seems to be widespread among guilds. Although demethylation has often been shown to occur in the presence of methanogens,48 methane production was low in Allen Bay (SI Figure S1), possibly reflecting an organic input limitation. In cases of methanogenesis limitation in marine sediments, FeRM may have the greatest potential for demethylation, as shown here, in the absence of other microbial guilds (discussed below). Photochemical demethylation has been identified as a crucial abiotic sink of MMHg in marine Arctic waters.6 Because the sediment slurries were incubated in the dark, photodemethylation would not have occurred. Despite this, demethylation was still observed (albeit to a lesser extent than in the biotic treatments) in the killed control suggesting the possibility of 1) an alternative abiotic pathway of demethylation, such as reaction with hydrogen sulphide49 or 2) the presence of spore-forming bacteria or another microbial guild resistant to formaldehyde and uncontrolled for by the other treatments. While the use of inhibitors and stimulants allowed for the determination of the methylation and demethylation potentials by each microbial guild, it could not accurately reflect the interactions that may occur between microbes in the natural sample. By comparing C1 to T2, competition between SRBs and methanogens appears to limit the accumulation of MM199Hg relative to when methanogens are inhibited. Likewise, ferrous iron (Fe(II)) production is reduced in the presence of SRB, potentially suggesting competition between SRB and FeRM for carbon sources and/or electron donors.50 The similarity between treatments T2 (methanogens inhibited) and T4 (methanogens inhibited with addition of sulfate) suggests that sulfate does not limit SRB activity within the system. This was not surprising given that sulfate is naturally
Figure 4. Maximum rates of sulphide production (μM h−1) by sediment slurry treatment at three (4, 12, 24 °C) incubation temperatures. Means (n = 2 replicated experiments) and error bars (±1 SD) are shown. Treatments are described in Table 1
at 4 and 12 °C (p > 0.05, for example, SI Figure S2, panel C), suggesting that sulphide concentrations only become high enough to limit MMHg production at higher temperatures. In fact, the increase in MM199Hg observed in C1 and T2 clearly precedes the increase in sulphides (SI Figure S2, panels A and B). At low sulphide levels (
