Mercury (micro)biogeochemistry in polar environments

Abstract

The contamination of polar regions with mercury that is transported as inorganic mercury from lower latitudes has resulted in the accumulation of methylmercury in the food chain of polar environments, risking the health of humans and wildlife. This problem is likely to be particularly severe in coastal marine environments where active cycling occurs. Little is currently known about how mercury is methylated in polar environments. Relating observations on mercury deposition and transport through polar regions to knowledge of the microbiology of cold environments and considering the principles of mercury transformations as have been elucidated in temperate aquatic environments, we propose that in polar regions (1) variable pathways for mercury methylation may exist, (2) mercury bioavailability to microbial transformations may be enhanced, and (3) microbial niches within sea ice are sites where active microorganisms are localized in proximity to high concentrations of mercury. Thus, microbial transformations, and consequently mercury biogeochemistry, in the Arctic and Antarctic are both unique and common to these processes in lower latitudes, and understanding their dynamics is needed for the management of mercury-contaminated polar environments.

mercury, microbial transformations, methyl mercury accumulation, polar regions, contamination, redox transformations

Introduction

Over the last few decades, concerns for the vulnerability of polar regions to organic and inorganic contaminants that originate from lower latitudes have increased. These contaminants may be subjected to direct, long-range atmospheric transport, repeated cycles of evaporation and condensation, or transport via migratory species such as seabirds (Blais et al., 2005; Macdonald et al., 2005), some being chemically transformed in the atmosphere before deposition. Mercury is among the most serious of these contaminants due to its accumulation in polar food chains resulting in health risks to both humans and wildlife (Macdonald et al., 2005). This problem is exacerbated by springtime mercury depletion events (MDEs) in the High Arctic (Schroeder et al., 1998; Lindberg et al., 2002), Subarctic (Dommergue et al., 2003) and the Antarctic (Ebinghaus et al., 2002), which results in rapid and massive deposition of ionic mercury, Hg(II), from the atmosphere. This springtime deposition is thought to be due to the oxidation of atmospheric elemental mercury, Hg(0), the form in which mercury is globally distributed, by reactive halogen radicals from sea-salt aerosols (Ariya et al., 2002a; Lindberg et al., 2002). Furthermore, the highest concentration of mercury ever reported in remote environment, up to 0.82 μg L⁻¹ (approx. 4 nmol L⁻¹), was found during springtime only in direct proximity of Arctic sea leads associated with surface vapour crystals such as frost flowers and surface hoar (Douglas et al., 2005). This corresponds to almost a 1000-fold increase compared to Arctic inland locations (St Louis et al., 2005) or to levels recorded during dark periods (Steffen et al., 2002). Furthermore, our own data suggest that oxidation processes are enhanced in Arctic coastal marine environments compared to inland systems (Poulain A.J. et al. unpublished). Together, these results point to coastal and marine environments as sites of intense mercury cycling and vulnerability to mercury toxicity.

Our concerns with mercury toxicity are focused on the production of the potent neurotoxic compound, methylmercury (MeHg), and its availability to aquatic food chains (Wren, 1986), as the consumption of contaminated fish and shellfish is considered the major route of human exposure to MeHg (Ratcliffe et al., 1996). Methylmercury manifests its toxicity as a variety of symptoms ranging from mild numbness of the extremities to blindness, and in severe cases, death (Clarkson, 1997, 2002). Recent research showing that mercury in predators occupying top levels of the Arctic food chain is almost exclusively methylated (Campbell et al., 2005) and that blood and fatty tissue of native human populations have elevated levels of mercury (Bjerregaard & Hansen, 2000; Van Oostdam et al., 2005; Butler Walker et al., 2006) clearly indicate that the dynamics and impact of mercury contamination in the Arctic are similar to these phenomena in temperate zones of the world. As atmospherically deposited mercury is mostly in its inorganic forms (Schroeder & Munthe, 1998; Raofie & Ariya, 2004; Lin et al., 2006), within-ecosystem transformations must play a critical role in the toxicity and distribution of mercury in polar regions.

In temperate zones, microbial activities critically impact MeHg accumulation by carrying out various transformations. Recent reviews on mercury cycling in the environment (Fitzgerald & Lamborg, 2003; O'Driscoll et al., 2005) and specifically on the role of microorganisms (Barkay et al., 2003, 2005) are available. Here we consider information on the role of microorganisms in the geochemical cycling of mercury in temperate regions together with information available from research on mercury and microbiology in polar environments. We synthesize these sources of information to propose junctures where microorganisms critically affect the geochemical cycle of mercury in polar regions (Fig. 1) and to identify research questions that address gaps in our understanding of how microorganisms modulate the toxicity and mobility of mercury in Arctic and Antarctic environments (Table 1).

1

The biogeochemical cycle of mercury in coastal marine environments in polar regions. The various transfer pathways and transformation reactions are described in the text.

1

Microbial transformations of Hg in polar regions, what is known about them and pending research questions and needs*

Microbial transformation	What is unique about this transformation in the Arctic	Questions/research needs
Methylation	Presence of diMeHg in coastal water (Pongratz & Heumann, 1999). MeHg present in snow and meltwaters. SRB and dsr genes not detected in soil where MeHg is formed (Loseto et al., 2004b).	What are the pathways for methylation? Which organisms methylate mercury in polar regions?
Demethylation	Oxidation of C1 compounds is slow in high latitudes (Hines & Duddleston, 2001). mer gene expression in Arctic biomass (Poulain et al., unpublished). Photoreduction of MeHg in epilimnitic lake water (Hammerschmidt & Fitzgerald, 2006).	What are the pathways for the degradation of MeHg in polar regions?
Hg(II) reduction	Hg-resistant organisms are common in ancient permafrost (Mindlin et al., 2005). High bioavailability in snow (Lindberg et al., 2002). Possible concentration of Hg(II) in proximity to active microbes in sea ice. mer gene expression in Arctic biomass (Poulain et al., in preparation).	Development of psychrophilic Hg biosensors. The interactions of microbes in sea ice with Hg. Measurement of mercury concentrations in the complex sea ice matrix. Further assessment of the evolution of Hg resistance in polar areas.
Hg(0) oxidation	High chloride concentrations in coastal marine environments induce abiotic oxidation of Hg(0).	A better understanding of Hg(0) oxidation in Hg biogeochemistry.

*

See text for details.

Microbial transformations in mercury biogeochemistry in polar environments

Our current view of the role of microorganisms in the cycling of mercury in the environment is based on studies that were initiated by the discovery of the toxicity of MeHg to consumers of contaminated fish and shellfish in the 1960s (Westöö, 1966). Results from environmental, geochemical, microbiological, biochemical and molecular studies have converged to establish our current view of the mercury biogeochemical cycle (Barkay et al., 2003, 2005). Within that paradigm, microorganisms impact the production of MeHg directly by methylation and MeHg degradation, and indirectly by controlling the supply of Hg(II), the substrate for methylation, by carrying out redox transformations that affect transitions between Hg(II) and Hg(0). These transformations and how they are likely to be impacted by the unique conditions of cold environments are discussed below.

Hg(II) methylation

That anaerobic microorganisms methylate mercury has been known for almost 40 years (Jensen & Jernelöv, 1969) and for the last 20 this activity has been attributed to sulfate-reducing bacteria (SRB) in anoxic sediments (Compeau & Bartha, 1985; Gilmour et al., 1992; King et al., 2000). The mechanism of methylation may (Choi et al., 1994) or may not (Ekstrom et al., 2003; Ekstrom & Morel, 2004) be related to the production of acetyl CoA and methylcobalamine (B12). Very recently, however, the possibility of methylation by iron-reducing bacteria has been proposed (Fleming et al., 2006) and awaits further confirmation. The methylation of Hg(II) by abiotic processes (Weber, 1993; Siciliano et al., 2005) may be indirectly related to biological activities because of the dependence of the processes on biological products such as dissolved organic matter. Formation of MeHg in the Arctic has been documented in wetland soils (Loseto et al., 2004b) and streams (Loseto et al., 2004a), in freshwater ponds (St Louis et al., 2005) and lakes and tundra watersheds (Hammerschmidt et al., 2006). In the later study, a mass balance analysis showed that sediment production of MeHg accounted for 80–91% of the whole lake MeHg production. The issue of which organisms methylate Hg(II) in the Arctic has been addressed by Loseto et al. (2004b). Based on a low abundance of SRB in soil samples and a failure to detect Deltaproteobacteria– the major taxonomic group among the bacteria to include SRB – and genes encoding for the disulfite reductase enzyme (Dsr) in DNA that was extracted from the soil microbial biomass, the authors concluded that methylation was not mediated by SRB. We have made similar observations with samples collected from an Adirondack watershed. Further examination of waterlogged soils, however, using the additions of substrates (sulfate) and specific inhibitors (molybdate) as was described by Gilmour et al. (1992) and Compeau and Bartha (Compeau & Bartha, 1985), implicated SRB in Hg(II) methylation, leading us to conclude that the sensitivity of molecular methods was not sufficient to detect SRB whose presence and activities resulted in MeHg production (Barkay T. and Hines M.E, unpublished). Indeed, SRB were readily detected when more sensitive molecular probes (Daly et al., 2000) were applied to the same samples (Yu R. and Barkay T., unpublished). Thus, the involvement of SRB in methylation in the Arctic, especially in sediments of coastal environments where they are likely to carry out the bulk terminal oxidation under anaerobic conditions, remains to be examined. This involvement is supported by observations that SRB are abundant in Arctic coastal marine sediments (Svalbard, Norway) as detected by FISH and dot blot hybridization (Ravenschlag et al., 2001), and that psychrophilic SRB isolated from the same sediments actively reduced sulfate at in situ temperatures (Knoblauch et al., 1999; Bruchert et al., 2001).

Observations on the occurrence and distribution of MeHg suggest that several novel methylation pathways may occur in the Arctic in addition to methylation in sediments, as follows.

1. (1)

   An atmospheric source of MeHg. One of the unique aspects of methylation and MeHg accumulation in coastal Arctic environments suggested by a large flux of MeHg at the initiation of snowmelt (Loseto et al., 2004a; St Louis et al., 2005) and a strongly supported positive correlation between MeHg and chloride in snow pack suggested a marine source. This source could possibly be due to the evasion of dimethylmercury (diMeHg), from leads and polynyas where it may be formed by phytoplankton in the water column [see below and (Pongratz & Heumann, 1999)], and its subsequent atmospheric photodegradation to monomethylmercury chloride (Niki et al., 1983) and deposition.

2. (2)

   Methylation in snow packs. An alternative explanation for the large flux of MeHg during snowmelt. Experiments using bioreporters (Selifonova et al., 1993; Golding et al., 2002) indicated that Hg(II) that is deposited during MDEs is highly bioavailable (Scott, 2001; Lindberg et al., 2002). Moreover, organic compounds [e.g. dicarboxylic acids (Kawamura et al., 1996)] are present in Arctic snow and may serve as a carbon source for microorganisms and as ligands for mercury complexation. Our direct bacterial counts by flow cytometry showed 2 × 10⁵ cells mL⁻¹ of melted snow from the High Canadian Arctic, and melted snow from Antarctica's dry valleys had 200–5000 cells mL⁻¹ (Alfreider et al., 1996; Carpenter et al., 2000; Segawa et al., 2005). Some of these microorganisms were metabolically active as indicated by the reduction of INT, a respiratory indicator (Alfreider et al., 1996) and by low, but detectable, levels of protein and nucleic acid synthesis at in situ temperatures (Carpenter et al., 2000). Microbes in snow, therefore, may methylate Hg(II).

3. (3)

   Atmospheric methylation of Hg. It has been long suggested that aerosols could support life (Gidlen, 1948) and together with recent exciting reports of organic compound transformations in aerosols by bacteria and fungi (Ariya et al., 2002b; Ariya & Amyot, 2004) may point to a possible atmospheric production of MeHg.

4. (4)

   Photomethylation. Recently described in a northern temperate ecosystem (Siciliano et al., 2005) and may be another pathway for methylation in snow where biological processes produce dissolved organic matter (Calace et al., 2005), the catalyst implicated in this process.

5. (5)

   Production of MeHg by phytoplankton in the marine water column. Pongratz and Heumann (Pongratz & Heumann, 1999) observed overlapping chlorophyll and diMeHg optima in-depth profiles from the Antarctic coastal marine environment and reported methylation by isolated pure cultures of phytoplankton. Thus, several niches within the coastal marine environments of polar regions may support production of MeHg, all requiring testing for a full understanding of the processes that result in MeHg production and availability to polar food chains.

Methylmercury degradation

The degradation of MeHg is the other half, next to methylation, of the equation that determines the production of this neurotoxic substance. Three processes, photodegradation (Sellers et al., 1996) and two microbially mediated ones (Schaefer et al., 2004; Barkay et al., 2005), are known for the degradation of MeHg. Photodegradation is the dominant mechanism for demethylation in surface water in many mercury impacted ecosystems and very recent and elegant work showed it to be the sole process for the degradation of MeHg in the eplimnitic water of a highly oligotrophic freshwater Arctic lake (Hammerschmidt & Fitzgerald, 2006). To the best of our knowledge MeHg degradation has not been examined in sediments or samples from coastal marine environments in polar regions.

Microbial pathways for the degradation of MeHg are distinguished by the gaseous carbon products of the degradation process; in oxidative demethylation carbon dioxide is produced and in the reductive process the product is methane. We (Schaefer et al., 2004) and others (Marvin-Dipasquale et al., 2000; Gray et al., 2004) have shown that the choice between these processes is to a large extent controlled by environmental factors. Reductive demethylation, an activity that is mediated by the organomercury lyase enzyme, which is a part of the mercury resistance (mer) system in bacteria (see below), is favored at high redox and at high concentrations of mercury, an effect that we have attributed to the dependence of mer operon gene expression on inducing concentrations of Hg(II) (Schaefer et al., 2004). While the absolute concentrations required for induction depends on the complicated issue of mercury bioavailability, observations of mer gene expression from temperate regions suggested that in lakes and streams with less than nM concentrations of Hg, as observed in most polar waters, expression is repressed (Poulain et al., 2004a; Schaefer et al., 2004). Thus, one would not expect reductive demethylation to be a dominant process in polar regions. However, this expectation is contradicted by a recent observation of mer transcripts in RNA extracts from microbial biomass collected in the Canadian High Arctic (see below).

Oxidative demethylation is favored at low redox and a broad range of mercury concentrations and is most likely related to C1 pathways in anaerobic prokaryotes (Marvin-Dipasquale & Oremland, 1998). The occurrence and rates of C1 metabolism in microorganisms from cold environments has been getting a lot of attention due to anticipated effects of global warming on the release of carbon from large frozen reservoirs in permafrost and polar tundra. While methanogenesis (Berestovskaia et al., 2005) using bicarbonate or acetate as substrates (Rivkina et al., 2004) and methanotrophy (Berestovskaia et al., 2005) were noted, rates were drastically impacted by a drop in the incubation temperature, and degradation of C1 compounds such as methylbromide or acetate, common in temperate soils (Hines et al., 1998), are rarely observed at high latitudes proximal to polar areas (Hines & Duddleston, 2001). Based on these observations, the likelihood for oxidative MeHg degradation in polar regions is low. Nevertheless, demethylation plays an important role in determining MeHg production and availability to food chains, and its occurrence and mechanisms in cold environments need to be addressed.

Redox transformations of inorganic mercury

Redox transformations between the ionic and elemental mercury forms affect MeHg production by controlling the amount of the substrate that is available for methylation (Fitzgerald et al., 1991). Among reduction processes, photoreduction dominates in surface water (Krabbenhoft et al., 1998; Amyot et al., 2004; O'Driscoll et al., 2004; Poulain et al., 2004a; Garcia et al., 2005; Zhang et al., 2006), has recently been discovered in snow (Lalonde et al., 2002, 2003) and is thought to mediate the evasion of most of the Hg(II) that is deposited from the atmosphere onto snow during MDEs (Lindberg et al., 2002; Dommergue et al., 2003). The reduction of Hg(II), nonrelated to the activity of mercury-resistant microorganisms (see below), can be associated to the activity of microorganisms in fresh and salt waters via pathways still to be determined but related to both heterotrophic and/or photosynthetic activity (Ben-Bassat & Mayer, 1978; Mason et al., 1995; Poulain et al., 2004a; Rolfhus & Fitzgerald, 2004). Reduction by the activities of mercury-resistant bacteria impacts the partition of mercury into the gaseous phase in some environments (Barkay, 1987; Barkay et al., 2005). This activity is mediated by the prokaryotic mercuric reductase enzyme (MerA), which is encoded by the merA gene, a part of the inducible mercury resistance (mer) operon (Barkay et al., 2003). This operon is broadly distributed among bacteria (Barkay et al., 2003) and archaea (Simbahan et al., 2005) from diverse environments (Osborn et al., 1997; Vetriani et al., 2005). The description of several mer operons in bacteria, one among them possibly the ancestor of the mer transposon, Tn21 (Kholodii et al., 2003), from 10 000–40 000 year-old Siberian permafrost (Mindlin et al., 2005), and preliminary results showing the presence of merA in bacteria from a 120 000-year-old glacial ice core (Lu-Irving P. and Barkay T., unpublished), suggest that mercury-resistant prokaryotes may be endemic to cold environments. Their role in mercury biogeochemistry in polar regions remains to be examined.

In addition to merA, the mer operon contains several additional genes that actively transport Hg(II) into the cytoplasm as well as those that regulate mer gene expression. Some operons also encode for the organomercury lyase and microorganisms carrying such mer operons reductively degrade MeHg (see above). The regulator of mer expression, MerR, plays a critical role in determining where and under what conditions Hg(II) reduction by MerA occurs. Expression is repressed in the absence of Hg(II) and is quantitatively induced in its presence (Summers, 1992; Brown et al., 2003). Because of this requirement for induction, MerA-mediated reduction has been considered of little relevance to transformations of mercury in natural environments (Morel et al., 1998). Indeed, a series of studies performed in several environments that were impacted by various sources of mercury showed mRNA transcripts of the merA gene in highly contaminated environments, whereas microbial biomass from environments with low levels of contamination contained low to nondetectable levels of these transcripts (Nazaret et al., 1994; Hines et al., 2000; Poulain et al., 2004a; Schaefer et al., 2004). Based on these observations, one would not expect merA expression and microbial reduction of Hg(II) in polar microbial communities where mercury concentrations, during most times of the year, are in the pM range (Steffen et al., 2002; St Louis et al., 2005).

Very recent observations of merA expression in biomass collected from coastal marine environments in the high Arctic during the summer of 2005 (Poulain A.J. et al. unpublished) challenge the paradigm that assigned MerA a minor role in mercury geochemistry (see above). This paradigm has been previously questioned by a report that suggested that MerA activities in protein extracts from lake microbial biomass were positively related to diel cycles of dissolved gaseous mercury [DGM, mostly comprised of Hg(0) in freshwaters, (Vandal et al., 1991)] accumulation in lake water that had pM total mercury concentrations (Siciliano et al., 2002). It is most likely that the concentration of bioavailable mercury, rather than that of total mercury, is what determines mer induction in a given environment (Barkay et al., 1998, 2005). If so, the high levels of bioavailable mercury in snow during atmospheric MDE (Scott, 2001; Lindberg et al., 2002) may explain the observed induction of merA. Development of mercury biosensors in psychrophilic bacteria, similar to the ones that have been employed for the detection and assessment of bioavailable mercury in temperate environments (Barkay et al., 1998), is needed as part of an approach for distinguishing bioavailable from total mercury in polar environments.

An alternative explanation for the induction of merA in high Arctic microbiota relates to the highly heterogeneous nature of microbial habitats in polar regions, most noticeably in sea ice (Thomas & Dieckmann, 2002; Mock & Thomas, 2005), leading to the formation of unique environments where mercury resistance might be essential for survival. The slow rates of transcript degradation previously reported in cold environments (Vlassov et al., 2005) might furthermore explain the detection of merA transcripts in sea ice associated biota. Sea ice, the habitat for most of the microbial biomass in coastal polar environments, may contain niches where both mercury and microorganisms are concentrated. It is likely that mercury, like other solutes in sea ice (Eicken, 2003), is highly concentrated in brine channels where actively metabolizing microorganisms are located (Deming, 2002; Junge et al., 2004). Furthermore, and as stated above, some of the highest concentrations of mercury ever reported in natural samples were encountered in sea ice formations such as frost flowers and surface hoar (Douglas et al., 2005). Our hypothesis on the localized proximity of microorganisms to mercury in brine channels is also supported by the observations that microorganisms in brine channels during winter are associated with particles (Junge et al., 2004), that a significant fraction of atmospherically derived mercury is bound to particles (Schroeder et al., 1998), and that mercury in snow - especially in marine environments - is almost exclusively associated with particles (Poulain A.J. et al., unpublished). Thus, the interaction of mercury with microorganisms in sea ice is likely a spatially fractured phenomenon and may be one of the most exciting aspects of future studies on mercury biogeochemistry in polar environments. Such studies will also bring new perspectives to mercury biogeochemistry at large, as spatial constrains on mercury microbe interactions have not been considered to date, although the chemical physical properties of mercury should clearly drive its heterogeneous distribution in any structured environment.

MerA-mediated reduction of Hg(II) may dramatically affect mercury cycling and thus MeHg production in polar regions. Our preliminary modeling effort suggests that the great majority of elemental Hg is of bacterial origin in arctic marine surface waters and further studies should explore its role at the Arctic scale. The microbial oxidation of Hg(0) to Hg(II) is the part of the mercury biogeochemical cycle about which we know the least. Most research efforts have to-date examined abiotic mechanisms of light and dark oxidation (Lalonde et al., 2001, 2004; Poulain et al., 2004b; Raofie & Ariya, 2004; Sheu & Mason, 2004; Garcia et al., 2005; Whalin & Mason, 2006). Bacterial enzymes known for their role in preventing oxidative damage, such as hydroperxidases, oxidize Hg(0) in organisms that are common in natural waters and soils (Smith et al., 1998). Siciliano et al. (2002), related the levels of mercury oxidases in lake microbial biomass to variations in DGM concentrations. How these microbially mediated oxidative processes affect mercury speciation in polar regions, and especially their impact on the fate of the volatile elemental mercury Hg(0), has not been examined.

Conclusion

The study of mercury biogeochemistry in polar environments is in its early stages, but the synthesis of information available from the study of mercury biogeochemistry in temperate regions in light of what we know about the distribution of mercury in polar regions and about microbiology in cold environments points to the uniqueness of mercury cycling in polar regions. As in temperate environments, MeHg is accumulated by aquatic food chains, but the sites where methylation occurs and the methylation pathways themselves may differ in polar areas from those in lower latitudes. A particularity of Arctic ecosystems resides in the enhanced vulnerability of marine/coastal environments to mercury toxicity, as evidence suggests that these sites are locations where mercury cycling is highly dynamic.

Most intriguingly, polar coastal marine environments are characterized by spatially and temporally fractured environments in term of their physical, chemical and biological features, and therefore are composed of a multitude of microbial niches. These unique niches may alter, or modulate, the pathways of microbial transformations of mercury relative to their characteristics in temperate environments. Specifically, high bioavailability of Hg(II) and the presence of niches where mercury and microorganisms are concentrated may alter the production of MeHg by enhancing bacterial reduction of Hg(II). Thus, our current state of knowledge provides us with a starting point for studies on mercury transformations at the world polar regions, and such studies promise to add new dimensions to our perception of the mechanisms and pathways that determine mercury toxicity and facilitate life in its presence.

Acknowledgements

The authors' research on mercury biogeochemistry is supported by the Environmental Remediation Science Program (ERSP), Biological and Environmental Research (BER), of the US Department of Energy, the National Science Foundation, and the Fond Québecois de la Recherche sur la Nature et les Technologies.

Author notes

Editor: Max Häggblom