Effect of exogenous and endogenous sulfide on the production and the export of methylmercury by sulfate-reducing bacteria

Mercury (Hg) is a global pollutant of environmental and health concern; its methylated form, methylmercury (MeHg), is a potent neurotoxin. Sulfur-containing molecules play a role in MeHg production by microorganisms. While sulfides are considered to limit Hg methylation, sulfate and cysteine were shown to favor this process. However, these two forms can be endogenously converted by microorganisms into sulfide. Here, we explore the effect of sulfide (produced by the cell or supplied exogenously) on Hg methylation. For this purpose, Pseudodesulfovibrio hydrargyri BerOc1 was cultivated in non-sulfidogenic conditions with addition of cysteine and sulfide as well as in sulfidogenic conditions. We report that Hg methylation depends on sulfide concentration in the culture and the sulfides produced by cysteine degradation or sulfate reduction could affect the Hg methylation pattern. Hg methylation was independent of hgcA expression. Interestingly, MeHg production was maximal at 0.1–0.5 mM of sulfides. Besides, a strong positive correlation between MeHg in the extracellular medium and the increase of sulfide concentrations was observed, suggesting a facilitated MeHg export with sulfide and/or higher desorption from the cell. We suggest that sulfides (exogenous or endogenous) play a key role in controlling mercury methylation and should be considered when investigating the impact of Hg in natural environments.


Introduction
Mercury (Hg) is of high concern due to its toxicity even at low concentration. It is persistent, highly volatile, and is able to convert into highly toxic methylmercury (CH 3 Hg or MeHg), a strong neurotoxin, that is bioaccumulated and biomagnified across the aquatic food web (Atchison and Hare 1994;Boudou and Ribeyre 1997;Farina et al. 2011;Rice et al. 2014). Biotic methylation of inorganic mercury (iHg(II)) by microorganisms is the primary source of MeHg (Jensen and Jernelöv 1969). The environmental and health risk associated to Hg contamination is highly dependent on environmental parameters that can either promote or limit the production of MeHg by microorganisms at the bottom of the food web. Current knowledge on Hg methylation is based on anaerobic microorganisms such as sulfate-reducing (SRB) (Compeau and Bartha 1985) and iron-reducing bacteria (Bravo et al. 2018;Kerin et al. 2006) although other microbial groups , like methanogens, are also known to methylate Hg. Two genes encoding HgcA (a putative methyl transferase involved in methyl transfer to iHg(II)) and HgcB (a Fe-S cluster protein involved in HgcA electron recycling) are required for Hg methylation in these microorganisms (Parks et al. 2013). However, no link between the expression of these genes and the presence of iHg(II) has been so far observed (Goñi-Urriza et al. 2015;Qian et al. 2018).
In SRB, Hg methylation was shown to be higher under non-sulfide-producing growth metabolism (non-sulfidogenic growth: fumarate respiration or fermentation) compared to sulfide producing condition (sulfidogenic growth: sulfate Responsible Editor: Philippe Garrigues respiration) (Bridou et al. 2011;Gilmour et al. 2011;Goñi-Urriza et al. 2015). A negative linear correlation between sulfide and Hg methylation was also observed in a millimolar range of sulfide (Benoit et al. 2001a). The decrease of Hg methylation was attributed to a variation in predicted HgS complexes, specifically to a decrease of the dissolved neutral HgS species (HgS 0 and Hg(SH) 2 0 ) (Benoit et al. 1999a(Benoit et al. , b, 2001a2001b;Drott et al. 2007). These neutral forms are suspected to be more available for the cell than the charged species and their decrease would negatively affect the Hg methylation. While it has been generally accepted that in SRB sulfides inhibit Hg methylation, it was observed recently in methanogens that sulfides at low concentration may promote mercury methylation (Gilmour et al. 2018). Similarly, Graham et al (2013) observed that sulfides enhanced Hg methylation of the sulfate reducer ND132 at low sulfide concentrations (< 200 µM sulfides). Moreover, it has been showed that the formation of nanoscale β-HgS(s) particles in the presence of sulfide could enhance Hg methylation (Deonarine and Hsu-Kim 2009;Gerbig et al. 2011;Zhang et al. 2012;Graham et al. 2012a, Graham et al. 2013Pham et al. 2014;Zhang et al. 2014) questioning the role of HgS complexes and specifically, the impact of neutral HgS species.
The effect of thiols on Hg methylation has been also investigated (Adediran et al. 2019;Liu et al. 2016;Schaefer and Morel 2009;Thomas et al. 2020), although they are present at much lower concentrations in aquatic environments compared to sulfate/sulfide (Bouchet et al. 2018;Liem-Nguyen et al. 2017a). Among these thiols, cysteine (at concentrations between 1 and 100 μM) increased the production of MeHg (Schaefer and Morel 2009). However, cysteine degradation by microbes produces sulfides (Graham et al. 2012a, b;Thomas and Gaillard 2017;Thomas et al. 2018). These produced sulfides are expected to limit Hg methylation, although they can form β-HgS that favor Hg bio-uptake (Thomas and Gaillard 2017;Thomas et al. 2018Thomas et al. , 2020. The effect of cysteine on Hg methylation is not yet fully understood. The sulfides produced by SRB from cysteine degradation (and not cysteine per se) could be responsible for the observed increase in Hg methylation.
Our hypothesis is that iHg(II) methylation is modulated by sulfide regardless of its origin, environmental or produced by the bacteria from cysteine or sulfate. However, whether sulfides enhance or inhibit Hg methylation may depends on the concentration of these molecules present in the cell environment. To support this hypothesis, in this work, (i) the potentials and rates of Hg methylation, (ii) the partitioning between extracellular and cell-associated/ particulate fractions of Hg species (to distinguish exported from the adsorbed/intracellular and particulate MeHg and iHg(II)), and (iii) the hgcA expression were measured in the SRB Pseudodesulfovibrio hydrargyri BerOc1 grown with different concentrations of added (exogenous) or produced (endogenous) sulfides along with various Hg concentrations. For this purpose, we used P. hydrargyri BerOc1 under three different conditions regarding sulfide contents: grown (i) under fumarate respiration (non-sulfidogenic metabolism) in the presence of a wide range of exogenous sulfide concentrations, (ii) under fumarate respiration in the presence of cysteine, and (iii) under sulfate reduction (sulfidogenic metabolism). Because Hg methylation depends on iHg(II) concentration (Gilmour et al. 2011;Isaure et al. 2020), the effect of sulfides (added or produced) was also examined at three different iHg(II) concentrations (from nM to µM). Our study provides new insights into the major role of sulfides on Hg methylation process at sulfide concentrations of environmental relevance, which is important to address the risks of Hg regarding the environmental and the health standpoints on this issue.
Cultures were performed with 40 mM pyruvate as carbon source and either with 40 mM sulfate (sulfidogenic condition) or with 40 mM fumarate (non-sulfidogenic metabolism). Under non-sulfidogenic growth, 0.1 mM cysteine, or a wide range of sulfide (Na 2 S) concentration (from 0.0005 to 5 mM) was added as sulfur source. In growth conditions requiring sulfide addition, sulfides were supplemented 24 h before the inoculation enabling equilibration with the culture medium maintained in anoxic/reduced condition. Cultures were inoculated at 0.03 optical density (OD) measured with a spectrophotometer at 600 nm, from pre-cultures under the same growth condition. Growth was then monitored by OD, and biomass production (ΔOD 600 ) was calculated as [OD (Tf) -OD (Ti)], where Tf corresponds to the end of the exponential phase and Ti to the time of the inoculation of the strain. Growth rate (µ (h −1 )) was obtained by dividing the produced biomass during the exponential phase ( ln(biomassproduction) ) by time of the exponential growth phase.
Cells numbering was performed by flow cytometry (BD Accuri C6 analysor, TBMCore). Aliquots of 1.6 mL of cultures were stored at − 80 °C in 5% (v/v) filtered formaldehyde until cells counting. Samples were tagged with 10X SYBR® (Invitrogen) for 15 min in the dark, and immediately counted using Trucounts beads (BD). Sulfide quantification was performed by Cline (Cline 1969) method using a spectrophotometer at 670 nm and adapted to growth conditions and sulfide concentrations expected. The detection limit was 0.92 µM.

Hg methylation assays and Hg species partitioning
All methylation assays were conducted in autoclaved Hungate tubes pre-washed by ultrasonication in 10% HNO 3 and HCl baths, and rinsed with ultrapure Milli-Q water. iHg(II) stock solution (5 mM HgCl 2 in 1% HCl) was used to yield final concentrations of 0.05 µM, 2 µM, or 5 µM. These concentrations are higher than the concentrations usually measured in natural environments but in line with Hg contents found in highly contaminated aquatic sediments (Higueras et al. 2006;Rimondi et al. 2015;Hellal et al. 2020). Since metacinnabar (β-HgS) is likely to precipitate in these conditions we have performed thermodynamic modelling using Visual Minteq software and its database to verify this point and to get a global view of the Hg species in solution. When necessary (cysteine calculation), Visual Minteq database has been implemented with logK determined by Liem-Nguyen et al. (2017b).
In order to avoid heterogeneity, media flushed with N 2 were inoculated with BerOc1 strain and distributed in tubes containing 100 µL of iHg(II) at the appropriate concentration. The fully filled tubes were closed with PTFE coated butyl stoppers, mixed and incubated in the dark at 37 °C. Inoculation and Hg addition were thus performed simultaneously (Online Resource 1a). Quantification of Hg species (i.e., inorganic mercury, iHg(II), and methylmercury, MeHg), sulfide concentration and cell numbering in the bulk culture were determined at both Ti (just after inoculation and Hg addition) and Tf (at the end of the exponential growth phase). For iHg(II) and MeHg quantification (bulk fraction), an aliquot of 500 µL of the cell culture was collected and digested in 50% (v/v) of 6 N nitric acid (HNO 3 ). The remaining part of the cultures was centrifuged (30 min, 6000 g at 4 °C) to separate supernatant and pellet fractions (Online Resource 1a). Five hundred microliters of the supernatant was collected and digested in 50% (v/v) 6 N HNO 3 (extracellular fraction). Samples were stored at 4 °C before the analyses. Abiotic controls without cells were performed under the same experimental conditions for the three different culture media. iHg(II) and MeHg concentrations were measured by capillary gas chromatography (GC TriPlus™ RSH™, Thermo Scientific) connected to an inductively coupled plasma mass spectrometer (ICPMS, X2-series, Thermo Electron) (Bridou et al. 2011;Pedrero et al. 2012). Quantification of each Hg species concentration was performed by isotopic dilution (Monperrus et al. 2005). Briefly, a known amount of fractions (bulk or supernatant fractions) was buffered at pH 3.9 using a 0.1 M acetic acid/acetate buffer and then spiked with a known quantity of isotopically enriched Hg species ( 199 iHg(II) and 201 MeHg) as internal standards. The recovery of enriched and certified Hg species solutions was of 109.4% ± 16.7% (n = 5). iHg(II) and MeHg were both ethylated using 5% (v/v) NaBEt 4 and extracted in isooctane by shaking vigorously during 20 min. Organic phase containing Hg species was collected and analyzed by GC-ICPMS. The absolute detection limit was of the order of magnitude of femtograms. Each fraction was measured two times (replicate sample precision: 4.1 ± 3.9%).
Hg methylation potential and Hg species partitioning were determined in the same batch culture. Methylation potentials were calculated in the bulk cultures by dividing MeHg concentration produced (Ti-Tf) by total Hg concentration measured at Ti. Partitioning of each Hg species (iHg(II) and MeHg) between cell-associated/particulate and extracellular fractions was calculated by dividing concentrations of Hg species measured in the supernatant (extracellular fraction) by concentrations of Hg species measured in the bulk fraction × 100. Percentage of Hg species in the pellet (i.e., cell-associated fraction (sorbed and intracellular) and particles in the medium) was determined by subtracting the percentage of Hg species in the supernatant fraction to 100%. Total Hg recovery was from 42 to 96.2% (Online Resource 3 and Online Resource Table 1). Part of Hg was likely lost by HgS(s) formation, and/or, by vessel wall adsorption supported by an almost systematic decrease in total Hg recovery between Ti (initial) and Tf (final) (Online Resource Table 1). The lost can be also due to reduction of iHg(II) into Hg(0) but was never demonstrated.
Spearman's correlation between the percentage of MeHg in extracellular fraction and sulfide concentrations was calculated with R software.

hgcA expression assays and analyses
hgcA expression assays were performed under various iHg(II) concentration exposures in non sulfidogenic condition with either cysteine or sulfide (0.1 mM). Growth was performed as previously described except that (i) iHg(II) was added at mid-exponential growth phase in the anaerobic chamber, and (ii) Hg incubation was carried out for 1 h (Online Resource 1b), because gene expression is the first step of a metabolic process and short lived (Norambuena et al. 2020). Based on Hg methylation results, hgcA expression level was measured at iHg(II) concentrations of 0.05 µM, 2 µM, and 5 µM with sulfide, and 0.005 µM, 0.05 µM, 0.5 µM, and 2 µM with cysteine. Controls without iHg(II) were conducted to determine the basal hgcA expression.
After 1 h of incubation with iHg(II), cultures were stopped with 3% (v/v) of RNA protect solution (5% phenol saturated with Na acetate 1 M pH 5.5 into 95% ethanol). Five milliliters of the culture was centrifuged (10 min, 5000 g), and total RNA was extracted using EXTRACT-ALL® Kit (eurobio) according to the kit protocol guidelines. RNA fraction was purified by removing DNA contamination with TURBO DNase RNase-free™ Kit (Ambion) and quantified with Quant-iT RNA Assay Kit (Invitrogen™) and Synergy HTX multi-mode reader (BioTek). Purified RNA was diluted at 20 ng/µL, and cDNA was synthesized using M-MLV-RT (Invitrogen).
hgcA transcripts were quantified by qRT-PCR (quantitative reverse-transcription PCR) in a Light Cycler 480 (Roche) using the DyNAmo™ ColorFlash SYBR® Green qPCR Kit (Finnzyme), in a final volume of 20 µL with 1 µL of cDNA and the appropriate primers (Goñi-Urriza et al. 2015). gyrB gene expression was used as housekeeping gene expression for normalization (Goñi-Urriza et al. 2015). hgcA and gyrB primers were specifically designed to amplified hgcA and gyrB genes in P. hydrargyri strain BerOc1 in a previous study (Goñi-Urriza et al. 2015). Real-time PCR was carried out with an initial denaturation of 7 min at 95 °C followed by 45 cycles of 10 s at 95 °C and 20 s at 60 °C. Melt characteristics of amplicons were checked at the end of the amplification. Assays were performed in three independent cultures (biological replicates) each one measured three times (technical replicates). A reaction mixture with DNA-free RNA was run as a control for detection of DNA contamination. Amplification efficiencies of hgcA and gyrB primers were determined with standard curves as previously described (Goñi-Urriza et al. 2015) and were of 1.90 and 1.95, respectively.
hgcA induction level was determined following double delta Ct (cycle threshold) procedure using gyrB expression as reference. Briefly, it consists of (i) a first normalization of the target gene expression (hgcA) with a reference gene expression (gyrB) and (ii) a second normalization of the Ct in the induced condition (growth with iHg(II)) with the Ct of the non-induced condition (growth without iHg(II)).

Hg methylation is driven by sulfides
The SRB P. hydrargyri strain BerOc1, able to use sulfate or fumarate as electron acceptors, was used in this study to determine the link between MeHg production and sulfide. The production of MeHg was assessed at three different iHg(II) concentrations (0.05, 2, and 5 µM of iHg(II)) under non-sulfidogenic growth (pyruvate (Pyr)/fumurate) with cysteine or sulfide added exogenously (0.1 mM) and under sulfidogenic growth (Pyr/sulfate). The chosen cysteine and sulfide concentrations correspond to the minimal sulfur concentration needed for optimal growth without iHg(II) (Online Resource 2a). The iHg(II) concentrations were chosen from a previous study that showed a maximum of mercury methylation rate at 5 µM of iHg(II) (Isaure et al. 2020). Sulfide concentrations measured at the end of the growth in the presence of cysteine (0.010 ± 0.003-0.020 ± 0.005 mM, Table 1) are consistent with previous sulfide concentrations reported for degradation of cysteine by bacteria, including the well-known Hg methylators Geobacter sulfurreducens PCA and Pseudodesulfovibrio mercurii ND132 (Graham et al. 2012a, b;Thomas and Gaillard 2017;Thomas et al. 2018). Moreover, for a given growth condition, addition of iHg(II) did not impact neither bacterial growth nor sulfide concentration (Table 1, Online Resource 2d-f).
Interestingly, the iHg(II) methylation potential under non-sulfidogenic growth with 0.1 mM sulfide was higher than in the other tested conditions, regardless of the iHg(II) concentration (Fig. 1a, Online Resource Table 1). While a progressive decrease of Hg methylation potential over an increase of iHg(II) concentrations was detected under sulfidogenic growth, a drastic decline of Hg methylation potential for concentrations above 0.05 µM iHg(II) was observed under non-sulfidogenic growth with cysteine (decline observed from 15 to < 1% of Hg methylation potential, starting from 0.05 of iHg(II) to 5 μM) (Fig. 1a). However, under non-sulfidogenic growth with sulfide only a moderate decline was observed above 2 µM iHg(II). Furthermore, the maximal rate of production of MeHg per cell (τ max calculation) is estimated between 50-and 100-fold higher ( Table 1) in presence of sulfide. The saturation of MeHg production (K Michaelis estimation) of BerOc1 grown with sulfide is reached at higher iHg(II) concentrations than when grown with sulfate (10 times lower than sulfide) or cysteine (1000 times lower than sulfide). Although, the estimation was performed with three points only, these results suggested that the Hg methylation mechanism is saturable in the various cell growth conditions tested as previously described under a different growth condition than tested here (Isaure et al. 2020). Abiotic controls confirmed that measured Hg methylation was a fully biotic process and that the observed differences between growth conditions were not related to differences in abiotic Hg methylation (Online Resource 3). Furthermore, our study reports that the probable concentration at which saturation of Hg methylation is reached and the capacity of a single cell to produce MeHg depends on growth conditions. A more detailed assessment of iHg(II) kinetics methylation is necessary to confirm these results.
Importantly, Hg methylation potentials were higher with 0.1 mM of sulfide, compared to cysteine condition (expected to favor Hg methylation (Schaefer and Morel 2009)) regardless of the iHg(II) concentration added (Fig. 1a). However, in presence of cysteine, around 10 µM of sulfides were measured in all the tested iHg(II) concentrations (Table 1). We thus hypothesized that the variation observed in Hg methylation depends on sulfide concentration (added or produced) rather than directly on the added cysteine. Therefore, we investigated iHg(II) methylation at varying concentration of sulfides (from 0.0005 to 5 mM). The range of sulfide concentrations were chosen following the measured sulfide concentration under cysteine and sulfate reduction growth (Table 1). Two iHg(II) concentrations, 0.05 and 2 µM iHg(II) (where the differences between iHg(II) potential are the most prominent (Fig. 1a)), were used to test the effect of sulfide concentration on iHg(II) methylation potentials. Addition of low exogenous sulfide concentrations (< 0.05 mM) limited BerOc1 growth only at 2 µM of Hg(II) while the addition of exogenous sulfide above 2.4 mM was toxic for the growth, regardless of the iHg(II) concentration added (Table 1, Online Resource 2b-c). At 0.05 µM of iHg(II), Hg methylation potentials increased with the increase of sulfide concentrations up to 0.5 mM, and then decreased at higher sulfide concentrations (Fig. 1b). The same pattern was observed with 2 µM of iHg(II) with a maximal Hg methylation reached at 0.1 mM of sulfides (Fig. 1c). Overall, our data showed that sulfide at concentration ranged from 0.1 to 0.5 mM favors Hg methylation in P. hydrargyri BerOc1 and that lower sulfide concentration (< 0.1 mM) and higher sulfide concentration (> 0.5 mM) from that range are disadvantageous to the process. Interestingly, the Hg methylation potential related to the sulfide concentration measured under cysteine and sulfate reduction growth conditions fits perfectly within this pattern (Fig. 1).
Thermodynamic calculation indicated that metacinnabar could precipitate for the two Hg concentrations (0.05 and 2 µM iHg(II)) and for all the tested sulfide concentrations (Online Table 1 Sulfide concentration, biomass production, growth rate, Hg methylation rate (τ), maximal Hg methylation rate (τ max ), and iHg(II) concentrations to reach half of the maximal methylation rates (K Mic ) measured/calculated in P. hydrargyri strain BerOc1 grown under nonsulfidogenic (Pyr/Fum) conditions (with cysteine or sulfide) or under sulfidogenic (Pyr/sulfate) condition a Sulfide concentration in cultures was measured at the end of the exponential growth phase (Tf). Limit of detection (LD). Biomass production (ΔOD600) was determined by subtracting optical density measured at the end of the exponential growth phase (Tf) from OD measured at the time of strain inoculation (Ti). Standard deviations were calculated from at least three independent replicates b Hg methylation rates where the MeHg concentration produced (Tf-Ti) was divided by Ȼ (concentration of cells at Tf) and the incubation time with iHg(II) c τ max : maximal Hg methylation rate was calculated based on a double-reciprocal or Lineweaver-Burk plot 1 = f 1 Hg(II)concentration d K Mic : Michaelis-Menten constant represents the iHg(II) concentration to reach half of the maximal methylation rate determined by the Lineweaver-Burk plot Resource 4a and b), even in the presence of cysteine (Online resource Table 3). The fraction of precipitated Hg as metacinnabar ranged from 100 to 98% with increasing sulfides (Online Resource 4c and d). Moreover, the nature of the main Hg species did not change with the increase of sulfide concentrations in all the tested iHg(II) concentrations (HgHS 2 − > Hg(HS) 2 (aq) > HgS 2 2− ) (Online Resource Table 2). The addition of cysteine (with 0.01 mM of sulfide, representative of the cysteine condition) in the calculations did not change the main Hg species compared to conditions with no cysteine (Online resource Table 3). Thus, the observed Hg methylation pattern might not be directly related to the Hg species given by the thermodynamic calculation. In our previous work on BerOc1, we identified nanosized β-HgS (Isaure et al. 2020). Other works stated that the size, nanostructure and ageing of the metacinnabar impacted the Hg methylation. Particularly, Hg methylation decreased from dissolved Hg/ nano HgS to microHgS (Zhang et al. 2012. Deposition of nanoHgS on bacterial surfaces might also enhance Hg methylation potential of the nanoHgS, either by facilitating dissolution or by direct uptake (Pham et al. 2014). In addition, disorders of HgS nanoparticles could be promoted by bacterial extracellular polymeric substances (EPS), resulting in enhanced MeHg potential, likely due to cell nanoparticles interactions (Zhang et al. 2020). Nano β-HgS likely precipitated in our experiment but the direct link between these particles and Hg methylation was not addressed in this work.
Several previous studies concluded that the presence of sulfide has an inhibitory effect on Hg methylation (Benoit et al. 1999a(Benoit et al. , 2001a. The inhibitory effect on Hg methylation at high sulfide concentration was for instance observed in Desulfobulbus propionicus (1pr3) by testing sulfide concentration from 0.05 to 2 mM with Hg concentrations ranging from 150 to 6800 µM (Hg-containing ores) (Benoit et al. 2001a). However, because only one concentration of sulfide below 0.1 mM was tested in this study, the beneficial effect of sulfide on Hg methylation was not noticed on the plot. Additionally, an optimum sulfide concentration (between 0.01 and 0.1 mM) favoring Hg methylation has been observed in methanogens with 0.01 µM of iHg(II) (Gilmour et al. 2018) and also in the sulfate-reducing bacterium ND132 (at roughly 0.5 mM) with 0.0005 µM of iHg(II) (Graham et al. 2013). It is important to note that iHg(II) concentration has little (if any) influence in the sulfide concentration favoring MeHg production. The modeling linking sulfate concentration to MeHg production in natural ecosystems also identified a similar pattern, with an optimum of Hg methylation ranging from 0.2 to 0.5 mM of sulfate (Gilmour and Henry 1991). Since the optimal concentration of sulfide favoring Hg methylation in our study is close to the optimal concentration of sulfate favoring Hg methylation in Gilmour and Henry study (Gilmour and Henry 1991), it would be interesting to perform the same comparative analysis between Hg methylation and sulfide (instead of sulfate) in natural environments. exogenously; gray triangle: measured in the cysteine growth condition; gray diamond: measured in the sulfidogenic growth condition). Inoculation and iHg(II) addition were performed at the same time. Pyruvate (Pyr) was used as carbon source at 40 mM for all the growth conditions and, sulfate and fumarate (Fum) were added at 40 mM as electron acceptor to promote sulfidogenic (Pyr/Sulfate) or non-sulfidogenic (Pyr/Fum) metabolisms, respectively. MeHg potentials (% of total recovered Hg converted to MeHg) were measured at the end of the exponential phase (Tf). Error bars represent the standard deviations of three independent replicates. NG no growth The mechanism by which S containing molecules favors Hg methylation is far from being understood. Thomas et al. (Thomas et al. 2020) identified HgS complexes in nongrowing Geobacter sulfurreducens PCA (an iHg(II) methylating bacterium) incubated with cysteine. In this study, the addition of cysteine also increased iHg(II) methylation and the authors suggested that the degradation of cysteine produces sulfide that forms HgS complexes and those complexes would favor iHg(II) methylation through a facilitated Hg uptake, as demonstrated in E. coli (Thomas et al. 2019). Unfortunately, sulfide production by G. sulfurreducens PCA was not provided. Very recently, Stenzler and coworkers (2022) demonstrated that the cysteine-induced biogenic production of H 2 S increased iHg(II) bioavailability in E. coli and suggested that regardless of its origins (biogenic production of H 2 S or exogenously amended H 2 S), sulfides could remotely alter iHg(II) speciation. Accordingly, the thermodynamic calculation of Hg species performed in our study indicated that the presence of sulfide at 0.01 mM (even in the presence of cysteine), drastically changed the Hg speciation compared to Hg speciation found with only cysteine (Online resource Table 3). Further spectroscopic analysis and characterization of HgS complexes under a range of sulfide concentrations would shed light on the role of sulfide at 0.1-0.5 mM in the production of HgS forms, particularly the nano β-HgS, and the increase of iHg(II) methylation.
Overall, how sulfide, sulfate (Gilmour and Henry 1991), and cysteine (Schaefer and Morel 2009) affect Hg methylation depends mainly on their concentrations. In marine environments, sulfate can be found at concentrations up to 20 mM (Jørgensen et al. 2019) and cysteine at fM to nM concentrations (Joe-Wong et al. 2012;Liem-Nguyen et al. 2017a;Bouchet et al. 2018;Adediran et al. 2019). At those concentrations, sulfate and cysteine are not expected to improve the MeHg production (Gilmour and Henry 1991;Schaefer and Morel 2009). However, other S-containing molecules, especially those involved in the sulfur biogeochemical cycle (sulfide, sulfite, thiosulfate, and polysulfide), reach concentrations between 0.1 and 1 mM in marine and freshwater sediments (Findlay 2016;Findlay and Kamyshny 2017;Gentès et al. 2017) and can all be transformed into sulfide by reduction or degradation. Here, we reported an increase of Hg methylation at sulfide concentration encountered in the environment. We hypothesize that in natural environments, sulfate will enhance Hg methylation because it favors SRB metabolism whereas sulfide (regardless of its origin, exogenous or endogenous) might govern the extent of Hg methylation potentials, at least for SRB.

Sulfides influence methylmercury partitioning
In a previous study, a link between Hg species partitioning and Hg methylation was observed with various thiol molecules (Schaefer and Morel 2009). Thus, to further understand the differences on Hg methylation observed with various S containing molecules and sulfide concentration in our study, we investigated the Hg species (iHg(II) and MeHg) partitioning between the extracellular and the cellassociated/particulate (cell-sorbed, intracellular and particulate) fractions in BerOc1 grown under the three growth conditions (Pyr/fumarate with cysteine or sulfide and Pyr/ sulfate) and with sulfide gradient under Pyr/fumarate condition. Distribution of Hg species was measured at the beginning of the growth (Ti, iHg(II), Online Resource 5) and at the end of exponential phase (Tf, iHg(II) and MeHg produced) at increasing iHg(II) concentrations (Figs. 2 and 3). Most of iHg(II) was distributed in the pellet at Ti (Online Resource 5), likely due to the precipitation of β-HgS. For a given growth condition, increasing iHg(II) concentrations did not affect neither iHg(II) (Fig. 2a) nor MeHg (Fig. 3a) partitioning at Tf. iHg(II) became more associated with the cells/particles over time (Online Resource 5a, Fig. 3a) and was mostly associated with this fraction at the end of the exponential growth phase (80-100%) regardless of the growth condition (Fig. 3a). On the contrary, MeHg distribution between extracellular and the cell fractions depends on the growth condition (Fig. 3a): while most of the MeHg was detected in the extracellular fraction under sulfidogenic growth (96 ± 2%), the proportion in the extracellular fraction reached only 30 ± 4% under non-sulfidogenic condition with cysteine. An intermediate distribution (50 ± 5%) of MeHg was observed in non-sulfidogenic growth with sulfide. Thus, the proportion of extracellular MeHg increased with the increasing sulfide concentrations measured in the cultures (Fig. 3a, Table 1). The dominance of produced MeHg in the extracellular medium has been already observed and was interpreted as an export of MeHg by the bacteria or desorption from the cells (Graham et al. 2012b;Liu et al. 2016;Pedrero et al. 2012). We thus infer that the export/desorption of MeHg from the bacterial cells can be favored by sulfide.
In order to test this hypothesis, we measured the distribution of Hg species in the extracellular and the cellassociated/particulate fractions in BerOc1 cells grown with increasing exogenous sulfide concentrations. At both 0.05 and 2 µM of iHg(II), iHg(II) became more associated to the cell/particulate fraction over time for all of the sulfide concentrations tested (Online Resource 5b-c and Fig. 2b, c), as observed for BerOc1 under Pyr/fumarate with cysteine and Pyr/sulfate conditions (Online Resource 5a and Fig. 2a). When the highest proportion of iHg(II) is in the pellet (and thermodynamic calculation estimates near 100% of metacinnabar precipitation (Online Resource 4)), the maximum of Hg methylation is measured, suggesting that the precipitation did not limit Hg bioavailability and Hg methylation. Thus, iHg(II) partitioning appear not to be influenced by sulfide concentration. Remarkably, the extracellular MeHg increased with the increase of sulfide concentrations regardless of the iHg(II) concentration added (Figs. 3b and c). A strong positive correlation was observed (Spearman correlation rho = 0.89, p value < 0.001). Previous studies showed an increase of the extracellular MeHg in presence of thiols (Ndu et al. 2016). More specifically, cysteine strongly enhanced the extracellular MeHg produced by G. sulfurreducens PCA and more slightly by Pseudodesulfovibrio mercurii ND132. Accordingly, a facilitated MeHg cell export and desorption from the cells was proposed via the cysteine (Lin et al. 2015). However, no data of sulfide contents was provided (Lin et al. 2015) and, in the light of the recent papers (Thomas et al. 2019), it may be possible that the facilitated export/desorption was the consequence of sulfide production due to cysteine degradation. Our results Fig. 2 Partitioning of inorganic mercury (iHg(II)) between extracellular (gray) and cell-associated/particulate (black) fractions at the end of the exponential growth phase (Tf). a The partitioning of iHg(II) at 0.05, 2, and 5 µM iHg(II) under non-sulfidogenic growth (Pyr/ Fum) with either 0.1 mM cysteine or 0.1 mM sulfide or under sulfidogenic condition (Pyr/sulfate). The partitioning of iHg(II) at either b 0.05 µM or c 2 µM of iHg(II) under Pyr/Fum with a range of sulfide concentrations (added as exogenous sulfide or measured in the cultures). The same cultures were used to determine Hg species in the bulk fraction at the end of the exponential growth phase and Hg species partitioning. After sampling for Hg species determination, cultures were centrifuged and the extracellular iHg(II) and MeHg were measured in the supernatant. Error bars represent the standard deviations of three independent replicates, each one measured three times Fig. 3 Partitioning of MeHg between extracellular (gray) and cellassociated/ particulate (black) fractions at the end of the exponential growth phase (Tf). a The partitioning of MeHg at 0.05, 2, and 5 µM iHg(II) under non-sulfidogenic growth (Pyr/Fum) with either 0.1 mM cysteine or 0.1 mM sulfide or under sulfidogenic condition (Pyr/sulfate). The partitioning of MeHg at either b 0.05 µM or c 2 µM of iHg(II) under Pyr/Fum with a range of sulfide concentrations (added as exogenous sulfide or measured in the cultures). The same cultures were used to determine Hg species in the bulk fraction at the end of the exponential growth phase and Hg species partitioning. After sampling for Hg species determination, cultures were centrifuged and the extracellular iHg(II) and MeHg were measured in the supernatant fraction. Error bars represent the standard deviations of three independent replicates, each one measured three times suggest that sulfide, either exogenous or endogenous (produced by the cell via cysteine degradation or sulfate reduction), could be a parameter involved in MeHg export and/or desorption from the cell.
A recent study on abiotic interaction of MeHg and sulfide predicted the formation of bismethylmercury sulfide as the dominant MeHg species in sulfidic solutions (Kanzler et al. 2018). In mammalian cells, the bismethylmercury sulfide formation from MeHg has been proposed as a mechanism of cell detoxification (Yoshida et al. 2011). In this study, sulfide (exogenous or produced via cysteine and homocysteine degradation) form bismethylmercury sulfide in the presence of MeHg, avoiding MeHg binding to cellular proteins (Yoshida et al. 2011). In our study, the progressive decrease of cell-associated/particulate MeHg observed when sulfide concentration increased could be explained by the formation of bismethylmercury sulfide that avoid the binding of MeHg to cellular proteins. However, it is difficult to speculate on the combined role of Hg methylation and MeHg export for Hg cell detoxification since no link between both Hg methylation potentials and MeHg export could be found. MeHg export probably protects BerOc1 cells against MeHg toxicity.

The increase of Hg methylation in presence of sulfide is not associated to hgcA overexpression
hgcAB are the only genetic determinisms identified as necessary for Hg methylation (Parks et al. 2013). To investigate if the increase of Hg methylation observed with 0.1 mM of sulfide was related to the overexpression of the enzymatic mechanism of Hg methylation, the expression of hgcA gene (involved in the methyl transfer to Hg) was measured in BerOc1. We selected cells growing under non-sulfidogenic growth with 0.1 mM sulfide and 0.1 mM cysteine, where Hg methylation was lower (Fig. 1a). Low hgcA overexpression (from 1.8 to 2.2) was observed for cysteine for all the iHg(II) concentrations tested (Fig. 4a). In contrast, no overexpression could be detected for 0.1 mM sulfide, where the methylation potentials were higher (Figs. 1a and b). Thus, no direct link could be established between hgcA expression and Hg methylation potentials. In addition, our results revealed that hgcA gene is expressed at basal level even at increasing iHg(II) concentrations. iHg(II) seems not control hgcA expression, unlike other metals (Zn, Cu, As, Fe) which regulate the expression of genes encoding enzymes controlling their transformation (Andrei et al. 2020;Andrews et al. 2003;Pederick et al. 2015;Silver 1996;Silver and Phung 2005). Nevertheless, the available data regarding the expression of hgcAB gene are so far insufficient to ascertain that hgcAB genes are not regulated. A transcriptional regulator, arsR, located upstream hgcAB genes and cotranscribed with hgcA clearly suggests a regulated process (Goñi-Urriza et al. 2020). Its implication in the regulation of hgcAB genes expression and the factors that trigger the ArsR response are still unknown and should be addressed in future studies.

Conclusion
In this work, the influence of sulfide (added exogenously or produced endogenously from cysteine or sulfate) on mercury methylation and Hg species partitioning was investigated using the model strain Pseudodesulfovibrio hydrargyri BerOc1. The mercury methylation assays showed that there is a range of sulfide concentration (0.1 to 0.5 mM) that enhance mercury methylation. Interestingly, sulfides produced by cysteine degradation or sulfate reduction seems also an important parameter influencing Hg Fig. 4 Expression level of hgcA gene in P. hydrargyri strain BerOc1 under non-sulfidogenic growth (Pyr/Fum) with a cysteine and b sulfide at various iHg(II) concentrations. iHg(II) was spiked at the beginning of the exponential growth phase and incubated for 1 h. A control condition was performed without iHg(II) (0 μM). Expression levels of hgcA gene were normalized using gyrB gene expression levels in the same condition, following double delta Ct method. Error bars represent the standard deviations of three independent replicates, each one measured three times methylation. However, experiments with a broader range of cysteine and sulfate would be necessary to confirm this hypothesis.
Importantly, partitioning assays also demonstrated that MeHg export increases with increasing sulfide concentration. A drastic shift was observed: low sulfide concentrations lead to MeHg being associated to the cells while high sulfide concentrations results in extracellular MeHg, likely exported from the cells. Despite the lack of correlation between MeHg export and Hg methylation in this study, such a shift in partitioning of MeHg, was never observed, and sulfide seems to be a key player governing MeHg production and export. The evaluation of the concentration of sulfide in natural environments is thus important and should provide information on the possibility of the neurotoxic methylmercury spread in the aquatic environment, in order to better evaluate the risks of environmental Hg contamination and its impact on public health.