Prevalence of Heterotrophic Methylmercury Detoxifying Bacteria across Oceanic Regions.

Microbial reduction of inorganic divalent mercury (Hg2+) and methylmercury (MeHg) demethylation is performed by the mer operon, specifically by merA and merB genes, respectively, but little is known about the mercury tolerance capacity of marine microorganisms and its prevalence in the ocean. Here, combining culture-dependent analyses with metagenomic and metatranscriptomic data, we show that marine bacteria that encode mer genes are widespread and active in the global ocean. We explored the distribution of these genes in 290 marine heterotrophic bacteria (Alteromonas and Marinobacter spp.) isolated from different oceanographic regions and depths, and assessed their tolerance to diverse concentrations of Hg2+ and MeHg. In particular, the Alteromonas sp. ISS312 strain presented the highest tolerance capacity and a degradation efficiency for MeHg of 98.2% in 24 h. Fragment recruitment analyses of Alteromonas sp. genomes (ISS312 strain and its associated reconstructed metagenome assembled genome MAG-0289) against microbial bathypelagic metagenomes confirm their prevalence in the deep ocean. Moreover, we retrieved 54 merA and 6 merB genes variants related to the Alteromonas sp. ISS312 strain from global metagenomes and metatranscriptomes from Tara Oceans. Our findings highlight the biological reductive MeHg degradation as a relevant pathway of the ocean Hg biogeochemical cycle.

Introduction

Mercury (Hg) is one of the most toxic, widespread, and worrisome contaminants[1,2] and is emitted to the atmosphere by natural sources, such as volcanoes and rock weathering, but particularly by anthropogenic activities. The rising Hg levels since the industrial era, estimated as an increase of 450% in the atmosphere,[3] makes the study of its biogeochemical cycle a major concern to the scientific community. As a consequence, the Minamata Convention, a global treaty to protect human and wildlife health and the environment from the adverse effects of mercury by, for example, reducing its atmospheric emissions, was held in 2013 and entered into force in August 2017.[4]

Emitted elemental (Hg0) and inorganic divalent (Hg2+) Hg can be deposited on land and in oceans by wet and dry depositions.[5,6] Inorganic divalent Hg in the ocean can then be volatilized back again to the atmosphere as Hg0[7] or can be methylated[8−12] to methylmercury (MeHg), which bioaccumulates and biomagnifies in aquatic food webs.[3,13,14] As a consequence, humans are exposed to this neurotoxicant mainly through fish and seafood consumption.[13,15,16] Methylmercury levels in the oceans vary with depth, and usually, measures are being reported low in open ocean surface waters, maximal in intermediate layers, especially in regions of low-oxygen and near or below the thermoclines (up to 1000 m depth), and low and relatively constant in deeper waters (>1000 m depth).[13,17,18] While Hg2+ methylation has been reported to occur in oxic and sub-oxic layers of the ocean water column[8,17,19−21] mainly associated with the microbial remineralization of sinking particulate organic matter,[17,18,22] much less is known about MeHg demethylation and Hg2+ reduction processes. Although MeHg demethylation and Hg2+ reduction processes can be photochemically mediated,[23−25] light penetration in the ocean water column is limited to 200 m;[26] thus, biological MeHg degradation and Hg2+ reduction processes likely govern in the ocean water column. Moreover, taking into account that different concentrations of MeHg can be found through the ocean water column,[20,27−29] it would be plausible to find microorganisms with Hg detoxification capacities, especially in deep aphotic zones. However, very few studies have unveiled and demonstrated the role of microorganisms in MeHg degradation in the vast deep ocean.

Biological MeHg demethylation and Hg2+ reduction detoxification processes can be mediated by different biotic processes, including reductive reactions mediated by the mer operon,[30,31] but also oxidative reactions[32,33] as well as the recently described MeHg degradation performed by methanotrophic bacteria.[34] In this article, we will focus only on the study of bacteria carrying the mer operon (i.e., biological reductive MeHg degradation pathways). While the operon can be composed of different sets of genes,[30,31] the key genes are merA and merB. The first one codifies a mercuric reductase and is responsible for the transformation of Hg2+ to the less harmful and volatile Hg0.[30] The merB gene encodes an organomercurial lyase enzyme that confers resistance to the organic MeHg form. It is responsible for its demethylation, releasing Hg2+, which is then reduced to Hg0 by the merA gene.[30] These machineries have been found in numerous microorganisms, including aerobic and anaerobic microbial species, although demethylation appears to be predominantly accomplished by aerobic organisms.[31,33] To date, very few studies have showed the presence of mer genes in oceanic waters, with the exception of some studies in the North Pacific and Arctic Oceans.[35] In order to build this gap in knowledge, the aim of this study was to demonstrate that marine bacteria that encode mer genes could degrade MeHg and to explore its prevalence in the open ocean, including deep-ocean waters. For this, we took advantage of the MARINHET[36] culture collection, which includes marine bacterial strains from a wide variety of oceanographic regions and depths. We analyzed 290 marine heterotrophic bacteria in order to: (i) detect the presence of merA and merB genes in different oceanographic regions and depths, (ii) assess the tolerance of an important fraction of strains that encode those genes to different concentrations of divalent inorganic Hg (Hg2+) and monomethylmercury (MeHg), (iii) describe the degradation potential for the most tolerant strains, and (iv) explore its prevalence and biogeography patterns across oceans and depths by retrieving merA and merB genes from marine microbial metagenomes (metaG) and metatranscriptomes (metaT) from the Tara Oceans expedition.[37,38]

Identification of Hg-resistant bacteria in contrasting aquatic ecosystems and the assessment of their tolerance to different concentrations of MeHg provide new opportunities to explore the ubiquity and prevalence of marine cultured bacteria with detoxification capacity in the open ocean (i.e., non-contaminated sites). On the other hand, this type of study sets the fundamentals for finding suitable microorganisms to be used for bioremediation strategies.

Materials and Methods

Primer Design for merA and merB Genes of Alteromonas and Marinobacter

Primers were designed in order to identify mercury-resistant bacteria among Alteromonas and Marinobacter strains in a culture collection because: (i) these genera are well known to encode the mer operon[39−43] in their genomes; (ii) they are the most common culturable heterotrophic bacteria living in open marine waters all around the world, as they have been isolated from a wide variety of marine environments,[41,44−49] and in the case of Alteromonas, it is one of the most ubiquitous cultured taxa in the ocean;[36] and (iii) we have access to a large number of isolates thanks to the MARINHET culture collection (see the Supporting Information for more details about the culture collection). For all these reasons, these two genera were a suitable target to prove that mer genes can be found across different oceanographic regions and depths but also to demonstrate the MeHg-degrading capability of different strains.

Specific primer pairs were designed separately for: (i) merA of Alteromonas, (ii) merA of Marinobacter, (iii) merA + merB of Alteromonas, and (iv) merA + merB of Marinobacter based on reference sequences downloaded in 2016 from the Integrated Microbial Genomes (IMG) database of the Joint Genome Institute (JGI). See Figure S1 and Table S1 for detailed information. Noteworthy, the primers designed in this study are useful for detecting exclusively Alteromonas and Marinobacter and might not be suitable for other taxa.

DNA Extraction and PCR Conditions

The primers previously designed were used for the screening of merA and merAB Hg-resistant genes in all Alteromonas and Marinobacter strains (n = 290) available at the MARINHET culture collection in 2016 (n = 1313). DNA of all the strains was extracted from 48 h liquid cultures grown in a Zobell broth medium (i.e., 5 g peptone, 1 g yeast extract in 750 mL of 30 kDa filtered seawater and 250 mL of Milli-Q water) using the DNeasy Blood & Tissue kit (Qiagen) following the manufacturer’s recommendations. Detailed PCR conditions are described in the Supporting Information. The PCR products were verified and quantified by agarose gel electrophoresis with a standard low-DNA mass ladder (Invitrogen). Purification and OneShot Sanger sequencing of merA and merAB genes products were performed by Genoscreen (Lille, France) with both forward and reverse primers. Geneious software v.11.0.5[50] was used for manual cleaning and quality control of the sequences.

Minimum Inhibitory Concentration Experiments

A total of 73 strains from the 290 isolates previously screened by PCR were subjected to minimum inhibitory concentration (MIC) assays, including all strains with positive results for merA and/or merAB presence except one that could not grow again from the cryostock. MIC assays were designed based on previous studies[43,51] in order to assess the tolerance of the marine strains to different concentrations of inorganic Hg [as mercury(II) chloride, HgCl2] and organic Hg (as methylmercury chloride, CH3HgCl) and thus to test the activity of merA and merB genes, respectively. A stock solution of HgCl2 was prepared at 500 μM with autoclaved Milli-Q water. Liquid cultures of the strains growing in the Zobell broth with an optical density (O.D. at 600 nm) of 0.1 were placed in 24-well plates and inoculated with 5, 10, 20, 25, and 50 μM HgCl2. In specific cases, growth was observed in all HgCl2 concentrations and further MIC assays were done to increase the final concentrations to 50, 60, 70, 80, 90, and 100 μM. The tolerance to CH3HgCl was also tested for the most tolerant strains to HgCl2. In these case, 24-well plates were inoculated with the stock solution to reach final concentrations of 2.5, 5, 10, 15, and 20 μM. In all plates, a positive control (liquid culture of the strains not amended with CH3HgCl or HgCl2) and a negative control (broth media without bacteria in order to check for possible environmental contamination) were included in the assays. Plates were sealed with parafilm and incubated at room temperature (RT, ∼20 °C) and kept in the dark for 72 h. Visual examination and OD measurements at 600 nm were done in a 24 h period using an automatic plate reader (Infinite M200, Tecan), and data were collected using the Magellan Data Analysis Software (Tecan Diagnostics).

Growth Curves

Growth curves were performed to characterize the growth rates of the most tolerant strain (ISS312) to different concentrations of CH3HgCl. We prepared 200 mL of liquid cultures in the Zobell broth supplemented with CH3HgCl at final concentrations of 0 (positive control), 1, 2.5, and 5 μM in triplicates. The initial OD at 600 nm of the cultures was 0.05 in order to assure enough concentration of cells for growth. Samples for OD measurements and for bacterial cell counts were taken approximately every 2 h. OD was measured at 600 nm with a spectrophotometer (Varian Cary 100 UV–Vis), and cells were stained with 4′,6-diamidino-2-phenylindole and counted with an automated microscope Zeiss Axio Imager Z2M[52,53] using the automated image analysis software ACME Tool (www.technobiology.ch). Predicted growth curves based on OD observations and kinetic values, including growth rates (μmax), carrying capacity (k), and lag phase time, were calculated with the R package Growthcurver v.0.3.0[54] and GrowthRates v.4.3 software.[55] For graphical representation, replicates of the different growth curve experiments at several CH3HgCl concentrations were averaged. Hence, mean OD and standard deviation were calculated for each time point of the curves.

From the 1 and 5 μM growth curves, 2 mL samples were also taken for characterizing CH3HgCl degradation rates at time 0, 6, 12, 24, and 48 h. Besides, in order to check the possibility that CH3HgCl was being abiotically removed, we also measured the CH3HgCl concentrations from samples taken from multiwell plate experiments, including different controls (more details in the Supporting Information). Descriptions of Hg species concentration measurements are also provided in the Supporting Information.

Phylogenetic Analyses

In order to show that MeHg degradation capacity is strain specific, a phylogeny of the isolates screened by PCR for merA and merAB genes was inferred from their partial 16S rRNA sequences in order to detect a possible clustering between all the positive strains. The closest sequence to each isolate 16S rRNA gene was obtained by BLASTn[56] against the SILVA v.132 database. Alignment of the isolates and reference sequences was performed with MUSCLE from the Geneious software v.11.0.5.[50] The alignment was trimmed to the common 16S rRNA gene fragment covered by both sets of sequences. Phylogeny was constructed using maximum-likelihood inference with RAXML-NG 0.9.0,[57] the GTR evolutionary model with optimization in the among-site rate heterogeneity model and the proportion of invariant sites (GTR + G + I), and 100 bootstrap replicates. In the same way, a phylogenetic tree was constructed with the partial 16S rRNA sequences of the positive isolates only. In this tree, the closest match in the SILVA v.132 database was also included. Presence of merA and merAB genes, origin of the strains, plus their tolerance to HgCl2 and MeHg were added with Interactive Tree of Life (ITOL).[58] Supplementary phylogenetic trees were also constructed using amino acid sequences of merA and merB genes (see the Supporting Information for details).

Fragment Recruitment Analyses of the Genome of ISS312 strain and MAG-0289 in Bathypelagic Metagenomes

The abundance of ISS312 strain across the global bathypelagic ocean was assessed thanks to the sequencing of its complete genome (Supporting Information) that it was taxonomically assigned toAlteromonas mediterranea. In addition, thanks to a previous study,[59] we could reconstruct a metagenome assembled genome (MAG-0289) also affiliated to the Alteromonas mediterranea and encoding merA and merB genes. Fragment recruitment analyses (FRA) of the ISS312 strain and the MAG-0289 was performed by mapping the metagenomic reads of 58 bathypelagic microbial metaG from 32 stations[59,60] from the Malaspina expedition, including free-living (FL) (0.2–0.8 μm) and particle-attached (PA) (0.8–20 μm) microbial communities. Analyses were done with BLASTn v2.7.1+.[56] Details of FRA are explained in the Supporting Information.

Detection of merA and merB Genes in Global Metagenomes and Metatranscriptomes from Tara Oceans Expedition

BLASTp analyses were performed using a conservative e-value (>1E-100) with the merA and merB sequences of the ISS312 genome against global prokaryotic metaG and metaT available in the Ocean Microbial Reference Gene catalog V2 (OMRGC.v2)[38] from the Tara Oceans expedition covering surface, deep chlorophyll maximum (DCM), and mesopelagic layers across oceanographic regions. The generation and annotation of OMRGC.v2, the taxonomic profiling of metagenomic and metatranscriptomic composition, and the normalization of metagenomic and metatranscriptomic profiles used for the extraction of the abundance of the merA and merB genes homologous were thoroughly explained in Salazar et al.[38] Analyses could be performed thanks to the Ocean Gene Atlas Web resource,[37] and abundance results from metaG and metaT were normalized based on the percentage of mapped reads.

Statistical Analyses

ANOVA tests from the stats package of the R statistical software[61] were performed in order to observe if the presence of merA and/or merAB genes in the studied strains was linked to specific oceanographic locations. Further, non-parametric Kruskal–Wallis test, from the stats package of the R statistical software,[61] was performed followed by the post hoc pairwise Wilcox test to see the differences between FRA results in different oceanographic regions and between FL and PA bacterial communities. To assess significance, all the statistical analyses were set to an alpha value of 0.05.

Nucleotide Accession Numbers

Mercury detoxification genes (merA and merAB) detected in this study through PCR were deposited in GenBank under accession numbers MW273028–MW273125. The A. mediterranea ISS312 genome was deposited in ENA under study accession number PRJEB46669.

Results and Discussion

Presence of merA and merB Genes among Alteromonas and Marinobacter Trains

Among the 290 strains of the MARINHET bacterial culture collection,[36] 244 were taxonomically classified as Alteromonas sp. and 46 as Marinobacter sp. (Table S2). These strains were isolated from different depths, including the surface, the deep chlorophyll maximum (DCM), and the bathypelagic zone. The strains here studied for merAB genes were isolated from different oceanic regions such as the North Western Mediterranean Sea (89), South (101) and North Atlantic (42), Indian (44), Arctic (7) and Southern (7) Oceans and included isolates from photic (160) and aphotic (130) layers (Table , Figure S2, and Tables S3, S4).

Table 1

Summary of the PCR Screening Results for merA and merAB in Alteromonas and Marinobacter Strainsa

				positives PCR for
genus	no. of tested strains	layer	ocean	merA	merAB	total strains with merA and/or merAB
Alteromonas	127	photic	Southern Ocean	1	0	33 (13.5%)
			Indian Ocean	0	1
			NW Mediterranean	2	0
			North Atlantic Ocean	5	0
			South Atlantic Ocean	4	0
	117	aphotic	South Atlantic Ocean	18	3
Marinobacter	33	photic	Southern Ocean	4	4	41 (89.1%)
			North Atlantic Ocean	1	0
			South Atlantic Ocean	26	16
	13	aphotic	NW Mediterranean	9	2
			North Atlantic Ocean	1	0
			South Atlantic Ocean	1	0

Photic includes surface and DCM isolates, while aphotic includes bathypelagic isolates. NW Mediterranean: North Western Mediterranean.

The functional screening of the merA and merB genes from the 244 Alteromonas and 46 Marinobacter strains revealed that 13.5% (32 out of 244) and 89.1% (41 out of 46) of the strains presented only merA, while only 1.6% (4 out of 244) and 47.8% (22 out of 46) presented both merA and merB genes (merAB) (Table ). No significant differences were found between depths or between oceans (ANOVA, P-value > 0.05), but in general, we found a higher proportion of positive strains coming from waters of the Southern Ocean (71% despite the lower number of strains tested) and the South Atlantic Ocean (48%), followed by those retrieved from the North Atlantic Ocean (17%) and the North Western Mediterranean Sea (13%).

High Variability of Mercury Tolerance within Marine Bacteria

It is unknown whether mercury tolerance is a conservative trait within marine bacterial strains of the same genera. To prove so, MIC experiments were done in all Alteromonas and Marinobacter isolates (73 strains) presenting merA and/or merAB detected by PCR, except one isolate that was not able to grow again from the cryostock. First, we tested the tolerance for inorganic mercury (HgCl2), and we observed that 32 Alteromonas and 41 Marinobacter strains displayed different levels of tolerance. MIC values ranged generally from 5 to 50 μM. Around 50% of the Alteromonas and Marinobacter strains tested presented a MIC of 20 μM, and one of the isolates stood out as it presented a tolerance to HgCl2 up to 70 μM (Table S5). Besides, we tested the tolerance to MeHg amended in the form of CH3HgCl for those strains already presenting a tolerance to inorganic mercury above 20 μM and encoding the merB gene. For Alteromonas, three strains taxonomically classified as A. mediterranea presented a high tolerance to CH3HgCl, growing at concentrations up to 10 μM (Figure ), but all the tested Marinobacter sp. strains did not show a substantial growth above 2.5 μM of CH3HgCl (Figure ).

Figure 1

Phylogeny of the 16 S rRNA gene of Alteromonas and Marinobacter positive strains for merA and/or merAB genes screening. First inner colored strip indicates genus of the strain. Second colored strip indicates the presence or absence of genes based on PCR results. Bars indicate results from the MIC experiments: yellow-green, HgCl2; purple, MeHg. Tolerance values are in μM. JX52807, Sulfitobacter, and IM32RT_ISS194 are outgroups of the tree. The numbers in the nodes represent bootstrap percentages >75%. Names in bold indicate reference sequences: A. australica, Alteromonas australica; A. mediterranea, Alteromonas mediterranea; unc, uncultured; M. hydrocarbono, Marinobacter hydrocarbonoclasticus.

If we take a look into the phylogenetic tree constructed with the 16S rRNA sequences (Figure ) of the strains presenting merA and/or merAB genes, different tolerances were found within the same phylogenetic cluster. For example, within the cluster of A. mediterranea, some strains presented a MIC of 20 μM, while the isolate that presented the highest tolerance (70 μM, ISS312) also belonged to the same Alteromonas species (Figure ). The same pattern occurred among Marinobacter isolates, where members of the Marinobacter hydrocarbonoclasticus cluster presented MIC values ranging from 10 to 50 μM (Table S5 and Figure ). The same pattern can be observed in the merA and merAB amino acid phylogenies (Figures S3 and S4).

The MIC value heterogeneity within strains belonging to the same phylogenetic cluster suggested that the level of Hg resistance was strain specific, and we probably retrieved different ecotypes within the same species with contrasting tolerances to Hg. Despite these differences between strains, the tolerances found for were similar to those found in other studies where Alteromonas(43,62−64) and Marinobacter(65) genera were also isolated from different marine ecosystems such as hydrothermal vents, estuaries, or contaminated sediments. However, to the best of our knowledge, this is the first study that addresses the tolerance of Marinobacter spp. and A. mediterranea isolated from the ocean to MeHg. Hence, we found out that a strain affiliated to A. mediterranea (ISS312) presented a MIC to HgCl2 higher than other strains already published, up to 70 μM, and we also determined that it was able to grow in the presence of MeHg, presenting a MIC up to 10 μM, representing a good candidate for future bioremediation studies in highly contaminated areas with both organic and inorganic mercuric compounds.

Description of the Highly Tolerant Alteromonas sp. Strain ISS312

Strain ISS312, isolated from South Atlantic bathypelagic waters at 4000 m and classified as A. mediterranea, displayed the highest tolerance to both HgCl2 (70 μM) and MeHg (10 μM). The growth rates of this strain at different concentrations of MeHg were assessed and included: a control without MeHg (0 μM) and with 1, 2.5, and 5 μM MeHg. Growth curves at 0 and 1 μM were very similar, as well as between 2.5 and 5 μM (Figure A). We observed that the major difference between growth curves was the length of the lag phase, where bacteria adapt themselves to the growth conditions. This phenomenon seems to be a common trait for Hg-resistant strains in the presence of toxic compounds, as this behavior has been repeatedly observed in different species of Pseudomonas sp., Alcaligenes sp., or Bacillus sp.[66−68] However, once the cultures started to grow, their growth rates (μmax) were very similar independently of their initial MeHg concentrations, ranging from 0.10 h–1 in the control to 0.09 h–1 at 5 μM. A stationary phase was reached in all concentrations at 80 h, even though at this time, cultures at higher concentrations of MeHg seemed to be only entering the plateau (Figure A). In addition, their carrying capacity (k), that is, the maximum population size of a species, was between 1.6 and 1.9 based on OD measures, revealing very similar values between tested concentrations, an observation also recurrently reported.[67,69] Transmission electron microscopy (TEM) observations of the ISS312 cultures growing at 0 and 5 μM of MeHg also showed similar morphology and ultrastructure of the cells (Figure A).

Figure 2

Growth effect of MeHg in ISS312 strain. (A) Growth kinetics of the A. mediterranea ISS312 strain in a Zobell broth containing MeHg [control (0 μM), 1, 2.5, and 5 μM]. μmax indicates the maximum growth rate for each MeHg concentration. TEM images of the strain growing at 0 μM and 5 μM are shown in the right side of the plot. Details for preparation and observation of samples for TEM are explained in the Supporting Information. (B) MeHg removal by strain ISS312 during the growth curve experiment at 5 μM. Mean and standard deviation from three replicates samples are shown in both graphs.

MeHg concentrations were measured in the isolate exposed to 5 μM at different incubation times. MeHg concentrations were reduced by 36% (2.6 μM) and 72% (1.1 μM) (Figure B) during the lag phase at 6 h and 12 h, respectively. After 24 h, when almost all MeHg was removed (98.2%), this microorganism began the exponential growth phase. After 48 h, MeHg could not be detected (Figure B and Table S6). It is important to highlight that we detected a certain level of abiotic MeHg degradation (up to 25%) in the medium exposed without bacteria and in the killed control (Table S7).

Global Distribution of the ISS312 Strain and MAG-0289 in the Bathypelagic Ocean

The biogeographic and size fraction distribution of the ISS312 genome was assessed in all available bathypelagic metaG of the Malaspina expedition[59] because strain ISS312 was originally retrieved from bathypelagic waters of the South Atlantic Ocean. We found that this strain belonging to the A. mediterranea species was distributed across all the temperate bathypelagic waters, including the Atlantic, the Pacific, and the Indian Oceans (Figure ). Its abundance, according to the data from the FRA, varied across ocean basins, and we found significant differences between the Pacific and the Brazil basins (P-value = 0.019), and between the Pacific and the Canary basins (P-value: 0.011), suggesting a higher abundance of this bacterium in the Atlantic Ocean. Despite finding these differences between oceans, we did not find significant differences between plankton size fractions, indicating that the isolate could be present in both the FL (0.2–0.8 μm) and the PA (0.8–20 μm) microbial communities (Figure ). Moreover, by using bathypelagic metaG, one MAG was reconstructed, the MAG-0289, which genome-aligned at 99% with the Alteromonas sp. ISS312 strain genome and showed a 99.34% average nucleotide identity. The FRA of the MAG-0289 and the ISS312 strain against the bathypelagic metaG displayed identical biogeographical patterns (Figure S5), indicating that the MAG-0289 is a good representative of the ISS312 strain.

Figure 3

World map showing the distribution of the A. mediterranea strain ISS312. Size of the dots indicate number of reads (×1000) and color indicates if the reads were recruited in the FL (0.2–0.8 μm) or in the PA (0.8–20 μm) microbial communities of the bathypelagic samples.

Our results confirm the prevalence and wide distribution of the A. mediterranea species carrying Hg-resistant genes across the global bathypelagic ocean but also shows its occurrence in both plankton microbial size fractions analysed. To the best of our knowledge, the detection and characterization of microorganisms encoding former genes have not been previously reported from non-contaminated deep open ocean waters at a large scale.

Recovery of merA and merB Genes from Related Lineages of ISS312 from Marine Microbial Metagenomes and Metatranscriptomes

A remaining question is whether merA and merB genes are active under natural ambient Hg concentrations. For answering this question, we have analyzed the available metaT of the Ocean Microbial Reference Gene Catalog V2 (OMRGC.v2)[38] from the Tara Oceans expedition through the Ocean Gene Atlas resource[37] covering surface, DCM, and mesopelagic layers across oceanographic regions. BLASTp search using the merA and merB genes from the ISS312 strain as input allowed us to extract 54 merA and 6 merB genes (e-value 1E-100) that ranged from 64 to 99.3% identities in their amino acid sequences and belonged to gammaproteobacterial lineages mainly from the Alteromonadales order and Halomonadaceae family (Table S8). These merA and merB transcripts were prevalent across oceans and depths (see Figure , Figures S6 and S7). We found that merB genes were more transcribed (Figure ) in the mesopelagic zone of the North Indian and Eastern South Pacific Oceans, which are well known regions of oxygen minimum zone areas, and also in some stations from surface South Atlantic waters. It is known that MeHg production presents its maximum concentrations usually near the thermocline and in regions with low oxygen concentrations.[20,70] Perhaps bacteria present in those regions encode and transcribe the Hg-resistant genes in order to cope with those more elevated MeHg concentrations.

Figure 4

Biogeography of the merA and merB transcripts across oceanic regions and depths from microbial metaT from the Tara Oceans expedition. (A–C) Abundance of the merA transcripts found on the surface, DCM, and mesopelagic samples. (D–F) Abundance of merB transcripts found on the surface, DCM, and mesopelagic samples.

Outlook

A combination of culture-dependent analyses with metagenomic and metatranscriptomic analyses from the global photic and aphotic oceans unveiled that biological reductive MeHg degradation capacities were widely distributed and active in the open ocean, especially in the mesopelagic zone. In particular, this study has uncovered that heterotrophic bacteria containing mer genes are present in the open ocean from different oceanographic regions and depths but also plankton size fractions. Also, this study reveals that strains closely related phylogenetically presented contrasting levels of tolerance to Hg, indicating that this biological capacity is strain specific. In particular, we show that the ISS312 strain isolated from bathypelagic waters of the South Atlantic Ocean presented a strong and fast capacity to degrade MeHg. Moreover, the ISS312 strain genome and related lineages harboring merA and merB genes are present and transcribed globally in marine samples across oceans and depths, including polar regions. This outcome has important implications in the biogeochemical cycle of Hg as it provides new understanding on the main players driving MeHg levels in the ocean and can ultimately help to improve current frameworks for marine food webs and human exposure to MeHg.