Chemical Composition of the MIW
Concentrations of the main constituents of concern were 0.354 mg-Se/l for total dissolved selenium and 54.9 mg-N/l for total nitrite-plus-nitrate. The pH of the MIW was neutral at 7. The MIW was high in total dissolved solids (TDS), to which the main contributors were calcium (Ca2+) (404 mg/l), sulfate (2030 mg/l), and magnesium (Mg2+) (390 mg/l) (Table 1). Hardness of the MIW expressed in terms of CaCO3 was very high compared to natural surface waters. Such high concentrations can lead to precipitation of minerals such as calcite that can cause scaling. High TDS aquatic environments can be challenging for growth and activity of microorganisms 36. Microorganisms from natural freshwater environments experience osmotic stress when introduced into high salinity environments, which can result in water flowing out of the cell into the surrounding environment. If the microbial cells cannot counter this effect, they become dehydrated, which interrupts their cellular functioning, growth, or even lead to cell lysis (Mark et al. 2016). Due to the vast differences between mine impacted and natural environments, we hypothesize that microorganisms native to the specific salinities and chemical compositions of MIW are better suited to inoculate bioreactors treating MIW, versus the more common practice of importing inoculants from bioreactors that treat wastewater other than MIW.
Nitrogen in the MIW is mostly present in the most oxidized form of nitrate (NO3-). The concentrations are higher than those typical for fresh water due to the use of N containing explosives in blasting. During blasting, not all N is converted to gaseous compounds and residual nitrate remains in the waste rock 37. This is usually flushed out quickly and is only high in drainage from active mine waste dumps but lower in drainage from decommissioned dumps. Presence of nitrate in MIW is problematic for removal of dissolved Se using biological processes as it is preferentially reduced by denitrifying organisms over selenate as an electron acceptor. In many cases nitrate inhibits selenate reduction, but this has been shown to depend on the microbial species present as some organisms have two enzymes, one that is specific for nitrate reduction and a separate enzyme for selenate reduction 38. Discovery and use of microorganisms that have the combined capability for nitrate and selenate reduction is an opportunity for exploring microbiomes in mine impacted environments due to exposure of native microorganisms to both of these electron acceptors at elevated above background concentrations.
Sulfate is the anion present in the MIW at the highest concentration, which is over 5,500 times more concentrated than selenate. This results from oxidation of sulfide minerals, such as pyrite, present in the waste rock 39. Penetration of water and oxygen into the waste rock, as well as presence of sulfur oxidizing bacteria, such as Thiobacillus ferrooxidans, contributes to sulfide mineral oxidation and conversion to soluble sulfate 40. Sulfate can be an electron acceptor for sulfate reducing bacteria that are obligately anaerobic microorganisms often active in certain types of MIW 41.
Table 1 Anion and total metal constituents and their concentrations (in mg/l) in the MIW
Constituent
|
Concentration (mg/l)
|
CaCO3 (hardness)
|
2610
|
Chloride (Cl)
|
19
|
Nitrate (as N)
|
54.9
|
Nitrite (as N)
|
0.029
|
Sulfate (SO4)
|
2030
|
Barium (Ba)-Total
|
0.0092
|
Calcium (Ca)-Total
|
404
|
Lithium (Li)-Total
|
0.124
|
Magnesium(Mg)-Total
|
390
|
Nickel (Ni)-Total
|
0.0327
|
Potassium (K)-Total
|
6.2
|
Selenium (Se)-Total
|
0.354
|
Silicon (Si)-Total
|
2.21
|
Sodium (Na)-Total
|
14.4
|
Strontium (Sr)-Total
|
0.250
|
Uranium (U)-Total
|
0.0310
|
Zinc (Zn) – Total
|
-
|
Removal of Nitrate-plus-nitrite-N and Total Dissolved Se from MIW by Three Different Inoculants
No decrease in total nitrate-plus-nitrite-N or total dissolved Se concentration was observed in the negative controls indicating the absence of abiotic mechanisms for reduction of these compounds in the MIW under these conditions (Figure 1). Some decreases in SCOD concentration were observed in the negative controls over time. Contamination of the negative controls was unlikely since less than 0.25 ng/µL DNA was extracted from passages 1 and 2 samples, and no measurable amounts of DNA were extracted from passages 3, 4 and 5 samples.
The positive control cultures from a denitrifying sludge confirmed the capability of denitrifying bacteria to reduce total nitrate-plus-nitrite-N concentrations by greater than 80% in each passage. This indicated that there were no compounds in the MIW that greatly inhibited microbial activity for denitrification. Some reduction in concentrations of soluble Se were measured in the positive control, which became more pronounced over time but remained low at only ~8% removal. This suggests that either nitrate levels were inhibitory for selenate reduction or very few species in the denitrifying sludge had the capacity for selenate reduction.
When the nutrient-amended MIW was inoculated with enriched microorganisms from Site A (natural vegetated wetland), denitrification took place to extents similar to those observed in the positive control in passages 1, 3 and 5 (Figure 1). Interestingly, the extents of denitrification were lower in passages 2 and 4. The extents of dissolved selenium removal were similar to those observed in the positive control. This suggests that these native mine site bacteria are similarly effective at denitrification as the off-site wastewater treatment denitrifying sludge bacteria, and both sources of inoculum were poor at achieving simultaneous removal of soluble Se and nitrate. In contrast, when the nutrient-amended MIW was inoculated with Site B (decommissioned tailings storage facility) enrichment bacteria, nitrate_plus_nitrite-N was often reduced faster than was observed in the other two treatments. Additionally, the extents of total dissolved selenium removal over each passage were all greater than those observed for Site A and the denitrifying sludge enrichment bacteria treatments. The performance of the Site B enrichment bacteria to remove selenium improved noticeably with increased passaging. Ten percent of dissolved selenium was removed in the third passage, which increased to 19% removal in the fourth passage and to over 50% of dissolved selenium removal in the last passage (Passage 5). The large range for the error bar for Site B enrichment treatments at the last time point for Passage 5 reflects the variation in performance of the triplicate cultures. Carbon source consumption was similar for all three bacterial treatments (SCOD versus time plots in Figure 1).
Overall, these observations suggest that native mine site microorganisms can achieve better removal of soluble Se in the presence of nitrate than an external consortium of microorganisms from a wastewater treatment bioreactor, but this depends on the composition of the native consortium. In this work comparing two mine site microbial consortia, the results indicate that the microorganisms enriched from a mine structure, the decommissioned tailings storage facility, were more capable of removing dissolved selenium in the presence of nitrate than a consortium from a natural vegetated wetland, even though both of them were impacted by MIW. This result was not anticipated based on previous testing of the enrichments in laboratory growth media, which indicated that Site A contained microorganisms capable of simultaneous nitrate and selenate reduction, and Site B contained microorganisms capable of rapid selenate reduction only when nitrate was almost completely reduced 24. Microbial enrichments can contain a wide variety of species with some members that are functionally important and others that are growing on the carbon source and nutrients without contributing to nitrate or selenate reduction. To gain a better understanding of why Site B enrichment was better at nitrate-plus-nitrite-N and selenate reduction than Site A, taxonomic microbial population composition profiling and genome resolved metagenomics were performed on the cultures as they were passaged.
Microbial Population Composition
A total of 68 culture samples were sequenced for SSU rRNA amplicons resulting in 1,422,752 good quality denoised reads. The average number of reads per sample was 26,549 +/- 7,946 for the denitrifying sludge, 27,717 +/- 7,336 for Site A enrichment, 23,329 +/- 3.920 for Site B enrichment and 1,859 +/- 1,337 for the abiotic controls. Even though the negative control samples yielded very little or no DNA, they were still subject to sequencing to confirm low possibility of cross contamination. The sequencing results for the negative controls were considered noise as none of the identified taxa were found in any of the other biological samples. Rarefaction curves for the samples indicated that sequencing depth was more than adequate to capture the diversity of species. The total number of unique ASVs (amplicon sequence variants) per culture type were 595 for the denitrifying sludge, 321 for the Site A enrichments and 320 for the Site B enrichments. Overall, the diversity was greater for the denitrifying sludge than the native mine site consortia, both of which were similarly diverse in terms of number of unique ASVs.
The ASVs were classified into 11 dominant genera (present at > 1% of the total population) (Figure 2). Microbial population compositions were reasonably consistent between culture replicates. All starting inoculants and subsequent cultures had genera Sulfurosprillum and Macellibacteroides in common. Sulfurospirillum species are members of the Epsilonproteobacteria and have been reported to grow on substrates such as selenate, as well as nitrate and sulfur compounds42. The species Sulfurospirillum barnesii SES3 was reported to be involved in simultaneous respiratory selenium oxyanion and nitrate reduction43. Despite Sulfurospirillum remaining dominant throughout all cultures, their presence did correlate with selenate reduction in all cultures. There is only one known Macellibacteroides species, M. fermentans, an obligate anaerobe from an abattoir wastewater treatment bioreactor 44. But this strain was not able to metabolize selenate, nitrite, or nitrate. Macellibacteroides sp. are closely related to species in the Parabacteriodes genus. According to the BioCyc database, Parabacteroides sp. have pathways for nitrate reduction 45. Members of Macellibacteroides genus were most prevalent in the Site B cultures and their percentage abundance remained the same throughout passaging.
The five dominant genera in the denitrifying sludge inoculant (Bacteroides, Escherichia, Macellibacteroides, Sulfurosprillum and Veillonella) remained dominant at similar abundances throughout all passages, with the only notable differences being appearance of Paracoccus in Passage 4 and decreasing abundances of Veillonella sp. over time. Paracoccus sp. are well known denitrifying bacteria, some members of which have the capability to reduce selenate22 . Some Veillonella species have been associated with Se reduction as they were abundant microorganisms in anaerobic granules that reduced selenium oxyanions to elemental selenium in a bioreactor46, but it is not known what specific metabolic role they were playing. Escherichia species, such as E. coli, have selenate reductases (Yfgk) that are associated with assimilation of Se into selenoproteins47, and some strains can reduce selenite to elemental Se nanoparticles48. Bacteroides species are involved in carbohydrate degradation and are not associated with denitrification or selenate reduction49. Note that these five genera that were represented in the dominant ASVs in the entire dataset, comprise just over 50% of the total bacterial population in the denitrifying bacteria sludge inoculant. Many other taxa were present in the sludge, but these did not become dominant in the MIW cultures.
Distinguishing genera in the Site A cultures were Sedimentibacter, Pseudomonas and Escherichia (Figure 2). Sedimentibacter and Pseudomonas were most prevalent in the earlier passages and diminished in abundance over time. Sedimentibacter are anaerobic freshwater bacteria that grow on amino acids and no species are known to reduce nitrate or selenate. However, some Pseudomonas sp. are well known for selenate reduction 23,50. Escherichia were more prevalent in the Site A cultures than the other two treatments.
The Site B enrichment culture microbial population was distinguished by presence of Youngibacter, Paracoccus, Macellibacteroides and Desulfomicrobium (Figure 2). There are two known species of Youngibacter, which is in the Firmicutes phylum, both of which are strictly anaerobic carbohydrate fermenters 51. Paracoccus, well known denitrifying bacteria, were more prevalent in the Site B cultures than in the other treatments. The sulfate reducing bacteria genus of Desulfomicrobium increased steadily in abundance in the final three passages of Site B cultures. Desulfomicrobium species can reduce selenate as well as sulfate52 and are also capable of denitrification (KEGG nitrogen metabolism pathway for Desulfomicrobium baculatum53). Abundances of the dominant genera in Site B cultures remained similar over passages 1 to 5.
Overall, several genera were dominant in all cultures with the capacity for denitrification and selenate reduction. Functional traits of the dominant organisms based on what is known about closely related characterized species is still speculative. We used genome-resolved metagenomics to align functional annotation with the taxonomic species dominating these cultures to seek a more accurate understanding of which organisms were contributing to denitrification and selenate reduction.
Genome-resolved Metagenomics
Whole genome sequencing was performed on Site A and Site B enrichment cultures, which included the starting inoculants and samples taken at the end of passages 3, 4 and 5. Library sizes varied from 1.26 to 4.0 Gbp. These were assembled into 138,845 contigs from 200 to 479,064 bp in length (N50 = 2,999) for a total length of 164,808,635 bp. Contigs were binned into 23 metagenome assembled genomes (MAGs) of which six were high quality (> 90% complete, < 10% contaminated) and nine were good quality (>75% complete, <10% contaminated).
As part of the SqueezeMeta pipeline, QC’ed raw reads were aligned onto the assembled contigs to determine relative abundances of genus-level assigned taxonomy (Figure 3). The dominant genera were those also detected in the SSU amplicon sequencing analysis. A larger abundance of Citrobacter were revealed in the WGS than in the SSU data. Few WGS reads were matched to Macellibacteriodes or Veillonella, although this depended on the method used for taxonomic assignment. The Metalign V2 algorithm detected some reads associated with Macillibacteroides (Figure S1). Few genomes for Macellibacteroides species limited the taxonomic annotation with WGS. However, there are 33 genomes for Veillonella species in the BioCyc genome database collection (accessed August 6, 2023).
The main taxonomic differences between Site A and Site B enrichment cultures are significant dominance of Sulfurospirilum by the end of the fifth passage in MIW in the former, and dominance of Paracoccus in the latter. By the fifth passage, Site B microbial population was more diverse than that observed in Site A. Sulfurospirilum became increasingly abundant in Site A enrichments with increased passaging in MIW, whereas they decreased in abundance in Site B over passaging.
The top seven best quality MAGs included species within the main dominant genera revealed by SSU and WGS (Figure 4). No good quality MAGs were obtained for Citrobacter, Desulfomicrobium, Pseudomonas and Sedimentibacter species. The good quality MAGs were screened for open reading frames (ORFs) coding for enzymes potentially involved in selenate reduction. Dissimilatory nitrate reductases, Nap (periplasmic) and Nar (membrane bound), that catalyse reduction of nitrate (NO3-) to nitrite (NO2-) were present in Sulfurospirilium related Bin030 (Nap), Paracoccus related Bin023 (Nap & Nar), E. coli related Bin009 (Nap & Nar) and Veillionellaceae related Bin006 (Nar). Auxiliary maturation protein, NapD was always present with NapA, the redox active subunit. Some nitrate reductases (from Paracoccus sp. for e.g.) are versatile and can also reduce selenate in addition to nitrate albeit at much lower activities 2254. The Paracoccus and E. coli related bins had both periplasmic and membrane bound nitrate reductases, Nap and Nar, plus many of the auxiliary proteins. The Sulfurospirilium related Bin was the only one containing a selenate reductase, SerABC (all subunits). The S. halorespirans PCE-M2 strain’s genome in BioCyc has a sulfate adenyltransferase (EC 2.7.7.4) that can reduce selenate to selenite. The E. coli related Bin had selenate reductases associated with Se assimilation pathways, Ygf and Ynf 55 (Accession ID G6845 in 45), with the latter highly similar to Dms (dimethyl sulfoxide reductase), which was also present in Bin009. All enzymes included in Figure 4 are members of the dimethyl sulfoxide (DMSO) reductase family, which are molybdopterin reductases known to be versatile and important in reduction of toxic elements 56. The Youngiibacter, Macellibacteroides and Bacteroides related bins did not have any putative selenate reductases. It is possible that other species present in the cultures whose genomes were not captured due to their low coverage are capable of selenate reduction. For instance, there are Citrobacter 57, Pseudomonas 23 and Desulfomicrobium 52 species that are known to reduce selenate.
Gene abundance coverage of the key enzymes in Figure 4 across passages were assessed (Figure 5). There was a wider distribution of putative selenate reductases in the Site A enrichment passages, some of which decreased in abundance and others that increased in abundance, whereas in Site B most enzymes potentially capable of selenate reduction were the nitrate reductases Nar and Nap. Based on this metagenomic analysis we can tentatively conclude that simultaneous nitrate and selenate reduction in MIW by Site B enrichment was due to the Paracoccus chartae species and nitrate reductases. A surprising finding was that despite dominance of Sulfurospirilum with putative selenate-specific reductase, SerABC, this did not necessarily correlate with simultaneous selenate and nitrate reduction. Conclusions from metagenomic analyses are hypothetical since no direct causation can be proven, however the information is useful for pursuing further bioreactor studies with the Site B enrichment or isolates of Paracoccus chartae from this enrichment. Since the enrichments were consortia of several different microorganism species, it is possible that synergistic interactions between several key taxa are needed to achieve simultaneous nitrate and selenate reduction. A useful follow up to this work would be to isolate the dominant microorganisms and then perform experiments with different combinations of isolates added into consortia. With respect to microbial community interactions, it is possible for species without pathways for nitrate or selenate reduction to be important if they play an auxiliary role, such as conversion of carbon sources to forms that are metabolized more readily by the functionally important species. In this study, lactate was used as the carbon source as we hypothesized that it would be directly metabolized by the bacteria reducing nitrate and sulfate. It would be important to further investigate why Sulfurospirillum sp. were dominant but did not correlate with efficient selenate reduction despite their genome encoding the SerABC enzyme. Future work using methods to track active organisms, for example using SIP 58 or BONCAT 59, would be very useful in addition to consortia studies.
Other Explanations for Observed Differences in Selenate Removal by Site A and Site B Enrichments
One consideration was if biomass concentrations were different in the two treatments, and this might have influenced selenate reduction rates and extents. It was not possible to measure accurately concentrations of microorganisms using OD600 or flow cytometry due to interference from precipitates. Additionally, the biomass amounts in these cultures were too low to capture with total volatile suspended solids mass analysis. Concentrations of DNA obtained from the same volume of suspended solids, using consistent sampling and extraction methods, were used as a proxy for cell biomass amount (Figure 6). Significantly greater amounts of DNA were extracted from samples taken from the Site B enrichment treatments, and this increased over time. It appears that the microorganisms in the Site B enrichment experienced better growth in the MIW milieu. One possible explanation for this observation might be that the natural wetland (Site A), although this site was impacted by mine leachate, supported microorganisms not able to adapt to the MIW used in the treatments, compared with the microorganisms from the decommissioned tailings storage facility. It might be that even subtle nuances in the chemistry of MIW from different locations on the same mine can influence microbiome composition and activity.
This explanation is somewhat supported by the results for denitrification (Figure 1), where it appeared that nitrate was reduced earlier within each passage in the Site B enrichment inoculated MIW cultures than in those inoculated with Site A enrichment or denitrifying sludge. Possibly the higher biomass amounts in Site B enrichment cultures achieved more rapid denitrification decreasing inhibition of selenate reduction by nitrate. But there was little difference between the nitrate reduction rates for all cultures over the last passage.
The MIW contained high concentrations of salts other than dissolved selenium and nitrate that might have influenced the types of microorganisms that grew in the cultures. For instance, sulfate concentration in the MIW was high. Sulfate is a less thermodynamically favourable electron acceptor than nitrate, selenate or selenite, and no production of sulfide was detected in any of the cultures. In high concentration sulfate MIW we often find specific taxonomic groups associated with sulfate reduction 60. One such genus, Desulfomicrobium, was among the dominant ASVs in the Site B enrichment cultures, which increased in abundance in the later passages. Reduction of selenium oxyanions by sulfate-reducing bacteria has been observed 61, and some studies found that presence of sulfate enhanced selenate removal 62. However, these researchers suggested that selenate reduction is coupled with oxidation of sulfide produced from sulfate reduction, as an abiotic mechanism for selenate removal.