High diversity, abundance, and expression of hydrogenases in groundwater (2024)

Search for other works by this author on:

Article history

Abstract

Generating hydrogen (H₂) from solar and wind energy, and subsequently storing it on a terawatt scale in the subsurface is currently considered a key aspect of the energy transition [1-3]. One of the potential challenges of this approach is the microbial oxidation of hydrogen, which could induce hydrogen loss [4-6]. Our recent work suggested a high potential for microbial hydrogen turnover in groundwaters, based on dissolved hydrogen concentrations, as well as detection and activity of hydrogenotrophic methanogens [7]. Here we explored the diversity and potential functions of hydrogenases with an expanded sample set encompassing 265 groundwater samples (geochemically characterised and amplicon sequenced, with additional 25 metagenomes and 5 proteomes) from 138 wells in Alberta (Canada), with sampling depths between 0 and 157m (Fig. S1, Tables S1, S2). The groundwaters displayed a range of oxidation states from oxic to completely reduced, accompanied with a wide range of sulfate (>10g/L to below detection) and methane concentrations (74mg/L to below detection).

The abundance and expression of different types of hydrogenases were estimated based on unassembled reads, assembled contigs, metagenome-assembled genomes (MAGs), and proteins in 25 groundwaters. Few [Fe]- and [FeFe]-hydrogenases were present in our data. Notably, the catalytic subunit of [NiFe]-hydrogenase was as abundant as the DNA-directed RNA polymerase (rpoB) (Fig. 1A, Table S3). In 12 out of 25 metagenomes, hydrogenase genes were more abundant than rpoB genes, indicating multiple copies of same subtypes or various subtypes of hydrogenases in a single genome. The ratio of hydrogenase over rpoB correlated positively with depth (P = 0.009, Fig. 1B). From metagenomes, 616 high-quality and 629 medium-quality MAGs were obtained. Hydrogenases were present in 481 out of 1245 MAGs, which together accounted on average for 50% of the relative abundance of all MAGs (Fig. 1C, Table S4). In eight samples, MAGs with hydrogenases accounted for >70% of all MAGs. Although conducting proteomics with groundwater samples is challenging due to low cell counts, we obtained proteomes of five groundwaters, showing hydrogenases accounted for 0.016–1.0% of all proteins (Tables S5–S9). In proteomes of individual species associated with our MAGs, the relative abundance of hydrogenases ranged from 0.0026% (Methylomonadaceae) to 5.9% (Methanobacteriaceae) (Fig. 1D). Interestingly, hydrogenase was more abundant in the three methanogen proteomes (>1.3%) than the 21 bacterial proteomes (<0.19%) (Tables S5–S9). Thus, hydrogenase genes might be one of the most prevalent genes in the subsurface and active expression indicated these genes were functional.

From the assembled contigs, 1545 [NiFe]-hydrogenase gene sequences were recovered (Supplementary Result 1), which displayed vast diversity (Table S10). These groundwater hydrogenase sequences considerably increased the number (1999) of [NiFe]-hydrogenase sequences present in a widely used database (Fig. 2) [8]. The newly discovered diversity, abundance, and expression were concentrated among a few specific subtypes of [NiFe]-hydrogenases, groups 1e, 3a, 3b, and 3d. Most of the new diversity was observed in group 3b, while group 3d was the most abundant in metagenomes and group 3a was the most abundant in proteomes.

Hydrogenases of groups 1e and 3b are associated with sulfur reduction [8, 9]. In our data, the abundance of groups 1e and 3b both positively correlated with sulfate concentration (P = 0.025 and 0.007, respectively, Fig. S2). Group 3b hydrogenases were occasionally observed in close proximity to sulfhydrogenase subunit delta and sulfite reductase subunit A (Tables S12, S13). Among high-quality MAGs, group 1e was exclusively present in members of Burkholderiales, and sometimes co-existed with group 3b. Group 3b was commonly detected in MAGs of sulfate-reducing microorganisms, particularly thirteen Desulfobacterota and three Thermodesulfovibrionia. These MAGs also encoded sulfate adenylyltransferase, adenylylsulfate reductase, and dissimilatory sulfite reductase (Table S11). Some group 3b hydrogenases were detected in MAGs of sulfur-oxidising microorganisms such as three Gallionellaceae that encoded sulfide:quinone oxidoreductase and sulfite dehydrogenase, and four Thiobacillaceae that contained the thiosulfate oxidation sox complex. These hydrogenases might also function alongside sulfur oxidation, coupled to oxygen or nitrate reduction. Many other group 3b were detected in genomes of microbial “dark matter” clades, such as Patescibacteria (14) and Omnitrophota (5), consistent with previous findings [10, 11].

Group 3a hydrogenases are associated with methanogenesis [8, 9] and were most abundant in our proteomes (Fig. 1D, Tables S5–S9). They were exclusively observed in MAGs of hydrogenotrophic methanogens, six Methanomicrobiales and four Methanobacteriales [12]. All Methanobacteriales also encoded tetrahydromethanopterin-reducing [Fe]-hydrogenases. Nine of them possessed tetrahydromethanopterin S-methyltransferase (Mtr), the key enzyme for hydrogenotrophic methanogenesis [13]. Two of them contained Methylamine:Coenzyme M Methyltransferase (mtbA), suggesting that they might produce methane from methylamine [14]. These results were consistent with previous research showing active conversion of CO₂ into methane in hydrocarbon reservoirs [15].

Group 3d is associated with fermentative metabolism and chemoautotrophy, interconverting electrons between hydrogen and NADH depending on cellular redox state [8, 9]. Group 3d was the most abundant subgroup in 15 out of 25 groundwater samples (Fig. 1A). Most group 3d hydrogenase genes were close to an NADP oxidoreductase gene (Tables S13, S14). 3d hydrogenase genes were present in 89 high-quality MAGs, with 12 of them encoding formate C-acetyltransferases or lactate dehydrogenases, both signature genes of fermentative metabolism. Of these MAGs, 21 were associated with methanotrophic Methylomonadaceae. For the other 68 MAGs, 40 of them contained both RuBisCO and phosphoribulokinase, indicating a functional Calvin cycle. For instance, MAGs associated with Rhodocyclaceae (12), Hydrogenophaga (7), Nitrosomonas (7), and Rhodoferax (5) fall into this category. Thus, it is likely that these chemolithoautotrophs can use hydrogen as an additional energy source, with the hydrogenase transferring electrons from H₂ to NAD⁺ to drive their Calvin cycles.

Two newly discovered groups of [NiFe]-hydrogenases further expanded diversity. The first was positioned near the root of the tree (Fig. 2). This group consisted of three sequences, exclusively found in Methanobacteriaceae. The other newly discovered group was near the root of group 3b, composed of six sequences, including five sequences affiliated with Anaerolineaceae and one affiliated with Bathyarchaeia.

Consistency in the types/subgroups of hydrogenases and metabolisms among MAGs with the same taxonomic identity was observed for common groundwater residents, which helped to extrapolate metagenomic findings to 265 amplicon-sequenced groundwater samples. For example, the total relative abundance of Methylomonadaceae bacteria (all 21 MAGs had hydrogenases) could reach 88.6% (Table S15). Members of Hydrogenophaga (8 out of 14 MAGs had hydrogenases) could be as abundant as 71.2%. These findings suggest a high potential for hydrogen consumption in sampled subsurface habitats.

While the subsurface ecosystems analysed here would not be suitable for hydrogen storage, our study adds to growing evidence that hydrogenases are diverse, functional and ubiquitous in subsurface environments [16-19]. However, as hydrogenases were most abundant in methanogen proteomes, this need not always be a barrier to hydrogen storage, since recovery of methane could still be a desirable outcome. Likely, any subsurface environment at a temperature conducive to life would harbor microorganisms that thrive on hydrogen.

Acknowledgements

The authors thank the Groundwater Observation Well Network team members of Alberta Environment and Protected Areas (https://www.alberta.ca/groundwater-observation-well-network.aspx) for providing access to groundwater monitoring wells, sampling, and providing highest quality groundwater samples, and for sharing measurement results and expertise. We would like to thank the University of Calgary’s Center for Health Genomics and Informatics for sequencing and informatics services. We thank Carmen Li for help with MiSeq sequencing. We also thank Daan R. Speth for help with phylogenetic analysis.

Author contributions

SL, MS, and MD designed the study. SL, DM, AK, PH, BM, and MD performed lab research. SL, DM, MS, and MD analyzed the data. PH and BM assisted with data interpretation. SL, MS, and MD wrote the manuscript.

Conflicts of interest

None declared.

Funding

This study was supported by the Natural Sciences and Engineering Research Council (NSERC) through a Discovery Grant and Canada Research Chair (CRC-2020-00257 to M.S.), the Canada Foundation for Innovation (CFI), the Digital Research Alliance of Canada, the Canada First Research Excellence Fund (CFREF), the Government of Alberta, and the University of Calgary.

Data availability

Amplicons in this study are under the Bioproject PRJNA861683 and PRJNA700657. Metagenomes and metagenome-assembled genomes are under the Bioproject PRJNA700657 (NCBI). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository [20] with the dataset identifier PXD044305.

References

© The Author(s) 2024. Published by Oxford University Press on behalf of the International Society for Microbial Ecology.

Issue Section:

Brief Communication

Email alerts

Email alerts

• Google Scholar

Citing articles via

More from Oxford Academic