Water column dynamics control nitrite-dependent anaerobic methane oxidation by Candidatus “Methylomirabilis” in stratified lake basins

Hydrochemistry and methane oxidation rates

The water column of the deep North Basin is considered meromictic (i.e., permanently stratified). At the time of sampling for methane oxidation rate measurements in November 2016, the redox transition zone extended from 79 to 105 m depth; as defined here, with an upper boundary set at O₂ < 5 µM, and a lower boundary set where oxygen sensitive reduced chemical species like Fe²⁺ or H₂S start to rise above background levels (Fig. 1A, Supplementary Fig. S3). Methane concentrations increased from very low levels (<0.1 µM) at the upper redox transition zone to 20 µM at 155 m. At this depth, nitrite was below the limit of detection (0.02 µM) and nitrate in the sub-micromolar range (<1 µM, Fig. 1B).

Vertical profiles of potential rates of methane oxidation (MOR) in the North Basin showed a bimodal pattern (Fig. 1C). The upper peak of methane oxidation rates (0.06 ± 0.01 µmol L⁻¹ d⁻¹) at 100 m was at least partly due to the activity of aerobic methanotrophs, thriving under microoxic conditions, as reported before [3]. However, methane oxidation continued below the redox transition zone, and a secondary methane oxidation rate maximum of 0.08 ± 0.07 µmol L⁻¹ d⁻¹ was detected at 125 m. A similar bimodal methane oxidation rate distribution has been observed in the North Basin water column before, whereby, at the time, both the oxycline and the two separate methane oxidation rate maxima were located at greater depths [3].

In the eutrophic South Basin, seasonal near-bottom anoxia typically starts to develop in early summer, and is associated with elevated turbidity in the deep waters. This benthic nepheloid layer extends from the lake ground up to the oxic/anoxic interface and consists of microbial biomass, produced locally and to large parts by methanotrophs [4]. Its development starts at the sediment-water interface and then progressively expands 10–20 m into the water column, following the rising redox transition zone [4]. During the time of sampling, in November 2016, the redox transition zone extended from 55.5–75 m (Fig. 2A and Supplementary Fig. S4), below which, in contrast to the North Basin, considerable amounts of nitrate (38–73 µM) and nitrite (1.2–3.9 µM) were present (Fig. 2B). Methane concentrations below the redox transition zone increased towards the sediment and reached levels that were comparable (28 µM) to those in the North Basin. In the South Basin, a single peak of methane oxidation rate (0.18 ± 0.1 µmol L⁻¹ d⁻¹) was observed at 70 m depth, towards the lower boundary of the redox transition zone (Fig. 2C). Although Type I methane-oxidizing bacteria (MOB) were shown previously to dominate the biomass in the benthic nepheloid layer, where the highest methanotrophic activity was observed, it remained unclear whether the observed activity was soley due to aerobic methanotrophs [4]. Particularly the presence of both nitrate/nitrite, but also sulfate, Fe(III)- and Mn(IV)oxides (Supplementary Fig. S4) within the benthic nepheloid layer bears the potential that methane could be oxidized anaerobically with either of these oxidants.

Methane oxidation within oxic/anoxic transition zones of other stratified lakes has often been attributed to aerobic methanotrophs [6, 70]. In lakes with shallow redox transition zones, cryptic oxygen production by phototrophs could sustain aerobic methane oxidation even in seemingly anoxic waters [5, 71]. At the depths of the redox transition zone in Lake Lugano, particularly in the North Basin, oxygen production by phototrophs is an unlikely mechanism. Alternatively, Blees et al. [3] suggested that aerobic methane oxidizers below the redox transition zone can survive prolonged periods of oxygen starvation, and can resume high methane oxidation activity upon episodic downwelling of oxygen, for example during cooling events. Yet potential mechanisms that inject oxygen to the deep hypolimnion were not investigated, and it remained speculative if, and to which depth, such events may occur. Thus, methane oxidation below the redoxcline in the Lake Lugano North Basin may indeed be anaerobic.

Evidence for nitrate/nitrite-dependent AOM

To test for the presence of active anaerobic methanotrophs, and to indentify potential oxidants for methane, we set up anoxic incubation experiments with ¹⁴CH₄ as substrate, different electron acceptors (nitrate, nitrite, sulfate), and concentrated biomass. The biomass was collected from 85–90 m in the South Basin, a depth well below the redox transition zone at this time of sampling, but where nitrate, nitrite, and sulfate were present. Biomass from the North Basin was collected at 105–110 m, where nitrite was undetectable but low levels of nitrate and sulfate were still present.

When biomass from the meromictic North Basin was used, we found that both nitrate (p < 0.01) and nitrite (p = 0.05) stimulated AOM rates significantly (Fig. 3, Table S4). Compared to the controls without additions, AOM rates increased by an average of 62% and 43% in the presence of nitrate (57.8 ± 10.8 µmol L⁻¹ d⁻¹) and nitrite (50.9 ± 10.8 µmol L⁻¹ d⁻¹), respectively. There was no significant difference between the controls and amendments with sulfate (Table S4B).

In the South Basin, no significant stimulation of methane oxidation occurred with any of the added electron acceptors (Fig. 3 and Table S4), suggesting that these are not immediate oxidants for methane, and AOM was not a major mode of methane removal in spite of the presence of nitrate, nitrite, and sulfate. However, the methane oxidation rates in the South Basin incubations increased with a longer incubation time (i.e., 32 days, data not shown), independently of the added compounds, including molybdate. The observed stimulation of AOM thus seems independent of sulfate-reducing bacteria. In turn, the question arises as to what was the electron acceptor supporting AOM under these conditions, and more generally, what was the oxidant for methane in the unamended controls? Non-zero AOM rates in controls are frequent [3, 72, 73], possibly resulting from minor oxygen contamination at the start of the experiment or during sampling. Although greatest care was taken to prevent oxygen contamination during preparation, incubation, and sampling, by executing all steps in a nitrogen-flushed anoxic chamber, we did not add a reducing agent to chemically remove any traces of oxygen. Thus, if trace amounts of oxygen were still present in the incubations, they might have served as substrate for the methane monooxygenase. After the oxygen-dependent initial hydroxylation of methane to methanol, its further transformation could proceed anaerobically by fermentation, whereby hydrogen, formate, acetate, and other compounds are produced [74]. While trace oxygen contamination may be the reason for non-zero AOM rates in the living controls, they can not explain the increase in AOM rates between 16 days and 32 days of incubation with South Basin biomass. A possibility, inferred from these incubation experiments, is that the methanotrophs oxidized methane with Fe(III)- or Mn(IV)-oxides [16, 18, 19], which were present in the South Basin (Supplementary Fig. S4). Growing experimental and metagenomic evidence suggests that fermentation and the potential for extracellular electron transfer (e.g., to Fe(III)- or Mn(IV)-oxides) are widespread among freshwater MOB [75, 76]. Particulate metal oxides would accumulate on the filters used to concentrate MOB biomass for the incubation experiments. Close spatial arrangement or even direct contact between methanotrophs and insoluble metal oxides on the filters may have stimulated AOM rates with time. Whereas this hypothesis still needs to be tested for Lake Lugano, incubation experiments with biomass from the North Basin clearly showed that both nitrate and nitrite enhanced methane oxidation under anoxic conditions, providing evidence for active N-AOM in the meromictic North Basin.

Abundance and diversity of methanotrophic bacteria

Methanotrophy is an important biogeochemical process at the redox transition zones of both lake basins. Up to 32.2% of the 16 S rRNA gene amplicons in the benthic nepheloid layer of the South Basin, and 11.2% at 95 m in the North Basin were related to methanotrophs (Supplementary Excel Table S1). Among the 41 identified ASVs of putative methanotrophs, gamma-proteobacterial type I MOB (31 ASVs) were by far the most important group in terms of relative abundance and diversity. Also, eight ASVs of alpha-proteobacterial type II MOB were detected, but they were generally low in abundance. Furthermore, two identified ASVs (ASV9 and ASV7279) were related to Ca. Methylomirabilis (Supplementary Excel Table S1), capable of mediating AOM with nitrite as oxidant [14]. No sequences of anaerobic methane-oxidizing Archaea, such as Ca. Methanoperedens, or representatives of the ANME-1, -2 or -3 groups, were detected in any of the samples.

The guild of methanotrophs in the lake was dominated by only seven highly abundant ASVs, with >1% relative abundance in at least one sample, including uncultured representatives of Methylobacter sp (ASV5, ASV18, ASV19, ASV42), Crenothrix sp. (ASV10, ASV91), as well as Ca. Methylomirabilis (ASV9). These seven taxa combined represented >96% of all sequence reads of methanotrophs (Supplementary Fig. S5, Supplementary Excel Table S1). All but Ca. Methylomirabilis were present in both the North Basin and the South Basin water column, coexisting, at varying proportions, respectively, in microoxic as well as in anoxic water layers (Fig. 4). Ca. Methylomirabilis, however, showed a clear habitat preference for the meromictic North Basin, where ¹⁴CH₄ incubations indicated active N-AOM, and where Ca. Methylomirabilis was with 6.5% at 95 m the most abundant MOB in November 2016 (Fig. 4).

All three MOB groups, Methylobacter sp, Crenothrix sp., and Ca. Methylomirabilis coexisted in the North Basin water column, where methane consumption rates were highest (Fig. 1, Supplementary Figs. S5 and S6), suggesting that both aerobic and anaerobic methanotrophs work in concert under the microoxic to anoxic conditions in the redox transition zone, and form an efficient pelagic methane filter (Fig. 1). The methane oxidation rate maximum coincided with the maximum abundance of Ca. Methylomirabilis, but nitrite, the prime electron acceptor for this organism was undetectable below 85 m depth at the time of sampling. Nitrite could be supplied by ammonia oxidizing archaea or bacteria (Supplementary Fig. S7), or, alternatively, by nitrate-reducing bacteria, as has been proposed previously to meet the nitrite demand of anammox bacteria [47]. In support of the second hypothesis, we find, by correlation-based network analysis, that Ca. Methylomirabilis forms a subnetwork together with some nitrate-reducing bacteria, including the denitrifying bacterium Sterolibacterium sp. (Fig. 5). In contrast, the typical nitrifiers are associated with, probably degrading, algal biomass. The other main methanotrophs, Crenothrix sp. and Methylobacter sp., form a separate subnet together with Methylotenera sp., suggesting a potential syntrophic interaction of methano- and methylotrophs as recently shown experimentally by ¹³CH₄ based DNA-SIP [46].

In the eutrophic, seasonally stratified South Basin, 16 S rRNA gene sequences of Ca. Methylomirabilis were not detected in the water column, although the chemical conditions seemed favorable for N-AOM, with nitrate and nitrite concentrations reaching 73 µM and 3.9 µM, respectively (Fig. 2). The two lake basins are hydrologically and microbiologically connected, and Ca. Methylomirabilis was present in South Basin surface sediments (Fig. 4C). The absence of Ca. Methylomirabilis in the South Basin water column is thus not the consequence of limited dispersion, or unsuitable chemical conditions. Rather, it is the limited period of time, during which these chemical conditions prevail, which prevents Ca. Methylomirabilis—and other slow growing organisms such as anammox bacteria [47]—from forming a stable and sizeable population in this basin. Indeed, the anoxic stratification period of ~5 months is short compared to an apparent doubling time of 104 days estimated for Ca. Methylomirabilis in the North Basin (Supplementary Fig. S8), and faster growing, and potentially metabolically more versatile taxa outcompete Ca. Methylomirabilis. In accordance with Blees et al. [4], we find that the assemblage of methanotrophs at the depth interval of maximum methane oxidation (65–75 m) was composed of Methylobacter sp. (e.g., ASV5, 14.5% and ASV18, 4.0%) followed by Crenothrix sp. (e.g., ASV10, 2.8%) (Fig. 4B and Supplementary Excel Table S1). This latter ASV10 is highly similar to Crenothrix sp. D3, which has been shown to contribute to methane oxidation in two other stratified lakes in Switzerland [77]. Despite their occurrence in anoxic waters, these gamma-proteobacterial MOB are still considered aerobic, as they use molecular oxygen for the particulate methane monooxygenase and the initial activation of methane. The apparent absence of true anaerobic methane oxidizers in the South Basin water column is consistent with the lack of significant stimulation of AOM with nitrate/nitrite in the incubations with biomass from this basin (Fig. 3 and Table S4). Nonetheless, genomes of several aerobic methanotrophs, including Crenothrix, encode putative nitrate (narG, napA), nitrite (nirS, nirK), and/or nitrogen oxide reductases (norB) [72, 75, 77, 78]. Methylomonas denitrificans, for example, can couple the oxidation of methane (and methanol) to the reduction of nitrate to nitrous oxide under severe oxygen limitation [78], but the oxidation of methane under completely anoxic conditions has, to our knowledge, not been demonstrated experimentally for Crenothrix or any of the gamma-proteobacterial MOB. Nonetheless, although a plausible metabolic mechanism is still missing, they have been shown to incorporate ¹³C into their membrane lipids during biosynthesis from isotopically labeled methane in anoxic lake-sediment incubations [79].

Water column stability as an ecological factor fostering nitrite-dependent anaerobic methane oxidation

After nearly 50 years of meromixis in the North Basin of Lake Lugano, two exceptionally strong mixing events in 2005 and 2006 led to a complete ventilation of the water column and a massive reduction of the pelagic methane inventory from 2800 tons to 3 tons within just 1 month [50, 51]. The water column re-stabilized rapidly in the following year, and remained stratified, with anoxia below 100–125 m depth until today (Fig. 6A). As a consequence of the intrusion of oxygen to deep hypolimnetic waters, ammonium was also oxidized almost completely, leading to transiently increased nitrite and, subsequently, nitrate concentrations below 125 m depth (Fig. 6B). While we have no DNA data from the mixing period itself, in 2009 we observed unusual vertical distribution patterns and high relative abundances of three dominating groups of MOB within and below the redox transition zone: Methylobacter sp., Crenothrix sp., and Ca. Methylomirabilis (Fig. 6C). Water column mixing, e.g., during fall overturn in eutrophic lakes, can lead to blooms of aerobic methanotrophs in the entire water column [80]. We suggest that this also happened during the water column mixing in 2005 and 2006, and that the assemblage of MOB at still high relative abundances below the redox transition zone is partly a legacy signal from that major bloom event some years before. While in the oxic water column, such signals are readily eliminated by the regular community turnover, it may have prevailed for a longer period of time under anoxic conditions, where community turnover is slower due to reduced or absent grazing pressure by protists. Moreover, increased levels of oxygen and nitrate below 125 m until 2008 (Fig. 6B), may have helped maintaining the high relative abundances of the three MOB groups.

In 2010, in contrast to the previous years, a distinct pattern was observed for the declining Ca. Methylomirabilis population with respect to the other MOB (Fig. 6C). This could hint to a narrower ecological niche for Ca. Methylomirabilis, depending more strictly on the presence (or cryptic formation) of oxidized nitrogen compounds, which, by 2010, had been depleted below 125 m (Fig. 6B). On the other hand, the persistence of Methylobacter sp. and Crenothrix sp. also in deeper waters, could be explained by a metabolic lifestyle independent of external electron acceptors and, possibly, even of methane [75]. Indeed, mixed-acid fermentation and hydrogen production has been proposed previously based on metagenomic evidence for both Crenothrix D3 from lake Zug [77], as well as for Methylobacter sp. [46].

More stable water column conditions between 2014–2016 led to a shallowing of the redox transition zone with only few transient injections of oxidants into water depths below 125 m (Fig. 6B). During that period of time, Ca. Methylomirabilis was consistently peaking within the redox transition zone at 110 m, 100 m, and 95 m depth, respectively, representing up to 3.6% of total sequences in September 2014, 5.4% in June 2015, and 6.7% in November 2016. The read-based relative abundances of Ca. Methylomirabilis were confirmed by qPCR (Supplementary Figs. S8 and S9), showing that Ca. Methylomirabilis 16 S rRNA gene abundances increased from 1.5·10⁴ copies/mL in September 2014 to 3·10⁶ copies/mL in November 2016. By that time, Ca. Methylomirabilis had become the most abundant MOB in the redox transition zone, suggesting that water column stability was a major environmental factor promoting the growth of this slow-growing nitrite-dependent anaerobic methanotroph.

Links between water column stability and the growth of Ca. Methylomirabilis seem additionally confirmed by the most recent data. Enhanced mixing in 2017 and 2018 resulted in a deepening of the redox transition zone (Fig. 6A, C), and the decline and spreading of Ca. Methylomirabilis (Fig. 6C). While the ventilation of the upper monimolimnic water levels led to increased nitrate concentrations from stimulated nitrification (Fig. 6B), Ca. Methylomirabilis did not seem to benefit from this situation, likely because the detrimental effects of oxygenation due to deeper mixing seemed to outweigh the positive effect of the enhanced oxidant supply. Indeed, experimental studies with a Ca. Methylomirabilis enrichment culture have revealed that even in presence of low oxygen levels, rates of methane and nitrite conversion are strongly reduced, and cell-division-associated genes are downregulated [81]. In contrast, the more dynamic water column conditions in 2017 and 2018 stimulated the faster growing aerobic MOB, Methylobacter sp. but also Crenothrix sp., which also dominate in the dynamic South Basin water column.

While under stable water column conditions, oxygen and methane are brought together through diffusive processes, and at low concentrations, mixing events lead to transient but often strongly increased solute concentrations that affect directly the kinetics and pathways of biogeochemical processes [82]. As for methanotrophic communities, stable environmental conditions with low substrate concentrations thus are selecting for MOB with higher enzyme affinities but slower growth. Conversely, mixing-induced higher substrate levels will select for fast responding, aerobic methanotrophs. It is thus likely the interplay between stable and intermittently perturbed environmental conditions, at the redox transition zone of the meromictic North Basin, that allow the long-term coexistence of aerobic and anaerobic MOB, Methylobacter sp., Crenothrix sp, and Ca. Methylomirabilis.

Source: Ecology - nature.com