Methane emissions from aquatic ecosystems including, lakes, reservoirs, and wetlands, have been shown to be a substantial (Cole et al., 2007; Bastviken et al., 2011; Deemer et al., 2016), and poorly constrained, contribution to global greenhouse gas production (Walter et al., 2007; Engram et al., 2020). Part of the challenge of constraining methane emissions is their inherently heterogeneous nature, both within and across locations (Musenze et al., 2014; Podgrajsek et al., 2014; Beaulieu et al., 2016). For example, within a single reservoir system, hotspots of methane production occur that are double or triple the rate of methane emissions of other regions in the reservoir (Berberich et al., 2020). Across locations, a range of physical and chemical parameters contribute to variation in methane production and consumption (Sobek et al., 2012; Beaulieu et al., 2014; Deemer et al., 2016; Grinham et al., 2018), but these general features often do not explain variation within a given system. Biological parameters, such as bioturbation, may offer insight into the variability of methane emissions both across sites and within the different habitats of a given system.

Bioturbation is the disruption of sedimentary deposits by living organisms (Kristensen et al., 2012). This can include small to large-scale disturbances and can range from the impacts of burrowing invertebrates to predators (e.g., fish) that feed on these invertebrates. Though the scale of effects vary, bioturbation often results in the mixing of geochemical or redox gradients and the introduction of oxygen to anoxic sediments (Bertics and Ziebis, 2009; Kristensen et al., 2012). In freshwater sediments, large-scale disruption of sediment by organisms such as benthic feeding fish can also enhance methane ebullition (Frei et al., 2007; Carey et al., 2018; Brown and Hershey, 2019), the release of gaseous methane bubbles that are formed and trapped in super saturating sediment conditions (Joyce and Jewell, 2003). Ebullition can result in increased total methane emissions from sediments, likely due to the escape of methane that would otherwise diffuse and be consumed by methane-oxidizing organisms (Frei and Becker, 2005; Frei et al., 2007; Datta et al., 2009). However, under some conditions, fish bioturbation has been shown to reduce methane emission, likely due to the exposure of anoxic sediments to oxygen which can inhibit methanogenic organisms (Oliveira Junior et al., 2019; Booth et al., 2021). Thus, it is not well understood how the degree or frequency of bioturbation impacts methane emissions through changing the balance of methanogenesis and methanotrophy in sediments and the physical perturbance of sediment to release methane bubbles via ebullition, nor how the structure of the microbial communities is impacted by bioturbation.

In most ecosystems, methane production (i.e., methanogenesis) is an anaerobic process performed by specific classes of Archaea that fall into seven orders: Methanobacteriales, Methanococcales, Methanomicrobiales, Methanosarcinales, Methanocellales, Methanopyrales, and Methanomassiliicoccales. Under standard conditions, methanogenesis conserves a small amount of energy and is traditionally thought to occur only in environments where more energetically favorable terminal electron acceptors, such as oxygen, nitrate, iron, and sulfate are absent (Zinder, 1993). In these anoxic environments, methanogenesis is a major driver of organic matter degradation and plays a critical role in decomposition. In addition to energetic constraints (and the likelihood of being out-competed for resources), the biochemistry involved in generating methane is highly sensitive to oxygen (Thauer et al., 2008). Many of the iron-sulfur proteins and cofactors involved in methanogenesis are damaged in the presence of oxygen, or can result in oxygen radical formation, which further constrains the growth of methanogens (Zinder, 1993). Oxygen tolerance in methanogens grown in culture or from sediment mesocosms has been shown to vary, ranging from exposure that results in complete loss of methanogenesis with prolonged recoveries post exposure (Yuan et al., 2009, 2011), to methanogens that contain and express genes to deal with oxygen exposure (i.e., superoxide dismutase) (Tholen et al., 2007). Even in tolerant organisms, oxygen presents a stress that can inhibit or reduce methane production.

This work investigates how sediment geochemistry and microbiology are impacted by mechanical disturbance (mimicking the sediment disruption caused by fish bioturbation of the sediment), building off our prior work investigating impacts on greenhouse gas emissions (Booth et al., 2021). Using sediments derived from Acton Lake (near Oxford, OH, United States), we constructed sediment microcosms that were exposed to different frequencies of disturbance to mimic different levels of bioturbation. Gas ebullition and diffusion measurements were taken for each treatment (Booth et al., 2021). Booth et al. (2021) observed enhanced methane ebullition at intermediate (7-day) disturbance frequencies (Figure 1A). No statistical difference was observed in terms of diffusion of methane related to disturbance frequency (Figure 1B). By measuring the gas volumes in our gas traps pre- and post-disturbance, we could link 65–75% of the gas ebullition in each disturbance treatment to disturbance events (Booth et al., 2021). Notably, other greenhouse gases were monitored in this work—specifically, CO₂, N₂O, and NO—but methane was the only greenhouse gas that demonstrated a statistically significant difference under varying bioturbation frequency (Booth et al., 2021). We hypothesized that oxygen exposure introduced by disturbance inhibited methanogenic microbial communities at the highest disturbance frequencies, resulting in lower total methane emissions. At intermediate frequencies, disruption of the sediment drove the escape of methane gas that would otherwise remain trapped in sediment and resulting in elevated methane emissions to lower frequency experiments. Here we combine microbial community analysis and quantitative PCR (qPCR) targeting the methanogenic community to investigate the corresponding impacts of sediment disturbance frequency on methanogenesis in disturbed sediments.

The impacts of bioturbation on sediment geochemistry and microbial processes are dependent on the magnitude and frequency of the sediment disturbance, though the explicit relationships between different modes of bioturbation and specific processes remain poorly characterized. As shown in our previous work, we investigated the relationship between sediment disturbances mimicking benthic feeding fish and greenhouse gas production. We found that methane emissions were enhanced at intermediate disturbance frequencies, and this was driven by the ebullition of methane (Figure 1; Booth et al., 2021). Given that the physical disruption of the sediment was shown to drive the majority of methane ebullition from sediment [up to 75% of the total emissions (Booth et al., 2021)], we predicted that the intermediate disturbance frequencies balanced production of methane (by strictly anaerobic methanogens), with ebullition that is enhanced by mechanical disturbance, while limiting residence time of methane and therefore consumption. This work sought to test the hypothesis that disturbance impacts the methane producing community, most likely through the periodic introduction of oxygen into sediments that were previously anoxic.

Voltametric measurements of sediment demonstrate that oxygen is consumed within the first 5 mm of sediment for all bioturbation frequencies. However, one can reasonably assume that mechanical bioturbation would have oxygenated the subsurface through the depth of mixing (at least 20 mm). Our measurements demonstrate that the oxygen profile is re-established within 1-day post-disturbance with oxygen falling below detection at 5 mm. Future work is planned to more rigorously characterize the timing of redox changes post-bioturbation at shorter timescales. Though the specific magnitude and timing of oxygen introduction and subsequent consumption back to a steady-state gradient have yet to be characterized, even short oxygen exposures have been shown to inhibit a wide range of methanogenic species (Zinder, 1993; Thauer et al., 2008). Most of the biochemical pathway involved in methanogenesis is sensitive to oxygen, as it contains highly reactive metal containing co-factors which are either easily oxidized or can result in Fenton reactions which produce destructive free radicals (Cedervall et al., 2010). While some species have been shown to contain strategies for dealing with the presence of oxygen, often its presence results in a complete shutdown of metabolic activity (Kato et al., 1993; Yuan et al., 2009; Freitag et al., 2010). In culture, multiple taxa have been shown to shut down methanogenesis for up to 3 days (Kiener and Leisinger, 1983) upon oxygen exposure. In the environment, a shutdown in metabolic activity would impact an organisms’ ability to grow and/or compete with other organisms for nutrients. This is consistent with the trend we observed in this work, with methanogen abundance decreasing with an increase in disturbance frequency as this would also increase the frequency of oxygen exposure to methanogenic organisms in the impacted area (i.e., 1 cm depth) compared to a region not impacted by disturbance (i.e., 3 cm).

The total microbial community was not significantly altered by the different disturbance treatments, however there was a trend observed in the methanogen-specific community that demonstrated a marked relative increase in a putatively oxygen-tolerant methanogen phylotype, candidatus Methanothrix paradoxum (Angle et al., 2017). Given that 16S rRNA abundance of ca. Methanothrix paradoxum did not change on an absolute scale (Figure 6), this suggest that other methanogen populations were more strongly impacted by the disturbance treatments. This phylotype was previously identified from sediments in an oxic and methane producing wetland, in Old Woman Creek. Other members of the Methanothix genera (previously Methanosaeta) are known for acetoclastic methanogenesis, often with a high affinity for acetate (Patel and Sprott, 1990; Ma et al., 2006). Candidatus M. paradoxum was named for the conundrum of it’s ability to make methane under functionally oxic conditions (Angle et al., 2017). Metatranscriptomic work for Old Woman Creek supported the active transcription of methanogenesis genes from oxic sediments and showed no evidence of oxygen exposure/stress. Nonetheless, given the genomic features, the authors suggest that it is unlikely that candidatus M. paradoxum is immune to oxygen, though perhaps can generate anoxic niches in these functionally oxic systems that facilitate generation of methane. Alternatively, these organisms could contain other not well characterized strategies for dealing with oxygen exposure that allow them to persist and actively generate methane in the top few cm of sediment impacted by disturbance.

One caveat to this work is that fish disturbance of sediments is linked to feeding behaviors and thus linked to the removal of invertebrates As reported in our prior work (Booth et al., 2021), the invertebrate communities were composed of oligochaetes, with relatively few stonefly larvae (Plecoptera) and snails. The majority of the invertebrate communities were composed of oligochaetes and no difference in the invertebrate communities were demonstrated due to disturbance (Booth et al., 2021). However in recent work looking at the impact of different fish and macroinvertebrate impacts on greenhouse gas emissions, invertebrate feeding fishes that reduced the number of tubifex worms were shown to increase mcrA gene copies, likely by reducing oxygen exposure from worm activities in the top cm of sediment (Colina et al., 2021). Notably fish activities in the Colina et al. (2021) work, did not impact ebullition of methane. This likely speaks to the minimal sediment disturbance due to the fish species and sizes used, which further speaks to the potential importance of degree of disturbance linked to fish activities. Additionally, the specific methanogen species were not quantified in the Colina et al. (2021) work and thus insights into their susceptibility to oxygen which may also play a role in the mcrA copy number dynamics (indicative of methanogen populations) are difficult to determine. For example, we highlight that the top few centimeters of sediment is an active and important region for methanogenic activity and makes significant contributions to the overall methane emissions in sediment. In this work no effects on the community or methanogenic population at 3 cm depths were noted, indicating that the quantifiable differences in methane emissions between disturbance treatments was due to microbial activity in the shallow sediment. As a difference was noted at 1 cm depth, this supports the important contributions of shallow methanogen populations to methane production, however the composition of the methanogenic microbial community and their corresponding tolerance to oxygen will likely impact the methanogenic activity in these shallow sediments.

Overall, this work supports the hypothesis that mechanical disturbance, or more particularly, oxygen exposure caused by bioturbation, can affect the methanogenic community of freshwater sediments and impact methane emissions. However, this work also supports that multiple factors are impacting methane emissions. We found a linear trend between the decrease in methanogens and an increased frequency of disturbance. This matched the trend seen in greenhouse gas emissions for high to moderate frequency disturbances. Specifically, 3 days of disturbance had significantly lower methane emissions than 7 days. However, the later time points had a trend in higher methanogen populations, but lower emissions than the 7-day treatment. We predict that the longer residence time of methane supported more methane diffusion (relative to ebullition) and stimulated the activity of methanotrophic populations which acted to reduce total methane emissions. We were unable to quantify a difference in the methanotrophic population across treatments at 1 cm sediment depths. However, this does not rule out that methanotrophic organisms in surface sediments with more consistent oxygen exposure are driving the enhanced methane consumption at lower disturbance frequencies. This hypothesis will be more fully tested in our follow up work which will broaden the range/depths of sediment samples tested, expand the frequencies of disturbance, and look more directly at the impacts of fish-specific bioturbation. Though this work is preliminary, given the interesting patterns observed and the likelihood of multiple processes impacting the methane emissions, it points to developing a mechanistic framework to better understand the impacts of processes like disturbance, bioturbation and sediment microbial ecology on the heterogenous nature of methane emissions in a range of freshwater ecosystems.

The datasets presented in this study can be found in online repositories. The nucleic acid data presented in this study are deposited in the NCBI repository, under bioproject number PRJNA766591 with associated SRA and biosample information.

AR helped plan the experiments, supervised lab work, mentored MU and LT, performed the data analysis and interpretation, and authored the manuscript. MU performed the mesocosm studies, collected the samples, extracted nucleic acids, helped with the data analysis, and provided the editorial comments. LT supervised nucleic acid work, helped with the data analysis, and provided the editorial comments. JS performed the redox sediment measurements, performed the data analysis and interpretation, and provided the editorial comments. GD mentored JS, helped with the data analysis and interpretation, and provided the editorial comments. MB designed the experimental set up, acquired funding, mentored MU, helped with the data analysis and interpretation, and provided the editorial comments. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.