Activity and abundance of methane-oxidizing bacteria on plants in experimental lakes subjected to different nutrient and warming treatments

Shallow lakes produce and emit substantial amounts of methane (CH 4 ). Part of the CH 4 produced in lakes is consumed by methane-oxidizing bacteria (MOB) present in the sediment and water column, thus reducing the overall CH 4 emissions. However, the role of aquatic plants as habitat for CH 4 oxidation by MOB is poorly understood. In this study, we compared CH 4 oxidation rates and MOB abundance associated with different types of aquatic plants (periphyton, filamentous algae


Introduction
Freshwater ecosystems are globally the largest source of natural or anthropogenic emissions of the greenhouse gas methane (CH 4 ) (Saunois et al., 2020), with average emissions of 398 Tg CH 4 emission per year (Rosentreter et al., 2021).Most of these emissions originate from lakes (Saunois et al., 2020), especially shallow lakes and ponds (< 0.001 km 2 ), which cover around 9 % of the global freshwater area (Holgerson and Raymond, 2016) and are estimated to produce about 37 % of total lake CH 4 emissions via the diffusive and ebullitive flux (Rosentreter et al., 2021).
In shallow lakes, most CH 4 is produced by microbial methanogenesis carried out by methanogenic archaea in the anoxic zone of the sediment (Bastviken et al., 2008).The produced CH 4 can be oxidized to CO 2 by anaerobic methane-oxidizing archaea or bacteria (Martinez-Cruz et al., 2018) or by aerobic methane-oxidizing bacteria (MOB) (Oremland et al., 1992;Wu et al., 2011).Methane-oxidizing bacteria can oxidize up to 90 % of the CH 4 reaching the oxic zone, creating an effective CH 4 filter (Bastviken et al., 2004(Bastviken et al., , 2002;;Oremland et al., 1992).Methane that escapes this oxidation filter is emitted to the atmosphere through diffusion or is released directly from the sediment as CH 4 -rich bubbles, called ebullition (DelSontro et al., 2011).
Previous studies have demonstrated that nutrient enrichment and the increase of dissolved organic carbon (DOC) derived from algae can promote CH 4 release from shallow lakes (Aben et al., 2017;Beaulieu et al., 2019;Deemer and Holgerson, 2021;West et al., 2016;Zhou et al., 2019).At the same time, although the mechanisms are still not clear, the presence of macrophytes is generally accompanied with a reduction in the rates of CH 4 emission (Benelli and Bartoli, 2021;Davidson et al., 2015;Li et al., 2021), with a marked decline in CH 4 ebullition when the abundance of macrophytes is high (Davidson et al., 2018).The decline in CH 4 ebullition could be connected to increased oxidation of the sediment root zone by radial oxygen losses from the macrophytes (Laanbroek, 2010).In addition, previous studies have demonstrated the presence and activity of MOB in biofilms from freshwater canals (Pelsma et al., 2021), but it is still unknown whether MOB are associated with filamentous algae and what their contribution to reducing dissolved CH 4 in the water column might be.Furthermore, MOB are present and active on both above and below ground parts of aquatic macrophytes (Cui et al., 2022(Cui et al., , 2020;;Faußer et al., 2012) and it has been demonstrated that MOB associated with submerged aquatic macrophytes can also reduce CH 4 emission (Sorrell et al., 2002;Yoshida et al., 2014), but the range of CH 4 -oxidation potential differs among species and the studied tissues (Heilman and Carlton, 2001).However, it is still not clear how different nutrient loading and warming treatments can influence MOB presence and activity associated with submerged aquatic plants and their contribution to oxidizing the produced CH 4 .
Here, we sought to address the following questions: (1) how do different functional groups of aquatic plants in shallow lakese.g.periphyton, filamentous algae and macrophytesdiffer in their potential CH 4 oxidation and MOB abundance?(2) How is MOB abundance and activity affected by long-term eutrophication and warming?(3) Are there other factors, such as plant species identity or different concentrations of CH 4 in the water that can create differences in the potential CH 4 oxidation of submerged aquatic plants?
Two nutrient levels and three warming scenarios have been continuously applied to the mesocosms, resulting in a total of six treatments, each with four replicates.The nutrient treatments consist of a low nutrient treatment, where no additional nutrients are added to the source water, and a high nutrient treatment, where 2.7 mg P m − 2 day − 1 and 108.6 mg N m − 2 day − 1 are added weekly to the mesocosms as Na 2 HPO 4 and Ca(NO 3 ) 2 solutions.The warming treatments consist of AMB (ambient), A2 (+2-3 • C) and A2 + (A2 +50 %, +3-5 • C).The degree of warming is based on scenario A2 of the regional Intergovernmental Panel on Climate Change (IPCC) (Cooper et al., 2002;Houghton et al., 2001;Liboriussen et al., 2005).
Samples of periphyton, filamentous algae and macrophytes from above-ground tissue and the rhizosphere were collected from the mesocosms.Potential CH 4 oxidation rates of the different plant types were determined through in vitro oxidation incubations and the abundance of MOB associated with them, through quantitative PCR.

Periphyton, filamentous algae and macrophytes sampling
Several plant types from each mesocosm were sampled between 24 August and 4 September 2020, and their potential CH 4 oxidation rate quantified.Specifically, we collected periphyton (defined as attached algae according to Vander Zanden and Vadeboncoeur, 2020), submerged filamentous algae and submerged macrophytesdivided into above-ground tissue and the rhizosphere (see Supplementary methods).
However, samples of all plant types could not be collected from all 24 mesocosms due to the absence of filamentous algae or macrophytes in some of them (Table 1).
To collect the periphyton samples, we sampled from metal strips (4 cm wide and 93.5 cm long) which have hung on the inside wall of each mesocosm since 2010.Two metal strips per mesocosm were sampled with the biomass scraped off with a razor blade from 0 to 20, 20-50 and 50-70 cm of depth below the water surface and then mixed.Around 2.5 g wet weight was collected for each strip in duplicates for each mesocosm.
Submerged filamentous algae were present in eight of the 24 mesocosms (five low nutrient, of which two AMB, one A2, two A2 + temperature treatment and three high nutrient ones, all A2 + temperature treatment, Table1), and in each of them 5 g wet weight of material was collected in duplicates for each mesocosm.
The species of submerged aquatic macrophytes present in the mesocosms are Potamogeton crispus L. (P.crispus) and Elodea canadensis Michx (E.canadensis).Submerged aquatic macrophytes are generally more frequent and abundant in the low nutrient mesocosms than in the high nutrient mesocosms.Conversely, the chlorophyll a concentrationan indicator of algal abundanceis generally higher in the high nutrient mesocosms.Macrophyte surveys were conducted separately for upper 50 cm and lower 50 cm of the 100 cm water column; thus, two portions for each mesocosm.For each portion, macrophyte coverage was estimated and macrophyte height was measured for each species (P.crispus and E. canadensis).Macrophyte abundance was quantified as percent volume inhabited (PVI) of the water column, based on coverage ( %) * macrophyte height (cm)/water depth (cm).The conversion between P. crispus PVI and biomass was based on Lauridsen (1990) and Lauridsen et al. (1994).For E. canadensis, the conversion was based on macrophytes harvested at the LMWE during 2004 (Lauridsen, unpubl. results).Accordingly, the harvested plant material was dried at 105 • C (P. crispus) or 60 • C (E. canadensis) for 24 h and then weighed.The dry biomass (g) of E. canadensis and P. crispus per mesocosm was calculated using the equations: y = 15.573* PVI, and: y = 0.7642 * PVI, respectively.
Aquatic macrophytes were present in 12 mesocosms and in three of them both species of macrophytes occurred.Potamogeton crispus was present in eight mesocosms of which, two were high nutrient mesocosms (one exposed to AMB and one to A2 + temperature treatment) and six were low nutrient (one exposed to AMB, two to A2 and three to A2 + temperature treatment).E. canadensis was found in seven low nutrient mesocosms (four exposed to AMB, one A2 and one A2 + temperature treatment, Table 1).Both species of macrophytes were collected if present in the same mesocosm.Samples were collected by hand, where possible, by plucking out the entire plant including the rhizosphere.We then placed the plant material directly in a tray and carefully separated it into above-ground tissue and rhizosphere, the latter defined as the roots and the soil adjacent to them (Bowen and Rovira, 1999).No biofilm was removed during the separation of the biomass, and the rhizosphere was gently rinsed in water from the respective mesocosm so that the loosely attached sediment was removed, but the rhizosphere was left largely intact before weighing.Five g (wet weight) of above-ground material (stems and leaves) was collected, with triplicates for each sample.For the rhizosphere, 2.5 g (wet weight) was collected, in duplicate for each mesoscom.
Straight after the collection, wet samples of periphyton, filamentous algae and macrophytes were weighed and placed in 120 ml glass bottles for the above-ground macrophyte tissue and in 20 ml for all the other plant types to set up oxidation incubations.For each sample, a subsample, which was kept frozen at − 20 • C in 5 ml Eppendorf tubes, was taken for molecular analysis (for details see Section 2.4).

Determination of potential CH 4 oxidation associated with periphyton, filamentous algae and macrophytes
Potential CH 4 oxidation rates associated with each plant type were measured by in vitro oxidation incubations.The material (2.5-5 g wet weight) was incubated in 20 or 120 ml glass bottles, depending on the plant type.The bottles contained the plant material exposed to ambient air and no water was added to the incubations.Afterwards, the bottles were closed with gastight butyl rubber stoppers and capped.Using previous studies as guidelines (Sorrell et al., 2002) and considering the dissolved concentrations of CH 4 in the mesocosms in the previous years between 2011 and 2019 (mean = 0.16 µmol CH 4 -C L − 1 , median = 0.25 µmol CH 4 -C L − 1 , Davidson, unpublished), the samples were all incubated with a headspace concentration of 150 ppm CH 4 , which corresponds to dissolved concentration of 0.22 µmol CH 4 -C L − 1 .The incubations were started within 2 h of the sample collection.The incubated samples were placed on a Laboshake shaker (Gerhardt, Germany) at 150 rpm in darkness for 12 h at 25 • C. The CH 4 concentration in the headspace was measured every two hours, t = 0, 2, 4, 6, 8, 10, 12 h, by directly injecting 30 µl of gas sample in a gas chromatograph SRI 8610 C (SRI Instruments, Earl St. Torrance, California, USA) equipped with a 3 ′ x 1/8" Silica Gel Packed Column and a flame ionization detector (FID).Certified CH 4 standards were used for calibration and validation.After the incubations, samples were dried at 60 • C for 24 h.Potential CH 4 oxidation rates were calculated from the slope of the linear regression of the decreasing CH 4 in the incubations and divided by the dry weight of the plant material added.To avoid underestimation of oxidation rates in case all the CH 4 present in the bottles was oxidized by the end of the incubations, only the points during which there was still a linear decrease in CH 4 were used to determine CH 4 oxidation rates and measurement points afterwards were discarded.In this case, at least 3 data points were used to determine the slope of the oxidation rates, else, all the 7 data points from the every-two-hour measurement were considered.
For the above-ground macrophyte tissue, we also determined the total CH 4 oxidation activity per mesocosm as µmol CH 4 h − 1 m − 2 .This was calculated from the product of the results from the oxidation incubation (µmol CH 4 h − 1 g − 1 ) and the total dry biomass of macrophytes per mesocosm as g m − 2 .Due to the paucity of samples in the high nutrient mesocosms, only samples present in the low nutrient mesocosms were considered to compare CH 4 oxidation potential among the two macrophyte species.When both species of macrophytes were present in the same mesocosm, the CH 4 oxidation activity of each species was then multiplied by its own biomass and the final value for each mesocosm was then given by the sum of the result of each species.
For the filamentous algae, due to the paucity of samples present at different nutrient and temperature levels, it was not possible to investigate the effect of nutrients or long-term warming on their CH 4 oxidation potential rates.

Molecular analysis
The extraction of DNA from periphyton, filamentous algae and macrophytes above-ground tissues and rhizosphere was done using the DNeasy PowerSoil kit (Qiagen, USA) with some modifications to optimize the DNA yield (see Supplementary methods).
We measured copy numbers of pmoA, a gene encoding the alpha subunit of particulate CH 4 monooxygenase (pMMO), which catalyses the initial step of CH 4 oxidation (Semrau et al., 1995), as a proxy for MOB abundance using quantitative PCR (qPCR), with the primers A189F and mmb661R (Costello and Lidstrom, 1999).The cycle had 3 min denaturation at 95 • C, 45 cycles of 10 s 95 • C, 15 s 58 • C, 25 s 72 • C and 10 s 82 • C, ending with a melt-curve in steps of 1 • C from 70 • C to 95 • C (Costello and Lidstrom, 1999).The DNA samples were diluted, and 1 ng of DNA was added to all reactions.All samples and standards were measured in triplicate for every run.qPCR was performed on a BioRad iQ5 Multicolor Real-Time PCR detection system (Vers.2.0, BioRad, Sweden) with iQ SYBR® Green Supermix as mastermix.Finally, we quantified pmoA copy numbers using a calibration curve, with a range from 10 to 10 6 copy number g − 1 dry weight (dw), and confirmed amplification specificity from melt curves.However, amplification of the pmoA gene was not possible in the above-ground macrophytes; due to the paucity of collected biomass, the yields of extracted DNA from the above-ground tissue of macrophytes were not enough to run amplifications.
Nutrients and long-term warming effect could not be investigated on the abundance of MOB in filamentous algae, due to the small sample size among treatments.For the macrophytesrhizosphere, only samples from low nutrient treatments were considered for the statistical analysis.

Table 1
Overview of the 24 mesocosms divided according to nutrient and temperature.For each mesocosm, it is indicated from which plant type the sample was collected and the species of macrophytes found, if any.The temperature treatments consist of AMB (ambient), A2 (+2-3 • C) and A2 + (A2 +50 %, +3-5

Methane concentrations and fluxes
Samples for dissolved CH 4 analysis were collected once on 26 August 2020, using the headspace method (Hope et al., 2004); 20 ml of in situ water from each mesocosm and 20 ml of N 2 were shaken and mixed in a closed 60-ml syringe for 60 s, after which the headspace was transferred to a 12 ml pre evacuated glass vial.Methane concentrations were determined using FID on a dual-inlet Agilent 7890 Gas Chromatograph (GC) system interfaced with a CTC Combi Pal autosampler (Agilent, Naerum, Denmark).
The ebullitive flux of CH 4 was captured in specially designed bubble traps, consisting of two inverted funnels for mesocosm leading to 100 ml polypropylene syringes that were held in place by a rubber bung.The funnels have a diameter of 120 mm and together cover an area of 0.0226 m 2 , which is 0.8 % of the total area of the mesocosms, a mesh placed in the neck of the funnel prevent animals from entering the trap.The volume trapped in the syringes was sampled on 26 August 2020, after two weeks of bubble collection.
The bubble sample was taken directly from the syringes connected to a three-way stopcock and then stored in pre-evacuated 12 ml vials (exetainer, Labco, UK).The CH 4 concentration in the samples was measured by directly injecting 30 µl into the SRI 8610 C GC.For the GC, certified CH 4 standards were used for calibration and validation (see Supplementary methods).
The total emitted fluxes of CH 4 for each mesocosm were then calculated as a sum of ebullitive and diffusive fluxes.

Statistical methods
Generalized least squares (GLS) was used to test for significant differences in potential CH 4 oxidation rates and pmoA copy numbers among plant types, nutrient levels, temperature treatments and macrophyte species using the nlme package (Pinheiro et al., 2017) in R version 4.0.3(R Core Team, 2020).To test for treatment effects (temperature and nutrients), the mean of the potential CH 4 oxidation rates of the multiple samples taken from an individual mesocosm was calculated.The rates from each mesocosms were tested by treatment (or species) by using a model containing a variance structure (in this case VarIdent) to account for the heterogeneity of the data across nutrient levels, plant types and species identity (Kariya and Kurata, 2004).When comparing rates of the two species of macrophytes present in the same mesocosm, a linear mixed-effects model (lme) was used to account for non-independence of the data where the mesocosm was considered as random effect.Linear regressions and descriptors of their strength (based on Ratner, 2009) were used to explore potential relationships between the CH 4 oxidation potential and the concentrations of dissolved CH 4 or total fluxes of CH 4 as well as between pmoA gene abundance and the concentrations of dissolved CH 4 or total fluxes of CH 4 .
The samples of periphyton had the lowest oxidation rates of all plant types (0.015 [0.0033-0.061]µmol CH 4 h − 1 g − 1 dw).Periphyton oxidation rates tended to be higher at high nutrient than at low nutrient but the difference between the two treatments was not significant (p = 0.077, Fig. 2a, b).In addition, neither long-term warming effects (p = 0.27) nor interactive effects between warming and nutrients (p = 0.22) were found on the oxidation rates of periphyton samples.Although not significantly different, filamentous algae had the highest oxidation rates among all the plant types considered (0.093, [0.020-0.25]µmol CH 4 h − 1 g − 1 dw) and the three samples from the high temperature (A2 +) and high nutrient level treatments generally had higher consumption rates than the low nutrient and different temperature treatments (Fig. 2c, d).
A similar pattern was found for the rhizosphere of the macrophytes, where P. crispus had higher values than E. canadensis, even when the two species coexisted (0.066 [0.033-0.13]µmol h − 1 g − 1 dw and 0.037 [0.025-0.053]µmol h − 1 g − 1 dw, respectively).In fact, considering the low nutrient treatment mesocosms only, our results showed significant differences in rhizosphere CH 4 oxidation associated with the two species at low nutrients (p = 0.034, Fig. 2g, h).No long-term warming effect on the oxidation rates of the selected samples was found (p = 0.22).However, due to the absence of samples from the high nutrient treatments (n = 2), the effect of nutrients and any interaction with long-term warming on oxidation rates on macrophytes (above-ground and rhizosphere) could not be determined.

qPCR targeting pmoAa proxy for MOB abundance
The presence of MOB in association with periphyton, submerged filamentous algae and the rhizosphere of submerged macrophytes was confirmed by qPCR analysis.The pmoA gene was amplified in all the samples with consistent abundance of CH 4 -oxidizing bacteria in a range between 0 and 10 8 copy number g − 1 dw for each plant type.Periphyton collected from high nutrient treatments had higher copy numbers of pmoA than those from low nutrient treatments (p < 0.001; Fig. 3a), but no long-term warming effects (p = 0.14) or interactive effects between warming and nutrients were found (p = 0.13).Even if not statistically significant, high nutrient samples of filamentous algae seemed to have  higher copy numbers than low nutrient samples (Fig. 3b).For the macrophyte -rhizosphere, P. crispus had higher pmoA copies than E. canadensis (p = 0.007, Fig. 3c), showing a similar pattern as for the CH 4 oxidation potential results.No long-term warming effect on the MOB abundance associated with the two species of macrophytesrhizosphere was found (p = 0.23).In general, no linear relationship was found between pmoA copy numbers and potential CH 4 oxidation activity (see Supplementary information).

Potential CH 4 oxidation and concentration of dissolved CH 4 in the water column
We looked at the relationship between dissolved CH 4 (for periphyton, filamentous algae and the above-ground macrophyte tissue, Fig. 4a, b, c) and total CH 4 fluxes (for the macrophyte rhizosphere, Fig. 4d) on the oxidation activity performed by MOB.A weak positive trend between concentration of dissolved CH 4 and CH 4 oxidation activity per unit area was found for filamentous algae and the rhizosphere, albeit not significant (Fig. 4b, d).For the macrophytes -rhizosphere, the mean of the oxidation rates between the two species was considered when both E. canadensis and P. crispus were present in the same mesocosm.A moderate positive relationship (r 2 = 0.33, p = 0.003) was recorded between the concentration of dissolved CH 4 and the CH 4 oxidation activity associated with the periphyton samples (Fig. 4a).Also for the above-ground macrophyte tissue, the oxidation rates were moderately positively influenced by the increase of dissolved CH 4 , showing a strong increase with enhanced dissolved CH 4 in the water (r 2 = 0.50, p = 0.011, Fig. 4c).In particular, only the two high nutrient samples had higher oxidation rates but also higher concentrations of CH 4 compared with the other mesocosms (Fig. 4c).

Effect of total CH 4 oxidation associated with macrophytesaboveground on concentration of dissolved CH 4
The relationship between total CH 4 oxidation of above-ground macrophyte tissue with the concentration of dissolved CH 4 in the water was not significant (p = 0.22, Fig. 5a) and we could not identify a pattern between the two variables.However, the two high nutrient samples of P. crispus, whilst having the highest concentrations of dissolved CH 4 and the highest oxidation rates per unit, they had low total oxidation capacity due their low total biomass.
In contrast, the relationship between total macrophyteaboveground biomass and dissolved CH 4 in each mesocosm had a significant negative relationship (p = 0.036), with a decreasing trend of concentration of dissolved CH 4 with the increase in the macrophytesaboveground biomass (Fig. 5b).

Potential CH 4 oxidation and MOB abundance associated with aquatic plants exposed to long-term eutrophication and warming
Potential CH 4 oxidation and MOB were detected on samples of periphyton, filamentous algae and two species of submerged macrophytes collected from shallow lake mesocosms.In particular, macrophytesabove-ground and macrophyte -rhizosphere had higher CH 4 oxidation rates than those of periphyton.We found that P. crispus had higher oxidation rates in the rhizosphere compared with E. canadensis throughout the low nutrient mesocosms as well as when the two species co-existed in the same mesocosm (Fig. 2).
Methane oxidation activity on the surface of submerged aquatic macrophytes recorded by previous studies in eutrophic systems (Heilman and Carlton, 2001;Sorrell et al., 2002;Yoshida et al., 2014) showed a high range of oxidation rates with higher rates for above-ground tissue ( < 0.01-37 µmol CH 4 h − 1 g − 1 dw) than for roots and washed roots (< 0.01-6.45µmol CH 4 h − 1 g − 1 dw).However, our results showed much lower values of potential CH 4 oxidation rates than the previous studies for both macrophytes above-ground tissues (0.022-0.14 µmol CH 4 h − 1 g − 1 dw) and rhizosphere (0.025-0.13 µmol CH 4 h − 1 g − 1 dw).These differences in oxidation rates may be due to differences in the concentrations of CH 4 used to incubate the plant tissues.While in Sorrell et al. (2002) and in Yoshida et al. (2014) the incubation concentrations were of about 3000 and 10000 ppm respectively, in our experiment it was of 150 ppm of CH 4 .The concentration used in our experiment was representative of the environment but at such concentration, the diffusion of CH 4 to the MOB might be the limiting factors for the CH 4 oxidation activity (Sorrell et al., 2002).This may also explain why we did not find any statistical difference between above-ground tissues and rhizosphere of macrophytes: at high concentrations of CH 4 , oxygen availability becomes the limiting factor for MOB communities, while at lower concentration, such as 150 ppm, the uptake of CH 4 might be the limiting factor (Sorrell et al., 2002).The amplification of the pmoA gene through qPCR proved the presence of MOB on periphyton, filamentous algae and the macrophyterhizosphere.In addition, where quantification of the pmoA gene was possible, similar patterns to those for the CH 4 oxidation potential results were found across the different plant types.A strong nutrient effect on the MOB abundance on periphyton was recorded, showing that the nutrient level in the water is associated with increased abundance of MOB on periphyton (Fig. 3a).However, the increase in MOB abundance on periphyton was not accompanied by an increase in CH 4 oxidation potential (see Supplementary results).This could be because of a shift in microbial community composition (Nijman et al., 2021), so that the abundance but not the activity of MOB increases, or because of the fact that copy numbers is not always linked to microbial activity (Rocca et al., 2015).For the filamentous algae, even though we could not test statistically the influence of nutrients on MOB abundance due to the scarcity of samples from high nutrient mesocosms (n = 3), the samples from the high nutrient treatments tended to have higher copy numbers than most of the samples from the low nutrient ones.The abundance of MOB associated with the macrophyte -rhizosphere was dependent on the species of macrophytes (Fig. 3c).In fact, for the macrophyte rhizosphere, P. crispus had significantly higher pmoA copy numbers than E. canadensis and this is consistent with our previous results on CH 4 oxidation potential rates, where P. crispus had also higher oxidation activity than of E. canadensis.
The lack of plant types from some of the mesocosms and therefore, from some of the treatments, complicated testing the effect of nutrients and long-term warming on the CH 4 oxidation activity and MOB abundance of the selected plant types.However, the influence of the nutrient level on the abundance of MOB associated with periphyton found here corresponds with previous studies evidencing a positive effect of eutrophication on the activity and abundance of MOB in freshwater lakes (Yang et al., 2019) and MOB associated with submerged aquatic plants (Yoshida et al., 2014).Furthermore, while temperature is one of the factors increasing microbial activity at the cellular level (Madigan et al., 2003) including CH 4 oxidation activity (Shelley et al., 2015;Zhu et al., 2020), our results did not demonstrate a long-term warming effect or any interaction with nutrients, when possible to test for it, for example for the periphyton samples.This may be because MOB associated with aquatic plants in the A2 and A2 + mesocosms might have adapted to the higher water temperature, as they were not exposed to a sudden temperature increase, or because of a possible shift of MOB composition but not of MOB activity (Nijman et al., 2021).It may also be that the long-term warming effect was overridden by other processes with a stronger influence, such as higher nutrient levels or higher CH 4 Circles represent data from the low nutrient samples and squares data from the high nutrient samples (a and b).For the macrophytes, blue: samples of E. canadensis, orange: samples of P. crispus, red: mesocosms where the two species were present at the same time (c and d).Only the average of the replicates for each sample is reported and a linear regression between the variables is shown.For the rhizosphere of the macrophytes, data are log10 transformed.
production (Davidson et al., 2015;Özen et al., 2013;Zhou et al., 2019).However, the samples of filamentous algae and macrophytes were unevenly distributed among the temperature treatments, which made it hard to generalize the possible effect of long-term warming, including its interactions with nutrients, on the oxidation performance.

Factors influencing potential CH 4 oxidation rates-macrophyte traits and concentration of dissolved CH 4 in the water
The higher potential CH 4 oxidation rates associated with P. crispus, than with E. canadensis, could be explained by several factors.The morphology and the physiology of the two species might influence the adaptation of MOB to these niches.P. crispus is generally characterised by ca.5-12 mm wide and 25-95 mm long leaves, while E. canadensis has smaller leaves with a length of ca.6-14 mm and a width of 0.8-2.3mm (Schou et al., 2017).Furthermore, P. crispus has a much lower biomass than E. canadensis when occupying the same volume (see methods).According to our results, samples of P. crispus had higher oxidation rates than E. canadensis, at least when they were both present in the same mesocosms, but the biomass of P. crispus at a given PVI was much lower.In addition, the scarcity of macrophyte samples from the high nutrient mesocosms prevented establishment of a strong relationship between nutrients and CH 4 oxidation rates; however, the samples from the two high nutrient mesocosms combined very low biomasses with the highest oxidation rates.This leads to the hypothesis that higher macrophyte biomass may increase the total abundance of MOB per surface but reduce the oxidation rates per unit of area because of higher competition to oxidize CH 4 .In emergent aquatic macrophytes, the presence and activity of MOB were controlled by the macrophyte oxygen transport capacity (Calhoun and King, 1997;Schipper and Reddy, 1996), and this could also be true for submerged aquatic macrophytes, where P. crispus and E. canadensis may have different oxygen transport rates.In fact, E. canadensis has less numerous and shorter roots and a much simpler anatomy than P. crispus (Hupfer and Dollan, 2003) and, consequently, a much smaller root surface area, which influences the internal oxygen transport rates as well as the oxygenation of the sediment.Due to the small sample size and the uneven distribution of the two species of macrophytes among the treatments, neither oxidation activity nor MOB abundance associated with macrophytes showed a temperature effect, this was true for both above-ground parts and the rhizosphere.However, temperature might play a role on increasing the abundance of MOB in freshwaters, as highlighted from recent studies (Nijman et al., 2021) and it may even interact differently among macrophyte species.
A relationship between the concentration of dissolved CH 4 in the water and the CH 4 oxidation activity associated with the different plant types was identified.This relationship was significant for the aboveground tissue of the macrophytes, showing that the CH 4 oxidation activity per unit area of macrophyte was most affected, when compared with the other plant types, by the increase in dissolved CH 4 in the water.In fact, previous research has identified a positive link between eutrophication and CH 4 emissions (Beaulieu et al., 2019;Davidson et al., 2018Davidson et al., , 2015;;Li et al., 2021;Sepulveda-jauregui et al., 2018), but our study suggested a positive influence of CH 4 concentration on CH 4 oxidation activity per unit area in macrophytes and, in particular, a connection between higher nutrient levels, higher dissolved CH 4 and the subsequently higher CH 4 oxidation activity per unit area.

Role of total CH 4 oxidation activity associated with macrophytes on concentration of dissolved CH 4 in shallow lakes
No significant positive or negative relationship was found between the total CH 4 oxidation potential of the macrophytesabove-ground and the concentration of dissolved CH 4 in the water (p = 0.22).However, the association between abundant macrophytes and reduced dissolved CH 4 concentrations was significant (p = 0.036, Fig. 5b) when biomass, rather than total above ground CH 4 oxidation rates, was used as the predictor.This is in agreement with previous work showing higher submerged aquatic macrophyte biomass significantly reduced dissolved CH 4 concentrations (Davidson et al., 2015).Our results suggest that although there is macrophyte associated CH 4 oxidation, other macrophyte associated processes contribute to the lower CH 4 concentrations and emissions and that a higher biomass of submerged macrophytes perhaps limits CH 4 production.For example, as a result of photosynthetic activity, the above-ground tissue of macrophytes releases oxygen directly into the water column (Kirk, 1994), while roots, the rhizosphere and rhizomes can oxygenate the sediment due to the internal oxygen transport of the macrophytes (Sand-jensen et al., 1982;Sorrell and Dromgoole, 1987).This may negatively affect the methanogenesis (Jespersen et al., 1998;Sutton-Grier and Megonigal, 2011) of mainly anaerobic Archaea (Bastviken et al., 2008).For this reason, the increase in the density of macrophytes, even if there are differences among species, would then create more niches or possible habitats for MOB that are oxygen-dependent organisms (King, 1992).
While our results show that increased nutrient levels increase MOB abundance and generally also CH 4 oxidation rates per gram associated with periphyton, the effect might be different for macrophyte abundance.In fact, it is known from previous studies that higher nutrient loads increase phytoplankton abundance in the water, and this reduces the water clarity and the light penetration and thus negatively influences macrophyte abundance (Søndergaard et al., 2021).Our study suggests that CH 4 oxidation activity associated with submerged aquatic macrophytes might contribute to reduce the total emissions of CH 4 and this applies mostly to shallow lakes where macrophytes tend to be abundant.At higher nutrient levels the total oxidation activity associated with macrophytes might not be sufficient to impact the total carbon cycle if the abundance of macrophytes decreases and CH 4 emissions increase (Davidson et al., 2018).
Similar considerations might be investigated for filamentous algae and macrophytes rhizosphere, by relating the total oxidation activity on the CH 4 budget.This requires volume to biomass conversion for which we did not possess data in the present study.
In conclusion, MOB on periphyton, filamentous algae and macrophytes oxidized CH 4 produced in shallow lakes, albeit with different rates among the plant types.MOB abundance increased in samples in high nutrient mesocosms, however, an increased nutrient load tends to reduce the abundance of macrophytes and, with it, the total CH 4 oxidation per mesocosm, but increase the oxidation activity per unit of area associated with them.In addition, even if CH 4 oxidation activity associated with plants may not be the only process involved in reducing CH 4 emissions, we demonstrate that the increase in plants biomass tends to reduce the concentration of dissolved CH 4 in the water.This highlights the importance of preserving or restoring lake conditions that are suitable for submerged aquatic macrophytes as both the above-ground tissue and the rhizosphere of these may potentially act as CH 4 sinks.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 1 .
Fig. 1.Potential CH 4 oxidation rates in each plant type after 12 h of incubation with 150 ppm of CH 4 .Letters above the boxplots indicate the results from the GLS test.Boxplots show the median, min and max values for each habitat.Different letters indicate significant differences.n indicates the number of samples per plant type.

Fig. 2 .
Fig. 2. Potential CH 4 oxidation rates per plant type after 12 h of incubation with 150 ppm of CH 4 at 25 • C. The left-hand panel shows bar plots of periphyton (a) and filamentous algae (c) plotted for the different nutrient and temperature treatments and of macrophytes -above-ground (e) and macrophytes -rhizosphere (g) samples divided by nutrient, temperature and species.For each plant type, each column or grouped column represents a mesocosm.The dots represent the duplicates or triplicates for each mesocosm and the column height is the average among the replicates.The right-hand panel shows boxplots of potential CH 4 oxidation for periphyton and filamentous algae divided by two nutrient levels (b and d) and macrophytes above-ground tissues and rhizosphere (f and h) divided by species (blue: Elodea canadensis, orange: Potamogeton crispus).Boxplots indicate median, whiskers, bars indicate min and max values.Letters above the boxplots indicate the results from the GLS test.Different letters indicate significant differences.n indicates the number of samples per plant type.*Differences between the nutrient treatments could not be tested because of paucity of samples at high nutrient.

Fig. 3 .
Fig. 3. Boxplots showing the median, min and max values of pmoA copy number g − 1 dw of periphyton (a), filamentous algae (b) and macrophytes -rhizosphere (c).Samples of periphyton and filamentous algae are grouped by high and low nutrient treatments, while samples of macrophytes -rhizosphere are grouped by species (blue: Elodea canadensis, orange: Potamogeton crispus).Letters above the boxplots indicate the results from the GLS test.Different letters indicate significant differences.The data from the three plant types are log10 transformed.n indicates the number of samples per plant type.*Differences between the nutrient treatments could not be tested because of paucity of samples at high nutrient.

Fig. 4 .
Fig. 4. Relationship between dissolved CH 4 and total CH 4 fluxes and the potential rates of CH 4 oxidation for periphyton (a), submerged filamentous algae (b), macrophytes -above-ground (c) and macrophytes -rhizosphere (d).Circles represent data from the low nutrient samples and squares data from the high nutrient samples (a and b).For the macrophytes, blue: samples of E. canadensis, orange: samples of P. crispus, red: mesocosms where the two species were present at the same time (c and d).Only the average of the replicates for each sample is reported and a linear regression between the variables is shown.For the rhizosphere of the macrophytes, data are log10 transformed.

Fig. 5 .
Fig. 5.The relationship between dissolved CH 4 in the water column and the total CH 4 oxidation activity of macrophytesabove-ground (a) and the total biomass of macrophytesabove-ground (b).The samples are divided by nutrient levels and macrophyte species.Circles represent data from the low nutrient samples and squares data from the high nutrient samples.Blue: mesocosms with E. canadensis, orange: mesocosms with P. crispus in the low and high nutrient treatments, red: mesocosms where the two species were present at the same time.Data are log10 transformed, and a linear regression between the two variables is shown.