Abundance and biogeography of methanogenic and methanotrophic microorganisms across European streams

Magdalena Nagler Universitat Innsbruck Nadine Praeg Universitat Innsbruck Georg H. Niedrist Universitat Innsbruck Katrin Attermeyer Uppsala Universitet Teknisk-naturvetenskapliga fakulteten Núria Catalán Universitat de Girona Francesca Pilotto Umea Universitet Catherine Gutmann Roberts Bournemouth University Christoph Bors (  [email protected] ) Universitat Koblenz-Landau https://orcid.org/0000-0001-7869-8590 Stefano Fenoglio Universita del Piemonte Orientale Miriam Colls Universitat de Girona Sophie Cauvy-Fraunié Irstea Centre de Lyon-Villeurbanne Brian Doyle Dundalk Institute of Technology Ferran Romero Universitat de Girona Björn Machalett Humboldt-Universitat zu Berlin Thomas Fuss Leibniz-Insitute of Freshwater Ecology and Inland Fisheries Adam Bednařík Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 1/35 Univerzita Palackeho v Olomouci Marcus Klaus Umea Universitet Peter Gilbert University of Islands and Highlands Dominique Lamonica Irstea Centre de Lyon-Villeurbanne Anna C. Nydahl Uppsala Universitet Clara Romero González-Quijano Leibniz-Institute of Freshwater Ecology and Inland Fisheries Lukas Thuile Bistarelli Leibniz-Institute of Freshwater Ecology and Inland Fisheries Lyubomir Kenderov Sofa University Elena Piano Universita degli Studi del Piemonte Orientale Jordi-René Mor Universitat de Girona Vesela Evtimova Bulgarian Academy of Sciences Elvira deEyto The Marine Institute Anna Freixa Universitat de Girona Martin Rulík Univerzita Palackeho v Olomouci Josephine Pegg Bournemouth University Sonia Herrero-Ortega Leiniz-Institute of Freshwater Ecology and Inland Fisheries Lea Steinle Universitat Basel Pascal Bodmer Universitat Koblenz-Landau

Research

Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 2/35 Keywords: Methanogenic archaea, methane oxidizing bacteria, biogeography, stream sediments, inland waters, potential methane production, potential methane oxidation

Posted Date: November 19th, 2019

DOI: https://doi.org/10.21203/rs.2.17386/v1

License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License

Version of Record: A version of this preprint was published at Journal of Biogeography on December 15th, 2020. See the published version at https://doi.org/10.1111/jbi.14052.

Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 3/35 Abstract

Background: Globally, streams emit signifcant amounts of methane, a highly potent greenhouse gas. However, little is known about the stream sediment microbial communities that control the net methane balance in these systems, and in particular about their distribution and composition at large spatial scales. This study investigated the diversity and abundance of methanogenic archaea and methane- oxidizing microorganisms across 16 European streams (from northern Spain to northern Sweden and from western Ireland to western Bulgaria) via 16S rRNA gene sequencing and qPCR. Furthermore, it examined environmental factors infuencing both abundance and community composition and explored the link to measured potential methane production and oxidation rates of the respective sediments.

Results: Our results demonstrated that the methanogenic and methanotrophic microbiomes of the studied European streams were linked to both the temperature and degree of anthropogenic alteration. The microbiomes could be separated into two to three groups according to environmental factors at both stream and catchment scales. Main methanogenic taxa found within more anthropogenically-altered, warm, and oxygen-poor environments were either Methanospirillum spp. or members of the families Methanosarcinaceae and Methanobacteriaceae . Within such environments, methane oxidizing communities were strongly characterized by members of the family Methylobacteriaceae ( Meganema spp. and Microvirga spp.). Contrastingly, communities in colder environments rich in oxygen and with relatively little anthropogenic impact at the catchment scale were characterized by the methanogenic Methanosaetaceae , Methanocellaceae and Methanoregulaceae and the methanotrophic Methyloglobulus spp ., members of the CABC2E06 group (all Methylococcaceae ) and by various Candidatus Methanoperedens. Overall, diversity of methanogenic archaea increased with increasing water temperature. Methane oxidizing communities showed higher diversities in southern sampling sites and in streams with larger stream areas and widths. Potential methane production rates signifcantly increased with increasing abundance of methanogenic archaea, while potential methane oxidation rates did not show signifcant correlations with abundances of methane oxidizing bacteria, presumably due to the more diverse physiological capabilities of this group.

Conclusions: We present the frst large scale overview of the large-scale microbial biogeography of two microbial groups driving the methane cycle dynamics within stream sediments and deduce the impact that future anthropogenic alterations may cause.

1. Background

Methane (CH4) exhibits a 34 times higher global warming potential than carbon dioxide (CO2)[1] due to its high efciency in capturing infrared radiation [2]. While sources of CH4 are manifold and may be natural or anthropogenic, current estimates suggest that 35–50 % of global CH4 emissions originate from natural sources [1]. Among these sources natural wetlands are the largest, followed by geological sources and freshwaters (running waters and lakes) [3], with running waters contributing 3 % of the global release orL o1a5d–in4g 0[M %at hoJfa txh]/eja xe/fouutpxu tf/rCoommm woentHlaTMndL/ afonndts /laTekXe/sfo [n4t,d 5at]a..js

Page 4/35 CH4 is almost exclusively produced by methanogenic archaea (MA) [6] during the fnal step of anaerobic degradation of organic matter, a complex process carried out by a consortium of hydrolytic, fermenting, syntrophic and acetogenic microorganisms [7]. MA metabolize substrates like hydrogen, CO2 and acetate to CH4, mainly via the hydrogenotrophic, and acetoclastic pathways of CH4 production [6]. Although methanogenesis is one of the main processes responsible for anaerobic organic matter mineralization in river sediments, little is known about the involved methanogens within this environment. Prior research on MA taxonomic distributions and community compositions has mainly focused on terrestrial habitats like soils, peat bogs, wetlands, and rice felds, and on marine and lacustrine sediments [8, 9]. Although representing important components of the global carbon cycle through carbon mineralization into CO2 and CH4 [5, 10], lotic waters have been scarcely addressed. Similar to lake sediments [11], frst studies on river sediments showed the dominance of Methanosarcinales and Methanomicrobiales [12–14], orders performing the acetoclastic and hydrogenotrophic CH4 production pathways, respectively.

Environments facilitating the occurrence of communities producing CH4 are also typical habitats for methane oxidizing microorganisms (MOX) [15], consuming the newly synthesized CH4 (and methanol) as a carbon and energy source. Some MOX may belong to the archaea, performing anaerobic CH4 oxidation and involving three distinct methanotrophic groups (ANME–1 to ANME–3), of which group ANME–2 is further divided into four sub-groups (ANME2 a, b, c and GOM Arc I) [16]. Furthermore, anaerobic CH4 oxidation can be performed by some single known bacterial species belonging to the phylum NC10 [17].

Aerobic CH4 oxidation, however, is carried out by a heterogeneous group of obligate or facultative methanotrophic bacteria (MOB) usually divided into three main types: Gammaproteobacteria (referred to as type I or type X), Alphaproteobacteria (type II) and Verrucomicrobia (type III), all with differing pathways in the specifc utilization of CH4 [15, 18]. Aerobic MOX often inhabit anoxic-oxic interfaces, such as the upper layers and surfaces of sediments. Most research on MOX in aquatic environments has been conducted on lake sediments, where a variety of different MOX genera were found to be predominant (i.e., Methylobacter, Methylomonas, Methylosarcina, Methylococcus, and Methylosoma) and type I MOX were more abundant than type II MOX [19, 20].

Environmental parameters such as oxygen levels, pH, sediment organic carbon and nitrogen content are highly spatially variable in stream ecosystems. Furthermore, transported and deposited sediments originate from the catchment through erosion processes and are thus determined by terrestrial geology, topography, climate, hydrology, vegetation and land-use. It is known that this multitude of parameters is paramount in determining the composition, activity and abundances of MA and MOX at a local scale, however, their role at a continental scale is still unknown.

While molecular approaches are helpful in analyzing community compositions and microbial abundances of MA and MOX, studying the net CH4 balance of an ecosystem also requires the quantifcation and investigation of CH4 production and oxidation processes. Studies measuring potential methane production (PMP) in streams, rivers [21, 22], or river impoundments [23, 24] are scarce and do noLot andeincge [sMsaathriJlayx l]/injakx/ PouMtpPut /rCaotmesm woniHthT MMLA/f oanbtsu/nTedXa/nfocnetdsa, tbau.jst rather focus on the emission of CH4 by these

Page 5/35 systems. However, some recent studies on PMP in stream sediments [12, 25, 26] showed two depth- related PMP maxima along the sediment profles dominated by hydrogenotrophic methanogenesis. While this study could not prove a relationship between MA abundance and PMP, others [27] observed highest PMP rates in the sediment of streams enriched in the methane-producing mcrA gene.. Data on potential

CH4 oxidation (PMO) in freshwater ecosystems are also scarce (but see [28–31]). Within coastal marine ecosystems, increased PMO rates were found close to the seafoor, but were not directly linked to CH4 concentrations [32]. In river sediment, PMO was found to increase with decreasing gravel size [33], and in lake waters CH4 concentrations and temperature both had a positive effect on PMO [34].

To our knowledge, an in-depth investigation of taxonomic distributions, abundances, and underlying environmental conditions of both microbial groups involved in the CH4 cycle has not been carried out on large continental (European) scale. Thus, the present study aims to (i) describe the biogeography of MA and MOX communities of stream sediments, (ii) identify the dominant species within each functional group, (iii) determine the driving environmental parameters for their community compositions and associated indicator species, and (iv) compare the absolute abundances of MA and MOX with their potential activities across 16 European streams.

2. Results

2.1. Environmental data

The principal component analysis (PCA) of environmental variables (Fig. 1; c.f. Tab. S1, S4) revealed the infuence of variables measured at stream-level on the frst axis, explaining 23.3 % of all variation in the data. Physical stream characteristics describing stream size, such as stream area, cross-sectional area, and slope were the best represented on the frst axis, while physical and chemical properties including pH, dissolved oxygen, and water temperature were less explained (Fig. 1). The second axis depicted 13.9 % of the variation in the environmental data and discriminated the samples mainly based on grain size distributions and organic carbon concentrations. Thus, differing sediment types from within streams (i.e., coarse, medium, and fne sediment patches) were mostly separated on axis 2, while axis 1 roughly ordered samples according to latitude (as represented by the site colors from bright (south) to dark (north)).

2.2. Alpha- and beta-diversity of MA and MOX communities in streams across Europe

Total 16S rRNA read numbers ranged from 73249 to 140000. Of all reads, a mean of 728 (1.0 %) belonged to MA and 799 (1.1 %) to MOX (Fig. 2a, c; Tab. S5). There were no signifcant differences in overall microbiome composition of all sequenced operational taxonomic units (OTUs) (including all bacteria and archaea) between samples from 0-3 cm and 6-9 cm sediment depths, as tested with PELoRaMdinAgN [MOVathAJ.ax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 6/35 Reads of methanogenic archaea were assigned to ten different families, with Methanoregulaceae, Methanosaetaceae, Methanosarcinaceae and Methanobacteriaceae being most abundant in terms of relative reads (Fig. 2a, b; Tab. S5). The diversity of MA ranged from 9.7 to 32.8 OTUs and over a Shannon index from 1.6 to 2.9 (Fig. 2a; Fig. S3; Tab. S5). MA species richness and Shannon diversity were positively correlated to water temperature (Pearsonsr = 0.45) and catchmentarea(Pearsons r = 0.48), respectively, while they were both negatively correlated to increasing percentages of natural land cover within the catchment (r = -0.48 and r = -0.5, resp.; Fig. 3).

For MOX, OTUs were assigned to 11 different families, with bacterial Methylococcaceae, unclassifed Methylococcales, Methylobacteriaceae and the archaeal clade GOM Arc I being among the most abundant (Fig. 2c, d; Tab. S5). Within MOX, signifcant differences in read numbers between both sampling depths (0-3 vs. 6-9 cm) were found for GOM Arc I as well as Nitrospirales inc. sed., yielding higher read numbers in deeper sediment samples (p < 0.01, Tab. S5). For other families including Methylocystaceae, CABC2E06, Crenotrichaceae and plW-20, differences were less pronounced (p = 0.014, p = 0.017, p = 0.034 and p = 0.023, respectively) but all families displayed lower read numbers in the deeper sediment layer. Similarly, read numbers of anaerobic methane oxidizers (GOM Arc I + Nitrospirales inc. sed.) were signifcantly higher within the deeper sediment layer (p < 0.01), while read numbers of aerobic methane oxidizers did not differ signifcantly between both sampling depths. MOX OTU numbers ranged from 14.7 to 65.7 and Shannon diversities from 1.4 to 3.2, with Shannon diversity being signifcantly correlated to geographical latitude (Pearsonsr = - 0.51), streamarea(Pearsons r = 0.51), wetted stream width (Pearsonsr = 0.49) and slope(Pearsons r = 0.46).

2.3. Correlations among taxonomic groups and with environmental variables

Exploring differences in OTU-based microbial community compositions without environmental constraints, the non-metric multidimensional scaling (NMDS) analysis showed a more homogeneous distribution of MA communities across European streams as compared to MOX communities, the latter forming clusters that are more distinct among sampling sites and roughly grouped according to their geographical latitude (Fig. S1). The differences of within-stream variability linked to both the sampling depth and the differences in the sediment grain size (i.e., coarse, medium, and fne) are clearly depicted by the individually sized convex hulls of each sampling site. According to the two-way-PERMANOVA, the afliation to a certain stream exhibited a signifcant infuence on both, MA and MOX (p < 0.01), while the sampling depth was found to be marginally infuential only for MOX communities (p = 0.0314) and no signifcant interaction effect was found. However, for both microbial communities, samples from the three Spanish streams (ESP2_1, ESP2_2, ESP2_3) exhibited distinct communities as compared to the rest of the samples. In addition, DEU2_1, AUT1_1 and GBR1_1 formed distinct clusters within both MOX and MA, while DEU1_1 and SWE1_1 formed separate clusters within MOX.

Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 7/35 Combined with signifcant environmental parameters, the forward selected explanatory variables chosen in the redundancy analyses (RDA) accounted for an adjusted explained variation of 42.3 % for MA and 55.3 % for MOX (Fig. 4a, b). Important MA taxa found within more anthropogenically-altered, warm environments with increased pH were represented by various Methanospirillum spp. as well as members of the family Methanosarcinaceae (Methanomethylovorans sp., Methanosarcina sp. and Methanolobus sp.) and by a variety of taxa belonging to Methanobacteriaceae (Fig. 4a). Within comparable environments, MOX communities were characterized by members of the family Methylobacteriaceae (Meganema sp. and Microvirga sp.; Fig. 4b).

Communities in well oxygenated streams within catchments dominated by natural land cover were characterized by MA groups Methanosaeta spp., Methanocellaceae and Methanoregulaceae, and by MOX representatives such as Methyloglobulus spp., members of the CABC2E06 group (all type I MOX) and, interestingly, by various Candidatus Methanoperedens spp. (Fig 3a, b).

For MA a distinct cluster formed in accordance with the NMDS for AUT2_1 and ESP2_1 but no characteristic community was defned (Fig. 4a). For MOX, however, a distinct community within a separate environment featuring high values for discharge, catchment area and fow velocity favored the occurrence of Methylocaldum spp. and Crenothrix spp. (Fig 4b) and included mainly sampling site DEU1_1.

For MA families, relative read numbers of Methanocellaceae and MHLsu47-B8a increased with increasing latitude, while Methanosarcinaceae decreased with latitude (Fig. 3a). Furthermore, Methanocellaceae and Thermoplasmatales inc. sed. decreased with increasing water temperature, while Methanocellaceae in addition decreased with organic carbon contents (Fig. 3a).

A large number of environmental variables signifcantly correlated with relative read numbers of various MOX-families (Fig. 3a). Among the most infuencing environmental variables (based on Pearson`s r) were water dissolved oxygen, stream area, and elevation of the sampling site. Dissolved oxygen was positively correlated with Methylococcaceae as well as unknown members of Methylococcales, and negatively correlated to both the Marine Methylotrophic group and Methylobacteriaceae. Stream area, however, was positively correlated to the Marine Methylotrophic group, Methylobacteriaceae and Methylocystaceae and elevation to Hyphomicrobiaceae and Nitrospirales inc. sed. (Fig. 3a).

Signifcant correlations among relative read abundances of different MA families were found for Methanosarcinaceae, being negatively correlated to Methanosaetaceae, MHLsu47-B8A and Rice Cluster II, while positive relationships were found between Methanomicrobiaceae and Rice Cluster II and between Methanocellaceae and Thermoplasmatales inc. sed. (Fig. 3b). Methanosarcinaceae were negatively correlated to several families (MHLsu47-B8A, Rice Cluster II and Methanosaetaceae; Fig. 3b) that have also been assigned to the contrasting environment by the RDA (Fig. 4a), while positive correlations between MA families were only found for families indicative of the same environment (Fig. 3), confrming the explanatory power of the RDA. Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 8/35 Among MOX families, relative read numbers of Methylobacteriaceae were positively correlated to Methylocystaceae and Nitrospirales inc. sed. and negatively to Methylococcaceae, while Methylococcaceae also showed a negative relationship with Hyphomicrobiaceae and Nitrospirales inc. sed.. The pLW-20 group was positively related to both CABC2E06 and Crenotrichaeceae, and Methylocystaceae to the Marine Methylotrophic group and to Xanthobacteriaceae (Fig. 3b).

Finally, positive and negative correlations were found between MA and MOX. Methylobacteriaceae showed a negative correlation with Methanoregulaceae, MHLsu47-B8A and Rice Cluster and a positive one with Methanosarcinaceae. Methanosaetaceae were positively related to both unknown Methylococcales and Methylococcaceae, while Thermoplasmatales inc. sed. showed a negative correlation with the Methylococcales but a positive one with Nitrospirales inc. sed. In addition, Xanthobacteriaceae and Methanocellaceae were positively correlated (Fig. 3b).

2.4. Methane production and oxidation potentials

-3 -1 -3 -1 PMP rates ranged from 0.000097 g CH4 m d (IRL1_1) to 9.46 g CH4 m d (CZE1_2) and PMO rates -1 -1 from 0.00017 % CH4 d (SWE1_1) to 0.32 % CH4 d (IRL1_1; Tab. S5). PMP rates did not differ signifcantly between both sediment depths, with mean values of 1.61 ± 0.69 at 0-3 cm and 0.56 ± 1.58 g -3 -1 CH4 m d at 6-9 cm.

2.5. Gene abundance data

While mcrA gene abundances ranged from 3.1x101 to 3.2x107 gene copies gFW-1, type Ia MOB gene abundances ranged from 2.82x102 to 2.55x107 and type II MOB from 5.98x102 to 6.8x107 gene copies gFW-1 (Tab. S5). Lower mean gene copy numbers were detected for all genes in the deeper sediment samples, however, no signifcant difference in gene numbers between both sampling depths was found (mcrA: 2.62x106 vs. 2.46x106 type Ia: 3.34x106 vs. 2.09x106 type II: 4.31x106 vs. 2.07x106).

Gene copy numbers of mcrA-genes were signifcantly correlated to both single MOB-deriving gene abundances (R² = 0.670 for type Ia, R² = 0.348 for type II, Fig. 5a, b) as well as to the cumulative MOB gene copy numbers (R² = 0.594, Fig. 5c), forming a LOG-LOG relationship. Similarly, mcrA gene copy numbers showed a signifcantly positive LOG-LOG correlation with PMP rates (R² = 0.393, Fig. 5d). A correlation between PMO rates and both MOB-related gene abundances was not detected (Fig. 5e, f).

While MA gene copy numbers did not differ between sampling depths, cumulative MOB gene copy numbers were signifcantly higher in upper sediment layers of each sample (paired Student’s t-test, p < 0.01, Fig. S3), although the overall difference between shallow and deep subsamples was only marginal.

Table 1. Generalized linear mixed models (GLMM) summary statistics, with proportions of variance in MA anLoda MdinOBg [M gaethnJea xc]/ojapxy/o nutupmut/bCeormsm eoxnpHlTaMinLe/dfo natts /sTtereXa/fmon-tsdcataal.ejs (among streams) and site-scale (within streams).

Page 9/35 GLMM are based on random and fxed effects, which explain differences on different levels (among and within stream reaches respectively). Mixed Effects Model R² gives the proportion of variance explained by both, random and fxed predictors, as outlined in [35].

Among streams Within streams Among and within streams Dependent random % fixed F p % Mixed Effects Model R² group predictor deviance predictors values deviance (conditional R²) explained explained MA Catchment 55 Org. C 17.5 1x10-4 11 0.65 area Sediment surface 5.1 0.027 7 area MOB Catchment 44 Org. C 24.1 4x10-6 15 0.54 area

Gene copies of MA and cumulative MOB differed signifcantly among sampled streams (Kruskal Wallis χ²

= 59 and 46, p < 0.05), with higher gene copy numbers in streams draining larger catchment areas (linear model, p < 0.05) (among stream differences, Tab. 1). Within streams MA gene copies increased additionally with higher organic carbon content and sediment surface area, and MOB gene copies with sediment organic carbon content (Tab. 1). The increasing R²-value of the models due to site-scale variables (R², from 0.55 to 0.65 for MA and 0.44 to 0.54 for MOB; Tab. 1) indicated positive correlations between gene copies and organic carbon content and/or surface area within streams, while the overall gene copy number differed among streams. No other parameter showed signifcant relationships with response measures, but nitrogen concentration was signifcantly correlated to organic carbon content and was therefore excluded from the model. Additionally, measured discharge correlated positively with catchment area and was therefore not included as an independent variable. The applied GLMM validation indicated no assumption violation regarding the residuals.

3. Discussion 3.1. Environmental constraints

A major interest in microbial ecology is the identifcation and conceptualization of biogeography, which is shaped by processes such as habitat fltering, dispersal, drift and mutation [36]. Although the analysis of drift and mutation is beyond the focus of the study, we could demonstrate a biogeographic pattern of methanotrophic and methanogenic communities associated with geographic dispersal and environmental infuences, identify dominant species within each functional group, and compare the absolute abundance of MA and MOX with their potential activities within 16 stream sediments across a distance of 2700 km from east to west and north to south. Based on the frst PCA-axis separating the stLroeaadmings [(MFaigth.J 1ax)],/ tjawx/oo ugtrpouut/pCosm wmitohn HdTifMfLe/rfionngts e/TnevXir/ofonnmtdaetna.tjas l conditions can be distinguished on either end of Page 10/35 this axis. “Environment 1” represents warm streams with large stream areas, high conductivity values, low oxygen concentrations, and high percentages of agricultural and urban land within their catchment (ESP 2_2, ESP 2_3, ESP 2_1). In contrast, “Environment 2” includes cold, medium-to-small sized streams with low conductivity and pH, high oxygen concentrations and low percentages of agricultural and urban land within their catchment (SWE 2_1, SWE1_1, GBR2_1, DEU2_1, IRL1_1). These two most diverging environments should be kept in mind for the following discussion, which evaluates if these encountered habitat diversities are mirrored in the biogeography of MA and MOX.

3.2. MA and MOX biogeography and diversity

3.2.1.Methanogenic archaea

The four dominating families of MA, Methanoregulaceae, Methanobacteriaceae, Methanosarcinaceae and Methanosaetaceae across the studied European streams were comparable to those found in other sediment-related studies [9, 11, 26]. Methanoregulaceae are known to be ubiquitous and abundant in many terrestrial and freshwater habitats and have been suggested as a proxy for freshwater infuence in the marine realm [37]. They perform hydrogenotrophic methanogenesis similar to Methanobacteriaceae, the dominant archaeal groups within lake ecosystems [38]. Methanosarcinaceae and Methanosaetaceae are abundant in river ecosystems [26]. Across European streams, their relative abundances were negatively correlated to each other probably due to their shared capability to perform acetoclastic methanogenesis and the related competition for the same niche. Of these, Methanosarcinaceae were relatively more abundant in southern sampling sites, probably due to their versatile metabolic capabilities allowing them to adapt to the presence of other substrates besides acetate [6]. Further, they likely experience growth advantages over Methanosaetaceae, as acetate has been shown to be a major precursor of CH4 particularly under low-temperature-conditions [39].

Higher temperature and larger catchment area both increased MA diversity and richness (Fig. 3), probably due to a higher variability in nutrient sources and subsequent niches to be exploited. Similarly, lower anthropogenic infuence within the catchment might lead to less nutrients entering running waters [40], explaining the overall lower diversity and richness of MA in streams with higher percentages of natural land within their catchment.

3.2.2.Methane oxidizing microorganisms

In accordance with previous studies, the predominant family of MOX across European streams was represented by Methylococcaceae [19, 20], whereas a considerable number of OTUs (30) was assigned to Methylococcales but could not be determined further (Fig. 2). The order and the family are both type I MOB, while Methylobacteriaceae, the second most abundant family, is a representative of type II MOB, wLitoha deinagc [hM taythpJea xp]/ejarfxo/orumtpiuntg/C aom dmifofneHreTnMtL p/faotnhtsw/TaeyX /ofof nCtdHa4ta o.jsxidation, and inhabiting a different niche. The

Page 11/35 general predominance of type I over type II MOB is consistent with previous fndings [20], while increased relative abundances of type II over type I MOB were observed in all Spanish and Austrian, as well as in one British stream (GBR_1_1), being refected by their distinct clustering in the NMDS (Fig. S1). For Spanish sites experiencing high agricultural infuence, this might be explained by a potential eutrophication that favors type II MOB [41].

Next to the aerobic type I and type II MOB, the anaerobic CH4 oxidizing archaeal GOM Arc I was also abundant, occupying anaerobic niches as corroborated by their higher abundance in deeper sediment layers. Such layers are typically lower in oxygen supply than the layers immediately at the sediment-water interface, what also explains the higher abundance of the second anaerobic MOX representative, Nitrospirales inc. sed. Results on aerobic MOX showed the opposite trend. Furthermore, GOM Arc I were positively correlated to higher sediment surface areas (i.e., smaller particle sizes) that might lead to more anaerobic niches within those sites.

Shannon diversity of MOX was mainly correlated to variables describing stream size (e.g., wetted stream width, stream area) and geographical latitude, with larger, southern streams exhibiting a higher diversity, probably due to a more diverse nutrient input owing to a larger catchment area [42].

3.3. Environmental parameters driving community composition, associated indicator species and abundances

3.3.1.Methanogenic archaea

Compared to MOX, MA showed a more homogeneous distribution across European streams, although PERMANOVA confrmed that there were signifcant differences among sampled streams. On the one hand, anthropogenically altered, warm streams (environment 1) with increased pH showed MA communities housing members mostly of the hydrogenotrophic Methanospirillaceae and Methanobacteriaceae and, the metabolically and physiologically versatile Methanosarcinaceae [43], which were also negatively correlated with geographical latitude (Fig. S2). On the other hand, acetoclastic Methanosaetaceae as well as hydrogenotrophic Methanocellaceae, Methanoregulaceae and MHLsu47- B8A [44, 45] occurred in more natural environments rich in oxygen and with an elevated water velocity and discharge (environment 2) (Fig. S2). Over past decades increased land use or land cover changes have been recorded and are mainly attributed to the conversion of natural ecosystems to agricultural or urban areas [46]. Our results suggest that a future increase of anthropogenically altered catchments might result in an altered MA community with increased abundances of species characteristic for environment 1. Due to their diverse metabolic capacities as well as the dependency of their metabolic activity on the type and amount of present substrates, altered CH4 production rates are difcult to estimate. However, according to our results, increasing stream temperatures might lead to higher abundances of methanogens that use CO2 and H2 as substrates, a highly efcient CH4-production Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 12/35 pathway [47]. On the other hand, relative abundances of Methanocellaceae as well as Rice Cluster II, a hydrogenotrophic group dominant in partially thawed, cold environments [48], typically decrease with increasing stream temperatures, and might thus be negatively impacted by climate change.

3.3.2.Methane oxidizing microorganisms

In this study, MOX communities were more distinct among streams than MA communities (Fig. S1). In addition, the sampling depth infuenced the species composition, probably due to oxygen supply variability with depth. As the major biological sink for CH4 is represented by MOX, these microorganisms attenuate the rivers great CH4 production potential within the water column and the sediment.

Indeed, redundancy analysis analyses showed oxygen to be an important environmental parameter shaping the MOX community compositions. Typical representatives of oxygen-rich and more natural environments (environment 2) were Methyloglobulus spp. and other representatives of the family Methylococcaceae as well as several OTUs assigned to CABC2E06, a suspected sister group of Crenothrix [49] (Fig. 4, Fig. S2). All these MOX belong to type I MOB, performing the ribulose monophosphate pathway of CH4 oxidation. Type I MOB (Gammaproteobacteria) are known to require oxygen for CH4 activation [50]. In a study on foodplains of west Siberian rivers, type I MOB dominated cold foodplains and featured a large proportion of psychrotolerant methanotrophs [51], which fts to their higher abundance in environment 2. Interestingly, Candidatus Methanoperedens, mediating nitrate- dependent anaerobic CH4 oxidation, was also typically representative of these oxygen-rich environments, although it was found to be relatively more abundant within the deeper sampling depth. This genus can counteract oxidative damage and adapt in environments where it is regularly exposed to oxygen [52].

Contrastingly, environment 1 harbored representatives of mostly type II MOB (Alphaproteobacteria) including Microvirga, Meganema and other representatives of the Methylobacteriaceae (Fig. S2). While Microvirga and Meganema are putatively methylotrophic, Methylobacteriaceae are facultative methylotrophs and able to grow on methanol and other one-carbon compounds as sources of energy and carbon [53]. Indeed, this group was more abundant in southern streams, with low oxygen, high conductivity, and larger stream areas (Fig. 3). However, among all detected MOB, this group was the least restricted to perform CH4 oxidation, and the only MOX group signifcantly negatively correlated to several MA-families. Expected future land use changes with increased percentages of agricultural or urban land in the catchment area combined with increasing temperatures due to climate change might therefore lead to a shift of MOX communities towards the increased importance of type II MOB exhibiting a lower CH4 oxidation capacity.

Besides the previously described environments, MOX communities formed a third cluster including streams characterized by high discharge, large catchment areas, and high stream orders (environment 3)

(Fig. S2). This environment is preferentially inhabited by Methylocaldum spp., a genus fxing CO2 in adLodaitdiiongn [Mtoa tChHJa4x ](/Gjarxo/ouuptp Xut)/,C aomndm oCnrHeTnMotLh/froixn ts/pTpe.X, /afonn tidmatpao.jsrtant type I MOB described in Swiss lakes forming

Page 13/35 multicellular and flamentous structures and exhibiting an unusual enzyme-clustering together with nitrifying comammox bacteria [49].

3.4. Gene abundances and their relation to PMO and PMP

The logarithmic relationship between MA and MOB gene copies is evident for both MOB types (Ia and II) and suggests a close dependency of CH4 oxidizing bacteria with the occurrence of methane-producing archaea. Based on the observed R² of this relationship, type Ia MOB seem to be more dependent on the occurrence of MA than type II MOB, which exhibits a weaker relationship. This might be because some representatives of type II MOB including the genera Methylocella, Methylocystis, Methylocapsa, and all Methylobacteriaceae, are facultative methanotrophs and thus capable of growing on carbon sources other than CH4 [54].

Sufcient CH4 availability is a prerequisite for the proliferation of type I methanotrophs. The steeper slope of the relationship between MA and type I MOB as compared to MA and type II MOB might be explained by the ‘low-afnity’ CH4 oxidation kinetics of type I methanotrophs which enables the high oxidation of

CH4 at concentrations that are well above atmospheric levels. In contrast, type II methanotrophs are at an advantage in niches where resources are more limiting [55]. Another explanation is that the ribulose monophosphate pathway of CH4 oxidation is performed by type I MOB and this pathway is considered more efcient than the serine pathway performed by type II MOB [34, 56].

According to the GLMM, organic carbon supply and thus nutrient availability increases the abundance of MA and MOB within stream sites, and MA abundance is also positively infuenced by sediment surface area, negatively correlated to particle size. The latter dependency may refect the general afnity of MA to anaerobic microhabitats formed to a higher degree in fne sediments [57]. The signifcant positive relationship between gene abundance and catchment area indicates that larger catchments receive more diverse nutrients and thereby provide higher nutrient availability and number of present niches for a wider range of organisms [58].

Across European stream sediments, gene abundances of MA showed a signifcant positive correlation with PMP rates (Fig. 5 d), clearly linking the microbial abundances to its function by forming a LOG-LOG relationship. However, a slope of <1 suggests that with more methanogenic gene copies, the PMP is slightly lower as compared to smaller gene copy numbers. Thus, increased gene copy numbers within the investigated environmental DNA are not linearly correlated to activity. This is most probably due to extracellular DNA that might be produced with increasing activity, or due to changing counts of mcrA gene copies in the genome of different MA species [59, 60]. Similarly, Mach et al. [26] were not able to link PMP and mcrA gene copy numbers in river sediments due to a regulation of PMP on the MA RNA or activity level rather than their absolute abundance. Studies linking the PMP function to microbial abundance are rather scarce, especially in freshwater systems. Xiao et al [61], however, found the highest PMP values in the top 0-2 cm layer of sulphate-rich marine surface sediments, in which they also found a Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 14/35 peak in abundance of methanogenic archaea. The highest PMP was also consistent with a strong expression of the mcrA gene in the oxygenated water column of an oligotrophic lake [62]. Other studies determining CH4 emissions rather than the process of PMP itself found signifcant relationships between abundance of methanogenic archaea and CH4 emissions in soils of an alpine wetland [63] and in natural and restored peatlands [64].

Contrary to a previous study [65], no signifcant relationship was found between PMO rates and MOB gene abundance across European stream sediments . There are several possible explanations for this: (i) RNA or activity level of MOB might be more relevant than absolute abundance [26], (ii) alternative pathways might be carried out to gain energy as many MOB are non-obligatory CH4 oxidizers, and (iii) current methods are unable to distinguish aerobic from anaerobic PMO, while the sediment composition might lead to a higher proportion of either through differences in oxygen availability [57]. The different pathways and efciencies of aerobic and anaerobic PMO might not allow a simple relationship between abundance and PMO.

4. Conclusions

Our results demonstrate that the methanogenic and methanotrophic microbiomes of the investigated European streams mainly group into two major habitat types described by differing abiotic conditions. Such distinguishable microbiomes suggest, that future climate- and land-use changes may infuence the prevailing microbes involved in the large scale stream-related CH4 cycle. Increasing water temperatures in combination with an intensifcation of agricultural and urban land use might thereby lead to higher abundances of highly efcient hydrogenotrophic CH4-producers (i.e., Methanospirillaceae,

Methanobacteriaceae and Methanosarcinaceae),, while for CH4-consumers the less efcient type II MOB might gain importance. This combination potentially leads to a future increase of CH4 release from stream ecosystems.

Furthermore, links between sediment organic carbon content with MA and MOB abundance, and between MA abundances and PMP emphasize the importance of nutrient input for the magnitude of processes involved in the CH4 cycle and thus for the overall CH4 emission from running waters. While the general importance of nutrient effects on CH4 cycling has been highlighted in many previous studies (reviewed by [5]), our results suggest, that these effects are mediated by the effects on the microbial communities driving this cycle. We recommend that this nutrient effect should be addressed more intensively in future studies, including nutrient sources such as phosphorous.

5. Explanatory Power Of The Study & Recommendations For Future Studies

On the basis of environmental variables (PCA, Fig. 1) and their combination with community data (RDA, Fig. 4), we were able to depict characteristic environments for the occurrence of specifc communities of Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 15/35 MA (two environments) and MOX (three environments). Due to time constraints of the experimental design utilized in this study (i.e., coordinated collaborative collection of all samples across Europe within a two-week period to minimize temporal variability), some potentially relevant environmental parameters such as light availability/canopy cover, nutrient availability of the porewater etc. could not be measured. Their inclusion into future studies is encouraged in order to refne the defnition of the environments described here, which nonetheless represent, to the best of our knowledge, the frst described microbiomes of stream MA and MOX communities.

Furthermore, with the aim to cover a wide span of sediment characteristics from coarse to fne, three sediment patches per stream stretch were sampled, while the same environmental data (e.g., fow velocity, surface water oxygen content) were attributed to all sediment patches within one stream (Figs. 1 and 3), neglecting sediment-patch-specifc microenvironments. Acknowledging this shortcoming we strongly recommend to measure physical and chemical parameters as well as fow conditions on a microhabitat scale in further studies.

6. Methods 6.1. Sampling strategy

We sampled 16 streams across 10 European countries (Tab. S1), using the network of the EuroRun project [66-68]. Within a representative 50 m section of the investigated stream [69] three areas were selected for sampling based on their differing sediment grain size; (1) coarse (gravel, >0.2 cm to 2 cm), (2) medium (sand and mud, >6 μm to 2 mm), and (3) fne (silt/loam/clay, <6 μm). If one sediment type was not present or could not be sampled, sediment patches with the widest possible range of sediment grain size were selected (coarse, medium, and fne sediment). Particle size distribution were additionally measured in the laboratory in order to check for misalignments and comparability of the categories across the streams. In most cases the categories did not agree with the measured grain sizes and, therefore, the three areas were considered as replicates of the same stream. Sediment surface area (measured) was used as parameter describing within-stream heterogeneity (see chapter 6.2.)

From each of the three selected sediment classifcations, 10 sediment cores were sampled with cut-off 100 ml syringes (diameter: 3.6 cm), avoiding sediment vertical profle disturbance or compression. Cores were then separated in the feld into three sediment layers (0-3 cm, 3-6 cm, and 6-9 cm below the surface water interface, SWI), and transferred to 30 ml glass vials. Following their collection samples were always transported in cooling boxes, and stored at 4°C, with shipping to the respective laboratories occurring the following day by express mail. Three replicates of the top (0-3 cm) and bottom (6-9 cm) sediment layers per patch were combined to form one composite sample of each depth, and a subsample was taken for DNA extraction (n = 96; stored at -20°C; processed between 7 and 21 days after sampling). Furthermore, three 0-3 cm sediment replicates per sediment patch were used for potential methane oxidation rate (PMO) measurements (n = 144; stored at 3 °C; processed between 5 and 8 days after sampling), and four seLdoaimdinegn [tM caothrJeasx ]o/jfa xa/lolu tthpuret/eC odmemptohnsH TpMerL /sfeondtism/TeenXt/ fponattdcahta w.jsere used for potential methane production rate

Page 16/35 (PMP) measurements (n = 384; stored at 4 °C; processed between 3 and 10 days after sampling). The samples for DNA-extraction and subsequent 16S rRNA and qPCR analyses were processed at the University of Innsbruck, the samples for PMO at the University of Basel, and the samples for PMP at the University of Koblenz-Landau.

6.2. Physical and chemical parameters

The upstream catchment area, altitude (meters above sea level), Strahler Stream order [70], and land use within the catchment were taken from [66], which covered the same sites (Tab. S1). Land use within the catchment was further simplifed after inspection via PCA (see section data analysis) by unifying the classes “urban” and “agriculture” as well as “forest” and “others” into “anthropogenically altered” and “natural”, respectively. As those two variables represented inverse values, only one of them (“natural”) was used for those analyses that could be infuenced by autocorrelated variables (i.e., the RDA and GLMM).

Three cross-sectional areas per stream were calculated by measuring the wetted stream width and the water depth at fve characteristic points of the riverbed (e.g., deepest point, slope toe) perpendicular to the main fow direction and summing the partial areas. The stream mean cross-sectional area (m2), water depth (m), and stream width (m) were computed by averaging the corresponding data obtained at the three cross-sections.

Stream area (m2) was calculated by measuring the wetted stream width every fve meters, calculating partial areas and summing them up. In the case of three streams in Spain (ESP2_1, ESP2_2. ESP2_3), only the frst fve partial areas were summed up and multiplied by two in order to get the area of the 50 m stream section.

Discharge (m3 s-1) was measured at one cross-section at an almost straight area of the studied 50 m stream section, without recirculating fows and without protruding elements creating local fow accelerations. Eleven depth measurements were performed, and at fve evenly spaced spots (wetted stream width divided by fve) fow velocities at 20 % and 80 % of the local water depth were measured (if water depth was below 20 cm, one measurement at 60% of the local water depth was performed; see supplementary Tab. S2 for devices used by the individual teams). Mean fow velocity (in m s-1) was calculated by dividing the discharge by the total cross-sectional area.

The slope (in %) of the 50 m stream sections was calculated by averaging fve to six slope measurements along the stream section using a water level gauge. Briefy, a transparent tube flled with water was placed along the stream (about 10 m apart) and the distance of the meniscus within the tube to the water surface was measured at both ends. The slope was calculated by subtracting the meniscus to water surface difference from upstream to downstream, divided by the distance between the measurements.

Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 17/35 Physical and chemical parameters (pH, oxygen (mg L-1), conductivity (µS cm-1), temperature (°C)) were measured during daytime in the middle of the stream, approximately 10 cm below the surface (see supplementary Tab. S2 for devices used by the individual teams).

Particle size distribution was measured from a subsample of each PMP sample at the Thuringian Particle Size Laboratory using a Beckman-Coulter LS 13320 PIDS following B Machalett, EA Oches, M Frechen, L Zöller, U Hambach, NG Mavlyanova, SB Markovic and W Endlicher [71], resulting in 116 reported particle size classes per sample. The Aqueous Liquid Module and an Auto-Prep station were used to measure each sample under equal conditions using a completely degassed and clean water supply and a defned standard operating method [72]. Prior to particle size analysis, representative sub-samples were treated with 11.9 % H2O2 and heated to a maximum temperature of 50° C, aiming at a gentle, yet effective chemical treatment to remove organic matter that may not alter the particle size of the primary sediment [73]. All samples were sieved using a 1800 µm sieve, in order to remove and quantify particles larger than 1800 µm. All weighted samples were dispersed, by spiking each sample with 1% ammonia solution

(NH4OH) and treating them for at least 12 hours in overhead tube rotators. Results are reported in volume percent for each individual particle size class [74]. Finally, the 116 particle size classes were grouped into four particle classes: clay (0,04 µm – 2,2076 µm), silt (2,2076 µm - 63,4192 µm), sand (63,4192 µm – 1800 µm) and > 1800 µm (i.e., coarse). Due to material shortage, this analysis was only conducted on 512 of the 574 collected samples. To estimate the area that can be potentially colonized by the microorganisms, a parameter describing the sediment surface area (cm² cm-3) was calculated using the mean size of each particle class assuming a spherical shape of all particles.

Sediment nitrogen and organic carbon content was measured from PMP samples after the incubation period, assuming the consumption under anoxic conditions in this period of time to be negligible. The sediment was freeze dried, sieved with a mesh size of 2 mm (D = 200 µm, ISO 3310-1, ATECHNIK, Germany) and the subsequent < 2 mm fraction homogenised by grinding. A known amount (15 – 20 mg) of ground material was weighed into silver boats, exposed to HCl (37 %, 12 M) vapor for 6 hours in order to remove carbonates [75] and dried again for 4 hours at 60 °C. Prior to dry combustion (Vario MICRO Cube, Elementar Analysensysteme GmbH, Germany), 20 – 30 mg tungsten trioxide (WO3) was added to the sample. Nitrogen and organic carbon content of the sediment is expressed as weight percent of dry sediment.

6.3. Potential methane production and potential methane oxidation

PMP was measured according to Wilkinson et al. [23]. Briefy, a subsample was transferred to an incubation vial (fushed with nitrogen and crimp sealed with a butyl rubber stopper) in a glove box fushed -1 with nitrogen gas under low oxygen conditions (ca. 0.5 mg O2 L ; atmospheric concentration ca. 12 mg -1 O2 L ), and incubated for about fve weeks at 16 °C in the dark. Weekly, 100 µL gas samples were Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 18/35 withdrawn from the headspace using a gas-tight syringe (1710RN, 100 µL, Hamilton, Switzerland) for analysis in a closed loop with an Ultraportable Greenhouse Gas Analyzer (UGGA, model 915-0011, Los Gatos Research Inc., Mountain View Calif., USA) [23, 76, 77]. PMP was calculated from the linear increase -3 -1 in CH4 partial pressure in the headspace over time (g CH4 m d ).

PMO was measured according to [78] and [79]. Briefy, samples were amended with trace amounts of 14C- labelled aqueous CH4 solution (10 µL, 24 kBq, ~75 nmol CH4, American Radiolabeled Chemicals) and incubated in the dark for ~3 days at room temperature. Then, incubations were stopped by fxing 14 extruded sediment samples in 20 mL 2.5% sodium hydroxide. PMO was assessed by CH4 combustion 14 -1 [80] and CO2-acidifcation [81] and expressed in % CH4 d . All samples were measured in triplicates and results were corrected for abiotic tracer turnover.

6.4. DNA Extraction

DNA extraction was carried out on approx. 500 mg of fresh sample of the composite sediment for each depth using the Nucleo Spin® Soil Kit (Macherey-Nagel, Düren, Germany), choosing buffer SL1 as the lysis buffer and following the manufacturer’s instructions. DNA quality was checked on 1 % (w/v) agarose gels using GelGreen™ staining (Biotium Inc., Fremont, Canada) and its quantity was measured using a Quantus™Fluorometer and the QuantiFluor® Dye System for double-stranded DNA (dsDNA) (Promega, Fitchburg, MA, USA).

6.5. Quantitative PCR

Quantitative PCR (qPCR) was conducted using the SensiFastTM SYBR No-Rox Kit (Bioline, UK) on a Corbett Life Science Rotor-GeneTMQ system (Qiagen, Germany). Total mcrA genes of methanogenic archaea (MA) were determined using the primer set mlas/mcrA-rev (469 bp) (Steinberg and Regan, 2009) and the abundance of pmoA gene of type Ia and type II methane-oxidizing bacteria was determined using the primer pairs A189f/Mb601r (432 bp) and II223f/II646r (444 bp) respectively, with cycling conditions summarized in Tab. S3. In contrast to the sequencing approach, the qPCR approach does not include anaerobic methane oxidizing archaea (ANME) and the quantifed group targeted with the pmoA is thus abbreviated as MOB (methane oxidizing bacteria). qPCR reactions (20 µL) contained 10 µL SensiFastTM SYBR No-Rox, 0.3 µM of each primer (10 µM), 5 mM

MgCl2, 0.04 % (v/v) bovine serum albumin, and 2 µL of DNA template. Non-template DNA and non- template controls (UltraPure DNase/RNase Free Distilled Water, Invitrogen, USA) were included in the runs. After amplifcation, PCR products were quality-controlled via melt-curve analysis. Pure culture DNA for the generation of calibration curves was extracted from Methanosarcina barkeri (DSMZ 800), Methylosarcina fbrata (DMSZ 13736) and Methylosinus sporium (DMSZ 17706) for MA and MOB type Ia and type II, respectively and PCR-amplifying them with the primer pairs stated above. Amplicons were Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 19/35 purifed using the GenEluteTM PCR Clean-Up Kit (Sigma-Aldrich, St. Louis, USA) and diluted to ft the range from 105 to 101 (MA) or 102 to 106 (MOB) gene copies µL-1 after measuring their concentrations using the QuantiFluor® Dye System for double-stranded DNA (dsDNA) (Promega, Fitchburg, MA, USA). All standard curves exhibited a high model ft R² > 0.999, and all qPCR runs were conducted with 40 cycles. Calculation of gene copy number per g fresh sample was performed using the Rotor-Gene 6000 Series Software 1.7 (Qiagen, Germany).

Finally, gene copies of type Ia and type II MOB were summed in order to represent the fnal gene copy number for MOB.

6.6. 16S rRNA gene sequencing and bioinformatics

A total of 96 DNA extracts resulting from (composite) triplicates of 16 streams and two depths was subjected to high-throughput amplicon sequencing on an Illumina MiSeq platform performing a 300 bp paired-end run and targeting bacteria and archaea via the V4 region of the 16S SSU rRNA gene (Microsynth, Austria) using the 515f/806r primer system [82].

Data were processed applying the CoMA pipeline [83]. In brief, merging of paired-end reads and trimming of barcodes- and primers were done using the recommended settings. High-quality reads were selected to show < 5 % deviation from the mode sequence length and at least 99.7 % base call accuracy (phred score ≥ 25). Sequences were aligned applying a 97 % similarity level and assigned taxonomically using the blast algorithm against SILVA SSU (release 123) and Greengenes (release 13_5) as primary and backup databases, respectively. Operational taxonomic units (OTUs) represented by only one read within all samples were excluded and samples were subsampled to 73249 reads after rarefaction analysis, thereby dropping two samples (FRA1_1_2_B and FRA1_1_2_C) with considerably lower read numbers (37630 and 34099 reads, respectively). From the resulting OTU-fles, MA and methane-oxidizing microorganisms ( containing both, archaea and bacteria and therefore abbreviated as MOX) were extracted referring to literature data [6, 15] and used for subsequent analyses.

6.7. Data analysis

6.7.1. Environmental data

To condense the multivariate dataset and explore the environmental conditions of the streams, a principal component analysis (PCA) was performed using the FactoMineR and factoextra packages in R [84-86].

6.7.2.Community data Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 20/35 Differences among all sequenced OTUs from different sediment depths (0-3 cm, 6-9 cm) were tested with PERMANOVA (using Adonis function) based on Bray-Curtis dissimilarity matrix of both MA and MOX communities using the vegan package in R [84, 87]. Additionally, variations in community composition as a function of stream and depth were tested applying a two-way PERMANOVA with 9999 permutations again based on Bray-Curtis dissimilarities of MA and MOX OTU communities using the vegan package in R [87]. Sampling depth itself did not signifcantly infuence the community composition of MA and MOX, thus all subsequent analyses were performed on sequencing data of both sampling depths (n = 96): α- diversity was inspected calculating Shannon diversities and OTU richness (Chao index) using the Mothur software pipeline v.1.39.0 (64 bit executable) and averaging their distribution on family-level in each stream (i.e., combining the three samples of supposedly different particle sizes). The top fve families for each stream in the upper and lower sediment sections were determined by calculating mean ranks of all families. β-diversities of both communities were studied performing an NMDS with Bray-Curtis dissimilarities on OTU-level of MA and MOX including both sediment layers.

Relationships between log-transformed relative read numbers of MA- and MOX-families with each other and with all environmental variables were examined by generating Pearson correlation tables with the Corrplot package, using a signifcance level of 1% [84, 88]. Infuence of environmental variables on communities of MA and MOX communities were determined via redundancy analyses (RDA) performed with 499 unrestricted permutations on log-transformed read numbers using Canoco5 [89]. Only one representative of each auto-correlating group of environmental variables identifed in the PCA (see above) was included for RDA analysis and signifcant environmental variables were selected via the default interactive forward selection process.

6.7.3 Gene abundance data

Relationships between MA and MOB, between PMP and MA, and between PMO and MOB were studied using generalized linear models (GLM) on log-transformed data. Differences of methanogenic and methanotrophic gene copy numbers at different sediment depths were evaluated and tested using pairwise comparisons (Wilcoxon signed-rank test) between both depths (0-3 and 6-9 cm) in all technical replicate samples with the package ggpubr [90] and visualized with ggplot2 [91]. Differences of the same gene copy numbers between distinct streams were evaluated using Kruskal-Wallis rank sum test and best correlating stream-scale variables were identifed with GLMs on the data aggregated by stream (n=16) with forward selection based on Bayesian information criterion, using a log-link-function and considering multicollinearity between predictors.

Generalized linear mixed models (GLMMs) were used to disentangle the site-specifc environmental drivers of MA and MOB gene copy numbers, while considering stream-scale differences (n = 96). Hence, stream identity was considered as random effect and the following site-specifc variables were included as fxed effects: concentration of organic carbon, substrate surface area, nitrogen content, and C:N ratio. Various models with different combinations of fxed effects were constructed using the packages lme4 Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 21/35 [92] and r2glmm [93], multicollinearity between variables was previously checked using the package corrplot [82]. The fnal GLMM for the two dependent microbial groups (MA and MOB) was selected based on the GLMMs ft by examining both the mixed effect models conditional R² according to [35] and the prediction quality of gene copy numbers (correlation between observed and predicted gene copies). The importance of each environmental variable was evaluated using the percentage of deviance explained.

7. Declarations

Ethics approval and consent to participate

Not applicable

Consent for publication

Not applicable

Availability of data and material

The datasets supporting the conclusions of this article are available in the NCBI repository under BioProject PRJNA578023, BioSample accessions from SAMN13046387 to SAMN13046482 and SRA IDs from 13046387 to 13046482, respectively (https://www.ncbi.nlm.nih.gov/sra/PRJNA578023).

Competing interests

The authors declare that they have no competing interests.

Funding

This study was fnancially supported by the German Research Foundation (grant BO 5050/1-1). NC hold a Beatriu de Pinos grant (BP 2016-00215) from the Government of Catalonia's Secretariat for Universities and Research of the Ministry of Economy and Knowledge.

Authors' contributions

MN analysed the DNA-samples, performed bioinformatics and statistics and wrote the major part of the manuscript, NP performed laboratory work, performed bioinformatics and statistics, helped with the interpretation of the results and wrote parts of the manuscript, GHN and NC performed statistical tests and wrote parts of the manuscript, KA generated the conceptual fgure and gave major inputs to the manuscript, BM performed laboratory analyses, LS performed laboratory analyses and wrote parts of the manuscript, PG performed language editing, and PB coordinated the project, was responsible for the study design, wrote major parts of the manuscript and discussed and interpreted the results. Besides NP, BM and LS, all authors selected sites, participated on the sampling campaign and data collection, and contributed to the different versions of the manuscript.

Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 22/35 Acknowledgements

We would like to thank Jihyeon Kim, Christian Noss, Merlin Rosner, Haiying Xia for their support during the campaign preparations, laboratory work, and raw data processing. Furthermore, we would like to thank Sebastian Hupfauf, Andreas Lorke and Jeremy Wilkinson for their input regarding bioinformatics, study design and technical questions.

8. References

1.Ciais P, Sabine C, Bala G, Bopp L, Brovkin V, Canadell J, Chhabra A, DeFries R, Galloway J, Heimann M et al: Carbon and other biogeochemical cycles. In: Climate Change 2013: the physical science basis Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Edited by Stocker TF, Qin D, Plattner G-K, Tignor M, Allen SK, Boschung J, Nauels A, Xia Y, Bex V, Midgley PM. Cambridge, United Kingdom and New York, NY USA; 2013.

2.Shindell DT, Faluvegi G, Koch DM, Schmidt GA, Unger N, Bauer SE: Improved Attribution of Climate Forcing to Emissions. Science 2009, 326(5953):716–718.

3.Hamdan LJ, Wickland KP: Methane emissions from oceans, coasts, and freshwater habitats: New perspectives and feedbacks on climate. Limnology and Oceanography 2016, 61(S1):S3-S12.

4.Saarnio S, Winiwarter W, Leitão J: Methane release from wetlands and watercourses in Europe. Atmospheric Environment 2009, 43(7):1421–1429.

5.Stanley EH, Casson NJ, Christel ST, Crawford JT, Loken LC, Oliver SK: The ecology of methane in streams and rivers: patterns, controls, and global signifcance. Ecological Monographs 2016, 86(2):146– 171.

6.Nazaries L, Murrell JC, Millard P, Baggs L, Singh BK: Methane, microbes and models: fundamental understanding of the soil methane cycle for future predictions. Environmental Microbiology 2013, 15(9):2395–2417.

7.Cicerone RJ, Oremland RS: Biogeochemical aspects of atmospheric methane. Global Biogeochemical Cycles 1988, 2(4):299–327.

8.Alpana S, Vishwakarma P, Adhya TK, Inubushi K, Dubey SK: Molecular ecological perspective of methanogenic archaeal community in rice agroecosystem. Science of The Total Environment 2017, 596– 597:136–146.

9.Wen X, Yang S, Horn F, Winkel M, Wagner D, Liebner S: Global Biogeographic Analysis of Methanogenic Archaea Identifes Community-Shaping Environmental Factors of Natural Environments. Frontiers in Microbiology 2017, 8(1339).

Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 23/35 10.Raymond PA, Hartmann J, Lauerwald R, Sobek S, McDonald C, Hoover M, Butman D, Striegl R, Mayorga E, Humborg C et al: Global carbon dioxide emissions from inland waters. Nature 2013, 503:355.

11.Bomberg M, Montonen L, Münster U, Jurgens GN: Diversity and function of archaea in freshwater habitats. Current Trends in Microbiology 2008, 4:61–89.

12.Chaudhary PP, Rulík M, Blaser M: Is the methanogenic community refecting the methane emissions of river sediments?—comparison of two study sites. MicrobiologyOpen 2017, 6(4):e00454.

13.Li Q, Wang F, Chen Z, Yin X, Xiao X: Stratifed active archaeal communities in the sediments of Jiulong River estuary, China. Frontiers in Microbiology 2012, 3(311).

14.Zeleke J, Lu S-L, Wang J-G, Huang J-X, Li B, Ogram AV, Quan Z-X: Methyl Coenzyme M Reductase A (mcrA) Gene-Based Investigation of Methanogens in the Mudfat Sediments of Yangtze River Estuary, China. Microbial Ecology 2013, 66(2):257–267.

15.Kalyuzhnaya MG, Gomez OA, Murrell JC: The Methane-Oxidizing Bacteria (Methanotrophs). In: Taxonomy, Genomics and Ecophysiology of Hydrocarbon-Degrading Microbes. Edited by McGenity TJ. Cham: Springer International Publishing; 2019: 1–34.

16.Timmers PHA, Welte CU, Koehorst JJ, Plugge CM, Jetten MSM, Stams AJM: Reverse Methanogenesis and Respiration in Methanotrophic Archaea. Archaea 2017, 2017:22.

17.Ettwig KF, Butler MK, Le Paslier D, Pelletier E, Mangenot S, Kuypers MMM, Schreiber F, Dutilh BE, Zedelius J, de Beer D et al: Nitrite-driven anaerobic methane oxidation by oxygenic bacteria. Nature 2010, 464:543.

18.Chistoserdova L, Vorholt JA, Lidstrom ME: A genomic view of methane oxidation by aerobic bacteria and anaerobic archaea. Genome Biology 2005, 6(2):208.

19.Rahalkar M, Bussmann I, Schink B: Methylosoma difcile gen. nov., sp. nov., a novel methanotroph enriched by gradient cultivation from littoral sediment of Lake Constance. International Journal of Systematic and Evolutionary Microbiology 2007, 57(5):1073–1080.

20.Costello AM, Auman AJ, Macalady JL, Scow KM, Lidstrom ME: Estimation of methanotroph abundance in a freshwater lake sediment. Environmental Microbiology 2002, 4(8):443–450.

21.Bednařík A, Blaser M, Matoušů A, Hekera P, Rulík M: Effect of weir impoundments on methane dynamics in a river. Science of The Total Environment 2017, 584–585:164–174.

22.Sanders IA: Emission of methane from chalk streams has potential implications for agricultural practices. Freshwater biology 2007, v. 52(no. 6):pp. 1176–1186–2007 v.1152 no.1176.

Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 24/35 23.Wilkinson J, Bodmer P, Lorke A: Methane dynamics and thermal response in impoundments of the Rhine River, Germany. Science of The Total Environment 2019, 659:1045–1057.

24.Wilkinson J, Maeck A, Alshboul Z, Lorke A: Continuous Seasonal River Ebullition Measurements Linked to Sediment Methane Formation. Environmental Science & Technology 2015, 49(22):13121–13129.

25.Brablcová L, Buriánková I, Badurová P, Chaudhary PP, Rulík M: Methanogenic archaea diversity in hyporheic sediments of a small lowland stream. Anaerobe 2015, 32:24–31.

26.Mach V, Blaser MB, Claus P, Chaudhary PP, Rulík M: Methane production potentials, pathways, and communities of methanogens in vertical sediment profles of river Sitka. Frontiers in Microbiology 2015, 6(506).

27.Crawford JT, Loken LC, West WE, Crary B, Spawn SA, Gubbins N, Jones SE, Striegl RG, Stanley EH: Spatial heterogeneity of within-stream methane concentrations. Journal of Geophysical Research: Biogeosciences 2017, 122(5):1036–1048.

28.Utsumi M, Nojiri Y, Nakamura T, Nozawa T, Otsuki A, Takamura N, Watanabe M, Seki H: Dynamics of dissolved methane and methane oxidation in dimictic Lake Nojiri during winter. Limnology and Oceanography 1998, 43(1):10–17.

29.Bastviken D, Cole JJ, Pace ML, Van de Bogert MC: Fates of methane from different lake habitats: Connecting whole-lake budgets and CH4 emissions. Journal of Geophysical Research: Biogeosciences 2008, 113(G2).

30.Bastviken D, Ejlertsson J, Tranvik L: Measurement of Methane Oxidation in Lakes: A Comparison of Methods. Environmental Science & Technology 2002, 36(15):3354–3361.

31.Oswald K, Milucka J, Brand A, Littmann S, Wehrli B, Kuypers MMM, Schubert CJ: Light-Dependent Aerobic Methane Oxidation Reduces Methane Emissions from Seasonally Stratifed Lakes. PLOS ONE 2015, 10(7):e0132574.

32.Steinle L, Maltby J, Treude T, Kock A, Bange HW, Engbersen N, Zopf J, Lehmann MF, Niemann H: Effects of low oxygen concentrations on aerobic methane oxidation in seasonally hypoxic coastal waters. Biogeosciences 2017, 14(6):1631–1645.

33.Shelley F, Abdullahi F, Grey J, Trimmer M: Microbial methane cycling in the bed of a chalk river: oxidation has the potential to match methanogenesis enhanced by warming. Freshwater Biology 2015, 60(1):150–160.

34.Hanson RS, Hanson TE: Methanotrophic bacteria. Microbiol Rev 1996, 60(2):439–471.

35.Nakagawa S, Schielzeth H: A general and simple method for obtaining R² from generalized linear mLioxaeddin-egf [fMeactthsJ amx]/ojadxe/olsu.t pMute/Cthomodmso ninH TEMcLo/lfoogntys /aTneXd/ fEovnotdlauttai.ojsn 2012, 4(2):133–142.

Page 25/35 36.Hanson CA, Fuhrman JA, Horner-Devine MC, Martiny JBH: Beyond biogeographic patterns: processes shaping the microbial landscape. Nature Reviews Microbiology 2012, 10:497.

37.Yang S, Liebner S, Winkel M, Alawi M, Horn F, Dörfer C, Ollivier J, He J-s, Jin H, Kühn P et al: In-depth analysis of core methanogenic communities from high elevation permafrost-affected wetlands. Soil Biology and Biochemistry 2017, 111:66–77.

38.Fan X, Xing P: Differences in the Composition of Archaeal Communities in Sediments from Contrasting Zones of Lake Taihu. Frontiers in Microbiology 2016, 7(1510).

39.Nozhevnikova AN, Nekrasova V, Ammann A, Zehnder AJB, Wehrli B, Holliger C: Infuence of temperature and high acetate concentrations on methanogenensis in lake sediment slurries. FEMS Microbiology Ecology 2007, 62(3):336–344.

40.Withers PJA, Lord EI: Agricultural nutrient inputs to rivers and groundwaters in the UK: policy, environmental management and research needs. Science of The Total Environment 2002, 282–283:9– 24.

41.Yang Y, Chen J, Tong T, Li B, He T, Liu Y, Xie S: Eutrophication infuences methanotrophic activity, abundance and community structure in freshwater lakes. Science of The Total Environment 2019, 662:863–872.

42.Wilhelm L, Besemer K, Fragner L, Peter H, Weckwerth W, Battin TJ: Altitudinal patterns of diversity and functional traits of metabolically active microorganisms in stream bioflms. The Isme Journal 2015, 9:2454.

43.Kurade MB, Saha S, Salama E-S, Patil SM, Govindwar SP, Jeon B-H: Acetoclastic methanogenesis led by Methanosarcina in anaerobic co-digestion of fats, oil and grease for enhanced production of methane. Bioresource Technology 2019, 272:351–359.

44.Oren A: The Family Methanoregulaceae. In: The Prokaryotes: Other Major Lineages of Bacteria and The Archaea. Edited by Rosenberg E, DeLong EF, Lory S, Stackebrandt E, Thompson F. Berlin, Heidelberg: Springer Berlin Heidelberg; 2014: 253–258.

45.Sakai S, Conrad R, Imachi H: The Family Methanocellaceae. In: The Prokaryotes: Other Major Lineages of Bacteria and The Archaea. Edited by Rosenberg E, DeLong EF, Lory S, Stackebrandt E, Thompson F. Berlin, Heidelberg: Springer Berlin Heidelberg; 2014: 209–214.

46.Hurtt GC, Chini LP, Frolking S, Bett RA, Feddema J, Fischer G, Fisk JP, Hibbard K, Houghton RA, Janetos A et al: Harmonization of land-use scenarios for the period 1500–2100: 600 years of global gridded annual land-use transitions, wood harvest, and resulting secondary lands. Climatic Change 2011, 109:117–161.

Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 26/35 47.Lackner N, Hintersonnleitner A, AWagner AO, Illmer P: Hydrogenotrophic Methanogenesis and Autotrophic Growth of Methanosarcina thermophila. Archaea 2018, 2018.

48.Mondav R, Woodcroft BJ, Kim E-H, McCalley CK, Hodgkins SB, Crill PM, Chanton J, Hurst GB, VerBerkmoes NC, Saleska SR et al: Discovery of a novel methanogen prevalent in thawing permafrost. Nature Communications 2014, 5:3212.

49.Oswald K, Graf JS, Littmann S, Tienken D, Brand A, Wehrli B, Albertsen M, Daims H, Wagner M, Kuypers MMM et al: Crenothrix are major methane consumers in stratifed lakes. The Isme Journal 2017, 11:2124.

50.Kits KD, Campbell DJ, Rosana AR, Stein LY: Diverse electron sources support denitrifcation under hypoxia in the obligate methanotroph Methylomicrobium album strain BG8. Frontiers in Microbiology 2015, 6(1072).

51.Oshkin IY, Wegner C, C. L, Glagolev MV, Fillipov IV, Pimenov NV, Liesack W, Dedysh SN: Gammaproteobacterial Methanotrophs Dominate Cold Methane Seeps in Floodplains of West Siberian Rivers. Microbial Ecology 2014, 80(19):5944–5954.

52.Guerrero-Cruz S, Cremers G, van Alen TA, Op den Camp HJM, Jetten MSM, Rasigraf O, Vaksmaa A: Response of the Anaerobic Methanotroph “Candidatus Methanoperedens nitroreducens” to Oxygen Stress. Applied and Environmental Microbiology 2018, 84(24):e01832–01818.

53.Kelly DP, McDonald IR, Wood AP: The Family Methylobacteriaceae. In: The Prokaryotes: Alphaproteobacteria and Betaproteobacteria. Edited by Rosenberg E, DeLong EF, Lory S, Stackebrandt E, Thompson F. Berlin, Heidelberg: Springer Berlin Heidelberg; 2014: 313–340.

54.Dedysh SN, Dunfeld PF: Chapter three - Facultative and Obligate Methanotrophs: How to Identify and Differentiate Them. In: Methods in Enzymology. Edited by Rosenzweig AC, Ragsdale SW, vol. 495: Academic Press; 2011: 31–44.

55.Cai Y, Zheng Y, Bodelier PLE, Conrad R, Jia Z: Conventional methanotrophs are responsible for atmospheric methane oxidation in paddy soils. Nature Communications 2016, 7:11728.

56.Kalyuzhnaya MG, Yang S, Rozova ON, Smalley NE, Clubb J, Lamb A, Gowda GAN, Raftery D, Fu Y, Bringel F et al: Highly efcient methane biocatalysis revealed in a methanotrophic bacterium. Nature Communications 2013, 4:2785.

57.Higashino M: Quantifying a signifcance of sediment particle size to hyporheic sedimentary oxygen demand with a permeable stream bed. Environmental Fluid Mechanics 2013, 13(3):227–241.

58.Kosek K, Luczkiewicz A, Kozioł K, Jankowska K, Ruman M, Polkowska Ż: Environmental characteristics of a tundra river system in Svalbard. Part 1: Bacterial abundance, community structure and nutrient levels. Science of The Total Environment 2019, 653:1571–1584. Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 27/35 59.Hildenbrand C, Stock T, Lange C, Rother M, Soppa J: Genome copy numbers and gene conversion in methanogenic archaea. J Bacteriol 2011, 193(3):734–743.

60.Nagler M, Podmirseg SM, Grifth GW, Insam H, Ascher-Jenull J: The use of extracellular DNA as a proxy for specifc microbial activity. Applied Microbiology and Biotechnology 2018, 102(6):2885–2898.

61.Xiao K-Q, Beulig F, Kjeldsen KU, Jørgensen BB, Risgaard-Petersen N: Concurrent Methane Production and Oxidation in Surface Sediment from Aarhus Bay, Denmark. Frontiers in Microbiology 2017, 8(1198).

62.Grossart H-P, Frindte K, Dziallas C, Eckert W, Tang KW: Microbial methane production in oxygenated water column of an oligotrophic lake. Proc Natl Acad Sci U S A 2011, 108(49):19657–19661.

63.Zhang Y, Cui M, Duan J, Zhuang X, Zhuang G, Ma A: Abundance, rather than composition, of methane- cycling microbes mainly affects methane emissions from different vegetation soils in the Zoige alpine wetland. MicrobiologyOpen 2019, 8(4):e00699.

64.Juottonen H, Hynninen A, Nieminen M, Tuomivirta TT, Tuittila E-S, Nousiainen H, Kell DK, Yrjälä K, Tervahauta A, Fritze H: Methane-cycling microbial communities and methane emission in natural and restored peatlands. Applied and environmental microbiology 2012, 78(17):6386–6389.

65.Ma K, Lu Y: Regulation of microbial methane production and oxidation by intermittent drainage in rice feld soil. FEMS Microbiology Ecology 2011, 75(3):446–456.

66.Bravo AG, Kothawala DN, Attermeyer K, Tessier E, Bodmer P, Ledesma JLJ, Audet J, Casas-Ruiz JP, Catalán N, Cauvy-Fraunié S et al: The interplay between total mercury, methylmercury and dissolved organic matter in fuvial systems: A latitudinal study across Europe. Water Research 2018, 144:172–182.

67.EuroRun [https://freshproject-eurorun.jimdo.com/eurorun/project-eurorun/]

68.Bodmer P, Attermeyer K, Pastor A, Catalán N: Collaborative Projects: Unleashing Early Career Scientists’ Power. Trends in Ecology & Evolution 2019, 34(10):871–874.

69.Meier C, Haase P, Rolauffs P, Schindehütte K, Schöll F, Sundermann A, Hering D: Methodisches Handbuch Fließgewässerbewertung. In: Handbuch zur Untersuchung und Bewertung von Fließgewässern auf der Basis des Makrozoobenthos vor dem Hintergrund der EG-Wasserrahmenrichtlinie. 2006.

70.Strahler AN: Hypsometric (area-altitude) analysis of erosional topography. GSA Bulletin 1952, 63(11):1117–1142.

71.Machalett B, Oches EA, Frechen M, Zöller L, Hambach U, Mavlyanova NG, Markovic SB, Endlicher W: Aeolian dust dynamics in central Asia during the Pleistocene: Driven by the long-term migration, seasonality, and permanency of the Asiatic polar front. Geochem Geophys Geosyst 2008, 9.

Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 28/35 72.Machalett B: Past Atmospheric Circulation Patterns and Aeolian Dust Dynamics Recorded in Eurasian Loess: Utilizing high-resolution particle size analysis and amino acid geochronology. Mensch & Buch Verlag, Berlin, Dissertation, 120 Seiten 2011, Dissertation.

73.Mikutta R, Kleber M, Kaiser K, Jahn R: Review: Organic matter removal from soils using hydrogen peroxide, sodium hypochlorite, and disodium peroxodisulfate. Soil Science Society of America Journal 2005, 69(1):120–135.

74.Machalett B: WEBINAR: Particle size characterization by laser diffraction analysis in geoscience and soil science - background, analyses, application, and interpretation. In: Particle Size Research Letters 1. Beckman Coulter Life Sciences; 2017.

75.Harris D, Horwáth WR, van Kessel C: Acid fumigation of soils to remove carbonates prior to total organic carbon or CARBON–13 isotopic analysis. Soil Science Society of America Journal 2001, 65(6):1853–1856.

76.Wilkinson J, Bors C, Burgis F, Lorke A, Bodmer P: Measuring CO2 and CH4 with a portable gas analyzer: Closed-loop operation, optimization and assessment. PLOS ONE 2018, 13(4):e0193973.

77.Wilkinson J, Bors C, Burgis F, Lorke A, Bodmer P: Correction: Measuring CO2 and CH4 with a portable gas analyzer: Closed-loop operation, optimization and assessment. PLOS ONE 2019, 14(3):e0206080.

78.Steinle L, Schmidt M, Bryant L, Haeckel M, Linke P, Sommer S, Zopf J, Lehmann MF, Treude T, Niemannn H: Linked sediment and water-column methanotrophy at a man-made gas blowout in the North Sea: Implications for methane budgeting in seasonally stratifed shallow seas. Limnology and Oceanography 2016, 61(S1):S367-S386.

79.Su G, Niemann H, Steinle L, Zopf J, Lehmann MF: Evaluating radioisotope-based approaches to measure anaerobic methane oxidation rates in lacustrine sediments. Limnology and Oceanography: Methods 2019, 17(8):429–438.

80.Treude T, Niggemann J, Kallmeyer J, Wintersteller P, Schubert CJ, Boetius A, Jørgensen BB: Anaerobic oxidation of methane and sulfate reduction along the Chilean continental margin. Geochimica et Cosmochimica Acta 2005, 69(11):2767–2779.

81.Blees J, Niemann H, Wenk CB, Zopf J, Schubert CJ, Kirf MK, Veronesi ML, Hitz C, Lehmann MF: Micro- aerobic bacterial methane oxidation in the chemocline and anoxic water column of deep south-Alpine Lake Lugano (Switzerland). Limnology and Oceanography 2014, 59(2):311–324.

82.Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Lozupone CA, Turnbaugh PJ, Fierer N, Knight R: Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proceedings of the National Academy of Sciences 2011, 108(Supplement 1):4516–4522.

83Lo.CadoiMngA [M—aCthoJmaxp]/jaarxa/otuivtpeu Mt/CiocmromboionHmTeM AL/nfoanltyss/Tise X[h/ftotnptsd:a/t/aw.jsww.uibk.ac.at/microbiology/services/coma.html] Page 29/35 84.R Core Team: R: A language and environment for statistical computing. In. Vienna, Austria: R Foundation for Statistical Computing; 2018.

85.Kassambra A, Mundt F: factoextra: Extract and Visualize the Results of Multivariate Data Analyses. In.; 2017.

86.Lê S, Josse J, Husson F: FactoMineR: An R Package for Multivariate Analysis. Journal of Statistical Software; Vol 1, Issue 1 (2008) 2008.

87.Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, D M, Minchin PR, O’Hara RB, Simpson GL, Solymos P et al: vegan: Community Ecology Package. In., 2.5–5 edn; 2019.

88.Wei T, Simko V, Levy M, Xie Y, Jin Y, Zemla J: Corrplot: Visualisation of a correlation matrix. In.; 2017.

89.Šmilauer P, Lepš J: Multivariate Analysis of Ecological Data using CANOCO 5, 2 edn. Cambridge: Cambridge University Press; 2014.

90.Kassambra A: ggpubr: “ggplot2” Based Publication Ready Plots. In., R package version 0.2. edn; 2018.

91.Wickham H: ggplot2: Elegant Graphics for Data Analysis. In. New York: Springer-Verlag; 2016.

92.Bates D, Mächler M, Bolker B, Walker S: Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software 2015, 67:1–48.

93.Jaeger B: 2glmm: Computes R Squared for Mixed (Multilevel) Models. In., version 0.1.2. edn; 2017: R package.

Figures

Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 30/35 Figure 1

Principal component analysis plot of the frst two dimensions based on all environmental variables. Sample sites are colored according to geographical latitude with SWE1_1 representing the northernmost and ESP2_3 the southernmost sampling points. Further information about units may be found in Tab. S1 and S4. Proportions of explained variance are given in brackets.

Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 31/35 Figure 2

Biogeography of (a, b) methanogenic archaea (MA) and (c, d) methane-oxidizing bacteria and archaea (MOX) from 0 to 9 cm sediment depth in streams across Europe a) and c) show relative α-diversities as well as absolute read numbers at family level with sampling sites ordered according to geographical latitude, with southernmost sites to the left and northernmost sites to the right. The line represents the Shannon Diversity Index. b) and d) show the top fve families across European streams selected by mean rank of each family among all sites. Pie charts show percentages of reads with regard to all reads belonging to MA or MOX.

Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 32/35 Figure 3

Pearson correlation heat map of microbial families and environmental parameters Heat map showing signifcant (p<0.01) Pearson correlation coefcients (r) of log10 transformed relative read abundances of MA and MOX families with environmental parameters (a) and among each other (b).

Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 33/35 Figure 4

Triplots of redundancy analysis of a) MA and b) MOX sequences Signifcant environmental parameters selected by forward selection are depicted and are grouping OTUs as well as samples towards their predominant environmental characteristics and preferred habitats, respectively. Explained variations of each axis are given as percentage of all Eigenvalues. The stream-ID color code refers to circles and the

OTLoUa dcinogl o[Mr actohJdaex ]t/oja xla/obuetplsut./CommonHTML/fonts/TeX/fontdata.js

Page 34/35 Figure 5

Gene abundance correlations Correlations of log-transformed mcrA- gene abundances of methanogenic archaea and type Ia MOB(a), type II MOB (b), cumulative MOB (c) and potential methane production rates (PMP) (d). Relation of type Ia (e) and type II MOB (f) and potential methane oxidation rates. MOB = methane oxidizing bacteria. Colors refer to sampled streams described in Figure 1, circles to sediments sampled from 0-3 cm depth and squares to sediments sampled from 6-9 cm depth. Slope and intercept of the ftted lines are only reported for signifcant relationships.

Supplementary Files

This is a list of supplementary fles associated with this preprint. Click to download.

SupplementarymaterialMBIOfnal.docx

Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js

Page 35/35