Vertical distribution of methane oxidation and methanotrophic response to elevated methane concentrations in stratified waters of the Arctic fjord Storfjorden ( Svalbard , Norway )

The bacterially mediated aerobic methane oxidation (MOx) is a key mechanism in controlling methane (CH 4) emissions from the world’s oceans to the atmosphere. In this study, we investigated MO x in the Arctic fjord Storfjorden (Svalbard) by applying a combination of radio-tracerbased incubation assays ( H-CH4 and C-CH4), stable CCH4 isotope measurements, and molecular tools (16S rRNA gene Denaturing Gradient Gel Electrophoresis (DGGE) fingerprinting,pmoAandmxaFgene analyses). Storfjorden is stratified in the summertime with melt water (MW) in the upper 60 m of the water column, Arctic water (ArW) between 60 and 100 m, and brine-enriched shelf water (BSW) down to 140 m. CH4 concentrations were supersaturated with respect to the atmospheric equilibrium (about 3–4 nM) throughout the water column, increasing from ∼ 20 nM at the surface to a maximum of 72 nM at 60 m and decreasing below. MO x rate measurements at near in situ CH 4 concentrations (here measured withH-CH4 raising the ambient CH 4 pool by < 2 nM) showed a similar trend: low rates at the sea surface, increasing to a maximum of ∼ 2.3 nM day−1 at 60 m, followed by a decrease in the deeper ArW/BSW. In contrast, rate measurements with C-CH4 (incubations were spiked with ∼ 450 nM of C-CH4, providing an estimate of the CH4 oxidation at elevated concentration) showed comparably low turnover rates ( < 1 nM day−1) at 60 m, and peak rates were found in ArW/BSW at ∼ 100 m water depth, concomitant with increasing13C values in the residual CH 4 pool. Our results indicate that the MO x community in the surface MW is adapted to relatively low CH 4 concentrations. In contrast, the activity of the deep-water MO x community is relatively low at the ambient, summertime CH 4 concentrations but has the potential to increase rapidly in response to CH 4 availability. A similar distinction between surface and deepwater MOx is also suggested by our molecular analyses. The DGGE banding patterns of 16S rRNA gene fragments of the surface MW and deep water were clearly different. A DGGE band related to the known type I MO x bacteriumMethylosphaerawas observed in deep BWS, but absent in surface MW. Furthermore, the Polymerase Chain Reaction (PCR) amplicons of the deep water with the two functional primers setspmoAandmxaFshowed, in contrast to those of the surface MW, additional products besides the expected one of 530 base pairs (bp). Apparently, different MO x communities have developed in the stratified water masses in Storfjorden, which is possibly related to the spatiotemporal variability in CH4 supply to the distinct water masses.


Introduction
Methane (CH 4 ) is a potent greenhouse gas with a global warming potential that exceeds carbon dioxide (CO 2 ) 23fold over a 100 yr timescale and is, after water vapor and CO 2 , the most important greenhouse gas (IPCC, 2007).Substantial research efforts have consequently been made to understand its sources and sinks.A large part of oceanic CH 4 is generated under reduced conditions in anoxic marine sediments, predominantly through microbially mediated Published by Copernicus Publications on behalf of the European Geosciences Union.

S. Mau et al.: Vertical distribution of methane oxidation
CO 2 reduction and disproportionation of methylated substrates (Whiticar, 1999;Hinrichs and Boetius, 2002;Formolo, 2010).Sedimentary CH 4 is also formed by thermal breakdown of organic matter and, although of lesser importance, serpentinization and Fischer-Tropsch reaction, both occurring at high temperature and pressure.In addition, conspicuous CH 4 concentration maxima in oxic water layers provided indications for CH 4 production under oxic conditions, possibly mediated by yet unknown microbes using dimethylsulfoniopropionate (DMSP) (Damm et al., 2010) or methylphosphonic acid (MPn) (Karl et al., 2008;Metcalf et al., 2012) as substrate.However, despite the apparent ubiquity of methanogenesis in marine systems and the large area covered by oceans, comparably little CH 4 is liberated from the oceans into the atmosphere because of microbial consumption (Reeburgh, 2007;IPCC, 2007).About 80% of sedimentary CH 4 is consumed in reduced sediments as a result of the anaerobic oxidation of methane (AOM) with sulfate as the terminal electron acceptor (Barnes and Goldberg, 1976;Reeburgh, 1976;Martens and Berner, 1977;Reeburgh, 2007;Knittel and Boetius, 2009).Finally, aerobic CH 4 -oxidizing bacteria at the sediment surface and/or in the water column (belonging to the Alpha-(type II) or Gammaproteobacteria (type I and type X)) consume CH 4 that has bypassed the anaerobic microbial filter according to the following reaction (Hanson and Hanson, 1996;Murrell, 2010): Several techniques have been used to quantify aerobic methane oxidation (MO x ) rates (Reeburgh, 2007).A common method is to incubate water column or sediment samples with radio-labeled tracers such as 14 C-CH 4 or 3 H-CH 4 (Reeburgh et al., 1991;Valentine et al., 2001;Niemann et al., 2006;Mau et al., 2012), which has proven to be highly sensitive.During the incubation, 14 C-CH 4 or 3 H-CH 4 is converted at the same rate as the natural, non-labeled CH 4 to 14 CO 2 and 14 C-biomass or 3 H 2 O.Despite the importance of water column MO x controlling oceanic CH 4 emission to the atmosphere, only a small number of water column MO x rate measurements exist, which is particularly true for high-latitude environments (Ward and Kilpatrick, 1990;Griffiths et al., 1982).The available data show a large scatter of rates over several orders of magnitude (Fig. 1), but factors controlling MO x activity such as temporal variations in CH 4 availability (e.g., Mau et al., 2007a, b;Damm et al., 2007) and the activity of the present MO x community during time periods with elevated CH 4 availability are not well constrained.Our aims were to investigate MO x rates at ambient and at elevated CH 4 concentrations as well as to determine differences between MO x communities thriving at different CH 4 concentrations in a natural marine environment.As a model system, we chose the fjord Storfjorden (Svalbard), which is characterized by seasonal stratification, separating distinct water masses with different CH 4 sources during summer time.

Study site
Storfjorden is located in the Svalbard Archipelago between the islands Spitsbergen, Barentsøya, and Edgeøya (Fig. 2).CH 4 concentrations in the fjord water exceed atmospheric equilibrium concentration throughout the water column by a factor of 2-16, although surface water CH 4 is of a different origin compared to the CH 4 in subsurface waters (Damm et al., 2008).Surface waters contain recently produced, 13 C-depleted CH 4 , which was proposed to result from a summer phytoplankton bloom producing methylated compounds such as DMSP, which is a potential substrate for methylotrophic methanogenesis.Other potential methanogenic substrates such as methylphosphonates were not investigated in the study area.A CH 4 productionremoval cycle appears to be established in the surface water as reflected by varying CH 4 concentrations and 13 C-CH 4 values (Damm et al., 2008).In contrast, deeper water contains CH 4 that is mixed into the bottom water as a result of brine-enriched shelf water (BSW) formation during wintertime causing enhanced turbulence and repeatedly occurring resuspension of sediments releasing CH 4 (Damm et al., 2007).The winter-released CH 4 is then trapped by increasing water stratification during warmer seasons, and ongoing CH 4 consumption leads to a 13 C-enriched isotopic signature of the residual CH 4 .During summertime, the water column is stratified, with surface melt water (MW) and intermediate Arctic water (ArW) constituting the upper water column, while denser BSW is restricted to deep basins (Loeng, 1991).
The residence time of the high-salinity water in deeper layers (90-246 days) is longer than the fjord's surface waters (51-141 days) (Geyer et al., 2009).

Sampling
Water samples were collected from nine stations in Storfjorden and at one open-ocean station (70 • 35.913N, 10 • 51.591 E) during a cruise with RV Heincke in August 2010 (Fig. 2, Table 1).The Storfjorden stations were aligned along the cyclonic coastal current flowing into Storfjorden along Edgeøya and out along Spitsbergen (Loeng, 1991;Skogseth et al., 2005) (Fig. 2).We intended to sample and compare the fjord's upper and lower water column because of the different CH 4 sources and water residence times.We sampled vertical profiles throughout the water column, thus recovering samples from MW, ArW, and BSW.All water masses were subsampled for chemical/biogeochemical analyses (method 2.3 and 2.4), but we focused on the MW and BSW for molecular analyses (method 2.5).Specific water depths were sampled with a CTD/rosette sampler equipped with twelve 5 L Niskin bottles, a Sea-Bird SBE 911 plus CTD and an SBE 43 oxygen sensor for online monitoring of salinity, temperature, pressure, and dissolved oxygen.

CH 4 concentrations and stable isotope composition
Aliquots of seawater were immediately subsampled from the Niskin bottles using 1 L glass bottles for measurements of in situ CH 4 concentrations.CH 4 was extracted from the water by vacuum-ultrasonic treatment within a few hours af- ter sampling (Schmitt et al., 1991).Hydrocarbon concentrations were measured with a Chrompack 9003 gas chromatograph (GC) equipped with a flame ionization detector (FID).
Duplicate measurements indicate an error of 5-10 % (Lammers and Suess, 1994).After GC analyses, an aliquot of the extracted CH 4 gas was transferred into pre-evacuated glass containers for stable carbon isotope analysis performed with an isotope ratio mass spectrometer (IRMS; Finnigan Delta XP plus) in our onshore laboratories.The extracted gas was purged and trapped with the PreCon equipment (Finnigan) to preconcentrate the sample.All isotopic ratios have an analytical error < 1 ‰ and are presented in the δ notation against the Vienna Pee Dee Belemnite (VPDB) standard.

Methane oxidation rates
MO x rates were determined from ex situ incubations of water samples in 100 mL serum vials.The vials were filled bubblefree from Niskin bottles and crimped with rubber stoppers (halogenated butyl elastomer).One set of samples was then incubated with 50 µL of 3 H-labeled CH 4 (160-210 kBq) in N 2 , and a second set was incubated with 10 µL of 14 Clabeled CH 4 (12-15 kBq). 3H-CH 4 tracer addition raised ambient CH 4 concentrations by 1-2 nM and 14 C-CH 4 addition by 440-540 nM.The samples were subsequently shaken for ∼ 10 min on an orbital shaker to facilitate tracer dissolution and then incubated in the dark at 2 • C. CH 4 oxidation rates (r ox ) were calculated assuming first-order kinetics (Reeburgh et al., 1991;Valentine et al., 2001): where k is the effective first-order rate constant calculated as the fraction of labeled CH 4 oxidized per unit time, and [CH 4 ] is the in situ CH 4 concentration.In order to determine a suitable incubation time period, we performed parallel time series incubations with samples collected from the fjord (stations 2 and 18) and from an open-water station (reference www.biogeosciences.net/10/6267/2013/Biogeosciences, 10, 6267-6278, 2013 station -RS).During each incubation series, tracer consumption was measured in duplicates after a time period of 0.5, 1, 2, 3, 4 and 5 days.In the CH 4 -rich waters of the fjord, our results showed a linear tracer consumption of about 5-15 % over the first three days of incubation (Fig. 3).A potential bias due to substrate limitation and/or variations in reaction velocity thus seems negligible, at least over a time period of 3 days, which we chose for our ex situ incubations.Just as the time series incubations, vertical distribution of MO x was determined in duplicates.
Incubations with 3 H-CH 4 and measurements of 3 H-CH 4 and 3 H-H 2 O were carried out according to Valentine et al. (2001) and Mau et al. (2012).Briefly, total activity ( 3 H-CH 4 + 3 H-H 2 O) was measured in 1 mL of sample aliquot by wet scintillation counting, and activity of 3 H-H 2 O was measured after sparging the sample for > 30 min with N 2 to remove remaining 3 H-CH 4 .
Incubations with 14 C-CH 4 were terminated by injecting 0.5 mL of 10 M NaOH and adding a 5 mL headspace so that the remaining 14 C-CH 4 accumulated in the headspace and the produced 14 C-CO 2 and 14 C biomass was trapped in the aqueous NaOH solution.Separation and activity measurement of 14 C-CH 4 and 14 C-CO 2 were carried out analogously to previous measurements of CH 4 turnover in sediments (Treude et al., 2003;Niemann et al., 2005).In short, 14 C-CH 4 in the headspace was combusted to 14 C-CO 2 , while 14 C-CO 2− 3 was converted to 14 CO 2 through acidification with HCI.In either case, 14 C-CO 2 was trapped in a solution of methoxyethanol and phenylethylamine, and the radioactivity was measured by wet scintillation counting.We also measured remaining radioactivity in the sample after 14 C-CH 4 and 14 CO 2− 3 removal (probably 14 C incorporated into biomass), which ranged between 4 and 84 % (average 35 %) of the total product.

Diversity of MO x community
The diversity of the natural bacterioplankton assemblages was examined by denaturing gradient gel electrophoresis (DGGE) based on the 16S rRNA gene.Immediately after sampling, bacterial cells were concentrated on Nuclepore filters (0.2 µm pore size) and the filters were stored frozen at −20 • C until DNA extraction.Total community DNA was extracted using the UltraClean Soil DNA Kit (MoBio Laboratories, USA).One to five microliters of DNA extract was applied as the template in the 16S rRNA gene specific PCR, with GM5 plus GC-clamp as the forward primer and 907RM as the reverse primer (Muyzer et al., 1993).PCR conditions were as described by Gerdes et al. (2005).PCR products (ca.500 bp) were analyzed by DGGE based on the protocol of Muyzer et al. (1993) using a gradient chamber.Clearly visible bands of the DGGE pattern were excised from the gel and reamplified by PCR (Gerdes et al., 2005) and sequenced.The 16S rRNA gene sequences were then assigned to the new higher-order taxonomy proposed in Bergey's taxonomic  outline of the "Prokaryotes" by the "Ribosomal Database Project (RDP) Classifier" (Wang et al., 2007).The sequences were further compared with those deposited in GenBank using the BLAST algorithm.
The presence of CH 4 -oxidizing bacteria in the communities was screened by the two functional primer sets "pmoA" and "mxaF", targeting the genes encoding subunits of the particulate methane monooxygenase (pMMO) and the methanol dehydrogenase (MDH), respectively.Both enzymes are key enzymes for methanotrophs (e.g., McDonald et al., 2008).However, the mxaF gene is also present in almost all other methylotrophic bacteria.The primer sets and amplification conditions employed in the gene-specific PCR reaction are described in Holmes et al. (1995) and McDonald and Murrell (1997), respectively.
The MW extended from the surface to ∼ 60 m water depth; this is the depth range where the thermocline is located and temperature decreased by ∼ 4 • C (Fig. 4a).In the MW, CH 4 concentrations increased from ∼ 20 nM at the surface to 72.3 nM at 60 m water depth (Fig. 5a).All concentrations were high and oversaturated with respect to the atmospheric equilibrium concentration of 3.3-3.9nM (at the relevant T /S conditions, Wiesenburg and Guinasso, 1979).Similar to concentrations, microbial oxidation rates determined with 3 Hand 14 C tracer increased with depth to 2.3 nM day −1 and 0.77 nM day −1 , respectively, at 60 m (Fig. 5b and c).In the MW, rates measured with 14 C tracer (440-540 nM 14 C-CH 4 added) were consistently lower than those determined with 3 H tracer (1-2 nM 3 H-CH 4 added).δ 13 C-CH 4 values in this water mass ranged between −43.5 and −53.6 ‰ (Fig. 5d).
In the ArW, (60 to ∼ 100 m water depth) oxygen concentrations decreased with depth from 350 to 320 µM (Fig. 4c) and CH 4 concentrations from 42 to 6.5 nM (Fig. 5a).Both MO x rates determined with 3 H and 14 C tracer showed a maximum at ∼ 80 m in this water mass (Fig. 5b and c).The stable carbon isotopic signature of CH 4 showed a strong shift from −46 to about −32 ‰ at ∼ 80 m (Fig. 5d).
The BSW (> 100 m water depth) was characterized by oxygen concentrations below 320 µM (Fig. 4c).CH 4 concentrations decreased slightly with depth, but were stable below 120 m (8-9 nM, Fig. 5a).MO x rates determined with 3 H-labeled CH 4 showed a similar trend to the CH 4 concentrations.However, while 3 H-MO x rates (approximately in situ rates) were low, rates determined with 14 C-labeled CH 4 (rates determined at elevated CH 4 concentrations) were comparably high, with a maximum of 1.9 nM day −1 at ∼ 100 m water depth (Fig. 5b and c).The carbon isotopic signature of the CH 4 decreased steadily from its maximum of −30 ‰ at 100 m to −39 ‰ in the lowermost sample (136 m, Fig. 5d).

DGGE of 16S rDNA
Similar to the biogeochemical results, the MW and BSW at the studied stations (stations. 2, 5, 12, 18, 19) showed distinct DGGE banding patterns (Fig. 6, Table 2) indicating that surface MW and deep waters were populated by different microbial communities.
The MW samples showed strong DGGE bands that we could assign to eukaryotic-chloroplast DNA (#3, #4) and to Alphaproteobacteria of the genera Phaeobacter and Sulfitobacter (#7, #8).The affiliation to the genus Phaeobacter was, however, relatively weak (0.51 confidence value, Table 2), indicating a possibly yet-undescribed bacteria type.Additional bands (#5, #9, and #11) could be assigned to the genera Fluviicola within the phylum Bacteroidetes, Haliea within the phylum Proteobacteria, and llumatobacter within the phylum Actinobacteria.Although we could measure CH 4 oxidation in the surface waters, the DGGE based on the 16S rRNA gene did not reveal known methanotrophs.
In contrast to the diverse MW community, all deep-water samples (station 12, 127 m; station 2, 138 m; station 18, 136 m) showed a quite low diversity with only two strong (# 6 and #7) and one weaker DGGE band (#10) (Fig. 6).Band #7 was also common in the upper water masses, while band #6 was only found in the BSW samples.This band could be affiliated with Methylosphaera, which is a known type I MO x bacterium (Bowman et al., 1997).However, the confidence value of 0.38 was relatively low (Table 2).The deepwater-specific band #10 could be assigned to the sulphate reducer Desulfobacca, also with a relatively low confidence level (0.19, Table 2).

Molecular marker genes of methanotrophs
The pmoA gene that encodes the alpha subunit of the particulate methane monooxygenase is a molecular marker gene Table 2. Classification of partial 16S rRNA sequences to bacterial taxa performed with the RDP Classifier (Wang et al., 2007).The confidence value (0-1) for assignment at the level of class and genus is given in parentheses. of methanotrophs (McDonald et al., 2008).In contrast to the 16S rRNA-based survey, the pmoA-based PCR yielded amplicons within all surface-and deep-water samples (Fig. 7) attesting to the ubiquitous presence of MO x communities in waters of Storfjorden.However, besides the expected product of 530 bp, all deep-water samples showed a further, longer amplicon.Nevertheless, none of the > 530 bp amplicons could be affiliated with known pmoA genes, which suggests either novel pmoA types or unspecific PCR products.A similar distinction of the water masses was also apparent from the distribution of the mxaF gene (Fig. 7) that encodes the enzyme methanol dehydrogenase, which catalyzes the second step in CH 4 oxidation.The mxaF gene was also found in all samples, but deep-water samples showed several additional, weak, and shorter mxaF bands.

Water column stratification and methane sources
Storfjorden water column mixing regimes were the subject of several previous publications (e.g., Haarpaintner et al., 2001;Skogseth et al., 2005;Fer, 2006).The fjord is a deep semi-enclosed basin in the Svalbard Archipelago characterized by brine formation as a result of ice formation in latent heat polynyas during wintertime (Haarpaintner et al., 2001).Descending brines induce strong vertical mixing (Jardon et al., 2011) and turbulence at the sediment-water interface.However, accumulation of brine in bottom waters also leads to a stabilization of the water column, which is further enhanced through a ∼ 60 m thick surface layer of relatively saltdepleted MW in summertime (Fig. 4).The residence time of the deep BSW is 90-246 days, relatively long compared to the 51-141 days of the surface water (Geyer et al., 2009), so that ongoing oxygen consumption leads to the comparably low oxygen levels that were detected previously (Anderson et al., 1988) and in this study.CH 4 concentrations in Storfjorden are generally high, at 6-72 nM.These elevated concentrations originate from microbial methanogenesis in the sediments and enhanced transport from sediments into the water column as a result of the descending brines inducing turbulence at the sedimentwater interface (Damm et al., 2007).However, CH 4 concentrations indicate a second CH 4 source at 40-60 m water depth (Fig. 5a).Here O 2 concentrations were high as well (Fig. 4c),  2. MW and BSW samples are framed by a light-blue and dark-blue rectangle, respectively.Dendrogram derived from UP-GMA cluster analysis with the similarity coefficient of Jaccard.
possibly indicating a maximum of phytoplankton.The second CH 4 source could thus be related to water column in situ production by yet unidentified microorganisms utilizing the phytoplankton metabolite DMSP as a carbon source (Damm et al., 2008) or microbially produced MPn as a potential phosphorus source (Karl et al., 2008;Metcalf et al., 2012).However, further investigations are required to determine the role of these compounds as potential CH 4 precursors at Storfjorden.
While a significant fraction of the CH 4 is consumed (see Sect. 4.2), Storfjorden is apparently a CH 4 source to the atmosphere (Damm et al., 2007), as indicated by CH 4 concentrations of up to 30 nM in the well-mixed surface layer.These concentrations are highly supersaturated with respect to the atmospheric equilibrium (3.3-3.9 nM, Wiesenburg and Guinasso, 1979).

Vertical distribution of methane oxidation
Maximum MO x rates in the water column of Storfjorden were ∼ 2 nM day −1 , and are thus very similar to MO x rates measured in the Santa Barbara Basin (Pack et al., 2011;Mau et al., 2012) and the Black Sea (Reeburgh et al., 1991), which are both well-known areas of large methane input (Fig. 1).However, our rates are 3 orders of magnitude lower compared to the measurements conducted after the Deep Water Horizon accident in the Gulf of Mexico, during which catastrophic amounts of hydrocarbons were released into the water column, triggering a rapid response in MO x activity (Valentine et al., 2010;Kessler et al., 2011).Other MO x rate measurements were conducted in Bristol Bay and the southeast Bering Sea (Griffith et al., 1982), the Cariaco Basin (Ward et al., 1987), Saanich Inlet (Ward, 1989 andWard andKilpatrick, 1990), in Southern California Bight (Ward, 1992;Pack et al, 2011;Heintz et al., 2012;Ward and Kilpatrick, 1993), in the Eel River basin (Valentine et al., 2001), the Gulf of Mexico (Kelley, 2003), and at hydrothermal vents at Juan de Fuca Ridge (de Angelis et al., 1991Angelis et al., , 1993).It appears that most of these MO x rates fall into the range between 0.001 and 10 nM day −1 and that MO x activity is elevated in ocean environments with high CH 4 concentrations.However, MO x activity in the surface- (Ward, 1992;Kelley, 2003, this study) and open ocean (Sauter et al., 2006) seems to be rather low.
In addition to these larger-scale trends, our results from Storfjorden indicate distinct and vertically separated regimes of CH 4 oxidation.This distinction is apparent when comparing MO x rates in deep BSW (> 100 m) with surface MW (< 60 m).The ArW (60-100 m) appears to be an intermediate between the two regimes (Fig. 5b and c).We incubated parallel samples with 3 H-and 14 C-labeled CH 4 .While absolute rate measurements with 3 H-CH 4 were moderate in ArW and BSW, rates with 14 C-CH 4 were elevated in these water masses.We suggest that this is related to the different amounts of CH 4 that were added as a result of 3 H-CH 4 compared to 14 C-CH 4 application.While in incubations with 3 H-CH 4 , the final CH 4 concentrations were only raised by < 2 nM, 14 C-CH 4 amendments lead to a CH 4 increase of ∼ 450 nM.It is therefore reasonable to assume that the activity of the deep-water MO x community was stimulated as a result of elevated CH 4 concentrations (Pack et al., 2011).This is most likely related to enzyme kinetics (Ward and Kilpatrick, 1990;Bender and Conrad, 1993;Smith et al., 1997), which can be described with the Michaelis-Menten model (Button, 1985;Johnson and Goody, 2011, translation of the 1913 Michaelis-Menten paper).The Michaelis-Menten relation shows that enzyme activity, expressed by the reaction rate, increases hyperbolically with substrate concentration but levels off once the enzymatic machinery involved in the metabolic pathway is saturated with substrate.Similar relations were found between cell-or community-specific rates and substrate concentrations (Button, 2010, and references therein).For a stable community, a maximum rate thus exists, which may only increase as a result of elevated enzyme concentrations (e.g., population growth) and/or optimization of cytoarchitectural components relevant for substrate metabolism (e.g., transporter system).We could show that substrate turnover rates were linear over the incubation time period of three days (Fig. 3).At least for our incubation experiments, it thus seems unlikely that the CH 4 amendments induced an increase in enzyme concentration or optimization of other parameters relevant for substrate metabolism.
The derivative of the Michaelis-Menton function for low substrate concentrations (CH 4 concentrations lower than half saturation constant, k m ) yields the first-order rate constant (k ), which, multiplied with the substrate concentration, defines the actual rate (r ox ; see Eq. 1).Consequently, under substrate-limiting conditions, k values are high but decrease if substrate concentrations approach enzyme saturation level.This relationship is depicted in Fig. 8.In MW (the fjord's surface layers), k values were high during 3 H-CH 4 incubations, i.e., without substantial CH 4 amendments, but the addition of CH 4 in the 14 C-CH 4 incubations led to a substantial decrease (5-10 fold) in k , which suggests enzyme saturation.On the other hand, the deep-water community in ArW and particularly in BSW appeared to operate at CH 4 concentrations below saturation because the addition of CH 4 through 14 C-CH 4 tracer application led to an increase in k compared to parallel incubations with 3 H-CH 4 .
It should be noted that rate measurements with two different substrate concentrations (we added 2 nM and 450 nM) are not useful for a kinetic study yielding k m and/or the max- imum reaction velocity (v max ).Nevertheless, adaptation to different substrate concentrations, as indicated by our results, can (at least in parts) be explained by the Michaelis-Menten concept.The MO x community inhabiting the deeper water masses, which are periodically exposed to high CH 4 concentrations, seems to be adapted to high CH 4 substrate levels.In contrast, the community inhabiting surface waters with generally low CH 4 concentrations seems not to be adapted to metabolize additional CH 4 , which could be related to a low enzymatic k m .Low k m values, which were found to range between 10 nM and 10 µM (e.g., Bender andConrad, 1993, Baani andLiesack, 2008), could indeed explain this phenomenon.However, the available k m values from the literature were determined from organisms found in terrestrial or freshwater environments or from cultured bacteria, which most likely do not represent the rather unknown marine communities.Furthermore, the enzymatic k m may not be the same as the apparent cell/community-based k m (see discussion by Button, 2010).The question remains as to why the MO x communities in deep and surface waters were apparently adapted to high and low CH 4 concentrations, respectively.Relatively low CH 4 concentrations in deeper water layers seem to be a regular feature of Storfjorden, at least during summertime (Damm et al., 2008).However, during wintertime, CH 4 export from the sediments is enhanced, leading to elevated CH 4 concentration of up to ∼ 60 nM with a δ 13 C-signature of −40 to −50 ‰ in deeper water layers of Storfjorden (Damm et al., 2007).It thus appears reasonable to assume that the deepwater community is adapted to comparably high wintertime CH 4 concentrations.In summertime, ongoing CH 4 oxidation leads to decreasing CH 4 concentrations and an increase in 13 C in the residual CH 4 (Fig. 5).In contrast, surface CH 4 seems only to increase strongly during summer (to ∼ 50 nM), potentially as a result of CH 4 production in the oxic water column.However, we cannot explain why surface-water methanotrophs appear not to have adapted to the high summertime CH 4 concentrations or possibly lack the ability to adapt.

Microbial community
Similar to the MO x regimes, the diversity of the bacterial assemblage was different when comparing surface MW to the deep BSW.Our DGGE analyses indicate a higher microbial diversity in surface-compared to the deep water (Fig. 6, Table 2).Nevertheless, we only found one band in the surface water (#9) and one band in the deep water (#6) that might be related to CH 4 oxidizers.Band #9 could be affiliated to the genus Haliea of which novel isolates were found to oxidize ethylene and to possess genes similar to those encoding pMMO (Suzuki et al., 2012).Band #6 could be assigned to a known aerobic methanotroph of the genus Methylosphaera (yet with a relatively low confidence value of 0.38).Species of the order Methylosphaera were previously found in Antarctic marine salinity, meromictic lakes (Bowman et al., 1997).The different patterns of MO x -related bands in surface-and deep water thus indicate the presence of different MO x communities in these water masses.
Similar to the 16S gene rRNA-based survey, the pmoA and mxaF gene analyses indicated differences between surfaceand deep-water masses (Fig. 7).All samples analyzed generated amplicons with both primer sets; the deep-water samples, however, showed an additional, longer pmoA band and several weak, shorter mxaF bands suggesting the presence of different, and possibly novel, pmoA-and mxaF-related gene sequences or unspecific PCR products.Similar suggestions (novel sequences or unspecific PCR products) were given in other studies using the pmoA primers, which we used, resulting in a limited coverage of the methanotrophic diversity and yielding PCR products that could also not be assigned to known pmoA types (Bourne et al., 2001;Henckel et al., 2000, Tavormina et al., 2008).Nevertheless, in addition to the 16S banding pattern and MO x rates at elevated CH 4 concentrations, this further indicates that surface-and deep waters comprise different MO x communities.
The question remains as to what are the driving mechanisms for the development of the MO x communities in the different water masses.Here, we suggest that resuspension of sediments as a result of turbulent mixing during wintertime could have inoculated the deeper water masses with sediment microbes including benthic MO x communities.These are of-ten distinct from planktonic communities (Bowman et al., 1997;He et al., 2012;Tavormina et al., 2008), and probably adapted to higher CH 4 concentrations.This scenario would also explain the presence of the sulfate reducer Desulfobacca in the oxic deep waters.Sulfate-reducing bacteria are usually adapted to an anoxic environment (e.g., sediments), and may tolerate only low O 2 levels, yet resting cells of sulfate reducers were also found in fully oxygenated waters (Hastings and Emerson, 1988;Teske et al., 1996).In contrast to the deep waters, the comparably short residence time of surface waters and the rather rapid exchange with the Barents Sea argues for a planktonic source of MO x communities in this water mass.

Conclusions
Incubation experiments with different substrate levels (here we used different tracers) are useful to identify distinct methanotrophic responses in different water masses.With respect to the natural CH 4 concentrations of our study site (< 80 nM, Fig. 5), we used 3 H-CH 4 amendments, which altered absolute CH 4 concentrations by 1-2 nM.The 14 C-CH 4 amendments, in contrast, increased CH 4 concentrations by 440-540 nM.The 3 H-CH 4 ex situ tracer incubations thus yield rates that may be similar to in situ rates. 14C-CH 4 ex situ rates were within the same order of magnitude as those determined with 3 H-CH 4 .Yet, because of the high CH 4 concentration increase during our incubations with 14 C-CH 4 , these ex situ rates rather provide an indication for the rate potential of the MO x community.Rate measurements typically provide a temporal snapshot, which is difficult to upscale, particularly in environments with spatiotemporally varying CH 4 fluxes.Knowledge on the MO x rate at elevated CH 4 concentrations, on the other hand, provides a means to estimate the response in MO x activity in relation to changing CH 4 fluxes.

Fig. 2 .
Fig. 2. Locations of the stations in Storfjorden.Stations are marked by white dots and station numbers.The coastal current is indicated by a dashed blue arrow.Contours are drawn every 100 m until 1000 m water depth Fig. 3. Time series incubation of water samples collected at (A) station 2 (77 • 5.226 N and 19 • 29.694 E) at 135 m water depth, at (B) station 18 (78 • 15.288 N and 19 • 29.070 E) at 50 m water depth, and at (C) a reference station (RS, 70 • 35.91 N and 10 • 51.59 E) at 101 m water depth. 14C-CH 4 and 3 H-CH 4 results are shown as black and gray circles, respectively.

Fig. 4 .
Fig. 4. Depth profiles of temperature (A), salinity (B), and oxygen concentrations (C), as well as a temperature-salinity graph with temperature-salinity ranges of the dominant water masses in Storfjorden (D).Stations 5 and 8 are less than 20 m deep and appear as dots in the temperature-salinity graph.

Fig. 6 .
Fig. 6.DGGE profile of 16S rRNA gene fragments of MW and BSW samples from different stations in Storfjorden.Numbers on the left-hand side of the lanes indicate excised and successfully sequenced DGGE bands, whose phylogenetic assignment is listed in Table2.MW and BSW samples are framed by a light-blue and dark-blue rectangle, respectively.Dendrogram derived from UP-GMA cluster analysis with the similarity coefficient of Jaccard.

Fig. 7 .
Fig. 7. Agarose-electrophoresis gels of PCR products of the pmoA and mxaF genes obtained from surface MW and deep BSW water samples of different stations in Storfjorden.MW and BSW samples are framed by a light-blue and dark-blue rectangle, respectively.

Fig. 8 .
Fig. 8.Comparison of rate constants (k ) determined with 3 H-CH 4and 14 C-CH 4 tracers.Straight line shows the 1:1 fit; that is, if k derived from both tracers were equal.Samples from surface melt water fall above this line (k determined by 3 H-CH 4 is higher than k derived by 14 C-CH 4 ) and samples from the deep brine-enriched shelf water mainly fall below this line (k determined by 14 C-CH 4 is higher than k derived by 3 H-CH 4 ).

Table 1 .
Locations of stations and performed analyses.