In this study, a multidisciplinary team of scientists aboard the research vessel (R/V) Electra participated on a cruise conducted in the Western Gulf of Finland during June 2017 and September 2018. Field sampling consisted of collecting sediment slices (top 0–2 cm) from eight stations along a coastal gradient (0–4 km from land, 10–45 m water depth) in the Storfjärden bay, Tvärminne, Finland (n = 3 per station; Fig. 1 and Table 1). The stations were divided into four offshore sites (stations 5, 7, 10, 13; 36–45 m deep) and four inshore sites (stations 11, 12, 15, 16; 10–28 m deep). Sediment was extracted for metagenomic DNA and RNA sequencing to identify CH4-related microbial populations and active metabolism. During the 2018 sampling campaign this data was coupled to i) real-time measurements of CH4 in the 0.5–1.0 m water surface (first presented in Humborg C, Geibel MC, Sun X, McCrackin M, Mörth C-M, Stranne C, Jakobsson M, Gustafsson B, Sokolov A, Norkko A and Norkko J ), and ii) acoustic data of the seafloor and bottom water to identify CH4 seeps. Furthermore, CTD casts in the study area were used to collect water column profiles of light (PAR) and oxygen concentrations.
Water column parameters. During the sampling campaigns salinity ranged 6.5–7.0 ppt, with higher salinity in the bottom water. In the surface water salinity did not differ between the inshore and offshore stations during sampling. Temperature ranged 3.4–8.9°C (2017 early June) and 6.02–15.82°C (2018 late September), with higher temperatures in the surface water. CTD profiles of the water column from twelve during the 2018 sampling campaign locations inside the study area showed that photosynthetically active radiation (PAR) light reached a water depth of 28 m at sites < 30 m water depth, and the inshore stations would therefore have been illuminated (Additional File 1: Fig. S1). The bottom water was oxic with oxygen concentrations between 7.6–8.6 ml/l in the study area as measure during 2018 (Additional File 1: Fig. S1).
CH4 concentrations in the surface water. CH4 concentrations in the surface water were higher in the inshore shallow stations close to land (23.4–40.6 nM, n = 4 stations) compared to the offshore areas (16.2–23.4 nM; n = 3 stations, Fig. 1 and Table 2).
Alpha and beta diversity. In the 0–2 cm sediment surface prokaryotic alpha diversity ranged between 5.3–6.2 (Shannon’s H, 16S rRNA gene DNA 2018 data) with no difference between inshore and offshore stations (Kruskal-Wallis test). NMDs of Bray-Curtis beta diversity showed that the offshore stations 7, 10, and 13 clustered differently compared to inshore stations 11, 12, 16, and 15 (PERMANOVA 9999 permutations, F = 8.5, P = 0.0001 for the whole model; Additional File 1: Fig. S2). A full list of the prokaryotic classifications and sequence counts is available in Additional File 2: Data S1.
Methanotrophic bacteria in inshore and offshore sediments. Gammaproteobacteria had the highest relative abundance of the prokaryotic community in the 0–2 cm sediment surface when comparing phyla and Proteobacteria classes between stations (Fig. 2A). In the metagenome 16 rRNA gene data the relative abundance of Gammaproteobacteria ranged between 30–51 % for all stations (Fig. 2A). The relative abundance of the Type I CH4 oxidizing bacteria Methylococcales (dominated by the family Methylomonaceae, with majority of sequences (up to 79.7 %) aligning to Methyloprofundus; Additional File 2: Data S1) was significantly higher in stations located offshore when compared to the more shallow inshore stations (16S rRNA gene DNA data, Kruskal-Wallis tests, df = 1, H = 19.7, P = 0.000009 (2017 DNA data), df = 1, H = 18.1, P = 0.000021 (2018 DNA); Fig. 2B). In the DNA data Methylococcales had a relative abundance up to 4.98 % of the whole microbial community (Additional File 2: Data S1). Furthermore, taxonomic classification of all metagenomic sequences against the NCBI RefSeq genome database (Additional File 2: Data S2) showed that Methylococcales was attributed a higher relative proportion of reads in the offshore stations when compared to the inshore stations (Kruskal-Wallis tests, 2017 DNA data, H = 17.4, P = 0.000030; 2018 DNA data, df = 1, H = 14.8, P = 0.000116; Fig. 3A). Similarly, based on mapping RNA reads against metagenome assembled pmoAB genes (Additional File 2: Data S3), more RNA reads were mapped in the offshore stations compared to the inshore stations (Kruskal-Wallis test, df = 1, H = 14.8, P = 0.000121; Fig. 3B and Table 2). To test that the higher relative abundance of methanotrophs were not an effect of sequencing depth the count data of Methylococcales 16S rRNA gene sequences, all other 16S rRNA gene sequences, and total library size was tested with DESeq2. The results showed that in the offshore stations the Methylococcales 16S rRNA gene counts had a log2 fold change of 10.5 (for year 2017) and 10.9 (year 2018) compared to the inshore stations. In contrast, the counts for other 16S rRNA gene sequences and the total library size had both a log2 fold change of 0.1 for both years (Additional File 1: Fig. S3).
CH4 concentrations measured in the 0.5–1.0 m water surface showed a negative relationship with the 16S rRNA gene relative abundance of Methylococcales in the sediment, with lower CH4 concentrations in the offshore stations (Fig. 4A) where Methylococcales activity was higher. Furthermore, CH4 concentration measured in the water column during the sampling campaign 2018 correlated negatively with the relative abundance of Methylococcales for the same-year DNA data (rho = -0.768, P = 0.000047, n = 21). That Methylococcales was associated with offshore sites further away from the coast was also indicated by positive correlations with water depth (2018 DNA data, rho = 0.818, P = 0.000006; 2017 DNA data, rho = 0.740, P = 0.000036). In addition to a higher relative abundance of Methylococcales, RNA transcripts attributed to the protein family AMO/pMMO also correlated negatively with measured concentrations of CH4 (based on classifying all paired-end merged RNA sequences, rho = -0.760, P = 0.000064, n = 21; See Additional File 2: Data S4 for all protein classifications; Fig. 4). RNA transcripts attributed to AMO/pMMO were also significantly higher in the offshore stations (FDR < 0.05, test between all stations individually; Fig. 4B and Table 1), while functional genes in the metagenome attributed to AMO/pMMO were available at all stations with little difference in CPM values (counts per million sequences) (1429–1652 CPM; Table 2), showing that the potential to oxidize CH4 was available at all sites. The soluble form of MMO was not detected in the RNA transcript dataset (Additional File 2: Data S4). AMO/pMMO sequences were classified against the UniProtKB/Swiss-Prot database to separate AMO and pMMO sequences. The results showed a large difference in pMMO CPM values between the offshore stations (8742 ± 2342 CPM, one standard deviation shown) compared to the shallower inshore stations (58 ± 175 CPM, Kruskal-Wallis test, df = 1, H = 14.7, P = 0.000124; Fig. 5). These pMMO sequences were affiliated with the reference species Methylococcus capsulatus in the UniProtKB/Swiss-Prot database. That methanotrophy was higher in the offshore stations compared to inshore was also supported with quantitative reverse transcription PCR (RT-qPCR) based on RNA samples and degenerate pmoA primers (Additional File 1: Data S3). The offshore stations had 0.003257 ± 0.001067 NRQ (normalized relative quantification) compared to the inshore stations with 0.000066 ± 0.000094 NRQ (One-Way ANOVA, F(1,19) = 108.0, P = 0.000000003; Fig. 6 and Table 2; results from non-degenerate primers are shown in Additional File 2: Fig. S4; NRQ is based on pmoA transcripts numbers normalized for 16S rRNA).
Because light has been indicated to inhibit CH4 oxidation we also analysed the amount of RNA transcripts attributed to proteins in the Gene Ontology (GO) category Photosynthesis (Table 2 and Additional File 2: Data S4). Photosynthesis proteins in the sediment surface had a negative correlation with both the relative abundance of Methylococcales (2018 data,rho = -0.615, P = 0.003), and AMO/pMMO enzymes (rho = -0.760, P = 0.00006, n = 21). Photosynthesis proteins were also negatively correlated with water depth (rho = ‑0.676, P = 0.0008, n = 21; Additional File 1: Fig. S5). Moreover, 18S rRNA data of diatoms with a higher relative abundance of benthic genera such as Amphora and Nitzschia in the inshore stations provides further indication that these stations were euphotic (Additional File 1: Fig. S6). This in accordance with the PAR data that indicated the inshore areas to be illuminated while offshore bottom zones were in darkness. A full list of proteins can be found in Additional File 2: Data S4 (RNA) and Additional File 2: Data S6 (DNA).
Our results clearly show that Methylococcales were the major methanotroph active in our sediments, while other methanotrophic bacteria were found to be absent in the DNA dataset. This included the Type II methanotrophic families Methylocystaceae and Beijerinckiaceae (belonging to Alphaproteobacteria), the Verrucomicrobia family Methylacidiphilaceae, and the NC10 phylum known to contain anaerobic methanotrophic bacteria (Additional File 2: Data S1).
Pore water ammonium concentrations. NH4+ analyses showed that the pore water concentration of NH4+ was higher in the offshore stations (308 ± 59 µM) compared to the inshore stations (196 ± 49 µM; One-Way ANOVA, F6,14 = 33.1, P = 0.00000017, with Tukey post hoc test between stations, P < 0.01; Table 2). The DNA dataset for both years 2017 and 2018 showed that aerobic ammonia oxidizing bacteria and archaea were present at all sites and had a significantly higher relative abundance in the inshore sites (inshore: 2017, 6.9 ± 1.5; 2018, 5.4 ± 1.5 %, compared to offshore: 2017, 4.2 ± 1.0; 2018, 3.4 ± 1.3 %) (Kruskal-Wallis tests, 2017, H = 14.9, P = 0.00010; 2018, H = 6.5, P = 0.011; Additional File 1: Fig. S7). However, AMO sequences showed no differences in CPM values between the offshore and inshore stations (196 ± 90 CPM, Kruskal-Wallis test, H = 0.4, P = 0.83; Fig. 5), suggesting that pore water NH4+ concentrations did not explain the difference in methanotrophic activity between inshore and offshore areas.
Methanotrophic and methanogenic archaea in the sediment. Archaea had a 0–2 % relative abundance in the 0–2 cm sediment (Fig. 2), and methanotrophic archaea (ANME) were not detected in the DNA data for both years (Additional File 2: Data S1). Similarly, methanogenic archaea (e.g. Methanobacteria, Methanomicrobia) were not detected in the DNA samples (Additional File 2: Data S1). These findings suggest that higher CH4 concentrations in the surface water in the study area was unrelated to methanogenic activity in the 0–2 cm sediment surface (e.g. if surface sediment had been oxygen deficient).
Methane escape from the sediment. Acoustic data of the seafloor and bottom water was collected in the study area during 2018, and CH4 seeps were defined as either trains of bubbles or bubble plumes (Fig. 7). The results showed that the prevalence of CH4 seeps in sediment surface was greater in shallow areas compared to deeper areas (Fig. 8A), further suggesting that CH4 availability did not explain the lower relative abundance of methanotrophs in the inshore stations. Moreover, the amount of CH4 seeps km-1 was negatively correlated with water depth (Pearson correlation, r = -0.83, P < 0.000001, n = 52; Fig. 8B).