Development of bacterial and archaeal communities in erupted subsurface muds at the Håkon Mosby mud volcano

The microbial oxidation of methane controls the emission of the greenhouse gas methane from the ocean floor. However, some seabed structures such as mud volcanoes have leaky microbial methane filters and can be important sources of methane. We investigated the disturbance and recovery of a methanotrophic mud volcano microbiome (Håkon Mosby mud volcano, 1250 m water depth), to assess time scales of community succession and function in the natural deep-sea environment. We analyzed 10 surface and 5 subsurface sediment samples across HMMV mud flows from most recently discharged subsurface muds towards old consolidated muds as well as one reference site (REF) located approximately 0.5 km outside of the HMMV. Surface samples were obtained in 2003, 2009 and 2010. The surface of the new mud flows at the geographical center was sampled in 2009 and 2010. Around 100 m south of the center, we sampled more consolidated aged muds in 2003 and 2010. Old mud flows were sampled around 300 m southeast and 100 m north of the geographical center in 2003, 2009 and 2010. Surface sediment samples (0-20 cm) were recovered either by TV-guided Multicorer or by push cores using the remotely operated vehicle Quest (Marum, University Bremen). Subsurface sediments of all zones (>2 m below sea floor) were obtained in 2003 by gravity corer. After recovery, sediments were immediately subsampled in a refrigerated container (0°C) and further processed for biogeochemical analyses or preserved at -20°C for later DNA analyses. Our study show that freshly erupted muds hosted heterotrophic deep subsurface communities, which were replaced by surface communities within a few years of exposure. Aerobic methanotrophy was established at the top surface layer within less than a year, followed by anaerobic methanotrophy, sulfate reduction and finally thiotrophy. Our data indicate that it takes decades in cold environments before efficient methanotrophic communities establish to control methane emission. The observed succession provides insights to the response time of complex deep-sea communities to seafloor disturbances.

Metadata und statistic analysis of archael and bacterial sequences originating from sediments of the Håkon Mosby mud volcano (Z1-Z3, all habitats)

DNA extraction was carried out as described on the MICROBIS project pages (http://icomm.mbl.edu/microbis ) using a commercially available extraction kit. We amplified the hypervariable regions V4-V6 of archaeal and bacterial 16S rRNA genes using PCR and several sets of forward and reverse primers (http://vamps.mbl.edu/resources/primers.php). Massively parallel tag sequencing of the PCR products was carried out on a 454 Life Sciences GS FLX sequencer at Marine Biological Laboratory, Woods Hole, MA, following the same experimental conditions for all samples. Sequence reads were submitted to a rigorous quality control procedure based on mothur v30 (doi:10.1128/AEM.01541-09) including denoising of the flow grams using an algorithm based on PyroNoise (doi:10.1038/nmeth.1361), removal of PCR errors and a chimera check using uchime (doi:10.1093/bioinformatics/btr381). The reads were taxonomically assigned according to the SILVA taxonomy (SSURef v119, 07-2014; doi:10.1093/nar/gks1219) implemented in mothur and clustered at 98% ribosomal RNA gene V4-V6 sequence identity. V4-V6 amplicon sequence abundance tables were standardized to account for unequal sampling effort using 1000 (Archaea) and 2300 (Bacteria) randomly chosen sequences without replacement using mothur and then used to calculate inverse Simpson diversity indices and Chao1 richness (doi:10.2307/4615964). Bray-Curtis dissimilarities (doi:10.2307/1942268) between all samples were calculated and used for 2-dimensional non metric multidimensional scaling (NMDS) ordinations with 20 random starts (doi:10.1007/BF02289694). Stress values below 0.2 indicated that the multidimensional dataset was well represented by the 2D ordination. NMDS ordinations were compared and tested using Procrustes correlation analysis (doi:10.1007/BF02291478). All analyses were carried out with the R statistical environment and the packages vegan (available at: http://cran.r-project.org/package=vegan), labdsv (available at: http://cran.r-project.org/package=labdsv), as well as with custom R scripts. Operational taxonomic units at 98% sequence identity (OTU0.03) that occurred only once in the whole dataset were termed absolute single sequence OTUs (SSOabs; doi:10.1038/ismej.2011.132). OTU0.03 sequences that occurred only once in at least one sample, but may occur more often in other samples were termed relative single sequence OTUs (SSOrel). SSOrel are particularly interesting for community ecology, since they comprise rare organisms that might become abundant when conditions change.16S rRNA amplicons and metagenomic reads have been stored in the sequence read archive under SRA project accession number SRP042162.