ORIGINAL RESEARCH ARTICLE published: 16 May 2014 doi: 10.3389/fmicb.2014.00231 Recoding of the stop codon UGA to glycine by a BD1-5/SN-2 bacterium and niche partitioning between Alpha- and Gammaproteobacteria in a tidal sediment microbial community naturally selected in a laboratory chemostat Anna Hanke 1, Emmo Hamann 1, Ritin Sharma 2,3, Jeanine S. Geelhoed 1, Theresa Hargesheimer 1, Beate Kraft 1, Volker Meyer 1, Sabine Lenk 1, Harald Osmers 1, Rong Wu 4, Kofi Makinwa 4, Robert L. Hettich 2,3, Jillian F. Banfield 5, Halina E. Tegetmeyer 1,6 and Marc Strous 1,6,7* 1 Microbial Fitness Group, Max Planck Institute for Marine Microbiology, Bremen, Germany 2 UT-ORNL Graduate School of Genome Science and Technology, University of Tennessee, Knoxville, TN, USA 3 Chemical Science Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA 4 Faculty of Electrical Engineering, Mathematics and Computer Science, Delft University of Technology, Delft, Netherlands 5 Department of Earth and Planetary Science, Department of Environmental Science, Policy, and Management, University of California, Berkeley, CA, USA 6 Center for Biotechnology, University of Bielefeld, Bielefeld, Germany 7 Department of Geoscience, University of Calgary, Calgary, AB, Canada Edited by: Sandy coastal sediments are global hotspots for microbial mineralization of organic Martin G. Klotz, University of North matter and denitrification. These sediments are characterized by advective porewater flow, Carolina at Charlotte, USA tidal cycling and an active and complex microbial community. Metagenomic sequencing Reviewed by: of microbial communities sampled from such sediments showed that potential sulfur Stefan M. Sievert, Woods Hole Oceanographic Institution, USA oxidizing Gammaproteobacteria and members of the enigmatic BD1-5/SN-2 candidate Steven Hallam, University of British phylum were abundant in situ (>10% and ∼2% respectively). By mimicking the Columbia, Canada dynamic oxic/anoxic environmental conditions of the sediment in a laboratory chemostat, *Correspondence: a simplified microbial community was selected from the more complex inoculum. Marc Strous, Department of Metagenomics, proteomics and fluorescence in situ hybridization showed that this Geoscience, University of Calgary, Campus Drive 7200 NW, Calgary, simplified community contained both a potential sulfur oxidizing Gammaproteobacteria T2N 1N4, Alberta, Canada (at 24 ± 2% abundance) and a member of the BD1-5/SN-2 candidate phylum (at e-mail: [email protected] 7 ± 6% abundance). Despite the abundant supply of organic substrates to the chemostat, proteomic analysis suggested that the selected gammaproteobacterium grew partially autotrophically and performed hydrogen/formate oxidation. The enrichment of a member of the BD1-5/SN-2 candidate phylum enabled, for the first time, direct microscopic observation by fluorescent in situ hybridization and the experimental validation of the previously predicted translation of the stop codon UGA into glycine. Keywords: continuous culture, enrichment, chemostat, Roseobacter, Maritimibacter, stop codon INTRODUCTION (Llobet-Brossa et al., 1998; Eilers et al., 2001). Potential sulfur oxi- The Wadden Sea along the northern European coast is the largest dizing Gammaproteobacteria make up for an important part (ca. tidal system worldwide and has been a UNESCO world heritage 39.6%) of all Gammaproteobacteria found on Janssand. Within area since 2009. It receives nutrients, mainly in the form of nitrate, Gammaproteobacteria these bacteria form a large phylogenetic phosphate, and silicate from a large catchment area in northern clade that also includes the ubiquitous marine SUP05 cluster Europe, stimulating growth of algae and other microorganisms in (Walsh et al., 2009), many bacterial symbionts of marine inverte- the surface water. Tidal pumping of this water through the per- brates (e.g., Kleiner et al., 2012) and some cultivated species such meable sediments of tidal flats leads to the continuous removal as Sedimenticola selenatireducens (Narasingarao and Häggblom, of nutrients by highly active indigenous benthic microbes. For 2006), Dechloromarinus chlorophilus (Coates and Achenbach, example, the measured in situ denitrification rates are very high, 2004)andThiotaurens thiomutagens (Flood, 2010). Many of up to 60 µmol m−2 h−1 (Kieskamp et al., 1991). these bacteria are known to be facultative aerobic chemolithoau- In the past decades the biogeochemistry and the micro- totrophs that can use inorganic electron donors such as sulfide bial diversity of the Wadden Sea tidal flat have been stud- and hydrogen to perform aerobic respiration and denitrifica- ied intensively. The microbial community in the upper oxic tion (Walsh et al., 2009; Stewart, 2011). The Flavobacteria on tidal flat sediments was shown to be dominated by populations the other hand are known to be involved in the degradation of of Gammaproteobacteria (Lenk et al., 2011)andFlavobacteria macromolecules such as polysaccharides (Kirchman, 2002). www.frontiersin.org May2014|Volume5|Article231| 1 Hanke et al. A simple marine microbial community To understand the overall function of individual bacterial sulfur oxidizing Gammaproteobacteria and a member of the species in complex natural microbial communities microbiology enigmatic bacterial BD1-5/SN-2 clade which lacks cultivated rep- has traditionally depended on the isolation of target microbes resentatives and was previously predicted to translate the stop in pure culture. Because such isolation is often unsuccessful, codon UGA into glycine. Proteomic analysis of the simplified metagenomic genome reconstruction (Tyson et al., 2004; Lo et al., microbial community enabled the experimental validation of this 2007; Wrighton et al., 2012; Castelle et al., 2013) and single cell prediction. genomics (Kalisky and Quake, 2011; Blainey, 2013; Rinke et al., 2013) have been used to unravel the metabolism of key commu- MATERIALS AND METHODS nity members without their isolation. Both approaches can yield SAMPLING AND INOCULATION near-complete (Tyson et al., 2004; Baker et al., 2010; Wrighton Sediment samples were taken from an intertidal flat in the central et al., 2012) and, in some cases, complete genomes (e.g., Albertsen German Wadden Sea known as “Janssand” located south of the et al., 2013; Castelle et al., 2013; Rinke et al., 2013). However, Eastern Friesian Island Spiekeroog (N: 053◦ 44 151/E: 007◦ 41 sequencing methods alone lack the ability to probe metabolic 945). For direct metagenomic sequencing one sample was col- function and provide limited insight into microbial interactions. lected on October 24 2009 (0–5 cm depth), and three samples on Both with conventional pure cultures and single cell meth- March 23 2010 (0–2 cm depth). To assess small scale differences, ods, the target microorganism is, literally, isolated from its natural the three March samples were taken form locations approxi- context and it is often difficult to understand the ecological niche mately 3 m apart. To analyze sequencing bias, two of the three of the isolated microorganism. For this, it would be ideal to study March samples were pooled, and the pool was divided into two the organism in the context of its natural habitat. Because of the samples, named Mar10/1a and Mar10/1b, to be sequenced in sep- dynamics and complexity of natural communities this is generally arate sequencing runs. Sediment samples were stored at −20◦C not straightforward. in 50 ml falcon tubes. For inoculation sampling was conducted Engineering of a simplified natural ecosystem in the labora- at low tide in May 2011 (15◦C, overcast). Using a trowel, sedi- tory can enable the study of uncultivated bacteria in the context ment from the upper 2 cm layer of Janssand was shovelled into of a simplified microbial community (e.g., Kartal and Strous, a cooling box and kept on ice during its 5 h transport to the 2008; Belnap et al., 2009). Such a simplified microbial commu- lab. It was diluted with artificially prepared seawater (33.4 g/l nity will assemble spontaneously from a more complex inoculum salt, Red Sea, Somerset, UK; in Milli-Q water) in a ratio of 1:1 by natural selection under the conditions applied. In combina- and mixed for 3 min using a drilling machine (PSB 850-2 RE, tion with metagenomic genome reconstruction, proteomic and 850 Watt, Bosch, Stuttgart, Germany) with a cement mixer. The metabolomic approaches this allows for characterization of over- turbid, sand free supernatant (the “cell extract”) was filled in all community metabolic function, interactions, and provides a 800 ml portions into 1 l transfusion bottles (Ochs Glasgerätebau, route to unravel the contribution of each of the individual mem- Bovenden/Lenglern, Germany) and pH was set to 8.1–8.4 with bers. Once an engineered system is available, key environmental 1 M NaOH solution. The cell extract was made anoxic by alter- factorsthatdefinetheecologicalnicheofselectedpopulations nately applying vacuum to 0.3 bar and argon to 1.2 bar, 3 times can be identified by manipulating the applied conditions. Such each. Each bottle was supplemented with NaNO3 stock solution manipulation is not possible when natural ecosystems are studied to reach a final concentration of 0.1 mM serving as electron accep- directly. tor. 50 mg/l cycloheximide (AppliChem, Darmstadt, Germany) To achieve a significant substrate turnover at low, near in situ was added and the cell extract was incubated at 4◦Cfor24htokill substrate concentrations, habitat engineering depends on contin- predatory eukaryotic