FEMS Microbiology Ecology, 91, 2015 doi: 10.1093/femsec/fiu022 Advance Access Publication Date: 8 December 2014 Research Article RESEARCH ARTICLE Ice formation and growth shape bacterial community structure in Baltic Sea drift ice Eeva Eronen-Rasimus1,2,∗ , Christina Lyra3 , Janne-Markus Rintala2,4 , Klaus Jürgens5 , Vilma Ikonen1,† and Hermanni Kaartokallio1 1 Marine Research Centre, Finnish Environment Institute (SYKE), Erik Palménin aukio 1, PO Box 140, Helsinki 00251, Finland, 2 Tvärminne Zoological Station, University of Helsinki, J.A. Palménin tie 260, FI-10900 Hanko, Finland, 3 Department of Food and Environmental Sciences, PO Box 56, Viikinkaari 9, FI-00014 University of Helsinki, Finland, 4 Department of Environmental Sciences, PO Box 65, Viikinkaari 1, FI-00014 University of Helsinki, Finland and 5 Leibniz Institute for Baltic Sea Research Biological Oceanography, Seestr. 15, 18119 Rostock, Germany ∗ Corresponding author: Marine Research Centre, Finnish Environment Institute (SYKE), Erik Palménin aukio 1, PO Box 140, Helsinki 00251, Finland. Tel. +358 (0) 40-182-31-72; Fax. +358 (0)9-54-90-22-90; E-mail [email protected] † Present address: Roal Oy, Tykkimäentie 15, FI-05200 Rajamäki, Finland. One sentence summary: Clear changes in the bacterial communities were observed in Baltic Sea drift-ice samples from open water to thick ice, leading to bacterial communities similar to those of polar sea ice. Editor: Patricia Sobecky ABSTRACT Drift ice, open water and under-ice water bacterial communities covering several developmental stages from open water to thick ice were studied in the northern Baltic Sea. The bacterial communities were assessed with 16S rRNA gene terminal-restriction fragment length polymorphism and cloning, together with bacterial abundance and production measurements. In the early stages, open water and pancake ice were dominated by Alphaproteobacteria and Actinobacteria, which are common bacterial groups in Baltic Sea wintertime surface waters. The pancake ice bacterial communities were similar to the open-water communities, suggesting that the parent water determines the sea-ice bacterial community in the early stages of sea-ice formation. In consolidated young and thick ice, the bacterial communities were significantly different from water bacterial communities as well as from each other, indicating community development in Baltic Sea drift ice along with ice-type changes. The thick ice was dominated by typical sea-ice genera from classes Flavobacteria and Gammaproteobacteria, similar to those in polar sea-ice bacterial communities. Since the thick ice bacterial community was remarkably different from that of the parent seawater, results indicate that thick ice bacterial communities were recruited from the rarer members of the seawater bacterial community. Key words: bacteria; sea ice; 16S rRNA; T-RFLP; cloning INTRODUCTION The Baltic Sea, a brackish subpolar semi-enclosed sea, is annually covered with seasonal ice to varying extent (Granskog, Kaartokallio and Kuosa 2010). Coastal areas are covered by fast ice, meaning that ice forms and melts almost at the same place, whereas in offshore areas, depending on water depth and ice thickness, ice is moved by winds and currents and is called drift ice. Dynamic ice growth conditions, such as rafting and Received: 4 July 2014; Accepted: 30 November 2014 C FEMS 2014. All rights reserved. For permissions, please e-mail: [email protected] 1 2 FEMS Microbiology Ecology, 2015, Vol. 91, No. 1 ridging, are common in Baltic drift ice, and the pancake ice cycle is important in turbulent waters similar to that in the Antarctic (Granskog et al., 2006). Pancake ice floes are round-shaped ice discs that form through accretion of ice crystals that move by waves and ocean swell. By bumping and grinding against one another, they can grow from centimetres to decimetres in size and eventually congeal into larger floes and finally to a continuous solid ice sheet (Petrich and Eicken 2010). Ice floes can slide one on the other, resulting in rafted ice. As a consequence of ridging, ice floes can pile against one another, forming pack ice fields. Baltic Sea ice is structurally similar to polar sea ice, despite its brackish nature (Granskog et al., 2010). However, brine channels are smaller and the brine salinity is lower compared to the polar counterparts (Granskog et al., 2006). When sea ice forms, inorganic and organic components dissolved in seawater are concentrated into brine that moves within channels and pores in the sea-ice matrix (Petrich and Eicken 2010). Sea ice serves as a habitat for diverse assemblages of auto- and heterotrophic organisms, including sea-ice bacteria (Thomas and Dieckmann 2002; Mock and Thomas 2005; Arrigo, Mock and Lizotte 2010; Caron and Gast 2010; Deming 2010). Bacterial growth in sea ice is regulated by multiple simultaneously acting environmental factors, such as salinity, temperature, nutrients (Pomeroy and Wiebe 2001; Kuosa and Kaartokallio 2006), as well as food web interactions, such as protistan grazing (Kaartokallio 2004; Riedel, Michel and Gosselin 2007). Similar to the open ocean, bacteria in sea ice participate in decomposing of particulate organic matter and recycling and production of dissolved organic matter (DOM). Bacteria are entrained from seawater into the forming seaice matrix and apparently suffer a transient reduction in activity, but the activity is recovered after the ice has consolidated (Grossmann and Gleitz 1993; Grossmann 1994; Grossmann and Dieckmann 1994; Helmke and Weyland 1995; Kaartokallio 2004; Kaartokallio et al., 2008; Deming 2010). In consolidated sea ice, psychrophilic bacteria replace psychrotrophic bacteria along with maturation of the ice cover (Helmke and Weyland 1995; Bowman et al., 1997a). Succession and development of bacterial communities in sea ice have been associated with high substrate concentrations and abundant attachment sites, such as particles and algae (Helmke and Weyland 1995; Bowman et al., 1997a; Junge et al., 2002; Junge, Eicken and Deming 2004; EronenRasimus et al., 2014). However, detailed description of bacterial succession at the community level in sea ice has been done only twice. In temperate Baltic Sea fast ice, the bacterial community changed throughout the winter (Kaartokallio et al., 2008) whereas in extremely cold Arctic drift ice, no change in bacterial community composition was observed during the lowproductive winter period (Collins, Rocap and Deming 2010). Despite the knowledge on succession in thick sea ice, bacterial communities in the various developmental stages of drift ice have not been studied. Bacterial communities in the Baltic Sea consist of both freshwater (e.g. Betaproteobacteria, Actinobacteria and Verrucomicrobia) and marine (Alphaproteobacteria, Bacteroidetes) bacteria adapted to ambient brackish water conditions (Hagström, Pinhassi and Zwiefel 2000; Pinhassi and Hagström 2000; Sipura et al., 2005; Riemann et al., 2008; Andersson, Riemann and Bertilsson 2010; Herlemann et al., 2011; Laas et al., 2014). Baltic Sea wintertime surface-water bacterial communities are dominated by Alphaproteobacteria and Actinobacteria (Laas et al., 2014). In general, the same bacterial groups are found in both water and ice (Kaartokallio, Laamanen and Sivonen 2005, 2008; Andersson et al., 2010; Herlemann et al., 2011; Laas et al., 2014). The phylum Bacteroidetes (e.g. class Flavobacteria) and classes Alpha- and Gammaproteobacteria dominate sea-ice bacterial communities, but less abundant groups such as Betaproteobacteria, Gram-positive bacteria and Archaea have also been found (Bowman et al., 1997a; Bowman, Brown and Nichols 1997b; Junge et al., 1998, 2002, 2004; Staley and Gosink 1999; Petri and Imhoff 2001; Brown and Bowman 2001; Brinkmeyer et al., 2003, 2004; Kaartokallio et al., 2005, 2008; Mock and Thomas 2005; Collins et al., 2010; Bowman et al., 2012). The same bacterial classes are described both in Baltic and polar sea ice, but their relative abundance varies, depending on sampling location and time. The phylum Bacteroidetes in sea ice is possibly favoured by decreasing temperature (Junge et al., 2004) and is apparently abundant in both Arctic and Baltic Sea ice during the cold winter months (Junge et al., 2004; Kaartokallio et al., 2008). In contrast, Gammaproteobacteria are most abundant in spring/summer sea ice and thus presumably favour high availability of organic substrates originating from ice algal growth (Junge et al., 2002; Brinkmeyer et al., 2003; Kaartokallio et al., 2008; Deming 2010; Bowman et al., 2012). Betaproteobacteria occur more often in Baltic and Arctic sea ice than in Antarctic ice, suggesting terrestrial influence on sea-ice bacterial communities (Brinkmeyer et al., 2003, 2004; Kaartokallio et al., 2008). Our aim was to investigate the bacterial community structure in various developmental stages of drift ice from open water to consolidated sea ice. The study was conducted aboard the RV Maria S. Merian in the Gulf of Bothnia, Baltic Sea in February– March 2006. The bacterial communities were studied with 16S rRNA gene terminal-restriction fragment length polymorphism (T-RFLP) and cloning together with bacterial abundance and production measurements. MATERIALS AND METHODS Study site and sampling The samples were collected from the Bothnian Sea and Bothnian Bay, Baltic Sea, aboard the RV Maria S. Merian in February–March 2006 (Fig. 1). The overall ice conditions during the cruise and sea-ice structures are described in detail in Rintala, Piiparinen and Uusikivi (2010). Open-water samples were retrieved with a conductivity–temperature–depth /Rosette sampler from three different depths: 2–5, 70–80 and 100–230 m at stations 22–24. Sea ice was sampled from stations 25–29 and 31 from a range of ice types including new ice, pancake ice, consolidated ice with two different developmental stages categorized as young ice and thick ice, as well as pack ice (see Rintala et al., 2010). Ice at stations 26 and 29 were rafted (see Rintala et al., 2010). For bacterial community analysis, new ice and pancake ice were collected from one station, young ice from two different stations and thick ice from three different stations (Table 1). Two replicates were collected from each station, except for pancake ice which had three replicates. New ice, young ice, thick ice and three out of five pack ice samples were drilled with a US Army Cold Regions Research and Engineering Laboratory-type power core auger (9 cm Ø, Kovacs Enterprise, Roseburg, OR, USA). The ice cores were cut with a handsaw into three sections so that the surface and bottom ice were approximately 10 cm in length and the middle part of the ice varied from 13 to 30 cm, except at station 25, where the ice was cut in only two pieces (Table 1). The remaining two pack ice cores were cut with a handsaw from the surface layer of the pack ice field. Three approximately 0.5-m-diameter pancake ice floes were collected from the ice edge, using a metal basket. Eronen-Rasimus et al. 3 Bacterial abundance St. 26 65 SWEDEN To determine the bacterial abundance, 20 ml of melted ice or water was fixed with 25% electron microscopy grade glutaraldehyde (1% final concentration; Sigma-Aldrich, St. Louis, MO, USA) and stored in a cool, dark place prior to analysis. Sample of 5 ml was filtered onto 0.2-μm-polycarbonate filters (Millipore, Darmstadt, Germany) and stained with 0.02% acridine orange (Hobbie, Daley and Jasper 1977). The total bacterial numbers were counted using a Leitz Aristoplan (Leica Microsystems, Bensheim, Germany) epifluorescence microscope equipped with Leitz a I3 filter and PL Fluotar 100X 12.5/20X oil immersion objective giving a total magnification of ×1000. At least 200 recorded cells in a minimum of 20 random fields were counted using a New Porton E11 counting grid (Graticules Ltd, Edenbridge, Kent, UK). St. 25 St. 28 St. 25 St. 31 64 Latitude ° N St. 29 63 Pancake ice St. 24 FINLAND 62 St. 23 61 St. 22 60 THE BALTIC SEA 59 58 18 20 22 24 26 Longtitude ° E Bacterial production 28 30 Figure 1. Map of the northern Baltic Sea, showing sampling stations during the cruise aboard RV Maria S. Merian during 28 February–9 March 2006. Open water was collected from stations 22 to 24, new ice and pack ice from station 28, young ice from station 25 and 31 and thick ice from station 26, 27 and 29. Pancake ice had no station number and is thus marked directly on the map. The ice samples were allowed to thaw in autoclaved plastic containers in darkness at 4◦ C (direct melting, Helmke and Weyland 1995; Kaartokallio 2004) and filtered subsequently after becoming fully melted. Under-ice water samples were collected directly from the drill holes into autoclaved plastic bottles. Slush was removed from the hole before sampling and the bottles were kept in darkness until further processing. For the DNA extractions, 470–1500 (mode 500) ml of melted sea ice and 1000 ml of under-ice water and open water were filtered through sterile 0.22 μm membrane filters (Ø 47 mm; Whatman, GE Healthcare, Little Chalfont, Kent, UK) and frozen immediately at −80◦ C. Bacterial net biomass production was measured using the 3 Hlabelled thymidine incorporation method (Fuhrman and Azam 1982). The samples were processed based on the description in Kaartokallio (2004): a piece of each ice section was crushed with a spike. Approximately, 7.5 g of crushed ice was weighed in a scintillation vial, after which 2 ml of 0.2μm-filtered seawater was added to better simulate the brine pocket salinity and ensure even distribution of labelled substrate. All the work was done in a climate room at 4◦ C. Three aliquots and a formaldehyde-killed absorption blank were amended with [methyl-3 H]-thymidine (PerkinElmer, specific activity 20 Ci mmol L−1 ) to a final concentration of 18 nM. The samples were incubated in the dark at 4◦ C for 17–22 h. The incubations were stopped by adding formaldehyde (final concentration of 5%). The standard cold trichloroacetic acid extraction method (Fuhrman and Azam 1980, 1982) was used for removal of unincorporated 3 H-thymidine. The samples were filtered onto 0.2 μm polycarbonate filters (Millipore, Darmstadt, Germany), after which the filters were placed into scintillation vials, scintillation cocktail (Irgasafe plus PerkinElmer Life and Analytical Sciences, Groningen, Netherlands) added and measured with a scintillation counter (PerkinElmer Life and Analytical Sciences, Groningen, Netherlands). The empirical conversion factor of 2 × 1018 cells per mol incorporated thymidine L−1 was used to estimate bacterial cell production (Smits and Riemann 1988). Environmental parameters DNA extraction and PCR amplification Inorganic nutrients including phosphate (PO4 3− ), ammonium (NH4 + ), nitrite (NO2 − ), nitrate (NO3 − ), as well as temperature, salinity and brine volume were measured (Table 1) and the structures of the sea ice cores were determined from one replicate from each station, except for that in pancake ice, nutrients were measured from all three replicates. The methods employed are described in Rintala et al. (2010). Elevated PO4 3− , NH4 + and NO3 − concentrations, high bulk salinity and brine volume percentage were detected in the snow ice (0–16 cm) layer at station 26, indicating incorporation of nutrients from atmospheric deposition into the ice via snow ice formation (see Rintala et al., 2010). In addition to the environmental parameters described above, also chlorophyll a (chl-a), algal, protozoan and metazoan biomasses were determined from the same sample set. The methods and results were reported and discussed earlier in Rintala et al. (2010). R DNA Isolation DNA was extracted from filters with a PowerSoil Kit (Mobio Laboratories, Inc., Carlsbad, CA, USA) as described by Eronen-Rasimus et al. (2014). The extracted DNA was used as a template (∼50 ng) in triplicate PCR reactions to amplify the 16S rRNA genes for the T-RFLP and clone libraries. For the T-RFLP, the primers FAM27f, FAM-GAGTTTGATCMTGGCTCAG with 6carboxyfluorescein label [Sait et al., 2003; high-performance liquid chromatography (HPLC)-purified, Oligomer Oy, Helsinki, Finland], and unlabelled 1406r, ACGGGCGGTGTGTRC, (Lane et al., 1985; HPLC-purified; Oligomer Oy) were used, whereas for the clone libraries both primers were unlabelled. The PCR reactions and purifications were performed, as described by Sinkko et al. (2011) with the following exceptions: three parallel PCR reactions from each sample were done in a reaction volume of 25 μl and DyNazymeTM EXT DNA polymerase was used (Finnzymes, Thermo Fisher Scientific, Vantaa, Finland). 0–0.10 0.10–0.40a 0.40–0.63 NA 0–0.12 0.12–0.25 0.25–0.46 NA 0–0.10 0.10–0.40 0.40–0.59 NA 0–0.12 0.12–0.27a 0.27–0.34 NA 0–0.12 0.12–0.25 NA 0.12 0.12 0.19 0.08 0.20 0.17 0.17 0.14 0.55 0.33 0.11 0.07 0.28 0.15 0.89 0.10 0.07 0.11 0.06 0.16c 0.14 0.07 – – – 8.89 3.76 2.21 1.38 8.86 32.14 9.60 0.78 9.28 11.48 1.45 3.28 8.94 7.62 4.14 9.31 1.77c 11.28 1.73 – – – – – – – – – NO3 − 0.04 0.03 0.03 0.03 – – – 0.02 – – – 0.02 0.03 0.01 0.01 0.03 – – 0.09 0.08c 0.01 0.03 – – – – – – – – – NO2 − NA = not applicable a indicates samples from which 16S rRNA gene clone libraries were derived b mean value from 2 replicates c mean value from 3 replicates under-ice water 29 thick ice under-ice water 27 thick ice under-ice water 26 thick ice under-ice watera 31 young ice under-ice water 25 young ice pancake icea 0–0.09 – – – 2a 70 210 24 open water 28 new ice under-ice water – – – 3 80 100 23 open water – – – PO4 3− 5 70 230 Depth from surface (m) 22 open water Station number and sample type 2.02 0.89 1.06 1.47 1.48 1.38 1.44 1.25 5.71 2.51 0.65 0.06 2.87 1.08 3.03 0.12 0.08 0.07 0.63 1.06c 1.17 – – – – – – – – – – NH4 + 6.50 7.58 4.05 31.76 6.29 3.30 3.32 29.95 8.14 4.97 2.89 24.78 7.05 1.94 5.58 33.52 – 3.62 32.21 4.48c 15.63 33.13 – – – – – – – – – SiO4 − 0.70 0.90 0.45 2.40 1.10 0.35 0.45 2.20 1.30 0.77 0.40 2.10 0.30 0.50 0.60 2.50 0.60 0.38 3.45 – 1.60 3.50 5.42 5.57 6.39 5.46 6.04 6.44 5.39 6.45 6.99 Salinity 0.33 0.64 1.39 – 1.18 0.94 0.97 – 1.12 0.58 0.98 – 0.52 0.98 2.02 – −5.45 −3.81 −2.00 – −4.80 −2.67 −2.30 – −6.80 −4.47 −1.32 – 0.83 3.48 – – 1.17 – NA NA NA −4.35 −3.07 −1.60 – −3.50 −0.53 – – −3.50 – 0.31 1.54 3.49 NA NA NA NA NA NA −0.14 2.61 4.29 0.57 2.84 3.47 Brine volume (%) Temperature (◦ C) 4.80 3.40 2.81 7.36 2.32 2.28 2.64 7.36 2.59 3.05 2.44 6.60 3.27 4.50 3.87 7.84 1.87 3.38 9.59 4.61c 6.28b – – – – – – – – – Bacterial abundance (x108 cells L−1 ) 0.30 2.74 2.58 – 4.97 4.76 1.14 – 2.22 6.80 8.43 – 0.27 0.21 0.72 – – – – – 0.44 – – – – – – – – – Tdr incorporation (x107 cells L−1 h−1 ) Table 1. Dissolved inorganic nutrient concentrations (μM), bulk salinities, temperatures, total bacterial abundance and bacterial cell production measured by thymidine (Tdr) incorporation from northern Baltic Sea water and sea-ice stations. Ice was rafted at stations 26 and 29 (see Rintala et al., 2010). 4 FEMS Microbiology Ecology, 2015, Vol. 91, No. 1 Eronen-Rasimus et al. T-RFLP The bacterial community dynamics and composition were determined with T-RFLP (Liu et al., 1997) and cloning. Restriction enzyme digestions and T-RFLP were performed at the Institute of Biotechnology, University of Helsinki, Finland, as described by Sinkko et al. (2011) with three different restriction enzymes: BsuRI, MspI and RsaI (Fermentas, Thermo Fisher Scientific, Burlington, ON, Canada). The peaks were cleared of background noise and normalized with the statistical methods developed by Abdo et al. (2006), written in Perl and R languages (R Development Core Team 2011). With BsuRI, fragments from 26.5 to 1000 base pairs (bp) and with MspI and RsaI from 49.5 to 1000 bp were included in normalization. Cloning and identification of terminal restriction fragments (T-RFs) Five clone libraries were constructed to identify the T-RFs and to determine the taxonomic classification for bacterial communities in different sample types. Cloning of the 16S rRNA gene sequences, plasmid extractions and sequencing were performed at the Institute of Biotechnology, University of Helsinki, Finland. Approximately, 950 bp from the 5 terminus of each 16S rRNA gene were sequenced and the sequences were corrected manually with the Staden Package 1.6.0 Gap v. 4.10 (Staden, Beal and Bonfield 1998; Staden, Judge and Bonfield 2003). Putative chimeras were checked with the Chimera Uchime algorithm in Mothur v.1.31.2. (Schloss et al., 2009). Taxonomic classification of the 16S rRNA genes was done with the naı̈ve Bayesian Classifier (v. 2.6, RDP training set 9, Wang et al., 2007) by applying an 80% confidence threshold, and the Seqmatch tool (release 11.1, v. 3, default options) of the Ribosomal Database Project (RDP, Cole et al., 2009). Since almost all of the sequence matches were affiliated with unknown bacterial sequences, Seqmatch was also carried out with default options, excluding uncultured sequences. Representative sequences were chosen by (1) lowest rank and (2) highest S ab score. In addition, the National Center for Biotechnology Information nucleotide Basic Local Alignment Search Tool (blastn), megablast, 2.2.29+, (Morgulis et al., 2008) was used to obtain the identity percentage for the identified matches from RDP Seqmatch. The 16S rRNA gene sequences were deposited in the European Molecular Biology Laboratory Nucleotide Sequence Database under accession numbers from LM651929 to LM652337. Shared richness based on the observed OTUs with cutoff level 0.03 was calculated using venn command in Mothur (Schloss et al., 2009). The chloroplast sequences were omitted from the clone library analyses. The T-RFs were identified with in silico-digested 16S rRNA clone libraries using the Restriction Enzyme Database (REBASE 7.11, version 1.20080403) virtual digest program (http://insilico.ehu.es/restriction/main/; Roberts et al., 2010) and verified in vitro. For the in vitro digestions, the 16S rRNA gene PCR amplifications, digestions and analyses from clones were performed as described above. Phylogenetic analysis of the 16S rRNA genes A phylogenetic neighbour-joining (NJ) tree was constructed to visualize the diversity in the clone libraries obtained. Approximately 950-bp-long sequences were aligned with RDPipeline aligner [INFERence of RNA Alignment (INFERNAL) version 1.1rc4, Nawrocki, Kolbe and Eddy 2009] of the RDP (Cole et al., 2009). A sequence from the archaeon Sulfolobus tokodaii (AB022438) was 5 used as an outgroup in the alignment. The bootstrapped (1000) NJ tree was constructed using PHYLogeny Interference Package (Phylip) 3.695 (Felsenstein 2005), with the Jukes–Kantor evolution model. The tree was visualized with the Interactive Tree Of Life online tool (Letunic and Bork 2007). Reference sequences for the phylogenetic tree were chosen based on Seqmatch results as described above. Statistical analysis A priori groups were determined, based on sample types. Generalized discriminant analysis based on distance (Canonical Analysis of Principal coordinates routine, Anderson and Robinson 2003) was performed to test whether the bacterial communities could be discriminated between a priori groups. Discriminant analysis was performed on the Bray–Curtis similarity matrix derived from square-root-transformed T-RF data. Square-root transformation was applied to reduce the contribution of dominant species, because the Bray–Curtis resemblance measure does not scale individual peaks by its total or maximum across all samples and our samples showed large differences between their relative abundances. P-values were calculated with 9999 permutations. For discriminant analysis, Plymouth Routines In Multivariate Ecological Research (PRIMER) v. 6 software (Clarke and Gorley 2006) with the add-on package PERmutational ANOVA/MANOVA+ (PERMANOVA+) (Anderson, Gorley and Clarke 2008) was used. Pairwise comparisons between clone library sequences were done with Mothur LIbrary SHUFFling (LIBSHUFF) (Schloss, Larget and Handelsman 2004) applying Bonferroni correction leading to a 0.005 significance level. RESULTS Bacterial abundance and production Bacterial abundance and production rates are presented in Table 1. The bacterial abundance was one-fourth higher in new ice than in pancake ice implying that bacteria suffered reduction in cell numbers immediately after ice formation. After consolidation of sea ice, the bacterial abundance did not notably change. Bacterial production was negligible in the early stages of ice formation but was clearly higher in thick ice, indicating active bacterial growth and biomass turnover in thick ice. Bacterial community composition and dynamics Based on LIBSHUFF analysis, all clone libraries differed significantly, except that for pancake ice, which did not differ from the water samples (Table 2, Bonferroni corrected P < 0.005). The sequences obtained were defined with the RDP Classifier and Seqmatch tools. The open-water, under-ice water and pancake ice bacterial communities were dominated by classes Alphaproteobacteria (21, 22 and 19% respectively) and Actinobacteria (15, 29 and 28% respectively; Table 3, Fig. 2). Most of the actinobacterial sequences in all three clone libraries had closest sequence match to Ilumatobacter coccineus (identity 95%; AP012057) extracted from Kumagawa River estuary sediment, Japan (Matsumoto et al., 2009) and most of the Alphaproteobacteria to Candidatus Pelagibacter sp. (SAR11, group III) (identity 99%; CP002511) from Arctic Ocean surface water (Oh et al., 2011). The young ice bacterial community was dominated by the class Gammaproteobacteria (53%; Table 3). The 6 FEMS Microbiology Ecology, 2015, Vol. 91, No. 1 Table 2. P-values of pairwise comparisons from LIBSHUFF analysis between five constructed 16S rRNA gene clone libraries from the northern Baltic Sea water and ice samples. The open-water clone library was constructed from the surface water from station 24, underice water from station 31, pancake ice from replicate three, middle part of the young ice from station 31 and middle part of columnar ice from station 29 (Table 1). Significant values applying Bonferroni correction (P < 0.005) are shaded in grey. gammaproteobacterial sequences showed closest sequence matching to Serratia marcescens (identity 99%; GQ889261) from soil and cyanobacteria-associated Acinetobacter johnsonii (identity 99%; DQ911549). In addition, Alphaproteobacteria (15%) and Planctomycetes (14%) were common in young ice (Table 3, Fig. 2). Similar to the water samples, most of the Alphaproteobacteria showed closest sequence matching to Candidatus Pelagibacter sp. (identity 99%; CP002511) (59). Planctomycetes showed closest sequence matching to uncultured Planctomycetaceae sequences (identity 99%; HM856408) from Yellowstone Lake, WY, USA (Clingenpeel et al., 2011). Overall, the bacterial community in thick ice differed from all other bacterial communities at the genus level (Tables 3 and 4). The thick ice bacterial community was dominated by the class Flavobacteria (36%) and most of the sequences showed closest sequence matching to Flavobacterium degerlachei (identity 99%; AJ557886) isolated from the Ace Lake microbial mat, Antarctica (Van Trappen et al., 2004). All flavobacterial sequences were affiliated with polar regions covering Flavobacterium spp. sequences from both Arctic and Antarctic habitats (Brinkmeyer et al., 2003; Van Trappen et al., 2004; Harding et al., 2011; Zhou et al., 2013; Prasad et al., 2014). The second most common classes in thick ice were Gammaproteobacteria (14%) and Betaproteobacteria (14%; Table 3, Fig. 2) with closest sequence matches to Psychromonas ingrahamii (Gammaproteobacteria; identity 99%; CP000510), Shewanella denitrificans (Gammaproteobacteria; identity 99%; AY771743) from the Table 3. A summary of bacterial taxa in 16S rRNA gene clone libraries derived from northern Baltic Sea water and sea-ice samples. Taxonomic classification was done with naı̈ve Bayesian Classifier (v. 2.6, RDP training set 9, Wang et al., 2007) of the RDP (Cole et al., 2009) with an 80% confidence threshold level. Phylum, class and lowest rank (>80%) Proteobacteria: Alphaproteobacteria: Rhodobacter Thalassospira Loktanella unclassified Rhodobacteraceae unclassified Rhodospirillaceae unclassified Sphingomonadaceae unclassified Acetobacteraceae unclassified Alphaproteobacteria Betaproteobacteria: Albidiferax Limnohabitans Methylotenera Georgfuchsia Acidovorax unclassified Comamonadaceae unclassified Methylophilaceae unclassified Oxalobacteraceae unclassified Burkholderiales unclassified Betaproteobacteria Gammaproteobacteria: Psychrobacter Psychromonas Acinetobacter Serratia Thalassolituus Rheinheimera Pseudomonas Haliea Shewanella Legionella unclassified Enterobacteriaceae unclassified Gammaproteobacteria Deltaproteobacteria: unclassified Deltaproteobacteria unclassified Proteobacteria Open water Under-ice water Pancake ice Young ice Thick ice ND ND 1 2 1 1 2 3 ND ND ND ND ND ND ND 16 1 ND ND ND ND ND ND 6 1 ND ND 1 ND 1 ND 12 1 1 ND ND ND ND ND ND ND 1 1 ND ND 1 ND ND ND ND ND ND 1 1 ND ND 1 ND ND 1 ND ND 1 ND ND ND ND ND ND ND ND ND ND ND 1 ND ND 1 ND ND 2 ND ND ND ND ND ND ND 1 2 ND ND ND ND ND ND ND ND ND 1 ND 2 2 ND ND ND ND 3 ND ND ND ND ND 5 ND ND ND ND ND ND ND ND ND ND ND 3 ND ND 35 11 2 ND 2 ND ND ND 1 1 ND 1 ND ND 1 ND ND 1 1 ND ND 1 1 4 ND 4 ND 2 ND 1 ND ND Eronen-Rasimus et al. 7 Table 3. continued. Phylum, class and lowest rank (>80%) Open water Under-ice water Pancake ice Young ice Thick ice 7 ND ND ND ND ND ND 7 1 ND 10 2 ND 1 5 ND 1 3 1 ND ND 2 ND ND 6 ND 1 ND ND ND ND 1 ND ND ND ND ND 1 2 1 ND ND ND ND 1 1 ND ND ND 1 2 ND ND ND ND 10 2 ND 1 ND ND 3 ND ND ND ND 1 1 ND ND 2 10 2 14 4 1 ND 1 ND 1 ND ND ND ND ND ND ND ND 1 ND Chloroflexi: Caldilineae: Caldilinea 1 1 1 ND ND Nitrospira: Nitrospira: Nitrospira ND ND ND ND 1 1 ND ND ND ND ND 1 ND ND 2 ND ND ND ND ND Firmicutes: unclassified Firmicutes 1 ND ND ND ND Cyanobacteria: GpIIa 2 1 ND ND ND ND ND 3 22 70 2 1 1 2 75 ND ND 6 44 80 ND ND 2 13 112 ND ND 4 36 72 Actinobacteria: Actinobacteria: Ilumatobacter Conexibacter Citricoccus unclassified Actinomycetales unclassified Micrococcineae unclassified Corynebacterineae unclassified Acidimicrobineae Bacteroidetes: Flavobacteria: Flavobacterium Polaribacter Wandonia unclassified Flavobacteriales unclassified Cryomorphaceae Sphingobacteria: Algoriphagus unclassified Bacteroidetes Planctomycetes: Phycisphaerae: Phycisphaera Planctomyces: Planctomyces Schlesneria unclassified Planctomycetaceae Lentisphaerae: Lentisphaeria: unclassified Lentisphaeria Verrucomicrobia: Verrucomicrobiae: unclassified Verrucomicrobia unclassified Verrucomicrobiaceae Other: OD1 genera incertae sedis TM7 genera incertae sedis unclassified Bacteria Chloroplast Total number of sequences Arctic and to Albidiferax ferrireducens (Betaproteobacteria; identity 97%; CP000267). A phylogenetic NJ tree was constructed to visualize diversity in different sample types (Fig. 3). In general, the various sample types were intermixed in the tree, but certain genera, such as Flavobacterium and Acinetobacter, also formed their own clades specific for only sea ice. The clone library sequences were also digested in vitro to identify T-RFs in the various sample types (Fig. 4). In general, the same bacterial groups that were identified with cloning were also identified with T-RFLP. Actinobacteria was the most common class in seawater samples, whereas Flavobacteria was the most common class in sea-ice samples. Alphaproteobacteria were more abundant in water than in sea ice. Planctomycetes were 8 FEMS Microbiology Ecology, 2015, Vol. 91, No. 1 100 % 90 % Alphaproteobacteria Betaproteobacteria 80 % Gammaproteobacteria Deltaproteobacteria 70 % Flavobacteria Sphingobacteria 60 % Actinobacteria Phycisphaerae 50 % Planctomycetacia Verrucomicrobiae 40 % Lentisphaeria Caldilineae 30 % Nitrospira OD1 genera incertae sedis 20 % TM7 genera incertae sedis Cyanobacteria 10 % Unclassified 0% Open water Under-ice water Pancake ice Young ice Thick ice Figure 2. Bacterial diversity of 16S rRNA gene clones from the northern Baltic Sea. The open-water clone library was constructed from the surface water from station 24, under-ice water from station 31, pancake ice from replicate three, middle part of the young ice from station 31 and middle part of thick ice from station 29 (Table 1). Taxonomic classification was done at class level with naı̈ve Bayesian Classifier (v. 2.6, RDP training set 9) (Wang et al., 2007) of the RDP (Cole et al., 2009) with an 80% confidence threshold level. Table 4. Total and shared number of bacterial 16S rRNA gene sequence OTUs between different sample types collected from northern Baltic Sea water and drift ice samples. OTUs are defined by 97% sequence similarity. most abundant in the under-ice water and young ice samples. In general, Gammaproteobacteria and Betaproteobacteria were equally distributed among seawater and ice samples. However, the pack ice at station 28 and young ice at station 31 were dominated by Gammaproteobacteria. Generalized discriminant analysis (Anderson and Robinson 2003) yielded results similar to those of the clone libraries, showing that bacterial communities were discriminated by the sample types (Fig. 5). Generalized discriminant analysis showed significant differences between a priori groups with all three restriction enzymes (P = 0.0001 with 9999 permutations). With the BsuRI T-RF data, the first eight axes (choice of m = 8) classified a priori groups 86.9% correctly and explained 78% of the total variability in the T-RF data. Since new ice (station 28) had formed in a refrozen ship lead (see Rintala et al., 2010), the new ice bacterial community resembled the under-ice water community, indicating no changes in community composition shortly after ice formation. DISCUSSION Clear changes in the bacterial communities were observed, both at the class and genus levels in Baltic Sea drift-ice samples from open water to thick ice. The results suggest that the bacterial communities in Baltic Sea drift ice developed over time, leading to bacterial communities similar to those of polar sea ice, although distinctive features in the communities were also found. The bacterial communities apparently went through reduction in activity in the beginning of ice formation as bacterial abundance decreased from new ice to pancake ice and bacterial production was very low. After consolidation of ice, the bacterial production was still low in young ice and increased only in thick ice. Reduction in bacterial activity during the ice formation has been also observed earlier in sea ice (Grossmann and Gleitz 1993; Grossmann 1994; Kaartokallio 2004, 2008). However, in experimental conditions activity was rapidly restored in the presence of abundant substrate (Eronen-Rasimus et al., 2014). The low bacterial production followed the chl-a concentration in ice 9 Eronen-Rasimus et al. atobac ter flum 057 inis Ilum atobac ter coc cineus YM16− EU91 304 9857 Uncul tured LM65 bacte 2224 rium, Ko ngsfjo LM65 rden 2328 LM65 2301 LM 6522 19 ccus sp 01 . V4 43 LM .M Myc O.2 65 0 ob 22 32 ac LM teria 65 20 ce 60 ae LM MI− 65 20 6.3 89 22 88 35 LM 65 65 20 LM 39 65 20 LM 52 65 20 04 LM 65 20 66 19 80 Mic roco 20 LM 65 65 65 46 LM 36 24 23 AJ 65 LM FN 55 LM 6521 LM 6520 76 LM 6521 55 LM 6520 50,LM 6520 90 00 LM6521 96 AB3603 43 Ilum AP012 LM652 082 LM652169 LM65209 3 LM65203 5 86 LM6521 61 LM652075 LM652048 LM652026 LM652223 LM652038 LM652045 LM652007 LM652002 LM652201 LM652172 AY151240 Synechococcus sp. MW72C6 LM652057 LM652217 LM652153 51 LM6520 3 LM65214 7 LM65194 19 LM652105 sp cI 3.a 2− LM 26 5 2 65 LM 6 28 FJ 31 88 42 Can did 20 65 20 23 LM 49 . 22 18 us at Pla hila op nkt lim 21 65 23 65 e Eg H− lM 04 20 65 LM LM 65 65 20 37 65 LM 53 Actinobacteria ii W 56 20 65 LM 07 23 65 LM 54 20 65 LM . M 28 LM 94 tica ne 72 LM 65 20 47 LM 65 20 s to ko da sp LM 65 yc es 160 lobu 65 s gracili ctom 97 an 65 21 ae LM Su lfo llu la 42 llo LM Pire LM 65 2 LM 65 20 LM 74 65 10 22 26 13 66 LM 65 ce 41 nsi sN BR C 5 97 eta yc 02 24 38 19 iku re 1 65 LM om nct Pla m 6521 LM Ye d ra 21 LM65 , ium 652 LM cter ba re ltu isp hae LM AB ,LM 97 cu Ph yc X81 95 0 Pl X8 54 22 65 LM 49 20 65 LM 26 23 65 LM 70 22 65 56 LM 19 65 LM 13 21 65 Un 19 65 08 01 23 38 2167 LM65 2087 LM65 red 64 M 1,L 85 4 20 HM 65 ,LM AP LM652 ultu 75 22 65 LM Thick ice LM6520 Unc Young ice a rospin 38 Nit 87 60 Pancake ice a nolic etha 42 aera 6522 osph LM Olig 11 8581 6520 AB55 LM 08 6523 LM 58 20 65 LM 67 20 65 LM 29 23 65 LM 80 21 M65 nt 4,L ve 14 al 652 rm ,LM he ot 15 dr 20 hy 65 ke ,LM e La 239 on wst 44 HM Under−ice water 652215 121,LM LM652 187 LM652 FR8650 Open water LM652046 100 Color ranges: 20 LM 65 ,LM 05 65 21 45 25 20 LM ium ba red cultu Un 70 1285 HM 54 6521 LM 15 6523 LM , Ta gu ng La lha ke cter 73 6522 LM 68 6519 2198 LM65 2195 LM65 2030 LM65 2175 LM65 LM652 257 111 LM652 276 LM652 LM652 183 LM652 281 AB1250 LM6522 44 sia enweek 62 Ow nge hongko nsis 62 LM6521 LM6523 32 LM65224 9 LM6 52246 6 LM65221 LM65231 2 3 LM65217 LM652256 LM651999 LM652278 LM651930 LM652262 Polarib JQ800145 LM652228 acter sp. KJF9−7 LM651958 LM652252 Bacteroidetes LM652231 LM652119 FR682719 LM652287 R−36233 Flavobacterium sp. LM651986 LM652271 LM652233 LM652263 LM651934 LM652292 AY771756 Flavobacterium degerlachei LM652299 JQ800275 Flavobacterium sp. KJF15−19 LM652237 LM652327 LM652235 LM651931 LM652311 Alphaproteobacteria LM652243 Z2 r tjernbergiae Acinetobacte KC213887 LM651955 LM651944 Gammaproteobacteria S TSA11 LM652303 * LM651970 LM651979 8 LM65224 AJ557886 Flavobact erium 1 LM65225 LM65197 2 98 LM6522 sonii r john etobacte 49 Acin 258 DQ9115 LM652 LM652 Betaproteobacteria * 324 LM6521 6522 LM FR69 2300 LM 0 s S2 en 85 cesc 22 mar 65 LM LM EU 37 23 65 LM 19 23 65 LM 90 22 65 LM 47 22 65 LM 41 22 19 65 as on m 41 32 19 65 LM 7 ii 3 ham gra in LM 03 20 65 6519 64 Rose ov ariu s sp . BB 19 94 74 LM652022 LM652205 LM652043 KC854922 Methylobacillus flagellatus LX158 LM652032 LM651993 LM652102 76 4 LM652190 LM651960 T118 LM65227 2 LM651 976 LM6521 ucens MWH−C ferrired Rhodof erax CP0002 67 ans sp. 6522 68 LM LM 6522 LM65 2029 26 AJ9380 25 Lim nohabit 63 90 LM65 221 2 LM651 939 CC IM LM65 2221 LM65 2148 LM 65 19 96 LM 6522 55 LM 6522 60 . BW DY −24 sp. cter ba lagi Pe 00 22 65 ,LM 06 20 65 as sp 52 us idat Pseu do nd ,LM 95 Ca 2193 70 11 12 6520 LM 22 65 27 5 DQ 46 LM 21 25 00 CP 65 LM 05 77 6522 HF6 80 55 20 65 LM 6523 LM 2110 LM65 2063 LM65 22 LM65 33 6520 LM 17 6520 LM 22 83,L LM6522 LM652309 LM65224 30 LM6522 LM6521 LM652156 LM652186 LM652120 LM652192 CP001672 Methylotenera mobilis JLW8 LM651998 LM652019 LM652269 009 LM652 e Bay sapeak 079 um, Che LM652 bacteri 79 ultured M6521 47 Unc 52084,L EU8013 14,LM6 M6520 mon ar aq uim era im Rh ein he 65 21 84 LM 65 20 34 LM 65 20 59 LM 65 20 LM 68 LM 65 23 65 22 20 ,LM 69 65 31 21 2 Th 50 LM alas 65 solit 19 67 uu LM s ol 65 eivo 19 rans 43 MIL −1 LM EF0 76 75 7 6521 81 2 06 C1 HTC 51 ue 19 iq 65 b 17 ru LM 23 te ac 65 ib 20 LM 22 elag 65 1P LM 19 10 41 AF5 21 8 65 215 LM 97 65 22 ,LM 65 34 LM 93 21 21 65 65 ,LM LM 40 20 65 LM 01 20 65 LM 27 ch LM 65 20 LM 64 65 20 16 LM ro Psy 10 05 00 CP s an 53 43 17 77 AY 53 22 fic itri den 8 65 LM ella −3 65 LM an ew Sh 97 3289 Rh od ob acte Unc r sp ultu . CR 22 red 07 64 JQ −9 ,LM ba 7 71 cter 65 21 21 ium 17 36 LM Unc ,N 65 orth ultu 20 Se 42 red LM a 65 ba 20 cter 13 AF0 ium 69 ,O 96 il co LM 3 ntam Unc 65 21 ultu inat 82 LM ed red 65 se 19 Rick aw LM 40 ater et 65 tsia 2 09 les LM 1 65 19 GQ 61 85 05 LM 45 65 Un 19 cu 84 ltu re d alp hap ro te ob ac te riu m ,B eri ng Se a JX01 LM 65 21 0 tia Serra 30 18 65 LM 61 8892 GQ LM65 23 LM 6521 64 LM 6522 04 36 6523 LM SW imari niv miense 202 LM65 2178 * 2036 LM65 /2−7 s 88 r acte hrob Psyc 65 3425 is AJ31 LM Psy 2805 21915 LM65 2250 2062 LM AY72 LM652 −242 LM65 dovastu m atsu LM651 980 267 LM652 s SW aensi namh ei LMG 57 AB3819 35 Rho LM652 211 323 LM652 acter chrob degerlach LM6519 74 Figure 3. Phylogenetic NJ tree of 16S rRNA gene sequences (∼900 bp) derived from the northern Baltic Sea water and drift-ice samples. Bootstrap values >50% are shown in grey circles. The S. tokodaii (AB022438) sequence was used as an outgroup in the alignment. (see results in Rintala et al., 2010) and therefore recovery after ice consolidation presumably followed the growth and adjacent introduction of new substrate from the ice algal community. The patterns in activity were concomitant with observed community changes in the different sample types. The openwater bacterial community differed significantly from all other bacterial communities, except that of pancake ice, indicating that parent water determines the early stages of sea-ice bacterial communities. In the open water, under-ice water and pancake ice, the classes Alphaproteobacteria and Actinobacteria dominated bacterial communities, which is in accordance with previous study of Baltic Sea wintertime seawater communities (Laas et al., 2014). Actinobacteria are often associated with soil and freshwater environments, but they are also common members of the Baltic Sea bacterioplankton community (Riemann et al., 2008; Holmfeldt et al., 2009; Andersson et al., 2010; Herlemann et al., 2011; Laas et al., 2014). Although few actinobacterial sequences/isolates have been found in both Arctic sea ice and Arctic melt ponds (Brinkmeyer et al., 2003, 2004; Collins et al., 2010; Bowman et al., 2012), as well as Antarctic (Bowman et al., 1997a,b; Junge et al., 1998; Brinkmeyer et al., 2003) and Baltic Sea ice (Kaartokallio et al., 2005), it seems that Actinobacteria are more common in less saline habitats. Most of the Alphaproteobacteria in water and pancake ice samples showed closest sequence identity to the Arctic Candidatus Pelagibacter sp., which is a member of the SAR11 clade (group III) (Giovannoni et al., 1990; Oh et al., 2011). SAR11 is a ubiquitous member of marine bacterial communities (Morris 24 25 31 26 27 28 29 28 25 31 26 27 29 81 112 169 182 187 230 250-252 271 279 290 299 321 448 591-592 612-614 907 new ice pancake ice young ice (surface) young ice (bottom) young ice (surface) young ice (middle) young ice (bottom) thick ice (surface) thick ice (middle) thick ice (bottom) thick ice (surface) thick ice (middle) thick ice (bottom) thick ice (surface) thick ice (middle) thick ice (bottom) packed ice (drilled) packed ice (saw n) Pl LM anc 65 tom 22 yc 18 eta cia (P lan cto m yc es ) Al LM pha Ac pr 6 o 5 22 teo LM tino 2 0 bac 65 bac 22 te ter 86 ria ia (A cti no m yc eta les G ) LMamm 65 apr 22 ot 47 eob ac ter G ia LMamm 65 apr 23 ot 22 eob ac ter ia 28 Fragment length Fl LM avo 65 bact 19 er 30 ia 23 Sample type open w ater 5 m open w ater 70 m open w ater 230 m open w ater 3 m open w ater 80 m open w ater 100 m open w ater 2 m open w ater 70 m open w ater 210 m under-ice w ater Fl LM avo 65 bact 21 eri 19 a 22 F LM lavo 65 bac 19 ter 68 ia Station FEMS Microbiology Ecology, 2015, Vol. 91, No. 1 A LM ctin 65 oba 20 ct 82 eri G a LMamm 65 apr 22 ot Ph 43 eob ac LM ycis ter 65 pha ia 21 er 44 ae (P B hy LM etap cis r 65 ot ph 22 eob ae 69 ac ra) ter ia (B ur kh ol de ria Al les p ) LM ha p 65 rot 20 eo 01 ba cte ria Al G LM pha LM amm pr a 6 65 pr 52 ote 19 ot o 15 b 96 eob 7 ac ter ac ter ia ia (T ha las so Ph lit LM yc uu 65 isph s) 22 ae 13 ra e( Ph yc isp ha era ) 10 Figure 4. Distribution and relative abundance of in vitro (digested) identified bacterial taxa from the northern Baltic Sea water and ice samples. The size of the symbol relates to the relative abundance of 16S rRNA gene T-RF. Taxonomic classification was done with naı̈ve Bayesian rRNA Classifier (v. 2.6, RDP training set 9, Wang et al., 2007) of the RDP (Cole et al., 2009) with an 80% confidence threshold level. The accession numbers indicate the identity of digested clone. Figure 5. Generalized discriminant analysis plot of BsuI-digested 16S rRNA gene sequence T-RFs, showing changes in the bacterial community according to different sample types from open water to thick ice. Only results with BsuI are shown, since it yielded the best taxonomic separation. et al., 2002), including Baltic Sea waters (Herlemann et al., 2011; Laas et al., 2014) and winter ice (Collins et al., 2010). Although SAR11 seems to persist in Arctic thick ice (Collins et al., 2010), in the Baltic Sea they appear to be most competitive in aquatic habitats. After consolidation of sea ice, the bacterial community in young ice was significantly different from other sample types. Young ice was dominated by Gammaproteobacteria. Overwhelming gammaproteobacterial dominance in the initial phases of ice formation was also seen previously in an experimental study where the mesocosms were enriched with DOM (Eronen-Rasimus et al., 2014), indicating that Gammaproteobacteria are capable of opportunistic growth in sea ice. Gammaproteobacterial sequences were affiliated with the genera Acinetobacter and Serratia. Acinetobacter has been previously found in sea ice (Brinkmeyer et al., 2003; Deming 2010) and from the Baltic Sea water column (Hagström et al., 2000; Labrenz, Jost and Jürgens 2007) whereas psychrotolerant Serratia has been found in Baltic Sea sediment (Edlund and Jansson 2006) but not in sea ice. Although the bacterial production in young ice was very low, it is possible that Gammaproteobacteria were temporarily active in the initial phases of the ice formation. As also chl-a concentration was low (see results in Rintala et al., 2010), Gammaproteobacteria would have been presumably utilizing some other than algal-derived substrate. However, it is also possible that these bacteria had been transported into growing sea ice e.g. from sediment or other external source with water currents and thus are not actively growing members of sea-ice bacterial communities. In addition to Gammaproteobacteria, young ice also had high numbers of Planctomycetes sequences, reported previously only once from a melt pond (Brinkmeyer et al., 2004) but not to our knowledge from sea ice. The sequences were affiliated with the class Phycisphaera, a recently accepted class in Planctomycetes (Fukunaga et al., 2009). High numbers of planctomycete sequences and T-RFs indicate their possible ecological importance in Baltic Sea ice. Planctomycetes are often associated with aggregates and algal blooms in marine environments (DeLong, Franks and Alldredge 1993; Pizzetti et al., 2011) and co-occur with Gammaproteobacteria in Baltic Sea waters (Riemann et al., Eronen-Rasimus et al. 2008). Thus, Plactomycetes probably have an ecology similar to that of Gammaproteobacteria, which occur in the sea ice in spring/summer presumably subsequent to an ice algal bloom (Junge et al., 2002; Brinkmeyer et al., 2003; Kaartokallio et al., 2008; Deming 2010; Bowman et al., 2012) and are possibly also permanent members of the Baltic Sea ice communities. Thick ice, which differed from all other ice and water types, was dominated by the class Flavobacteria, with sequences most closely related to Flavobacterium from both the Arctic and Antarctic. The class Flavobacteria, of the phylum Bacteroidetes, is the other major bacterial group inhabiting sea ice along with the Proteobacteria (Bowman et al., 1997a,2012; Staley and Gosink 1999; Brown and Bowman 2001; Junge et al., 2002, 2004; Brinkmeyer et al., 2003; Kaartokallio et al., 2005, 2008). The second most common group in thick ice was Gammaproteobacteria and the sequences were classified to the genera Psychromonas and Shewanella. Both genera are typical sea-ice bacteria (Bowman et al., 1997a,b, 2012; Brown and Bowman 2001; Junge et al., 2002; Brinkmeyer et al., 2003; Deming 2010) and their prevalence together with Flavobacteria indicates a shift in bacterial community composition from seawater communities to sea-ice communities. The betaproteobacterial sequences were as abundant as those of Gammaproteobacteria and thus Betaproteobacteria are likely permanent members of sea-ice bacterial communities, as also previously shown in Baltic and Arctic sea ice (Brinkmeyer et al., 2003; Kaartokallio et al., 2008; Collins et al., 2010) as well as in melt ponds (Brinkmeyer et al., 2004). In general, the water and thick ice communities differed at the genus level, suggesting that most of the community members in the thick ice represented the rarer members of the open-water bacterial communities. The result is an example of the rare biosphere concept (Sogin et al., 2006; Pedros-Alio 2006, 2012) showing how bacterial communities have potential to respond to changing environmental conditions implying ability of minor community members to become dominant when environmental conditions change. Overall, the bacterial communities within each sample type were very similar to each other, despite the long geographical distance between sampling stations, suggesting that selection pressure caused by ice formation and ice maturation is a stronger community-shaping factor than geographical distance. Thus, based on this and previous studies, plausible explanations for the differences observed in bacterial communities from open water to thick ice include formation and maturation of ice and subsequent selection of psychrotolerant and psychrotrophic bacteria, substrate availability by sea-ice algae, as well as resource competition. Measured bacterial production and observed community changes together indicate an active bacterial growth and biomass turnover in thick ice. Substrate availability and maturation of sea ice are known to shape bacterial communities (Helmke and Weyland 1995; Pomeroy and Wiebe 2001; Kaartokallio et al., 2008; Eronen-Rasimus et al., 2014). For example, the role of sea-ice maturation was associated with changes in bacterial community composition when the communities were observed throughout the ice-covered season (January–March) in Baltic Sea fast ice (Kaartokallio et al., 2008). In cold Arctic drift ice, the bacterial communities did not change in contrast to those of the parent seawater, indicating that selection during freezing processes is relatively minor (Collins et al., 2010). However, the authors speculated that copiotrophic spring/summer sea-ice bacteria, such as Gammaproteobacteria, appear in sea ice as a result of competitive outgrowth of oligotrophic bacteria, similar to our conclusions. 11 In conclusion, the parent water bacterial community probably determines the sea-ice community at the early stages of sea-ice development, despite the change in temperature and salinity during sea-ice formation. Along with ice consolidation, the sea-ice bacterial communities changed remarkably together with increasing bacterial production, pointing to a temporal development in Baltic Sea drift-ice bacterial communities. In addition, the results suggest that sea-ice communities are possibly formed by taxa that were present in very low numbers in the parent water community and not readily detected with the methods employed in this study. The consolidated sea-ice bacterial community was dominated by typical sea-ice-associated flavobacterial and gammaproteobacterial genera, pointing to a similar community structure in polar sea ice, despite the brackish nature of Baltic Sea water. However, the classes Betaproteobacteria and Planctomycetes, which seldom occur in polar sea ice, are probably permanent members of the Baltic Sea ice bacterial communities, showing that the Baltic Sea also has unique features of its own, probably influenced by the low salinity in its northern parts. In all, sea-ice formation is a phenomenon where a natural marine environment is fundamentally changed by physical forcing and sea-ice bacteria show substantial adaptive capacity to respond this environmental change. Thus, we suggest that sea-ice bacteria are recruited from the rare members of the parent water, and that sea ice provides an excellent example of the importance of the total existing biodiversity in marine bacterial community formation. ACKNOWLEDGEMENTS This work was supported by the Walter and Andree de Nottbeck Foundation. The authors thank the crew of the RV Maria S. Merian for technical support during the cruise, and Erika Trost and Dr Falk Pollehne for performing nutrient analyses from sea ice samples. Also, we want to thank Prof. Kaarina Sivonen for providing facilities for the molecular work and Riitta Autio for support in various stages during the study. Conflict of interest statement. None declared. REFERENCES Abdo Z, Schuette UME, Bent SJ, et al. Statistical methods for characterizing diversity of microbial communities by analysis of terminal restriction fragment length polymorphisms of 16S rRNA genes. Environ Microbiol 2006;8:929–38. Anderson MJ, Gorley RN, Clarke KR. PERMANOVA+ for PRIMER: Guide to Software and Statistical Methods. Plymouth, UK: Primer-E Ltd, 2008. Anderson MJ, Robinson J. Generalized discriminant analysis based on distances. Aust N Z J Stat 2003;45:301–18. Andersson AF, Riemann L, Bertilsson S. Pyrosequencing reveals contrasting seasonal dynamics of taxa within Baltic Sea bacterioplankton communities. ISME J 2010;4:171–81 Arrigo KR, Mock T, Lizotte MP. Primary producers in sea ice. In: Thomas DN, Dieckmann GS (eds). Sea Ice. 2nd edn, Oxford, UK: Wiley-Blackwell Publishing, 2010, 283–325. Bowman JP, Brown MV, Nichols DS. Biodiversity and ecophysiology of bacteria associated with Antarctic sea ice. Antarct Sci 1997a;9:134–42. Bowman JP, McCammon SA, Brown MV, et al. Diversity and association of psychrophilic bacteria in Antarctic sea ice. Appl Environ Microb 1997b;63:3068–78. 12 FEMS Microbiology Ecology, 2015, Vol. 91, No. 1 Bowman JS, Rasmussen S, Blom N, et al. Microbial community structure of Arctic multiyear sea ice and surface seawater by 454 sequencing of the 16S RNA gene. ISME J 2012;6:11–20. Brinkmeyer R, Glöckner FO, Helmke E, et al. Predominance of beta-proteobacteria in summer melt pools on Arctic pack ice. Limnol Oceanogr 2004;49:1013–21. Brinkmeyer R, Knittel K, Jurgens J, et al. Diversity and structure of bacterial communities in Arctic versus Antarctic pack ice. Appl Environ Microb 2003;69:6610–9. Brown MV, Bowman JP. A molecular phylogenetic survey of sea-ice microbial communities (SIMCO). FEMS Microbiol Ecol 2001;35:267–75. Caron DA, Gast RJ. Heterotrophic protists associated with sea ice. In: Thomas DN, Dieckmann GS (eds). Sea Ice. 2nd edn, Oxford, UK: Wiley-Blackwell Publishing, 2010, 327–56. Clarke KR, Gorley RN. Primer v6: User Manual/Tutorial. Plymouth, UK: Primer-E Ltd, 2006. Clingenpeel S, Macur RE, Kan J, et al. Yellowstone Lake: highenergy geochemistry and rich bacterial diversity. Environ Microbiol 2011;13:2172–85. Cole JR, Wang Q, Cardenas E, et al. The ribosomal database project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res 2009;37:D141–5. Collins RE, Rocap G, Deming JW. Persistence of bacterial and archaeal communities in sea ice through an Arctic winter. Environ Microbiol 2010;12:1828–41. DeLong E, Franks D, Alldredge A. Phylogenetic diversity of aggregate-attached vs free-living marine bacterial assemblages. 1993;38:924–34. Deming JW. Sea ice bacteria and viruses. In: Thomas DN, Dieckmann GS (eds). Sea Ice. 2nd edn, Oxford, UK: Wiley-Blackwell Publishing, 2010, 247–82. Edlund A, Jansson JK. Changes in active bacterial communities before and after dredging of highly polluted Baltic Sea sediments. Appl Environ Microb 2006;72:6800–7. Eronen-Rasimus E, Kaartokallio H, Lyra C, et al. Bacterial community dynamics and activity in relation to dissolved organic matter availability during sea-ice formation in a mesocosm experiment. MicrobiologyOpen 2014;3:139–56. Felsenstein J. PHYLIP (Phylogeny Inference Package) version 36 Distributed by the author Department of Genome Sciences. Seattle, WA: University of Washington, 2005. Fuhrman JA, Azam F. Bacterioplankton secondary production estimates for coastal waters of British Columbia, Antarctica, and California. Appl Environ Microb 1980;39:1085–95. Fuhrman JA, Azam F. Thymidine incorporation as a measure of heterotrophic bacterioplankton production in marine surface waters: evaluation and field results. Mar Biol 1982;66:109–20. Fukunaga Y, Kurashashi M, Sakiyama Y, et al. Phycispaera mikurensis gen. nov., sp. nov, isolated from a marine alga, and proposal of Phycisphaeraceae fam. nov., Phycisphaerales ord nov., and Phycisphaerae classis nov., in the phylum Planctomycetes. J Gen Appl Microbiol 2009;55:267–75. Giovannoni S, Britschgi T, Moyer C, et al. Genetic diversity in Sargasso Sea bacterioplankton. Nature 1990;345:60–3. Granskog MA, Kaartokallio H, Kuosa H. Sea ice in non-polar regions. In: Thomas DN, Dieckmann GS (eds). Sea Ice. 2nd edn, Oxford, UK: Wiley-Blackwell Publishing, 2010, 531–77. Granskog MA, Kaartokallio H, Kuosa H, et al. Sea ice in the Baltic Sea—a review. Estuar Coast Shelf S 2006;70:145–60. Grossmann S. Bacterial activity in sea ice and open water of the Weddel sea, Antarctica: a microautoradiographic study. Microb Ecol 1994;28:1–18. Grossmann S, Dieckmann GS. Bacterial standing stock, activity, and carbon production during formation and growth of sea ice in Weddel sea, Antarctica. Appl Environ Microb 1994;60:2746–53. Grossmann S, Gleitz M. Microbial responses to experimental sea ice formation: implications for the establishment of Antarctic sea-ice communities. J Exp Mar Biol Ecol 1993;173:273–89. Hagström Å, Pinhassi J, Zwiefel U. Biogeographical diversity among marine bacterioplankton. Aquat Microb Ecol 2000;21:231–44. Harding T, Jungblut AD, Lovejoy C, et al. Microbes in high arctic snow and implications for the cold biosphere. Appl Environ Microb 2011;77:3234–43. Helmke E, Weyland H. Bacteria in sea ice and underlying water of the eastern Weddel sea in midwinter. Mar Ecol-Prog Ser 1995;117:269–87. Herlemann D, Labrenz M, Jürgens K, et al. Transition in bacterial communities along the 2000 km salinity gradient of the Baltic Sea. ISME J 2011;4:171–81. Hobbie JE, Daley RJ, Jasper S. Use of Nuclepore filters for counting bacteria by fluorescence microscopy. Appl Environ Microb 1977;33:1225–8. Holmfeldt K, Dziallas C, Titelman J, et al. Diversity and abundance of freshwater actinobacteria along environmental gradients in the brackish northern Baltic Sea. Environ Microbiol 2009;11:2042–54. Junge K, Eicken H, Deming JW. Bacterial activity at −2 to −20◦ C in Arctic wintertime sea ice. Appl Environ Microb 2004;70:550–7. Junge K, Gosink J, Hoppe H-G, et al. Arthrobacter, Brachybacterium and Planococcus isolates identified from antarctic sea ice brine description of Planococcus mcmeekinii, sp. nov. Syst Appl Microbiol 1998;21:306–14. Junge K, Imhoff F, Staley T, et al. Phylogenetic diversity of numerically important Arctic sea-ice bacteria cultured at subzero temperature. Microb Ecol 2002;43:315–28. Kaartokallio H. Food web components, and physical and chemical properties of Baltic Sea ice. Mar Ecol Prog-Ser 2004;273: 49–63. Kaartokallio H, Laamanen M, Sivonen K. Responses of Baltic Sea ice and open-water natural bacterial communities to salinity change. Appl Environ Microb 2005;71:4364–71. Kaartokallio H, Tuomainen J, Kuosa H, et al. Succession of seaice bacterial communities in the Baltic Sea fast ice. Polar Biol 2008;31:783–93. Kuosa H, Kaartokallio H. Experimental evidence on nutrient and substrate limitation of Baltic sea sea-ice algae and bacteria. Hydrobiologia 2006;554:1–10. Laas P, Simm J, Lips I, et al. Spatial variability of winter bacterioplankton community composition in the Gulf of Finland (the Baltic Sea). 2014;129:127–34. Labrenz M, Jost G, Jürgens K. Distribution of abundant prokaryotic organisms in the water column of the central Baltic Sea with an oxic-anoxic interface. Aquat Microb Ecol 2007;46: 177–90. Lane DJ, Pace B, Olsen GJ, et al. Rapid determination of 16S ribosomal RNA sequences for phylogenetic analyses. P Natl Acad Sci USA 1985;82:6955–9. Letunic I, Bork P. Interactive Tree Of Life (iTOL): an online tool for phylogenetic tree display and annotation. Bioinformatics 2007;23:127–8. Liu WT, Marsh TL, Cheng H, et al. Characterization of microbial diversity by determining terminal restriction fragment length polymorphisms of genes encoding 16S rRNA. Appl Environ Microb 1997;63:4516–22. Eronen-Rasimus et al. Matsumoto A, Kasai H, Matsuo Y, et al. Ilumatobacter fluminis gen nov, sp nov, a novel actinobacterium isolated from the sediment of an estuary. J Gen Appl Microbiol 2009;55:201–5. Mock T, Thomas DN. Recent advances in sea-ice microbiology. Environ Microbiol 2005;7:605–19. Morgulis A, Coulouris G, Raytselis Y, et al. Database indexing for production MegaBLAST searches. Bioinformatics 2008;24:1757–64. Morris RM, Rappe MS, Connon SA, et al. SAR11 clade dominates ocean surface bacterioplankton communities. Nature 2002;420:806–10. Nawrocki EP, Kolbe DL, Eddy SR. Infernal 10: inference of RNA alignments. Bioinformatics 2009;25:1335–7. Oh HM, Kang I, Lee K, et al. Complete genome sequence of strain IMCC9063, belonging to SAR11 subgroup 3, isolated from the Arctic Ocean. J Bacteriol 2011;193:3379–80. Pedrós-Alió C. Marine microbial diversity: Can it be determined? Trends Microbiol 2006;14:257–63. Pedrós-Alió C. The rare bacterial biosphere. Annu Rev Mar Sci 2012;4:449–66. Petri R, Imhoff JF. Genetic analysis of sea-ice bacterial communities of the western Baltic Sea using an improved double gradient method. Pol Biol 2001;24:252–7. Petrich C, Eicken H. Growth, structure and properties of sea ice. In: Thomas DN, Dieckmann GS (eds). Sea Ice. 2nd edn, Oxford, UK: Wiley-Blackwell Publishing, 2010, 23–77. Pinhassi J, Hagström Å. Seasonal succession in marine bacterioplancton. Aquat Microb Ecol 2000;21:245–56. Pizzetti I, Fuchs BM, Gerdts G, et al. Temporal variability of coastal Planctomycetes clades at Kabeltonne station, North Sea. Appl Environ Microb 2011;77:5009–17. Pomeroy LR, Wiebe WJ. Temperature and substrates as interactive limiting factors for marine heterotrophic bacteria. Aquat Microb Ecol 2001;23:187–204. Prasad S, Manasa P, Buddhi S, et al. Diversity and bioprospective potential (cold-active enzymes) of cultivable marine bacteria from the subarctic glacial Fjord, Kongsfjorden. Curr Microbiol 2014;68:233–8. R Development Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing. 2011, ISBN 3-900051-07-0. Available online at http://www.R-project.org/. Riedel A, Michel C, Gosselin M. Grazing of large-sized bacteria by sea-ice heterotrophic protists on the Mackenzie shelf during the winter-spring transition. Aquat Microb Ecol 2007;74:3–4. Riemann L, Leitet C, Pommier T, et al. The native bacterioplankton community in the central baltic sea is influenced by freshwater bacterial species. Appl Environ Microb 2008;74: 503–15. Rintala J-M, Piiparinen J, Uusikivi J. Drift-ice and under-ice water communities in the Gulf of Bothnia (Baltic Sea). Polar Biol 2010;33:179–91. 13 Roberts RJ, Vincze T, Posfai J, et al. REBASE-a database for DNA restriction and modification: enzymes, genes and genomes. Nucleic Acids Res 2010;38:D234–6. Sait L, Galic M, Strugnell RA, et al. Secretory antibodies do not affect the composition of the bacterial microbiota in the terminal ileum of 10-week-old mice. Appl Environ Microb 2003;69:2100–9. Schloss PD, Larget BR, Handelsman J. Integration of microbial ecology and statistics: a test to compare gene libraries. Appl Environ Microb 2004;70:5485–92. Schloss PD, Westcott SL, Ryabin T, et al. Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microb 2009;75:7537–41. Sinkko H, Lukkari K, Jama AS, et al. Phosphorus chemistry and bacterial community composition interact in brackish sediments receiving agricultural discharges. PLoS ONE 2011;6:e21555. Sipura J, Haukka K, Helminen H, et al. Effect of nutrient enrichment on bacterioplankton biomass and community composition in mesocosms in Archipelago Sea, northern Baltic. J Plankton Res 2005;27:1261–72. Smits JD, Riemann B. Calculation of cell production from [3H]thymidine incorporation with freshwater bacteria. Appl Environ Microb 1988;54:2213–9. Sogin ML, Morrison HG, Huber JA, et al. Microbial diversity in the deep sea and the underexplored ‘rare biosphere’. P Natl Acad Sci USA 2006;103:12115–20. Staden R, Beal KF, Bonfield JK. Computer methods in molecular biology. In: Misener S, Krawetz SA (eds). Bioinformatics Methods and Protocols. Totowa, NJ: Humana Press Inc, 1998, 115–30. Staden R, Judge DP, Bonfield JK. Managing sequencing projects in the GAP4 environment. In: Krawetz SA, Womble DD (eds). Introduction to Bioinformatics: A Theoretical and Practical Approach. Totowa, NJ: Humana Press Inc, 2003, 327–44. Staley JT, Gosink JJ. Poles apart: biodiversity and biogeography of sea ice bacteria. Annu Rev Microbiol 1999;53:189–215. Thomas DN, Dieckmann GS. Antarctic Sea ice—a habitat for extremophiles. Science 2002;295:641–4. Van Trappen S, Vandecandelaere I, Mergaert J, et al. Flavobacterium degerlachei sp nov, Flavobacterium frigoris sp, nov,m and Flavobacterium micromati sp, nov,, novel psychrophilic bacteria isolated from microbial mats in Antarctic lakes. Int J Syst Evol Micr 2004;54:85–92. Wang Q, Garrity GM, Tiedje JM, et al. Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ Microb 2007;73:5261–7. Zhou MY, Wang GL, Li D, et al. Diversity of both the cultivable protease-producing bacteria and bacterial extracellular proteases in the coastal sediments of King George island, Antarctica. PLoS ONE 2013;8:E79668.
© Copyright 2026 Paperzz