Physicochemical control of bacterial and protist ...

bs_bs_banner

Environmental Microbiology (2015)

doi:10.1111/1462-2920.12865

Physicochemical control of bacterial and protist community composition and diversity in Antarctic sea ice

Anders Torstensson,1* Julie Dinasquet,2† Melissa Chierici,3,4 Agneta Fransson,5,6 Lasse Riemann2 and Angela Wulff1 Departments of 1Biological and Environmental Sciences and 6 Earth Sciences, University of Gothenburg, Göteborg SE-40530, Sweden. 2 Marine Biological Section, Department of Biology, University of Copenhagen, Helsingør DK-3000, Denmark. 3 Department of Chemistry and Molecular Biology, University of Gothenburg, Göteborg SE-40530, Sweden. 4 Institute of Marine Research, Tromsø NO-9294, Norway. 5 Norwegian Polar Institute, Fram Centre, Tromsø NO-9296, Norway. Summary Due to climate change, sea ice experiences changes in terms of extent and physical properties. In order to understand how sea ice microbial communities are affected by changes in physicochemical properties of the ice, we used 454-sequencing of 16S and 18S rRNA genes to examine environmental control of microbial diversity and composition in Antarctic sea ice. We observed a high diversity and richness of bacteria, which were strongly negatively correlated with temperature and positively with brine salinity. We suggest that bacterial diversity in sea ice is mainly controlled by physicochemical properties of the ice, such as temperature and salinity, and that sea ice bacterial communities are sensitive to seasonal and environmental changes. For the first time in Antarctic interior sea ice, we observed a strong eukaryotic dominance of the dinoflagellate phylotype SL163A10, comprising 63% of the total sequences. This phylotype is known to be kleptoplastic and could be a significant primary producer in sea ice. We conclude Received 30 January, 2015; revised 20 March, 2015; accepted 31 March, 2015. *For correspondence. E-mail anders.torstensson@ bioenv.gu.se; Tel. +4631 786 6596; Fax +4631 786 2560. †Present address: Marine Biology Research Division, Scripps Institution of Oceanography, UCSD, La Jolla, CA 92037, USA.

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd

that mixotrophic flagellates may play a greater role in the sea ice microbial ecosystem than previously believed, and not only during the polar night but also during summer when potential food sources are abundant. Introduction The brine channels in sea ice provide a unique microhabitat for ice-associated microbial communities. Despite extreme physicochemical conditions experienced (Thomas and Dieckmann, 2002), a diverse range of organisms thrive in sea ice, encompassing members of multiple trophic levels, such as metazoans, unicellular algae, protozoa, cyanobacteria, bacteria, fungi and viruses (Horner et al., 1992; Bachy et al., 2011; Bowman et al., 2012). The large extent of sea ice and its high productivity makes it one of the most important polar ecosystems, acting both by seeding pelagic phytoplankton blooms and as an important direct food source for higher trophic levels (Lizotte, 2001; Riaux-Gobin et al., 2011). In addition, sea ice algae and bacteria play an important role in carbon biogeochemistry in polar areas (Fransson et al., 2013; Vancoppenolle et al., 2013). For instance, ice algae may contribute up to 25% of the annual primary production of ice-covered waters and provide an important food source for krill populations, especially during seasons when phytoplankton biomass is scarce in the water column (Kottmeier and Sullivan, 1987; O’Brien, 1987; Arrigo and Thomas, 2004). In addition, approximately 20–30% of the primary production is cycled through heterotrophic bacteria (Staley and Gosink, 1999), and bacterial production can even exceed primary production in thick Antarctic pack ice (Grossmann and Dieckmann, 1994). A number of species–energy hypotheses have sought to explain global biodiversity patterns on land and in the ocean, such as the ‘physical tolerance hypothesis’ (Currie et al., 2004), the ‘thermoregulatory loads hypothesis’ (Lennon et al., 2000) and the ‘range limitation hypothesis’ (Evans and Gaston, 2005). These hypotheses all assume that warmer locations support higher species diversity due to higher activity and lower physiological cost for endotherms. Nevertheless, these hypotheses have

2 A. Torstensson et al. received recent critique for being too general (Clarke and Gaston, 2006) and are not applicable to all microorganisms (Sharp et al., 2014). It seems likely that the extreme environment of sea ice has selected for unique microbial communities, and that environmental conditions, such as temperature and salinity, exert strong control selection pressures on microbial composition and diversity. However, extremophiles may not follow the general trends of higher species diversity at higher temperatures (Sharp et al., 2014). With the recent advancement of highthroughput sequencing, detailed examination of microbial diversity is now feasible. Still, there is very little published data from high-throughput sequencing of microbial diversity in sea ice (Bowman et al., 2012). Diatoms are by far the most studied protist group in sea ice (Arrigo et al., 2010; Torstensson et al., 2012; 2013). Nevertheless, studies have shown that other protist groups may be more common than previously believed, such as heterotrophic flagellates (Caron and Gast, 2010; Bachy et al., 2011; Paterson and Laybourn-Parry, 2012). Microzooplankton grazing is an important trophic link in planktonic ecosystems (Schmoker et al., 2013) and may play a significant role in carbon cycling in sea ice. However, heterotrophic protists have received relatively little attention in sea ice compared with autotrophic algae. Climate change is significantly affecting the extent and character of sea ice. A reduction of multiyear ice in the Arctic may affect the microbial community structure (Bowman et al., 2012; Hatam et al., 2014). Further effects of climate change on sea ice that will affect the sea ice microbial community may include more melt ponds, longer melt seasons, thinner and warmer ice (Markus et al., 2009; Frey et al., 2011; Arrigo et al., 2012). In turn, this could lead to changes in food web structure and elemental cycling in polar areas (Kirchman et al., 2009). To address how polar regions may respond to environmental and seasonal changes, key environmental drivers affecting community composition need to be identified. Although sea ice covers up to 7% of the earth’s surface during winter, molecular approaches to sea ice microbial diversity still represent a rather unexplored field (Deming, 2010). Bacterial diversity is believed to significantly promote ecosystem functioning (Bland et al., 1997), and changes in diversity may have significant effects on the whole polar ecosystem. To understand how microbial communities respond to climate change, for example, it is essential to examine their sensitivity to changes in the physicochemical environment. During the Oden Southern Ocean 2010/2011 cruise (Fig. 1, Table 1), we investigated environmental controls of diversity and community composition of bacteria and eukaryotes in Antarctic sea ice. By using 454-sequencing of the 16S and 18S rRNA

genes, we examined how key environmental variables (temperature, brine salinity, pH) affect composition and diversity of the sea ice microbial ecosystem. Although rapid warming is occurring in the western Antarctic oceans (Meredith and King, 2005), relatively little data on sea ice microbes are available from these regions.

Results Bacterial abundance and community composition Bacterial cell counts varied between 1.2 × 105 and 6.6 × 106 cells ml−1 in bulk ice, with the lowest abundance at station 15 (McMurdo Sound, Table 2). After quality control, a total of 177 775 partial 16S rRNA gene sequences remained among 26 samples (3149–13 127 sequences per sample). In total, the 16S rRNA sequences yielded 671 unique operational taxonomic units (OTUs) based on 97% similarity. Rarefaction curves for 16S rRNA genes normalized to 3149 sequences tended to plateau for most samples (Fig. S1A). A diverse range of bacterial taxa was observed in the sea ice samples (Fig. 2A). The bacterial phylum Proteobacteria was numerically dominant and represented 62.9% of the total sequences, with 356 unique OTUs. The Proteobacteria was dominated by the subclasses α-, β-, δ- and γ-Proteobacteria in most samples (Fig. 2A). The class Flavobacteria dominated (on average 26.2% of the total sequences, mainly Polaribacter) the second largest phylum Bacteroidetes. The most abundant order was Alteromonadales (of γ-Proteobacteria) containing genera common to sea ice, such as Pseudoalteromonas (12.0% of all sequences), Glaciecola (10.2% of all sequences) and Colwellia (1.3% of all sequences).

Eukaryotic community composition For eukaryotes, 321 068 partial 18S rRNA gene sequences among 25 samples remained after quality control (2812–31 801 sequences per sample). The sequences generated 383 unique OTUs (97% similarity). One of the 18S rRNA samples was lost due to technical errors. Rarefaction curves normalized to 2812 sequences tended to plateau for most samples (Fig. S1B). The protist community (Fig. 2B) was dominated by dinoflagellates, representing 69% of the total sequences. Ninety-two per cent of the dinoflagellate reads were closely related to the SL163A10 clade (Fig. 3A). The diatoms were the second most abundant group representing 23% of the total OTUs relative abundance across all stations. Pennate diatoms (Fig. 3B) were dominated by the classes Bacillariophyceae and Fragilariophyceae and accounted for 70% of the total diatom relative abundance.

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

Physicochemical control of sea ice microbial communities

3

Fig. 1. Sea ice sampling sites during the Oden Southern Ocean 2010/2011 cruise. Data of average sea extent from December 2010 were provided by the National Snow and Ice Data Center (http://nsidc.org/). Map was created using the Quantarctica QGIS-package, developed by the Norwegian Polar Institute (www.quantarctica.org).

Table 1. Station list and environmental parameters.

Station

Date

Latitude (°S)

Longitude (°W)

Ice thickness (cm)

Snow depth (cm)

Temperature (°C)

Bulk salinity

pHT

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

16 December 2010 17 December 2010 18 December 2010 19 December 2010 20 December 2010 21 December 2010 26 December 2010 27 December 2010 29 December 2010 30 December 2010 2 January 2011 2 January 2011 4 January 2011 6 January 2011 10 January 2011

68.35 69.28 70.01 70.91 72.25 72.46 72.57 72.11 72.03 72.10 72.05 72.48 73.26 75.33 77.35

102.1 103.0 106.6 111.9 115.3 114.1 116.6 118.6 123.1 127.1 132.4 135.4 139.2 149.2 E 165.4

70–90 80 69 149 92–95 102–106 232 146–152 119 192 90–95 132 85 140 155–157

31 20 30 80 35 32 43 55 25 35 10 17 0 38 0

−1.3 −1.4 −1.5 −1.8 −1.3 −1.4 −1.7 −1.7 −1.8 −1.7 −1.3 −1.3 −0.5 −1.2 −1.3

2.3 3.4 2.2 3.6 4.1 6.0 7.3 4.0 8.1 4.1 5.0 4.3 4.3 3.8 4.2

8.357 8.203 8.566 8.498 NA 8.665 NA NA 8.327 NA 8.905 8.239 9.041 8.379 8.474

pHT is reported at 15°C. See Fig. 1 for station numbers.


4

A. Torstensson et al.

Table 2. List of biological parameters. Station

n

Chlorophyll a

Fucoxanthin

Peridinin

Bacterial abundance

Chao1 (16S)

Chao1 (18S)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

3 1 1 1 2/1 2 1 2 2 1 2 2 2 2 2

NA 2.0 1.3 0.3 2.4 (±0.1) 14.3 (±0.1) 0.5 11.0 (±2.8) 6.4 (±0.1) 2.2 15.4 (±18.7) 9.1 (±5.2) 1.1 (±0.2) 11.8 (±0.8) 0.2 (0.04)

NA 3.6 4.2 0.5 5.1 (±0.3) 28.6 (±3.4) 0.6 17.4 (±0.6) 10.1 (±1.0) 3.2 21.0 (±26.2) 18.0 (±0.8) 1.8 (±0.2) 19.4 (±1.4) 0.1 (±0.07)

NA 0.3 0.7 0 0.08 (±0.1) 0.7 (±0.2) 0 0.1 (±0.2) 0.08 (±0.1) 0 1.4 (±2.5) 0.7 (±0.6) 0.09 (±0.1) 0.5 (±0.04) 0.0 (±0.007)

NA 8.2 13.4 5.8 9.4 (±2.6) 12.6 (±1.2) 5.6 5.4 (±2.6) 19.8 (±16.1) 9.6 36.6 (±19.6) 59.2 (±9.2) 5.6 (±2.5) 9.9 (±3.7) 1.4 (±0.3)

199 (±20) 246 228 374 215 (±24) 280 (±117) 371 288 (±40) 346 (±9) 336 154 (±10) 176 (±43) 150 (±25) 185 (±0.1) 93 (±10)

129 (±25) 85 144 109 98 107 (±20) 91 120 (±25) 138 (±30) 90 115 (±62) 72 (±12) 77 (±8) 122 (±17) 67 (±10)

Data are presented as average (±standard deviation) concentration of photosynthetic pigments (μg l−1 bulk ice), bacterial cell counts (105 cells ml−1 bulk ice), and Chao1 taxonomic richness for 16S and 18S rRNA. Pigment and bacterial abundance data were lost due to technical errors for station 1, and one sample of 18S taxonomic richness was lost at station 5.

Environmental parameters shaping microbial communities The replicated ice cores showed high similarity in bacterial and eukaryotic community composition, respectively (Fig. 4), indicating that the sea ice was homogenous with low patchiness. Hence, comparisons between separate biological and chemical samples can be addressed. Bottom sea ice temperature ranged between −0.5°C and −1.8°C, with the warmest temperature at station 13 (Table 1). In total, pH was measured

at 11 stations (Table 1) and did not have an effect on bacterial and eukaryotic richness and diversity (P > 0.05). A multiple linear regression on taxonomic richness (expressed as the number of unique 16S OTUs) versus sea ice temperature and bulk salinity was performed (R2 = 0.51). Bulk salinity was not significantly correlated (P = 0.42) with taxonomic richness and only explained 1.4% of the variation in the model. However, taxonomic richness showed positive correlation with calculated brine salinity (P < 0.0001, Fig. 5B) and negatively correlated

A Others Verrucomicrobia

75

δ-Proteobacteria β-Proteobacteria

50

Actinobacteria α-Proteobacteria

25 0

B

γ-Proteobacteria Bacteroidetes

Sample

Others and unclassified

100

Other Stramenopiles Alveolata unclassified

75

Protalveolata Ciliophora

50

Picozoa Opisthokonta

25

Cryptophyta 0

Cercozoa 1a 1b 1c 2a 3a 4a 5a 5b 6a 6b 7a 7b 8a 9a 9b 10a 11a 11b 12a 12b 13a 13b 14a 14b 15a 15b

Relative abundance (%)

Fig. 2. Relative abundance obtained by 454-sequencing on 15 sites where each individual sample is represented by the letters a, b or c. The relative abundance represents the proportion of all sequences from a given sample. (A) Major bacterial taxonomic divisions (phyla and Proteobacteria subclasses). (B) Eukaryotic major divisions. Sample 5b was lost due to technical errors. The two ‘Others’ groups represent taxonomic groups accounting for < 1% of total sequences in all samples.

1a 1b 1c 2a 3a 4a 5a 5b 6a 6b 7a 7b 8a 9a 9b 10a 11a 11b 12a 12b 13a 13b 14a 14b 15a 15b

Relative abundance (%)

100

Sample

Bacillariophyta Dinophyta


Physicochemical control of sea ice microbial communities

5

Fig. 3. Maximum likelihood tree of OTUs closely related to Alveolates (A) and Stramenopiles (B). Only OTUs representing more than 0.1% of the total relative abundance of 18S rRNA gene reads are included; the number of reads per OTU across all samples is presented in the parenthesis. Highlighted in grey are the two main OTUs representing 75% of the total relative abundance across all samples. Reference sequences are indicated in italic. Bootstrap values (n = 1000) are indicated at nodes; scale bar represents changes per positions.

Temperature(°C) −1.1

B

A

●

● ● 6a 6b ●●

9b

−0.5

●

●

● 3a ● 1c 1a 2a 1b Cluster I

−1.0

● ● 14b

5a5b ● 14a 12b

4a ●

Cluster II

●

11a

●

15a

15b Cluster III

● ● 11b ● ● 13a

13b 12a

Stress = 0.132 −1.0

−0.5

0.0 0.5 nMDS 1

nMDS 2

8a

● ● 8b ● 10a ● 9a● 7a

1.0

1.5

9b 9a ●

7a ●

4a −0.5

0.5

●

●

● ●

0.5

Cluster IV

nMDS 2 0.0

−0.4

0.0

1.0

−1.8

●

●

● 8a 8b

● ● ●13a 11b12a13b

●

1a ● 2a ●● 1c ● 1b 5a ●

Cluster II ● ● 12b 11a

●

10a

●

15b 15a Cluster III

●

6b ● 3a 6a Cluster I −0.5

●

● ●

14a 14b

0.0 nMDS 1

Stress = 0.18 0.5

1.0

Fig. 4. Non-metric multidimensional scaling of (A) bacterial communities in 26 sea ice samples and (B) eukaryotic community in 25 samples, based on Bray–Curtis dissimilarity. The colour of each sample represents sea ice temperature and the ellipses for each cluster show 95% confidence interval of standard deviation.



Taxonomic richness (#OTUs)

6

400

R2 = 0.491

A

P < 0.0001

300 200 100 0 -2.0

-1.5

-1.0

-0.5

0.0

Taxonomic richness (#OTUs)

Sea ice temperature (°C) 400

R2 = 0.557 P < 0.0001

B

300 200

Bacteria and protist co-occurrence

100 0

0

10

20

30

40

Shannon’s diversity H'

Brine salinity 8

In addition, station 1–3 formed a separate cluster (Cluster I, Fig. 4A), and the samples from station 15 (McMurdo Sound) differed from all other samples in terms of both bacterial and eukaryotic community composition (Cluster III, Fig. 4). Station 15 had the lowest bacterial and eukaryotic richness, 72 and 52 OTUs respectively. According to the canonical correspondence analysis (CCA), dinoflagellates did not relate with chl a concentration (Fig. 6B), suggesting that diatoms were the main primary producers of the sea ice. However, sea ice thickness was a main driver for explaining Alveolata dominance (Fig. 6A). Concentrations of the photosynthetic pigments chl a, fucoxanthin and peridinin are shown in Table 2. Chl a concentration correlated significantly with fucoxanthin concentration (P < 0.0001, R2 = 0.918).

R2 = 0.320 P < 0.003

C

6

A CCA of the major (> 1% of total sequences, Table 3) OTUs revealed co-occurrence of protists and bacteria (Fig. 6B). Phylogenetic analysis revealed that OTU# 236 was closely related to the dinoflagellate Polarella glacialis (Fig. 3A), and explained the occurrence of bacterial OTUs from the genera Pseudoalteromonas, Alteromonas and Oleispira (Fig. 6B). Furthermore, the abundance of an OTU from Winogradsky, Thiohalorhabdales and Rhodobacteraceae was related to two diatom OTUs, 509 and 263 (closely related to Fragilariopsis spp., Figs 3B and 6B).

4

Discussion

2 0 -2.0

-1.5

-1.0

-0.5

0.0

Sea ice temperature (°C) Fig. 5. Environmental control of 16S rRNA gene richness and diversity. (A) Temperature dependency of taxonomic richness (expressed as number of OTUs). (B) Temperature dependency of calculated brine salinity. (C) Temperature dependency of Shannon’s diversity index.

with increasing temperature (P < 0.0001, Fig. 5A). Shannon’s diversity index was also negatively correlated with temperature (P < 0.004, Fig. 5C). We found no correlations between bacterial abundance and temperature (P = 0.15), and between chlorophyll a (chl a) and temperature (P = 0.07). In addition, no correlation between sampling date and sea ice temperature could be detected (P > 0.05). Following non-metric multidimensional scaling of bacterial community composition, clustering revealed that the coldest samples (Cluster IV) separated from the remaining samples (Fig. 4A).

In this study, we have explored environmental control of the microbial community structure of Antarctic sea ice. We used 454-sequencing of the 16S rRNA and 18S rRNA genes to understand how microbial communities are affected by changes in the physicochemical properties of sea ice. Understanding the environmental regulations of microbial communities is a key factor to understand the implications of seasonal and environmental changes on polar ecosystems. Sea ice bacterial community Our findings show that sea ice bacterial taxonomic richness is consistently related to temperature, and consequently brine salinity. The lowest sea ice temperatures recorded during our cruise, close to the freezing point of oceanic water, generated the highest 16S taxonomic richness and diversity. Taxonomic richness and Shannon’s index gradually decreased when sea ice temperature approached 0°C. Physical loss of protists and bacteria due to brine drainage and ice melt can be disregarded


Physicochemical control of sea ice microbial communities 0.8 A

7

1.0 B

Temp

509

Bacteria Peridinin

pH

236

Depth

263

Chl a Fucoxanthin 454 251

Salinity 62

−0.6 Pseudo-F = 3.0, P = 0.019 −0.6

1.0

300

−0.6 Pseudo-F = 3.2, P = 0.001 −1.0

1.0

Fig. 6. Canonical correspondence analysis (CCA) of eukaryotic and bacteria communities. (A) 18S rRNA gene composition and the environmental variables: sea ice depth, bulk pH, salinity, sea ice temperature, bacterial biomass and photosynthetic pigment concentrations (chlorophyll a, fucoxanthin and peridinin). (B) Association of the major bacterial (points) and protist (arrows) OTUs (> 1% of total sequences). See Table 3 for OTU list. Pseudo-F statistic was obtained in CANOCO 5 by testing all constrained axes using the Monte Carlo permutation procedure (n = 1000).

since chl a and the standing stock of bacteria did not decrease with higher temperature. In addition, the variation in temperature was not dependent on sampling date, showing that the results were not artefacts from sampling during 25 days of gradual summer ice melt. These results rather illustrate the strong acclimation and adaptation of sea ice microorganisms to the extreme environment of Southern Ocean sea ice. Temperature and salinity are probably the two most important factors controlling bacterial diversity in bottom sea ice, although temperature ranges are much lower in the interior ice compared with winter ice and at the surface of Arctic sea ice (Collins et al., 2010; Ewert and Deming, 2014). Due to the well-

drained summer sea ice, containing brine with low salinity (down to a salinity of 10.1), temperatures close to 0°C could be observed. Since the calculated brine salinity in sea ice is controlled by temperature due to the freezingpoint depression, it is impossible to separate their single effects in this study. It is well known that increasing temperature may alter community composition and diversity. General theories suggest that environments with higher temperatures support larger species diversity (Lennon et al., 2000; Currie et al., 2004; Evans and Gaston, 2005; Fuhrman et al., 2008). Conversely, these general assumptions do not seem to apply to extreme environments, such as

Table 3. Operational taxonomic unit (OTU) list and deepest taxonomy identified for the most abundant OTUs (> 1% of total sequences). 16S

18S

OTU #

Taxa

% of total

OTU #

Taxa

% of total

3 54 188 208 219 465 476 669 670 724 788 832 873 962 1015

γ-Proteobacteria Winogradskyella sp. γ-Proteobacteria Polaribacter irgensii Loktanella sp. Rhodobacteraceae Rhodobacteraceae Thiohalorhabdales Rhodobacteraceae Alteromonadaceae Flavobacteriaceae Pseudoalteromonas sp. Alteromonas sp. Saprospiraceae Oleispira sp.

4.8 1.7 1.4 10.6 2.1 1.2 2.5 1.2 3.9 9.0 2.2 12.0 1.7 2.1 2.6

62 236 251 263 300 454 509 529

Bacillariophyta Gymnodiniphycidae Bacillariophyta Bacillariophyta SL163A10 Gymnodiniphycidae Coscinodiscophyta Bacillariophyta

1.7 2.3 1.0 11.7 63.3 1.9 2.2 1.4

Taxonomic assignment was performed against the Greengenes 13_5 and PR2 databases for 16S and 18S respectively.


8


geothermal springs (Sharp et al., 2014) and sea ice (this study). The upper part of the sea ice exerts substantial stress on sea ice communities, as temperature and salinity are more variable and extreme (Ewert and Deming, 2014). However, interior sea ice is a much more thermally stable environment than, for instance, air or snow, with summer interior temperatures generally ranging between 0°C and −2.5°C (Fransson et al., 2011). Therefore, strong acclimation and adaptation to the sea ice environment are expected, and a temperature/salinity dependency is not surprising. Nevertheless, just as in geothermal springs (Sharp et al., 2014), we show much stronger relationships between temperature and diversity than previously reported for marine bacterioplankton (Fuhrman et al., 2008). Thus, Antarctic sea ice bacterial communities may be extremely sensitive to changes in environmental conditions. By using 454-sequencing, pack ice was shown to hold a much larger bacterial diversity than previously believed using traditional techniques. Brinkmeyer and colleagues (2003) reported Shannon indices of bacteria from Southern Ocean pack ice between 1.127 and 0.812, based on 16S rRNA gene clone libraries. By using 454-sequencing, we found Shannon indices that are six times higher than the previous estimates. However, we also observed some spatial differences in community composition that could be related to differences in the early colonization of the ice. In addition, the samples from the station at McMurdo Sound were different than all other samples, both in terms of biomass and community composition. Land-fast (e.g. station number 15, McMurdo Sound) and multiyear ice (Bowman et al., 2012) probably display very different microbial diversity and densities compared with pack ice due to differences in their physicochemical properties. Adaption to an extreme environment may result in low resilience to, for example, environmental change. All major bacterial taxa (> 1% of total sequences) found in this study have previously been reported to be associated with sea ice (Deming, 2010; Maas et al., 2011; Bowman et al., 2012). The dominant genus, Pseudoalteromonas, contains several psychrophilic species ubiquitous to sea ice and has been described as one of the most culturable bacteria from sea ice (Yu et al., 2015). In addition, an psychrophilic Antarctic member of Oleispira, isolated from sea ice (Oleispira antarctica), has been characterized to be hydrocarbonoclastic (Yakimov et al., 2003). Generally, hydrocarbon-degrading bacteria occur in very low abundance in the absence of oil pollution. Although very low levels of oil is present in Antarctic sea ice, other hydrocarbons, such as volatile halogenated organic compounds (halocarbons), are abundant and believed to be involved in several biological pathways (Granfors et al., 2013). Such halocarbons are highly reactive and are related to the destruction of ozone in the

stratosphere, and the specific species involved in production and degradation are not fully identified. Hence, hydrocarbonoclastic bacteria, such as Oleispira, could be involved in the biological degradation of halocarbons, a process that should be further investigated. Sea ice protists community To our knowledge, this is the first study in Antarctica demonstrating dominance of dinoflagellates in bottom sea ice. It is generally believed that diatoms are the most abundant eukaryotic organism group in interior sea ice, and that heterotrophic organisms are mainly restricted to the upper ice (Arrigo et al., 2010; Eddie et al., 2010). However, there may be temporal deviations from this pattern. For instance, Bachy and colleagues (2011) reported dominance of alveolates (> 70%) using 18S rRNA gene clone libraries from various parts of Arctic sea ice at the end of the polar night. The authors explained the dominance of mixotrophic dinoflagellates by the fact that the samples were collected after several months of darkness. On the other hand, in our case, the samples have experienced sunlight for several months. According to the CCA, the relative abundance of Dinophyta also increased with sea ice thickness, suggesting that heterotrophic dinoflagellates could become more dominant under darker conditions when less light is available for photosynthesis. The finding of dominance of dinoflagellates in our samples could be biased by the large genome and multiple ribosomal gene copies per genome in dinoflagellates (Zhu et al., 2005) or due to preferential amplification of dinoflagellates. However, the V4 primers used by us have previously been shown to efficiently target diverse eukaryotes and be less biased towards dinoflagellates compared with V9 primers (Stoeck et al., 2010). Nevertheless, the dominant dinoflagellate identified (63% of the total OTUs relative abundance) was closely related to a dinoflagellate found to be very abundant in Antarctic water (Ross Sea), especially in pack ice environments with abundance comparable with planktonic bloom conditions (Gast et al., 2006). This suggests that it was indeed very abundant in our samples, and this was not due to polymerase chain reaction (PCR) or sequencing bias. Dinoflagellates related to this clade were also found abundant in deep Antarctic waters (Lopez-Garcia et al., 2001), around the Antarctic Peninsula (Luria et al., 2014), and were isolated from the Ross Sea (Gast et al., 2006). This shows that this specific dinoflagellate could play a key role in food web dynamics and elemental cycling in Antarctic waters. A prevalence of dinoflagellates may also be explained by the high abundance of bacteria in some samples. Bacterial cells were up to 70 times enriched in sea ice brine compared with under ice water (A. Torstensson,


Physicochemical control of sea ice microbial communities unpubl. data), and could be important prey for some dinoflagellates. According to the CCA, Alveolata did not seem to relate to chl a, suggesting that the dinoflagellates in this study could be hetero- or mixotrophic. Both concentration of the diatom-specific pigment fucoxanthin and Bacillariophyta relative abundance in the sequence dataset were strongly related to chl a, implying that diatoms were the main primary producers in the sea ice. The diatoms that were observed in this study were mainly pennate and are commonly associated with sea ice, such as the major diatom OTU that was closely related with Fragilariopsis curta and Fragilariopsis kerguelensis. Furthermore, Gast and colleagues (2007) suggested that the dominant dinoflagellate SL163A10 engage in kleptoplasty of Phaeocystis antarctica chloroplasts, which could complicate the interpretation of pigments. For instance, the use of the signature light harvesting pigments in dinoflagellates, peridinin, as an indicator of kleptoplastic dinoflagellates may be misleading. Although dinoflagellates were prevalent, very low concentrations of peridinin were observed and agrees with the findings in Luria and colleagues (2014). Kleptoplasty could possibly explain why low concentrations of peridinin are observed in these areas, and this would support that the majority of the observed dinoflagellates were indeed kleptoplastic. Microzooplankton has recently been estimated to graze more than 60% of the daily global planktonic primary production and is extremely important in global carbon budgets (Schmoker et al., 2013). Sea ice algae contribute to a large part of the Southern Ocean primary production (Arrigo and Thomas, 2004). However, very little is known about microzooplankton grazing in sea ice due to the complexity of estimating in situ production rates in sea ice (Mock, 2002; Caron and Gast, 2010). The dominant kleptoplastic dinoflagellate does not seem to be actively grazing (Gast et al., 2007), although many other mixotrophic flagellates can be significant grazers in the Southern Ocean (Gast et al., 2014). Kleptoplasty and microzooplankton grazing may have a significant role in sea ice ecology and biogeochemistry, and therefore needs to be addressed in order to correctly account for sea ice in global carbon budgets. Naked dinoflagellates have previously been observed in the slush layer of sea ice (Garrison and Buck, 1989; Gast et al., 2006). By using 18S rRNA gene clone libraries, Gast and colleagues (2006) reported the dominance of a novel dinoflagellate phylotype closely related to Karenia and Karlodinium species, in the slush layer of sea ice in the Ross Sea. Until now, this phylotype has been believed to be rare in interior Antarctic sea ice (Gast et al., 2006). Other dinoflagellates, such as Polarella glacialis, have also been observed in high abundance and restricted to the upper section of land-fast ice (Montresor et al., 1999). It is still unknown why the prevalence of

9

dinoflagellates in sea ice that have been reported using molecular techniques (Gast et al., 2006; Bachy et al., 2011) is not reflected in traditional microscopy approaches. One possible explanation could be that dinoflagellates are undersampled due to their fragile cell structures, and may not stay structurally intact during the stressful processing techniques of sea ice samples (e.g. thawing or centrifuging) (Garrison and Buck, 1986). In addition, the lack of distinctive morphologies of gymnodinoid dinoflagellates makes them difficult to identify using light microscopy (Gast et al., 2006). The presence of dinoflagellates and their ecological importance in Southern Ocean sea ice may be underestimated and needs to be further addressed, especially in summer pack ice. Bacteria and protist co-occurrence We observed some co-occurrences between protists and bacteria that could explain their prevalence. One may, for instance, speculate that the release of inorganic nutrients associated with bacterial mineralization of organic matter would benefit specific eukaryotes and eukaryotic community structure. This may extend the summer bloom when nutrient levels are generally low (Fransson et al., 2011). For instance, we observed an association between the taxa closely related to the dinoflagellate Polarella glacialis and bacteria from Pseudoalteromonas, Alteromonas and Oleispira. A potential symbiotic relationship between specific bacteria and protists may be prevalent and could be further explored in situ by single cell sequencing (Martinez-Garcia et al., 2012). Hence, other factors than the physicochemical environment may structure sea ice microbial communities. Concluding remarks Our results show that both the diversity and complexity of sea ice microbial communities may be even greater than previously believed. Sea ice microbial communities are important producers in the Southern Ocean, and sea ice plays an important role in global energy budgets and ocean–atmosphere interactions (Arrigo and Thomas, 2004; Arrigo et al., 2010; Fransson et al., 2011). Since sea ice microbial communities also seed the pelagic production (Riaux-Gobin et al., 2011) and provide a direct link to higher trophic levels through grazing (O’Brien, 1987), changes in these communities may have significant consequences for biogeochemistry of polar ecosystems as such. For the first time at the ice/water intersect, we observed eukaryotic dominance of a dinoflagellate ubiquitous in Antarctic water masses. We suggest that dinoflagellates may play an important role in bottom sea ice ecology, and that kleptoplastic dinoflagellates are more


10


widespread in sea ice than previously believed. We have also shown that the sea ice bacterial community is strongly dependent on environmental conditions, such as sea ice temperature and brine salinity. Climate and seasonal changes result in warmer and less saline brine, which could reduce the diversity, richness and composition of psychro- and halophilic sea ice bacteria. This may, in turn, impact food web dynamics and elemental cycling in polar oceans. Experimental procedures

(CCAGCASYGCGGTAATTCC) and TARR (ACTTCGTTCT TGATYRA), as described in Stoeck and colleagues (2010). For both bacteria and eukaryotes, the amplification was run with the regular primers followed by a second step with the primers complemented with 454-adapters and samplespecific barcodes (Berry et al., 2011). For each sample, triplicate PCR products from independent runs were pooled prior to being purified (Agencourt AMPure XP kit, Beckman Coulter) and quantified using PicoGreen (Molecular Probes). The samples were mixed in equimolar amounts and sequenced from the reverse primer direction using Roche/ 454 GS FLX Titanium technology (National High-Throughput DNA Sequencing Center, University of Copenhagen).

Study site Phylogenetic analysis Sea ice samples were collected from 15 sites for 25 days between December 2010 and January 2011 in the Amundsen and Ross seas (Table 1, Fig. 1). A sampling plot with homogeneous sea ice was chosen using terrestrial LiDAR surveys (Weissling and Ackley, 2015). Separate sea ice cores for biological and chemical analyses were collected within a radius of < 1 m using a 12 cm diameter ice corer. In total, 26 ice cores were sampled for biological analyses (community composition, photosynthetic pigments and bacterial abundance), which resulted in duplicated cores from most stations (Table 2). The cores were immediately covered in opaque black plastic bags for protection against sunlight and cut into 10 cm sections. To minimize osmotic and thermal stress, the lowermost sections were placed in 1 l, 0.2 μm filtered seawater (∼ 4°C) and thawed in darkness for 12–20 h. Sea ice temperature was recorded immediately after recovery of cores used for physicochemical analyses (temperature, salinity, pH) using a digital thermistor (Ama-digit ad 15 th, Amarell GmbH & Co, Kreuzwertheim, Germany). The samples were vacuum-packed in gas-tight Tedlar© bags and left to thaw overnight. Salinity in the melted sea ice (bulk salinity) was measured using a conductivity meter (Cond 310i, WTW GmbH, Weilheim, Germany) and converted into brine salinity together with sea ice temperature, using the equations described in Frankenstein and Garner (1967): vb/v = Si(0.0532–4.919/Ti), for ice temperatures (Ti) between −22.9°C and −0.5°C, and where vb/v is the brine volume (percent of total core volume), and Si is the bulk ice salinity. Brine salinity was subsequently calculated from the brine volume and bulk salinity.

Bacterial and eukaryotic community composition A total of 400–600 ml melted sample was immediately filtered on 0.2 μm Supor200® filter (PALL Life Science) after thawing, using acid-washed filtration equipment. The filters were stored at −80°C until extraction. The DNA was extracted using an enzyme/phenol-chloroform protocol (Riemann et al., 2000) but with a 30 min lysozyme digestion at 37°C and an overnight proteinase K digestion (20 mg ml−1 final concentration) at 55°C (Boström et al., 2004). Bacterial 16S rRNA genes were PCR-amplified using puReTaq ReadyTo-Go PCR beads (GE Healthcare), 0.06 ng DNA μl−1, and primers Bakt_341F (CCTACGGGNGGCWGCAG) and Bakt_805R (GACTACHVGGGTATCTAATCC). Eukaryotes 18S rRNA genes (V4 regions) were PCR-amplified with TARF

Sequences were analysed and processed using the Quantitative Insights Into Microbial Ecology software (QIIME v1.4, Caporaso et al., 2010b) with default settings, excluding sequences < 350 bp or > 450 bp. Flowgrams were denoised directly in the pipeline (Reeder and Knight, 2010). All singletons were removed. Sequences were clustered into OTUs at 97% pairwise identity using the seed-based UCLUST algorithm, and representative sequences from each bacterial OTU aligned to the Greengenes imputed core reference alignment (DeSantis et al., 2006) (http://greengenes.lbl.gov) using PYNAST (Caporaso et al., 2010a). Eukaryotes OTUs were aligned to SILVA 119 imputed core reference alignment (Quast et al., 2013b). Chimeras were removed using CHIMERA SLAYER (Haas et al., 2011). Taxonomy assignments were made using the Ribosomal Database Project (RDP) classifier (Wang et al., 2007) against the database Greengenes 13_5 (McDonald et al., 2012) for 16S sequences, and both SILVA 119 (Quast et al., 2013a) and PR2 (Guillou et al., 2013) databases for 18S sequences. All sequences obtained in this study have been deposited in the European Nucleotide Archive-Short Read Archive under Accession Number PRJEB7907. Maximum likelihood trees were computed with MEGA6 (Tamura et al., 2013).

Photosynthetic pigments Directly after thawing, 70–185 ml of melted sea ice was filtered on 25 mm GF/F filters. The filters were immediately flash-frozen in liquid nitrogen and stored at −80°C until extraction. Cells were extracted in 2 ml acetone/methanol (80:20) while sonicated using a Vibra-cell sonicating probe, operating at 80% in 5-s pulses. High performance liquid chromatographic analyses of the extracts were performed according to Wright and Jeffrey (1997), using an absorbance diode array-based detector (Spectraphysics UV6000LP). A 150 × 3.0 mm Phenomenex Kinetex 2.6 μ C18 100A column was used for separation. Pigments were identified by their retention time and absorbance spectra (400–700 nm). DHI Water and Environment, Denmark, provided pigment calibration standards for quantification of pigment concentrations.

Bacterial abundance Samples of melted sea ice were fixed in 1% glutaraldehyde (final concentration) and stored at −80°C. Bacteria were


Physicochemical control of sea ice microbial communities stained with SYBR Green Nucleic Acid Gel Stain (Invitrogen, Eugene, OR, USA) for 10 min in darkness. Counting was performed with a FACScalibur flow cytometer (BD Biosciences, Mountain View, USA). Flow rate was determined with an internal standard of ultrasonicated 1 μm FluoSpheres (Invitrogen). Concentration of the internal standard was determined by triplicated measurement together with BD Trucount absolute counting beads (BD Biosciences).

pH pH in bulk ice was measured spectrophotometrically in a 1 cm cell using sulfonephthalein dye, m-cresol purple, as an indicator (Clayton and Byrne, 1993). Prior to analysis, the samples were thermostated to 15°C. The perturbation of seawater pH caused by the addition of the indicator solution was calculated and corrected for using the method described in Chierici and colleagues (1999). The analytical precision was estimated at ± 0.001 pH units, which was determined by a series of 10 analyses of one sample.

Statistical analyses Canonical correspondence analyses were performed in CANOCO 5 to explore potential linkages between community composition and environmental variables. Bootstrapping was performed using Monte Carlo permutation procedure, executed with 1000 iterations for each CCA. Non-metric multidimensional scaling and cluster analysis were performed using Bray–Curtis dissimilarity with the vegan package in R (www.r-project.org). Linear regression analyses were used to describe the relationships between taxonomic richness and diversity and environmental parameters. A probability level (P) of < 0.05 was used for statistical significance.

Acknowledgements We are grateful to the captain and crew of the IB Oden for their support. We thank the Swedish Polar Research Secretariat and the US National Science Foundation for the cruise planning and logistical support before and during the Oden Southern Ocean expedition 2010/2011. This project was funded by the Swedish Research Council projects (20078365, 2008-6228 and 2009-2994), the BONUS+ project BAZOOCA (Baltic Zooplankton Cascades) through the Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning, FORMAS (2008-1893 and 20081882), and grants from the Royal Society of Arts and Sciences in Gothenburg (KVVS) and The Danish Council for Independent Research, Natural Sciences (09-066396). The authors declare no conflict of interest.

References Arrigo, K.R., and Thomas, D.N. (2004) Large scale importance of sea ice biology in the Southern Ocean. Antarct Sci 16: 471–486. Arrigo, K.R., Mock, T., and Lizotte, M.P. (2010) Primary producers and sea ice. In Sea Ice. Thomas, D.N., and

11

Dieckmann, G.S. (eds). Oxford: Wiley-Blackwell, pp. 283– 325. Arrigo, K.R., Perovich, D.K., Pickart, R.S., Brown, Z.W., van Dijken, G.L., Lowry, K.E., et al. (2012) Massive phytoplankton blooms under Arctic Sea ice. Science 336: 1408. Bachy, C., López-Garcia, P., Vereshchaka, A., and Moreira, D. (2011) Diversity and vertical distribution of microbial eukaryotes in the snow, sea ice and seawater near the north pole at the end of the polar night. Front Microbiol 2: article 106. doi: 10.3389/fmicb.2011.00106. Berry, D., Ben Mahfoudh, K., Wagner, M., and Loy, A. (2011) Barcoded primers used in multiplex amplicon pyrosequencing bias amplification. Appl Environ Microbiol 77: 7846–7849. Bland, J.F., Maberly, S.C., and Cooper, J.I. (1997) Microbial diversity and ecosystem function. Oikos 80: 209–213. Boström, K.H., Simu, K., Hagström, A.A., and Riemann, L. (2004) Optimization of DNA extraction for quantitative marine bacterioplankton community analysis. Limnol Oceanogr Methods 2: 365–373. Bowman, J.S., Rasmussen, S., Blom, N., Deming, J.W., Rysgaard, S., and Sicheritz-Ponten, T. (2012) Microbial community structure of Arctic multiyear sea ice and surface seawater by 454 sequencing of the 16S RNA gene. ISME J 6: 11–20. Brinkmeyer, R., Knittel, K., Jürgens, J., Weyland, H., Amann, R., and Helmke, E. (2003) Diversity and structure of bacterial communities in Arctic versus Antarctic pack ice. Appl Environ Microbiol 69: 6610–6619. Caporaso, J.G., Bittinger, K., Bushman, F.D., DeSantis, T.Z., Andersen, G.L., and Knight, R. (2010a) PyNAST: a flexible tool for aligning sequences to a template alignment. Bioinformatics 26: 266–267. Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E.K., et al. (2010b) QIIME allows analysis of high-throughput community sequencing data. Nat Methods 7: 335–336. Caron, D.A., and Gast, R.J. (2010) Heterotrophic protists associated with sea ice. In Sea Ice. Thomas, D.N., and Dieckmann, G.S. (eds). Oxford: Wiley-Blackwell, pp. 327– 356. Chierici, M., Fransson, A., and Anderson, L.G. (1999) Influence of m-cresol purple indicator additions on the pH of seawater samples: correction factors evaluated from a chemical speciation model. Mar Chem 65: 281–290. Clarke, A., and Gaston, K.J. (2006) Climate, energy and diversity. Proc Biol Sci 273: 2257–2266. Clayton, T.D., and Byrne, R.H. (1993) Spectrophotometric seawater pH measurements: total hydrogen ion concentration scale calibration of m-cresol purple and at-sea results. Deep Sea Res Part I Oceanogr Res Pap 40: 2115–2129. Collins, R.E., Rocap, G., and Deming, J.W. (2010) Persistence of bacterial and archaeal communities in sea ice through an Arctic winter. Environ Microbiol 12: 1828–1841. Currie, D.J., Mittelbach, G.G., Cornell, H.V., Field, R., Guégan, J.-F., Hawkins, B.A., et al. (2004) Predictions and tests of climate-based hypotheses of broad-scale variation in taxonomic richness. Ecol Lett 7: 1121–1134. Deming, J.W. (2010) Sea ice bacteria and viruses. In Sea Ice. Thomas, D.N., and Dieckmann, G.S. (eds). Oxford: WileyBlackwell, pp. 247–282.


12


DeSantis, T.Z., Hugenholtz, P., Larsen, N., Rojas, M., Brodie, E.L., Keller, K., et al. (2006) Greengenes, a Chimerachecked 16S rRNA gene database and workbench compatible with ARB. Appl Environ Microbiol 72: 5069–5072. Eddie, B., Juhl, A., Krembs, C., Baysinger, C., and Neuer, S. (2010) Effect of environmental variables on eukaryotic microbial community structure of land-fast Arctic sea ice. Environ Microbiol 12: 797–809. Evans, K.L., and Gaston, K.J. (2005) Can the evolutionaryrates hypothesis explain species-energy relationships? Funct Ecol 19: 899–915. Ewert, M., and Deming, J.W. (2014) Bacterial responses to fluctuations and extremes in temperature and brine salinity at the surface of Arctic winter sea ice. FEMS Microbiol Ecol 89: 476–489. Frankenstein, G., and Garner, R. (1967) Equations for determining the brine volume of sea ice from -0.5 °C to -22.9 °C. J Glaciol 6: 943–944. Fransson, A., Chierici, M., Yager, P.L., and Smith, W.O. (2011) Antarctic sea ice carbon dioxide system and controls. J Geophys Res Oceans 116: C12035. doi:10.1029/2010JC006844. Fransson, A., Chierici, M., Miller, L.A., Carnat, G., Shadwick, E., Thomas, H., et al. (2013) Impact of sea-ice processes on the carbonate system and ocean acidification at the ice-water interface of the Amundsen Gulf, Arctic Ocean. J Geophys Res Oceans 118: 7001–7023. Frey, K.E., Perovich, D.K., and Light, B. (2011) The spatial distribution of solar radiation under a melting Arctic sea ice cover. Geophys Res Lett 38: L22501. Fuhrman, J.A., Steele, J.A., Hewson, I., Schwalbach, M.S., Brown, M.V., Green, J.L., and Brown, J.H. (2008) A latitudinal diversity gradient in planktonic marine bacteria. Proc Natl Acad Sci U S A 105: 7774–7778. Garrison, D., and Buck, K. (1986) Organism losses during ice melting: a serious bias in sea ice community studies. Polar Biol 6: 237–239. Garrison, D., and Buck, K. (1989) The biota of Antarctic pack ice in the Weddell sea and Antarctic Peninsula regions. Polar Biol 10: 211–219. Gast, R.J., Moran, D.M., Beaudoin, D.J., Blythe, J.N., Dennett, M.R., and Caron, D.A. (2006) Abundance of a novel dinoflagellate phylotype in the Ross Sea, Antarctica. J Phycol 42: 233–242. Gast, R.J., Moran, D.M., Dennett, M.R., and Caron, D.A. (2007) Kleptoplasty in an Antarctic dinoflagellate: caught in evolutionary transition? Environ Microbiol 9: 39–45. Gast, R.J., McKie-Krisberg, Z.M., Fay, S.A., Rose, J.M., and Sanders, R.W. (2014) Antarctic mixotrophic protist abundances by microscopy and molecular methods. FEMS Microbiol Ecol 89: 388–401. Granfors, A., Karlsson, A., Mattsson, A., Smith, W.O., and Abrahamsson, K. (2013) Contribution of sea ice in the Southern Ocean to the cycling of volatile halogenated organic compounds. Geophys Res Lett 40: 1–6. Grossmann, S., and Dieckmann, G.S. (1994) Bacterial standing stock, activity, and carbon production during formation and growth of sea ice in the Weddell Sea, Antarctica. Appl Environ Microbiol 60: 2746–2753. Guillou, L., Bachar, D., Audic, S., Bass, D., Berney, C., Bittner, L., et al. (2013) The Protist Ribosomal Reference

database (PR2): a catalog of unicellular eukaryote small sub-unit rRNA sequences with curated taxonomy. Nucleic Acids Res 41: D597–D604. Haas, B.J., Gevers, D., Earl, A.M., Feldgarden, M., Ward, D.V., Giannoukos, G., et al. (2011) Chimeric 16S rRNA sequence formation and detection in Sanger and 454pyrosequenced PCR amplicons. Genome Res 21: 494– 504. Hatam, I., Charchuk, R., Lange, B., Beckers, J., Haas, C., and Lanoil, B. (2014) Distinct bacterial assemblages reside at different depths in Arctic multiyear sea ice. FEMS Microbiol Ecol 90: 115–125. Horner, R., Ackley, S., Dieckmann, G., Gulliksen, B., Hoshiai, T., Legendre, L., et al. (1992) Ecology of sea ice biota. Polar Biol 12: 417–427. Kirchman, D.L., Moran, X.A.G., and Ducklow, H. (2009) Microbial growth in the polar oceans – role of temperature and potential impact of climate change. Nat Rev Microbiol 7: 451–459. Kottmeier, S.T., and Sullivan, C.W. (1987) Late winter primary production and bacterial production in sea ice and seawater west of the Antarctic Peninsula. Mar Ecol Prog Ser 36: 287–298. Lennon, J.J., Greenwood, J.J.D., and Turner, J.R.G. (2000) Bird diversity and environmental gradients in Britain: a test of the species–energy hypothesis. J Anim Ecol 69: 581– 598. Lizotte, M.P. (2001) The contributions of sea ice algae to Antarctic marine primary production. Am Zool 41: 57–73. Lopez-Garcia, P., Rodriguez-Valera, F., Pedros-Alio, C., and Moreira, D. (2001) Unexpected diversity of small eukaryotes in deep-sea Antarctic plankton. Nature 409: 603–607. Luria, C.M., Ducklow, H.W., and Amaral-Zettler, L.A. (2014) Marine bacterial, archaeal and eukaryotic diversity and community structure on the continental shelf of the western Antarctic Peninsula. Aquat Microb Ecol 73: 107– 121. Maas, E.W., Simpson, A.M., Martin, A., Thompson, S., Koh, E.Y., Davy, S.K., et al. (2011) Phylogenetic analyses of bacteria in sea ice at Cape Hallett, Antarctica. N Z J Mar Freshwater Res 46: 3–12. McDonald, D., Price, M.N., Goodrich, J., Nawrocki, E.P., DeSantis, T.Z., Probst, A., et al. (2012) An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. ISME J 6: 610–618. Markus, T., Stroeve, J.C., and Miller, J. (2009) Recent changes in Arctic sea ice melt onset, freezeup, and melt season length. J Geophys Res Oceans 114: C12024. Martinez-Garcia, M., Brazel, D., Poulton, N.J., Swan, B.K., Gomez, M.L., Masland, D., et al. (2012) Unveiling in situ interactions between marine protists and bacteria through single cell sequencing. ISME J 6: 703–707. Meredith, M.P., and King, J.C. (2005) Rapid climate change in the ocean west of the Antarctic Peninsula during the second half of the 20th century. Geophys Res Lett 32: L19604. Mock, T. (2002) In situ primary production in young Antarctic sea ice. Hydrobiologia 470: 127–132.


Physicochemical control of sea ice microbial communities Montresor, M., Procaccini, G., and Stoecker, D.K. (1999) Polarella glacialis, gen. Nov., sp. Nov. (dinophyceae): Suessiaceae are still alive! J Phycol 35: 186–197. O’Brien, D.P. (1987) Direct observations of the behavior of Euphausia superba and Euphausia crystallorophias (Crustacea: Euphausiacea) under pack ice during the Antarctic Spring of 1985. J Crust Biol 7: 437–448. Paterson, H., and Laybourn-Parry, J. (2012) Sea ice microbial dynamics over an annual ice cycle in Prydz Bay, Antarctica. Polar Biol 35: 993–1002. Quast, C., Pruesse, E., Yilmaz, P., Gerken, J., Schweer, T., Yarza, P., et al. (2013a) The SILVA ribosomal RNA gene database project: improved data processing and webbased tools. Nucleic Acids Res 41: 590–596. Quast, C., Pruesse, E., Yilmaz, P., Gerken, J., Schweer, T., Yarza, P., et al. (2013b) The SILVA ribosomal RNA gene database project: improved data processing and webbased tools. Nucleic Acids Res 41: D590–D596. Reeder, J., and Knight, R. (2010) Rapidly denoising pyrosequencing amplicon reads by exploiting rankabundance distributions. Nat Methods 7: 668–669. Riaux-Gobin, C., Poulin, M., Dieckmann, G.S., Labrune, C., and Vétion, G. (2011) Spring phytoplankton onset after the ice break-up and sea-ice signature (Adélie Land, East Antarctica). Polar Res 30: article 5910. doi: 10.3402/ polar.v30i0.5910. Riemann, L., Steward, G.F., and Azam, F. (2000) Dynamics of bacterial community composition and activity during a mesocosm diatom bloom. Appl Environ Microbiol 66: 578– 587. Schmoker, C., Hernández-León, S., and Calbet, A. (2013) Microzooplankton grazing in the oceans: impacts, data variability, knowledge gaps and future directions. J Plankton Res 35: 691–706. Sharp, C.E., Brady, A.L., Sharp, G.H., Grasby, S.E., Stott, M.B., and Dunfield, P.F. (2014) Humboldt’s spa: microbial diversity is controlled by temperature in geothermal environments. ISME J 8: 1166–1174. Staley, J.T., and Gosink, J.J. (1999) Poles apart: biodiversity and biogeography of sea ice bacteria. Annu Rev Microbiol 53: 189–215. Stoeck, T., Bass, D., Nebel, M., Christen, R., Jones, M.D., Breiner, H.W., and Richards, T.A. (2010) Multiple marker parallel tag environmental DNA sequencing reveals a highly complex eukaryotic community in marine anoxic water. Mol Ecol 19: 21–31. Tamura, K., Stecher, G., Peterson, D., Filipski, A., and Kumar, S. (2013) MEGA6: molecular evolutionary genetics analysis version 6.0. Mol Biol Evol 30: 2725–2729.

13

Thomas, D.N., and Dieckmann, G.S. (2002) Antarctic sea ice – a habitat for extremophiles. Science 295: 641– 644. Torstensson, A., Chierici, M., and Wulff, A. (2012) The influence of increased temperature and carbon dioxide levels on the benthic/sea ice diatom Navicula directa. Polar Biol 35: 205–214. Torstensson, A., Hedblom, M., Andersson, J., Andersson, M.X., and Wulff, A. (2013) Synergism between elevated pCO2 and temperature on the Antarctic sea ice diatom Nitzschia lecointei. Biogeosciences 10: 6391–6401. Vancoppenolle, M., Meiners, K.M., Michel, C., Bopp, L., Brabant, F., Carnat, G., et al. (2013) Role of sea ice in global biogeochemical cycles: emerging views and challenges. Quat Sci Rev 79: 207–230. Wang, Q., Garrity, G.M., Tiedje, J.M., and Cole, J.R. (2007) Naïve Bayesian classifier for rapid assignment of rRNA Sequences into the new bacterial taxonomy. Appl Environ Microbiol 73: 5261–5267. Weissling, B.P., and Ackley, S.F. (2015) Spectral analysis of Amundsen Sea pack ice roughness and estimates of airice drag coefficient. Ann Glaciol 56 (in press). Wright, S., and Jeffrey, S. (1997) High-resolution HPLC system for chlorophylls and carotenoids of marine phytoplankton. In Phytoplankton Pigments in Oceanography. Jeffrey, S., Mantoura, R., and Wright, S. (eds). Paris: UNESCO, pp. 327–341. Yakimov, M.M., Guiliano, L., Gentile, G., Crisafi, E., Chernikova, T.N., Abraham, W.-R., et al. (2003) Oleispira antarctica gen. nov., sp. nov., a novel hydrocarbonoclastic marine bacterium isolated from Antarctic coastal sea water. Int J Syst Evol Microbiol 53: 779–785. Yu, Z.-C., Chen, X.-L., Shen, Q.-T., Zhao, D.-L., Tang, B.-L., Su, H.-N., et al. (2015) Filamentous phages prevalent in Pseudoalteromonas spp. confer properties advantageous to host survival in Arctic sea ice. ISME J 9: 871–881. Zhu, F., Massana, R., Not, F., Marie, D., and Vaulot, D. (2005) Mapping of picoeukaryotes in marine ecosystems with quantitative PCR of the 18S rRNA gene. FEMS Microbiol Ecol 52: 79–92.

Supporting information Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Fig. S1. Rarefaction curves. A. Bacteria normalized to 3149 sequences. B. Eukaryotes normalized to 2812 sequences.
