Microbial community structure of Arctic multiyear sea ice and surface seawater by 454 sequencing of the 16S RNA gene

Abstract

Dramatic decreases in the extent of Arctic multiyear ice (MYI) suggest this environment may disappear as early as 2100, replaced by ecologically different first-year ice. To better understand the implications of this loss on microbial biodiversity, we undertook a detailed census of the microbial community in MYI at two sites near the geographic North Pole using parallel tag sequencing of the 16S rRNA gene. Although the composition of the MYI microbial community has been characterized by previous studies, microbial community structure has not been. Although richness was lower in MYI than in underlying surface water, we found diversity to be comparable using the Simpson and Shannon's indices (for Simpson t=0.65, P=0.56; for Shannon t=0.25, P=0.84 for a Student's t-test of mean values). Cyanobacteria, comprising 6.8% of reads obtained from MYI, were observed for the first time in Arctic sea ice. In addition, several low-abundance clades not previously reported in sea ice were present, including the phylum TM7 and the classes Spartobacteria and Opitutae. Members of Coraliomargarita, a recently described genus of the class Opitutae, were present in sufficient numbers to suggest niche occupation within MYI.

Figure 1

Distribution of genetic distances for the MYI and seawater communities. The shaded region indicates the distribution of distances for the seawater community. The multi-modal distribution is a result of the dominance of a single genus (Pelagibacter) and the two-domain structure of the community. The white lateral line indicates the distribution of distances for the MYI community and the solid vertical line, the mean distance. The dashed vertical line shows the mean distance for the seawater community, excluding the domain Archaea.

Figure 2

Numerically dominant clades in the MYI and seawater communities. Gammaproteobacteria and Flavobacteria dominated the ice community; Alphaproteobacteria and Archaea, the seawater community. Deltaproteobacteria and Archaea comprised the smallest fractions of the MYI community; Opitutae, of the seawater community.

Figure 3

Divergence of the 16S rRNA gene using a taxonomy-based method. Divergence of the 16S rRNA gene (unique reads) with abundance (total reads) for seawater (top): r=0.96, P<10^–4, df=130 and for MYI (bottom): r=0.96, P<10^–4, df=117. The MYI community appears more divergent than the seawater community, with more new clades formed from increasing membership: the slopes of the lines are significantly different (Student's t-test, t=15.06, P<10^–4, df=247).

Figure 4

Divergence of the 16S rRNA gene using a taxonomy-independent analysis. The number of OTUs in the seawater community for varying definitions of OTU (0–20% difference in read sequence) was normalized to the lower richness observed in the MYI community. The accumulation of new OTUs at low percentage divergence appears faster for the seawater than sea ice community (a). Error bars indicate 95% confidence. The slope of a best-fit line (shown with log scale on both axis) for values from both environments (b) is significantly different than one (df=18, P=10⁻⁴).