Estelle S. Kilias

• This doctoral thesis aimed at the establishment of molecular tools (ARISA and 454-pyrosequencing) for protist diversity assessments in polar regions and at the application of these tools for studying protist diversity in the Fram Strait and in the Central Arctic Ocean. In this thesis, three hypotheses were put forward: i) Molecular surveys of genetic protist diversity, obtained by 454-pyrosequencing, constitute an adequate tool for assessing natural protist diversity ii) Water masses in the Arctic Ocean and in the Fram Strait are characterized by distinct protist communities and iii) Complex hydrographical and environmental situations can be evaluated via genetic information. A comprehensive study of protist diversity is required, because so far investigations were biased towards big size cells (≥2 µm). A correct identification of smaller cells is almost impossible due to the small size and lack of morphological markers. However, a proper survey of protist diversity demands the inclusion of all size classes. This is of particular relevance, considering the fact that small cells were observed to dominate protist assemblages at certain times, according to abiotic circumstances. The Arctic Ocean constitutes in two ways an important research area: on the one hand, it experiences intense variations in the light regime based on seasonality and sea ice; and on the other hand, it is assumed to be affected more severely by climate change than other world oceans. Since marine microorganisms are highly responsive to environmental forcing, changes will likely impact the protist community structures. Against the background of ongoing environmental changes in the Arctic, a study of protist diversity is further crucial in order to get a baseline for the assessment of future community structure changes. i) The assay of the 454-pyrosequencing suitability was carried out by using different wide-employed methods for reconciliation, sharing the same (clone library and ARISA) and different drawbacks (light microscopy and HPLC). The use of 18S rRNA clone library sequencing thereby, aimed to compare but also to complement the 454-pyrosequencing data, because of the longer sequence lengths that allow a more comprehensive taxon detailed analysis. However, the clone library approach was not suitable neither for 454-pyrosequencing comparison nor for 454-pyrosequencing complementation. While one approach, comprising the picoplankton fraction, was biased against haptophytes, the other approach, comprising the whole size fraction was biased against diatoms. Moreover, despite the comparable high number of ~140 clones per library, the abundant biosphere of 454-pyrosequencing was not comprehensively recovered. Hence, a consult of previous data for 454-pyrosequencing gathered by clone library is not advisable or has to be interpreted with caution. The second molecular method, ARISA, reflected community structure shifts that were indeed recovered by 454-pyrosequencing. The indicative limitation of ARISA on sequence length variances however, makes the method more suitable for a preceded sample selection than for a robust 454-pyrosequencing support. The assessment with traditional methods as light microscopy and HPLC presented good analogies. Since light microscopy is biased against small cells and HPLC against heterotrophic protists, the comparison was adapted to the respective limitations. A quantitative comparison of the diatom assemblage showed similar percentages within light microscopy and 454-pyrosequencing (≥10 µm filter), and approved the use of diatoms for the evaluation of 454-pyrosequencing accuracy. The comparison with HPLC was further in accordance with 454-pyrosequencing and agreed in the portions of autotrophic protists. In summary, three out of four evaluation methods presented good analogies with 454-pyrosequencing data and approved the suitability of the molecular method for assessing natural protist diversity. ii) To address the hypothesis if water masses host specific protist communities, 454-pyrosequencing was tested in different hydrographic environments for different size classes. In the process, the picoplankton community structure was investigated at four stations in the Eastern Fram Strait during the expedition ARK XXIV/2. The community structure of all protists was analyzed at five stations in the Western Fram Strait (ARK XXV/2) and at eight stations in the Central Arctic Ocean (ARK XXVI/3). A relation of water mass, distinguished by abiotic factors such as temperature, salinity and/or nutrients, and protist community structure was observed in all three studies. However, while the water mass regimes in the Fram Strait promoted distinct community structures in protist assemblages of the pico size spectrum and the entire size spectrum, populations of the Central Arctic Ocean showed a less definite association. In the Eastern Fram Strait picoplankton community (0.2-3 µm) in the Atlantic Water was mostly dominated by Phaeocystis cells and in the ice-covered station by small dinophytes (e.g. Dinophyte 1). The influence of cold, coastal water at one station however, shifted the community structure from a Phaeocystis-based to a Micromonas-based protist assemblage. In the Western Fram Strait, the protist community showed a distribution pattern that also corresponded to the different water properties. While Polar Water (EGC) was dominated by diatoms in recently light exposed stations and by dinoflagellates in a station longer exposed to light, the Atlantic Water (WSC) presented a high dominance of Micromonas that was not associated with the light availability. The two different protist communities observed in the Polar Water, however, suggest a high influence of light availability on the protist assemblage by promoting a protist succession. In the Central Arctic Ocean, in contrast, protist communities showed a less pronounced relation to the four water masses (Atlantic Water, Pacific Water, Mixed Water I and II) with no switch of dominant protist community members. All water masses were mainly dominated by dinoflagellates (e.g. Syndiniales 2), and at two stations by Micromonas. In principle, different water masses were reflected more significantly by whole community structure changes (ARISA) than by the appearance or disappearance of single protists (454-pyrosequencing) in the abundant biosphere. In particular, protist communities of the so-called Mixed Water I and II were difficult to separate from Atlantic Water and Pacific Water communities. The formation of mixed water, e.g. the mixing of two separate water masses, resulted in a combination of the abiotic characteristics (temperature, salinity, and nutrients) and of the protist communities. Moreover, the high sea-ice concentration and thus low light availability in the Central Arctic Ocean represented a strong selective force that eventually led to a unification of the protist communities. iii) 454-pyrosequencing revealed an adequate tool not just for investigating the protist diversity but also for reflecting hydrographical situations, as the recirculation of AW in the Fram Strait by protist community structure shifts. The hydrographic system of the Central Arctic Ocean, including Atlantic, Pacific, and mixed water masses, was not that strongly reflected by community structure shifts. The consistent presence of various ice concentrations strongly controlled the community composition and promoted heterotrophic and/or mixotrophic cells. In this regard, the Arctic protist assemblage presented high contributions of dinoflagellates under higher sea-ice concentrations (low light areas) and a higher contribution of diatoms under low sea-ice concentrations (high light areas). The strong response of the Arctic protist assemblage to the changing light conditions finally hampered the identification of water mass associated protist communities and hence, the reflection of the hydrographical situations. This thesis showed the suitability of 454-pyrosequencing for molecular studies of protist diversity and biogeography, independent of organisms cell size, or organisms nutritional strategy (hetero-, mixo-, or autotrophy). The application of 454-pyrosequencing, to study Arctic protist distribution facilitated to reveal even complex hydrographical situations and indicated the presence of different microbial habitats in polar regions, determined by water mass properties (T, S, and nutrients) and altered by sea ice concentration. This hypothesizes an improved differentiation of the microbial habitats in the Central Arctic Ocean under continuously sea ice decrease. Moreover, the study of the rare biosphere revealed a constant distribution of taxonomic groups. In contrast to the abundant biosphere, the rare biosphere did not respond to changing nutrient or sea ice concentrations.