Polar Biology

, Volume 29, Issue 8, pp 662–667

Dense populations of Archaea associated with the demosponge Tentorium semisuberites Schmidt, 1870 from Arctic deep-waters


    • Institute of Biogeochemistry and Marine ChemistryUniversity of Hamburg
    • Research Center Ocean MarginsUniversity of Bremen
  • Friederike Hoffmann
    • Center of Geosciences, GeobiologyUniversity of Göttingen
    • Max Planck Institute for Marine Microbiology
  • Nadia-Valérie Quéric
    • Alfred-Wegener-Institute for Polar and Marine Research, Deep Sea Research
  • Karen von Juterzenka
    • Alfred-Wegener-Institute for Polar and Marine Research, Deep Sea Research
    • Institute for Polar EcologyChristian-Albrechts-University of Kiel
  • Joachim Reitner
    • Center of Geosciences, GeobiologyUniversity of Göttingen
  • Walter Michaelis
    • Institute of Biogeochemistry and Marine ChemistryUniversity of Hamburg
Original Paper

DOI: 10.1007/s00300-005-0103-4

Cite this article as:
Pape, T., Hoffmann, F., Quéric, N. et al. Polar Biol (2006) 29: 662. doi:10.1007/s00300-005-0103-4


The associated microbial community in the mesohyl of the Arctic deep-water sponge Tentorium semisuberites Schmidt, 1870 (Hadromerida, Demospongiae) is dominated by Archaea. This is the result of an integral approach applying analyses of microbial lipid biomarkers as well as microscopic investigations using differential fluorescence in situ hybridisation with universal probes and counterstaining with 4′,6′-diamidino-2-phenylindole (DAPI) on sponge sections based on samples collected in the Greenland Sea in 2001, 2002 and 2005. The distribution of isoprenoidal C40 hydrocarbons of the biphytane series suggests that affiliates of both major archaeal kingdoms, the Crenarchaeota and the Euryarchaeota, are present in the choanosome of T. semisuberites. Positive signals using the oligonucleotide probe ARCH915 indicate high numbers of Archaea in the mesohyl of this sponge. Based on optical estimations 70–90% of all microbial DAPI signals accounted for archaeal cells. Archaea in these high proportions have never been described in an Arctic deep-sea hadromerid sponge, nor in any other demosponge species. Similar observations in specimens collected over a time scale of 4 years suggest permanent sponge-Archaea associations.


Over wide areas of the Arctic deep-sea, sponges represent the majority of benthic biomass, even though the diversity is low (Barthel and Tendal 1993). The sediments of the deep Arctic basins are populated by species of Demospongiae, Calcarea and Hexactinellida. Tentorium semisuberites Schmidt, 1870 (Hadromerida, Demospongiae) represents one of eight core species of the typical Norwegian and Greenland Sea abyssal sponge-association (Barthel and Tendal 1993) and it is a widely distributed element of polar suspension feeder communities at deep slopes, sea mounts and rises (Kröncke 1998; own observations). In the absence of appropriate hard substrates, which is a common feature for the central Greenland Sea, sponges are partly buried and anchor in the sediment with root-like structures (Barthel and Tendal 1993; own observations).

Numerous sponges host large amounts of microorganisms in their mesohyl, which can amount up to 50% or more of the sponge biomass (Vacelet 1975; Reitner 1993). Development of molecular and biochemical techniques has opened new possibilities to explore these sponge microbial communities. For numerous bacteriosponges, the microbial community is species-specific, including Cyanobacteria (Wilkinson 1983) and diverse heterotrophic Bacteria (Manz et al. 2000; Hentschel et al. 2002). Up to now, indications for Archaea as endobionts were reported for several Demospongiae and Hexactinellida (Preston et al. 1996; Webster et al. 2001, 2004; Margot et al. 2002; Pape et al. 2004). However, since sponge-associated Archaea eluded laboratory cultivation so far, their metabolic properties and ecological functions remain to be determined.

In our previous studies on sponge-associated microorganisms, lipid biomarker analysis proved to be a rapid and effective tool to trace archaeal membrane constituents incorporated in sponge tissues (Thiel et al. 2002; Pape et al. 2004). Here, we show a strongly Archaea-dominated microbial association of the Arctic deep-water sponge T. semisuberites using lipid biomarker analysis and fluorescence in situ hybridisation (FISH).



Samples were collected in 2001 and 2005 from the French research vessel “L′Atalante” using pushcores or a slurp gun operated by the ROV “Victor” and in 2002 from the German research vessel “Polarstern” by employing an Agassiz Trawl. Sampling sites were situated in the Fram Strait on the continental rise off Svalbard (78°36′N; 05°04′E at 2,310 m depth in 2001, 79°06′N; 04°22′E at 2,340 m depth in 2002, 79°07′N; 04°13 at 2,440 m depth in 2005, respectively). For biomarker analysis, sponges were immediately frozen and stored at −20°C. Other specimens were fixed either in 2% formaldehyde/0.04% glutaraldehyde in filter-sterilized seawater or in 4% formaldehyde (Manz et al. 2000) and subsequently dehydrated in an increasing ethanol concentration series for FISH. Samples were stored in 70% EtOH for 18 months (2001), 6 months (2002) and 1 week (2005).

Lipid biomarker analysis of sponge tissues

For lipid biomarker analysis about 20 specimens sampled in 2002 were thoroughly rinsed with distilled water. Sediment particles were removed from the tissue surfaces with tweezers and by centrifugation. The sponge bodies were pooled and lyophilized and, subsequently, the dry weight (DW) of the tissues was determined as a whole. The sponge bodies were crushed with a blender and successively extracted with solvents of decreasing polarity (methanol, methanol/dichloromethane [1:1; v:v] and dichloromethane) by means of ultrasonification. The organic extracts were combined and treated with HI in order to cleave ether bonds of archaeal polar lipids. The alkyl iodides generated were transformed to their deuterated compound analogues using LiAlD4 in dry tetrahydrofurane under an Ar-atmosphere.

Compounds were analysed by gas chromatography and gas chromatography-mass spectrometry (Pape et al. 2005). Structural elucidation was conducted on a Micromass Quattro II mass spectrometer interfaced with a HP6890 gas chromatograph. Identification of deuterated biphytanes was achieved by comparison of mass spectra and retention characteristics with published data of non-deuterated biphytanes (DeLong et al. 1998) and by co-injection with an authentic standard mixture (kind donation of P. Adam, Université Louis Pasteur, Strasbourg). Quantification of individual biphytanes was achieved by adding tetratriacontane (n-C34) as internal standard of a known concentration prior to GC analysis.

Fluorescence in situ hybridisation on sponge sections

Longitudinal 5-μm sections of whole sponges were prepared from six specimens (two from each sampling campaign) embedded in paraffin (Hoffmann et al. 2003). FISH was performed on 4–10 sponge sections per specimen (Manz et al. 2000). Entire sections, including all parts of the sponge body, were investigated. The following oligonucleotide probes labelled with Cy3, Oregon Green or FLUOS (MetaBion, Germany) were used to identify sponge microbes and their spatial distribution within the sponge: Probe ARCH915 targeting Archaea (Stahl and Amann 1991) and probe EUB I-III (Daims et al. 1999) targeting Bacteria. Formamide concentrations in hybridisation solution and wash buffer were 35% v:v to test for specificity, and 20% v:v to give optimal brightness. To detect accidental binding of probe ARCH915 to Bacteria (Pernthaler et al. 2002), these probes were also applied simultaneously with different labels. All FISH preparations were counterstained with 1 μg ml−1 4′,6′-diamidino-2-phenylindole (DAPI). Results were always corrected by subtracting signals observed with probe NON-EUB338 (Amann et al. 1990). Samples were viewed under a Zeiss axioplan microscope (Zeiss, Oberkochen, Germany). Micrographs were taken using a Zeiss Axiocam MRC camera and processed with AxioVision and ADOBE Photoshop software.

Results and discussion

After cleavage of ether bonds, four isoprenoidal C40 hydrocarbons of the biphytane series were detected in the total organic tissue extract of T. semisuberites. The biphytanes possessed none to three internal cycloalkyl rings (Fig. 1 Top). The sites of ether bonding which link the biphytanes to glycerol moieties are determined by the deuteration experiment with LiAlD4. As exemplified by the biphytane with three alkylrings (Fig. 1 Bottom), the molecular ions (M+) and characteristic fragment ions (m/z) of deuterated biphytanes show higher values of 2 and 1 atomic mass units, respectively, compared to those of non-deuterated biphytanes. By labelling the cleaving positions, an ether bonding of the biphytanes as alkyldiethers to Archaea-derived core membrane lipids (Chappe et al. 1980) is shown.
Fig. 1

Top: Molecular structures of ether-released isoprenoidal C40 hydrocarbons detected in Tentorium semisuberites. Structures of the acylic C40:0 and the internally cyclized C40:1, C40:2 and C40:3 are shown. Bottom: Mass spectrum and inferred fragmentation pattern of the dideuterated biphytane with three cycloalkyl rings (C40:3, Crenarchaeol). For comparison the mass spectrum of the non-deuterated C40:3 obtained by analysis of an authentic standard is given. Diagnostic ion fragments comprising terminal molecular units, e.g. m/z=165/166/167 and m/z=263, are shifted by one mass unit due to the deuterium incorporation. Since the ion fragment m/z=193 is not affected by the deuterium labelling it most likely derives from the inner part of the molecule

The distribution pattern of the biphytanes in T. semisuberites (Fig. 2) concurs with previous observations on several sponges of the classes Hexactinellida and Demospongiae (Pape et al. 2004). The series is dominated by the acyclic biphytane (C40:0) followed by the bicyclic C40:2 (19.0% relative to the abundance of C40:0). The monocylic C40:1 was the least abundant biphytane in the tissue (9.2% relative to the abundance of C40:0). With regard to phylogenetic affiliations the vast majority of marine Archaea appear to belong to either the Crenarchaeota or the Euryarchaeota. As indicated by its mass spectral characteristics, the tricyclic biphytane is a derivative of crenarchaeol—a characteristic feature for representatives of the Crenarchaeota subgroup (Sinninghe Damsté et al. 2002). However, compared to biphytane distributions reported for enrichments of the sponge-associated crenarchaeote Cenarchaeum symbiosum (DeLong et al. 1998; Fig. 2), the acyclic C40:0 was found in relatively higher amounts. High concentrations of this compound are frequently observed in pure cultures of methanogenic euryarchaeotes (Koga et al. 1993). Therefore, the biphytane distribution pattern in T. semisuberites may indicate the presence of sponge-associated methanogenic euryarchaeotes.
Fig. 2

Distribution patterns (left) of ether-released isoprenoidal C40 hydrocarbons detected in T. semisuberites. The distributions are given as relative abundance to C40:0. For comparison the relative abundances of biphytanes in the sponge symbiont Cenarchaeum symbiosum (Crenarchaeota) are shown (right, data adapted from DeLong et al. 1998)

For a comparison of archaeal cell densities previously estimated on biphytane concentrations in tissues of several hexactinellids and demosponges (Pape et al. 2004), biphytane concentrations were calculated. For T. semisuberites, total biphytanes yielded concentrations of about 66.4 μg per g DW sponge matter and were much higher than in the hexactinellid Aulosaccus mitsukuri which contained 2.4 μg total biphytanes per g DW sponge matter (Pape et al. 2004). To the best of our knowledge, such high concentrations of archaeal lipids associated to sponge tissues have not been reported so far.

The proposed high archaeal cell densities associated with the tissues of T. semisuberites are supported by our differential FISH analyses (Fig. 3a–c). DAPI counterstaining on all sponge sections showed high densities (Fig. 3b) of associated microorganisms in the entire mesohyl of T. semisuberites. As obtained by repeated optical estimation, 70–90% of all microbial DAPI signals corresponded to those obtained with the molecular probe ARCH915 targeting Archaea (Fig. 3a–c). Surprisingly, most of the remaining non-eukaryotic DAPI signals in the sponge mesohyl were not targeted by the bacterial probe EUB I-III. Positive EUB signals were sparse (<1% of DAPI), and did not overlap with the Archaea signals. A drawback of FISH with fluorescently monolabelled oligonucleotide probes is that some slowly-growing microorganisms with small cell sizes are not viewed by this approach and hence underestimated (Pernthaler et al. 2002). The identity of the remaining microbial DAPI signals, which are neither targeted by the probes for Archaea nor Bacteria is therefore still uncertain. Signals with ARCH915 and corresponding DAPI signals were observed in the dense mesohyl between spicule tracts throughout the entire sponge body. Variations by means of cell density occurred within the sections, but not between samples.
Fig. 3

Epifluorescence micrographs of identical tissue sections of T. semisuberites. Numerous sponge-associated Archaea (small signals, indicated by arrows) are visualized by fluorescence in situ hybridisation (FISH) with the Cy3-labelled oligonucleotide probe ARCH915 (a). With DAPI staining (b), nuclei of sponge cells (ce) are visible in addition to the microbial DAPI signals in the tissue surrounding the canals (ca). Image overlay (c) shows that a high percentage of microbial DAPI signals corresponds to the Archaea signals. Bar=10 μm

Results from our biomarker studies in combination with microscopic investigation using FISH on several specimens collected in different years indicate a microbial community dominated by Archaea. Archaeal associates have only been detected in a few demosponge species so far, affiliated with the orders Hadromerida, Poecilosclerida, Halichondrida or Dictyoceratida (Preston et al. 1996; Webster et al. 2001, 2004; Margot et al. 2002). A more widespread occurrence of associated Archaea in both Demospongiae and Hexactinellida and a comparatively higher affinity of hexactinellid sponges to Archaea were demonstrated using lipid biomarker analysis (Thiel et al. 2002; Pape et al. 2004).

Archaea associated with demosponges were reported to co-occur with bacterial communities (Preston et al. 1996; Webster et al. 2004; Pape et al. 2004). The combination of molecular with microscopic approaches revealed the accumulation of Archaea in distinct regions of the sponge body such as the pinacoderm (Webster et al. 2001) or collagen fibres (Margot et al. 2002). We present the first description of high proportions of Archaea inhabiting the entire mesohyl. In studies on archaeal gene sequences retrieved from sponge tissues, crenarchaeotal sequences are more frequently reported than those derived from euryarchaeotes (Preston et al. 1996; Margot et al. 2002; Webster et al. 2004). Members of Crenarchaeota and Euryarchaeota were observed in oxic marine environments, e.g. as part of the picoplankton (Pernthaler et al. 2002), but also from anoxic settings (Madrid et al. 2001). Webster et al. (2001) detected gene sequences affiliated with methanogen euryarchaeotes in the tissue of a dictyoceratid demosponge, and proposed the presence of anoxic microniches within this species. Suboxic and anoxic niches have been detected in some sponge species with a dense mesohyl structure (Hoffmann et al. 2005a, b), and may also occur in T. semisuberites. Thus, the relatively high abundance of acyclic C40:0 suggests the presence of euryarchaeotal, possibly methanogenic associates.

Archaea are uncommon partners for invertebrate symbioses and the ecological function of Archaea in T. semisuberites is unknown. Witte (1995) observed a conspicuously low amount of choanocyte chambers in T. semisuberites, which is consistent with our microscopic observation of choanocyte chambers being small (20 μm diameter) and sparse. Reduced choanocyte chambers may imply the possible importance of “microbial farming”, i.e. that this species rather depends on its associated microbes than on food particles obtained by filtration. This kind of biomass transfer has been described for various sponge species associated with heterotrophic and autotrophic microbes (Ilan and Abelson 1995; Vacelet et al. 1996).

The bacterial and archaeal community in sediments near T. semisuberites appeared to be more complex in composition compared to adjacent benthic microhabitats (Quéric, unpublished data), indicating a direct influence of sponges on the microbial deep-sea communities. Ongoing investigation will elucidate the phylogenetic affiliation of the Archaea associated with T. semisuberites as well as their significance for the metabolism of the sponge-microbe association. Comparative phylogenetic studies will focus on the relation of the associated Archaea in T. semisuberites to Archaea from the surrounding sediments, to address their impact on Arctic deep-sea benthic communities in general.


Sampling took place during annual revisits to the long term monitoring station “AWI-Hausgarten”. We thank the crews of the R/Vs “Polarstern” and “L′Atalante” for assistance with sampling, Sabine Beckmann for analytical work and Wolfgang Dröse as well as Angela Scharfbillig for preparation of sponge sections. Three anonymous reviewers have greatly improved the quality of the original manuscript. Andrew K. Sweetman is kindly acknowledged for proofreading of the manuscript. This paper represents publication no. 54 of the research program BOSMAN (03F0358 A and C). Financial support was provided by the Bundesministerium für Bildung und Forschung (BMBF), Germany. F.H. was funded by the Max Planck Society and by the Deutsche Forschungsgemeinschaft (DFG – Project HO 3293/1–1) during preparation of the manuscript.

Copyright information

© Springer-Verlag 2006