A Metataxonomic Approach Reveals Diversified Bacterial Communities in Antarctic Sponges

Marine sponges commonly host a repertoire of bacterial-associated organisms, which significantly contribute to their health and survival by producing several anti-predatory molecules. Many of these compounds are produced by sponge-associated bacteria and represent an incredible source of novel bioactive metabolites with biotechnological relevance. Although most investigations are focused on tropical and temperate species, to date, few studies have described the composition of microbiota hosted by Antarctic sponges and the secondary metabolites that they produce. The investigation was conducted on four sponges collected from two different sites in the framework of the XXXIV Italian National Antarctic Research Program (PNRA) in November–December 2018. Collected species were characterized as Mycale (Oxymycale) acerata, Haliclona (Rhizoniera) dancoi, Hemigellius pilosus and Microxina sarai by morphological analysis of spicules and amplification of four molecular markers. Metataxonomic analysis of these four Antarctic sponges revealed a considerable abundance of Amplicon Sequence Variants (ASVs) belonging to the phyla Proteobacteria, Bacteroidetes, Actinobacteria and Verrucomicrobia. In particular, M. (Oxymycale) acerata, displayed several genera of great interest, such as Endozoicomonas, Rubritalea, Ulvibacter, Fulvivirga and Colwellia. On the other hand, the sponges H. pilosus and H. (Rhizoniera) dancoi hosted bacteria belonging to the genera Pseudhongella, Roseobacter and Bdellovibrio, whereas M. sarai was the sole species showing some strains affiliated to the genus Polaribacter. Considering that most of the bacteria identified in the present study are known to produce valuable secondary metabolites, the four Antarctic sponges could be proposed as potential tools for the discovery of novel pharmacologically active compounds.


Introduction
The Antarctic region comprises ice shelves, waters and all the island territories in the Southern Ocean, covering about 10% of the total world ocean's area. The Antarctic sarai (Calcinai and Pansini, 2000), and Haliclona (Rhizoniera) dancoi (Topsent, 1901), and Poecilosclerida with one species, Mycale (Oxymycale) acerata (Kirkpatrick, 1907) ( Table 1). BLAST similarity search totally agreed with the morphological identification obtained for B4 and D4 samples. Molecular analysis confirmed B4 species as M. (Oxymycale) acerata, with CO1 primers that were the most specific (98% of pairwise identity) in comparison to 18S, 28S and ITS molecular markers. Similarly, CO1 also appeared to be the best molecular marker for the identification of the sponge D4, with a highest sequence similarity to H. pilosus (98% sequence identity). Regarding sample C6, molecular markers identified the genus corresponding to Haliclona, with the most striking result achieved using the 18S marker (92% similarity to Haliclona sp.). Unfortunately, it was not possible to identify this sponge at the species level, because there are no other available sequences on GenBank for H. (Rhizoniera) dancoi. Similarly, the results achieved with sample D6 were partially unclear, since several genera at low-sequence similarity were observed from BLAST outputs. In fact, the sequences of M. sarai, identified by spicule observations, are still not uploaded in GenBank (see Tables S1-S4; details on the alignments are reported in Figures S1-S3).
In addition, M. (Oxymycale) acerata, H. pilosus and H. (Rhizoniera) dancoi revealed an abundance of both Deltaproteobacteria and Acidimicrobiia. Manhattan algorithm indicated that M. sarai clustered separately in comparison to the others, with H. pilosus and H. (Rhizoniera) dancoi resulting as the most similar in terms of species structure and abundance ( Figure 1).

Discussion
In the present study, we analyzed the species composition and abundance of the associated microbiota from four Antarctic sponges, M. (Oxymycale) acerata, H. (Rhizoniera) dancoi, H. pilosus and M. sarai, collected from Tethys Bay (Victoria Land, Antarctica). In particular, the associated community of M. (Oxymycale) acerata, collected from site 1 ( Table 1), was similar to H. (Rhizoniera) dancoi and H. pilosus (Figure 1), retrieved from site 2 (Table 1). Interestingly, at this latter site, we collected the sponge M. sarai, whose species abundance was found statistically different by Manhattan clustering analysis ( Figure 1).
(Oxymycale) acerata and other Antarctic species by metagenomic approaches [23,27,28,[30][31][32]48]. These bacteria were frequently found to be the dominant bacterial phyla in marine ecosystems [49]. In particular, Proteobacteria showed different functions in host, including nitrogen fixation, and were involved in host defense mechanisms [50]. Furthermore, some bacteria were described as highly specialized hydrocarbon degrading microorganisms [51,52] and their wide distribution may be due to a strong positive interaction in environments where bacteria represent a fundamental source of nutrients, such as the case of Antarctica. This finding could be corroborated by results revealing that these bacteria are able to adapt to extreme environments, including polar habitats [53][54][55][56]. Concerning their biotechnological potential, genome-mining approaches reported several biosynthetic gene clusters (BGCs) encoding for bioactive molecules from marine Proteobacteria (reviewed by Buijs et al. [57]). However, there is no direct 100% correlation between the presence of a certain BGC, a bacterial genus and a bioactive metabolite. BGCs can be silent in certain conditions and, hence, methods should be developed to unlock their silent potential [58], to observe the production of a particular compound and induce the desired bioactivity. The most common approach known to discover new metabolites is the "OSMAC" (one strain many compounds) approach. The term OSMAC was coined for the first time by Zeeck and co-workers [59], indicating the ability of single strains to produce different metabolites when cultivated under different conditions. Examples are the use of different culturing strategies to trigger the production of secondary metabolites such as changing culturing conditions (e.g., nutrients or light exposure), mimicking environmental stressors and co-culturing with other species.
On the whole, several species belonging to Gammaproteobacteria and Alphaproteobacteria isolated from sponges and soils showed antibacterial, antiviral, antifungal and antiprotozoal activities that make them suitable tools in drug discovery research fields [36,43,[60][61][62][63]. In particular, the Gammaproteobacteria of the genus Endozoicomonas, identified from M. (Oxymycale) acerata in the present work ( Figure 2; Table S5), was found to induce antimicrobial activities [64,65].
Always M. (Oxymycale) acerata showed a relative abundance of a Gammaproteobacteria belonging to the genus Colwellia ( Figure 2; Table S5), which is extremely interesting since it was recently proposed as a useful tool for the bioremediation of nitrogen pollutants [66]. Previous investigations also demonstrated that a sponge-associated Colwellia sp. produces several extracellular polymeric substances (EPSs) with potential use in the production of cosmeceutical and nutraceutical ingredients [35,67].
M. (Oxymycale) acerata also revealed some bacterial strains classified as Fulvivirga and Ulvibacter, two genera included into Bacteroidetes, the second most abundant phylum found in the samples under analysis (Figure 2; Table S5). Genome-mining approaches coupled to chemical analyses revealed the presence of some amine acylated desferrioxamine siderophores from Fulvivirga sp. with anticancer properties [68]. Similarly, Ulvibacter species, already observed in Antarctic habitats [69], belong to the family Flavobacteriaceae, whose biotechnological applications are well-documented. In fact, several polysaccharidedigesting enzymes together with antibiotics and other bioactive compounds, such as quercetin (known for its antioxidant, anti-inflammatory, chemopreventive properties), were isolated [70,71].
ASV's data showed bacteria of the family Rhodobacteraceae (class Alphaproteobacteria) in H. (Rhizoniera) dancoi and H. pilosus (Figure 2; Tables S6 and S7). Several genera were recognized as a huge source of novel bioactives, especially Pseudovibrio species living in seawater and through symbiotic relationships with sponges, tunicates and corals [85,86]. For example, H. pilosus specifically hosted the genus Roseobacter, which was also studied for its antimicrobial properties [87,88]. The hydrocarbon-degrading Gammaproteobacteria of the genus Pseudohongiella, with potential use in the bioremediation of anthropogenic contaminants [89,90], were also revealed in H. (Rhizoniera) dancoi and H. pilosus (Figure 2; Tables S6 and S7).
Less abundant members living within H. pilosus and H. (Rhizoniera) dancoi belonged to the classes Nitrospinia (phylum Nitrospinae) and Nitrosophaeria (phylum Thaumarchaeota) (Figure 2; Tables S6 and S7). Recent investigations already reported low percentages of Nitrospinia from H. pilosus, H. (Rhizoniera) dancoi and other Antarctic species [26,29]. Concerning the capability to produce molecules with biotechnological potential, very little information is available so far. In a recent study, BLASTp search against the Integrated Microbial Genomes (IMG) database identified a Pseudoalteromonas luteoviolacea gene encoding for a L-amino acid oxidase (LAAO) with antimicrobial properties in the genome of a strain belonging to the phylum Nitrospinae [91].
ASV's analysis of the three sponges, M. (Oxymycale) acerata, H. pilosus and H. (Rhizoniera) dancoi, displayed Deltaproteobacteria (Figure 2; Tables S5-S7) belonging to the phylum Proteobacteria, that, as mentioned above, produce interesting bioactive metabolites [57]. For instance, H. pilosus exhibited some strains included into the genus Bdellovibrio ( Figure 3; Table S7), which is an obligate predator of other Gram-negative bacteria that was proposed for possible biotechnological applications toward medicinal, agricultural and industrial fields [113][114][115].  Table S8). Some data demonstrated that these cold-adapted bacteria produced interesting EPSs molecules with protective effects on human skin and anti-aging properties [116,117].

Sponge Collection
Four sponge samples, reported as B4, C6, D4 and D6, were collected by scuba divers in November-December 2018 at two sites of  (Figure 3). Samples were immediately washed at least three times with filter-sterilized natural seawater to remove transient and loosely attached bacteria and/or debris [14,27]. Firstly, a small fragment of each sponge was preserved in 70% ethanol for taxonomic identification. Specimens were then placed into individual sterile tubes and kept in RNAlater © at −20 • C until transported to the Stazione Zoologica Anton Dohrn (Naples, Italy).
Sponge slides of spicules are deposited at the Italian National Antarctic Museum (MNA, Section of Genoa, Italy). The MNA voucher codes of the sponges investigated in the present work are reported in Table 1.

Morphological Analysis of Spicules
The taxonomic identification was conducted at the species level. Small fragments of each sponge were heat-dissolved in nitric acid, rinsed in water and dehydrated in ethanol. Then, spicules were mounted on slides for microscopic analyses, following standard methods [118]. The skeletal architecture was examined under light microscope and hand-cut sections of sponge portions were made as described in Hooper [119].
The taxonomic classification follows the updated nomenclature reported in the World Porifera Database (WPD) [120].

DNA Extraction and PCR Amplification
About 10 mg of tissue were used for DNA extraction by using QIAamp ® DNA Micro kit (QIAGEN), according to the manufacturer's instructions. DNA quantity (ng/µL) was evaluated by NanoDrop spectrophotometer. PCR reactions were performed on the C1000 Touch Thermal Cycler (BioRad) in a 30 µL reaction mixture final volume including about 50-100 ng of genomic DNA, 6 µL of 5X Buffer GL (GeneSpin Srl, Milan, Italy), 3 µL of dNTPs (2 mM each), 2 µL of each forward and reverse primer (25 pmol/µL, Table 1 Sequences of PCR primers are reported in Supporting Information (Table S9). PCR products were run on 1.5% agarose gel and the fragment length was evaluated by using 100 bp DNA ladder (GeneSpin Srl, Milan, Italy). PCR products were purified using QIAquick Gel Extraction Kit (Qiagen), according to the manufacturer's instructions. PCR amplicons were then sequenced in both strands through Applied Biosystems (Life Technologies) 3730 Analyzer (48 capillaries). Sequences produced were~300-650 bases long in average with more than 97.5% accuracy, starting from PCR fragments. The total 18S, 28S, ITS and CO1 region were submitted to GenBank using Basic Local Alignment Search Tool (BLAST) [125] and then aligned with highly similar sequences using MultiAlin (http://multalin.toulouse.inra.fr/multalin/ accessed on 29 January 2021) [126].

Metagenomic DNA Extraction and Illumina MiSeq Sequencing
Genomic DNA for 16S rRNA sequencing was performed from about 250 mg of tissue by using DNeasy ® PowerSoil ® Pro Kit (QIAGEN), according to the manufacturer's instructions. Extractions were performed using both internal and external sponge tissue in order to obtain the whole bacterial community. DNA quantity (ng/µL) and quality (A260/280, A260/230) were evaluated by NanoDrop spectrophotometer, whereas DNA integrity was checked on 0.8% agarose gel electrophoresis in TAE buffer (40 mM Trisacetate, 1 mM EDTA, pH 8.0). 20 µL of samples (30 ng/µL final concentration) were subjected to 16S V3-V4 rRNA gene library preparation and sequencing (Bio-Fab Research, Rome, Italy). Illumina adapters overhang nucleotide sequences were added to the gene specific primer sequences targeting the V3-V4 region [127]. After 16S amplification, a PCR clean-up was done to purify the V3-V4 amplicon from free primers and primerdimer species. A subsequent limited cycle amplification step was performed to add multiplexing indices and Illumina sequencing adapters by using a Nextera XT Index Kit. Finally, libraries were normalized and pooled by denoising processes (Table S10), and sequenced on Illumina MiSeq Platform with 2x300 bp paired-end reads. Taxonomy was assigned using "home made" Naive Bayesian Classifier trained on V3-V4 16S sequences of SILVA 132 database [128]. QIIME 2 (Quantitative Insights Into Microbial Ecology) platform [129] was used for microbiome analysis from raw DNA sequencing data. QIIME analysis workflow was performed by demultiplexing, quality filtering, chimera removal and taxonomic assignment. The full dataset of raw data has been deposited in the SRA database (submission ID: SUB8701897; BioProject ID: PRJNA687362).

Statistical Analysis
ASVs distribution and frequency in the whole dataset and for each sample are reported in the Supporting Information (Figures S5 and S6).
Heatmap was generated by using Heatmapper Software available at http://www. heatmapper.ca/ accessed on 29 January 2021 [130]. The number of features observed for each identified taxa were normalized as Log10 and scaled by column. Hierarchical clustering was applied on both rows and columns by average linkage method. For computing distance between rows and columns, Manhattan distance measurement algorithm was performed.

Conclusions
Our metataxonomic analysis highlights the occurrence of dominant and locally enriched microbes in the Antarctic sponges M. (Oxymycale) acerata, H. (Rhizoniera) dancoi, H. pilosus and M. sarai, characterized by morphological and molecular approaches. This can be considered a starting point in the understanding of the global Antarctic microbiome in a more complete perspective, given the scarce information in the literature for extreme environments such as the Antarctica. According to the microbial community identified, the biotechnological value should not be underestimated. In fact, our findings open new perspectives concerning the possible role of these Antarctic sponges and their symbiotic bacteria as a source of bioactive compounds. Further studies will be devoted to bioassay-guided fractionations for identifying new potential drugs useful in pharmaceutical, nutraceutical and cosmeceutical fields.