Introduction

The Planctomycetes comprise one phylum of the PVC superphylum that also contains the phyla Verrucomicrobia, Chlamydiae, and Lentisphaerae (Wagner and Horn 2006). These remarkable bacteria are generally characterized by budding reproduction, membrane-bounded compartments within the cells, and absence of peptidoglycan in their proteinaceous cell walls (Ward et al. 2006). Other cellular eukaryote-like features, presence of endocytosis-like protein uptake, and membrane coat-like proteins (Lonhienne et al. 2010; Santarella-Mellwig et al. 2010) were observed in one genus evidencing the evolutionary relevance of this group.

Molecular ecology and culture studies have revealed the widespread distribution of Planctomycetes in a broad range of terrestrial and aquatic environments and in association with several organisms namely sponges, the giant tiger prawn, corals, macroalgae, and the human gut (Nelson et al. 2010; and see Fuerst and Sagulenko 2011 for a review). Higher representation of Planctomycetes was observed in marine sediments (Rusch et al. 2003; Musat et al. 2006; Chipman et al. 2010; Hu et al. 2010) and on the biofilm of the kelp Laminaria hyperborea (Bengtsson and Ovreas 2010; Bengtsson et al. 2010) and Fucus vesiculosus (Lachnit et al. 2011) when compared with coastal waters or the marine pelagic environment (Rusch et al. 2007). But the highest density of Planctomycetes in natural environments was found in the hindgut of soil-feeding termites Cubitermes spp., where they reach concentrations of 2.6 × 109 cells mL−1 (Kohler et al. 2008).

Planctomycetes isolates have been obtained from oligotrophic to highly polluted environments suggesting a broad metabolic adaptation. They are believed to play an important role in the global carbon and nitrogen cycles (Strous et al. 1999; Glockner et al. 2003) through, for example, the mineralization of marine snow particles (DeLong et al. 1993) and the anaerobic ammonia oxidation—anammox (Strous et al. 1999). Furthermore, the unexpected presence of key genes for 110 sulfatases in Rhodopirellula baltica (Glockner et al. 2003) suggests their involvement in the transformation of inorganic sulfur. Up-to-now, the majority of the known Planctomycetes strains is chemoheterotrophic or chemoautotrophic (anammox). Despite the information gathered in the last years, the physiology and metabolism of Planctomycetes are still not well understood (Ward et al. 2006), and an effort to improve the number of new cultivable strains is important. Improvements in the isolation methodology of Planctomycetes have been achieved (Schlesner 1994; Zengler et al. 2002; Winkelmann and Harder 2009; Lage and Bondoso 2011).

In the last decade, the great diversity of the microbial communities associated with aquaculture systems have been revealed using molecular tools (Tal et al. 2003; Cytryn et al. 2005; Sugita et al. 2005; Michaud et al. 2009; Gao et al. 2012). Although not the most frequent, Planctomycetes seem to be usual inhabitants of aquaculture environments (Sugita et al. 2005; Bissett et al. 2006; Kawahara et al. 2009; Michaud et al. 2009; Gao et al. 2012). In this study, we describe the isolation of Planctomycetes from the sediment of a treatment water recycling tank of a marine fish farm.

Materials and methods

Isolation, cultivation, and maintenance

Samples were collected from a fish farm located in the north region of Portugal (41°27′10″N and 8°46′27″W) nearby the sea. The fish tanks used for the growth of turbot, Scophthalmus maximus, European sea bass, Dicentrarchus labrax, and sole, Solea senegalensis, were feed with a mixture of seawater captured directly from the sea and the water coming from the treatment water recycling tanks. The water coming from the fish tanks goes through sedimentation and nitrification in the treatment water recycling tank where samples were collected. The dimensions of the tank are 10 × 25 × 1.2 m. The temperature of the water varied between 7 and 15 °C (mean 11 °C), and the pH was 8. The samples (water and sediments, n = 1) were collected directly using sterile 1-L glass bottles or 50-mL sterile collection vials and brought immediately to the laboratory. The sediments, 0.1 g, were washed in sterile seawater to remove the remaining water from the tank and resuspended in 1 mL of PBS with vortexing to suspend the attached bacteria. About 500 mL of the water was concentrated by centrifugation at 5,000 rpm during 30 min, and the concentrated cells were resuspended in 1 mL PBS. Aliquots of the prepared sediment and water samples (100 μl) were spread into five solid media, modified M13, M14, Zobell, M590 (Table 1), and R2A (BD-Difco™). For the selective isolation of Planctomycetes, the media were supplemented with 200 μg mL−1 ampicillin, 500 μg mL−1 streptomycin, and 20 μg mL−1 cycloheximide. The cultures were incubated in the dark at 20 °C, for one month. Screening for Planctomycetes was done by light microscopy. Pure cultures were obtained by subcultivation in the same isolation media and routinely maintained in M13. For long-term purpose storage, pure cultures were stored in M13 liquid medium with 20 % (w/v) glycerol and kept at −80 °C.

Table 1 Media used for isolation and cultivation
Full size table

Electron microscopy

For transmission electron microscopy, cells were harvested from 3-day-old cultures in M13, fixed in 2.5 % (w/v) glutaraldehyde in marine buffer (pH 7.2) for 2 h, and post-fixed in 1 % (v/v) osmium tetroxide in the water for 4 h and 1 % uranyl acetate for 1 h. Cells were dehydrated through a graded ethanol series, then propylene oxide, and embedded in Epon resin. Ultrathin sections were stained for 10 min in 1 % (v/v) uranyl acetate and for 10 min in Reynolds lead citrate. The sections were examined in a JEOL 100CXII transmission electron microscope.

Molecular identification of the isolates

Identification of the isolates was done by 16S rRNA gene analysis after amplification with the primers 27f and 1492r (Lane 1991) by direct colony PCR as described by Lage and Bondoso (2011). Briefly, the PCR mixture consisted of 1× PCR buffer, 1.5 mM MgCl2, 1 unit of GoTaq Flexi DNA Polymerase, 200 μM of each deoxynucleoside triphosphate, 2 μM of each primer, and material from a single colony to a final volume of 50 μl. Thermal PCR profile was done in MyCycler™ Thermo Cycler (Bio-Rad) and consisted in an initial denaturing step of 5 min at 95 °C, 30 cycles of 1 min at 94 °C, 1 min at 52 °C, and 90 s at 72 °C, and a final extension of 5 min at 72 °C. The obtained PCR products were purified using the GFX™ PCR DNA and Gel Band purification Kit (Amersham Biosciences), cloned into pGEM-T Easy vector (Promega) following standard protocols, and sequenced with universal primers M13F and M13R. Sequencing was performed at STAB Vida (Oeiras, Portugal). The 16S rDNA sequences obtained were assembled with Vector NTI Advance™ 10.3, compared with known sequences in GenBank and aligned using ClustalW (Thompson et al. 1994). The sequences were corrected manually using the alignment generated by ClustalW. Phylogenetic and molecular evolutionary analyses were conducted using MEGA version 5.03 (Tamura et al. 2007). Phylogenetic trees were generated with different calculation methods including neighbor joining (NJ), maximum parsimony (MP), and maximum likelihood (ML) to test for stability of the tree. Final ML tree applying Jukes–Cantor model was calculated with 33 sequences. Validation of reproducibility of the branching patterns was made by bootstrap based on 1,000 resamplings.

ERIC-PCR fingerprinting

DNA of the isolates was extracted using E.Z.N.A Bacterial DNA Isolation Kit (Promega) according to the manufacturer’s instructions from 7-day cultures on agar plates. The PCR was performed as described before (Rademaker and De Bruijn 1997) with the pair of primers ERIC1R and ERIC2 (Versalovic et al. 1991), and 100 ng of template DNA was used in each reaction. The annealing temperature was 48 °C. The PCR was carried out in a MyCycler™ Thermo Cycler (Bio-Rad). Each ERIC-PCR product (ten microliters) was separated by electrophoresis in a 2 % agarose gel at 60 V for 90 min in Tris–acetate–EDTA buffer. The gel was stained with GelRed 1:100,000 (v/v) during 45 min, and images were captured using GenoPlex (VWR). Computer-assisted analysis of the fingerprintings was performed using the software BioNumerics 6.5 (Applied Maths, Belgium).

Nucleotide sequence accession numbers

The 16S rRNA gene sequences of the isolated strains determined in this study were submitted to the GenBank and have been given accession numbers DQ851134, EF012748 to EF012750, and EF421447 to EF421453.

Results and discussion

Isolation and morphological identification

After about 1-month incubation, eleven pink colonies appeared on the solid medium M13 inoculated with sediment sample from the water recycling treatment tank. These isolates were designated as strains OJF1, 4, 5, 6, 9, 10, 11, 12, 14, and 15. No isolation was achieved on the other solid media or from the water column of the tank. The selectivity of the medium was provided by the ampicillin, a β-lactamic antibiotic that is ineffective against Planctomycetes, as they do not possess peptidoglycan in their cell walls, and by streptomycin as described by Schlesner (1994). These antibiotics prevented the overgrowth of other bacteria. The grown colonies were small round and convex with a granular appearance and a pink pigmentation characteristic of some Planctomycetes.

The cells of all the isolates were generally pear-shaped having the rosette-like cell aggregation (Fig. 1) with a variable number of cells characteristic of Rhodopirellula-like strains (Schlesner et al. 2004; Ward et al. 2006; Winkelmann and Harder 2009). Cells were about 0.7 × 1.5 μm. They reproduced by budding. In young cultures, many unicellular forms were mobile as observed by light microscopy. This is the first phase of some Planctomycetes life cycles, which occur prior to the loss of cell mobility when the cells become sessile (Staley 1973; Gade et al. 2005b). Under the electron microscope, cells presented polarity due to the asymmetric location of the DNA and fimbriae (Fig. 2). On this pole, crateriform structures are also evident. These morphologic characteristics are representative of this phylum and were indicative of the presence of Planctomycetes.

Fig. 1
figure 1

Phase contrast micrograph of Rhodopirellula sp. strain OJF1. Note many unicellular forms, the rosettes of variable number of cells, and a budding cell (arrow). Bar 2 μm

Full size image
Fig. 2
figure 2

Ultrathin section of Rhodopirellula sp. strain OJF1. On the reproductive pole, the cells present fimbriae (F) and crateriform structures (arrows). DNA is located asymmetrically nearby this pole. Bar 0.5 μm

Full size image

Isolation was only obtained in M13 medium, the media that has been used for the successful isolation and cultivation of strains of Rhodopirellula (Schlesner 1986; Winkelmann and Harder 2009). Medium M14 is appropriate for Blastopirellula (Schlesner 1986) and did not provide the isolation of Rhodopirellula spp. The other three tested media commonly used in isolation from marine environments (R2A, M590, and Zobell) were unable to provide growth probably due to the lack of vitamin B12 needed by Rhodopirellula spp. (Schlesner et al. 2004).

Molecular identification by 16S rDNA and ERIC fingerprinting of the strains

The analysis of the 16S rRNA gene confirmed the affiliation of all the obtained isolates to the Planctomycetes (Fig. 3). The strains were affiliated to Rhodopirellula and were divided in two clusters. Cluster A, which has a 99.5–99.9 % similarity to R. baltica, represents strains of this species. Cluster B, although also closely related to R. baltica, only has 97.7–97.9 % similarity to this species. This degree of similarity may be indicative of a new species of Rhodopirellula. Stackebrandt and Ebers (2006) considered a threshold of 98.7 % similarity in the 16S rRNA gene sequence as the separation for a new species. A similarity of 99.8–100 % exists among the members of cluster B. This putative new species has also been found in the epiphytic community of macroalgae from the north coast of Portugal (Lage and Bondoso 2011). Despite the differences in the geographic and environmental conditions, strain OJF1 is very similar to strain Rhodopirellula sp. LF2 as confirmed by the BOX-PCR fingerprinting (data not shown). Rhodopirellula sp. LF2 was isolated from the surface of Laminaria sp. collected in Foz, Porto (Lage and Bondoso 2011). R. baltica is a cosmopolitan species of Planctomycetes as confirmed by 16S rRNA gene sequence analysis, DNA–DNA hybridization, and MLSA analysis (Schlesner et al. 2004; Winkelmann et al. 2010). It was found for the first time as a free-living organism in the water column in the Baltic Sea (Schlesner 1994) and, posteriorly, in the tissue of the Mediterranean sponge Aplysina aerophoba (Gade et al. 2004), in marine snow (Pacific Ocean) (DeLong et al. 1993), in the biofilm of macroalgae (Bengtsson and Ovreas 2010; Lage and Bondoso 2011), and in several European Seas (Winkelmann and Harder 2009).

Fig. 3
figure 3

Phylogenetic 16S rDNA tree generated by maximum-likelihood analysis based on Jukes–Cantor model indicating the relationship between the strains isolated to members of the Planctomycetes. Isolates obtained in this study are shown in bold. GenBank accession numbers are shown in parenthesis. The anammox “Candidatus” genera were used as outgroup. The tree with the highest log likelihood is shown. The percentage of trees in which the associated taxa clustered together is shown above the branches. Bar 0.05 substitutions per 100 nucleotides

Full size image

Intraspecies variability was accessed with the ERIC-PCR fingerprinting (Fig. 4). With an 80 % similarity cutoff in the ERIC-PCR fingerprinting, it is possible to differentiate 9 genotypes. In cluster A, we can discriminate 5 genotypes, with OJF6 and OJF14 being part of the same genotype; in cluster B, we can see 4 genotypes in which strains OJF9 and OJF10 are the same. This fingerprinting analysis indicated that strains OJF1 and OJF12 are more similar to R. baltica strains than to cluster B strains. Simpson index for the whole community is 0.018, with cluster B presenting a higher diversity (0.05) than cluster A (0.033). Genetic diversity in Rhodopirellula was also reported by Winkelmann et al. (2010), confirming the presence of different genotypes within this genus.

Fig. 4
figure 4

Dendrogram of ERIC-PCR patterns of the isolated Rhodopirellula sp. strains, using curve-based Pearson-UPGMA method. The clusters A and B based on the 16S rDNA phylogenetic tree are evidenced. Gel images were analyzed with Bionumerics 6.5. Numbers on the tree refer to the similarity between the isolated strains

Full size image

Planctomycetes in aquacultures environments

Several Planctomycetes were reported in association with aquacultures environments namely Planctomyces, Pirellula, Rhodopirellula, and the “Candidatus” anammox (Sugita et al. 2005; Bissett et al. 2006; Kawahara et al. 2009; Michaud et al. 2009; Gao et al. 2012) reinforcing our findings. The ecological role of Planctomycetes, namely R. baltica, in fish farm environments may be due to their ability to perform the dissimilatory nitrate reduction to ammonium (DNRA) as reveled by in silico analysis of R. baltica genome (Mohan et al. 2004; Simon et al. 2011). This hypothesis is supported by the fact that in carbon-rich habitats, like fish farm sediments, DNRA pathway is greater than denitrification (Christensen et al. 2000; Nizzoli et al. 2006). Furthermore, they might also be important players in mineralization of organic matter as heterotrophic bacteria are known to play a key role in the biodegradation of organic material in marine environments (Deming and Baross 1993). R. baltica is a known specialized aerobic carbohydrate utilizer (Schlesner et al. 2004; Gade et al. 2005a), and strains belonging to Cluster B also utilized a wide variety of carbohydrates (unpublished data).

Information on the occurrence of antibiotic-resistant bacteria in marine sediments near fish farms and their relationship with population composition in the microbial community is still scarce (Schmidt et al. 2000; Tendencia and de la Peña 2001). According to Chelossi et al. (2003), in Italy, ampicillin is one of the most employed antibiotics against fish pathologies due to Vibrio spp., Pasteurella piscicida, and Myxobacterium spp. This and other classes of antibiotics, namely streptomycin, are used in aquaculture practice for the control of diseases (Kummerer 2009). Our work evidences the presence of Rhodopirellula spp. on the treatment of effluents or recirculation water in marine hatcheries. Their ability to degrade a variety of carbon sources (Gade et al. 2005a) and their natural resistance to high levels of streptomycin and β-lactamic antibiotics may be indicative of an ecological role in these kinds of habitats.

Conclusions

The isolated strains from the sediments of a water recycling treatment tank of a marine fish farm were identified as Planctomycetes by the combination of their morphological characteristics and antibiotic resistance. Further investigation using molecular techniques revealed the presence of two phylogenetic groups belonging to Rhodopirellula. A potential new species of Rhodopirellula was isolated. The ERIC fingerprinting showed a higher diversity within the isolates than the one revealed by the 16S rDNA.