Community composition of the Planctomycetes associated with different macroalgae

Abstract

Insights into the diversity of marine natural microbial biofilms, as for example those developing at the surface of marine macroalgae, can be obtained by using molecular techniques based on 16S rRNA genes. We applied denaturing gradient gel electrophoresis (DGGE) with 16S rRNA genes–specific primers for Planctomycetes to compare the communities of these organisms developing on six different macroalgae (Chondrus crispus, Fucus spiralis, Mastocarpus stellatus, Porphyra dioica, Sargassum muticum, and Ulva sp.) sampled in spring 2012 in two rocky beaches in the north of Portugal. Planctomycetes can be one of the dominant organisms found in the epibacterial community of macroalgae, and we wanted to determine the degree of specificity and the spatial variation of these group. Shannon diversity indexes obtained from the comparison of DGGE profiles were similar in all the macroalgae, and in both sites, F. spiralis was the algae presenting lower Planctomycetes diversity, while M. stellatus and P. dioica from Porto showed the highest diversity. The analysis of DGGE profiles, including anosim statistics, indicate the existence of a specific Planctomycetes community associated with the algal host, likely independent of geographical variation. Sequencing of DGGE bands indicated that Planctomycetes communities were highly diverse, and some Operational Taxonomic Units seemed to be specifically associated with each macroalgae.

Planctomycetes, macroalgae, biofilm, denaturing gradient gel electrophoresis, host specificity

Introduction

Planctomycetes is a widespread deep-branching phylum of Bacteria present in many diverse habitats, although, in general, it appears in relatively low abundances in environmental samples (Rusch et al., 2007). They are part of the PVC superphylum together with Chlamydiae, Verrucomicrobia, Lentisphaera, and the Candidate groups Poribacteria, OP3 and WWE (Wagner & Horn, 2006). The phylum Planctomycetes has been poorly studied, although in the last decade, it has recalled attention mostly due to the unusual presence of characteristics previously found in eukaryotic cells (Devos & Reynaud, 2010). Examples of these particular features include the presence of proteinaceous cell walls without peptidoglycan, budding reproduction, a complex cell plan (Fuerst & Sagulenko, 2011), endocytosis-like protein uptake (Lonhienne et al., 2010) and membrane coat-like proteins (Santarella-Mellwig et al., 2010). The cosmopolitan distribution of Planctomycetes suggests a wide capacity to adapt distinct habitats. Members of this group are usually found in marine (Vergin et al., 1998; Shu & Jiao, 2008), brackish (Schlesner, 1994;Zeng et al.,2013a), and fresh waters (Pizzetti et al., 2011; Pollet et al., 2011) as well as in terrestrial ecosystems (Wang et al., 2002; Buckley et al., 2006). They are also reported to be present in extreme environments such as hypersaline (Burns et al., 2004; Baumgartner et al., 2009) and acidophilic (Ivanova & Dedysh, 2012) habitats and glacial waters (Zeng et al.,2013b). Planctomycetes can also be found in association with eukaryotic hosts, like ants (Eilmus & Heil, 2009), invertebrates (Fuerst et al., 1991; Chaiyapechara et al., 2012), sponges (Pimentel-Elardo et al., 2003; Mohamed et al., 2008; Zhu et al., 2008; Ouyang et al., 2010; Sun et al., 2010; Sipkema et al., 2011; Webster et al., 2011; Costa et al., 2012), ascidians (Oliveira et al., 2013), corals (Webster & Bourne, 2007), macrophytes (Hempel et al., 2008; He et al., 2012), lichens (Grube et al., 2012), Sphagnum peat bogs (Kulichevskaya et al., 2006), and with the rhizosphere of several plants (Da Rocha et al., 2009). Several studies have shown that Planctomycetes are frequent in the epibacterial community of several macroalgae (Longford et al., 2007; Bengtsson et al., 2010; Burke et al., 2011; Lachnit et al., 2011; Lage & Bondoso, 2011; de Oliveira et al., 2012; Hollants et al., 2013; Miranda et al., 2013). They are the dominant group in the kelp Laminaria hyperborea with values that can reach 51% of the total bacterial community (Bengtsson & Ovreas, 2010). An advantage for this colonization is the presence in Planctomycetes of a high number of sulfatase genes (Wegner et al., 2013), which are involved in the degradation of the sulphated polymers produced by the algae.

Denaturing gradient gel electrophoresis (DGGE) is a molecular technique that separates similar sized DNA fragments based on their G + C content. It was first described by Muyzer et al. (1993) as a community fingerprinting technique using the 16S rRNA gene to estimate bacterial diversity in environmental samples and identify the dominant uncultivable taxa. It is a relatively easy, reproducible, reliable, and fast technique. Although DGGE does not allow a full taxonomic assignment, it has successfully been applied in the comparison of bacterial communities as they vary through time and space. DGGE has been used with marine bacterial assemblages (Murray et al., 1996; Moeseneder et al., 1999; Riemann et al., 1999; Schauer et al., 2000), marine picoeukaryotic assemblages (Dı́ez et al., 2001), with the bacteria of the surface mucus layer of coral species (Morrow et al., 2012), the cyanobacterial epiphytes on macroalgae (Ohkubo et al., 2006) or the microbial community inhabiting sponges (Li et al., 2007; Thiel et al.,2007a,b) and corals (Webster & Bourne, 2007).

Mühling et al. (2008) developed primers to apply the technique to particular bacterial groups. One of the primer sets was developed for Planctomycetes and it was further used by Pollet et al. (2011). Here, we optimized the Planctomycetes-specific PCR-DGGE developed by Mühling et al. (2008) and Pollet et al. (2011) to investigate the host-specific association of Planctomycetes with six different macroalgae belonging to the phyla Heterokontophyta, Chlorophyta and Rhodophyta and their spatial variation in two nearby locations of the north coast of Portugal.

Materials and methods

Macroalgae sampling and site locations

Macroalgae were collected in May 2012 in tidal pools of beaches in Porto (41º09′ N, 8º40′ W) and Carreço (41º44′ N, 8º52′ W). Fresh vegetative thalli of Chondrus crispus, Fucus spiralis, Mastocarpus stellatus, Porphyra dioica, Sargassum muticum, and Ulva sp. were collected in triplicate in sterile plastic bags with seawater and transported to the laboratory within 1–2 h. The algae used in this study were phylogenetically affiliated to Heterokontophyta (F. spiralis and S. muticum), Chlorophyta (Ulva sp.), and Rhodophyta (C. crispus, M. stellatus, and P. dioica). The algae M. stellatus was sampled only in Porto, as it was absent in Carreço. However, it was used for comparison with the other macroalgae from Porto. Temperature, salinity, and pH were measured at the sampling sites. Once in the laboratory, the algae were rinsed in sterile natural seawater to remove loosely attached bacteria and frozen at −20 °C until DNA extraction was performed.

DNA extraction

Genomic DNA of the bacterial communities associated with the macroalgae was extracted with UltraClean^® Soil DNA Isolation Kit (MoBio laboratories, Inc.). Ten circles from each specimen were cut with a circular 0.5 cm diameter cork borer and used for extraction. DNA extraction was performed according to manufacturer's instructions with the exception that the tubes containing the beads solution and the macroalgae pieces were initially vortexed for 15 min in a Disruptor Cell Genie.

PCR-DGGE fingerprinting

Geographical and macroalgae host variations of the Planctomycetes communities were assessed by DGGE with the Planctomyces-specific pairs of primers described in Table ⁰⁰⁰¹. To determine the best method to visualize the DGGE profiles, three approaches were tested: (1) an initial PCR with the specific Planctomycetes pair of primers 352F/920R, followed by a nested PCR with the pair of primers 518f-GC/907R, as described by Mühling et al. (2008); (2) direct PCR with the pair of primers 352F-GC/920R, as described by Pollet et al. (2011); and (3) an initial PCR with the pair of universal primers 9bfm/1512R, followed by a nested PCR with the primers 352F with a GC clamp and 920R using the previous PCR product diluted 200× as template. The DGGE profiles were then compared, and the method that yielded more defined and clear bands was selected. The variability of the Planctomycetes microbial community on macroalgae was analyzed by performing DGGE on 16S rRNA gene fragments of three individuals of the same algae collected in both locations. Polymerase chain reactions were performed in 50 μL mixtures containing 1× Green GoTaq^® Flexi Buffer, 2 mM MgCl₂, 200 μM of each deoxynucleotide, 1 mM of each primer, 1 mg mL⁻¹ of bovine serum albumin, 2 units of GoTaq^® DNA Polymerase, and 20 ng of DNA template. PCR conditions for the primer pair 9bfm/1512R were performed according to Mühling et al. (2008) and consisted in an initial denaturation step of 4 min at 96 °C followed by 30 cycles of 1 min at 96 °C, 1 min at 52 °C and 90 s at 74 °C, and a final extension of 10 min at 74 °C. The thermal PCR profile for the nested PCR with the pair of primers 352F-GC/920R was performed as described by Pollet et al. (2011) and consisted in an initial denaturation step of 5 min at 96 °C, 10 cycles of 1 min at 96 °C, 1 min at 68 °C (the temperature was reduced by 1 °C in each cycle) and 1 min at 72 °C, followed by 20 cycles of 1 min at 96 °C, 1 min at 58 °C and 1 min at 72 °C and a final extension of 5 min at 72 °C.

Oligonucleotides used for PCR-DGGE

Primers	Sequence (5′–3′)	Target organism	References
9bfm	GAGTTTGATYHTGGCTCAG	Bacteria	Mühling et al. (2008)
1512R	ACGGHTACCTTGTTACGACTT	Universal (bacteria and archaea)	Mühling et al. (2008)
PLA352F	GGC TGC AGT CGA GRA TCT	Planctomycetales	Mühling et al. (2008)
PLA920R	TGT GTG AGC CCC CGT CAA	Planctomycetales	Mühling et al. (2008)
518f-GC	CCAGCAGCCGCGGTAATCGCCCGCCGCGCGCGGCGGGCGGGGCGGGGGCACGGGGGG	Bacteria	Muyzer et al. (1993)
907r	CCGTCAATTCMTTTGAGTTT	Bacteria	Muyzer et al. (1998)
GC-tail	CGC CCG CCG CGCCCC GCG CCC GTC CCG CCG CCC CCC GGG CG		Pollet et al. (2011)

Used in combination with the primer PLA352F.

The PCR products were verified and quantified by agarose gel electrophoresis with a low DNA mass ladder standard (Invitrogen). About 800 ng of the PCR products from these mixtures were run in a DGGE gel at 60 °C with a CBS Scientific system as previously described by Pollet et al. (2011) using a 50–70% gradient (6% acrylamide) at 120 V (18 h). A ladder made from a mixture of individual isolates of Planctomycetes previously isolated from macroalgae was also loaded in the extremities of the gel. The gel was stained with SybrGold (Molecular Probes) for 45 min, rinsed with 1× Tris-acetate-EDTA buffer, removed from the glass plate to a UV-transparent gel scoop, and visualized under UV light in a ChemiDoc system (Bio-Rad). The DGGE images were analyzed using the quantityone software (Bio-Rad).

Sequence and analysis of DGGE bands

Representative DGGE bands were excised from the gel and re-amplified with the pair of primers 352F/920R with the PCR conditions described previously. The resulting PCR product was purified and sequenced at Macrogen Europe using the primer 352F. The sequences obtained from the DGGE gels were manually cleaned in Sequence Analysis 5.2 and blasted against the 16S rRNA gene database in RDP. The closest relatives were downloaded and aligned with the band sequences in clustalw. The resulting alignment was used to construct an optimum maximum-likelihood tree to determine the phylogenetic position of the obtained sequences. The sequences were deposit in GenBank under accession numbers KF364635–KF364661.

Data analysis and statistical treatment of DGGE profiles

Digitized DGGE images were analyzed using the quantityone software (Bio-Rad). Similarity of resulting banding patterns was assessed by constructing a matrix taking into account the presence or absence of individual bands in each sample and their relative intensity in each lane. Based on this matrix, a Bray–Curtis similarity matrix was produced, and then, two types of analyses were run to plot the results. A multidimensional scaling (MDS) analysis modeled the variability of the patterns by representing them as points in a low-dimensional space. On the other hand, the Cluster analysis grouped the samples on a dendrogram. These statistical analyses were run using the software tool Plymouth Routines in Multivariate Ecological Research (primer6). Comparisons between the sampling sites and macroalgae hosts were made using analyses of similarity (anosim) in which an R value of 1 indicates maximum variation between groups and an R value of 0 indicates no differences between groups.

Results

DGGE profiles of Planctomycetes associated with the macroalgae

The first approach to study the Planctomycetes communities in the surface of macroalgae following the PCR method described by Mühling et al. (2008) (method 1) appeared not to be appropriate for our samples. The resulting PCR amplicons yielded multiple faint and undefined bands when visualized in the agarose gel (details not shown). With the other two approaches, the resulting PCR product showed only one clear band with the expected size. The bands obtained with the nested PCR with the pair of primers 352F-GC/920R (method 2) were more intense than the ones from the direct PCR (method 3; details not shown). Furthermore, the number of bands obtained in the DGGE profiles with the nested PCR protocol was higher than the one obtained with the direct PCR approach, although these were apparently more defined and without noise background (Fig. ⁰⁰⁰¹a). Some of the representative bands from both, the nested and direct PCR-DGGE, were sequenced and all matched to the Planctomycetes (details not shown). Further analyses with PCR-DGGE of the Planctomycetes community on macroalgae were thus based on the nested PCR approach.

(a) DGGE profiles of 16S rRNA gene amplified from different algae with the direct PCR protocol and specific pair of primers for Planctomycetes (left), or the nested approach (right). Fs, Fucus spiralis; Lh, Laminaria sp.; Pd, Porphyra dioica; U1 and U2, Ulva sp.; L, Ladder. (b) DGGE fingerprinting profiles of the Planctomycetes community associated with three individuals (1, 2 and 3) of F. spiralis (F), Ulva sp. (U) and Chondrus crispus (Cc).

The intraspecies variability of the Planctomycetes community between different individuals of the same macroalgae is shown in Fig. ⁰⁰⁰¹b which shows DNA extracted and amplified from three individuals of the algae C. crispus, F. spiralis, and Ulva sp. sampled at two different sites. Overall, the DGGE profiles of the triplicates of each alga were similar, with the exception of individuals U1 and Cc1 from Porto, and the relative band abundances were also similar, suggesting that the Planctomycetes communities are consistent within different individuals of the same species of algae and do not present large intra-individual variations.

The DGGE profiles of the Planctomycetes communities associated with the six different macroalgae from Carreço and Porto are shown in Fig. ⁰⁰⁰². A total of 53 bands were identified in the gel. The bands identified in each DGGE profile were assumed to be different Operational Taxonomic Units (OTUs), and the intensity of each band was considered to provide the relative abundance of each OTU.

(a) DGGE fingerprinting profiles of the planctomycetes community associated with Sargassum muticum (Sm), Porphyra dioica (Pd), Chondrus crispus (Cc), Ulva sp. (U), and Fucus spiralis (Fs) from Carreço and Porto. L, Ladder. The arrows refer to the bands excised and sequenced. (b) Dendrogram of DGGE profiles of the planctomycetes communities, based on Bray–Curtis similarity.

Based on these assumptions, the number of dominant OTUs (S) and the Shannon Diversity Index was determined for each macroalga (Table ⁰⁰⁰²). The highest planctomycetes diversity was found in the red macroalgae P. dioica (H′ = 3.12) and M. stellatus (H′ = 3.02) both sampled in Porto. The other macroalgae showed lower planctomycetes diversity with Shannon indexes ranging from 2.4 to 2.9. Overall, there were no evident differences in the diversity of the Planctomycetes communities in Carreço (mean H′ = 2.71) and Porto (mean H′ = 2.75) among different macroalgal species.

Number of bands (S) observed in each macroalga and respective Shannon diversity index (H′)

Macroalgae	Site	OTUs (S)	Shannon Index (H′)
Sargassum muticum	Carreço	18	2.85
Porto	19	2.82
Porphyra dioica	Carreço	14	2.48
Porto	25	3.12
Chondrus crispus	Carreço	19	2.85
Porto	15	2.62
Ulva sp.	Carreço	19	2.84
Porto	15	2.56
Fucus spiralis	Carreço	14	2.54
Porto	12	2.38
Mastocarpus stellatus	Porto	23	3.02

The analysis of the DGGE gel showed the existence of Planctomycetes strains that are present in the majority of the macroalgae such as bands 10/28, 7, 25, and 4/12, while others were only found associated with a specific algal species in both locations, like band 23 (P. dioica) and bands 16/30 (F. spiralis).

Host-specific Planctomycetes community

The band profile of the DGGE gel was used to construct a resemblance matrix that originated a dendrogram showing the clustering of the samples (Fig. ⁰⁰⁰²b). With the exception of P. dioica, the community profiles of Planctomycetes associated with the different macroalgae were clustered according to the algal host phylum and not according to the sampling site. Ulva sp. showed a high similarity between individuals from both locations (> 60%) while S. muticum and C. crispus exhibit a similarity of c.50% between samples. In Fig. ⁰⁰⁰²b, it is possible to visualize two major branches, one containing the DGGE profiles of Rhodophyta (C. crispus and M. stellatus) and Chlorophyta algae and another one consisting of algae from the phylum Heterokontophyta. The profiles from the red algae C. crispus and the green algae Ulva sp. were grouped according to the host species and did not change with the geographical location. Porphyra dioica and F. spiralis harbored a specific community that differed according to the habitat sampled. These findings were confirmed by statistical analysis with anosim (a test to verify significant differences between two or more groups of samples) that confirmed the results discussed previously. Samples were grouped by ‘Site’ and ‘macroalgae species’ as factors. There were no statistical differences in the DGGE profiles between both sites (R = −0.056, P = 0.5) while between the different algae the profiles were significantly distinct (R = 0.536, P = 0.006).

A better visualization of the similarity between the Planctomycetes communities in each macroalgae was obtained with a nonmetric multidimensional scaling (nMDS, Fig. ⁰⁰⁰³). The nMDS plot showed two distinctly separated groups. The Planctomycetes community on P. dioica from Carreço was clearly distinct from the other macroalgae communities (Fig. ⁰⁰⁰³). This group was expanded in a second nMDS plot that shows two clusters and isolated samples from the macroalgae P. dioica (Porto) and F. spiralis (Porto) with no similarity to the other groups. Ulva sp. (Porto and Carreço), C. crispus (Porto and Carreço), and M. stellatus (Porto) were grouped together with a similarity of 40% and S. muticum (Porto and Carreço) shared a similarity of 40% with F. spiralis (Carreço).

nMDS plots evidencing the clustering of the macroalgae-inhabiting Planctomycetes samples based on the DGGE profiles. (a) General MDS plot with all the samples. (b) MDS subgroup plot from samples inside the square in (a). ◼, Carreço; ▲, Porto; —, 20% similarity; ---, 40% similarity.

Taxonomic affiliation of the bands

To identify the major groups of Planctomycetes associated with the macroalgae, the most defined and representative bands were extracted from the gel and sequenced. Sequences that showed double peaks were eliminated from the analysis. A total of 30 band sequences were included in the final dataset. The sequences were all phylogenetically affiliated to the phylum Planctomycetes confirming the specificity of the primers used (Table S1). Furthermore, some bands in different lanes but in the same position were cut and sequenced to confirm that they represented the same OTU. This was the case of the pairs of bands 4/12, 10/28, and 16/30 that appeared in the phylogenetic tree with a value of similarity higher than 99.6%. The majority of the bands sequenced represented distinct OTUs, with the exception of the bands 12, 14, and 27 that shared a similarity higher than 99.4% and 20/21 that shared 100% similarity in the 16S rRNA gene. The closest relatives of the bands were mainly uncultured Planctomycetes obtained from the surface of L. hyperborea, Fucus vesiculosus, and Ulva australis (Table S1). The closest cultured relatives were isolated strains from macroalgae surface (Lage & Bondoso, 2011), including strain Pd1 (from P. dioica), strains LF1 and LF2 (from Laminaria sp.), strains FC18 and FF4 (from F. spiralis), and strain UC8 (from Ulva sp.).

The phylogenetic tree obtained (Fig. ⁰⁰⁰⁴) showed that the OTUs were distributed in four major clusters (groups A–D) and were affiliated to genera Rhodopirellula, Planctomyces and with two unclassified Planctomycetes genera. For an easier interpretation of the results, the 16S rRNA gene sequences were grouped at 98% similarity, value indicative of ‘species’. Group A can be divided into two different ‘species’ (A1 and A2) and contained the majority of the bands sequenced. OTUs belonging to this group were found in all the algae sampled suggesting a widespread distribution. It includes strains mainly isolated from macroalgae, and it is phylogenetically related to Planctomycete sp. FC18 that was originally isolated from the surface of F. spiralis from Carreço (Lage & Bondoso, 2011). ‘Species’ A1 can be found in a wide variety of habitats, including macroalgae, microbial mats, methane see sediments, seafloor lavas, sponges, and oil-polluted sediments indicating that these strains are widely distributed and also that they can adapted to extreme and polluted environments. However, ‘species’ A2 was only found in L. hyperborea, F. spiralis and ocean water around Enteromorpha prolifera, suggesting that these strains are specifically associated with macroalgae. Group B sequences were phylogenetically related to the genus Rhodopirellula and contain three different OTUs closely related to other clones obtained from several habitats including macroalgae and sponges. Group C contains ‘species’ related to Planctomyces maris, and other clones mainly described from macroalgae. OTUs represented by bands 6 and 7 were only found in the macroalga L. hyperborea and were present in all the DGGE profiles indicating a possible specific association of these OTUs with macroalgae. Group D consisted of two different ‘species’ that were phylogenetically close to the Anammox genera [responsible for ANaerobic AMMonium Oxidation (Strous et al., 1999)] and presented < 80% similarity in the 16S rRNA gene to the described genera of Planctomycetes. This low value could be indicative of a distinct order of Planctomycetes that remain yet to be isolated.

Maximum-Likelihood tree of 16S rRNA gene sequences extracted from DGGE bands (in bold), and their phylogenetic relation to other members of Planctomycetes and closest uncultured representatives. Strains in gray represent the clones identified from the surface of macroalgae. The numbers beside nodes are the percentages for bootstrap analyses; only values above 50% are shown. Scale bar = 0.05 substitutions per 100 nucleotides. The different groups are presented on the right. Anammox 16S rRNA gene sequences were used as outgroup. The bands were named after the macroalgae they were sequenced from (Cc, Chondrus crispus; Fs, Fucus spiralis; Ms, Mastocarpus stellatus; Pd, Porphyra dioica; Sm, Sargassum muticum; U, Ulva sp.) and the sampling site (C, Carreço and P, Porto). L1, L2 and L3 correspond to the bands sequenced from the standard lane.

Discussion

This is the first culture-independent study exclusively focused on the distribution of Planctomycetes in the epiphytic microbial community of several co-occurring macroalgae. In the last years, members of the Planctomycetes have been reported to be associated with macroalgae (Longford et al., 2007; Fukunaga et al., 2009; Bengtsson & Ovreas, 2010; Bengtsson et al., 2010; Burke et al., 2011; Lachnit et al., 2011; Lage & Bondoso, 2011; Miranda et al., 2013). Several novel taxa of Planctomycetes have been isolated from the surface of macroalgae (Fukunaga et al., 2009; Lage & Bondoso, 2011), but it is well known that isolated strains do not always represent the whole community (Rappé & Giovannoni, 2003). With this study, we aimed to characterize the whole Planctomycetes community associated with macroalgae. Planctomycetes are known to contain a high number of sulfatases genes that could play a major role in the degradation of the sulfated polysaccharides abundant in the algae walls (Wegner et al., 2013), which prompted us to investigate the Planctomycetes community in these hosts. PCR-DGGE fingerprinting has been used in the microbial ecology study of bacterial communities associated with eukaryotic hosts, like sponges (Webster et al., 2011), algae (Lachnit et al., 2009; Tujula et al., 2010), and corals (Webster & Bourne, 2007) allowing to address questions such as the spatial and temporal variations and the determination of the host-specific bacterial community. In this study, we used PCR-DGGE with specific primers for Planctomycetes to explore the epiphytic community of this group on different co-occurring macroalgae from different phyla and from two different locations. We showed that there were significant differences in the DGGE profiles of the six macroalgae, indicating the existence of a Planctomycetes-specific community associated with each macroalgae. The results also indicated that individuals from the same algae, but different locations were more similar to each other than to other algae in the same location. Furthermore, it was shown that the intraspecies variability of the planctomycetes communities within the same species was not significant, although Ulva sp. and C. crispus showed some variation. Similar banding patterns of DGGE of the whole bacterial community on individuals of the same algae species in the same habitat have been reported previously. Longford et al. (2007) showed a minimum of 60% similarity between the bacterial communities in the epiphytic communities of the algae Delisea pulchra and 70% similarity in Ulva sp. Tujula et al. (2010) reported that the differences in the individuals DGGE profiles of U. australis from different tidal pools where not greater than the differences between individuals collected in the same tidal pool. The results presented in this study extend the absence of intraspecific variation within a given macroalgae to a specific group of bacteria.

Using PCR-DGGE of the 16S rRNA gene, we determined the composition of each Planctomycetes community profile of six different co-occurring macroalgae. Each band was assigned to a different OTU, although we found that in some cases different bands were very similar (higher than 99.4%). This does not exactly mean that they represent clones of the same species. In the case of planctomycetes, it was already shown that they present a high genetic diversity at the ecotype level, determined by Multilocus Sequence Analysis and Enterobacterial Repetitive Intergenic Consensus PCR, even when the 16S rRNA gene similarity between two isolates is higher than 99.5% (Winkelmann et al., 2010; Lage et al., 2012; Cayrou et al., 2013). However, we treated each band as a distinct OTU. In terms of richness, there was not a significant difference between the algae, indicating that Planctomycetes can easily colonize all the host species, and that there is a high diversity of species. This is in agreement with results obtained from culture- dependent methods in which different taxa of planctomycetes were isolated from different macroalgae, independently of the host species (Lage & Bondoso, 2011). Culture-independent methods have also shown that planctomycetes can be found in association with F. vesiculosus, Gracilaria vermicuphylla, Ulva intestinalis (Lachnit et al., 2011), Ulva spp. (Hengst et al., 2010), Macrocystis pyrifera (Michelou et al., 2013), U. australis (Longford et al., 2007; Burke et al., 2011), D. pulchra (Longford et al., 2007), L. hyperborea (Bengtsson & Ovreas, 2010), and Porphyra umbilicalis (Miranda et al., 2013). Interestingly, they were absent in other algae, for example, Saccharina japonica (Balakirev et al., 2012), Laminaria saccharina (Staufenberger et al., 2008). This can be due to the primers used, as some are known to contain mismatches to the phylum Planctomycetes or PCR conditions. As an example, an in-depth study where a large number (c.16 000 sequences) of clones from the whole bacterial community of U. australis (Burke et al., 2011) was sequenced showed the presence of 3.4% Planctomycetes clones, while a PCR-DGGE study made in the same algae indicated the absence of this phylum (Tujula et al., 2010).

Cluster analysis of the DGGE profiles showed the existence of a host-specific community of Planctomycetes. The banding patterns of the algae C. crispus, Ulva sp., and S. muticum were more similar to each other than to the ones from other algae in the same location. Furthermore, they were grouped according to the phyla of host macroalgae, with the exception of P. dioica, suggesting the existence of epiphytic Planctomycetes shared among taxonomically closely related hosts. The influence of the host in the bacterial community of macroalgae has been reported previously, although these studies address the whole bacterial community and not only the Planctomycetes. For example, Fucus serratus, F. vesiculosus, L. saccharina, Ulva compressa, Delesseria sanguinea, and Phycodrys rubens from North and Baltic Seas have epibacterial communities that differed less between both locations than between algae from the same place (Lachnit et al., 2009). These authors also found that DGGE profiles from the algae under study were grouped according to the host phylum in both locations. In our study, the DGGE fingerprinting profiles exhibited by Rhodophyta (with the exception of P. dioica from Carreço) were more similar to the ones of Chlorophyta. This finding was confirmed also by 16S rRNA gene clone libraries of C. crispus and Ulva sp. that showed several OTUs shared in common by both algae (J. Bondoso, F. Godoy-Vitorino, V. Balague, P. Gasol and O.M Lage, unpublished data). In the same Lachnit et al. (2009) study, the DGGE profiles of the bacterial community associated with Rhodophyta were more similar to the ones of Chlorophyta. Other studies on bacterial epiphytes from macroalgae also found host specificity of the bacterial communities. The red alga Bonnemaisonia asparagoides exhibited a different profile of bacterial species from other two coexisting red algae, Lomentaria clavellosa and Polysiphonia stricta, and there was not significant intraspecific differences between localities (Nylund et al., 2010). The bacterial composition on L. saccharina and Dyctyosphaeria ocellata were very similar between different habitats (Staufenberger et al., 2008; Sneed & Pohnert, 2011). Hengst et al. (2010) reported a strong effect of the algal hosts Ulva spp., Scytosiphon lomentaria, and Lessonia nigrescens in the bacterial community associated with these species. Host-specific associations have also been reported for sponges (Hentschel et al., 2002; Taylor et al., 2007) and diatoms (Grossart et al., 2005). Although a recent metagenomic study on U. australis-associated bacteria suggested that the differences among macroalgae are dependent on bacterial species functionality and not on its taxonomy, our results reinforce the importance of the host phylogeny in other macroalgae-associated bacterial communities and suggest that macroalgae modulate the bacterial community associated with their surface through different intrinsic biological, physical and chemical characteristics. Furthermore, the Planctomycetes communities from the surface of macroalgae would follow the same pattern as the whole bacteria community, presenting a host-specific association rather than a spatial specific distribution.

The phylogenetic composition of the Planctomycetes community in the macroalgae was found to be very diverse. Eleven different ‘species’ were retrieved from the DGGE-extracted bands, based on a 98% threshold. However, not all the bands could be extracted and sequenced, which could influence the results obtained indicating that not all the Planctomycetes community was covered. The most abundant group was group A, composed of two distinct ‘species’ and 11 OTUs. 16S rRNA gene clone libraries obtained from L. hyperborea surfaces also showed prevalence of strains belonging to this group, which accounted for 72–97.8% in the libraries (Bengtsson & Ovreas, 2010). Similarly, in F. vesiculosus, 16S rRNA gene clone libraries revealed that the majority of the Planctomycetes clones belonged to this group (Lachnit et al., 2011). So far, there is only one cultured representative in this group, isolated from the surface of F. spiralis from Carreço, Planctomycete sp. FC18 (Lage & Bondoso, 2011). FC18 exhibited a 16S rRNA gene similarity between 94% and 98% to the uncultured strains of this group. Some of the OTUs present in group A, in particular ‘species’ A2, appeared to be present only in macroalgae samples thus suggesting a specific association to macroalgae of these planctomycetes. Group B, closely related to Rhodopirellula spp., contains three different ‘species’ and were phylogenetically closer to other cultured Planctomycetes isolated from the surface of macroalgae that are currently being described as novel genera. Although Rhodopirellula sp. and in particular R. baltica have been shown to be widely associated with macroalgae, surprisingly none of the DGGE bands matched this group. One possibility for this fact could be that not all the DGGE bands could be sequenced. ‘Species’ belonging to group C, affiliated to P. maris, were also retrieved from L. hyperborea, F. vesiculosus, and P. dioica. Two of the OTUs were exclusively found in L. hyperborea. Group D was composed of two species loosely related with the order Planctomycetales, probably indicating the presence of a novel order of Planctomycetes that would be found mainly associated with macroalgae. The closest relative was only one uncultured sequence, with 93–94% similarity in the 16S rRNA gene to group D OTUs.

The results presented in this study showed a host-specific community of Planctomycetes associated with macroalgae. Furthermore, the Planctomycetes communities were highly diverse, and some of the OTUs were found to be specifically associated with macroalgae. Because DGGE does not allow a full taxonomic study and assignment of the communities, further studies are needed to investigate whether these specific OTUs are associated with a specific host, or are widely distributed among the macroalgae.

Acknowledgements

This research was supported by the European Regional Development Fund (ERDF) through the COMPETE – Operational Competitiveness Programme and national funds through FCT – Foundation for Science and Technology, under the project PEst-C/MAR/LA0015/2013 and by EU project Micro3B. The first author was financed by FCT (PhD grant SFRH/BD/35933/2007).

Supporting Information

Additional Supporting Information may be found in the online version of the article:

Table S1. Taxonomic affiliation of the DGGE bands sequences with the closest uncultured and isolated organisms.

Author notes

Editor: Riks Laanbroek