Abstract
The living prokaryotic microbiome of the calcified geniculate (articulated) red alga, Corallina officinalis from the intertidal seashore is characterised for the first time based on the V6 hypervariable region of 16S rRNA. Results revealed an extraordinary diversity of bacteria associated with the microbiome. Thirty-five prokaryotic phyla were recovered, of which Proteobacteria, Cyanobacteria, Bacteroidetes, Actinobacteria, Planctomycetes, Acidobacteria, Verrucomicrobia, Firmicutes and Chloroflexi made up the core microbiome. Unclassified sequences made up 25% of sequences, suggesting insufficient sampling of the world's oceans/macroalgae. The greatest diversity in the microbiome was on the upper shore, followed by the lower shore then the middle shore, although the microbiome community composition did not vary between shore levels. The C. officinalis core microbiome was broadly similar in composition to those reported in the literature for crustose coralline algae (CCAs) and free-living rhodoliths. Differences in relative abundance of the phyla between the different types of calcified macroalgal species may relate to the intertidal versus subtidal habit of the taxa and functionality of the microbiome components. The results indicate that much work is needed to identify prokaryotic taxa, and to determine the nature of the relationship of the bacteria with the calcified host spatially, temporally and functionally.
Keywords: Actinobacteria, Bacteroidetes, Corallina officinalis, Cyanobacteria, prokaryotes, Proteobacteria
The first characterisation of the living prokaryotic microbiome of an intertidal calcified geniculate red alga Corallina officinalis reveals a rich diversity of bacteria with possible adaptation to the intertidal.
INTRODUCTION
Marine macroalgae (red, green and brown seaweeds) host a wide range of microbial organisms (e.g. Egan et al. 2013; Cavalcanti et al. 2014) that include both epi- and endobiotic prokaryotes and eukaryotes, although bacteria are the dominant active group (Oliveira et al. 2012). The composition of macroalgal bacterial microbiomes (hereafter referred to as microbiomes) is significantly different to the bacterial community found in the surrounding seawater, may be at least partially species-specific (Barott et al. 2011; Oliveira et al. 2012; Cavalcanti et al. 2014) and is likely to consist of both generalist and specialist populations (Egan et al. 2013). Nevertheless, Hollants et al. (2013), who reviewed 161 macroalgal–bacterial studies, reported that a core microbiome appeared to exist at the phylum level, consisting of the Gammaproteobacteria, Bacteroidetes (CFB group), Alphaproteobacteria, Firmicutes and Actinobacteria. At present, however, there is little evidence to suggest that individual bacterial species are necessarily host specific (Egan et al. 2013). Studies have demonstrated that for some macroalgal species, the composition of the bacterial microbiome can differ across individual thalli (Staufenberger et al. 2008; Barott et al. 2011), between individuals of the same species and over large spatial (e.g. west versus south east Australia for Ulva australis; Burke et al. 2011a), and temporal (seasonally) scales (e.g. Lachnit et al. 2009; Tujula et al. 2010; Miranda et al. 2013).
Algal–bacterial associations represent complex and highly dynamic relationships (Miranda et al. 2013). Associations have been categorised as beneficial (mutualistic), neutral (commensal) or harmful (parasitic), but the nature of the impact of these associations on either the host or microorganism remains unclear (e.g. Gachon et al. 2010). In addition, changes are likely to occur in the nature of associations over the course of the life history of both the host and microbiome, complicating simple categorisation. Although we do not fully understand the ecological roles of macroalgal microbiomes (Cavalcanti et al. 2014), increasing evidence indicates that the epiphytic bacteria provide important services to hosts that are vital to their health, performance and resilience (Harder et al. 2012; Egan et al. 2013; Singh and Reddy 2014 and references therein).
Given the significant ecological importance of calcified ‘coralline’ macroalgae (Corallinales, Rhodophyta) (Nelson 2009), interest has recently focussed on cataloguing and gaining an understanding of their associated microbiomes (Cavalcanti et al. 2014). Coralline algae are the most abundant and important calcifying macroalgae worldwide (Silva and Johansen 1986; Nelson 2009) and include crustose coralline algae (CCA), free-living coralline algae (rhodolith/maerl) and geniculate (articulated) turfing algae (Irvine and Chamberlain 1994; Brodie et al. 2013). These form a cosmopolitan group of marine flora, ubiquitous in intertidal and shallow subtidal habitats from polar to tropical regions, where they act as autogenic ecosystem engineers (Littler et al. 1985; Kamenos, Moore and Hall-Spencer 2004; Nelson 2009; Brodie et al. 2014).
Coralline algae form one of the most important structural elements in many coastal zones, are significant contributors to carbon fluxes in these areas, and are heavily involved in the recruitment and metamorphosis of many commercially important species (Johansen 1981; Morse and Morse 1984; Martin and Gattuso 2009; Couto, Neto and Rodrigues 2010; McCoy and Kamenos 2015 and references therein). Information on the composition and ecological interactions of calcified macroalgal microbiomes, and the contribution of the microbiome to host fitness and viability, is therefore highly important for the on-going conservation and sustainable use of these biological resources (Cavalcanti et al. 2014). This is especially pertinent given the current high-risk faced by calcifying macroalgae from both local, e.g. sedimentation, eutrophication, change in freshwater flows and global, e.g. climate change and ocean acidification perturbations (Koch et al. 2013; Brodie et al. 2014; McCoy and Kamenos 2015).
Microbiome studies of coralline algae using advanced genomic techniques have tended to concentrate on those of CCAs associated with reef building corals. Barott et al. (2011) studied the microbial diversity of functional groups of benthic reef algae and found that Proteobacteria and Cyanobacteria were abundant on CCA species, but that these corallines also supported considerable biodiversity of Firmicutes and Chloroflexi. Notable in Barott et al.'s (2011), study was the large proportion (∼18%–25%) of the microbiome that could not be classified beyond Bacteria. Other CCA studies have concentrated on their role in larval settlement, environmental change and ocean acidification. Webster et al. (2011) demonstrated that elevated seawater temperature of 2°C –4°C above the mean maximum annual sea surface temperature caused a shift in the microbial community from Alphaproteobacteria to Bacteroidetes, with subsequent downstream impacts on coral recruitment. An increase in Bacteroidetes has been shown to be indicative of stress such as elevated sea surface temperature or disease in other marine organisms (Webster, Cobb and Negri 2008). Webster et al. (2013) also demonstrated that reduced pH and increased CO2 altered the biochemistry of the CCA or its microbial associates with concern for the development, maintenance and recovery of coral reefs. Sneed, Ritson-Williams and Paul (2015), who studied the bacterial assemblages of CCAs of coral reefs in Belize, showed that two CCA species, which facilitate larval settlement, had a high abundance of Operational taxonomic units (OTUs) related to bacteria that inhibit the growth and/or biofilm formation of coral pathogens, whereas two other species on which fewer larvae settle had more OTUs related to known coral pathogens and cyanobacteria.
Studies of the microbiome of rhodoliths and geniculate coralline algae using similar genomic techniques are more limited. The microbial composition and functional components of rhodoliths from the Abrolhos Bank off eastern Brazil revealed remarkable homogeneity across sites and depths, indicating tight host-microbiome specificity (Cavalcanti et al. 2014). Bacteria comprised 70% of the microbiome, with the abundance of Alphaproteobacteria, Firmicutes and Actinobacteria typical of other red macroalgal microbiomes, whilst Betaproteobacteria and Deltaproteobacteria were particularly enriched in these rhodoliths. Additionally, Cavalcanti et al. (2014) highlighted the presence of major functional groups related to microbial-induced organomineralisation, suggesting an important but still unappreciated role of microbes in carbonate precipitation. For geniculate species, Huggett et al. (2006), working in a tropical location, reported that the bacterial biofilms of a species identified as C. officinalis (although almost certainly misidentified, as C. officinalis is a temperate species virtually confined to the North Atlantic; see Williamson et al. 2015), Amphiroa anceps, and mixed CCA, were important for successful settlement of marine larvae. Certain genera of Gammaproteobacteria were highly implicated in the induction of larval settlement, notably Pseudoalteromonas, with Vibrio, Shewanella, Photobacterium and Pseudomonas also important in larval settlement in a species-specific manner (Huggett et al. 2006).
The aim of the present study was to characterise the prokaryotic microbiome of the geniculate (articulated) coralline alga C. officinalis using the next-generation molecular techniques. Given its significant ecological importance in temperate coastal habitats and potential vulnerability to on-going anthropogenic change, C. officinalis has been the focus of several recent studies that have resolved questions regarding its phylogenetics and distribution, ecophysiology and environmental tolerances and potential responses to future change (Coull and Wells 1983; Kelaher 2002, 2003; Hofmann, Straub and Bischof 2012; Hofmann et al. 2012; Brodie et al. 2013, 2014; Williamson et al. 2014a,b, 2015). To date, however, no information is available on the composition of the C. officinalis microbiome (assuming misidentification in Huggett et al. 2006). To this end, the bacterial microbiome of C. officinalis was characterised across different shore heights of a rocky intertidal reef off the southwest UK, using the V6 hypervariable region of the 16S rRNA gene.
METHODS
Field sample collection
Corallina species are epilithic in rock pools in the UK, with C. officinalis typically found across the entire littoral zone of rocky sheltered coastlines (Brodie et al. 2013). For the present study, specimens of C. officinalis were collected from lower (∼2.5 m above Chart Datum [CD]), middle (∼4 m above CD) and upper shore (∼5 m above CD) rock pools across the species’ zone in the intertidal at Combe Martin, North Devon (51°12′13N 4°2′19W) on 14 November 2012 (Fig. 1). Two discrete individual specimens of C. officinalis were collected from each of the three shore heights, ensuring no overlapping of basal discs. To ensure coverage of the complete microbiome associated with C. officinalis fronds (see example microbiome; Fig. 1 panels c and d), three subsamples (∼1 g) were taken from each frond, resulting in 18 subsamples from a total of six replicate fronds. Excess water was removed by gently shaking, and each subsample was placed into a sterile 15 ml Falcon tube containing 5 ml of LifeGuard preservation solution, ensuring the samples were covered. Samples were immediately placed on dry ice for transportation to the laboratory, where they were stored at –80°C prior to RNA extraction.
Figure 1.
Corallina officinalis habitat and microbiome; (a) south west UK sampling site at Combe Martin (red dot); (b) C. officinalis turfing assemblage in intertidal rock pool at Combe Martin; (c–e) scanning electron micrographs (scale bars indicate 1 μm) showing example C. officinalis microbiome containing e.g. single-celled diatoms (mainly Cocconeis spp.) (c) mixed cyanobacterial filaments (c and d) and mixed bacterial assemblages (c–e).
RNA extraction and amplification
During the present study, RNA was extracted in order to sample live organisms and to avoid dead material that may have accumulated on or in the calcified frond of C. officinalis. Samples were homogenised by grinding, and RNA extraction was undertaken using Power Biofilm RNA isolation kits (MoBio Laboratories, Carlsbad, CA, USA), according to the manufacturer's instructions. Three RNA extractions were performed for each of the 18 subsamples and pooled per subsample. Multiple cDNA samples were amplified using SuperScript® III reverse transcriptase (Invitrogen Thermo Fisher Scientific, Waltham, MA, USA) for each pooled RNA subsample, and subsequently combined. To capture the prokaryotic diversity, the V6 hypervariable region of the 16S rRNA gene, homologous to positions 967 to 1046 of the E. coli SSU rRNA sequence (reference U00096), was amplified using nine primers (Table 1b; see Huber et al. 2007). For each of the 18 subsamples, a unique 6 bp barcode was annealed to the 5′ end of each primer (Table 1a), allowing amplification of the V6 region of each subsample using uniquely barcoded primers. PCRs were replicated and pooled to avoid bias.
Table 1b.
The five forward (those with F in primer code) and four reverse (those with R in primer code) unique primers used (pooled) with each of the 18 C. officinalis samples to amplify the V6 region.
| Primer code | Primer |
|---|---|
| 967F_PP | 5′-CNACGCGAAGAACCTTANC-3′ |
| 967F_UC1 | 5′-CAACGCGAAAAACCTTACC-3′ |
| 967F_UC2 | 5′-CAACGCGCAGAACCTTACC-3′ |
| 967F_UC3 | 5′-ATACGCGARGAACCTTACC-3′ |
| 967F_AQ | 5′-CTAACCGANGAACCTYACC-3′ |
| 1046R | 5′-CGACAGCCATGCANCACCT-3′ |
| 1046R_PP | 5′-CGACAACCATGCANCACCT-3′ |
| 1046R_AQ1 | 5′-CGACGGCCATGCANCACCT-3′ |
| 1046R_AQ2 | 5′-CGACGACCATGCANCACCT-3′ |
Table 1a.
Barcodes used for each C. officinalis individual for each shore region. Every barcode is matched with every primer (Table 1b) and anneals to the 5′ end.
| Lower shore | Middle shore | Upper shore | |||
|---|---|---|---|---|---|
| Barcode | Plant | Barcode | Plant | Barcode | Plant |
| AAGCGT | Plant 6 | GAACGA | Plant 4 | ATTGGC | Plant 2 |
| ACCCGT | Plant 6 | GAATGT | Plant 4 | CACTGT | Plant 2 |
| TACGGA | Plant 6 | TACAAG | Plant 4 | GATCTG | Plant 2 |
| GCAGTA | Plant 5 | CTGATC | Plant 3 | ACATCG | Plant 1 |
| GCGATT | Plant 5 | GTAGCC | Plant 3 | CGTGAT | Plant 1 |
| GCTTAC | Plant 5 | TCAAGT | Plant 3 | TGGTCA | Plant 1 |
Sequencing
Sequencing was undertaken at the University of Exeter Sequencing Service. 1.25 ng amplicon DNA was diluted with 150 μl EB (Elution Buffer) before end-repair, adenylation and ligation of Nextflex adapters using the SPRIworks library preparation protocol without size selection (Beckmann Coulter, Brea, CA, USA). The unamplified library was checked for the absence of primer dimers and quantified using a Bioanalyser (Agilent Technologies, Santa Clara, CA, USA) 7500 chip. The library was diluted to 10 nM. 2 ul of the diluted library was denatured and diluted to 3.5 pM. The final library was sequenced (100 paired end) across 2 lanes of a Rapid Run flow cell with on board clustering using Illumina HiSeq 2500 Rapid Run cluster kit and SBS (Sequencing by Synthesis) reagents v1. RTA basecalling was performed with v1.13 and CASAVA 1.8.2 was used to demultiplex reads allowing zero mismatches in the barcodes.
Data analysis
Initial sequence examination was performed using Geneious v.8.0.5 (Biomatters); raw sequence data are available in the NCBI Sequence Read Archive under accession SRP076344. Mothur v.1.33.3 (Schloss et al. 2009) was used for curation and analysis of sequences following the Mi-Seq SOP pipeline (http://www.mothur.org/wiki/MiSeq_SOP). Sequences that may contain homopolymer sequences or ambiguous nucleotides (Ns) were removed. Sequences were aligned against the Silva database v.119 in Mothur, and pre-clustered if they were within one base difference of a more abundant sequence. Chimeras were detected using Mothur's implementation of the UCHIME algorithm and subsequently removed. Taxonomic classification was performed, using the classify.seqs command with an 80% bootstrap cutoff, and undesirable domains (i.e. non-bacterial) removed from subsequent analyses. OTUs were formed at 0.03 distance, generally considered to correspond to species level classification, using the average neighbour algorithm.
Following normalisation of data to the sample size of the smallest library (n = 276556), i.e. random resampling of each library to the extent of the sample size of the smallest library, OTUs were used to construct rarefaction curves, to estimate richness (Chao 1) and to calculate diversity indices (Shannon and Simpson). Differences in richness and diversity indices between shore heights (lower, middle and upper) were analysed using linear mixed-effects models with restricted maximum likelihood (REML) criterion, with shore height as fixed effect (three levels) and subsample as the random term (three levels), using the lmer function of R (v. 3.0.2, R Core Team 2013) package lme4 (Bates, Maechler and Bolker 2013).Venn diagram representation of the OTU associations within and between shore heights was constructed in R using the package VennDiagram (Chen 2015) based on outputs from the Mothur Venn command.
RESULTS AND DISCUSSION
Prokaryotic diversity
The living prokaryotic microbiome of the calcified geniculate red alga Corallina officinalis has been characterised for the first time using next-generation sequencing of the V6 hypervariable region of 16S rRNA. From an initial ∼45 million raw sequencing reads, a total of 18 761 729 contigs were produced, of which 2573 585 were unique. After filtering sequences (trimming, chimera removal and clustering), 9359 534 remained, of which 368 160 were unique (mean length 64 bp long). Alignment to the Silva v.119 reference database and removal of non-bacterial sequences, resulted in 362 545 unique bacterial sequences. The complete taxonomic classification of sequences is provided in (Table S1, Supporting Information). Thirty-five bacterial phyla were recovered (Table 2), of which the top nine, Proteobacteria, Cyanobacteria, Bacteroidetes, Actinobacteria, Planctomycetes, Acidobacteria, Verrucomicrobia, Firmicutes and Chloroflexi make up the core with between 99.53% (lower shore) and 99.94% (upper shore) of all classified sequences (summary, Table 3). However, the Proteobacteria and Cyanobacteria dominated, together comprising ∼84% of lower shore, 81% of middle shore and 80% of upper shore bacterial sequences (Fig. 2), and the Bacteroidetes and Actinobacteria accounted for most of the remaining classified sequences, with c. 15% on the lower shore, 18% on the middle shore and 16% on the upper shore.
Table 2.
Mean proportion of RNA sequences and number of OTUS for six subsamples per shore height for the phyla of Bacteria recovered from C. officinalis. ‘-’ represents samples where the proportion of sequences is <0.01.
| Sequences (% ± SE) | OTUS (number ±SE) | |||||
|---|---|---|---|---|---|---|
| Shore level | ||||||
| Phylum | Lower | Middle | Upper | Lower | Middle | Upper |
| Proteobacteria | 36.05 (±1.35) | 27.97 (±2.13) | 35.67 (±0.61) | 10323 (±549) | 5977 (±519) | 11865 (±659) |
| Cyanobacteria | 30.20 (±3.45) | 28.80 (±2.59) | 23.86 (±2.05) | 875 (±29) | 531 (±31) | 760 (±81) |
| Unclassified | 20.40 (±5.04) | 29.34 (±1.72) | 25.34 (±3.71) | 8704 (±388) | 5885(±314) | 11824(±587) |
| Bacteroidetes | 6.05 (±0.45 | 9.37 (±2.04) | 8.37 (±0.84) | 2860 (±229) | 2480 (±356) | 3612 (±205) |
| Actinobacteria | 5.75 (±0.52) | 3.35 (±0.49) | 4.21 (±0.30 | 923 (±44) | 530 (±62) | 838 (±37) |
| Planctomycetes | 0.70 (±0.07) | 0.50 (±0.08) | 1.21 (±0.13) | 529 (±49) | 298 (±46) | 781 (±71) |
| Acidobacteria | 0.28 (±0.05) | 0.15 (±0.02) | 0.41 (±0.04) | 159 (±22) | 83 (±6) | 302 (±15) |
| Verrucomicrobia | 0.27 (±0.01) | 0.26 (±0.01) | 0.31 (±0.01) | 150 (±7) | 110 (±11) | 197 (±12) |
| Firmicutes | 0.16 (±0.15) | 0.08 (±0.03) | 0.08 (±0.03) | 66 (±26) | 40 (±7) | 82 (±7) |
| Chloroflexi | 0.10 (±0.03) | 0.03 (±0.00) | 0.22 (±0.02) | 66 (±7) | 38 (±6) | 130 (±17) |
| Nitrospirae | 0.01 (±0.01) | 0.02 (±0.00) | 0.27 (±0.04) | 10 (±2) | 10 (±1) | 30 (±2) |
| Lentisphaerae | 0.01 (±0.00) | 0.01 (±0.00) | 0.03 (±0.01) | 18 (±1) | 12 (±2) | 38 (±7) |
| Gemmatimonadetes | 0.01 (±0.00) | – | 0.02 (±0.00) | 10 (±1) | 6 (±1) | 16 (±2) |
| Armatimonadetes | – | – | – | – | – | – |
| Fusobacteria | – | – | – | 5 (±1) | 5 (±1) | 8 (±2) |
| Spirochaetae | – | – | – | 2 (±0) | 2 (±0) | 3 (±1) |
| Candidate_division_BRC1 | – | – | – | – | – | 1 (±0) |
| Candidate_division_OD1 | – | – | – | 3 (±1) | 3 (±1) | 6 (±1) |
| Candidate_division_OP11 | – | – | – | – | – | – |
| Candidate_division_OP3 | – | – | – | 1 (±0) | 2 (±0) | 3 (±1) |
| Candidate_division_TM7 | – | – | – | 1 (±0) | 1 (±0) | 1 (±0) |
| Candidate_division_WS3 | – | – | – | 1 (±0) | 1 (±0) | 8 (±1) |
| Candidate_division_WS6 | – | – | – | – | – | – |
| Chlamydiae | – | – | – | 3 (±1) | 2 (±0) | 4 (±0) |
| Chlorobi | – | – | – | 4 (±0) | 2 (±1) | 9 (±1) |
| Deferribacteres | – | – | – | – | – | 1 (±0) |
| Deinococcus–Thermus | – | – | – | 1 (±0) | – | 1 (±0) |
| Elusimicrobia | – | – | – | – | – | – |
| Fibrobacteres | – | – | – | – | 1 (±0) | 2 (±0) |
| Hyd24-12 | – | – | – | – | – | – |
| NPL-UPA2 | – | – | – | – | 1 (±0) | 1 (±0) |
| SHA-109 | – | – | – | 2 (±0) | 1 (±0) | 1 (±0) |
| TA06 | – | – | – | – | – | 1 (±0) |
| TM6 | – | – | – | – | – | 1 (±0) |
| Tenericutes | – | – | – | 1 (±1) | – | – |
Table 3.
The number of classes, orders, families and genera recovered for the top 9 phyla of the C. officinalis microbiome.
| Phylum | No. of classes | No. of orders | No. of families | No. of generaa | Unclassified genera (%) |
|---|---|---|---|---|---|
| Proteobacteria | 14 | 75 | 205 | 520 (170) | 32.69 |
| Cyanobacteria | 4 | 6 | 8 | 18 (7) | 38.90 |
| Bacteroidetes | 11 | 14 | 42 | 166 (36) | 21.69 |
| Actinobacteria | 8 | 20 | 57 | 118 (37) | 31.36 |
| Planctomycetes | 9 | 17 | 18 | 31 (18) | 58.10 |
| Verrucomicrobia | 7 | 9 | 15 | 21 (10) | 47.62 |
| Acidobactera | 5 | 19 | 30 | 38 (28) | 73.68 |
| Firmicutes | 5 | 9 | 34 | 112 (22) | 19.64 |
| Chloroflexi | 12 | 6 | 6 | 6 (6) | 100 |
Numbers in brackets are the numbers of unclassified genera.
Figure 2.
Relative abundance (% sequences) of the top nine phyla recovered from the C. officinalis bacterial microbiome from upper (Up), middle (Mid) and lower shore (Low) fronds.
The results have revealed a diverse assemblage of bacteria found in the microbiome of C. officinalis. The largest number of OTUs was recorded for the Proteobacteria with 10 323 (±549) on the lower shore, 5977 (±519) on the middle shore and 11 865 (±659) on the upper shore (Table 2). The Cyanobacteria, with 875 (±29) OTUs on the lower shore, 531 (±31) on the middle shore and 760 (±81) on the upper shore, had far fewer OTUs than the Bacteroidetes, which had 2860 (±229) on the lower shore, 2480 (±356) on the middle shore and 3612 (±205) on the upper shore, despite Cyanobacteria comprising a much greater proportion of sequences at each shore level. Similarly, the Planctomycetes, which made up <1% of all sequences, demonstrated a relatively large number of OTUs, particularly on the lower and upper shore: 529 (±49) on the lower shore, 298 (±46) on the middle shore and 781 (±71) on the upper shore.
Given that rarefaction curves did not saturate for samples assessed from any shore height (Fig. 3), our data suggests that much diversity remains to be discovered. Additionally, approximately one-quarter of the bacterial sequences were unable to be classified to a bacterial phylum, reflecting the findings of Miranda et al. (2013) for the microbiome of Porphyra umbilicalis (Rhodophyta), whereby 25.2% of sequences recovered for the 16S V5V6 region were unclassified. Possible reasons for these findings include insufficient sampling of the oceans’ microorganisms and/or the macroalgae.
Figure 3.
Rarefaction analysis of V6 samples from each shore height showing the mean (±95% C.I) number of OTUs (0.03 distance) as a function of sequencing depth.
Shore height differences
Among all six plants collected from different intertidal habitats, rarefaction analysis indicated that greater diversity was recovered from the upper shore, followed sequentially by the lower shore and middle shore (Fig. 3). These results were supported by analysis of observed (OTU number) and expected (Chao1) richness, and Shannon diversity calculated for the C. officinalis microbiome (Table 4), which demonstrated significantly increased species richness and diversity on upper shore, as compared to lower and middle shore, sequentially, and observed OTUs (F2,9 = 42.86; P < 0.0001), Chao1 richness (F2,9 = 40.21; P < 0.0001) and Shannon diversity (F2,9 = 20.01; P < 0.0001) (post-hoc Tukey; P < 0.05 in all cases). In contrast, however, no significant difference in Simpsons diversity index was apparent between shore heights (P = 0.421).
Table 4.
Observed (number of OTUS) and estimated (Chao1) bacterial richness and diversity (Shannon and Simpson indices) of C. officinalis microbiome for each shore height (mean ± SE). n = 6. Significance based on linear mixed-effects analysis in relation to shore height (***P = <0.001; NS = not significant).
| Shore level | ||||
|---|---|---|---|---|
| Diversity index | Lower | Middle | Upper | Significance |
| No. of OTUs | 24 733.33 (±1013.89) | 16 003.67 (±685.93) | 30 509.67 (±1494.60) | *** |
| Chao1 | 71 537.00 (±3649.29) | 43 599.00 (±1925.56) | 90 275.75 (±4912.90) | *** |
| Shannon diversity | 4.89 (±0.10) | 4.51 (±0.08) | 5.55 (±0.027) | *** |
| Simpson diversity | 0.08 (±0.01) | 0.09 (±0.01) | 0.05 (±0.01) | NS |
Venn diagram representation of the OTU associations within and between shore heights (Fig. 4) indicates a microbiome that is common to the three shore levels for C. officinalis of 4249 OTUs, representing 43.17% of lower shore, 53.75% of middle shore and 35.61% of upper shore total OTU communities. Across shore heights, the largest number of OTUs was observed on C. officinalis in the upper shore (5152), followed by the lower (3376) and middle (1604) shore. A greater shared number of OTUs was apparent between the upper and lower shore (1347) than the upper and middle shore (1182), with the smallest shared OTUs apparent between the lower and middle shore (869).
Figure 4.
Venn diagram showing the number of OTUs (0.03 distance) shared between and among shore heights. For example, 4249 OTUs are shared by at least one sample from all shore heights.
Across the intertidal, our data indicated a relatively conserved C. officinalis microbiome, although there is evidence of subtle differences in diversity. Rarefaction analysis, observed (OTU) and expected (Chao1) richness, and Shannon diversity indices were significantly different between shore heights, all highlighting increased richness and diversity on the upper shore. The presence in our data of shared OTUs across shore heights was in contrast to Miranda et al. (2013) for the microbial community of P. umbilicalis, whereby only one common OTU was apparent across all twelve blades examined. Similarly, Burke et al. (2011b) reported only six common OTUs across samples of the green alga Ulva australis, though with the presence of a highly shared functional genome, which, Burke et al. (2011a) suggested, allows sub-sets of bacteria to be functionally equivalent. Our data do, however, reflect the findings of Cavalcanti et al. (2014), who demonstrated impressively homogenous communities of bacteria across seven rhodoliths from different sites and water depths off the eastern coast of Brazil.
Additional to our core C. officinalis microbiome, differences identified in richness and diversity across shore heights may have been related to gradients of stress occurring across the rocky intertidal. The duration of tidal emersion increases with shore height, along with consequent exposure to e.g. irradiance, temperature and salinity fluctuations (Williamson et al. 2014a). Greater abiotic stress experienced in the upper intertidal may thus have served to maintain a greater diversity in the bacterial communities of C. officinalis, whereas more stable lower and middle shore environments may have facilitated dominance by certain bacterial groups, reducing overall richness and diversity.
Classes and orders of the Corallina officinalis core microbiome
The abundance (% of sequences) and number of OTUs for classes of the top nine phyla with a relative abundance ≥0.5% sequences are displayed in Table 5. The majority of the nine core phyla were represented at each shore height by just one or two dominant classes with the exception of the Planctomycetes, which were relatively evenly spread across five classes, although making up less than 1% of the diversity. All phyla had a proportion of sequences that were unclassified, with the Cyanobacteria having the highest (63%–70%) and the Proteobacteria the lowest (4%–6%).
Table 5.
Proportion of sequences (% ±SE) and number of OTUs (number ±SE) recovered for each class for the top nine phyla for six subsamples per shore height across the three shore heights sampled, for sequences with a relative abundance of ≥0.5%. The three right hand columns represent abundance of each phylum and class at the different shore heights; l = lower shore, m = middle shore, u = upper shore; dark red = >50%, red = >10%–50%, orange = >1%–10% and light orange = 0%–1%.
|
For classes of the top nine phyla, the largest number of OTUs was found in the Gammaproteobacteria (1750–3388) and Alphaproteobacteria (2854–5290) of the Proteobacteria, with the second largest number of OTUs observed in the Flavobacteria and unclassified Bacteroidetes (Table 2). The Firmicutes and Chloroflexi had the smallest numbers of OTUs. With few exceptions, all classes with ≥0.5% proportion of sequences had representatives at all shore heights. Two exceptions were the class SGST604 of the Planctomycetes and the class Erysipelotrichia of the Firmicutes, which were both absent from the middle shore.
Table 6 shows the proportion of sequences and OTUs across shore heights for the orders with ≥0.5% sequence representation in at least one shore height for the classes shown in Table 5. With the exception of the Opitutales, which was absent from the middle shore, all orders were represented at each shore height. For the top nine phyla, 90 orders were recovered with ≥0.5% of sequences.
Table 6.
The proportion of sequences (±SE) and number of OTUs (±SE) for sequences recovered for each order with a relative abundance of ≥0.5% of the classes for the top nine phyla for six subsamples per shore height across the three shore heights sampled. The three right-hand columns represent abundance of each order at the different shore heights; l = lower shore, m = middle shore, u = upper shore, dark red = >50%, red = >10%–50%, orange = >1%–10%, light orange = 0%–1% and white = 0.
|
Core microbiome and comparison with calcified and fleshy red algae
The overall composition of the C. officinalis microbiome determined in the present study was comparable to that previously reported for both calcified and fleshy red algal species (Table 7), although differences were apparent in the relative abundance of bacterial phyla. Hollants et al.'s (2013) core community differed notably from the C. officinalis microbiome in the absence of Cyanobacteria and greater relative proportion of Firmicutes. The Proteobacteria dominated for all the calcified algae and the microbiome of the geniculate C. officinalis are shown here to be similar to those for the CCAs of Barott et al. (2011) and Webster et al. (2013), despite the latter's limited list. However, the Firmicutes (as with Hollant et al.'s (2013) core microbiome) and Chloroflexi in particular were relatively more abundant than for Barott et al.'s (2011) CCAs. Sneed, Ritson-Williams and Paul's (2015) CCAs surface microbiomes of the two species which facilitated larval settlement and two that did not (results not shown in Table 7) were similarly dominated by Proteobacteria and had relatively high numbers of Actinobacteria, but Bacteroidetes were not among the most dominant bacteria phyla found. In comparison to the microbiome of the free-living rhodoliths from the east coast of Brazil (Cavalcanti et al. 2014), C. officinalis demonstrated a reduced prevalence of Firmicutes, although Planctomycetes were present in its core microbiome.
Table 7.
Comparison of top phyla of the C. officinalis prokaryotic microbiome in order of relative abundance in relation to fleshy and calcified red algae. Data adapted from where comparable microbiome results are available.
| Calcified red algae | |||||||
|---|---|---|---|---|---|---|---|
| Geniculate | Crustose coralline algae (CCAs) | Uncalcified red algae | |||||
| Core community | C. officinalis | CCAs | Hydrolithon onkodes | Rhodoliths | D. pulchra | L. dendroidea | P. umbilicalis |
| Hollants et al. 2013 | This study | Barott et al. (2011) | Webster et al. (2013) | Cavalcanti et al. (2014) | Longford et al. (2007) | Oliveira et al. (2012) | Miranda et al. (2013) |
| V6 region of 16S rRNA; Highseq | V1-V3 region of 16SrRNA gene | 16S rRNA | Total DNA | 16S rDNA; DGGE | Transcriptomic profile | V8 & V5V6 region of 16S rDNA | |
| Proteobacteriac | Proteobacteria | Proteobacteria | Proteobacteria | Proteobacteria | Proteobacteriaa | Cyanobacteria | Bacteroidetes |
| Bacteroidetes | Cyanobacteria | Cyanobacteria | Bacteroidetes | Actinobacteria | Bacteroidetes | Proteobacteriaa | Proteobacteria |
| Proteobacteriad | Bacteroidetes | Firmicutes | Cyanobacteria | Firmicutes | Planctomycetes | Actinobacteria | Actinobacteria |
| Firmicutes | Actinobacteria | Chloroflexi | Cyanobacteria | Chloroflexi | Firmicutes | Chloroflexi | |
| Actinobacteria | Planctomycetes | Bacteroidetes | Bacteroidetes | Cyanobacteria | Bacteroidetes | Planctomycetes | |
| Acidobacteria | Actinobacteria | Actinobacteria | Acidobacteria | Firmicutes | |||
| Verrucomicrobia | Planctomycetes | Verrucomicrobia | Spirochaetes | Deinococcus–Thermusb | |||
| Firmicutes | Chlorobi | Deinococcus–Thermus | |||||
| Chloroflexi | |||||||
Gammaproteobacteria and Alphaproteobacteria dominant.
Identified from the V8 hypervariable region only.
Gammaproteobacteria.
Alphaproteobacteria.
Considering fleshy red macroalgae, the core microbiome of the subtidal species Delisea pulchra was comparable to that reported here for C. officinalis, although Cyanobacteria were less prominent, and Firmicutes were lacking in D. pulchra (Longford et al. 2007). Oliveira et al.'s (2012) microbiome of the red alga Laurencia dendroidea, an intertidal to subtidal (to 3 m) species, was dominated by nitrogen fixing Cyanobacteria and aerobic heterotrophic Proteobacteria, similar to C. officinalis, whereby the abundance of the class Cyanobacteria was proportional to that found in L. dendroidea (32%–41% C. officinalis, 36% L. dendroidea), and comparable orders and genera were recovered. However, differences were apparent in the relative abundance of the remaining bacterial phyla comprising the L. dendroidea microbiome, with Actinobacteria and Firmicutes making up the bulk, in contrast to Bacteroidetes and Actinobacteria in C. officinalis. The relative dominance of Bacteroidetes and Chloroflexi also differed between the microbiomes of C. officinalis and the intertidal P. umbilicalis. Bacteroidetes was the most dominant component of the P. umbilicalis microbiome, though the second most common phylum for C. officinalis, and Chloroflexi were also abundant in P. umbilicalis though much less so in C. officinalis (Miranda et al. 2013).
CONCLUSIONS
We have been able to characterise the Corallina officinalis microbiome and show similarities and differences in the bacterial composition in a relation to both calcified and fleshy red algae. It is clear, however, that a much work is needed to identify prokaryotic taxa, and to determine the nature of the relationship of the bacteria with the calcified host spatially, temporally and functionally. Future assessment will benefit from increased temporal and spatial resolution, and interspecies comparisons at the same locations, across a greater range of macroalgal functional forms.
Supplementary Material
Acknowledgments
We are grateful to Professor Tom Richards, University of Exeter, Dr Konrad Paszkiewicz and Dr Karen Moore, Exeter Sequencing Service and Wellcome Trust Biomedical Informatics and Guy Leonard (University of Exeter) for their assistance with the project.
SUPPLEMENTARY DATA
FUNDING
This work was supported by a Systematics and Taxonomy Research Scheme grant (SynTax), () and ().
Conflict of interest. None declared.
REFERENCES
- Barott KL, Rodriguez-Brito B, Janouškovec J, et al. Microbial diversity associated with four functional groups of benthic reef algae and the reef-building coral Montastraea annularis. Environ Microb. 2011;13:1192–204. doi: 10.1111/j.1462-2920.2010.02419.x. [DOI] [PubMed] [Google Scholar]
- Bates D, Maechler M, Bolker B. lme4: Linear-mixed Effects Models Using S4 Classes. 2013. http://CRAN.R-project.org/package=lme4 (25 May 2016, date last accessed) [Google Scholar]
- Brodie J, Walker R, Williamson C, et al. Epitypification and redescription of Corallina officinalis L., the type of the genus, and C. elongata Ellis & Solander (Corallinales, Rhodophyta) Cryptogamie, Algol. 2013;34:49–56. [Google Scholar]
- Brodie J, Williamson CJ, Smale D, et al. The future of the NE Atlantic benthic flora in a high CO2 world. Ecol Evol. 2014;4:2787–98. doi: 10.1002/ece3.1105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Burke C, Steinberg P, Rusch D, et al. Bacterial community assembly based on functional genes rather than species. P Natl Acad Sci USA. 2011a;108:14288–93. doi: 10.1073/pnas.1101591108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Burke C, Thomas T, Lewis M, et al. Composition uniqueness and variability of the epiphytic bacterial community of the green algal Ulva australis. ISME J. 2011b;4:301–11. doi: 10.1038/ismej.2010.164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cavalcanti GS, Gregoracci GB, dos Santos EO, et al. Physiologic and metagenomic attributes of the rhodoliths forming the largest CaCO3 bed in the South Atlantic Ocean. ISME J. 2014;8:52–62. doi: 10.1038/ismej.2013.133. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chen H. Package ‘VennDiagram’. 2015. https://cran.rproject.org/web/packages/VennDiagram/VennDiagram.pdf. [Google Scholar]
- Coull BC, Wells JBJ. Refuges from fish predation - experiments with phytal meiofauna from the New Zealand rocky intertidal. Ecology. 1983;64:1599–609. [Google Scholar]
- Couto RP, Neto AI, Rodrigues AS. Metal concentration and structural changes in Corallina elongata (Corallinales, Rhodophyta) from hydrothermal vents. Mar Pollut Bull. 2010;60:509–14. doi: 10.1016/j.marpolbul.2009.11.014. [DOI] [PubMed] [Google Scholar]
- Egan S, Harder T, Burke C, et al. The seaweed holobiont: understanding seaweed–bacteria interactions. FEMS Microbiol Rev. 2013;37:462–76. doi: 10.1111/1574-6976.12011. [DOI] [PubMed] [Google Scholar]
- Gachon CMM, Sime-Ngando T, Martina Strittmatter M, et al. Algal diseases: spotlight on a black box. Trends Plant Sci. 2010;15:633–40. doi: 10.1016/j.tplants.2010.08.005. [DOI] [PubMed] [Google Scholar]
- Harder T, Campbell AH, Egan S, et al. Chemical mediation of ternary interactions between marine holobionts and their environment as exemplified by the red alga Delisea pulchra. J Chem Ecol. 2012;38:442–50. doi: 10.1007/s10886-012-0119-5. [DOI] [PubMed] [Google Scholar]
- Hofmann L, Straub S, Bischof K. Competition between calcifying and noncalcifying temperate marine macroalgae under elevated CO2 levels. Mar Ecol-Prog Ser. 2012;464:89–105. [Google Scholar]
- Hofmann L, Yildiz G, Hanelt D, et al. Physiological responses of the calcifying rhodophyte, Corallina officinalis (L.), to future CO2 levels. Mar Biol. 2012;159:783–92. [Google Scholar]
- Hollants J, Leliaert F, De Clerck O, et al. What we can learn from sushi: a review on seaweed–bacterial associations. FEMS Microbiol Ecol. 2013;83:1–16. doi: 10.1111/j.1574-6941.2012.01446.x. [DOI] [PubMed] [Google Scholar]
- Huber JA, Welch DBM, Morrison HG, et al. Microbial population structure in the deep marine biosphere. Science. 2007;318:97–100. doi: 10.1126/science.1146689. [DOI] [PubMed] [Google Scholar]
- Huggett MJ, Williamson JE, de Nys R, et al. Larval settlement of the common Australian sea urchin Heliocidaris erythrogramma in response to bacteria from the surface of coralline algae. Oecologia. 2006;149:604–19. doi: 10.1007/s00442-006-0470-8. [DOI] [PubMed] [Google Scholar]
- Irvine LM, Chamberlain YM. Seaweeds of the British Isles: Vol 1 Rhodophyta Part 2B Corallinales, Hildenbrandiales. London: HMSO; 1994. pp. 1–276. [Google Scholar]
- Johansen HW. Corallina Algae, A First Synthesis. Boca Raton, Florida: CRC Press; 1981. p. [vii]+239. [Google Scholar]
- Kamenos NA, Moore PG, Hall-Spencer JM. Nursery-area function of maerl grounds for juvenile queen scallops Aequipectenopercularis and other invertebrates. Mar Ecol Prog Ser. 2004;274:183–9. [Google Scholar]
- Kelaher BP. Influence of physical characteristics of coralline turf on associated macrofaunal assemblages. Mar Ecol Prog Ser. 2002;232:141–8. [Google Scholar]
- Kelaher BP. Changes in habitat complexity negatively affect diverse gastropod assemblages in coralline algal turf. Oecologia. 2003;135:431–41. doi: 10.1007/s00442-003-1196-5. [DOI] [PubMed] [Google Scholar]
- Koch M, Bowes G, Ross C, et al. Climate change and ocean acidification effects on seagrasses and marine macroalgae. Glob Chang Biol. 2013;19:103–32. doi: 10.1111/j.1365-2486.2012.02791.x. [DOI] [PubMed] [Google Scholar]
- Lachnit T, Blümel M, Imhoff JF, et al. Specific epibacterial communities on macroalgae: phylogeny matters more than habitat. Aquatic Biol. 2009;5:181–6. [Google Scholar]
- Littler MM, Littler DS, Blair SM, et al. Deepest known plant life discovered on an uncharted seamount. Science. 1985;227:57–9. doi: 10.1126/science.227.4682.57. [DOI] [PubMed] [Google Scholar]
- Longford SR, Tujula NA, Crocetti GR, et al. Comparisons of diversity of bacterial communities associated with three sessile marine eukaryotes. Aquat Microb Ecol. 2007;48:217–29. [Google Scholar]
- McCoy S, Kamenos NA. Coralline algae (Rhodophyta) in a changing world: Integrating ecological, physiological, and geochemical responses to global change. J Phycol. 2015;51:6–24. doi: 10.1111/jpy.12262. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Martin S, Gattuso J-P. Response of Mediterranean coralline algae to ocean acidification and elevated temperature. Glob Change Biol. 2009;15:2089–100. [Google Scholar]
- Miranda LN, Hutchison K, Grossman AT, et al. Diversity and abundance of the bacterial community of the red macroalga Porphyra umbilicalis: did bacterial farmers produce macroalgae? PLoS One. 2013;8:e58269. doi: 10.1371/journal.pone.0058269. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Morse ANC, Morse DE. Recruitment and metamorphosis of Haliotis larvae induced by molecules uniquely available at the surface of crustose red algae. J Exp Mar Biol Ecol. 1984;75:191–215. [Google Scholar]
- Nelson WA. Calcified macroalgae - critical to coastal ecosystems and vulnerable to change: a review. Mar Fresh Res. 2009;60:787–801. [Google Scholar]
- Oliveira LS, Gregoracci GB, Silva GGZ, et al. Transcriptomic analysis of the red seaweed Laurencia dendroidea (Florideophyceae, Rhodophyta) and its microbiome. BMC Genomics. 2012;13:487. doi: 10.1186/1471-2164-13-487. [DOI] [PMC free article] [PubMed] [Google Scholar]
- R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2013. [Google Scholar]
- Schloss PD, Westcott SL, Ryabin T, et al. Introducing mothur: Open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microb. 2009;75:7537–41. doi: 10.1128/AEM.01541-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Silva PC, Johansen HW. A reappraisal of the order Corallinales (Rhodophyceae) Brit Phycol J. 1986;21:245–54. [Google Scholar]
- Singh RP, Reddy CRK. Seaweed-microbial interactions: key function of seaweed-associated bacteria. FEMS Microb Ecol. 2014;88:213–30. doi: 10.1111/1574-6941.12297. [DOI] [PubMed] [Google Scholar]
- Sneed JM, Ritson-Williams R, Paul VJ. Crustose coralline algal species host distinct bacterial assemblages on their surfaces. ISME J. 2015;9:2527–36. doi: 10.1038/ismej.2015.67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Staufenberger T, Thiel V, Wiese J, et al. Phylogenetic analysis of bacteria associated with Laminaria saccharina. FEMS Microb Ecol. 2008;64:65–77. doi: 10.1111/j.1574-6941.2008.00445.x. [DOI] [PubMed] [Google Scholar]
- Tujula NA, Crocetti GR, Burke C, et al. Variability and abundance of the epiphytic bacterial community associated with a green marine Ulvacean alga. ISME J. 2010;4:301–11. doi: 10.1038/ismej.2009.107. [DOI] [PubMed] [Google Scholar]
- Webster NS, Cobb RE, Negri AP. Temperature thresholds for bacterial symbiosis with a sponge. ISME J. 2008;2:830–842. doi: 10.1038/ismej.2008.42. [DOI] [PubMed] [Google Scholar]
- Webster NS, Soo R, Cobb R, et al. Elevated seawater temperature causes a microbial shift on crustose coralline algae with implications for the recruitment of coral larvae. ISME J. 2011;5:759–70. doi: 10.1038/ismej.2010.152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Webster NS, Uthicke S, ES Botté, et al. Ocean acidification reduces induction of coral settlement by crustose coralline algae. Glob Change Biol. 2013;19:303–15. doi: 10.1111/gcb.12008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Williamson CJ, Brodie J, Goss B, et al. Corallina and Ellisolandia (Corallinales, Rhodophyta) photophysiology over daylight tidal emersion: interactions with irradiance, temperature and carbonate chemistry. Mar Biol. 2014a;161:2051–68. doi: 10.1007/s00227-014-2485-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Williamson CJ, Njorka J, Perkins R, et al. Skeletal mineralogy of geniculate corallines: providing context for climate change and ocean acidification. Mar Ser. 2014b;531:71–84. [Google Scholar]
- Williamson CJ, Walker RH, Robba L, et al. Towards resolution of diversity in Corallina (Tribe Corallineae, Rhodophyta) and related genera. Phycologia. 2015;54:2–11. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.