Consistency in microbiomes in cultures of Alexandrium species isolated from brackish and marine waters

Abstract

Phytoplankton and bacteria interactions have a significant role in aquatic ecosystem functioning. Associations can range from mutualistic to parasitic, shaping biogeochemical cycles and having a direct influence on phytoplankton growth. How variations in phenotype and sampling location, affect the phytoplankton microbiome is largely unknown. A high-resolution characterization of the bacterial community in cultures of the dinoflagellate Alexandrium was performed on strains isolated from different geographical locations and at varying anthropogenic impact levels. Microbiomes of Baltic Sea Alexandrium ostenfeldii isolates were dominated by Betaproteobacteria and were consistent over phenotypic and genotypic Alexandrium strain variation, resulting in identification of an A. ostenfeldii core microbiome. Comparisons with in situ bacterial communities showed that taxa found in this A. ostenfeldii core were specifically associated to dinoflagellate dynamics in the Baltic Sea. Microbiomes of Alexandrium tamarense and minutum, isolated from the Mediterranean Sea, differed from those of A. ostenfeldii in bacterial diversity and composition but displayed high consistency, and a core set of bacterial taxa was identified. This indicates that Alexandrium isolates with diverse phenotypes host predictable, species-specific, core microbiomes reflecting the abiotic conditions from which they were isolated. These findings enable in-depth studies of potential interactions occurring between Alexandrium and specific bacterial taxa.

Figure 1

Dissimilarity analysis, nMDS, using a distance matrix with Bray Curtis dissimilarity of all strains, using R version 3.3.2, with packages Vegan (Oksanen et al., 2008) and Cluster (Maechler et al., 2016). Inherent stress: 0.09. Samples with > 162 000 read‐pairs were included in the analysis (excluding samples 18 (AOKAL), 20 (AOPL), 25 (LUSI1) and 26 (LUSI6), having < 32 000 read‐pairs; Supporting Information Table S2). The colours represent location of sampling: A. ostenfeldii – strains were isolated as resting cysts from four locations in the Baltic Sea Proper (AOF – Föglö Archipelago, Åland, AOKAL – Kalmar strait, Sweden, AOPL – Hel, Poland, AOVA – Valleviken, Sweden) (Tahvanainen et al., 2012); or level of anthropogenic impact (LUSI index): A. minutum/tamarense – strains were isolated from the north west Mediterranean Sea as vegetative cells, from locations with varying levels of anthropogenic impacts estimated using the LUSI index, 1–6 (E. Flo, unpubl.; Supporting Information Fig. S4); and sample ID for each strain (Supporting Information Fig. S4 and Table S2). The confidence limit was set at 0.80.

Figure 2

Composition of microbiomes of Alexandrium isolates at (A) Class and (B) Family level. Samples are grouped by dissimilarity (Supporting Information Fig. S3); AOA‐D: A. ostenfeldii group A–D, AMA‐B: A. minutum/tamarense group A–B. The cultures were harvested during late exponential phase using 0.2 μm, 47 mm filters and DNA was extracted (Boström et al., 2004). The V3‐V4 region of the 16S rRNA gene was amplified using primers 341F (CCTACGGGNGGCWGCAG) and 805R (GACTACHVGGGTATCTAATCC) and sequenced using Illumina MiSeq (Herlemann et al., 2011; Hugerth et al., 2014). The read‐pairs were clustered into OTUs using Usearch v8.1 (radius 1.5) (Edgar, 2013) corresponding to ~97% sequence identity and singletons were removed (Supporting Information Table S2). OTUs were classified using SILVA db 123 SSURef NR99; (Quast et al., 2013) using SINA v. 1.2.13 (Pruesse et al., 2012). The graph was constructed using ggplot2 (v. 2.2.1) (Wickham, 2009). The numbers for each sample specify the ID of each strain given in Supporting Information Fig. S4. The sequence data have been submitted to the European Nucleotide Archive (ENA) database under accession numbers ERS1617530‐ERS1617557.

Figure 3

Dynamics of OTUs with a relative abundance > 0.2% and a similarity of ≥ 95% to members from the core microbiome of A. ostenfeldii, in (A) a naturally occurring spring bloom, Prodiversa, 2013 (Bunse et al., 2016; Godhe et al., 2016) and (B) a time series covering April–October in a coastal to off‐shore transect, Planfish, 2010–2011 (Legrand et al., 2015), both from the Baltic Sea Proper. Relative abundance (%) of OTUs on the left y‐axis, the dinoflagellate biomass dynamics from the corresponding datasets on the right y‐axis (%). The legend lists Planfish/Prodiversa‐OTUs named by corresponding OUTs from the A. ostenfeldii dataset, including their taxonomic affiliation at family level. The relative abundances of dinoflagellate biomass were calculated based on the carbon content of dinoflagellates per total phytoplankton carbon, including the species: Dinobryon spp, Dinophysis sp, Gymnodiniales spp, Gymnodinium spp, Gyrodinium spp, Heterocapsa rotundata, Katodinium glaucum, Peridinella catenata, Peridinella (single cell), Protoperidinium spp and Scrippsiella. Plots were made using ggplot2 (v. 2.2.1) (Wickham, 2009) in R 3.4.0.