Friends and foes: symbiotic and algicidal bacterial influence on Karenia brevis blooms

Abstract

Harmful Algal Blooms (HABs) of the toxigenic dinoflagellate Karenia brevis (KB) are pivotal in structuring the ecosystem of the Gulf of Mexico (GoM), decimating coastal ecology, local economies, and human health. Bacterial communities associated with toxigenic phytoplankton species play an important role in influencing toxin production in the laboratory, supplying essential factors to phytoplankton and even killing blooming species. However, our knowledge of the prevalence of these mechanisms during HAB events is limited, especially for KB blooms. Here, we introduced native microbial communities from the GoM, collected during two phases of a Karenia bloom, into KB laboratory cultures. Using bacterial isolation, physiological experiments, and shotgun metagenomic sequencing, we identified both putative enhancers and mitigators of KB blooms. Metagenome-assembled genomes from the Roseobacter clade showed strong correlations with KB populations during HABs, akin to symbionts. A bacterial isolate from this group of metagenome-assembled genomes, Mameliella alba, alleviated vitamin limitations of KB by providing it with vitamins B₁, B₇ and B₁₂. Conversely, bacterial isolates belonging to Bacteroidetes and Gammaproteobacteria, Croceibacter atlanticus, and Pseudoalteromonas spongiae, respectively, exhibited strong algicidal properties against KB. We identified a serine protease homolog in P. spongiae that putatively drives the algicidal activity in this isolate. While the algicidal mechanism in C. atlanticus is unknown, we demonstrated the efficiency of C. atlanticus to mitigate KB growth in blooms from the GoM. Our results highlight the importance of specific bacteria in influencing the dynamics of HABs and suggest strategies for future HAB management.

Keywords: HABs; algicidal bacteria; dinoflagellates; harmful algal blooms; phytoplankton–bacteria interactions; vitamins.

Figure 1

Influence of microbial communities from the GoM on the growth of KB. (A) Experimental schematic illustrating the sequential filtration done to acquire BF and VF microbial communities and seawater (seawater control) fractions from KB GoM blooms with collection times. Subsequently, these fractions were incubated with KB cultures. “Seawater control” and an additional KB culture control (control), which received no additions, served as negative controls. Incubations were ultimately used for DNA sequencing and bacterial isolation. Illustration created with BioRender.com. (B) Cell densities of KB in CH1 for each condition. (C) Total bacterial cell density in CH1 across each condition. Arrows indicate the time points selected for DNA extraction. All error bars represent ranges of duplicate cultures.

Figure 2

Bacterial diversity and composition in CH1 and CH2 experiments. PCoA based on 16S rRNA ASVs (A) and shotgun metagenomes (B) for CH1 (circles) and CH2 (triangles) experiments. Dots are colored according to their respective groups, with 95% confidence ellipses provided for each experiment (P = .001 for A, P = .002 for B). (C) Bar plot depicting the taxonomic composition of the top 20 bacterial classes in CH1 and CH2 based on metagenomic data. Key bacterial groups are highlighted: Pseudomonadota (squares), PVC group (circles), and FCB group (stars).

Figure 3

Phylogenomic relationships and relative abundance of bacterial MAGs across CH1 and CH2. (A) Unrooted maximum-likelihood phylogenomic tree derived from 71 conserved bacterial markers in anvi’o v7, visualized with iTOL v5. Outer track: pie charts represent MAG abundance across the control, BF and VF groups. Middle track: colors indicate bacterial class. Innermost track: triangles denote MAG source (CH1 or CH2). Circle dots at branches indicate bootstrap values >95%, and diamond dots at nodes represent individual MAGs. (B) MAG relative abundances in CH1 (left) and CH2 (right) experiments. Abundances are represented as z-score transformed positive or negative values. The class of each MAG is represented by the left column in each heatmap. ^* indicates consensus MAGs between CH1 and CH2.

Figure 4

Abundance of potential KB symbiotic MAGs during a GoM KB bloom. (A) Monthly cell counts of KB and total phytoplankton collected at station EH25 in the GoM. Background shade colors represent different bloom intensity levels as defined by the Florida Fish and Wildlife Research Institute (https://myfwc.com/research/redtide/statewide/). (B) Top graphs: Relative abundance of 5 potential symbiotic MAGs based on metagenomic read recruitment overlaid with KB cell counts. Bottom graphs: Linear regression fit of MAG relative abundance and KB cell counts. Each dot represents a monthly sample. The solid lines represent the linear fit, while the dotted lines depict the 95% confidence intervals of the fit.

Figure 5

Interaction of KB with select bacterial strains from CH1. (A) Maximum-likelihood phylogenetic tree based on partial 16S rRNA sequences, depicting the taxonomy of bacteria isolated from CH1. Bacterial classes are differentiated by branch colors, while different genera are highlighted in different background colors. Bacterial isolates are in bold. Circle dots at branches indicate bootstrap values >90%. (B) Co-culture experiments of KB with three bacterial isolates. Karenia growth was monitored using relative chlorophyll a fluorescence (RFU). Significant differences of RFU are labeled where they first occur, with P < .05 (^*), and P < .0001 (^****). (C) Algicidal ratios exhibited by P. spongiae CE15 and C. atlanticus CE21 in co-culture with KB. ^****P < .0001. (D) Co-culture of KB under vitamins B₁, B₇ and B₁₂ limitation in the presence or absence of M. alba CE5. KB was cultured under vitamin-limited conditions for 20 days (left curve) followed by fresh culture inoculation into vitamin-limited media with or without M. alba CE5 (right curves). All error bars represent standard deviations from triplicate experiments.

Figure 6

Deciphering the molecular mechanisms of putative symbiotic and algicidal bacteria associated with KB. (A) Pangenomic analysis of 2957 gene clusters across 20 MAGs and isolate genomes. The central dendrogram shows hierarchical clustering based on gene cluster presence (dark) or absence (opaque). Each track represents a MAG or genome, while track colors indicate the putative type of interaction with KB (pink = symbiotic, light blue and dark blue = algicidal). Important genes from different genomic groups are annotated around the pangenome’s periphery and colored into four distinct groups based on their functions. Only gene clusters present in ≥5 MAGs/genomes are shown; unique gene clusters are not displayed. (B) Cobalamin biosynthesis pathway in putative symbiotic MAGs and M. alba CE5. Gene names in red are fully present in the pink tracks in (A), while names highlighted in green are absent in putative algicidal genomes. (C) Inhibition of KB by P. spongiae CE15 extracellular protease. KB cell counts and protease activity were measured under conditions of filtrate (F), protease inhibitor PMFS (I), and marine broth (MB). P values were evaluated through one-way ANOVA, followed by Dunnett’s multiple-comparison test. (D) Geographic distribution of the algicidal serine protease, EspI, from P. spongiae CE15 in the Tara Oceans database. Colors in the donut plots represent the taxonomic distribution of this protein in different taxa in surface waters (from 0.22 to 3 μm). Numbers denote Tara Oceans stations. (E) Inhibition of KB natural populations during a bloom in the GoM by C. atlanticus CE21. Water was collected at station TC16 and incubated for 24 hours (see Materials and methods). Error bars represent the SD of triplicate incubations.