A bloom of a single bacterium shapes the microbiome during outdoor diatom cultivation collapse

Abstract

ABSTRACT

Algae-dominated ecosystems are fundamentally influenced by their microbiome. We lack information on the identity and function of bacteria that specialize in consuming algal-derived dissolved organic matter in high algal density ecosystems such as outdoor algal ponds used for biofuel production. Here, we describe the metagenomic and metaproteomic signatures of a single bacterial strain that bloomed during a population-wide crash of the diatom, Phaeodactylum tricornutum, grown in outdoor ponds. 16S rRNA gene data indicated that a single Kordia sp. strain (family Flavobacteriaceae) contributed up to 93% of the bacterial community during P. tricornutum demise. Kordia sp. expressed proteins linked to microbial antagonism and biopolymer breakdown, which likely contributed to its dominance over other microbial taxa during diatom demise. Analysis of accompanying downstream microbiota (primarily of the Rhodobacteraceae family) provided evidence that cross-feeding may be a pathway supporting microbial diversity during diatom demise. In situ and laboratory data with a different strain suggested that Kordia was a primary degrader of biopolymers during algal demise, and co-occurring Rhodobacteraceae exploited degradation molecules for carbon. An analysis of 30 Rhodobacteraceae metagenome assembled genomes suggested that algal pond Rhodobacteraceae commonly harbored pathways to use diverse carbon and energy sources, including carbon monoxide, which may have contributed to the prevalence of this taxonomic group within the ponds. These observations further constrain the roles of functionally distinct heterotrophic bacteria in algal microbiomes, demonstrating how a single dominant bacterium, specialized in processing senescing or dead algal biomass, shapes the microbial community of outdoor algal biofuel ponds.

IMPORTANCE

Aquatic biogeochemical cycles are dictated by the activity of diverse microbes inhabiting the algal microbiome. Outdoor biofuel ponds provide a setting analogous to aquatic algal blooms, where monocultures of fast-growing algae reach high cellular densities. Information on the microbial ecology of this setting is lacking, and so we employed metagenomics and metaproteomics to understand the metabolic roles of bacteria present within four replicated outdoor ponds inoculated with the diatom Phaeodactylum tricornutum. Unexpectedly, after 29 days of cultivation, all four ponds crashed concurrently with a “bloom” of a single taxon assigned to the Kordia bacterial genus. We assessed how this dominant taxon influenced the chemical and microbial fate of the ponds following the crash, with the hypothesis that it was primarily responsible for processing senescent/dead algal biomass and providing the surrounding microbiota with carbon. Overall, these findings provide insight into the roles of microbes specialized in processing algal organic matter and enhance our understanding of biofuel pond microbial ecology.

INTRODUCTION

Algal-bacterial interactions fundamentally influence aquatic biogeochemistry and ecology (1–4). Algae support a diverse bacterial community (the algal microbiome) by exuding dissolved organic matter (DOM), and bacteria support algae through metabolite production and nutrient remineralization (5–11). Bacteria can also negatively impact algal hosts through competition or directly via algicidal mechanisms (10, 12, 13). A better understanding of the feedback resulting from algal-bacterial interactions during algal growth and demise is needed to predict the role of the algal microbiome in natural and engineered ecosystems. Moreover, uncovering how bacteria interact with each other within the algal microbiome will allow more accurate quantification of resource recycling across the algal lifecycle.

Algal microbiomes are functionally diverse, leading to partitioning of DOM processing and assimilation (14–16). Furthermore, a temporal succession of different bacterial taxa is well documented within marine algal blooms (17–20). Generally, in early algal growth, the DOM pool is composed of low molecular weight molecules (e.g., amino/organic acids, sugars) exuded by the algae (2, 14, 21). These, in turn, enrich for small-molecule users such as the Rhodobacteraceae, a group commonly found in algal microbiomes (14, 22). During algal demise, high molecular weight molecules (e.g., polysaccharides, proteins, lipids) from lysis or exudation dominate the DOM pool (14, 23) and are directly consumed by biopolymer degraders like the Flavobacteriaceae (19, 22, 24). Changes in DOM composition over time lead to a predictable succession of bacteria along the algal lifecycle.

The mass cultivation of microalgae in outdoor raceway ponds is considered to be an economically viable option for the production of biofuels, a renewable source of energy (25, 26). However, these systems are susceptible to changes in the surrounding environment, increasing the potential for system-wide algal pond crashes (27). These crashes can result from abiotic (e.g., temperature fluctuations and other suboptimal weather conditions) and/or biotic (e.g., viral, bacterial, eukaryotic infection) factors, all of which can have differing downstream effects on algal pond biogeochemistry (27, 28). However, the agent(s) causing pond crashes are rarely identified despite the need to mitigate these events. Moreover, because pond biogeochemistry is majorly dependent on the activity of heterotrophic bacteria via remineralization of carbon and nitrogen, for example, an understanding of how the microbiome recycles residual resources following pond crashes is needed.

Here, we document the microbial underpinnings associated with an algal population crash that occurred during an outdoor cultivation effort. We followed a month-long time course of algal raceway ponds inoculated with two distinct host algae: Phaeodactylum tricornutum (a diatom) and Microchloropsis salina (a chlorophyte). Using molecular tools (16S rRNA gene amplicon sequencing, metagenomics, and proteomics), we characterized the microbiome dynamics across time for the two algal hosts. A crash was observed only in P. tricornutum ponds after ~29 days of cultivation, which corresponded with a bloom in a single amplicon sequence variant assigned to the genus Kordia (family Flavobacteriaceae). Although we were unable to isolate the bacterium to test whether it could cause algal demise in the laboratory, we suspected that this bacterium contributed to P. tricornutum demise, as strains within the Kordia genus are capable of killing this model diatom (29) and other algae (30). Nevertheless, the profound shift in the microbiome during the pond crash led us to investigate potential mechanisms contributing to Kordia’s dominance and its downstream impact on the surrounding microbiota. We initially hypothesized that microbial taxonomic and functional composition would mirror DOM availability along the algal life cycle. We also performed a laboratory experiment to help explain patterns observed in the community-wide response to algal demise, suggesting cross-feeding is a mechanism in structuring the microbial community when algae die, specifically between Flavobacteriaceae and Rhodobacteraceae. Collectively, these results provide insight into the ecophysiology of Kordia sp. and emphasize the importance of considering the role of bacterial-bacterial interactions within the algal microbiome, both in nature and industrial settings, to understand the fate of algal-derived carbon.

MATERIALS AND METHODS

Algal raceway pond time series design

Eight 557 L algal raceway (ARW) ponds were inoculated with either M. salina (CCMP 1776, formerly Nannochloropsis salina; ARW4, 6, 9, & 11) or P. tricornutum (“Flour Bluff” isolate; ARW1, 5, 7, and 12) using natural, diatomaceous earth filtered seawater from Laguna Madre, Corpus Christi, TX, on 27 October 2015 (Fig. S1). Cultures were amended with 2.0 mM NH₄Cl, 2.0 mM pH-balanced H₃PO₄, and 0.07 mM FeSO₄. The first harvest (6 November 2015) removed 75% of the cultures with media replacement and was followed by 50% harvests (with media replacement) on 10, 13, 17, 20, and 24 November. Samples for ash-free dry weight (AFDW) were collected on 27 October (day 1), 2 November (day 7), 11 November (day 16), 17 November (day 22), 24 November (day 29), 27 November (day 32). Due to visible clearing of P. tricornutum biomass, samples for AFDW were only collected for M. salina ponds on 2 December (day 37) and 3 December (day 38; Fig. S1). Samples for sequencing were collected from all ponds on days 1, 7, 29, and 38 (Fig. S1). pH, salinity, temperature, and weather station data were collected daily. Each raceway had a pinpoint marine pH controller that operated a solenoid to inject CO₂ into the raceway through a ceramic airstone based on raceway pH. Water samples were shipped overnight to Lawrence Livermore National Laboratory (LLNL) on blue ice in polycarbonate clear bottles and processed for DNA extraction as outlined below. The remaining live samples were incubated in plant growth chambers under a 14/10 L/D cycle at 200 µmol quanta m⁻² s⁻² and 22°C for 1 week. To test whether a member of the bacterial community could be responsible for crashes, a sample from the ARW7 pond from day 32 was filtered with a 0.8 µm syringe filter, and 1 mL of this filtrate was added to a 100 mL flask of a previously axenic P. tricornutum CCMP 2561 culture in instant ocean F/2 medium. This flask showed visibly lower algal cell density and a cloudy color after incubation for 1 week, compared to a no-addition flask. This addition was repeated two more times, after which the crashing could no longer be propagated. Because we lost the activity before attempting to cultivate the microbe on standard media, we were unable to isolate the bacterium.

16S/18S rRNA gene amplicon sequencing

Overnight shipped samples were immediately filtered onto 0.22 µm pore size SUPOR filters to capture both algal and bacterial biomass and extracted for DNA using the Qiagen DNEasy kit. The 16S and 18S rRNA genes were amplified using 16S V4 primers 515F (31) and 806R (32) and 18S rRNA V4 primers 565F and 908R (33). Amplicons were sequenced through the Joint Genome Institute’s Community Sequencing Program using the Illumina MiSeq (JGI CSP). DADA2 v1.6.0 (34) was used for quality control, and removeBimeraDenovo was used to remove chimeras. Silva v138 (35) was used to assign taxonomy to the 18S/16S amplicon sequence variants (ASVs). The 16S ASVs were further assigned with the RDP classifier v2.11 (36) against training set 16.

Metagenomes

To generate a library of metagenome-assembled genomes (MAGs) in the ponds, total DNA extracted from two out of four P. tricornutum ponds (ARW1 and ARW7) was submitted for metagenome sequencing through JGI’s CSP with the Illumina NovaSeq S4 (Fig. S1). Raw reads were quality controlled using the JGI Standard Operating Procedure (37). Two co-assemblies were generated using QC’d reads across each pond using metaSPAdes v3.13.0 with kmers 21, 33, and 55 (38). Binning was done inside metawrap v1.3 (39) with metabat v2.12.1, maxbin v2.2.6, and concoct v1.0.0. Bins were assessed with checkM v1.1.3 (40), and those with a completion >50% and contamination <10% were kept and reassembled as individual assemblies using SPAdes v3.13.0 (41). Resultant MAGs were dereplicated into a set of non-redundant MAGs using dRep v2.2.3 (42). To estimate MAG relative abundances, QC’d reads were mapped to the MAGs with bbmap v35.85 (43) and perfectmode = t. MAG abundance was calculated using the mean of the median fold coverage (“MAG abundance”) in each sample (44).

MAG taxonomy was assigned with GTDB-tk v.1.0.0 r202 using default parameters (45). The PATRIC pipeline (46, 47) was used for gene calling and functional annotation. GATOR (https://github.com/jeffkimbrel/gator) was used in parallel to assess metabolic pathway completeness, curated for pathways relevant to algal microbiomes and unique carbon/energy utilization pathways. Carbohydrate-active enzyme (CAZyme) Hidden Markov models (HMMs) were queried using dbcan12 with dbCAN2 (48). Orthofinder 2.5.5 (49–51) was used to find orthogroups, and a species tree was generated using default parameters and annotated using iTol v5 (52). Genomes for 6 Kordia species (Kordia algicida OT-1 [GCF_000154725.1], Kordia periserrulae [GCF_003054265.1], Kordia antarctica [GCF_009901525.1], Kordia jejudonensis [GCF_001005315.1], Kordia zhangzhouensis [GCF_001005305.1], and Kordia sp. SMS9 [GCF_003352465.1]) were downloaded from NCBI. Amino acid sequences of these strains and our Kordia sp. MAG (“mARW1_16”) were clustered with Orthofinder v2.3.11 (49) to obtain a species phylogenomic tree. Average nucleotide identity (ANI) was calculated on the genome sequences using FastANI v1.3 (53).

Metaproteomes

Total proteins were extracted following Mayali et al. (11). A Waters nano-Acquity dual pumping UPLC system (Milford, MA) was configured for online trapping of a 5 µL injection at 5 µL/min with reverse-flow elution onto the analytical column at 300 nL/min. MS analysis was performed using a Q-Exactive HF mass spectrometer (Thermo Scientific, San Jose, CA) outfitted with a home-made nano-electrospray ionization interface (see Supplemental methods for more detail). Protein spectra were mapped to a database of assembled MAGs (1,134,297 sequences), a P. tricornutum genome [10,408 sequences, Phaeodactylum tricornutum CCAP1055.1 (54)], public M. salina/Nannochloropsis genomes (10,964 sequences, NCBI TaxID 5748) obtained in March 2021, and 16 common contaminant proteins. Sequences were processed using Protein Digestion Simulator (https://github.com/PNNL-Comp-Mass-Spec/Protein-Digestion-Simulator). MSGF+ (55) was used to identify peptides against the custom protein database in target/decoy mode with 20 ppm parent ion tolerance, tryptic rule without post-translational modifications considered. Best MSGF+ search matches were filtered at 1% FDR, and MASIC (56) was used to pull identified peptide abundances. Only protein-specific peptides (peptides unique to the protein in the whole protein collection) were used in subsequent analysis and aggregation. Protein abundance is reported as normalized spectral abundance factor (NSAF) as described in reference 57. The Kordia sp. (MAG “mARW1_16”) expressed proteins were manually assigned KEGG functional categories using the PATRIC gene calls. GATOR pathways were merged with this set of proteins, and data were plotted using ggplot2 (58).

Cross-feeding laboratory experiment

We tested whether pre-conditioning of P. tricornutum lysate (analogous to demise conditions) by Kordia algicida OT-1, as the primary degrader, enhanced the subsequent growth of a representative Rhodobacteraceae strain (Sulfitobacter sp. N5S). OT-1 was purchased from the Japanese National Institute of Technology and Evaluation (NITE) Biological Resource Center (NBRC Accession #100033). We annotated the public genome as described above. Strain N5S was isolated from Bodega Bay, CA (38.332615, –123.048296) by spreading whole seawater onto marine agar plates, and its genome was sequenced by JGI (NCBI Accession # PRJNA581036). See the Supplemental Methods for details.

P. tricornutum CCMP 2561 lysate was generated using an axenic exponentially growing culture in sterile F/2 media as described in Supplemental Methods (axenicity checked with DAPI staining and fluorescence microscopy). Algal cells were pelleted by spinning at 4,500 × g for 8 min, and the spent media discarded to remove algal exudate. Cells resuspended in sterile F/2 were lysed using an Ultrasonic Processer XL sonicator (Misonix) and filtered through a 0.8 µm syringe filter to remove intact cells. The final lysate was then incubated with or without K. algicida to produce “abiotic” and Kordia “conditioned” lysate. This experiment was carried out in triplicate in 3 mL sterile borosilicate tubes at 22°C in the dark for 5 days. Growth was monitored using OD₆₀₀ (BioTek Cytation 5 plate reader, Agilent) and flow cytometry. Strains OT-1 and N5S were then inoculated into 0.2 µm filtered abiotic versus conditioned lysate in five replicates in a 96-well plate (rounded bottom). Negative controls (no OT-1 and N5S bacteria added) and a positive control (high-nutrient Zobell media) were included. OD₆₀₀ absorbance was collected every 20 min for 72 h using a BioTek Cytation 5 plate reader at 22°C in the dark. Flow cytometry samples were collected at T = 0 h and T = 72 h.

Statistical methods

The R software platform (59) was used for statistics and data visualization (ggplot2 [58]). Differential abundance of ASVs comparing day 1 versus day 38 was done with ANCOM-BC (v.2.2.2) (60). The VEGAN (v.2.6-4) (61)) package was used for the following: nonmetric multidimensional scaling (NMDS) based on Bray-Curtis distances, adonis function for significance of clusters, and diversity function for Shannon and Simpson’s indices. To compare normalized protein abundance across time, an ordinary one-way ANOVA followed by the Benjamini-Hochberg adaptive false discovery rate (FDR) procedure to correct for multiple comparisons was conducted using GraphPad Prism 9. For the conditioning experiment, unpaired two-tailed t-tests were used to compare cell abundances at T = 72 h using GraphPad Prism 9. Growth rates were calculated on linear regressions of exponential phase growth (OD₆₀₀) using GraphPad Prism 9.

RESULTS

Differential persistence of the two algal species and evidence of P. tricornutum population decline in the outdoor algal raceway ponds

During concurrent cultivation of 8 co-located ARW ponds (four inoculated with Phaeodactylum tricornutum, and four inoculated with Microchloropsis salina), an algal crash occurred only in the four P. tricornutum ponds after 29 days of semi-continuous cultivation. A visible change in color from dark brown (healthy P. tricornutum biomass) to a milky white was noted. A steady decline in algal biomass (~0.3 to <0.1 g/L) across the P. tricornutum ponds occurred (Fig. S2; Table S1), but not for M. salina after day 29 (Fig. S2). During cultivation, pond salinity ranged from 31 to 35 ppt, pH from 7.38 to 8.6, and water temperature from 18.0°C to 28.0°C (Table S2).

To further quantify P. tricornutum demise, we used 18S rRNA gene profiling, focusing on the genera assignments Microchloropsis or Phaeodactylum within their respective ponds. Microchloropsis ponds exhibited stable host 18S rRNA relative abundance, whereas P. tricornutum ponds significantly declined in the relative abundance of host 18S rRNA at day 38 (5.8-log fold change between day 1 versus day 38 [ANCOMBC, P_adj = 0.0001]; Fig. 1A; Table S3). Proteomics also indicated a declining trend in proteins identified as P. tricornutum on days 29 and 38, with negligible proteins identified as M. salina (Fig. S3). Therefore, we refer to this event as a population-wide “crash,” with “demise” referring to the steady decline in ash-free dry weight of P. tricornutum after day 16 (Fig. S2).

Fig 1

Algal host strain and decline in P. tricornutum over time influence microbiome composition within the algal raceway ponds. (A) Relative abundance of representative host 18S rRNA ASVs between ponds inoculated with M. salina versus P. tricornutum. Error bars represent the standard deviation in relative abundance between four replicate ponds. 18S rRNA amplicons assigned to either Phaeodactylum or Microchloropsis are shown in either brown or green, where all other 18S rRNA amplicons are shown as “Other” in gray. (B) NMDS ordination based on the Bray-Curtis dissimilarity matrix of 16S rRNA bacterial communities between samples inoculated with M. salina (circles) and P. tricornutum (triangles) over timepoint (indicated by color); 95% confidence ellipses (stat_ellipse) are grouped by algal host inoculum. Stress = 0.115.

Bacterial community composition responded to algal host and health

Each algal host had distinct microbiome compositions that shifted with time (Fig. 1B), especially pronounced for P. tricornutum ponds (Fig. S4; presumably due to demise). NMDS ordination clustered 16S rRNA gene microbiome profiles based on algal host (Fig. 1B; ADONIS P < 0.0001). The dominant bacterial groups differed between the host alga (Fig. S5). M. salina ponds were dominated by ASVs assigned to the Alpha- and Betaproteobacteria, Sphingobacteriia, and “unclassified” Pseudomonadota (Fig. S5). The P. tricornutum ponds were dominated by ASVs assigned to the Alpha-, Gamma-, and Betaproteobacteria at days 1 and 7 but shifted to predominantly Flavobacteriia and Alphaproteobacteria classes at days 29 and 38 (Fig. S5). We focused further analyses on the P. tricornutum pond microbiome due to evidence of algal population demise. Based on NMDS clustering, two major clusters of ASVs in the P. tricornutum microbiome emerged (Fig. S4a; ADONIS P < 0.005). These clusters corresponded to P. tricornutum growth phases that occurred in the four replicated ARW ponds: “growth” (days 1 and 7) and “demise” (days 29 and 38). The alpha-diversity of the demise microbiome was significantly reduced compared to the growth microbiome (Fig. S4B).

Taxon-specific response to P. tricornutum demise—identification of a highly abundant bacterium, Kordia sp.

Examination of individual ASVs across the P. tricornutum growth phases revealed an ASV-specific succession (Fig. 2). One ASV assigned to the genus Kordia (ASV_4) dominated the demise phase (reaching up to 93% relative abundance of the 16S rRNA reads, Fig. 2), and increased 12.5-log fold on day 38 versus day 1 (ANCOMBC, P_adj = 3.47e−75, Table S4). ASV_4 was nearly undetectable during growth timepoints (Fig. 2) and had no detectable reads in the M. salina microbiome (Fig. S6). Aside from ASV_4, the demise microbiome was primarily composed of many ASVs assigned to the order Rhodobacterales, with ASV_24 (Nioella sp.) reaching up to ~6% relative abundance within a sample (Fig. 2). By contrast, the growth microbiome had greater diversity of order-level taxonomic affiliations, with non-Kordia Flavobacteriales, Enterobacterales, Cytophagales, and Sphingomonadales contributing high relative abundances (Fig. 2). Several Rhodobacterales-, Hyphomicrobiales-, and Nitrosococcales-assigned ASVs persisted throughout both phases (Fig. 2).

Fig 2

Shift in microbiome composition during P. tricornutum demise corresponds to a bloom of an individual ASV assigned to the genus, Kordia, and co-occurring taxa. Each pond replicate is shown on the x-axis, and individual ASVs (shown with ASV number and lowest taxonomic assignment) are on the y-axis. The timepoints are grouped and color-coded by algal growth phase, which is based on NMDS clustering of the relative abundance of ASVs in Fig. S4 (growth [purple line] versus demise [orange line]). Each bubble represents the relative abundance value for each replicate, and the averaged relative abundance across all the plotted samples for each ASV is shown on the bar graph to the right. Bubbles are color-coded by the taxonomic assignment to the order level according to the SILVA v138 database.

MAGs were generated for two of four raceways from the P. tricornutum ponds (“ARW1” and “ARW7”, Table S5). A total of 143 non-redundant MAGs were assembled, some of which cluster with bacterial strains previously isolated from these ponds and co-cultured with P. tricornutum (11, 62) (Fig. 3). The phyla Myxococcota, Planctomycetes, Bdellovibrionota, Patescibacteria, Bacteroidota, Gammaproteobacteria, and Alphaproteobacteria were also detected (Fig. 3). Within the Alphaproteobacteria, there was high representation (33 MAGs) of the Rhodobacteraceae (Fig. 3). The Kordia MAG (mARW1_16, Fig. 3) had the highest MAG abundance values across the P. tricornutum ponds (average 28% relative abundance, Fig. 3). We also assessed coverage of metagenome reads mapped to the MAGs, with the caveat that these were only done in duplicate. However, like the amplicon data, Flavobacteriaceae and Rhodobacteraceae family-level assignments were the dominant groups during demise, whereas the growth phase had greater diversity (Fig. S7). However, MAGs of the Pirellulaceae, Rhizobiaceae, and Phycisphaeraceae families were also present in the demise phase in lower abundances (Fig. S7).

Fig 3

Distribution of MAGs detected in the P. tricornutum pond metagenomes. The phylogenetic tree shows the relationship between the MAGs based on orthogroups (see Materials and Methods). Cultivated isolates are denoted by their strain identification around the tree as follows: PT13A: Oceanicaulis PT13A, ARW1R1: Algoriphagus ARW1R1, ARW1Y1: Muricauda ARW1Y1, ARW7G5Y1: Arenibacter ARW7G5Y1, PTEAB7WZ: Devosia PTEAB7WZ, ARW1T: Stappia ARW1T, PT6CLA: Rhodophyticola PT6CLA, PT4BL: Yoonia PT4BL, EA3: Pusillimonas EA3, EA2: Alcanivorax EA2, PT6ES: Henriciella PT6ES, EA10: Tepidicaulis EA10, PT13C1: Roseibium PT13C1, 11-3: Thalassospira 11-3, PT3-2: Marinobacter 3-2, ARW7G5W: Muricauda ARW7G5W, PT4D: Oceanicaulis PT4D, N2S: Roseobacter N2S, N5S: Sulfitobacter N5S. The relative abundance range of MAGs averaged across all samples is shown as bubbles surrounding the tree. MAG completeness shown as a heatmap ranging from 50% to 100% completeness. Clades are color-coded by Class-level GTDB taxonomy, and to the Family level for Alphaproteobacteria.

Members of the Kordia genus have been implicated in the demise of marine harmful algal blooms (63, 64), and the Kordia MAG was similar to a known algicidal Kordia-sequenced genome (K. algicida OT-1; Fig. S8) with 2,996 shared orthologs and an average nucleotide identity of 81% (65). Both the OT-1 genomes and the Kordia MAG include a protease previously hypothesized to be involved in algicidal activity (29). Although we lack direct evidence that Kordia was responsible for P. tricornutum's demise, it is reasonable to hypothesize that it played a role in this process. One indication is that a single genotype dominated the microbial community prior to and during the diatom collapse, similar to other bacterial pathogens (66, 67). In addition, previous investigations have implicated this genus in the demise of natural algal blooms, both through laboratory experiments and quantification throughout the bloom cycle (30). Furthermore, at the time of sampling the ponds, we were able to propagate algicidal activity from a live and 0.8 µm filtered ARW7 sample into previously axenic P. tricornutum cultures, suggesting a bacterial component could recreate an algal culture crash. We were not successful in maintaining this algicidal community and thus could not directly attribute P. tricornutum death to Kordia.

Expressed proteins mapped to Kordia sp. provide insight into mechanisms leading to its dominance during algal pond demise

Using metaproteomics of P. tricornutum pond samples over time, we identified expressed proteins mapping to the Kordia MAG (mARW1_16) to gain insight into the metabolic activity of Kordia during diatom demise. There were 1,728 of 4,252 (40.64%) proteins in the mARW1_16 MAG detected in the metaproteome (Fig. S9; Table S6). Proteins in the Transcription/Translation, Transport, Amino acid metabolism, and Central carbon metabolism categories were most detected, suggesting the Kordia population was metabolically active (Fig. 4A). The next most abundant categories were the Peptidases/proteases & inhibitors and Complex carbon metabolism (Fig. 4A). Furthermore, the Complex carbon metabolism category harbors the most abundant protein detected (based on normalized spectral abundance factor [NSAF]) across demise/crash samples: a SusC-family Ton-B-dependent transporter (Fig. 4B). The gene encoding this protein is located nearby genes encoding glycoside hydrolase family 2 TIM barrel-domain, beta-glucanase, and a putative lamarinase targeting beta-glucan, all detected in the Kordia proteome (Table S7). Likewise, several proteins assigned as SusD surface glycan-binding PUL component (Complex carbon metabolism) were abundant (Fig. 4B).

Fig 4

Abundant protein functional categories and proteins mapped to the Kordia MAG indicate antagonism and biopolymer degradation. (A) Breakdown of the most detected protein functional categories mapped to Kordia. The "Other" category specifies categories with less than 10 detected proteins across all metaproteomics samples. Proteins with "unknown" or "ambiguous" function were excluded for visual purposes. Protein assignments are available in Table S6. (B) The topmost abundant Kordia proteins detected across the demise samples (days 29 and 38) that had an average NSAF ≥ 10. Average NSAF across all demise samples (n = 5) is shown, and their standard deviations are represented as error bars.

The next two most abundant proteins belonged to stress response, defense, and virulence (Fig. 4B). The lysine-epsilon oxidase protein (LodA), known for its antimicrobial activity resulting from hydrogen peroxide production (68), was among the top three most abundant Kordia proteins expressed across algal demise (Fig. 4B; Table S6). In addition, a putative type VI secretion system (T6SS) component was the fourth most abundant protein (Fig. 4B) and is located nearby two expressed T6SS baseplate protein components (TssJ and VgrG) and many other encoded T6SS components (Table S6). It is believed that algicidal Kordia members secrete a protease involved in algal killing (29); thus, we searched for this protein in the proteome. This protein was detected in five samples, all of which were in the demise phase (Table S6). However, it was detected in low abundance (total spectral counts of 23, Table S6), potentially because we only sampled the intracellular proteome, missing the extracellular fraction where it is most likely to be detected.

Genomic analysis indicates Kordia fills a specific niche compared to the co-occurring Rhodobacteraceae community

We initially hypothesized that microbial taxa present during the demise phase may have metabolized complex biomolecules expected to be released from dead or dying diatom cells. We then asked whether taxa in the demise microbiome, other than Kordia sp., were specialized to acquire carbon from biopolymers. We examined the CAZyme profiles of P. tricornutum pond MAGs, focusing on comparing the Kordia sp. MAG to all MAGs assigned to the Rhodobacteraceae family, as these two groups dominated the demise microbiome (Fig. 2; Fig. S7). We counted all CAZyme hits to glycoside hydrolases (GH) and polysaccharide lyases (PL) and summarized them at the individual MAG level (Fig. 5A). We found that Kordia sp. harbored ~255 of these CAZymes, while the Rhodobacteraceae MAGs averaged ~61 CAZymes per MAG (n = 30 MAGs; Fig. 5A). In the Kordia MAG specifically, CAZymes were predicted to target a diverse range of substrates such as beta-glucan (>10), chitin/peptidoglycan (>7), glycogen (5), and several others (Fig. 5B). Many of these were also detected abundantly in the metaproteome during the demise phase (Table S8).

Fig 5

Kordia fills a distinct niche compared to Rhodobacteraceae during algal demise by encoding abundant CAZymes targeting a diverse range of predicted biopolymers. (A) All Rhodobacteraceae non-redundant MAGs at ≥75% completion (n = 30) assembled from either the ARW1 or ARW7 ponds are shown for comparison with the Kordia MAG (boldened). MAG IDs are shown in brackets, followed by their genus-level assignment. Only CAZyme hits to glycoside hydrolases (“GH”) or polysaccharide lyases (“PL”) are included in the summarized CAZyme counts per MAG, and are color-coded by their predicted substrate. Emboldened MAGs indicate taxa that were active during the demise phase, as indicated by metaproteomics (see Fig. S9). (B) Specific CAZyme HMM hits and dbCAN substrate annotations for Kordia (mARW1_16). Hits with no substrate predicted (n = 81) were removed for visual purposes.

Some Rhodobacteraceae MAGs harbored a diverse repertoire of CAZymes (e.g., mARW1_75, mARW1_60); however, MAGs with specific genus-level matches to the ASVs grouped in the demise phase were relatively depleted in diverse CAZymes (Fig. 5A). Furthermore, we did not find many Rhodobacteraceae CAZymes expressed in the proteome (spectral count ≤4 in a given sample; Table S8). Those Rhodobacteraceae CAZymes with the highest total spectral counts were found during the growth phase timepoints (Table S8). By contrast, the most abundantly detected proteins belonging to Rhodobacteraceae MAGs during the demise phase were transporters for sugars and amino acids, such as a glycerol ABC transporter substrate-binding component (mARW1_12.peg.2571), glutamate/glutamine/aspartate/asparagine ABC transporter substrate-binding component (mARW1_23.peg.3083), TRAP-type C4-dicarboxylate transport system periplasmic component (mARW1_20.peg.1354, mARW1_12.peg.44), and a fructose ABC transporter substrate-binding component (mARW1_23.peg.3255, Table S6).

Sulfitobacter, an ecologically relevant Roseobacter-clade member, required Kordia to process P. tricornutum lysate

As Rhodobacteraceae taxa present in the demise microbiome generally lacked the capability to degrade complex carbon sources, we hypothesized that the activity of Kordia sp., a specialized biopolymer degrader, might allow for the growth of these small-molecule users during diatom demise. To test this, we performed a laboratory experiment using a representative Kordia strain (K. algicida OT-1) and an isolate of the Rhodobacteraceae family (Sulfitobacter sp. N5S). This Rhodobacteraceae strain was selected because it is related to relatively abundant MAGs (Sulfitobacter sp. [mARW1_117] and Phaeobacter sp. [mARW1_12]; Fig. 3) in the P. tricornutum ponds and could be classified as a strict small molecule user due to a lack of CAZymes in its genome (Fig. 6A). Furthermore, this genus falls within the Marine Roseobacter Clade, an ecologically relevant and ubiquitous group found in marine environments (69, 70). This isolate also grows in co-culture (growth rate of 0.35 day⁻¹) with P. tricornutum (growth rate of 0.35 day⁻¹) at perpetuity with no added organic matter (Fig. S11).

Fig 6

Sulfitobacter sp. N5S (a representative Rhodobacteraceae strain) requires Kordia to pre-condition P. tricornutum lysate for growth in culture. (A) CAZyme summary of representative lab strains of Kordia (K. algicida OT1) and a Rhodobacteraceae (Sulfitobacter sp. N5S) used for the experiment. (B) Growth of OT1 versus N5S on P. tricornutum lysate either pre-conditioned by OT-1 (“Conditioned”) or without any added bacteria (“Abiotic” control). Growth curves show the mean and standard deviation (error bars, n = 5) of OD₆₀₀ values taken every 20 minutes. The boxplots to the right show the cell abundances at the final timepoint (n = 4). P-values are generated using unpaired two-tailed t-tests.

K. algicida OT-1 has been shown to have algicidal activity toward P. tricornutum when grown in high-nutrient media (29) and harbors a diverse repertoire of CAZymes targeting complex carbon sources (Fig. 6A). We cultivated K. algicida strain OT-1 in P. tricornutum lysate, an analog to the demise phase of the algal ponds, for 5 days. Then, we filter-sterilized the spent media (“conditioned” lysate) to remove Kordia cells and inoculated this conditioned lysate with Sulfitobacter strain N5S. There was a significantly (P < 0.0001; Table S9) reduced growth rate of Sulfitobacter N5S in the abiotic lysate (0.02 day⁻¹) compared to its growth rate in conditioned lysate (0.09 day⁻¹), with significantly different final cell abundances between the treatments (P < 0.0001; Fig. 6B). K. algicida did not have significantly different (P = 0.0614; Table S9) growth rates in both its own conditioned lysate (0.10 day⁻¹) and abiotic lysate (0.08 day⁻¹) with no significant difference in the final cell abundances (P = 0.5003; Fig. 6B). This suggests Sulfitobacter was dependent on Kordia to access the organic matter present in P. tricornutum lysate. Furthermore, the finding that Kordia did not grow differentially under conditioned versus abiotic lysate suggests that the algal-derived carbon was not fully metabolized during the first incubation, and Kordia limited its growth through an unknown mechanism.

Genomic capability to metabolize alternative compounds in marine Rhodobacteraceae

Members of the Rhodobacteraceae family are globally abundant in the oceans and well known for having the ability to rapidly respond to increased substrate availability and acquire carbon and energy from diverse sources (69, 71, 72). We hypothesized that the prevalence of the Rhodobacteraceae across the algal growth phases is due, in part, to their ability to grow on diverse carbon and energy sources that are typically not available to other microbes when preferred resources (e.g., sugars and amino acids) are rapidly consumed during competition. Thus, we screened all MAGs from the P. tricornutum ponds for pathways that confer the ability to utilize aromatic carbohydrates, fix inorganic carbon, and gain energy from light (Fig. 7). We found that the ability to degrade aromatic compounds such as 4-hydroxybenzoic acid (4HBA), coumaric acid, and phenylacetic acid was common, as was aerobic anoxygenic phototrophy for light-driven metabolism (Fig. 7A). However, 90%–100% of the MAGs in the Rhodobacteraceae, Rhizobiaceae, and Stappiaceae families contained the complete set of genes for carbon monoxide dehydrogenase (COX; coxL + coxM + coxS). Moreover, we detected protein expression in the metaproteome of multiple COX subunits across these families, with mARW1_20 (Nioella sedimensis), mARW1_25 (unclassified Rhodobacteraceae), and mARW7_5 (Hoeflea sp.) recruiting all three COX subunits (Fig. 7B). The expression patterns of individual COX proteins across timepoints were variable, especially during the early growth phase (highest expression on day 1; Fig. S12). On the other hand, we detected sparse protein expression of aromatic compound degradation and light-driven metabolism (Fig. S13). This suggests that carbon monoxide utilization was widespread and active in this group, especially during the growth phase.

Fig 7

Carbon monoxide (CO) oxidation as a potential alternative carbon and energy-generating mechanism used by Rhodobacteraceae in algal ponds. (A) Metabolic potential across all P. tricornutum pond MAGs to use alternative carbon and energy sources. All screened pathways are shown (x-axis), where dark blue tiles indicate 100% of pathway detection (i.e., all known genes involved in that pathway were detected) per-MAG (y-axis, grouped by family). (B) Protein presence/absence of each CODH subunit across MAGs in the metaproteome. Only MAGs that had a detected protein mapped to CODH out of 13 samples are shown.

DISCUSSION

Our study documents a serendipitous algal population crash that occurred during a multi-omics assessment of outdoor P. tricornutum ponds. Initially, we set out to characterize the microbiome dynamics of two algal species cultivated outdoors in pilot-scale biofuel production ponds. Despite close proximity to one another and the open nature of the ponds, we observed distinct microbiomes, shown to be driven by host-specific physiology (73). However, a drop in bacterial diversity during P. tricornutum demise corresponded with a “bloom” of Kordia, a genus belonging to the phylum Bacteroidetes and the Flavobacteriaceae family. The Kordia genus has been readily cultured from a wide range of aquatic environments and exhibits significant taxonomic and genomic diversity (74, 75). As the Kordia sp. reached dominant abundances (up to 93% of the bacterial community), we collected metagenomes and metaproteomes to identify features that may have contributed to its numerical dominance in the outdoor P. tricornutum ponds. In addition, we propose that Kordia acted as a primary biopolymer degrader, influencing the surrounding microbiota during diatom demise by providing partially degraded products. Specifically, cross-feeding of carbon between biopolymer degraders like Kordia (Flavobacteriaceae) and copiotrophic Rhodobacteraceae may be an ecologically relevant interaction that occurs during algal bloom senescence and die-off.

The genus Kordia is primarily known to exhibit an algicidal phenotype against diatom hosts, including P. tricornutum, through protease-mediated interactions (29). Genomic analysis of the Kordia MAG confirmed the presence of an extracellular metalloprotease putatively involved in algal-specific lysis. However, this protein was not detected in high abundance in the metaproteome (total count of 23 across all demise samples) as only intracellular protein pools were collected due to the sampling procedure that only retained biomass (>0.22 µm). Nevertheless, protease/peptidase-like proteins comprised a major fraction of the Kordia proteome during algal demise. These proteins have widely ranging substrate specificities and roles (e.g., antimicrobial, protein folding, biopolymer breakdown, pathogenesis [76]), and particular algicidal proteases could be species-specific, which complicates comparisons with previous studies to validate algicidal activity (77). Although we were able to replicate the diatom crash by incubating the <0.8 µm microbial fraction of crashed ponds with an axenic culture of P. tricornutum, we were unable to attempt isolation due to a loss of activity after several transfers. Therefore, although we lack causal evidence that Kordia was directly responsible for the diatom pond collapse, no other data suggest other mechanisms. The 18S rRNA gene amplicon data did not identify high abundances of known algal grazers or pathogens such as rotifers, aphelids, or chytrids during the demise phase. Diatom-infecting viruses exist (78), and it is possible that viral lysis was a mechanism of culture demise. Slowly decreasing temperature through the fall season could also have contributed to increased stress in the diatom cells, which could have made them more susceptible to biotic stress.

The metaproteomic data suggest Kordia employed mechanisms to inhibit the growth of other microorganisms. One of the most abundantly detected proteins was a putative lysine oxidase (LodA, aka marinocine), a protein shown to be antimicrobial via hydrogen peroxide (H₂O₂) produced during amino acid deamination (79). This L-amino acid oxidase hydrolyzes L-amino acids within cells or extracellularly, producing H₂O₂, which is bactericidal to a wide range of bacteria (79). LodA-driven H₂O₂ production can lead to bactericidal behavior between species under competition and can result in phenotypic variation within biofilm communities (80). Due to the broad-range antimicrobial activity of this oxidase, it is known to also be autotoxic (79). If this process occurred during the Kordia bloom, the population was able to withstand the damaging impacts of H₂O₂. Indeed, Kordia expressed detectable levels of potential detoxification proteins, namely catalase-peroxidase (KatG). The production of these antibacterial pathways suggests that Kordia was inhibiting the growth of other bacteria, which is in agreement with the decreased diversity of the microbial community during the algal demise phase. However, these pathways could also have been directly involved in the algal demise, but the literature is lacking on this topic. In addition, we found several type VI secretion system (T6SS) proteins expressed by Kordia. T6SS is involved in anti-bacterial and eukaryotic antagonism (81–83), a potential strategy to reduce bacterial competition or gain limiting resources (84). It is possible that Kordia employed T6SS activity against P. tricornutum; however, T6SS-guided algal killing has not yet been investigated (84).

In addition to antagonism, Kordia’s bloom during algal demise may be attributed to its superior ability to access diatom-derived biopolymers, and it simply outgrew its competitors. Kordia belongs to the Flavobacteriaceae (phylum Bacteroidota), a family known to degrade and consume biopolymers in aquatic ecosystems (85, 86). Similar to a previous Kordia genomic analysis (74), we found many polysaccharide-degrading CAZymes (87) in the Kordia MAG. Furthermore, the most abundant Kordia proteins detected were assigned to several polysaccharide utilization loci systems used to extracellularly hydrolyze and import biopolymer carbon using TonB transport systems (24, 86). Although the substrates cannot currently be determined based on sequences alone, these proteins are located upstream of expressed CAZymes putatively targeting beta-glucans commonly found in diatom cell walls (88). It is not clear whether these systems were involved in diatom lysis or if they were activated following lysis; however, these observations supported our initial hypothesis that biopolymer degradation and the Flavobacteriaceae would be prevalent during algal demise (14, 17, 18).

During algal demise, the diversity of other putative biopolymer degraders present initially during growth was reduced, and only one other Flavobacteriaceae (Muricauda sp.) remained. We hypothesize that Kordia sp. removed competing bacteria, either by being optimized to demise conditions and/or by antibacterial modes (e.g., LodA and T6SS). However, several distinct genera of the Rhodobacteraceae co-occurred with Kordia, some assigned to the Marine Roseobacter Clade (e.g., Roseovarius and Phaeobacter [69, 70]). Significant metabolic versatility within the Rhodobacteraceae may confer adaptation to the widely ranging conditions along algal blooms (14, 18, 19). Nevertheless, the Rhodobacteraceae commonly specialize in the uptake of algal-released compounds during growth and development (20, 22). Specifically, Rhodobacteraceae taxa are enriched in transporters for dicarboxylic acids, sugars, dimethylsulfoniopropionate, and polyamines (71, 89, 90). Our analysis of polysaccharide CAZymes did not indicate any unique divergence from typical Rhodobacteraceae (71, 91, 92) to access carbon directly from biopolymers enriched during algal demise.

The appearance of distinct Rhodobacteraceae co-occurring with Kordia conflicted with our initial hypothesis that they would be more diverse during algal growth. Instead, our results collectively indicate that the co-occurrence of distinct Rhodobacteraceae with Kordia sp. during algal demise may be a consequence of cross-feeding. Cross-feeding is a ubiquitous phenomenon that structures microbial communities through cooperation (93), and previous studies have speculated that cross-feeding between Rhodobacteraceae and Flavobacteriaceae led to transparent exopolymer particle degradation within a diatom bloom (94). When biopolymers are abundant, community assembly generally follows “primary degraders” that initially colonize these resources (95, 96). Extracellular biopolymer hydrolysis releases oligo- and monomers into the environment, which are then consumed by “exploiters” (direct uptake) and/or “scavengers” (uptake of exuded metabolites [95, 97–99]). Indeed, based on proteomics transporter expression, Rhodobacteraceae members were actively transporting small molecules such as sugars and amino acids during algal pond demise, which may be a result of the degradation of biopolymers by Kordia. Our laboratory experiment supported this idea by demonstrating that the pre-processing of P. tricornutum lysate by Kordia (a primary degrader) allowed the growth of a representative Marine Roseobacter Clade strain (Sulfitobacter sp. N5S) under conditions analogous to algal demise. It is not known whether Sulfitobacter benefited from direct breakdown products, other exuded or released metabolites, or antimicrobial compound reduction. Yet, these results demonstrate how interspecies (inter-family in this case) interactions, driven by Kordia, can structure the microbiome. Previously, the addition of Kordia algicida to a natural phytoplankton community removed dominant species, released algal nutrients via lysis, and was correlated with increased dissolved organic carbon concentrations, possibly due to biopolymer degradation (100). It is possible that Rhodobacteraceae populations were exploiting diatom-stored nutrients released via lysis in the ARW ponds; however, our experimental system suggests that Sulfitobacter required Kordia conditioning to access DOM. A follow-up experiment might involve testing this hypothesis under mixed culture conditions (99) to investigate ecological interactions between Kordia and Rhodobacteraceae. Furthermore, the structuring of microbiota across the algal life cycle in general likely reflects different cross-feeding networks between microbes that exchange organic molecules and is not necessarily restricted to algal demise conditions.

Although many Rhodobacteraceae taxa emerged during algal demise, other Rhodobacteraceae were also present during the P. tricornutum growth stage, alluding to this group’s generalist lifestyle. We probed the genomic potential across 32 Rhodobacteraceae MAGs that were present in the P. tricornutum ponds to gain insight into potential metabolic strategies of this diverse and abundant community. Oxidation of the trace gas carbon monoxide (CO) can generate supplementary energy or carbon for persistence under “survival” conditions when other sources are limited (101, 102). For Roseobacter Clade members, the CO oxidation pathway is common and hypothesized to be used primarily for energy due to a lack of carbon fixation pathways (71). Here, we propose that CO may have been a viable source of energy, and potentially carbon, for the Rhodobacteraceae populations in the ponds. Indeed, a previous proteomic assessment done by our research group on Rhodophyticola sp. PT6CLA (Rhodobacteraceae) co-cultured with P. tricornutum in the laboratory revealed that CoxMSL was a highly expressed protein cluster (11), suggesting this may be a common biogeochemical transformation occurring in algal microbiomes. This suggests that CO oxidation may contribute significantly to carbon flow in algal ponds and could be a mechanism to reduce carbon emissions into the atmosphere, if the CO oxidation results in CO₂ that is then fixed by microalgae. In addition, aromatic compound degradation pathways were also widespread among the Rhodobacteraceae MAGs, specifically for the degradation of coumaric acid to 4-hydroxybenzoate (4HBA) and 4HBA to protocatechuate, aromatics previously found to be exuded in abundance by P. tricornutum (16). Overall, these observations collectively emphasize the role of flexible carbon and energy metabolisms in structuring the microbiome within algal ponds and provide insight into potential survival mechanisms used by the Rhodobacteraceae when under competition for resources.

Conclusions

Our characterization of a diatom population crash in outdoor ponds suggests Kordia sp. is both an environmentally and industrially relevant organism that can structure both algal and bacterial communities. Our analysis provides insight into potential virulence pathways employed by Kordia that could have led to P. tricornutum demise and/or the removal of bacterial competitors. Although we did not definitively demonstrate that Kordia activity directly led to the pond crash, the metagenomic and metaproteomic data highlight potential markers indicative of ponds that have a likelihood of crashing. This is not only informative for biofuel pond monitoring but also to better understand the state of the organic matter pool of crashed systems following the activity of microbes like Kordia. It is also relevant to the marine biological carbon pump that sequesters carbon through the sinking of dead phytoplankton, as Kordia and organisms with similar ecological roles can lead to algal lysis and control the microbial community that degrades algal lysates. Lastly, we highlight features of the globally relevant Rhodobacteraceae population, suggesting they are well adapted to respond to rapid changes in algal growth and sustain abundances using alternative carbon and energy sources.

References

1 Cole, JJ. . 1982;. Interactions between bacteria and algae in aquatic ecosystems. Annu Rev Ecol Syst; 13, :291–314. doi: [DOI: https://dx.doi.org/10.1146/annurev.es.13.110182.001451]

2 Seymour, JR , Amin, SA , Raina, J-B , Stocker, R. . 2017;. Zooming in on the phycosphere: the ecological interface for phytoplankton-bacteria relationships. Nat Microbiol; 2, :17065. doi: [DOI: https://dx.doi.org/10.1038/nmicrobiol.2017.65] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28555622]

3 Cirri, E , Pohnert, G. . 2019;. Algae-bacteria interactions that balance the planktonic microbiome. New Phytol; 223, :100–106. doi: [DOI: https://dx.doi.org/10.1111/nph.15765] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30825329]

4 Kuhlisch, C , Shemi, A , Barak-Gavish, N , Schatz, D , Vardi, A. . 2024;. Algal blooms in the ocean: hot spots for chemically mediated microbial interactions. Nat Rev Microbiol; 22, :138–154. doi: [DOI: https://dx.doi.org/10.1038/s41579-023-00975-2] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37833328]

5 Wheeler, PA , Kirchman, DL. . 1986;. Utilization of inorganic and organic nitrogen by bacteria in marine systems. Limnol Oceanogr; 31, :998–1009. doi: [DOI: https://dx.doi.org/10.4319/lo.1986.31.5.0998]

6 Croft, MT , Lawrence, AD , Raux-Deery, E , Warren, MJ , Smith, AG. . 2005;. Algae acquire vitamin B12 through a symbiotic relationship with bacteria. Nature; 438, :90–93. doi: [DOI: https://dx.doi.org/10.1038/nature04056] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/16267554]

7 LE, Antoun, H , Bashan, Y. . 2008;. Involvement of indole‐3‐acetic acid produced by the growth‐promoting bacterium Azospirillum spp. in promoting growth of Chlorella vulgaris. J Phycol; 44, :938–947. doi: [DOI: https://dx.doi.org/10.1111/j.1529-8817.2008.00533.x] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27041612]

8 Amin, SA , Green, DH , Hart, MC , Küpper, FC , Sunda, WG , Carrano, CJ. . 2009;. Photolysis of iron-siderophore chelates promotes bacterial-algal mutualism. Proc Natl Acad Sci USA; 106, :17071–17076. doi: [DOI: https://dx.doi.org/10.1073/pnas.0905512106] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19805106]

9 Amin, SA , Hmelo, LR , van Tol, HM , Durham, BP , Carlson, LT , Heal, KR , Morales, RL , Berthiaume, CT , Parker, MS , Djunaedi, B , Ingalls, AE , Parsek, MR , Moran, MA , Armbrust, EV. . 2015;. Interaction and signalling between a cosmopolitan phytoplankton and associated bacteria. Nature; 522, :98–101. doi: [DOI: https://dx.doi.org/10.1038/nature14488] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26017307]

10 Ramanan, R , Kim, B-H , Cho, D-H , Oh, H-M , Kim, H-S. . 2016;. Algae-bacteria interactions: evolution, ecology and emerging applications. Biotechnol Adv; 34, :14–29. doi: [DOI: https://dx.doi.org/10.1016/j.biotechadv.2015.12.003] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26657897]

11 Mayali, Xavier , Samo, TJ , Kimbrel, JA , Morris, MM , Rolison, K , Swink, C , Ramon, C , Kim, Y-M , Munoz-Munoz, N , Nicora, C , Purvine, S , Lipton, M , Stuart, RK , Weber, PK. . 2023;. Single-cell isotope tracing reveals functional guilds of bacteria associated with the diatom Phaeodactylum tricornutum. Nat Commun; 14, : 5642. doi: [DOI: https://dx.doi.org/10.1038/s41467-023-41179-9] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37704622]

12 Mayali, X , Azam, F. . 2004;. Algicidal bacteria in the sea and their impact on algal blooms. J Eukaryot Microbiol; 51, :139–144. doi: [DOI: https://dx.doi.org/10.1111/j.1550-7408.2004.tb00538.x] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15134248]

13 Meyer, N , Bigalke, A , Kaulfuß, A , Pohnert, G. . 2017;. Strategies and ecological roles of algicidal bacteria. FEMS Microbiol Rev; 41, :880–899. doi: [DOI: https://dx.doi.org/10.1093/femsre/fux029] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28961821]

14 Buchan, A , LeCleir, GR , Gulvik, CA , González, JM. . 2014;. Master recyclers: features and functions of bacteria associated with phytoplankton blooms. Nat Rev Microbiol; 12, :686–698. doi: [DOI: https://dx.doi.org/10.1038/nrmicro3326] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25134618]

15 Ferrer-González, FX , Widner, B , Holderman, NR , Glushka, J , Edison, AS , Kujawinski, EB , Moran, MA. . 2021;. Resource partitioning of phytoplankton metabolites that support bacterial heterotrophy. ISME J; 15, :762–773. doi: [DOI: https://dx.doi.org/10.1038/s41396-020-00811-y] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33097854]

16 Brisson, V , Swink, C , Kimbrel, J , Mayali, X , Samo, T , Kosina, SM , Thelen, M , Northen, TR , Stuart, RK. . 2023;. Dynamic Phaeodactylum tricornutum exometabolites shape surrounding bacterial communities. New Phytol; 239, :1420–1433. doi: [DOI: https://dx.doi.org/10.1111/nph.19051] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37301990]

17 Riemann, L , Steward, GF , Azam, F. . 2000;. Dynamics of bacterial community composition and activity during a mesocosm diatom bloom. Appl Environ Microbiol; 66, :578–587. doi: [DOI: https://dx.doi.org/10.1128/AEM.66.2.578-587.2000] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/10653721]

18 Pinhassi, J , Sala, MM , Havskum, H , Peters, F , Guadayol, O , Malits, A , Marrasé, C. . 2004;. Changes in bacterioplankton composition under different phytoplankton regimens. Appl Environ Microbiol; 70, :6753–6766. doi: [DOI: https://dx.doi.org/10.1128/AEM.70.11.6753-6766.2004] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15528542]

19 Teeling, H , Fuchs, BM , Becher, D , Klockow, C , Gardebrecht, A , Bennke, CM , Kassabgy, M , Huang, S , Mann, AJ , Waldmann, J , et al.. 2012;. Substrate-controlled succession of marine bacterioplankton populations induced by a phytoplankton bloom. Science; 336, :608–611. doi: [DOI: https://dx.doi.org/10.1126/science.1218344] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22556258]

20 Pontiller, B , Martínez-García, S , Joglar, V , Amnebrink, D , Pérez-Martínez, C , González, JM , Lundin, D , Fernández, E , Teira, E , Pinhassi, J. . 2022;. Rapid bacterioplankton transcription cascades regulate organic matter utilization during phytoplankton bloom progression in a coastal upwelling system. ISME J; 16, :2360–2372. doi: [DOI: https://dx.doi.org/10.1038/s41396-022-01273-0] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35804052]

21 Hellebust, JA. . 1965;. Excretion of some organic compounds by marine phytoplankton. Limnol Oceanogr; 10, :192–206. doi: [DOI: https://dx.doi.org/10.4319/lo.1965.10.2.0192]

22 Kieft, B , Li, Z , Bryson, S , Hettich, RL , Pan, C , Mayali, X , Mueller, RS. . 2021;. Phytoplankton exudates and lysates support distinct microbial consortia with specialized metabolic and ecophysiological traits. Proc Natl Acad Sci USA; 118, : e2101178118. doi: [DOI: https://dx.doi.org/10.1073/pnas.2101178118] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34620710]

23 Biddanda, B , Benner, R. . 1997;. Carbon, nitrogen, and carbohydrate fluxes during the production of particulate and dissolved organic matter by marine phytoplankton. Limnol Oceanogr; 42, :506–518. doi: [DOI: https://dx.doi.org/10.4319/lo.1997.42.3.0506]

24 McKee, LS , La Rosa, SL , Westereng, B , Eijsink, VG , Pope, PB , Larsbrink, J. . 2021;. Polysaccharide degradation by the Bacteroidetes: mechanisms and nomenclature. Environ Microbiol Rep; 13, :559–581. doi: [DOI: https://dx.doi.org/10.1111/1758-2229.12980] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34036727]

25 Singh, A , Nigam, PS , Murphy, JD. . 2011;. Mechanism and challenges in commercialisation of algal biofuels. Bioresour Technol; 102, :26–34. doi: [DOI: https://dx.doi.org/10.1016/j.biortech.2010.06.057] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20609580]

26 Chisti, Y. . 2016;. Large-scale production of algal biomass: raceway ponds, p 21–40. In Bux, F , Chisti, Y. (ed), Algae biotechnology: products and processes;. Springer: .

27 Carney, LT , Lane, TW. . 2014;. Parasites in algae mass culture. Front Microbiol; 5, : 278. doi: [DOI: https://dx.doi.org/10.3389/fmicb.2014.00278] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24936200]

28 Lane, TW. . 2022;. Barriers to microalgal mass cultivation. Curr Opin Biotechnol; 73, :323–328. doi: [DOI: https://dx.doi.org/10.1016/j.copbio.2021.09.013] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34710649]

29 Paul, C , Pohnert, G. . 2011;. Interactions of the algicidal bacterium Kordia algicida with diatoms: regulated protease excretion for specific algal lysis. PLoS One; 6, : e21032. doi: [DOI: https://dx.doi.org/10.1371/journal.pone.0021032] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21695044]

30 Imai, I , Sunahara, T , Nishikawa, T , Hori, Y , Kondo, R , Hiroishi, S. . 2001;. Fluctuations of the red tide flagellates Chattonella spp. (Raphidophyceae) and the algicidal bacterium Cytophaga sp. in the Seto Inland Sea, Japan. Mar Biol; 138, :1043–1049. doi: [DOI: https://dx.doi.org/10.1007/s002270000513]

31 Parada, AE , Needham, DM , Fuhrman, JA. . 2016;. Every base matters: assessing small subunit rRNA primers for marine microbiomes with mock communities, time series and global field samples. Environ Microbiol; 18, :1403–1414. doi: [DOI: https://dx.doi.org/10.1111/1462-2920.13023] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26271760]

32 Apprill, A , McNally, S , Parsons, R , Weber, L. . 2015;. Minor revision to V4 region SSU rRNA 806R gene primer greatly increases detection of SAR11 bacterioplankton. Aquat Microb Ecol; 75, :129–137. doi: [DOI: https://dx.doi.org/10.3354/ame01753]

33 Stoeck, T , Bass, D , Nebel, M , Christen, R , Jones, MDM , Breiner, H-W , Richards, TA. . 2010;. Multiple marker parallel tag environmental DNA sequencing reveals a highly complex eukaryotic community in marine anoxic water. Mol Ecol; 19 Suppl 1, :21–31. doi: [DOI: https://dx.doi.org/10.1111/j.1365-294X.2009.04480.x] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20331767]

34 Callahan, BJ , McMurdie, PJ , Rosen, MJ , Han, AW , Johnson, AJA , Holmes, SP. . 2016;. DADA2: high-resolution sample inference from Illumina amplicon data. Nat Methods; 13, :581–583. doi: [DOI: https://dx.doi.org/10.1038/nmeth.3869] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27214047]

35 Quast, C , Pruesse, E , Yilmaz, P , Gerken, J , Schweer, T , Yarza, P , Peplies, J , Glöckner, FO. . 2013;. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res; 41, :D590–D596. doi: [DOI: https://dx.doi.org/10.1093/nar/gks1219] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/23193283]

36 Wang, Q , Garrity, GM , Tiedje, JM , Cole, JR. . 2007;. Naive bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ Microbiol; 73, :5261–5267. doi: [DOI: https://dx.doi.org/10.1128/AEM.00062-07] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/17586664]

37 Huntemann, M , Ivanova, NN , Mavromatis, K , Tripp, HJ , Paez-Espino, D , Tennessen, K , Palaniappan, K , Szeto, E , Pillay, M , Chen, IMA , Pati, A , Nielsen, T , Markowitz, VM , Kyrpides, NC. . 2016;. The standard operating procedure of the DOE-JGI metagenome annotation pipeline (MAP v.4). Stand Genomic Sci; 11, : 17. doi: [DOI: https://dx.doi.org/10.1186/s40793-016-0138-x] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26918089]

38 Nurk, S , Meleshko, D , Korobeynikov, A , Pevzner, PA. . 2017;. metaSPAdes: a new versatile metagenomic assembler. Genome Res; 27, :824–834. doi: [DOI: https://dx.doi.org/10.1101/gr.213959.116] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28298430]

39 Uritskiy, GV , DiRuggiero, J , Taylor, J. . 2018;. MetaWRAP-a flexible pipeline for genome-resolved metagenomic data analysis. Microbiome; 6, :158. doi: [DOI: https://dx.doi.org/10.1186/s40168-018-0541-1] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30219103]

40 Parks, DH , Imelfort, M , Skennerton, CT , Hugenholtz, P , Tyson, GW. . 2015;. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome Res; 25, :1043–1055. doi: [DOI: https://dx.doi.org/10.1101/gr.186072.114] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25977477]

41 Prjibelski, A , Antipov, D , Meleshko, D , Lapidus, A , Korobeynikov, A. . 2020;. Using SPAdes de novo assembler. Curr Protoc Bioinform; 70, : e102. doi: [DOI: https://dx.doi.org/10.1002/cpbi.102]

42 Olm, MR , Brown, CT , Brooks, B , Banfield, JF. . 2017;. dRep: a tool for fast and accurate genomic comparisons that enables improved genome recovery from metagenomes through de-replication. ISME J; 11, :2864–2868. doi: [DOI: https://dx.doi.org/10.1038/ismej.2017.126] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28742071]

43 Bushnell, B. . 2014;. BBTools software package.

44 Morris, MM , Kimbrel, JA , Geng, H , Tran-Gyamfi, MB , Yu, ET , Sale, KL , Lane, TW , Mayali, X. . 2022;. Bacterial community assembly, succession, and metabolic function during outdoor cultivation of Microchloropsis salina. mSphere; 7, : e00231-22. doi: [DOI: https://dx.doi.org/10.1128/msphere.00231-22] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35730934]

45 Chaumeil, P-A , Mussig, AJ , Hugenholtz, P , Parks, DH. . 2020;. GTDB-Tk: a toolkit to classify genomes with the genome taxonomy database;. Oxford University Press: .

46 Brettin, T , Davis, JJ , Disz, T , Edwards, RA , Gerdes, S , Olsen, GJ , Olson, R , Overbeek, R , Parrello, B , Pusch, GD , Shukla, M , Thomason, JA, 3rd, Stevens, R , Vonstein, V , Wattam, AR , Xia, F. . 2015;. RASTtk: a modular and extensible implementation of the RAST algorithm for building custom annotation pipelines and annotating batches of genomes. Sci Rep; 5, :8365. doi: [DOI: https://dx.doi.org/10.1038/srep08365] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25666585]

47 Snyder, EE , Kampanya, N , Lu, J , Nordberg, EK , Karur, HR , Shukla, M , Soneja, J , Tian, Y , Xue, T , Yoo, H , et al.. 2007;. PATRIC: the VBI PathoSystems resource integration center. Nucleic Acids Res; 35, :D401–D406. doi: [DOI: https://dx.doi.org/10.1093/nar/gkl858] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/17142235]

48 Zhang, H , Yohe, T , Huang, L , Entwistle, S , Wu, P , Yang, Z , Busk, PK , Xu, Y , Yin, Y. . 2018;. dbCAN2: a meta server for automated carbohydrate-active enzyme annotation. Nucleic Acids Res; 46, :W95–W101. doi: [DOI: https://dx.doi.org/10.1093/nar/gky418] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29771380]

49 Emms, DM , Kelly, S. . 2019;. OrthoFinder: phylogenetic orthology inference for comparative genomics. Genome Biol; 20, :238. doi: [DOI: https://dx.doi.org/10.1186/s13059-019-1832-y] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31727128]

50 Emms, DM , Kelly, S. . 2015;. OrthoFinder: solving fundamental biases in whole genome comparisons dramatically improves orthogroup inference accuracy. Genome Biol; 16, : 157. doi: [DOI: https://dx.doi.org/10.1186/s13059-015-0721-2] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26243257]

51 Emms, DM , Kelly, S. . 2017;. STRIDE: species tree root inference from gene duplication events. Mol Biol Evol; 34, :3267–3278. doi: [DOI: https://dx.doi.org/10.1093/molbev/msx259] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29029342]

52 Letunic, I , Bork, P. . 2021;. Interactive Tree Of Life (iTOL) v5: an online tool for phylogenetic tree display and annotation. Nucleic Acids Res; 49, :W293–W296. doi: [DOI: https://dx.doi.org/10.1093/nar/gkab301] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33885785]

53 Jain, C , Rodriguez-R, LM , Phillippy, AM , Konstantinidis, KT , Aluru, S. . 2018;. High throughput ANI analysis of 90K prokaryotic genomes reveals clear species boundaries. Nat Commun; 9, : 5114. doi: [DOI: https://dx.doi.org/10.1038/s41467-018-07641-9] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30504855]

54 Bowler, C , Allen, AE , Badger, JH , Grimwood, J , Jabbari, K , Kuo, A , Maheswari, U , Martens, C , Maumus, F , Otillar, RP , et al.. 2008;. The Phaeodactylum genome reveals the evolutionary history of diatom genomes. Nature; 456, :239–244. doi: [DOI: https://dx.doi.org/10.1038/nature07410] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/18923393]

55 Kim, S , Pevzner, PA. . 2014;. MS-GF+ makes progress towards a universal database search tool for proteomics. Nat Commun; 5, :5277. doi: [DOI: https://dx.doi.org/10.1038/ncomms6277] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25358478]

56 Monroe, ME , Shaw, JL , Daly, DS , Adkins, JN , Smith, RD. . 2008;. MASIC: a software program for fast quantitation and flexible visualization of chromatographic profiles from detected LC-MS(/MS) features. Comput Biol Chem; 32, :215–217. doi: [DOI: https://dx.doi.org/10.1016/j.compbiolchem.2008.02.006] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/18440872]

57 Florens, L , Carozza, MJ , Swanson, SK , Fournier, M , Coleman, MK , Workman, JL , Washburn, MP. . 2006;. Analyzing chromatin remodeling complexes using shotgun proteomics and normalized spectral abundance factors. Methods; 40, :303–311. doi: [DOI: https://dx.doi.org/10.1016/j.ymeth.2006.07.028] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/17101441]

58 Wilkinson, L. . 2011;. Ggplot2: elegant graphics for data analysis by WICKHAM, H;. Oxford University Press: .

59 R Core Team R. 2013;. R: a language and environment for statistical computing.

60 Lin, H , Peddada, SD. . 2020;. Analysis of compositions of microbiomes with bias correction. Nat Commun; 11, :3514. doi: [DOI: https://dx.doi.org/10.1038/s41467-020-17041-7] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32665548]

61 Dixon, P. . 2003;. VEGAN, a package of R functions for community ecology. J Vegetation Science; 14, :927–930. doi: [DOI: https://dx.doi.org/10.1111/j.1654-1103.2003.tb02228.x]

62 Samo, TJ , Kimbrel, JA , Nilson, DJ , Pett-Ridge, J , Weber, PK , Mayali, X. . 2018;. Attachment between heterotrophic bacteria and microalgae influences symbiotic microscale interactions. Environ Microbiol; 20, :4385–4400. doi: [DOI: https://dx.doi.org/10.1111/1462-2920.14357] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30022580]

63 Coyne, KJ , Wang, Y , Johnson, G. . 2022;. Algicidal bacteria: a review of current knowledge and applications to control harmful algal blooms. Front Microbiol; 13, : 871177. doi: [DOI: https://dx.doi.org/10.3389/fmicb.2022.871177] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35464927]

64 Sohn, JH , Lee, J-H , Yi, H , Chun, J , Bae, KS , Ahn, T-Y , Kim, S-J. . 2004; . Kordia algicida gen. nov., sp. nov., an algicidal bacterium isolated from red tide. Int J Syst Evol Microbiol; 54, :675–680. doi: [DOI: https://dx.doi.org/10.1099/ijs.0.02689-0] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15143006]

65 Lee, HS , Kang, SG , Kwon, KK , Lee, J-H , Kim, S-J. . 2011;. Genome sequence of the algicidal bacterium Kordia algicida OT-1. J Bacteriol; 193, :4031–4032. doi: [DOI: https://dx.doi.org/10.1128/JB.05241-11] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21622754]

66 Achtman, M. . 2012;. Insights from genomic comparisons of genetically monomorphic bacterial pathogens. Phil Trans R Soc B; 367, :860–867. doi: [DOI: https://dx.doi.org/10.1098/rstb.2011.0303] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22312053]

67 Cai, L , Chen, Y , Xiao, S , Liu, R , He, M , Zhang, R , Zeng, Q. . 2023;. Abundant and cosmopolitan lineage of cyanopodoviruses lacking a DNA polymerase gene. ISME J; 17, :252–262. doi: [DOI: https://dx.doi.org/10.1038/s41396-022-01340-6] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36357781]

68 Lucas-Elío, P , Gómez, D , Solano, F , Sanchez-Amat, A. . 2006;. The antimicrobial activity of marinocine, synthesized by Marinomonas mediterranea, is due to hydrogen peroxide generated by its lysine oxidase activity. J Bacteriol; 188, :2493–2501. doi: [DOI: https://dx.doi.org/10.1128/JB.188.7.2493-2501.2006] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/16547036]

69 Buchan, A , González, JM , Moran, MA. . 2005;. Overview of the marine Roseobacter lineage. Appl Environ Microbiol; 71, :5665–5677. doi: [DOI: https://dx.doi.org/10.1128/AEM.71.10.5665-5677.2005] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/16204474]

70 Luo, H , Moran, MA. . 2014;. Evolutionary ecology of the marine Roseobacter clade. Microbiol Mol Biol Rev; 78, :573–587. doi: [DOI: https://dx.doi.org/10.1128/MMBR.00020-14] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25428935]

71 Moran, MA , Belas, R , Schell, MA , González, JM , Sun, F , Sun, S , Binder, BJ , Edmonds, J , Ye, W , Orcutt, B , Howard, EC , Meile, C , Palefsky, W , Goesmann, A , Ren, Q , Paulsen, I , Ulrich, LE , Thompson, LS , Saunders, E , Buchan, A. . 2007;. Ecological genomics of marine Roseobacters. . Appl Environ Microbiol; 73, :4559–4569. doi: [DOI: https://dx.doi.org/10.1128/AEM.02580-06] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/17526795]

72 Wagner-Döbler, I , Biebl, H. . 2006;. Environmental biology of the marine Roseobacter lineage. Annu Rev Microbiol; 60, :255–280. doi: [DOI: https://dx.doi.org/10.1146/annurev.micro.60.080805.142115] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/16719716]

73 Kimbrel, JA , Samo, TJ , Ward, C , Nilson, D , Thelen, MP , Siccardi, A , Zimba, P , Lane, TW , Mayali, X. . 2019;. Host selection and stochastic effects influence bacterial community assembly on the microalgal phycosphere. Algal Res; 40, :101489. doi: [DOI: https://dx.doi.org/10.1016/j.algal.2019.101489]

74 Lim, Y , Kang, I , Cho, J-C. . 2020;. Genome characteristics of Kordia antarctica IMCC3317^T and comparative genome analysis of the genus Kordia. Sci Rep; 10, : 14715. doi: [DOI: https://dx.doi.org/10.1038/s41598-020-71328-9] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32895436]

75 Royo-Llonch, M , Ferrera, I , Cornejo-Castillo, FM , Sánchez, P , Salazar, G , Stepanauskas, R , González, JM , Sieracki, ME , Speich, S , Stemmann, L , Pedrós-Alió, C , Acinas, SG. . 2017;. Exploring microdiversity in novel Kordia sp. (Bacteroidetes) with proteorhodopsin from the tropical indian ocean via single amplified genomes. Front Microbiol; 8, : 1317. doi: [DOI: https://dx.doi.org/10.3389/fmicb.2017.01317] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28790980]

76 Rao, MB , Tanksale, AM , Ghatge, MS , Deshpande, VV. . 1998;. Molecular and biotechnological aspects of microbial proteases. Microbiol Mol Biol Rev; 62, :597–635. doi: [DOI: https://dx.doi.org/10.1128/MMBR.62.3.597-635.1998] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/9729602]

77 Syhapanha, KS , Russo, DA , Deng, Y , Meyer, N , Poulin, RX , Pohnert, G. . 2023;. Transcriptomics-guided identification of an algicidal protease of the marine bacterium Kordia algicida OT-1. Microbiologyopen; 12, : e1387. doi: [DOI: https://dx.doi.org/10.1002/mbo3.1387] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37877654]

78 Coy, SR , Gann, ER , Pound, HL , Short, SM , Wilhelm, SW. . 2018;. Viruses of eukaryotic algae: diversity, methods for detection, and future directions. Viruses; 10, : 487. doi: [DOI: https://dx.doi.org/10.3390/v10090487] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30208617]

79 Lucas-Elio, P , Hernandez, P , Sanchez-Amat, A , Solano, F. . 2005;. Purification and partial characterization of marinocine, a new broad-spectrum antibacterial protein produced by Marinomonas mediterranea. Biochim Biophys Acta Gen Subj; 1721, :193–203. doi: [DOI: https://dx.doi.org/10.1016/j.bbagen.2004.11.002]

80 Mai-Prochnow, A , Lucas-Elio, P , Egan, S , Thomas, T , Webb, JS , Sanchez-Amat, A , Kjelleberg, S. . 2008;. Hydrogen peroxide linked to lysine oxidase activity facilitates biofilm differentiation and dispersal in several gram-negative bacteria. J Bacteriol; 190, :5493–5501. doi: [DOI: https://dx.doi.org/10.1128/JB.00549-08] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/18502869]

81 Ray, A , Schwartz, N , de Souza Santos, M , Zhang, J , Orth, K , Salomon, D. . 2017;. Type VI secretion system MIX-effectors carry both antibacterial and anti-eukaryotic activities. EMBO Rep; 18, :1978–1990. doi: [DOI: https://dx.doi.org/10.15252/embr.201744226] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28912123]

82 Jiang, F , Waterfield, NR , Yang, J , Yang, G , Jin, Q. . 2014;. A Pseudomonas aeruginosa type VI secretion phospholipase D effector targets both prokaryotic and eukaryotic cells. Cell Host Microbe; 15, :600–610. doi: [DOI: https://dx.doi.org/10.1016/j.chom.2014.04.010] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24832454]

83 Coulthurst, S. . 2019;. The type VI secretion system: a versatile bacterial weapon. Microbiology (Reading); 165, :503–515. doi: [DOI: https://dx.doi.org/10.1099/mic.0.000789] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30893029]

84 Kempnich, MW , Sison-Mangus, MP. . 2020;. Presence and abundance of bacteria with the type VI secretion system in a coastal environment and in the global oceans. PLoS One; 15, : e0244217. doi: [DOI: https://dx.doi.org/10.1371/journal.pone.0244217] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33351849]

85 Fernández-Gómez, B , Richter, M , Schüler, M , Pinhassi, J , Acinas, SG , González, JM , Pedrós-Alió, C. . 2013;. Ecology of marine bacteroidetes: a comparative genomics approach. ISME J; 7, :1026–1037. doi: [DOI: https://dx.doi.org/10.1038/ismej.2012.169] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/23303374]

86 Grondin, JM , Tamura, K , Déjean, G , Abbott, DW , Brumer, H. . 2017;. Polysaccharide utilization loci: fueling microbial communities. J Bacteriol; 199, :00860–16. doi: [DOI: https://dx.doi.org/10.1128/JB.00860-16]

87 Huang, L , Zhang, H , Wu, P , Entwistle, S , Li, X , Yohe, T , Yi, H , Yang, Z , Yin, Y. . 2018;. dbCAN-seq: a database of carbohydrate-active enzyme (CAZyme) sequence and annotation. Nucleic Acids Res; 46, :D516–D521. doi: [DOI: https://dx.doi.org/10.1093/nar/gkx894] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30053267]

88 Myklestad, SM. . 1989;. Production, chemical structure, metabolism, and biological function of the (1→ 3)-linked, β3-D-glucans in diatoms. Biol Oceano; 6, :313–326. doi: [DOI: https://dx.doi.org/10.1080/01965581.1988.10749534]

89 Newton, RJ , Griffin, LE , Bowles, KM , Meile, C , Gifford, S , Givens, CE , Howard, EC , King, E , Oakley, CA , Reisch, CR , Rinta-Kanto, JM , Sharma, S , Sun, S , Varaljay, V , Vila-Costa, M , Westrich, JR , Moran, MA. . 2010;. Genome characteristics of a generalist marine bacterial lineage. ISME J; 4, :784–798. doi: [DOI: https://dx.doi.org/10.1038/ismej.2009.150] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20072162]

90 Poretsky, RS , Sun, S , Mou, X , Moran, MA. . 2010;. Transporter genes expressed by coastal bacterioplankton in response to dissolved organic carbon. Environ Microbiol; 12, :616–627. doi: [DOI: https://dx.doi.org/10.1111/j.1462-2920.2009.02102.x] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19930445]

91 Cha, Q-Q , Liu, S-S , Dang, Y-R , Ren, X-B , Xu, F , Li, P-Y , Chen, X-L , Wang, P , Zhang, X-Y , Zhang, Y-Z , Qin, Q-L. . 2023;. Ecological function and interaction of different bacterial groups during alginate processing in coastal seawater community. Environ Int; 182, :108325. doi: [DOI: https://dx.doi.org/10.1016/j.envint.2023.108325] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37995388]

92 Priest, T , Vidal-Melgosa, S , Hehemann, J-H , Amann, R , Fuchs, BM. . 2023;. Carbohydrates and carbohydrate degradation gene abundance and transcription in Atlantic waters of the Arctic. ISME Commun; 3, : 130. doi: [DOI: https://dx.doi.org/10.1038/s43705-023-00324-7] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/38071398]

93 D’Souza, G , Shitut, S , Preussger, D , Yousif, G , Waschina, S , Kost, C. . 2018;. Ecology and evolution of metabolic cross-feeding interactions in bacteria. Nat Prod Rep; 35, :455–488. doi: [DOI: https://dx.doi.org/10.1039/c8np00009c] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29799048]

94 Taylor, JD , Cottingham, SD , Billinge, J , Cunliffe, M. . 2014;. Seasonal microbial community dynamics correlate with phytoplankton-derived polysaccharides in surface coastal waters. ISME J; 8, :245–248. doi: [DOI: https://dx.doi.org/10.1038/ismej.2013.178] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24132076]

95 Enke, TN , Datta, MS , Schwartzman, J , Cermak, N , Schmitz, D , Barrere, J , Pascual-García, A , Cordero, OX. . 2019;. Modular assembly of polysaccharide-degrading marine microbial communities. Curr Biol; 29, :1528–1535. doi: [DOI: https://dx.doi.org/10.1016/j.cub.2019.03.047] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31031118]

96 Pontrelli, S , Szabo, R , Pollak, S , Schwartzman, J , Ledezma-Tejeida, D , Cordero, OX , Sauer, U. . 2022;. Metabolic cross-feeding structures the assembly of polysaccharide degrading communities. Sci Adv; 8, : eabk3076. doi: [DOI: https://dx.doi.org/10.1126/sciadv.abk3076] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35196097]

97 Cordero, OX , Datta, MS. . 2016;. Microbial interactions and community assembly at microscales. Curr Opin Microbiol; 31, :227–234. doi: [DOI: https://dx.doi.org/10.1016/j.mib.2016.03.015] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27232202]

98 Datta, MS , Sliwerska, E , Gore, J , Polz, MF , Cordero, OX. . 2016;. Microbial interactions lead to rapid micro-scale successions on model marine particles. Nat Commun; 7, : 11965. doi: [DOI: https://dx.doi.org/10.1038/ncomms11965] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27311813]

99 D’Souza, G , Schwartzman, J , Keegstra, J , Schreier, JE , Daniels, M , Cordero, OX , Stocker, R , Ackermann, M. . 2023;. Interspecies interactions determine growth dynamics of biopolymer-degrading populations in microbial communities. Proc Natl Acad Sci USA; 120, : e2305198120. doi: [DOI: https://dx.doi.org/10.1073/pnas.2305198120] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37878716]

100 Bigalke, A , Meyer, N , Papanikolopoulou, LA , Wiltshire, KH , Pohnert, G. . 2019;. The algicidal bacterium Kordia algicida shapes a natural plankton community. Appl Environ Microbiol; 85, : e02779-18. doi: [DOI: https://dx.doi.org/10.1128/AEM.02779-18] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30737345]

101 King, GM , Weber, CF. . 2007;. Distribution, diversity and ecology of aerobic CO-oxidizing bacteria. Nat Rev Microbiol; 5, :107–118. doi: [DOI: https://dx.doi.org/10.1038/nrmicro1595] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/17224920]

102 Cordero, PRF , Bayly, K , Man Leung, P , Huang, C , Islam, ZF , Schittenhelm, RB , King, GM , Greening, C. . 2019;. Atmospheric carbon monoxide oxidation is a widespread mechanism supporting microbial survival. ISME J; 13, :2868–2881. doi: [DOI: https://dx.doi.org/10.1038/s41396-019-0479-8] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31358912]