Bacterial community diversity and potential eco-physiological roles in toxigenic blooms composed of Microcystis, Aphanizomenon or Planktothrix

Cyanobacterial toxicity, cyanotoxins, and their impact on aquatic ecosystems and human health are well documented. In comparison, less is known about bloom-associated bacterial communities. Co-occurring bacteria can influence bloom development, physiology and collapse, and may also provide a niche for pathogenic bacteria. Existing research focuses on the cyanosphere of Microcystis-dominated blooms, despite the increasing prevalence of filamentous genera (Aphanizomenon and Planktothrix). This pilot study aimed to broaden our understanding of the bacterial consortia attached to morphologically distinct cyanobacteria (coccoid and filamentous) dominating phytoplankton communities and to explore their potential roles in amplifying the impacts of cyanobacterial blooms. We investigated four shallow freshwater bodies across three continents and two climate zones: an urban pond in the USA, a dammed reservoir and a natural lake in Poland, and an urban water body in Singapore. Amplicon sequencing (16S rRNA gene) was used to characterize bacterial communities, while shotgun metagenomics identified nitrogen- and phosphorus-cycling genes to infer potential eco-physiological functions. Cyanobacteria dominated bacterioplankton assemblages at all sites (>35.6%), with bloom composition influencing toxigenic profiles. A mixed bloom of Microcystis, Snowella, and Aphanizomenon had the broadest range of cyanotoxin synthetase genes (mcyE, cyrJ, anaF and sxtA). Microcystis blooms correlated with increased Roseomonas, while Planktothrix co-occurred with Flavobacterium – both bacteria likely contribute to nutrient-cycling within blooms and represent potential opportunistic pathogens for aquatic organisms and humans. The Microcystis cyanosphere exhibited the highest number of significant positive correlations with bacteria (19 relations), compared to Planktothrix and Aphanizomenon (11 and 2 relations, respectively). Non-diazotrophic blooms of Microcystis and Planktothrix showed greater abundances of nitrogen – (ureB, glnA, narB, and narHZ) and phosphorus-cycling genes (phoBHPR and ppk1), indicating a strong dependence on associated bacteria for nutrient acquisition compared to diazotrophic Aphanizomenon. These findings suggest that Aphanizomenon-dominated blooms may be sustained by simpler microbiomes. Our results provide preliminary evidence of cyanosphere heterogeneity potentially shaped by the dominance or coexistence of three morphologically and eco-physiologically distinct genera of cyanobacteria. A comprehensive knowledge of the taxonomy and functional roles of bloom-associated microbiomes is therefore essential to understand bloom activity, evaluate the environmental threat, and develop effective strategies for prevention and mitigation.

1 Introduction

Harmful cyanobacterial blooms (HCBs) are common threats resulting from the eutrophication of freshwater ecosystems, a situation exacerbated by increased anthropogenic development (anthropopressure) and global climate change (Paerl, 2014; Paerl and Barnard, 2020; Burford et al., 2020). Cyanotoxins are pollutants produced by cyanobacteria that represent a potential hazard to human and aquatic ecosystem health and include hepatotoxic microcystins (MCs) and cylindrospermopsins (CYN), and neurotoxic anatoxins (ATX) and saxitoxins (SXT) (Svirčev et al., 2019). Blooms dominated by coccoidal colonies of Microcystis spp. are among the most well documented due to their ubiquitous presence worldwide and their ability to produce MCs – the most common and well-studied cyanotoxin (Harke et al., 2016). Recent studies have reported that HCBs in freshwater ecosystems are shifting from Microcystis-dominated populations to other types of filamentous cyanobacteria, such as Aphanizomenon spp. (Parulekar et al., 2017; Zhang et al., 2020) and Planktothrix spp., both of which can produce CYN, ATX, and SXT (Wang et al., 2021). The toxicity of CYN (LD₅₀ = 6.9 μg g⁻¹), ATX (LD₅₀ = 20–200 μg g⁻¹) and SXT (LD₅₀ = 10 μg g⁻¹) has been reported to be as high or higher than that of MCs (LD₅₀ = 50–600 μg g⁻¹) (Christensen and Khan, 2020). However, there are few studies focusing on microbial and ecological factors associated with the domination of HCBs by filamentous forms of cyanobacteria. Therefore, additional studies could help identify other factors that elevate the environmental and health threats related to HCBs beyond our knowledge of their potential toxigenicity.

Interactions of bacteria with cyanobacteria (interactome) are important elements that alter the ambient chemical composition, physiology, and development of HCBs in freshwater ecosystems (Cook et al., 2020; Mankiewicz-Boczek and Font-Nájera, 2022). The bacteria thrive embedded in the surrounding extracellular polysaccharide space of cyanobacterial cells (EPS mucilage), known as the cyanosphere, where synergistic and antagonistic interactions occur that can alter the fate of the blooms (Le et al., 2022a). These interactions can shape the ecosystem diversity of bacterioplankton, including the dominance of cyanobacterial taxa in phytoplankton communities (Cook et al., 2020). Recent studies have investigated the diversity of bacteria attached to cyanobacteria. However, most have been focused on HCBs dominated by Microcystis spp. (Le et al., 2024). There is a lack of information regarding blooms that are dominated by filamentous cyanobacteria, such as Aphanizomenon spp. and Planktothrix spp. Microcystis blooms have been found to contain a high diversity of associated bacterial communities because they produce thick mucilage with rich EPS substances (Le et al., 2022a). Only a few studies have described the associated bacterial communities of Aphanizomenon spp. (Parulekar et al., 2017; Underwood et al., 2024) or Planktothrix spp. (Wang et al., 2021) and selected potential ecological functions. These studies reported that Flavobacterium and Rheinheimera were highly abundant in Aphanizomenon-dominated blooms, while Flavobacterium was also prominent in a Planktothrix-dominated bloom. However, these bacterial communities have not been directly compared with those associated with Microcystis-dominated blooms. There is a need to address this knowledge gap, since filamentous cyanobacteria (e.g., Aphanizomenon, Pseudanabaena, Dolichospermum and Raphidiopsis) are expected to occur more frequently in the context of global climate change and increasing anthropopressure, potentially outcompeting species like Microcystis (Gao et al., 2025).

HCBs can harbor pathogenic microorganisms, which is attributed to the ability of cyanobacteria to develop dense colonies in surface waters. Berg et al. (2009) isolated many pathogenic bacterial strains, e.g., Aeromonas, Vibrio, Acinetobacter and Pseudomonas spp., from colonial forms of cyanobacteria obtained from different Nordic lakes and the Baltic Sea. Unfortunately, the dominant cyanobacterial genera in these phytoplankton communities were not reported. To our knowledge, the presence of pathogenic bacteria has not been well characterized for HCBs dominated by coccoidal forms of Microcystis or filamentous forms of Aphanizomenon and Planktothrix. All three cyanobacteria are known to form dense colonial aggregations in surface waters, which indicates their potential ability to harbor these organisms.

Cyanobacteria are known to exhibit diverse nutrient affinities and nutrient cycling strategies, including phosphorus (P) and nitrogen (N). Non-diazotrophic cyanobacteria, such as Microcystis spp., have a higher affinity for ammonium and adapt well to N-rich environments (Bell et al., 2018). Microcystis spp. have higher P-affinity, that allows them to better thrive under P-limiting conditions than many other cyanobacteria (Harke et al., 2012). Non-diazotrophic cyanobacteria, such as Planktothrix spp., are well adapted to N- and P-rich environments, and can persist year-round in temperate freshwater bodies (Kokociński et al., 2011; Mankiewicz-Boczek et al., 2011; Wejnerowski et al., 2024). Conversely, diazotrophic cyanobacteria, such as Aphanizomenon spp. thrive in N-limiting conditions because of their ability to fix gaseous N (De Nobel et al., 1995; De Nobel et al., 1997). In summary, Microcystis, Planktothrix and Aphanizomenon may use different strategies for nutrient utilization, but these strategies are not well understood in the context of their dominance or co-occurrence in blooms.

Attached bacteria can play important roles in the transformation of unavailable nutrients into forms accessible to cyanobacteria during the development of HCBs (Yang et al., 2021; Wan et al., 2022; Yan et al., 2023). It has been hypothesized that bacteria may influence the biological strategies of cyanobacteria to cope with nutrient limitations in the environment. Nevertheless, their roles have been scarcely described in the literature, and most research has focused on intraspecific strategies of cyanobacteria. Microcystis blooms have been associated with high abundances of bacterial nitrogen decomposition genes (glutamate synthesis) and assimilatory and dissimilatory nitrate reduction to ammonium genes (ANRA and DNRA), regarded as aiding Microcystis spp. to utilize ammonium in N-rich environments. Microcystis blooms were also linked to high abundances of bacterial P-solubilizing genes, such as alkaline phosphatases, which are important for P transformation and availability in P-limiting environments (Yang et al., 2021; Wan et al., 2022; Yan et al., 2023; Cai et al., 2024). Wang et al. (2021) identified that Planktothrix, in a bloom co-dominated with Microcystis, was associated with the expression of microbial genes involved in N-synthesis and decomposition and alkaline phosphatases, which was attributed to a N-rich and P-limiting environment. To our knowledge, there are no studies describing the role of associated bacteria and availability of nutrients in blooms dominated by Aphanizomenon spp. The limited knowledge on the role of bacteria in the transformation and availability of nutrients highlights the need for further investigation, especially because nutrients play a key role in the development of HCBs (Paerl and Barnard, 2020).

The present study aimed to increase our knowledge of bacterial community diversity and their potential roles influencing the threat of HCBs dominated or mixed by coccoidal Microcystis or filamentous Aphanizomenon and Planktothrix. Our freshwater study sites are in three different countries: The Lake in Central Park (USA; U.S. Geological Survey station 404630073580801; U.S. Geological Survey, 2025), Raczyńskie Lake and Sulejów Reservoir (Poland), and an urban water body in Singapore. This study is preliminary, and the findings presented here are intended to provide baseline information for future comparative analyses. The investigation will be expanded through increased sampling efforts and inclusion of additional freshwater bodies, in accordance with the objectives of the ongoing CyMiBiom project (2024-2025). We focused on describing four elements that may increase the threat of HCBs to environmental health: (i) cyanobacterial toxigenicity, (ii) the diversity of associated bacterial taxa, (iii) presence of pathogenic bacteria, and (iv) abundance of bacterial nutrient-cycling genes. These elements will enhance our understanding of the eco-physiological roles of bacterial communities, allowing the development of hypotheses about the diverse relationship between three different cyanobacterial genera and their attached consortia. To accomplish our objectives, amplicon sequencing with the 16S rRNA gene was used to characterize bacterioplankton communities and shotgun metagenomic constructs were used to mine for the presence of N- and P-cycling and toxigenic genes. This knowledge is essential to understand the pervasive threat of variable HCBs in freshwater ecosystems worldwide.

2 Materials and methods

2.1 Study sites and sample processing

Four freshwater bodies, known to harbor different communities of bloom-forming cyanobacteria, were selected for our study (Supplementary Figure S1). The Lake in Central Park (site TL) in the City of New York, New York, is within the most visited park in the USA and has a high density of walking trails (Fisher, 2011). The Lake is an artificial impoundment used for recreational boating by thousands of people annually. Near shore monitoring in the last decade has shown that annual summer HCBs are highly toxic and dominated by Microcystis viridis (Flanzenbaum et al., 2022). Raczyńskie Lake (site RA) in Poland is a natural freshwater body formed by the Kamionka River in western Poland that has gradually changed from an agricultural to a recreational area, with substantial tourism pressure. HCBs in Raczyńskie Lake have been reported to contain Aphanizomenon flos-aquae, Planktothrix agardhii, Dolichospermum affine and Microcystis aeruginosa (Kowalczewska-Madura et al., 2022). The Sulejów Reservoir (site SU) in Poland is in the Pilica River catchment in central Poland originally designated for power generation, flood control, recreation and as an alternative water supply for the City of Łódź (Izydorczyk et al., 2008). Sulejów Reservoir is known to experience intense summer HCBs of Aphanizomenon flos-aquae and Microcystis spp. (Jaskulska et al., 2021; Mankiewicz-Boczek and Font-Nájera, 2022). In Singapore, the selected freshwater body (site SP) is an urban reservoir (hereafter referred to as “Singapore urban reservoir”) constructed by damming a river mouth located near the city of Singapore. Singapore urban reservoir is used as a source of drinking water supply and for recreational activities such as fishing and picnicking. The phytoplankton communities in this reservoir have changed, shifting from Microcystis-dominated blooms to mixed assemblages of Microcystis and filamentous cyanobacteria, such as Anabaena (Te and Gin, 2011).

Water samples were collected in shallow (<2 m), nearshore areas at each site, 0–50 cm below the surface. Three surface grab samples (up to 1 L each) were taken from each water body and homogenized in a large vessel that was pre-rinsed several times with local freshwater. Homogenized water was stored in plastic containers (up to 3 L), that were previously washed with ethanol and deionized water and preserved in cold (4 °C) and dark conditions until transport to local laboratories. Sampling was conducted during hot and sunny conditions in 2023, which favored the visible green coloration of water and the formation of cyanobacterial surface scums (Supplementary Figure S2). For the water bodies located in temperate zones, sampling took place at the end of the summer season on September 5th for RA, and on September 20th for SU and TL. In contrast, sampling at SP, which is located in the tropical zone was carried out on November 16th, coinciding with the end of an inter-monsoon period (Table 1). Physico-chemical parameters were measured on site using a multiparameter probe (YSI EXO2, YSI, Inc., Yellow Springs, OH, USA), including temperature, pH, dissolved oxygen concentration and specific conductance. Samples were processed for different analyses immediately upon return to the local laboratories. For dissolved nutrient composition, water was filtered through GF/C Whatman filters (0.45 μm) and the filtrate was frozen (−20 °C) until further analysis at local laboratories. For DNA analysis, 30–100 mL of surface water were filtered through 1.2 μm mixed cellulose ester (MCE) filters, in triplicate, to collect the particle-attached microbial fraction (Mankiewicz-Boczek and Font-Nájera, 2022). Laboratory controls, in duplicate, were prepared with the filtration of 100 mL of deionized water through 0.22 μm MCE filters at each local laboratory. These controls were used to describe background contamination, such as bacterial DNA in deionized water or laboratory materials and surfaces. Filters for DNA analysis were dried for 1 hour in aseptic conditions and shipped to Poland in ambient temperature for further analyses.

Table 1

Table 1. Surface water environmental parameters of studied sites.

2.2 Nutrient analyses

Water samples were immediately filtered and processed at local laboratories. For TL, orthophosphate (P-PO₄³⁻) concentration was measured at the U.S. Geological Survey National Water Quality Laboratory (Denver, CO, USA) by discrete analyzer phosphomolybdate formation and colorimetry (detection limit of 4 μg L⁻¹), according to Fishman (1993). Nitrogen species were analysed as N-NH₄⁺ with the salicylate-hypochlorite colorimetry method (detection limit of 20 μg L⁻¹), and NO₃⁻ + NO₂⁻ with the diazotization colorimetry method (detection limit of 40 μg L⁻¹), according to Fishman (1993). In Poland and Singapore, ion chromatography was used to estimate the concentration of above-mentioned nutrients according to Gągała et al. (2014) (limit of detection for all nutrients was 1 μg L⁻¹).

2.3 DNA isolation

Filters containing the particle-attached microbial fraction were placed into separate bead tubes and DNA was extracted with the specifications in the Dneasy® PowerWater® Kit (Qiagen, Hilden, Germany). DNA quality and quantity were measured with a NanoDrop microvolume spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and stored at −20 °C for further analyses.

2.4 Sequencing analyses

2.4.1 16S rRNA amplicon sequencing

DNA libraries (16S rRNA) were prepared with a 25 μL Polymerase Chain Reaction (PCR) mixture containing 12.5 μL of 2 x Phanta Master Mix (Vazyme, China), 2.5 μL of DNA template (5 ng μL⁻¹), 0.5 μL of primers (10 μM μL⁻¹) specific to V3-V4 regions of the 16S rRNA gene (Klindworth et al., 2013), designed with Illumina overhang adapters following the 16S Metagenomic Sequencing Library Preparation Guide (Illumina, 2023). Amplicon PCR conditions included 95 °C for 3 min, 25 cycles of 95 °C for 30 s, 55 °C for 30 s and 72 °C for 30 s, and a final extension of 72 °C for 5 min. PCR products were cleaned with Sera-Mag beads (Cytiva, UK) and analysed by electrophoresis on 1.5% agarose gel. PCR products were indexed with Nextera XT index kit (Illumina, Inc., San Diego, CA, USA) using the above-mentioned specifications for PCR with an 8-cycle amplification program. Products were quantified by quantitative PCR (qPCR) using NEBNext Library Quant kit for Illumina (New England Biolabs, Inc., Ipswich, MA, USA). Libraries were sequenced (2 × 250 bp paired-end reads) using NovaSeq 6000 SP Reagent kit (Illumina, USA) on the Novaseq 6000 platform (Illumina, USA) with an expected amount of 100,000 pair-end reads per sample. Amplicon sequencing was prepared in triplicates for each site.

2.4.2 Shotgun metagenomics

DNA libraries were prepared with the NexteraXT DNA library preparation kit (Illumina, USA) according to the manufacturer’s instructions (Illumina, 2021). The quality of libraries was assessed with Agilent Technologies (Santa Clara, CA, USA) Bioanalyzer 2,100 system with Agilent High Sensitivity DNA kit. Libraries were normalized with qPCR, similarly, as performed with amplicon sequencing. Metagenomics were performed by shotgun sequencing concurrently in the same Illumina platform during amplicon sequencing, also for 2 × 250 bp paired-end reads and maximum assumption of 50 million paired reads per sample. Shotgun metagenomics was prepared in duplicate for each site, reflecting the pilot nature of this study and the higher resource demands of shotgun sequencing relative to 16S rRNA amplicon sequencing.

2.4.3 Bioinformatic analyses

For 16S rRNA analysis, reads were edited with the package DADA2 1.32 (Callahan et al., 2016) in R environment 4.3.2 (R core Team, 2023), including filtering, trimming, error analysis, merging, chimera removal and allocation of taxa. The SILVA 138 database was utilized for amplicon sequence variants (ASV) taxonomical assignation by similarity (≥97% homology) to bacteria.¹ Sequences assigned to mitochondria and chloroplasts were removed. The phyloseq package 1.42 was used for data visualization (McMurdie and Holmes, 2013). Original datasets were uploaded to the NCBI Sequence Read Archive² as FASTQ files under the project PRJNA1241204, with three biosamples for TL (SAMN47547428 – SAMN47547430), RA (SAMN47547431 – SAMN47547433), SU (SAMN47547434 – SAMN47547436) and SP (SAMN47547437 – SAMN47547439).

For shotgun metagenomic analysis, sequences were trimmed to reduce sample contamination – a palindromic algorithm was used for sequence alignment with appropriate adaptors (TrimmomaticPE v0.39 + dfsg-2; Bolger et al., 2014). Trimmed sequences were merged using Vsearch command, and shotgun metagenomes were mined for the presence of cyanotoxin-producing genes and nutrient-transforming genes (N- and P-cycling genes) using BLAST via the DIAMOND tool in the conda environment for Python 2.7.18 (Python Software Foundation, 2020). For cyanotoxin synthesis, genes involved in microcystins (mcyE), cylindrospermopsins (cyrJ), anatoxins (anaC), and saxitoxins (sxtA) synthesis were included. Cyanotoxin genes were utilized to describe the potential toxigenicity of cyanobacterial blooms. For nitrogen (N) – cycling, several genes involved in N-decomposition and ammonium assimilation (ureB, glnA, gcvT, gltB, gltD, and amt), nitrification (amoA, nxrA, and nxrB), complete nitrification COMAMMOX (_camoA), anaerobic nitrification ANAMMOX (hzsA and hzo), assimilatory or dissimilatory nitrate reduction to ammonium – ANRA and DNRA (napA, narG, narZ, narH, narB, nrfA, and nrfH), denitrification (nirS, nirK, norB, norR, and nosZ) and nitrogen fixation (nifK and nifB) were included. For phosphorus (P) – cycling, genes involved in phosphorus solubilization (phoB, phoR, phoD, phoH, phoL, and phoP), phosphorus accumulation in the form of polyphosphate (ppk1), and polyphosphate solubilization and release of phosphorus (ppx) were included. The genes were used to describe potential eco-physiological roles of bacterioplankton communities in nutrient-transformation processes. Gene accession numbers are summarized in Supplementary Table S1. Gene counts (hit numbers) were filtered according to recommended parameters (sequence length ≥25 aa, match score ≥50 and e-value ≤1 × 10⁻¹⁰) and normalized to the 16S rRNA gene count pooled from shotgun metagenomic datasets. Original datasets as FASTQ files were uploaded to the NCBI Sequence Read Archive² under the project PRJNA1241753, with two biosamples for TL (SAMN47568682 – SAMN47568683), RA (SAMN47568684 – SAMN47568685), SU (SAMN47568686 – SAMN47568687) and SP (SAMN47568688 – SAMN47568689).

2.5 Data analyses

Rarefaction curve analysis was employed to estimate the sequencing depth of amplicons. The indices Shannon-Weiner (H′), abundance-based coverage estimator (ACE) richness, evenness (J) and Simpson’s dominance (D) were used to describe the diversity of bacterioplankton communities (refer to Supplementary Table S2). Principal component analysis (PCA) was used to describe differences between study sites based on the bacterioplankton communities and the composition of nutrient-cycling genes. Due to the large amount of ASVs, PCA was constructed with the taxa that contributed with the highest abundances (>1% of relative abundance per sample). Significant differences in diversity indices and spatial ordination of the samples were described with Kruskal–Wallis and Dunn’s post hoc tests (p < 0.05). Spearman’s correlation (r) was used to describe the relationships within the microbiological community (cyanobacteria and other bacteria). Bonferroni corrections were applied to p-values to reduce the finding of false positive and negatives. Graphs were prepared with the statistical software PAST 4.12b (Hammer et al., 2001).

3 Results

3.1 Environmental characterization of surface waters

Physico-chemical parameters and nutrient composition varied across water bodies (Table 1). The temperature showed clear differences according to latitude, with temperatures ranging from 20.5 to 23.6 °C for temperate water body sites (TL, RA and SU) and to the highest of 30.2 °C for the tropical site SP (Table 1). The generally observed basic pH (7.3–8.3) and high concentrations of dissolved oxygen (8.47–9.91 mg L⁻¹) in all water bodies are typical during daylight hours in surface waters containing HCBs (Table 1). Specific conductance was more variable, with the highest values observed in the water bodies located in Poland (RA = 538.0 μS cm⁻¹ and SU = 267.6 μS cm⁻¹) (Table 1). In the case of nutrients, the TL, SP and RA were characterized with the highest P-PO₄³⁻ concentrations (105.6, 34.00 and 15.62 μg L⁻¹, respectively) (Table 1). Notably, TL, SP and RA were an order of magnitude higher than observed in SU (1.53 μg L⁻¹ of P-PO₄³⁻), and concentrations in TL were about 3 to 7 times higher than in SP and RA, respectively (Table 1). All nitrogen species were below the limit of detection in TL and SP (Table 1). In contrast, concentrations were much higher in the Polish sites (RA and SU). The N-NO_x concentrations were similar in RA and SU (34.16 and 36.87 μg L⁻¹ of N-NO₃⁻_, and 0.030 and 0.053 μg L⁻¹ of N-NO₂⁻, respectively; Table 1), but N-NH₄⁺ concentration was about 3-times higher in SU than RA (28.69 and 9.50 μg L⁻¹, respectively; Table 1).

3.2 Description of toxic cyanobacterial communities

A database containing 3,512,878 raw reads was obtained from 12 samples using 16S rRNA amplicon sequencing. A total of 2,780,855 good quality sequences were retrieved after standard quality control, of which 2,661,437 sequences were classified to bacteria (Supplementary Table S3). Classified sequences were composed of 6,798 ASVs belonging to 50 phyla, 107 classes, 227 orders, 309 families and 540 genera. Rarefaction curve analysis showed that all samples reached a plateau, indicating that the sequencing depth was adequate to describe overall bacterial diversity (Supplementary Figure S3). Laboratory controls were estimated to contain <75 copies of 16S rRNA gene per ng DNA using qPCR, which represents roughly 3.38 × 10⁻⁴% of the average number of good quality reads with assigned taxonomy used in the present study (Supplementary Table S3). DNA libraries were not obtained from laboratory controls due to low DNA material (<3.1 ng uL⁻¹); therefore, we expected minimal background laboratory contamination. Classified reads were shown as relative abundances of the 16S rRNA gene representing different bacterioplankton communities (Figure 1). Generally, bacterial communities were dominated by Cyanobacteria (35.6–68.7%), followed by Proteobacteria (15.4–24.4%) and Bacteroidetes (4.5–31.4%) (refer to Figure 1a; Supplementary Table S4). Despite cyanobacterial dominance in all samples, cyanobacterial community composition differed among study sites (Figure 1b; Supplementary Table S5). TL had the highest relative abundance of Microcystis PCC-7914 (~ 33.4%), while the other sites were mixed between filamentous and coccoidal forms of cyanobacteria (Figure 1b; Supplementary Table S5). RA had high relative abundance of Microcystis PCC-7914 (~22.8%), followed by Aphanizomenon NIES81 (~18.9%) and Snowella 0TU37S04 (~11.9%) (refer to Figure 1b; Supplementary Table S5). The cyanobacterial composition in SU was dominated by Aphanizomenon MDT14a (~38.2%) and had lower abundance of Microcystis PCC-7914 (~7.5%) (refer to Figure 1b; Supplementary Table S5). Finally, SP had the highest relative abundance of Planktothrix NIVA CYA 15 (~17.6%), mixed with lower amounts of Snowella 0TU37S04 (~4.2%) and Microcystis PCC-7914 (~3.6%) (refer to Figure 1b; Supplementary Table S5). Microcystis PCC-7914 was the only cyanobacterial taxon commonly shared among all study sites.

Figure 1

Figure 1. Relative abundances (%) of bacterial communities using 16S rRNA amplicon sequencing. Abundances were described to the level of phylum (a), and in the case of genus, cyanobacteria were displayed separately from other types of bacteria (b). Unclassified sequences and other taxa with lower relative abundance (<1%) were excluded in (b). Relative abundances were estimated for each study site (TL: The Lake in Central Park – New York, USA; RA: Raczyńskie Lake, Poland; SU: Sulejów Reservoir, Poland; SP: Singapore urban reservoir) using sample replicates (n = 3).

On average, a total of 2.65 × 10⁷ reads per sample were obtained from shotgun metagenomes, with good quality reads reaching 1.68 × 10⁷. Total merged reads were 32.3–50.5% of good quality reads, which were used for formal analyses. Editing features and error statistics for shotguns are presented in Supplementary Table S6. A total of 2,000 hit numbers, representing selected cyanotoxin genes, were obtained from shotguns and normalized to the 16S rRNA (refer to Supplementary Table S7). Potential toxigenicity of HCBs in selected study sites was also described (Figure 2). Overall, mcyE gene counts were notably higher than other cyanotoxin genes at all sites. TL had the highest mcyE gene relative abundance compared to the other study sites, followed by RA, SU and, finally, SP (Figure 2a). Correlation analysis revealed that Microcystis PCC-7914 was the only cyanobacterial taxon that positively correlated with the normalized abundance of mcyE (r = 0.96, p = 0.047) (refer to Figure 2b; Supplementary Table S8). The genes anaC, cyrJ and sxtA were all highest at RA compared to the other sites. RA was the only site where all four cyanotoxin genes were detected, and that cyrJ was not detected in TL and SU, and sxtA was not detected in SU and SP (Figure 2a). Strong positive correlations were observed between the relative abundance of Aphanizomenon NIES81 and normalized gene counts of anaC (r = 0.76), cyrJ (r = 0.73), and sxtA (r = 0.75), although these were not significant after Bonferroni correction (Figure 2b; Supplementary Table S8). Pseudanabaena PCC7929 and Cuspidothrix LMECYA 163 correlated positively to genes anaC, cyrJ, and sxtA, but were not considered for further analysis because they were detected in contrasting lower abundances (up to 2.6 and 1.2%, respectively) and were not significant after p-value correction.

Figure 2

Figure 2. The overall relative abundance of cyanotoxin synthetase genes normalized to the total number of 16S rRNA (a), and their relation (non-parametric Spearman’s correlation) to the abundance of cyanobacterial taxa in (b). Bars and error lines were estimated for each study site (TL: The lake in Central Park – New York, USA; RA: Raczyńskie Lake, Poland; SU: Sulejów Reservoir, Poland; SP: Singapore urban reservoir) using sample replicates (n = 2).

3.3 Description of bacterial communities associated with cyanobacteria

The composition and diversity of bacterial communities – other than cyanobacteria – are shown with the use of the 16S rRNA amplicon sequencing (Figure 3; Supplementary Tables S9–S11). The composition of bacteria appeared more evenly distributed and diverse at TL compared to the other sites (Figure 3a). In fact, TL showed the highest values of Shannon-Wiener diversity (H = 4.4 ± 0.02), richness (ACE = 460.3 ± 13.3) and evenness (J = 0.72 ± 0.0002) (Figure 3b). Bacterial richness at TL was notably higher than all other sites, including RA (ACE = 357.5 ± 3.7), SU (ACE = 325.9 ± 11.8), and SP (ACE = 317.2 ± 20.7) (Figure 3b). The lowest diversity and evenness were observed at SP (H = 2.8 ± 0.1 and J = 0.49 ± 0.02, respectively; Figure 3b). In contrast, for Simpson’s dominance, SP showed the highest index value (D = 0.203 ± 0.015), which was notably different from SU (D = 0.104 ± 0.006), RA (D = 0.048 ± 0.006), and TL (D = 0.024 ± 0.001) (Figure 3b). The dominance indices at SP and SU were attributed to high relative abundance of two bacterial taxa: Flavobacterium and Rheinheimera (38.4 and 16.8%, respectively, at SP and 29.5 and 6.0%, respectively at SU; Figure 1a). Finally, a higher dominance value for RA compared to TL was attributed to high relative abundances of an undetermined taxon of Caldilineaceae (13.1%) and Roseomonas (11.7%; Figure 3a). There were no evident dominant bacterial taxa for TL (Figure 3a). Despite clear differences in diversity indices across sites, after applying the Bonferroni correction, statistically significant differences were observed only between TL and SP across Shannon-Wiener diversity, evenness and Simpson’s dominance (p = 0.01341 for all three indices; refer to Figure 3b; Supplementary Table S11).

Figure 3

Figure 3. Abundance and diversity of bacterial communities associated with cyanobacteria through the use of 16S rRNA amplicon sequencing. Bacteria not belonging to cyanobacteria were displayed to the level of genus (a) in triplicate samples (n = 3), and diversity indices (b) using averaged samples (n = 3). Analysis was conducted to the lowest reliable taxonomic rank (genus level). Higher-level taxa indicate unclassified members at the corresponding taxonomic rank. Unclassified sequences and other taxa with lower relative abundance (<1%) were excluded in (a) for better visualization, while all classified data were utilized for diversity analysis in (b). Significant differences among sites (TL: The Lake in Central Park – New York, USA; RA: Raczyńskie Lake, Poland; SU: Sulejów Reservoir, Poland; SP: Singapore urban reservoir) were estimated with non-parametric Kruskal–Wallis and Dunn’s post hoc test (p < 0.05). Sites with the same case letters (a–c) presented no significant differences.

3.4 Relations between cyanobacterial and other bacterial taxa

Differences in sites regarding the composition of bacterioplankton communities were described with principal component analysis (PCA; Figure 4). Scores and loadings for the construction of the PCA are presented in Supplementary Table S12, and the statistical analysis for the differentiation among sites in Supplementary Table S13. The horizontal axis (PC1–52.7% of variability) indicated significant differences in bacterioplankton communities for TL and RA when compared to SU and SP (Figure 4b). TL and RA where clearly distinguished by the cyanobacteria Microcystis PCC7914 and associated bacteria belonging to Roseomonas, an unclassified taxon of Burkholderiaceae and Sphingobacteriales_env.OPS17 (Figure 4a). RA was further differentiated from TL by the cyanobacterium Aphanizomenon NIES81 and other associated bacteria belonging to an unclassified taxon of Caldilineaceae (Figure 4a). SU was strongly differentiated solely by the cyanobacterium Aphanizomenon MDT14a, which was associated with Paucibacter (Figure 4a). The vertical axis (PC2–34.2% of variability) indicated that SP was the most different in bacterioplankton composition when compared to the other three sites (Figure 4a). This site was differentiated by Planktothrix NIVA CYA 15 and associated bacterial taxa belonging to Rheinheimera and Flavobacterium, and to a shorter extent, Rhodoferax and Emticicia (Figure 4a).

Figure 4

Figure 4. Principal component analysis (PCA) displaying the spatial distribution of study sites according to bacterioplankton composition (a). Unclassified sequences and other taxa with lower relative abundance (<1%) were excluded for better visualization. Significant differences (p < 0.05) among sites were estimated with non-parametric Kruskal–Wallis and Dunn’s post hoc test for the PC1 scores (b). Triplicate samples (1–3 of the same color) was utilized per study site (TL: The Lake in Central Park – New York, USA; RA: Raczyńskie Lake, Poland; SU: Sulejów Reservoir, Poland; SP: Singapore urban reservoir).

Correlation analysis between cyanobacteria and associated bacterial taxa were described (Figure 5; Supplementary Table S14) and revealed that Microcystis PCC7914 and Planktothrix NIVA CYA 15 had the highest number of significant correlations with associated bacterial taxa (Figure 5; Supplementary Table S14). Microcystis strongly correlated with 25 non-cyanobacterial taxa, of which 19 were significant positive correlations (Figure 5). In contrast, Planktothrix NIVA CYA 15 strongly correlated with 13 non-cyanobacterial taxa, of which 11 were significant positive correlations (Figure 5). Fewer significant positive and negative correlations were observed for Aphanizomenon strain, NIES81 (2 and 1, respectively, Figure 5).

Figure 5

Figure 5. Relations between cyanobacterial and other bacterial taxa for the study sites. Non-parametric Spearman’s correlations between the relative abundance of each cyanobacteria strain and other bacteria were utilized. Analysis was conducted to the lowest reliable taxonomic rank (genus level). Higher-level taxa indicate unclassified members at the corresponding taxonomic rank.

3.5 Nutrient-transforming genes in metagenomic constructs

The relative abundance of selected nitrogen cycling genes and normalized abundances to the 16S rRNA of total bacteria (Supplementary Table S15) were used in a PCA (Figure 6), with coordinates 1 and 2 showing the highest variability for the discrimination of samples and replicates (up to 87.85%). Scores and correlations for the construction of PCA are described in Supplementary Table S16. Genes involved in N-degradation (ureB, glnA, gcvT), DNRA (napA and narGHZ) and ANRA (narB) were detected across all sites. For TL, ureB, glnA, and narB were most abundant (Figure 6a), which separated TL from the other sites (Figure 6b). At RA and SU, the highest abundances were of the N-fixing gene nifK (Figure 6a), which separated the Polish sites from the others (Figure 6b). At SP, the abundances of several unique genes (narZH, norB, nosZ, nxrA, nxrB, and _camoA) were higher than the other sites (Figure 6a). Nitrifying genes were generally depleted in TL, RA, and SU (Figure 6a) compared to SP. ANNAMOX genes were not detected in any samples.

Figure 6

Figure 6. Relative abundances of nitrogen cycling genes in the study sites: (a) Heat map representing the differences in relative abundances of genes normalized with 16S rRNA of total bacteria; (b) principal components analysis (PCA) revealing the genes influencing differences among the study sites. Relative abundances were estimated using sample replicates (n = 2) for study sites (TL: The Lake in Central Park – New York, USA; RA: Raczyńskie Lake, Poland; SU: Sulejów Reservoir, Poland; SP: Singapore urban reservoir).

The relative abundance of selected P-cycling genes and normalized abundances to the 16S rRNA of total bacteria (Supplementary Table S15) were used in a PCA (Figure 7) with coordinates 1 and 2, showing the highest variability for the discrimination of samples and replicates (up to 92.07%). Scores and correlations for the construction of PCA are described in Supplementary Table S16. Sites TL and SP were clearly more abundant in P-cycling genes (Figure 7a). Both sites had similar high abundances of P-solubilizing genes phoL and phoR, and the P-accumulating gene ppk1 (Figure 7a). TL had the highest abundances of phoB and phoP (Figure 7a), which differentiated TL from the other sites (Figure 7b). In contrast, SP had the highest abundance of phoH (Figure 7a), which was a factor in the differentiation of SP (Figure 7b). RA and SU had the lowest relative abundances of P-cycling genes (Figure 7a), clustering together separate at the opposite side of the PC1 axis (Figure 7b).

Figure 7

Figure 7. Relative abundances of phosphorus cycling genes in the study sites: (a) Heat map representing the differences in relative abundances of genes normalized with 16S rRNA of total bacteria; (b) principal components analysis (PCA) revealing the genes influencing differences among the study sites. Relative abundances were estimated using sample replicates (n = 2) for study sites (TL: The lake in Central Park – New York, USA; RA: Raczyńskie Lake, Poland; SU: Sulejów Reservoir, Poland; SP: Singapore urban reservoir).

4 Discussion

The diversity and potential eco-physiology of global bacterioplankton communities during HCBs was investigated through synoptic sampling of four shallow freshwater sites located in three different continents. While TL and SP exhibited high P-PO₄³⁻ concentrations, SU and RA had very low P-PO₄³⁻ but measurable dissolved N-species (Table 1). Despite differences in nutrient conditions, all study lakes had phytoplankton communities dominated by cyanobacteria (35.6–68.7%; Figure 1a). However, the sites differed substantially in the cyanobacteria and associated bacterial community composition, toxigenicity, and eco-physiological functions.

It is important to acknowledge that this is a preliminary study based on a limited number of samples. While the dataset was sufficient for comparative analysis, results were interpreted with caution. Despite low sample sizes, the study provides baseline information and generates hypotheses about bacterial taxa potentially associated with cyanobacterial blooms. These findings provide a starting point for a broader, ongoing investigation with expanded sampling efforts across a wider range of freshwater systems.

4.1 Diversity of cyanobacteria and their toxigenicity

Colonial coccoidal forms of Microcystis were present in all sites, which is congruous with its ubiquitous nature and likely occurrence in eutrophic waters (Harke et al., 2016). At TL, Microcystis was dominant (Figure 1b) – an observation that was previously recorded annually from 2015 to 2020 (Flanzenbaum et al., 2022). The other three sites had higher abundances of colonial filamentous cyanobacteria (Aphanizomenon or Planktothrix; Figure 1b); in fact, RA and SU have had previously reported blooms composed of mixed assemblages (Microcystis and Aphanizomenon; Kowalczewska-Madura et al., 2022; Mankiewicz-Boczek and Font-Nájera, 2022). At SP, Planktothrix was dominant over Microcystis, which corresponds to previous observations that Microcystis blooms in this urban reservoir are being replaced by filamentous cyanobacteria (Te and Gin, 2011, and personal written communication of unpublished monitoring data by Te S. H. with Public Utilities Board of Singapore, 2024).

The cyanobacteria from the four sites had variable toxigenic potential (Figure 2a); however, all had the potential to produce MCs (Figure 2a). The strongest correlation between the mcyE gene and the abundance of Microcystis was at TL (Figure 2b). Flanzenbaum et al. (2022) corroborated elevated annual MCs and toxigenic potential at TL; they reported the highest gene copy numbers of mcyE occurred at mid- and post-summer seasons (up to 8.22 × 10⁵ copies mL⁻¹) when Microcystis dominated. Planktothrix is also well known to produce MCs, but a significant negative correlation between the mcyE gene and this cyanobacterium was observed at SP. In contrast, Microcystis correlated with the mcyE gene indicating that this coccoidal cyanobacteria is responsible for the potential production of MCs at SP. RA had the highest toxigenic potential for CYN, ATX, and STX, which was associated with the highest diversity and relative abundances of cyanobacterial taxa (Figure 1b). Interestingly, Aphanizomenon is one of the cyanobacterial genera implicated in the production of CYN, ATX, and STX (Wang et al., 2021). However, the positive correlations observed between cyanotoxin genes and Aphanizomenon NIES81 were not significant after p-value Bonferroni correction in this study (Figure 2b). Despite this, the presence of diverse toxicity potential in filamentous forms—particularly when co-occurring with hepatotoxic Microcystis colonies—may pose a greater environmental threat than blooms dominated by a single cyanobacterial type. Additionally, filamentous cyanobacteria are known to produce a greater array of cyanotoxins that are known to have higher toxicity than MCs (Christensen and Khan, 2020). In another study, a mixed bloom of Dolichospermum, Aphanizomenon, Planktothrix, and Microcystis in Utah Lake (USA) revealed a similar toxigenic potential (MCs, CYN and ATX) using constructed metagenomes (Li et al., 2023), raising concerns about the usage of water resources. In 2018, restoration efforts to reduce the occurrence of HCBs have been implemented in RA including phosphorus inactivation and biomanipulation techniques that resulted in water quality improvements during that summer. However, reductions in these restoration efforts during the subsequent years contributed to water quality deterioration and HCB recurrence (Kowalczewska-Madura et al., 2022). In this study, the presence of mixed cyanobacterial taxa and strong potential toxigenicity for RA indicates that sustainable restoration efforts applied with the same intensity as in 2018 could reduce the occurrence of HBCs at RA.

4.2 Diversity and eco-physiological traits of associated bacteria

Bacterial communities play an important role in the development and decay of HCBs; therefore, knowledge of their association with particular cyanobacterial taxa could reveal potential eco-physiological roles that can exacerbate health threats. In this study, bacterial communities were most diverse in a Microcystis-dominated HCB (Figure 3). Extracellular polysaccharides substances (EPS), in the mucilage of Microcystis helped explain high bacterial diversity indices. Pannard et al. (2016) demonstrated that EPS production in Microcystis aeruginosa was significantly higher than that of the filamentous cyanobacteria, Limnothrix sp. and Planktothrix agardhii in controlled laboratory assays. High EPS production in coccoidal forms of cyanobacteria was attributed to higher cellular surface-to-volume ratios, allowing more area for bacterial attachment. Furthermore, higher carbon (C): N ratios in EPS of M. aeruginosa suggest a potential role as a C-rich substrate that could support heterotrophic bacterial communities. In the current study, the previously mentioned factors contribute to a deeper understanding of the reduced diversity of bacterial communities associated with HCBs dominated by the filamentous cyanobacterium Planktothrix (Figure 3). This knowledge can be further extended to blooms dominated by Aphanizomenon, a cyanobacterium for which EPS production has not specifically been compared to coccoidal forms of cyanobacteria.

Among the sites, the cyanosphere was composed of unique communities of bacterial taxa, depending on the dominant cyanobacteria. Microcystis blooms were interrelated with Roseomonas and the family Burkholderiaceae in TL and RA (Figure 4a). Roseomonas contains genes involved in many metabolic pathways for obtaining energy from organic carbon (Cai et al., 2024), and Burkholderiaceae has numerous saprophytic strains known to utilize decaying organic matter (Coenye, 2014). Several other studies have described the interrelation of Microcystis with Roseomonas (Chun et al., 2020; Chen et al., 2020; Kim et al., 2019; Pérez-Carrascal et al., 2021) and Burkholderiaceae (Chun et al., 2020; Jankowiak and Gobler, 2020), although potential eco-physiological roles were not explicitly mentioned. Moreover, Microcystis was interrelated to Sphingobacteriales_env.OPS17 in TL (Figure 4a), another bacterial group described as Microcystis bloom specialists, favoring the uptake of cyanobacterial exudates and decaying material (Bagatini et al., 2014; Parulekar et al., 2017). The order Sphingobacteriales has been documented to contain microcystin-degrading bacteria (Kohler et al., 2014; Mankiewicz-Boczek and Font-Nájera, 2022), which was related to high MCs potential in Microcystis blooms (Figures 1b, 2a). These three bacterial taxa likely consist of opportunistic populations that thrive on the carbon-rich content of EPS within the Microcystis cyanosphere. Furthermore, it is probable that these three bacterial taxa play a substantial role in nutrient cycling processes within this environment. In the Planktothrix bloom in SP, a strong relation with Rheinheimera and Flavobacterium was observed (Figure 4a). Rheinheimera strains have been isolated from HCBs, and experimental laboratory assays have demonstrated their algicidal effects on M. aeruginosa (Wu et al., 2019), which suggests that this bacterium may regulate the growth of Planktothrix blooms in SP. Flavobacterium includes strains that are capable of cyanotoxin degradation and other recalcitrant and problematic organic compounds (Berg et al., 2009). Elevated dominance of Flavobacterium in Planktothrix blooms suggest that this bacterium plays a potential role in nutrient-cycling in the Planktothrix cyanosphere. Finally, in the case of Aphanizomenon (which was most dominant in SU), a strong relation was observed with Paucibacter (Figure 4a) – another potential algicidal bacteria shown to inhibit the growth of Microcystis and degrade cyanotoxins in a mesocosm study (Le et al., 2022b). Paucibacter has also been linked to a decline in cell growth of filamentous Dolichospermum blooms (Le et al., 2023), suggesting that it could be important for regulating Aphanizomenon blooms in SU. These results suggest that Microcystis blooms may have potentially stronger cyanobacterial-bacterial interactions, as indicated by the higher number of significant positive correlations observed between Microcystis and its attached bacterial consortia compared to Planktothrix or Aphanizomenon (Figure 5).

4.3 Potential pathogenic bacteria

HCBs could be sinks for dangerous pathogenic bacteria; however, 16S rRNA gene-based identification can only indicate the presence of genera with potential pathogenic members and cannot confirm pathogenicity at the strain level. Nevertheless, the cyanosphere has been described as an ideal environment for bacteria to confer resistance to many antimicrobial agents (Zhang et al., 2020). Knowledge about their presence within HCBs is essential because of additional threats to the health of ecosystems. In the present study, bacterial genera listed as priority pathogens by the World Health Organization (2024) – such as Mycobacterium, Pseudomonas, and Acinetobacter – were detected. However, their relative abundances were consistently low across all sites (<0.5% of relative abundances) and therefore, do not represent an elevated threat. Other bacteria that were frequently detected or associated with HCBs may pose a greater potential threat as pathogens. For example, Roseomonas was linked to Microcystis blooms in TL and RA (Figure 4a). Despite their ubiquity in natural environments, several strains have been isolated from clinical samples and are recognized as causes of opportunistic infections in humans (Ioannou et al., 2020). Roseomonas has been commonly associated with Microcystis blooms in previous studies (refer to section 4.2), although potential pathogenicity has not been properly addressed. In contrast, Flavobacterium may be an important potential pathogen for Planktothrix blooms in SP (Figure 4a). Flavobacterium is cosmopolitan in freshwater environments, nevertheless, most pathogenic strains have been isolated from animal tissue, and most cases of infection have been reported in freshwater fish (McBride, 2014). Zhang et al. (2020) reported the potential pathogenicity of Flavobacterium in Planktothrix blooms, and revealed that this bacterial group was associated with elevated amounts of antibiotic resistance genes. These observations underscore the need to further investigate the potential pathogenicity of Roseomonas and Flavobacterium, because they are key members of the consortia associated with coccoid and filamentous cyanobacterial blooms and appear to play important roles in nutrient cycling (refer to section 4.2).

4.4 Nutrient-transforming genes

The detection of nutrient-transforming bacterial genes varied significantly between sites dominated by non-diazotrophic Microcystis and Planktothrix (such as TL and SP, respectively) and those where diazotrophic Aphanizomenon was more abundant (RA and SU in Poland) (Figures 6, 7). In relation to N-cycling processes, the most abundant genes (ureB, glnA, gcvT, narB, and narZHG) indicated a potential preference for ammonium utilization by phytoplankton communities dominated by cyanobacteria. At TL and SP, N levels were below detection limits (Table 1), suggesting rapid N uptake by non-diazotrophic cyanobacteria. Coping mechanisms for N-uptake in Microcystis blooms at TL were associated with high relative abundances of N-synthesis and N-degradation genes (ureB and glnA) and ANRA genes (narB) (Figure 6). These genes are crucial to converting unavailable N to ammonium, which is largely regarded as the preferred N-species for Microcystis (Yan et al., 2023). In fact, these genes have been detected in other freshwater bodies with Microcystis blooms such as Lake Erie (OH, USA), Lake Agawam (NY, USA), and William H Harsha Lake (Lake Harsha, OH, USA) (Steffen et al., 2015; Gobler and Jankowiak, 2022; Wang et al., 2021). Wan et al. (2022) suggested that the N-degradation gene (glnA) and nitrate reduction gene (narB) contributed to ammonium accumulation in Tangzun and Zhiyin Lakes (China), resulting from the decay of a Dolichospermum bloom. This accumulation likely facilitated the subsequent appearance of Microcystis in late summer. In the Planktothrix bloom in SP, the DNRA gene (narHZ) was more abundant than other sites. Previous studies have suggested that Microcystis blooms harbor a large quantity of bacteria that perform DNRA, which is essential for N acquisition in non-diazotrophic cyanobacteria in N-limited environments (Yang et al., 2021; Yan et al., 2023; Wan et al., 2022; Cai et al., 2024). Observations in those studies indicated that Planktothrix might have similar N-scavenging strategies as have been previously demonstrated for Microcystis, especially considering that they both are non-diazotrophic. Furthermore, denitrifying genes (norB and nosZ) and nitrifying genes (nxrA, nxrB and _camoA) were more abundant in Planktothrix blooms compared to Microcystis (Figure 6). Denitrifying and nitrifying bacteria may compete with Microcystis for nutrients, because they eliminate nitrogen in freshwater ecosystems. Similar observations have been described in previous studies, where nitrifying and denitrifying genes were generally depleted during a Microcystis bloom (Yang et al., 2021; Cai et al., 2024), indicating that Planktothrix may rely on different mechanisms for N obtainment that have not been identified or described. However, the dominance of Planktothrix in SP was not as pronounced as that of Microcystis in TL (Figure 1b). Consequently, it is possible that a greater number of N-cycling processes may have been observed because of reduced competitive pressures among bacterioplankton communities. In contrast, Aphanizomenon blooms at the Polish sites (RA and SU) were associated with the highest abundances of the N-fixing gene nifK (Figure 6) which probably contributed to the highest observed concentrations of N-species for those sites (up to 28.69 μg L⁻¹ of N-NH₄⁺ and 36.87 μg L⁻¹ of N-NO₃⁻, Table 1). Other studies have shown that metagenomes of Aphanizomenon and Dolichospermum blooms were comprised of N-fixing genes that were considered important biological factors supporting N-processes for other bacterioplankton (Li et al., 2023; Yang et al., 2021; Wan et al., 2022). In our study, N-fixing genes likely contributed to the sustained presence of mixed blooms of diazo- and non-diazotrophic cyanobacteria in RA and SU.

The non-diazotrophic taxa Microcystis and Planktothrix exhibited higher abundances of phosphorus cycling genes in their bacterial consortia. Alkaline phosphatases in particular, were abundant in Microcystis and Planktothrix blooms (phoBP and phoHR, respectively, Figure 7). Alkaline phosphatases are generally enriched in Microcystis blooms and have been associated with their ability to thrive in P-limited environments (Yang et al., 2021; Cai et al., 2024). However, both TL and SP exhibited high concentrations of phosphates (105.6 μg L⁻¹ and 34.00 μg L⁻¹ of P-PO₄³⁻, respectively, Table 1) and were not P-limited. The shallowness and proximity to the sediments at these sites facilitates rapid microbial solubilization of organic phosphorus to inorganic forms (Flanzenbaum et al., 2022). Wan et al. (2022) suggests that the phenomenon may also be related to detritus from Microcystis decomposition, leading to phosphorus release and solubilization. While there are no specific observations for Planktothrix blooms, our results suggest a similar dynamic to that observed in Microcystis-dominated blooms. Furthermore, non-diazotrophic cyanobacteria were associated with high abundances of the ppk1 gene (Figure 6), highlighting their potential role in phosphorus accumulation and removal from water. However, the P-rich environments of TL and SP suggest that phosphorus is not a limiting factor for the growth of these cyanobacterial communities. In fact, both Microcystis and Planktothrix are known to thrive in P-rich waters (Kurmayer et al., 2016; Flanzenbaum et al., 2022; Wejnerowski et al., 2024). Finally, the low abundance of phosphorus-cycling genes associated with Aphanizomenon at RA and SU likely demonstrates the less complex microbial communities observed in their cyanosphere. However, additional studies would be needed to confirm this hypothesis because it was beyond the scope of our study.

Our study results underscore the importance of understanding the diversity of bacterial communities and genes associated with HCBs, as by revealing critical biological elements that enhance the persistence, toxicity and spread of HCBs in diverse ecosystems worldwide. This knowledge could help in elucidate the specific molecular mechanisms that enable cyanobacterial persistence in aquatic ecosystems, which could be used for developing more effective systemic solutions for the prevention, control and mitigation of HCBs. Our work represents an initial effort in the characterization of potential hazard elements associated with HCBs on a global scale, that could be extended to additional freshwater bodies in other countries in the near future.

Data availability statement

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Materials and Methods section (Bioinformatic analyses).

Author contributions

JM-B: Funding acquisition, Resources, Writing – review & editing, Conceptualization, Project administration, Writing – original draft, Methodology, Investigation, Supervision. AF-N: Investigation, Formal analysis, Writing – review & editing, Data curation, Visualization, Writing – original draft. KG: Writing – review & editing, Resources. JG: Resources, Writing – review & editing. DS: Writing – review & editing. RG: Formal analysis, Writing – review & editing. JK: Writing – review & editing, Formal analysis. ST: Formal analysis, Writing – review & editing. MK: Writing – review & editing, Formal analysis. MSk: Formal analysis, Writing – review & editing. MSe: Writing – review & editing, Formal analysis. FL-H: Formal analysis, Writing – review & editing.

Funding

The author(s) declare that financial support was received for the research and/or publication of this article. This research was funded by the National Science Centre, project no. 2022/47/NZ8/00689, CyMiBiom, “Intercontinental comparison of bacterial and archaeal communities associated with cosmopolitan cyanobacterium Microcystis – unveiling their ecological roles in anthropopressure and climate change.”

Acknowledgments

Nutrient and field parameter data from TL are available from the USGS Water Data for the Nation: National Water Information database (U.S. Geological Survey, 2025). Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government. Environmental data concerning RA and SU water bodies are not publicly available; they are stored in an internal database at the UNESCO Chair in Ecohydrology and Applied Ecology at the University of Lodz, Poland. The environmental monitoring data for SP are agency in-house records and are not publicly available. Access may be requested from the agency subject to data sharing policies. The authors sincerely thank Dr. hab Mikołaj Kokociński from the Department of Hidrobiology at the Adam Mickiewicz University in Poland, for his valuable help in sample collection.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The authors declare that no Gen AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmicb.2025.1655370/full#supplementary-material

Footnotes

^https://www.arb-silva.de/

^https://www.ncbi.nlm.nih.gov/sra

References

Bagatini, I. L., Eiler, A., Bertilsson, S., Klaveness, D., Tessarolli, L. P., and Henriques Vieira, A. A. (2014). Host-specificity and dynamics in bacterial communities associated with bloom-forming freshwater phytoplankton. PLoS One 9:e85950. doi: 10.1371/journal.pone.0085950

Crossref Full Text | Google Scholar

Bell, T. A., Sen-Kilic, E., Felföldi, T., Vasas, G., Fields, M. W., and Peyton, B. M. (2018). Microbial community changes during a toxic cyanobacterial bloom in an alkaline Hungarian lake. Antonie Van Leeuwenhoek 111, 2425–2440. doi: 10.1007/s10482-018-1132-7,

PubMed Abstract | Crossref Full Text | Google Scholar

Berg, K. A., Lyra, C., Sivonen, K., Lars, P., Suomalainen, S., Tuomi, P., et al. (2009). High diversity of cultivable heterotrophic bacteria in association with cyanobacterial water blooms. ISME J. 3, 314–325. doi: 10.1038/ismej.2008.110

Crossref Full Text | Google Scholar

Bolger, A. M., Lohse, M., and Usadel, B. (2014). Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114–2120. doi: 10.1093/bioinformatics/btu170,

PubMed Abstract | Crossref Full Text | Google Scholar

Burford, M. A., Carey, C. C., Hamilton, D. P., Huisman, D. P., Paerl, H. W., Wood, S. A., et al. (2020). Perspective: advancing the research agenda for improving understanding of cyanobacteria in a future of global change. Harmful Algae 91:101601. doi: 10.1016/j.hal.2019.04.004

Crossref Full Text | Google Scholar

Cai, H., McLiman, C. J., Jiang, H., Chen, F., Krumholz, L. R., and Hambright, K. D. (2024). Aerobic anoxygenic phototrophs play important roles in nutrient cycling within cyanobacterial Microcystis bloom microbiomes. Microbiome 12:88. doi: 10.1186/s40168-024-01801-4

Crossref Full Text | Google Scholar

Callahan, B. J., McMurdie, P. J., Rosen, M. J., Han, A. W., Johnson, A. J. A., and Holmes, S. P. (2016). DADA2: high-resolution sample inference from Illumina amplicon data. Nat. Methods 13, 581–583. doi: 10.1038/nmeth.3869,

PubMed Abstract | Crossref Full Text | Google Scholar

Chen, S., Yan, M., Huang, T., Zhang, H., Liu, K., Huang, X., et al. (2020). Disentangling the drivers of Microcystis decomposition: Metabolic profile and co-occurrence of bacterial community. Sci. Total Environ. 739:140062. doi: 10.1016/j.scitotenv.2020.140062,

PubMed Abstract | Crossref Full Text | Google Scholar

Christensen, V. G., and Khan, E. (2020). Freshwater neurotoxins and concerns for human, animal, and ecosystem health: A review of anatoxin-a and saxitoxin. Sci. Total Environ. 736:139515. doi: 10.1016/j.scitotenv.2020.139515,

PubMed Abstract | Crossref Full Text | Google Scholar

Chun, S. J., Cui, Y., Lee, J. J., Choi, I. C., Oh, H. M., and Ahn, C. Y. (2020). Network analysis reveals succession of Microcystis genotypes accompanying distinctive microbial modules with recurrent patterns. Water Res. 170:115326. doi: 10.1016/j.watres.2019.115326,

PubMed Abstract | Crossref Full Text | Google Scholar

Coenye, T. (2014). “The family Burkholderiaceae” in The prokaryotes. eds. E. Rosenberg, E. F. DeLong, S. Lory, E. Stackebrandt, and F. Thompson (Berlin, Heidelberg: Springer).

Google Scholar

Cook, K. V., Li, C., Cai, H., Krumholz, L. R., Hambright, K. D., Paerl, H. W., et al. (2020). The global Microcystis interactome. Limnol. Oceanogr. 65, S194–S207. doi: 10.1002/lno.11361

Crossref Full Text | Google Scholar

De Nobel, W. T., Huisman, J., Snoep, J. L., and Mur, L. R. (1997). Competition for phosphorus between the nitrogen-fixing cyanobacteria Anabaena and Aphanizomenon. FEMS Microbiol. Ecol. 24, 259–267. doi: 10.1111/j.1574-6941.1997.tb00443.x

Crossref Full Text | Google Scholar

De Nobel, W. T., Staats, N., and Mur, L. R. (1995). Competition between nitrogen-fixing cyanobacteria during phosphorus-limited growth. Water Sci. Technol. 32, 99–101. doi: 10.1016/0273-1223(95)00685-0

Crossref Full Text | Google Scholar

Fisher, C. (2011). Nature in the city: urban environmental history and Central Park. OAH Mag. Hist. 25, 27–31. doi: 10.1093/oahmag/oar038

Crossref Full Text | Google Scholar

Fishman, M. J. (Ed.) (1993). Methods of analysis by the U.S. Geological Survey National Water Quality Laboratory—Determinations of inorganic and organic constituents in water and fluvial sediments. Reston, Virginia, USA: U.S. Geological Survey Open-File Report, 93–125, 217.

Google Scholar

Flanzenbaum, J. M., Jankowiak, J. G., Goleski, J. A., Gorney, R. M., and Gobler, C. J. (2022). Nitrogen limitation of intense and toxic cyanobacteria blooms in lakes within two of the most visited parks in the USA: the Lake in Central Park and Prospect Park Lake. Toxins 14:684. doi: 10.3390/toxins14100684,

PubMed Abstract | Crossref Full Text | Google Scholar

Gągała, I., Izydorczyk, K., Jurczak, T., Pawełczyk, J., Dziadek, J., Wojtal-Frankiewicz, A., et al. (2014). Role of Environmental Factors and Toxic Genotypes in the Regulation of Microcystins – Producing Cyanobacterial Blooms. Microb. Ecol. 67, 465–479. doi: 10.1007/s00248-013-0303-3,

PubMed Abstract | Crossref Full Text | Google Scholar

Gao, G., Bai, D., Li, T., Jia, Y., Li, J., Wang, Z., et al. (2025). Understanding filamentous cyanobacteria and their adaptive niches in Lake Honghu, a shallow eutrophic lake. J. Environ. Sci. 152, 219–234. doi: 10.1016/j.jes.2024.05.010,

PubMed Abstract | Crossref Full Text | Google Scholar

Gobler, C. J., and Jankowiak, J. G. (2022). Dynamic Responses of Endosymbiotic Microbial Communities Within Microcystis Colonies in North American Lakes to Altered Nitrogen, Phosphorus, and Temperature Levels. Front. Microbiol. 12:781500. doi: 10.3389/fmicb.2021.781500,

PubMed Abstract | Crossref Full Text | Google Scholar

Hammer, O., Harper, D. A. T., and Ryan, P. D. (2001). PAST: paleontological statistics software package for education and data analysis. Palaeontol. Electron. 4, 1–9. Available online at:https://palaeo-electronica.org/2001_1/past/issue1_01.htm (Accessed July, 2024).

Google Scholar

Harke, M. J., Berry, D. L., Ammerman, J. W., and Gobler, C. J. (2012). Molecular response of the bloom-forming cyanobacterium, Microcystis aeruginosa, to phosphorus limitation. Microb. Ecol. 63, 188–198. doi: 10.1007/s00248-011-9894-8,

PubMed Abstract | Crossref Full Text | Google Scholar

Harke, M. J., Steffen, N. M., Gobler, C. H., Otten, T. G., Wilhelm, S. W., Wood, S. A., et al. (2016). A review of the global ecology, genomics, and biogeography of the toxic cyanobacterium Microcystis spp. Harmful Algae 54, 4–20. doi: 10.1016/j.hal.2015.12.007

Crossref Full Text | Google Scholar

Illumina. (2021). Nextera XT DNA Library Preparation Guide (document#15031942 v06). Illumina Inc. Available online at: https://support-docs.illumina.com/LP/NexteraXTRef/Content/LP/Nextera/XT/Protocol.htm (Accessed December, 2023).

Google Scholar

Illumina. (2023). 16S Metagenomic Sequencing Library Preparation Guide (Preparing 16S ribosomal RNA gene amplicons for the Illumina MiSeq system) (document # 15031942 v07). Available online at: https://support.illumina.com/documents/documentation/chemistry_documentation/16s/16s-metagenomic-library-prep-guide-15044223-b.pdf (Accessed December, 2023).

Google Scholar

Ioannou, P., Mavrikaki, V., and Kofteridis, D. P. (2020). Roseomonas species infections in humans: a systematic review. J. Chemother. 32, 226–236. doi: 10.1080/1120009X.2020.1785742,

PubMed Abstract | Crossref Full Text | Google Scholar

Izydorczyk, K., Jurczak, T., Wojtal-Frankiewicz, A., Skowron, A., Mankiewicz-Boczek, J., and Tarczyńska, M. (2008). Influence of abiotic and abiotic factors on microcystin content in Microcystis aeruginosa cells in a eutrophic temperate reservoir. J. Plankton Res. 30, 393–400. doi: 10.1093/plankt/fbn006

Crossref Full Text | Google Scholar

Jankowiak, J. G., and Gobler, C. J. (2020). The composition and function of microbiomes within Microcystis colonies are significantly different than native bacterial assemblages in two North American lakes. Front. Microbiol. 11, 1–26. doi: 10.3389/fmicb.2020.01016

Crossref Full Text | Google Scholar

Jaskulska, A., Font-Nájera, A., Czarny, P., Serwecińska, L., and Mankiewicz-Boczek, J. (2021). Daily dynamic of transcripts abundance of Ma-LMM01-like cyanophages in two lowland European reservoirs. Ecohydrol. Hydrobiol. 21, 543–548. doi: 10.1016/j.ecohyd.2021.07.003

Crossref Full Text | Google Scholar

Kim, M., Shin, B., Lee, J., Park, H. Y., and Park, W. (2019). Culture-independent and cultre-dependent analyses of the bacterial community in the phycosphere of cyanobloom-forming Microcystis aeruginosa. Sci. Rep. 9:20416. doi: 10.1038/s41598-019-56882-1,

PubMed Abstract | Crossref Full Text | Google Scholar

Klindworth, A., Pruesse, E., Schweer, T., Peplies, J., Quast, C., Horn, M., et al. (2013). Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Res. 41, 1–11. doi: 10.1093/nar/gks808

Crossref Full Text | Google Scholar

Kohler, E., Villiger, J., Posch, T., Derlon, N., Shabarova, T., Morgenroth, E., et al. (2014). Biodegradation of microcystins during gravity-driven membrane (GDM) ultrafiltration. PLoS One 9:e111794. doi: 10.1371/journal.pone.0111794,

PubMed Abstract | Crossref Full Text | Google Scholar

Kokociński, M., Stefaniak, K., Izydorczyk, K., Jurczak, T., Mankiewicz-Boczek, J., and Soininen, J. (2011). Temporal variation in microcystin production by Planktothrix agardhii (Gomont) Anagnostidis and Komárek (Cyanobacteria, Oscillatoriales) in a temperate lake. Ann. Limnol. Int. J. Lim. 47, 363–371. doi: 10.1051/limn/2011046

Crossref Full Text | Google Scholar

Kowalczewska-Madura, K., Kozak, A., Kuczyńska-Kippen, N., Dondajewska-Pielka, R., and Gołdyn, R. (2022). Sustainable restoration as a tool for the improvement of water quality in a shallow, hypertrophic lake. Water 14:1005. doi: 10.3390/w14071005

Crossref Full Text | Google Scholar

Kurmayer, R., Deng, L., and Entfellner, E. (2016). Role of toxic and bioactive secondary metabolites in colonization and bloom formation by filamentous cyanobacteria Planktothrix. Harmful Algae 54, 69–86. doi: 10.1016/j.hal.2016.01.004,

PubMed Abstract | Crossref Full Text | Google Scholar

Le, V. V., Ko, S. R., Kang, M., Lee, S. A., Oh, H. M., and Ahn, C. Y. (2022b). Algicidal capacity of Paucibacter aquatile DH15 on Microcystis aeruginosa by attachment and non-attachment effects. Environ. Pollut. 302:119079. doi: 10.1016/j.envpol.2022.119079,

PubMed Abstract | Crossref Full Text | Google Scholar

Le, V. V., Ko, S. R., Kang, M., Oh, H. M., and Ahn, C. Y. (2023). Effective control of harmful Microcystis blooms by paucibactin A, a novel macrocyclic tambjamine, isolated from Paucibacter aquatile DH15. J. Clean. Prod. 383:135408. doi: 10.1016/j.jclepro.2022.135408

Crossref Full Text | Google Scholar

Le, V. V., Srivastava, A., Ko, S. R., Ahn, C. Y., and Oh, H. M. (2022a). Microcystis colony formation: Extracellular polymeric substance, associated microorganisms, and its application. Bioresour. Technol. 360:127610. doi: 10.1016/j.biortech.2022.127610,

PubMed Abstract | Crossref Full Text | Google Scholar

Le, V. V., Tran, Q. G., Ko, S. R., Oh, H. M., and Ahn, C.-Y. (2024). Insights into cyanobacterial blooms through the lens of omics. Sci. Total Environ. 934:173028. doi: 10.1016/j.scitotenv.2024.173028,

PubMed Abstract | Crossref Full Text | Google Scholar

Li, H., Bhattarai, B., Barber, M., and Goel, R. (2023). Stringent Response of Cyanobacteria and Other Bacterioplankton during Different Stages of a Harmful Cyanobacterial Bloom. Environ. Sci. Technol. 57, 16016–16032. doi: 10.1021/acs.est.3c03114,

PubMed Abstract | Crossref Full Text | Google Scholar

Mankiewicz-Boczek, J., and Font-Nájera, A. (2022). Temporal and functional interrelationships between bacterioplankton communities and the development of a toxigenic Microcystis bloom in a lowland European reservoir. Sci. Rep. 12:19332. doi: 10.1038/s41598-022-23671-2,

PubMed Abstract | Crossref Full Text | Google Scholar

Mankiewicz-Boczek, J., Gagała, I., Kokociński, M., Jurczak, T., and Stefaniak, K. (2011). Perennial toxigenic Planktothrix agardhii bloom in selected lakes of Western Poland. Environ. Toxicol. 26, 10–20. doi: 10.1002/tox.20524,

PubMed Abstract | Crossref Full Text | Google Scholar

McBride, M. J. (2014). “The family Flavobacteriaceae” in The prokaryotes. eds. E. Rosenberg, E. F. DeLong, S. Lory, E. Stackebrandt, and F. Thompson (Berlin, Heidelberg: Springer).

Google Scholar

McMurdie, P. J., and Holmes, S. (2013). Phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One 8:e61217. doi: 10.1371/journal.pone.0061217,

PubMed Abstract | Crossref Full Text | Google Scholar

Paerl, H. W. (2014). Mitigating harmful cyanobacterial blooms in a human- and climatically-altered world. Life. 4, 988–1012. doi: 10.3390/life4040988

Crossref Full Text | Google Scholar

Paerl, H. W., and Barnard, M. A. (2020). Mitigating the global expansion of harmful cyanobacterial blooms: Moving targets in a human- and climatically-altered world. Harmful Algae 96:101845. doi: 10.1016/j.hal.2020.101845,

PubMed Abstract | Crossref Full Text | Google Scholar

Pannard, A., Pedrono, J., Bormans, M., Brian, E., Claquin, P., and Lagadeuc, Y. (2016). Production of exopolymers (EPS) by cyanobacteria: impact on the carbon-to-nutrient ration of the particulate organic matter. Aquat. Ecol. 50, 29–44. doi: 10.1007/s10452-015-9550-3

Crossref Full Text | Google Scholar

Parulekar, N. N., Kolekar, P., Jenkins, A., Kleiven, S., Utkilen, H., Johansen, A., et al. (2017). Characterization of bacterial community associated with phytoplankton blooms in a eutrophic lake in South Norway using 16S rRNA gene amplicon sequence analysis. PLoS One 12:e0173408. doi: 10.1371/journal.pone.0173408

Crossref Full Text | Google Scholar

Pérez-Carrascal, O. M., Tromas, N., Terrat, Y., Moreno, E., Giani, A., Braga Marques, L. C., et al. (2021). Single -colony sequencing reveals microbe-by-microbiome phylosymbiosis between the cyanobacterium Microcystis and its associated bacteria. Microbiome 9:194. doi: 10.1186/x40168-021-01140-8

Crossref Full Text | Google Scholar

Python Software Foundation (2020). Python 2.7.18 documentation. Available online at: https://docs.python.org/2.7/ (Accessed August 2024).

Google Scholar

R core Team (2023). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing.

Google Scholar

Steffen, M. M., Belisle, B. S., Watson, S. B., Boyer, G. L., Bourbonniere, R. A., and Wilhelm, S. W. (2015). Metatranscriptomic evidence for co-occurring top-down nd bottom-up controls on toxic cyanobacterial communities. Appl. Environ. Microbiol. 81, 3268–3276. doi: 10.1128/AEM.04101-14,

PubMed Abstract | Crossref Full Text | Google Scholar

Svirčev, Z., Lalić, D., Bojadžija Savić, G., Tokodi, N., Drobac Backović, D., Chen, L., et al. (2019). Global geographical and historical overview of cyanotoxin distribution and cyanobacterial poisonings. Arch. Toxicol. 93, 2429–2481. doi: 10.1007/s00204-019-02524-4,

PubMed Abstract | Crossref Full Text | Google Scholar

Te, S. H., and Gin, K. Y.-H. (2011). The dynamics of cyanobacteria and microcystin production in a tropical reservoir of Singapore. Harmful Algae 10, 319–329. doi: 10.1016/j.hal.2010.11.006

Crossref Full Text | Google Scholar

U.S. Geological Survey (2025). USGS Water Data for the Nation: National Water Information System database. Reston, Virginia, USA: U.S. Geological Survey.

Google Scholar

Underwood, J. C., Hall, N. C., Mumford, A. C., Harvey, R. W., Bliznik, P. A., and Jeanis, K. M. (2024). Relation between the relative abundance and collapse of Aphanizomenon flos-aquae and microbial antagonism in Upper Klamath Lake, Oregon. FEMS Microbiol. Ecol. 100:fiae043. doi: 10.1093/femsec/fiae043,

PubMed Abstract | Crossref Full Text | Google Scholar

Wan, L., Cao, L., Song, C., Cao, X., and Zhou, Y. (2022). Regulation of the nutrient cycle pathway and the microbial loop structure by different types of dissolved organic matter decomposition in lakes. Environ. Sci. Technol. 57, 297–309. doi: 10.1021/acs.est.2c06912,

PubMed Abstract | Crossref Full Text | Google Scholar

Wang, K., Mou, X., Cao, H., Struewing, I., Allen, J., and Lu, J. (2021). Co-occurring microorganisms regulate the succession of cyanobacterial harmful algal blooms. Environ. Pollut. 288:117682. doi: 10.1016/j.envpol.2021.117682,

PubMed Abstract | Crossref Full Text | Google Scholar

Wejnerowski, Ł., Dulić, T., Akter, S., Font-Nájera, A., Rybak, M., Kamiński, O., et al. (2024). Community Structure and Toxicity Potential of Cyanobacteria during Summer and Winter in a Temperate-Zone Lake Susceptible to Phytoplankton Blooms. Toxins (Basel) 16:357. doi: 10.3390/toxins16080357,

PubMed Abstract | Crossref Full Text | Google Scholar

World Health Organization (2024). WHO bacterial priority pathogens list, 2024: bacterial pathogens of public health importance, to guide research, development and strategies to prevent and control antimicrobial resistance. Geneva, Switzerland: World Health Organization.

Google Scholar

Wu, Q., Zhang, X., Jia, S., Li, J., and Li, P. (2019). Effects of the cultivable bacteria attached to Microcystis colonies on the colony size and growth of Microcystis. J. Freshw. Ecol. 34, 663–673. doi: 10.1080/02705060.2019.1665115

Crossref Full Text | Google Scholar

Yan, Z., Liu, Z., Jia, Z., Song, C., Cao, X., and Zhou, Y. (2023). Metabolites of extracellular organic matter from Microcystis and Dolichospermum drive distinct modes of carbon, nitrogen, and phosphorus recycling. Sci. Total Environ. 865:161124. doi: 10.1016/j.scitotenv.2022.161124,

PubMed Abstract | Crossref Full Text | Google Scholar

Yang, L., Cao, X. Y., Chen, X. Y., Deng, Q. H., Wan, L. L., Li, X. W., et al. (2021). Community composition and functional genes explain different ecological roles of heterotrophic bacteria attached to two bloom-forming cyanobacterial genera. Sci. Total Environ. 758:143850. doi: 10.1016/j.scitotenv.2020.143850,

PubMed Abstract | Crossref Full Text | Google Scholar

Zhang, Q., Zhang, Z., Lu, T., Peijnenburg, W. J. G. M., Gillings, M., Yang, X., et al. (2020). Cyanobacterial blooms contribute to the diversity of antibiotic-resistance genes in aquatic ecosystems. Commun. Biol. 3:737. doi: 10.1038/s42003-020-01468-1,

PubMed Abstract | Crossref Full Text | Google Scholar