Whole-Genome Sequencing of Bacterial Isolates That Degrade the Cyanobacterial Toxin Microcystin-LR

ABSTRACT

ANNOUNCEMENT

Cyanobacteria can form large aggregations (i.e., harmful algal blooms [HABs]) that foul water bodies and threaten human health by releasing cyanotoxins, including microcystin-LR (MC-LR) (1). MC-LR causes many human health problems, including liver cancer, and its cyclic structure makes it stable in the environment (2,–5). MC-LR threatens drinking water around the world, and while municipal water treatment facilities have treatment options for removing MC-LR, these processes have limited capacity, are inhibited by organic materials, are expensive, and generate other toxic by-products which must be further mitigated (2, 6, 7). Bioremediation—the use of microbes to remove contaminants or toxic materials—has been proposed as a potential solution for mitigating HAB toxins (8). For example, a Sphingomonas sp. was found to degrade MC-LR using the mlrABCD operon, and sand filters containing this bacterium removed 90% of MC-LR from contaminated water (9, 10). Subsequently, other MC-LR-degrading bacteria have been reported (11). We previously isolated and provided preliminary species identification of 13 bacterial isolates from Lake Erie that degraded MC-LR into nontoxic fragments (12). Interestingly, degradation was observed only when isolates were combined into small groups of bacteria, not by individual bacterial isolates. Additionally, mlrABCD-related genes were not detected in our isolates, indicating degradation by one or more unique pathways. Here, genomic sequencing was performed to better characterize these isolates and identify putative MC-LR degradation pathways.

Isolates were plated onto Reasoner’s 2A (R2A) agar medium (BD) containing 10 μg/L MC-LR (Cayman Chemical) and incubated at room temperature for 48 to 72 h, and genomic DNA was extracted from single colonies using the NucleoSpin microbial DNA kit (Macherey-Nagel). Libraries were constructed using the Nextera XT kit, and single-end sequencing was performed using the Illumina MiSeq v3 SE150 platform. The total number of reads generated for each isolate is listed in Table 1. The raw sequencing reads were filtered using BBDuk v38.79 (sourceforge.net/projects/bbmap) and Trimmomatic v0.36 (13), with default parameters. The filtered reads were assembled using SPAdes v3.13.0 (14), with a minimum contig size of 500 bp and >5× genome coverage. For 12 of the 13 isolates, genome assembly resulted in <350 contigs (Table 1). Because of high contig numbers, isolate ODNR6CL was resequenced using the Illumina v2 Micro PE150 platform and subjected to the same assembly parameters described above, resulting in 34 final contigs (Table 1). Genus (≥25% nucleotide identity) and species (≥97% nucleotide identity) designations were assigned by submitting 10-kb sequences from each isolate for NCBI BLASTn analysis (15) (Table 1). Compared to previous 16S rRNA genus and species assignments (12), genome sequencing resulted in 6 genus changes and 5 species changes (Table 1). The ranges of genera/species, genome sizes, predicted coding genes, and GC contents highlight the diversity of MC-LR-degrading bacteria (Table 1). The bacterial genomes were putatively annotated using NCBI PGAP (16). Confirming our previous results (12), no known mlrABCD orthologs were identified in the draft genomes. Efforts are currently focused on identifying putative MC-LR-degrading enzymes and confirming these predictions by transcriptomic analyses.

ACKNOWLEDGMENTS

REFERENCES