Multi-year molecular quantification and 'omics analysis of Planktothrix-specific cyanophage sequences from Sandusky Bay, Lake Erie

Abstract

Introduction: Planktothrix agardhii is a microcystin-producing cyanobacterium found in Sandusky Bay, a shallow and turbid embayment of Lake Erie. Previous work in other systems has indicated that cyanophages are an important natural control factor of harmful algal blooms. Currently, there are few cyanophages that are known to infect P. agardhii, with the best-known being PaV-LD, a tail-less cyanophage isolated from Lake Donghu, China. Presented here is a molecular characterization of Planktothrix specific cyanophages in Sandusky Bay.

Methods and results: Putative Planktothrix-specific viral sequences from metagenomic data from the bay in 2013, 2018, and 2019 were identified by two approaches: homology to known phage PaV-LD, or through matching CRISPR spacer sequences with Planktothrix host genomes. Several contigs were identified as having viral signatures, either related to PaV-LD or potentially novel sequences. Transcriptomic data from 2015, 2018, and 2019 were also employed for the further identification of cyanophages, as well as gene expression of select viral sequences. Finally, viral quantification was tested using qPCR in 2015-2019 for PaV-LD like cyanophages to identify the relationship between presence and gene expression of these cyanophages. Notably, while PaV-LD like cyanophages were in high abundance over the course of multiple years (qPCR), transcriptomic analysis revealed only low levels of viral gene expression.

Discussion: This work aims to provide a broader understanding of Planktothrix cyanophage diversity with the goals of teasing apart the role of cyanophages in the control and regulation of harmful algal blooms and designing monitoring methodology for potential toxin-releasing lysis events.

Keywords: PaV-LD; Planktothrix agardhii; cyanophage; metagenome; metatranscriptome; qPCR.

Figure 1

Map of sampling locations in Sandusky Bay. The inset shows the location of Sandusky Bay in relation to the western basin of Lake Erie. Sites were chosen to provide representation of the width and depth of the bay. Sites were ODNR4 (41.453333, −82.960767), Edison bridge (41.480156, −82.834328), ODNR1 (41.477367, −82.739783), and EC1163 (41.469000, −82.715000).

Figure 2

Quantitative PCR of PaV-LD like cyanophages and their host, Planktothrix agardhii, across five bloom-forming years in Sandusky Bay (Table 1). Planktothrix agardhii was quantified using the single copy housekeeping gene, rpoC1. PaV-LD like cyanophages were quantified using the major capsid protein gene sequence (PaVLD_ORF073R). Samples were particulate associated (0.22 μm filter) or free cyanophage (cation charged filter). Samples were analyzed as biological duplicates and standard deviation is indicated by the black bars.

Figure 3

Contigs with CRISPR-cas identified viral signatures were aligned with the published genome of PaV-LD. 24 contigs were closely related (greatest identity % ≥ 85). White regions indicate 100% similar, light grey regions indicate greater than 80% similar, while both dark grey and black regions indicate discordant regions (insertions, deletions, and nucleotide substitutions). Reference sequence is highlighted in yellow and includes gene annotations. Coverage is noted using blue.

Figure 4

Proteomic viral tree (VIPtree) generated from metagenome generated contigs with positive CRISPR-cas spacer hits and viral signatures as identified by VirSorter. These 24 contigs can be placed into one of six proteomic groups, which may represent fragments of novel cyanophage genomes, or different segments of the same virus.

Figure 5

Nucleotide comparison of mapped contigs to PaVLD_ORF073R. Reference sequence is highlighted in yellow and includes gene annotations. Black segments in the non-highlighted sequences indicate points of difference, grey segments indicate similar regions. Identity is displayed at the top of the figure. Green sites indicate the same residue across all sequences while yellow sites have 30%–100% identity.

Figure 6

Transcripts of putative viral sequences categorized by group and normalized by whole genome expression of Planktothrix agardhii. Relative transcript abundance is presented as reads per kilobase of transcript per million mapped reads (RPKM). Viral groups with RPKM consistently less than 10% of the host RPKM were removed. Dashed line represents a viral to host transcript ratio of 1, where higher values indicate an increased likelihood of widespread active infection.