Metagenomic assembly of new (sub)arctic Cyanobacteria and their associated microbiome from non-axenic cultures

Abstract

Cyanobacteria form one of the most diversified phylum of Bacteria. They are important ecologically as primary producers, for Earth evolution and biotechnological applications. Yet, Cyanobacteria are notably difficult to purify and grow axenically, 1 and most strains in culture collections contain heterotrophic bacteria that were likely associated to Cyanobacteria in the environment. Obtaining cyanobacterial DNA without contaminant sequences is thus a challenging and time-consuming task.Here, we deploy a metagenomic pipeline that enables the easy recovery of highquality genomes from non-axenic cultures. We tested this pipeline on 17 cyanobacterial cultures from the BCCM/ULC public collection and generated novel genome sequences for 15 arctic or subarctic strains, of which 14 early-branching organisms that will be useful for cyanobacterial phylogenomics. In parallel, we managed to assemble 31 co-cultivated bacteria from the same cultures and showed that they mostly belong to Bacteroidetes and Proteobacteria, some of them being very closely related in spite of geographically distant sampling sites. ImportanceComplete genomes of cold-adapted Cyanobacteria are underrepresented in databases, due to the difficulty to grow them axenically. In this work, we report the genome sequencing of 12 (sub)arctic and 3 temperate Cyanobacteria, along with 21 Proteobacteria and 5 Bacteroidetes recovered from their microbiome. Following the use of a state-of-the-art metagenomic pipeline, 12 of our new cyanobacterial genome assemblies are of high-quality, which indicates that even non-axenic cultures can yield complete genomes suitable for phylogenomics and comparative genomics. From a methodological point of view, we investigate the fate of SSU rRNA (16S) genes during metagenomic binning and observe that multi-copy rRNA operons are lost because of higher sequencing coverage and divergent tetranucleotide frequencies. Moreover, we devised a measure of genomic identity to compare metagenomic bins of different completeness, which allowed us to show that 2 Cyanobacteria-associated bacteria can be highly related in spite of considerable distance between collection points.

A constrained SSU-rRNA phylogeny reveals the unsequenced diversity of photosynthetic Cyanobacteria (Oxyphotobacteria)

“…However, for the purpose of this study aimed at pinpointing large patches of unsequenced diversity, a coarse-grained clustering was sufficient. The cluster file produced by CD-HIT was retrieved and 13 representative sequences initially selected by CD-HIT were replaced by genomic sequences from the same CD-HIT clusters, so as to be able to enforce the phylogenomic constraints from Cornet et al (2018b) [28] when inferring the rRNA tree. This process led to a dataset of 2302 OTUs.…”

“…The tree was inferred with RAxML v8.1.17 [36] under a GTR+Γ4 model with a 100x rapid bootstrap analysis, using a constraining tree of 75 different Oxyphotobacteria. The constraining tree was a phylogenomic tree based on a concatenation of 675 genes (>170,000 unambiguously aligned amino-acid positions; [28]). However, four organisms (Leptolyngbya sp.…”

“…The absence of rRNA genes is explained by the fact that the majority of these genomes (143 out of 193) were assembled with a metagenomic pipeline, according to NCBI metadata. This phenomenon was noticed in Cornet et al (2018a) [38] and is due to the frequent loss of rRNA sequences during metagenomic assembly [28]. Given that neither the isolation source nor the assembly pipeline were easy to determine, we decided not to consider these 193 genomes in our analysis by precautionary principle.…”

“…To alleviate the stochastic error plaguing single-gene analyses [27], we constrained our SSU rRNA (16S) phylogeny using a more reliable multi-gene phylogeny computed on a subset of the OTUs at hand. To this end, we used the phylogeny of Oxyphotobacteria developed in Cornet et al (2018b) [28], the largest ever-built in terms of conserved sites (>170,000 unambiguously aligned amino-acid positions from 675 genes). This strategy allowed us to easily compare the genome diversity available for phylogenomic studies to the more completely sampled rRNA diversity of Oxyphotobacteria.…”

A constrained SSU-rRNA phylogeny reveals the unsequenced diversity of photosynthetic Cyanobacteria (Oxyphotobacteria)

“…To alleviate the stochastic error plaguing single-gene analyses [ 27 ], we constrained our SSU rRNA (16S) phylogeny using a more reliable multi-gene phylogeny computed on a subset of the OTUs at hand. To this end, we used the phylogeny of Oxyphotobacteria developed in Cornet et al [ 28 ], the largest ever-built in terms of conserved sites (> 170,000 unambiguously aligned amino-acid positions from 675 genes). This strategy allowed us to easily compare the genome diversity available for phylogenomic studies to the more completely sampled rRNA diversity of Oxyphotobacteria.…”