Genome diversity of marine phages recovered from Mediterranean metagenomes: Size matters

Abstract

Marine viruses play a critical role not only in the global geochemical cycles but also in the biology and evolution of their hosts. Despite their importance, viral diversity remains underexplored mostly due to sampling and cultivation challenges. Direct sequencing approaches such as viromics has provided new insights into the marine viral world. As a complementary approach, we analysed 24 microbial metagenomes (>0.2 μm size range) obtained from six sites in the Mediterranean Sea that vary by depth, season and filter used to retrieve the fraction. Filter-size comparison showed a significant number of viral sequences that were retained on the larger-pore filters and were different from those found in the viral fraction from the same sample, indicating that some important viral information is missing using only assembly from viromes. Besides, we were able to describe 1,323 viral genomic fragments that were more than 10Kb in length, of which 36 represented complete viral genomes including some of them retrieved from a cross-assembly from different metagenomes. Host prediction based on sequence methods revealed new phage groups belonging to marine prokaryotes like SAR11, Cyanobacteria or SAR116. We also identified the first complete virophage from deep seawater and a new endemic clade of the recently discovered Marine group II Euryarchaeota virus. Furthermore, analysis of viral distribution using metagenomes and viromes indicated that most of the new phages were found exclusively in the Mediterranean Sea and some of them, mostly the ones recovered from deep metagenomes, do not recruit in any database probably indicating higher variability and endemicity in Mediterranean bathypelagic waters. Together these data provide the first detailed picture of genomic diversity, spatial and depth variations of viral communities within the Mediterranean Sea using metagenome assembly.

Fig 1. Taxonomic analyses based on metagenomic reads.

Taxonomic affiliation was obtained matching the reads against the NR database (>50% identity, >50% alignment). (A) Bacterial, archaeal, eukaryal and viral box plots were done using the 24 metagenomic data sets described in S1 Fig. (B) Upper panel shows the percentage of viral reads classified for each sample. Relative abundance of prokaryotic viral families is showed in more detail in the bottom panel. Using only reads assigned as a viral origin (C) UPGMA taxonomic tree and (D) PCoA were inferred with the cluster analysis option in MEGAN6 and a Bray-Curtis ecological distance matrix. Samples highlighted in blue and red correspond to samples obtained from the 5–20 μm and <0.22 μm filters, respectively. UP: Upper Photic, DCM: Deep Chlorophyll Maximum, LP: Lower Photic, DEEP: deep samples, MIX: mixed samples (in winter, without stratification).

Fig 2. Distribution of metagenomic viral contigs by length (Kb) and GC content (%).

Bacterial phages are open circles and colored according to the host assigned.

Fig 3. Reconstruction of Metagenome-Assembled Viral Genomes (MAVGs) using sequences belonging to Cluster-2.

A nucleotide comparison of several highly related contigs coming from different metagenomic samples is shown. Selected genes are labeled and colored uniformly.

Fig 4

(A) Phylogenetic tree of the cyanobacterial phage genomes obtained with 46 reference cyanophage genomes using a concatenation of 33 conserved proteins among them based on sequence similarities against the pVOG database. (B) Unrooted maximum likelihood phylogenetic trees of conserved proteins among magrovirus, halovirus and Med-OCT2015-90m-C1 (C) Fragment recruitment plot of Med-OCT2015-90m-C1 genome from Med-OCT2015-90m metagenome.

Fig 5. Recruitment of the Viral clusters (VCs) and singletons that recruits more than 10 RPKG (Reads per Kilobase of genome and Gigabase of metagenome) in at least two stations of the Tara Oceans viromes.

Left axis and dotted line indicates depth of sample. Upper panel shows the normalized value of total VCs by the number of stations belonging to SRF (surface) DCM (deep chlorophyll maximum) and DEEP (deep). Region abbreviations are as follows: ANTA, Antarctic Province; ARAB, Northwest Arabian Sea Upwelling Province; BENG, Benguela Current Coastal Province; CHIL, Chile-Peru Current Coastal Province; EAFR, Eastern Africa Coastal Province; FKLD, Southwest Atlantic Shelves Province; ISSG, Indian South Subtropical Gyre Province; MEDI, Mediterranean Sea Black Sea Province; MONS, Indian Monsoon Gyres Province; REDS, Red Sea; SATL, South Atlantic Gyral Province.