Vibrio cyclitrophicus phage encoding gene transfer agent fragment, representing a novel viral family

Highlights

• •A novel Vibrio cyclitrophicus phage vB_VviC_ZQ26 was isolated from the coastal water of Qingdao, China (120.327°E, 36.065°N).
• •Phage vB_VviC_ZQ26 represents a new family-level cluster, named Coheviridae, within the Caudoviricetes.
• •Two gene transfer agent proteins are encoded by phage vB_VviC_ZQ26, including baseplate attachment domain and central domain.
• •Phage vB_VviC_ZQ26 has a low relative abundance in the marine environment but one of the family members has high abundance.

Abstract

Vibrio is a prevalent bacterial genus in aquatic environments and exhibits diverse metabolic capabilities, playing a vital role in marine biogeochemical cycles. This study isolated a novel virus infecting Vibrio cyclitrophicus, vB_VviC_ZQ26, from coastal waters near Qingdao, China. The vB_VviC_ZQ26 comprises a linear double-stranded DNA genome with a length of 42,982 bp and a G + C content of 43.21 %, encoding 72 putative open reading frames (ORFs). Transmission electron microscope characterization indicates a siphoviral-morphology of vB_VviC_ZQ26. Nucleic-acids-wide analysis indicates a tetranucleotide frequency deviation for genomic segments encoding putative gene transfer agent protein (GTA) and coil-containing protein, implying divergent origins occurred in different parts of viral genomes. Phylogenetic and genome-content-based analysis suggest that vB_VviC_ZQ26 represents a novel vibriophage-specific family designated as Coheviridae. From the result of biogeographic analysis, Coheviridae is mainly colonized in the temperate and tropical epipelagic zones. This study describes a novel vibriophage infecting V. cyclitrophicus, shedding light on the evolutionary divergence of different parts of the viral genome and its ecological footprint in marine environments.

1. Introduction

Viruses are the most abundant ‘life forms’ in the ocean, one of the leading causes driving global biogeochemical cycles and manipulating the vast genetic diversity on the Earth (Suttle, 2005; Roux et al., 2016). Viruses regulate microbial communities and influence marine biogeochemical cycles by “virus shunt” and “virus shuttle” (Zimmerman et al., 2020). Prokaryotic phages are the most abundant and widely distributed in the ocean (Dion et al., 2020). As the significant predators of bacteria (Breitbart et al., 2018), viruses also mediate horizontal gene transfer (HGT) among bacteria, and impact marine carbon sequestration through the “biological pump” and “microbial carbon pump” (Weinbauer et al., 2007). The advent of transmission electron microscopy (TEM) provides the opportunity to directly observe the abundant phages (Williamson et al., 2012). Next-generation sequencing (NGS) development gradually unravels the cryptic diversity of viral genomes (Subirats et al., 2016). Benefiting from this technology, we will facilitate a better understanding of the origin and evolution of viruses and their effects on microbial communities.

V. cyclitrophicus is a species of the genus Vibrio, in the family Vibrionaceae of order Vibrionales under the class Gammaproteobacteria (de Souza Valente and Wan, 2021). Members of the genus Vibrio are Gram-negative bacteria that exhibit a curved or straight morphology. Most species are motile in liquid media due to the presence of a single polar flagellum (Li et al., 2019). This genus is ubiquitous in aquatic and marine environments. While many species are non-pathogenic, some can cause disease when present in high abundance (Kwan and Bolch, 2015). V. cyclitrophicus was first isolated from creosote-contaminated marine sediments and described as a novel species in 2001. It was found to utilize two- and three-ring polycyclic aromatic hydrocarbons (PAHs) as a carbon and energy source (Hedlund and Staley, 2001). Although PAH contamination can be toxic, these compounds also represent a rich source of nutrients for certain microorganisms (de la Pena et al., 2001). Many V. cyclitrophicus have been isolated from marine copepod and abalone feces (Vater et al., 2016; Nuttall et al., 2019). In addition, V. cyclitrophicus can infect Mytilus coruscus, leading to a decrease in diversity and increased mortality in haemolymph microbiome, indicating that it may have a significant impact on the aquaculture industry (Li et al., 2019). Siphoviral-morphologic viruses infecting Vibrio species have been extensively studied. However, our understanding of viruses that infect V. cyclitrophicus remains limited. To date, 49 V. cyclitrophicus isolates have been deposited in GenBank, but only four complete genome sequences of viruses infecting this species are available. These phages were collected in 2010 and have not been updated since then. Also, most phages are not specifically classified and have very limited genomic features.

Gene transfer agents proteins are particles that contain a random piece of the genome of the producing cell, and four unrelated GTAs have been identified: RcGTA in Alphaproteobacteria, Dd in Deltaproteobacteria, VSH-1 in the Spirochaete, and VTA in the archaeon (Lang et al., 2017). These types of genes are likely to derive from altruistic suicidal bacteria, providing enhanced nutrient utilization or alleviating nutrient deficiencies for other relatives or populations of bacteria (Kogay et al., 2020). All known GTAs possess tailed-phage structures and are presumably released into the environment through the lysis of the host cell (Lang et al., 2012). Analysis of phages carrying homologous GTA genes suggests that GTA is likely to be a prophage remnant (Huang et al., 2011). However, few GTAs have been reported to be present in Gammaproteobacteria.

In this study, we isolated and characterized a new virus, vB_VviC_ZQ26, infecting Vibiro cyclitrophicus. It represents a new siphoviral-morphologic viral cluster that could be considered a novel family designated as Coheviridae. The vB_VviC_ZQ26 shows a high level of genomic similarity with the other three siphoviral-morphologic viruses, which are not assigned to a defined taxon by the International Committee on Taxonomy of Viruses (ICTV). Furthermore, the genomes of Coheviridae were annotated with two GTAs fragment, which is poorly linked to the virus genome. This study fills the gap in understanding the isolation, cultivation, genomic, and evolution of Vibirophage, and provides new insight into viral-host interactions.

2. Materials and methods

2.1. Isolation of viruses and host

The vB_VviC_ZQ26 and its host, V. cyclitrophicus, were collected from the coastal water of Qingdao, China (120.327°E, 36.065°N). Seawater was diluted to 10⁻⁵ in a stepwise manner, and 200 μL of the diluted sample was inoculated onto 2216E solid medium using the plate coating method. The sample was then incubated at 28 °C and 120 rpm. V. cyclitrophicus ZQ26 was purified more than three times using a 2216E medium. The 16 s rRNA of the purified V. cyclitrophicus ZQ26 was amplified using PCR and sequenced. A BLASTn search revealed that the 16SrRNA similarity between V. cyclitrophicus ZQ26 and V. cyclitrophicus ECSMB14105 was 98.85 %. Phylogenetic analysis of the 16S rRNA sequence showed that V. cyclitrophicus ZQ26 and V. cyclitrophicus AB682659 formed a clade.

The double-layer plate method was used in viral isolation (Jamalludeen et al., 2007). Sampled seawater was filtered through 0.22 μm pore-size membranes as virion suspension. A suspension was made by mixing 200 mL of sewage sample filtrate and 200 mL of bacterial in the logarithmic growth phase in a cryopreservation tube and left to stand at 25 °C for 30 min. The mixture was then injected into 4.5 mL of semisolid medium melted at 50 °C and poured onto the surface of the solid medium. After culturing the agar plate in an incubator at 25 °C for 24 h, the formation of plaques was observed. If there was a plaque on the plate, it was picked out and placed in 1 mL of SM Buffer (100 mM NaCl, 8 mM MgSO₄·7H₂O, 50 mM Tris-Cl, pH = 7.5), and then filtered into a 0.22 mm PES Millipore filter. After the filtrate was gradually diluted, the above procedure was repeated. The infection step was repeated at least three times to ensure that the phage solution was completely purified.

2.2. Purification

The purified virion suspension was used to infect the host culture. The 400 ml lysis solution was treated with 10 % (w/v) PEG6000 at 4 °C overnight to obtain concentrated virions. The PEG6000 treated lysis solution was centrifuged (4500 g) for 30 min at 4 °C. The supernatant was then gently poured out without disturbing the deposit (Ul Haq et al., 2012). The precipitate was resuspended in 5 mL SM buffer to obtain concentrated virions, which were stored in SM buffer at 4 °C.

2.3. Transmission electron microscopy

Concentrated and purified virion suspensions were stained with 2 % (w/v) phosphotungstic acid (pH=7.5) and observed through 100 kV transmission electron microscopy (JEOLJEM-1200EX, Japan). The morphological measurements were made using ImageJ (v1.54). A grating scale with known line spacings was used as an internal calibration standard to ensure accurate size determination.

2.4. One-Step growth curve, thermal and pH stability assays

To conduct the one-step growth curve assay, the bacterial solution and purified phage solution were mixed with MOI 0.1 and incubated at 28 °C for 20 min. After centrifugation, the remaining unadsorbed phage was washed and precipitated three times with a 2216E liquid medium. Samples were taken every 5 min for the first 30 min and every 10 min for 30 to 90 min. The virus titer was determined by the double-layer agar plaque method, and the growth curve was plotted by counting the number of plaques. Three parallel experiments were performed to ensure the accuracy of the results. In the pH stability experiment, the SM buffers with pH 2∼12 were prepared, and the phage solution samples (initial titer ∼10⁹ PFU/mL) were diluted tenfold with corresponding buffers and then left to stand at room temperature for 2 h. The mixtures of the logarithmic growth period host solution and the treated phage sample solution were added to the semi-solid medium and then poured into the cured medium to form a bilayer plate. The phage was incubated overnight at a constant temperature to determine its activity. Three parallel samples were set up to ensure the accuracy of the experiment. To reveal the stability of the phage at different temperatures, the phage solution (initial titer ∼ 10⁹ PFU/ml) was respectively treated at −20 °C, 4 °C, 25 °C, 35 °C, 45 °C, 55 °C, 65 °C, 75 °C, 85 °C, and 100 °C for 2 h and the treated phage samples were subjected to gradient dilution. The phage activity was determined using the double-layer plate method for three parallel experiments, which plots the temperature-dependent phage growth curve by counting the number of spots on the plate.

2.5. Viral genome sequencing and annotation

The Virus DNA Kit (OMEGA) was utilized to perform DNA extraction of vB_VviC_ZQ26. Viral genomic next-generation sequencing (NGS) was powered by BioZeron Company. Raw data were assembled by the software ABySS (v1.3.7) (Simpson et al., 2009) to obtain initial genomic fragments, whose gaps were filled by GapClose (v1.12) software (Luo et al., 2012) with default parameters. Viral genome assembled completeness (occurrence of directed terminal repeat) is assessed by CheckV (v1.0.1) (Nayfach et al., 2021). Putative open reading frames (ORFs) were predicted through RAST (https://rast.nmpdr.org/rast.cgi) (Brettin et al., 2015) and PRODIGAL (v2.6.3) (Hyatt et al., 2010). ORFs functions were annotated by BLASTp (v2.12.0+) and hidden-Markov-Model-based search against nr database (https://blast.ncbi.nlm.nih.gov/), and Pfam-A (v35.0) (Madeira et al., 2022) database with 1e-5 as the E-value threshold of alignments.

DeepTMHMM (https://dtu.biolib.com/DeepTMHMM) was applied to predict alpha and beta transmembrane (TM) protein (Hallgren et al., 2022). Protein fold structure was analyzed by Phyre2 (http://www.sbg.bio.ic.ac.uk/phyre2/) under Batch processing (Kelley et al., 2015). The homologous protein of hypothetical protein, coil protein, and TMhelix-containing protein was detected against the UniRef90 database (https://www.uniprot.org/uniref/) using Position-Specific Iterated BLAST (PSI-BLAST), with parameters: 1e-5 as the e-value, 1000 as the maximum number of iterations, 30 % as the minimum query coverage. The HHpred (Zimmermann et al., 2018) was applied to detect the conserved protein domain in each ORF. The result of HHpred was manually corrected. The tRNA gene was predicted through tRNAscan-SE (http://lowelab.ucsc.edu/tRNAscan-SE/) (Chan and Lowe, 2019). The genome organization diagram was plotted by the ‘gggene’ package powered by R (v4.2.2). The G-C skew was calculated and visualized by Genskew (https://genskew.csb.univie.ac.at/webskew) (Lu and Salzberg, 2020).

2.6. Protein three-dimensional structure prediction and comparison

Further analysis of putative gene transfer agent protein 3D structure with the help of ColabFold (Mirdita et al., 2022), followed by comparison and visualization with RCSB PDB website (https://www.rcsb.org/). The amino acids of ORF18 and ORF20 were joined, the ColabFold predicted protein parameters were auto, and the best model with the highest pLDDT score was selected. The RcGTA baseplate protein 3D structure (6teh) as a template and alignment algorithm was jFATCAT (rigid).

2.7. vB_VviC_ZQ26-guided mining of viral genome

The viruses that have a close genome-wide relationship with vB_VviC_ZQ26 were mined from nr database and IMG/VR (v.3) dataset through BLASTp-based algorithm with the parameters: 1e-5 as the E-value, minimum 30 % protein sharing with vB_VviC_ZQ26 for target genomes. Separately, homologous viruses in the NR protein database were detected by a single gene phylogenetic tree. Three conserved proteins of vB_VviC_ZQ26 were performed by Diamond BLASTp against the nr database. The top 50 hits from the result were selected to calculate the maximum likelihood phylogenetic tree through IQ-Tree2 (Minh et al., 2020). The conditions for defining a homologous virus are that they are monophyletic group on the phylogenetic tree, so we used viruses existing on monophyletic clades for subsequent analysis.

2.8. Genome-content-based analysis of vB_VviC_ZQ26 in the background of Caudoviricetes

A total of 4457 Caudoviricetes reference genomes were downloaded from GenBank. In accordance with the result of phylogeny, all uncultivated genomes that have a close relationship with vB_VviC_ZQ26 from IMG/VR were combined. Genome-wide redundancy was removed by CD-HIT (v4.8.1) (Li and Godzik, 2006) with parameters: -c 0.99 -aL 0.9 -aS 0.9 -s 0.99. All-to-all -BLASTp (evalue < 1e-5) was used to align all proteins powered by Diamond. Proteins were clustered using the Markov clustering algorithm (MCL) to generate protein clusters (PCs) based on the E-values of all-to-all-BLASTp. Weights between different genomes were described by similarity scores calculated by vConTACT (v2.0) (Jang et al., 2019). Viral clusters (VCs) were assigned using ClusterONE (Nepusz et al., 2012). All genomes that have weights with Coheviridae were selected to perform network analysis, which was visualized by Gephi (v0.9.7).

2.9. Phylogenetic and comparative genomic analysis

A proteomic tree of Coheviridae was generated based on proteome-wide sequence similarity by ViPTree (https://www.genome.jp/viptree/) (Nishimura et al., 2017). The comparative genomic analysis of members of Coheviridae (vB_VviC_ZQ26, MG592659, MG592487, MG592566) was performed by tBLASTx. The comparative genome organization diagram was visualized by DiGAlign (v1.3).

A genome-wide phylogenetic tree was calculated by Virus Classification and Tree Building Online Resource (VICTOR) (Meier-Kolthoff and Göker, 2017). The tree contained four families (newly identified Coheviridae, Chaseviridae, Drexierviridae, Peduoviridae) and six floated genera associated with vB_VviC_ZQ26 in the genome-content-based network. The phylogenetic tree applied GBDP_D6 model was selected and visualized by iTol (Letunic and Bork, 2021) (v5.0). According to the proposal for the classification of ICTV, forming a monophyletic clade from three hallmark gene trees (major capsid protein, terminase large subunit protein, and portal protein) could be the criteria for establishing a novel viral family (2021.001BA.v1.abolish_Caudovirales). Hence, the phylogenetic tree of these genes from Coheviridae with outgroups was calculated by IQ-Tree2 (Minh et al., 2020) (v2.0.3).

The Average Nucleotide Identity (ANI) showed the homogenetic relationship at the nucleotide level and demonstrated the division of genus and species. The ANI of members of Coheviridae was calculated using the Virus Intergenomic Distance Calculator (VIRIDIC) (Moraru et al., 2020).

2.10. The ecological footprint of Coheviridae

The Global Ocean Viromes 2 (GOV2) (Sunagawa et al., 2020) dataset provides insight into the biogeographical distribution of Coheviridae in marine environments, which have been divided into five viral ecological zones, including Antarctic (ANT), Arctic (ARC), temperate and tropical epipelagic (EPI), bathypelagic (BATHY), and tropical mesopelagic (MES). The relative abundance of Coheviridae was calculated by CoverM (v0.3.0) with parameters: 95 % minimum reads identity and 75 % minimum reads alignment. The reads from the 154 GOV2 dataset were mapped to the Coheviridae genome by minimap2 (v2.17). The representative marine viruses were used as references, including the cyanophages (P-SSP7, P-SSM7, S-SSM7, and S-SM2 (Kang et al., 2013; Zhao et al., 2013)), pelagiphage (HTVC010P and HTVC011P), other two vibriophages and five phages are associated with Coheviridae. The relative abundance of these viruses was calculated by log₁₀(X + 1), and visualized by histogram and heatmap.

To assess the potential influence of the above 17 virus distribution patterns by different environmental factors, we performed a Canonical Correlation Analysis (CCA) by abundance. We removed GOV stations without environmental factor parameters and finally selected 128 stations which include eight environmental factors: Temperature, Ammonium, Nitrate at 5 m, Nitrite at 5 m, Iron at 5 m, Oxygen, Salinity, and Chlorophyll a. First, the relative abundances of the 17 viruses among the 128 stations were converted to TPM. The values of TPM were used as weights to calculate the relative environmental factor indices corresponding to each virus at each station. Their sum is the total environmental factor index, which can express the environmental factor preference of viruses in the distribution. CCA was calculated using vegan (2.6.8) based on a matrix consisting of 17 viruses and eight environmental factors and the relative abundance of 17 viruses at 128 stations. Global Monte-Carlo hypothesis testing was performed for all constrained axes by 999 permutations, and the significance of the results was assessed by p-value.

2.11. Tetranucleotides correlations analysis

Twenty-two segments were sliced from genomes of Coheviridae (window size: 10 kbp, step size: 2 kbp). The genomes were first linked together to maintain their integrity, with each 2 kbp of the genomes contained in a corresponding 10 kbp fragment (Teeling et al., 2004). For each segment, 256 possible combinations of tetranucleotide frequency ("AAAA" to "TTTT") were calculated and normalized using the z-scoring algorithm. Pearson's correlation coefficients (R-values) were calculated from an array of z-scores compared with the whole genome of the virus (Duhaime et al., 2011).

3. Results and discussion

3.1. Morphology and Characterization of vB_VviC_ZQ26

The V. cyclitrophicus phage vB_VviC_ZQ26 was isolated from the surface coastal seawater of Qingdao, China (120.327°E, 36.065°N), with V. cyclitrophicus ZQ26 as the host. TEM results show that the viral particles of vB_VviC_ZQ26 have a siphoviral morphology, with an icosahedral capsid measuring 34 nm ± 1.2 in cross-sectional diameter and a long, non-contractile tail (average length: 69 nm ± 1.5) (Fig. 1). The TEM image also displays that vB_VviC_ZQ26 possesses a distinct long tail structure, which is composed of a sheath and baseplate. From the experimental results of viral one-step growth, temperature and pH stability, the characteristics of vB_VviC_ZQ26 were revealed. The results of one-step growth curve show that the phage vB_VviC_ZQ26 has a 30 min long latent phase when the phage potency hardly changes and the progeny phages are being synthesized. During 30 to 180 min, the following rapid growth period is lysis phase, which is due to the release of the progeny phages into the environment by lysis. The phage potency reaches a stable level and enters a stable phase after 180 min. The titer of the phage vB_VviC_ZQ26 peaked at pH 7, and the phage activity decreased gradually with increasing and decreasing pH, indicating that there is an optimal pH for vB_VviC_ZQ26 phage to maximize its activity and this phage is not suitable for survival under extreme acids and bases conditions. For temperature stability, the suitable temperature of phage stability is 4 °C. The phage titer remains at a high level when the temperature is between −20 °C and 65 °C, but it drops sharply after 65 °C. The optimum temperature of vB_VviC_ZQ26 maintaining infectious activity is around 85 °C.

Fig. 1.

A Morphology of phage vB_VviC_ZQ26 lysate was stained with 2 % phosphotungstic acid. B One-step growth curve of phage vB_VviC_ZQ26. C The curve of pH stability. D thermal stability of phage vB_VviC_ZQ26.

3.2. The genomic characteristics of vB_VviC_ZQ26

According to sequencing and assembly results, vB_VviC_ZQ26 has a double-strand DNA genome of 42,982-bp with a G + C content of 43.21 %. No tRNA gene is encoded by the genome. The cumulative G + C skew analysis reflected the origin and terminus of phage genome replication (Fig. 2B), with the origin of replication located at position 25,100 nt, and the terminus of replication at the region 40,600 nt. Two breaking points could be identified at the above regions and demonstrate an asymmetric base composition, while the annotation results of genomes also suggested replication protein-encoding (25,704–29,007 nt) and provided additional proof for the origin of replication.

Fig. 2.

A Linear genome map of vB_VviC_ZQ26. Putative functional categories were defined according to annotation, and are represented by different colors. A tetra-nucleotide correlation was shown. The weaker correlations were circled by a red ellipse. B The phage genome sequence was analyzed using cumulative GC-skew. A window size and step of 100 bp were used to calculate the global minimum and maximum displayed in the cumulative graph. The GC-skew and cumulative GC-skew are shown as blue and red lines, respectively. The minimum and maximum of GC-skew can help predict the origin of replication and terminus location.

The genome of vB_VviC_ZQ26 contains 72 predicted ORFs, of which 32 coding regions (44 %) matched homologous proteins of unknown functions (Fig. 2A). The remaining 40 ORFs with specific functions were divided into five different modules: Lysis (one ORF), Packaging (three ORFs), Replication (11 ORFs), Structure (nine ORFs), Transcribing regulation (three ORFs). Of the total coding genes, 61 % (44 genes) were located on one strand, while the remaining 22 genes were located on the complementary strand. These genes in complementary strands were discontinuous and primarily associated with replication.

3.3. Genes related to the packaging and DNA replication genes

The DNA replication module contains restriction enzymes, nucleases, replication factors, and DNA-binding proteins which are encoded by ORFs 23–24, 26–27, 30, 34, 42–43, 45, and 70–71 (Fig. 2A). The module possesses a gene encoding the HNH endonuclease protein (ORFs 23 and 43), which is a common family of site-specific DNA endonucleases in viruses (Kala et al., 2014). They were characterized by the sequence motifs “His-Asn-His” and occur in group 1 and group 2 introns. HNH proteins also function as small nucleic acid-binding proteins, which are associated with DNA repair and replication (Dalgaard et al., 1997). Cyc-His-Cys-Cys (CHC2-type) zinc finger protein (ORF 30) was a relatively small protein motif of DNA primase (Pan and Wigley, 2000) which contains a finger-like bulge that tandem contact with the target molecule to act as a DNA-binding function. Its homologs had stable structures and wide functions including DNA recognition, RNA packaging, transcriptional activation, etc. (Laity et al., 2001).

The packaging module consists of genes encoding a portal protein, terminase large subunit (TerL), and teminase small subunit (TerS). Following DNA replication, the phage DNA is transported into an empty capsid through a pore formed by the portal protein (Lurz et al., 2001). TerL directly binds to the portal protein and possesses ATPase activity to transfer phage DNA into an empty capsid as well as cut the phage genome to initiate and terminate the packaging reaction. Meanwhile, TerS is essential for binding to the packaging initiation sites and activating TerL's activity (Baumann and Black, 2003).

3.4. Structural module in the genome of phage vB_VviC_ZQ26

The structural module of vB_VviC_ZQ26 comprises encoding to head-related proteins and tail-related proteins. Tetra correlations between the whole genome of vB_VviC_ZQ26 and the per 10 kb genome segments are drawn in Fig. 2A. The Pearson correlation coefficients of structural genes and the whole genome were all higher than 0.9, indicating a strong association between these structural genes and the entire genome of vB_VviC_ZQ26.

The tails of phages play an important role in host infection, they are critical for recognizing and attaching to the cell membrane as well as injecting DNA into the host cell (Nobrega et al., 2018). A total of three tail-related genes were identified and distributed throughout the genome (ORF 9–10, 15). ORF 9 was annotated as a tail completion protein from vibriophages 1.197.A._10 N.286.54.F2, which plays a fundamental role in the joining the head to the tail, which is the essential step in viral particle assembly (Auzat et al., 2014). The phage tail tube protein (ORF 10) represents a short all β-domain that forms a tube structure with a wide central channel for DNA ejection (Bebeacua et al., 2013). The prophage tail length tape measure protein (TMP) was detected to be homologous to the protein in vibriophages BUCT006 and is important for the determination and assembly of phage tail length, when this gene is mutated, the tail fibers also become shorter (Pedersen et al., 2000). The rich and diverse tail-related proteins of the vB_VviC_ZQ26 genome are associated with the formation of the tail structure and its interaction with the host (Taslem Mourosi et al., 2022).

3.5. Transcription-related genes of phage vB_VviC_ZQ26

Transcription is a vital process that regulates protein expression at the RNA level. In the transcription-related module, three genes have been detected in the vB_VviC_ZQ26 genome. One of them encodes the winged helix-turn-helix (W-HTH) DNA-binding domain protein (ORF 5), which originates from arsenical resistance operon repressor (ArsR) and similar metal-regulated homodimeric repressors in prokaryotes. This gene has been annotated in InterPro (IPR011991), and it is believed to be a bacterial transcription regulatory protein that contains two HTH domains that appear to dissociate from DNA when under metal ion conditions (Kar et al., 2001). Another gene in the transcription-related module of vB_VviC_ZQ26's genome is related to the restriction alleviation protein (ORF 11), which belongs to the Lar protein family. It can alleviate the restriction and enhance modification, thus contributing to the escape mechanisms involved in host-encoded restriction and modification (R-M) (King and Murray, 1995). C-5 cytosine-specific DNA methylases (ORF 59) are also a component of R-M systems and provide valuable functions for the manipulation of DNA (Cheng, 1995).

3.6. Unclassified and lysis genes

The genome encoded phage_lysozyme protein (ORF 2), a protease containing 144 amino acids, which assists in liberating mature phage particles from the cell wall by degrading the peptidoglycan. This protein has been identified in Pfam (PF00959) dataset, which includes lambda phage lysozyme and endolysin from E. coli (Weaver and Matthews, 1987). At the end of the lytic cycle, lysozyme breaks down the cell wall, releasing progeny phages (Zhang et al., 2022).

The unclassified genes module contains different genes that could not be assigned to a specific category. We have classified these genes by the proteins they encode into three sub-categories: coil containing proteins (ORFs 17, 31, 60–61, and 63), TMhelix containing proteins (ORFs 1, 47, 49, 51, 57, and 68), and domain-containing proteins without known functional classifications: putative gene transfer agent protein (GTAs, ORF18, 20). The result of the tetra correlations analysis showed that the Pearson correlation coefficients for the proteins without known functional classifications were all below 0.9, with a minimum value of 0.71, and demonstrated the lower adaptive ability of these genes to the phage genome.

3.7. Protein structure of gene transfer agent fragment and evolutionary analysis

The GTAs fragment (ORF18, ORF20) are part of the baseplate protein, HHpred annotation results show that these two putative GTAs genes encode tail proteins from the lambda phage, this may account for the observed low tetranucleotide correlation with the viral genome. Attachment of the GTA baseplate to the tail is enabled by the binding of the attachment domain and central domain (Bárdy et al., 2020). The core of the RcGTA baseplate contains hub protein and Megatron proteins, which are decorated with tail fibers. The structure of the ORF18 and ORF20 are very similar to the central domain in Megatron proteins and attachment domain in hub proteins of the RcGTA baseplate proteins (Fig. 7A). Attachment domains of vB_VviC_ZQ26 hub protein and the central domain of the Megatron protein of vB_VviC_ZQ26 can be superimposed onto the part of the RcGTA baseplate (Fig. 7B), which provides evidence of domain swapping. Although the proteins of RcGTA baseplate and vB_VviC_ZQ26 putative GTAs share less than 10 % sequence identity, the attachment domain alignment with RMSD of the corresponding atoms of 5.682 score (438 to 438 atoms) and RMDS score of the central domain is 4.247 (715 to 715 atoms).

Fig. 7.

A: Side-view of the RcGTA and complexes of ORF18, 20. RcGTA baseplate particles contain central domain (yellow) in megatron protein and an attachment domain (green) in hub protein. The iron-sulfur cluster (red) is indicated, wrapped by four conserved cysteines. ORF18, 20 complexes are also annotated with the same colors. B: Domain swapping among baseplate particle proteins of RcGTA and ORF18, 20. ORF18 and ORF20 can be superimposed onto the central domain and attachment domain of RcGTA baseplate particle shown in gray (PDB 6TEH).

Evolutionary analysis of ORF18 and ORF20 revealed distinct distribution patterns of these two proteins across taxonomic lineages (Supplementary Fig. 3). ORF18 is prevalent in Gammaproteobacteria, but less so in viruses. In contrast, ORF20 is more broadly distributed, being found not only in Gammaproteobacteria but also in Alphaproteobacteria and Betaproteobacteria. Monophyletic clade of different taxonomic lineages suggests that ORF18 is evolutionarily conserved and may have adapted to hosts or viruses over time, retaining a single evolutionary feature. Evidence of potential horizontal gene transfer was also detected, suggesting the possibility of such transfer between Gammaproteobacteria and viruses. Although there is no cooperation between the phylogenetic trees of ORF18 and ORF20, some branch lineages do exhibit congruence. In particular, the branch lineages of the Coheviridae are cooperative, indicating stable inheritance of ORF18 and ORF20 among this type of vibriophages. Furthermore, viruses infecting both Alphaproteobacteria and Gammaproteobacteria were present in a clade, indicating that phages may have mediated the exchange of GTAs fragments. Notably, ORF20 forms a unique monophyletic group in the phylogenetic tree, which is characterized by the presence of Coheviridae family and vibriophage BUCT006.

3.8. Phage vB_VviC_ZQ26 represents a novel viral family

Single-gene and whole-genome phylogenetic trees were used to infer the phylogenetic and genomic relationship between vB_VviC_ZQ26 and other phages. The phylogenetic trees of portal protein, terminase large subunit protein (TerL), and major capsid protein (MCP) were calculated using IQ-Tree2, respectively. In the portal protein tree, vB_VviC_ZQ26 and seven other vibriophages formed a monophyletic clade (Fig. 3A). Similarly, the TerL protein tree shows that eight vibriophages were clustered with vB_VviC_ZQ26. In the MCP tree, four vibriophages were clustered with vB_VviC_ZQ26. The phylogenetic monophyly of all tree viral hallmark genes suggested a novel viral family-level cluster. These viruses included vibriophage 1.291.O._10 N.286.55.F6 (MG592659.1), vibriophage 1.197.A._10 N.286.54.F2 (MG592566.1), and vibriophage 1.113.A._10 N.286.51.E7 (MG592487.1).

Fig. 3.

Unrooted phylogenetic trees of vB_VviC_ZQ26 were created using the maximum likelihood method based on amino acid sequences of the phage's major capsid protein, terminase large subunit protein, and portal protein. Different colors represent different classifications: gray for phages and other colors for different bacterial phyla. B A gene content-based viral network was created for the vibriophage vB_VviC_ZQ26. Viruses at different family levels are shown in different colors, with black representing uncultured viruses. Dotted lines enclose different virus groups.

The cluster analysis was used to further identify the taxonomic status. All viruses linked to Coheviridae from Genbank (order Caudoviricetes) and IMGVR (v3) dataset were selected and presented in the genome-content-based network (Fig. 3B). The three vibriophages and vB_VviC_ZQ26 were clustered as VC_12, which exhibited no direct linkages with the three known members of the family (Chaseviridae, Drexieviridae, and Peduoviridae) under the order Caudoviricetes. Some of the other metagenome contigs were assigned to VC_0 and were not eligible for classification within the same family as vB_VviC_ZQ26. Although DTR_578,389 and VC_12 were closely related, they were not assigned as VC_12.

The whole-genome phylogenetic trees were constructed to confirm that VC_12 provided evidence of a novel viral family (Fig. 4A). This result shows that the Coheviridae occupied a monophyletic clade. Within this clade, the four viruses had similar phylogenetic distances, and genetic distances from the last common ancestors (LCA) were 0.16, 0.2, 0.18, and 0.19, respectively.

Fig. 4.

Four phages were proven to be a family through phylogenetic analysis and average nucleotide identity. Phylogenetic analysis was performed with other related viruses in the network using genome-wide sequence similarity values computed by VICTOR. Mark the Coheviridae family with a red dashed line. B A heat map was created based on ANI values calculated using VIRIDIC.

Average nucleotide identity (ANI) analysis was performed to identify the genus and species boundaries (Jain et al., 2018) within Coheviridae. The highest ANI is 61.1 % (between vB_VviC_ZQ26 and vibriophage 1.291.O._10 N.286.55.F6); the lowest ANI (between vB_VviC_ZQ26 and vibriophage 1.197.A._10 N.286.54.F2) is 52 % (Fig. 4B). The result further demonstrates that vB_VviC_ZQ26 and the three other viruses are closely related and belong to a new family.

3.9. Comparative genomic analysis between Coheviridae

Comparative genomic analysis is a valuable tool for exploring the core genes of viruses and defining virus taxa (Chan et al., 2014). The candidate family members were selected for the core genes clustering and comparative genomic analysis. All viral clusters associated with the candidate family (VC_12) were used for protein clustering by The Markov Cluster Algorithm (MCL). In Fig. 5A, the results revealed distinct core gene modules among the different viral clusters, and the core region of VC_12 was marked with a red dotted box. Comparative genomic analysis results demonstrated that the candidate family had similar gene distribution patterns with six gene modules that were collinear across the genomes (Fig. 5B). In the packaging module, the genes encoding TerS, TerL, and portal protein were found to be collinear on the genome. The locations of core genes (Fig. 5B underlined in yellow) had a similar distribution pattern across the genome, with consecutive core genes appearing on coil protein, hypothetical protein, and TMhelix-containing protein. Some tail-associate proteins annotated by BLASTp (tail-completion protein, Phage tail tube protein, and Prophage tail length tape measure protein) are also encoded by the core genes of the Coheviridae family. These findings further support the significance of the long tails of these phages.

Fig. 5.

Comparative genomic analysis on vB_VviC_ZQ26 to confirm core genes of family Coheviridae and gene covariation of members within this family. A All viruses related to vB_VviC_ZQ26 were analyzed for core genes and displayed in heatmaps. Viruses with different VCs were clustered for proteins using the MCL algorithm and are shown in different colors. Different VCs contain different protein families, with red dashed boxes indicating the protein clusters of the Coheviridae family. B A gene covariance analysis was performed for the family Coheviridae. Different protein types are shown in different colors, with yellow underlines representing family core genes.

3.10. Ecological distribution of Coheviridae in the ocean

The biogeographic distribution of Coheviridae and other viruses was characterized by 154 viral metagenomes from Global Ocean Viromes (GOV2.0) database, which divided the global marine habitats into five major regions: Antarctic (ANT), Arctic (ARC), Temperate and tropical epipelagic (EPI), Temperate and tropical mesopelagic (MES) and Bathypelagic (BATHY) (Gregory et al., 2019). In addition, two vibriophages, two Synechococcus phages, two Prochlorococcus phages, two Pelagibacter phages and six phages with similar distances on the phylogenetic tree were selected as references. The results confirmed that Pelagibacter phage HTVC011P/HTVC010P, Prochlorococcus phage P-SSP7/P-SSM7, and Synechococcus phage S-SSM7/S-SM2 were highly abundant in all five major regions (Kang et al., 2013; Zhao et al., 2013), and other phages showed no abundance in the BATHY. Coheviridae were more abundant in EPI and MES (Except for vibriophage 1.291.0.10 N.286.55.F6, because its abundance in MES is zero) than those similar phages in the global ocean and shared a similar abundance distribution pattern except vibriophage 1.113.A._10 N.286.51.E7 (Fig. 6A). The abundance of vibriophage 1.113.A._10 N.286.51.E7 exhibited the highest abundance of all the Coheviridae in the MES region (5.76), temperate and tropical epipelagic zone (4.51), and also had some abundance in the ANT (3.77) as well as ARC (4.65). These results suggested that V. cyclitrophicus phages were widely distributed in surface seawater and could modulate microbial abundance and community structure by infecting microorganisms, highlighting their potential impact on marine population health (Li et al., 2019).

Fig. 6.

A: The relative abundance of vibriophage vB_VviC_ZQ26 and three other members of the Coheviridae family was calculated in 154 viromes from the Global Ocean Viromes 2 (GOV 2.0) data set. The abundance was expressed as TPM (transcripts per million mapped reads) and calculated using CoverM. The values were normalized by the number of databases in each viral ecological zone (VEZ) and transformed by log10 (X + 1). Representative Pelagibacter phages, cyanophages, Puniceispirillum phage HMO-2011, two phages similar to vB_VviC_ZQ26, and other vibriophages were added as references. The five marine VEZs are Arctic (ARC), Antarctic (ANT), temperate and tropical epipelagic (EPI), temperate and tropical mesopelagic (MES), and bathypelagic (BATHY). B The CCA (canonical correlation analysis) of Coheviridae and other viruses, indicated the correlation between viruses and different environmental factors. The arrows represent the eight environmental factors (Salinity, Temperature, Iron, Ammonium, Nitrite, Nitrate, Oxygen, Chlorophyll a.) and the length of the arrow represents the intensity of the effect of the environmental factors on the virus type, with the longer the length the greater the effect.

CCA analyses revealed the distributional characteristics of viruses and the correlation between environmental factors and their abundance (Fig. 6B). The distributional characteristics of vibriophage 1.113.A._10 N.286.51.E7 differed from those of other members of the Coheviridae family, as it clustered with several viruses known for their high abundance in marine environments. We hypothesize that the presence of GTA segments in vibriophage 1.113.A._10 N.286.51.E7 may have facilitated the interspersion of bacterial genomes, potentially contributing to a broad distribution of this bacterium in oceanic habitats.

4. Conclusion

Vibrio serves as an excellent model system due to its ability to thrive across a wide range of environmental temperatures and salinities, both of which are critical environmental factors influencing microbial community composition (Lozupone and Knight, 2007; Herlemann et al., 2011). Of particular interest is vB_VviC_ZQ26, a phage that infects V. cyclitrophicus and has been meticulously characterized in terms of morphology, genomic traits, and phylogenetic features. vB_VviC_ZQ26 is the isolated phage that infects V. cyclitrophicus, which was described in detail in terms of morphology, genomic characteristics and phylogenetic features. It has a large number of tail proteins in its genome and a unique low adaptation module compartment. We also identified the presence of GTA fragments in vibriophages, opening up a new sight for investigating the swap of GTA between phages and bacterium, as well as the distribution of GTA. Moreover, our analysis revealed a novel family, Coheviridae, in the evolutionary lineage, which had no uncultured virus sequence. The study provided a detailed account of the diversity, genomic evolution, abundance and distribution details of phages infecting V. cyclitrophicus. Considering the wide distribution and significant roles of Vibrio in marine, the isolation, diversity and ecological roles of vibriophage could be studied more specifically. Furthermore, our findings demonstrate the potential of the combination of metagenomics with phage isolation to improve our knowledge of the diversity and functions of marine viruses.

Funding

This work was supported by the Laoshan Laboratory (No. LSKJ202203201), Natural Science Foundation of China (No. 42188102, 42120104006, 41976117, and 42176111), and the Fundamental Research Funds for the Central Universities (202172002, 201812002, 202072001 and Andrew McMinn).

Conflict of interest

The authors declare that they have no conflict of interest regarding this study.

Ethical approval

This article does not contain any studies with animals or human participants performed by any of the authors.

Authors’ contributions

YL (Yantao Liang) and MW planned, supervised, and coordinated the study and revised the manuscript. YX, KM and XZ performed phage isolation, main experiments, and bioinformatic analyses, annotated the genome, and drafted the manuscript. KZ, TW, HZ and YD take TEM figures of phage vB_VviC_ZQ26. ZW, YL and HS critically evaluated the manuscript. AM helped to modify the language of the manuscript.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

• The genome sequence of phage vB_VviC_ZQ26 has been deposited in GenBank under accession number OP918135.