Biological and Genomic Characterization of a Novel Jumbo Bacteriophage, vB_VhaM_pir03 with Broad Host Lytic Activity against Vibrio harveyi

Vibrio harveyi is a Gram-negative marine bacterium that causes major disease outbreaks and economic losses in aquaculture. Phage therapy has been considered as a potential alternative to antibiotics however, candidate bacteriophages require comprehensive characterization for a safe and practical phage therapy. In this work, a lytic novel jumbo bacteriophage, vB_VhaM_pir03 belonging to the Myoviridae family was isolated and characterized against V. harveyi type strain DSM19623. It had broad host lytic activity against 31 antibiotic-resistant strains of V. harveyi, V. alginolyticus, V. campbellii and V. owensii. Adsorption time of vB_VhaM_pir03 was determined at 6 min while the latent-phase was at 40 min and burst-size at 75 pfu/mL. vB_VhaM_pir03 was able to lyse several host strains at multiplicity-of-infections (MOI) 0.1 to 10. The genome of vB_VhaM_pir03 consists of 286,284 base pairs with 334 predicted open reading frames (ORFs). No virulence, antibiotic resistance, integrase encoding genes and transducing potential were detected. Phylogenetic and phylogenomic analysis showed that vB_VhaM_pir03 is a novel bacteriophage displaying the highest similarity to another jumbo phage, vB_BONAISHI infecting Vibrio coralliilyticus. Experimental phage therapy trial using brine shrimp, Artemia salina infected with V. harveyi demonstrated that vB_VhaM_pir03 was able to significantly reduce mortality 24 h post infection when administered at MOI 0.1 which suggests that it can be an excellent candidate for phage therapy.


Introduction
The financial losses in aquaculture due to outbreaks of bacterial diseases are estimated to be in the range of billion US dollars globally. Disease outbreaks are among the most important threats for the economic sustainability of the aquaculture sector [1,2]. An important bacterial pathogen in aquaculture is Vibrio harveyi, which is a halophilic Gram-negative bacterium causing vibriosis disease in marine finfish, crustacean and molluscan species [3,4]. Vibrio harveyi is ubiquitous in the aquatic environment and can survive without a host. It is an opportunistic pathogen that will induce disease when the water temperature is optimal for its growth and at the same time its hosts are stressed [5]. Vibrio harveyi has also been increasingly reported in the Mediterranean aquaculture [6][7][8].
Intensification of aquaculture has been regularly considered as a major contributing factor to the outbreaks. In addition, the rising of the seawater temperature globally and climate change have also been associated with increasing Vibrio spp. detection in the environment [9][10][11][12]. Therefore, there is a high risk of more vibriosis outbreaks in the future. Antibiotics such as tetracyclines, fluoroquinolones and beta-lactamases have been extensively used and, in some instances, misused (or abused) in aquaculture as prophylactic or therapeutic means to control vibriosis [13]. As a consequence, resistant bacterial subpopulations develop rapidly in aquaculture through the exposure to subclinical dosages of antibiotic residues [14][15][16][17][18][19]. It is essential to reduce drastically the use of antibiotics in aquaculture or to be considered as a last resort option. To control bacterial diseases in aquaculture with reduced reliance on antibiotics, a working alternative is urgently needed. Bacteriophages or phages are ubiquitous viruses that exclusively infect bacteria. Recently, phage therapy has been revisited as a potential alternative to control bacterial diseases in aquaculture [20][21][22]. Phages are also the most abundant and diverse biological entities on earth therefore, finding phages that infect a specific strain of bacteria is relatively easy. Recent experimental phage therapy studies have also demonstrated positive results in controlling important bacterial fish pathogens such as Flavobacterium psychrophilum and Vibrio spp. [23][24][25][26][27]. However, phage therapy requires comprehensive knowledge of the applied bacteriophages therefore a full characterization of their biological and genomic attributes that includes their infectivity, lifestyle, stability, and possible virulence and antibiotic resistance encoding genes [28]. This study aimed to the isolation and characterization of a lytic phage against V. harveyi that could be used for phage therapy in aquaculture.

Isolation and Morphology of vB_VhaM_pir03
vB_VhaM_pir03 was isolated from an environmental sample collected from the Port of Piraeus, Athens, Greece against Vibrio harveyi type strain DSM19623. A single plaque of vB_VhaM_pir03 was carefully isolated and purified through six times propagation. Throughout the propagation steps, vB_VhaM_pir03 showed a consistent plaque morphology. In the double layer agar plating assay, vB_VhaM_pir03 produced a pinhole-type plaque formation with a diameter of 0.27 ± 0.05 mm. We found that a comparison between the use of LB agar and diluted (LB/2) agar as the bottom layer for plating showed that a higher bacteriological nutrient composition reduced the visibility and plaque size of vB_VhaM_pir03 but not the actual count (data not shown). Transmission electron microscopy (TEM) showed that vB_VhaM_pir03 has a short neck, contractile tail and an icosahedral capsid ( Figure 1) which indicated that its morphology is close to the phages of the Myoviridae family. Structural measurements of vB_VhaM_pir03 revealed relatively large virion dimensions.

Host Range and Efficiency of Plating (EOP) of vB_VhaM_pir03 against Multiple Antibiotic Resistant Strains
In the host range test (Table 1), vB_VhaM_pir03 was able to infect 31 out of 51 strains used. vB_VhaM_pir03 infected 21 of the 23 strains of V. harveyi, three of the seven strains of V. alginolyticus, the single strain of V. campbellii, both strains of V. owensii and four of the ten other unidentified presumptive Vibrio spp. There were no susceptible strains from V. parahaemolyticus, V. anguillarum and V. splendidus. EOP of vB_VhaM_pir03 was high for eight strains of V. harveyi, two strains of V. alginolyticus and two strains of unidentified Vibrio spp. and moderate for nine strains. For the antibiotic susceptibility tests (Table 1), all the phage susceptible strains were determined to be completely resistant against ampicillin. Eight strains were resistant against oxytetracycline, nine strains were resistant against oxalinic acid, nine strains were resistant against florfenicol, fifteen were resistant against sulfamethoxazole/trimethoprim and five strains were resistant against flumequine. Overall, vB_VhaM_pir03 was shown to be very effective against five multiple antibiotic-resistant strains and moderately effective against six multiple antibiotic-resistant strains.  Exposure to different temperatures ( Figure 2a) showed that vB_VhaM_pir03 was stable up to 35 • C. Significant reduction (one-way ANOVA, p < 0.05) of its titer was observed between 40 to 45 • C while complete inactivation of vB_VhaM_pir03 was observed from 50 • C and above. When exposed to 0.001% benzalkonium chloride, BKC (Figure 2b), vB_VhaM_pir03 titer was significantly reduced (one-way ANOVA, p < 0.05) compared to the control. However, vB_VhaM_pir03 was complete inactivated when exposed to other organic solvents. Effect of organic solvents to the stability of vB_VhaM_pir03. Incubation with LB was used as control. SE bars were shown for the mean of n = 3. Statistical significance indicated by different superscript letters was determined at p < 0.05.

Adsorption Time and One-Step Growth of vB_VhaM_pir03
In the adsorption time assay (Figure 3a), it was estimated that the time required for 90% of the vB_VhaM_pir03 to irreversibly bind to bacterial host was 6 min. One step growth assay ( Figure 3b) revealed that vB_VhaM_pir03 has a latent phase (period of between irreversible binding of vB_VhaM_pir03 to host cell until phage bursts) of 40 min. The rise phase (start of phage release from infected host until no more phages were released from its infected host) was estimated between 40 to 70 min. The plateau phase (period that indicated no more phages were released from the infected host cells) was reached at 70 min. In this assay, the burst size (number of new infective particles produced per each infected bacterial cell) of vB_VhaM_pir03 was 75 virions.

In Vitro Cell Lysis
In vitro lysis assay with DSM19623 ( Figure 4) showed that vB_VhaM_pir03 was able to lyse the host bacterial population from MOI 0.1 to 10 after 18 h of incubation. Bacterial population infected at MOI 10 initially showed the lowest growth after 4 h of incubation but after 6 h the infected bacterial populations showed similar growth curves until 18 h of incubation to the other two MOIs used. Overall, vB_VhaM_pir03 still managed to decrease the bacterial population of DSM19623 by an approximate 40% at all MOIs compared to the uninfected population after 18 h of incubation. For in vitro lysis of vB_VhaM_pir03 with other strains, infected bacterial populations of Vh5, SerKid SA1.1, SA1.2 and SA4.1 (Supplementary Figure S1) showed similar growth patterns with DSM19623. The bacterial populations infected at MOI 10 initially showed the lowest growth however, all the infected bacterial populations eventually showed a similar growth curve pattern. For strains SNGR, VhP1 Liv, SA5.

Whole Genome Sequencing and Assembly
The sequenced genome of vB_VhaM_pir03 produced 41,500,540 clean reads with an average read length of 150 bp and 96.13% correct base calls. The GC content (%) was 43.6%. The per base calls scores produced good per sequence quality scores with a median of 36 for 150 bp reads. The per base sequence content and per sequence GC content showed that there was no bias in the proportion of each base position calls for four normal DNA bases or contamination during library preparation for vB_VhaM_pir03 sequencing. Finally, the per base N content result showed that no N substitutions were made which indicated that the sequencer had sufficient confidence to make base call. The genome of vB_VhaM_pir03 was assembled into a single contig with a minimum genome coverage of 5×. The total genome length of vB_VhaM_pir03 was 286,284 bp. A total of 99.91% of the raw reads were mapped back to the assembled genome resulting to an average coverage depth of 21,669×. In addition, the vB_VhaM_pir03 genome does not have any termini and was found to be terminally redundant and circularly permuted.

Genomic Features of vB_VhaM_pir03
The genome size of vB_VhaM_pir03 (286,284 bp) indicated that it is a jumbo phage (phages with total genome length more than 200,000 bp). The gene-coding potential of the global genome is 96

Genomic Arrangement and Functional Annotations of vB_VhaM_pir03
Generally, the genome of vB_VhaM_pir03 did not have any modular arrangement ( Figure 5). However, genes encoding for head and tail proteins were arranged in subclusters while genes encoding for DNA replication and nucleotide metabolism proteins were scattered. Other genes that were functionally annotated are as in Table 2.           (ORFs 55, 139, 141, 142, 144, 150, 152, 165 and 166). The large terminase subunit involved for DNA packaging for tailed phages was identified at ORF 57. Interestingly, Proline-Alanine-Alanine-aRginine (PAAR) repeat proteins (ORF 10 and 11), a sharp conical structure for penetration of host cells [30] were also identified adjacent to the tail, baseplate protein and membrane puncturing device proteins. In addition, a prohead core protein protease which functions to facilitate the transition of a prohead or procapsid to a mature capsid [31]

Miscellaneous Proteins
Several enzymes for lysis of host bacteria were also identified as glycoside hydrolase (ORF 28, 168, 182) which functions to degrade the host bacterial cell wall prior to phage burst. Additionally, two ORFs with no previous phage-associated descriptions were identified in vB_VhaM_pir03 genome which are ORF 200 with a structural homolog to palindromic amphipathic repeat coding elements (PARCEL) protein and ORF 64 with a structural homologs to Ro60-related proteins.

Genomic Synteny of vB_VhaM_pir03 with Other Similar Phages
Following whole genome alignment with the most similar phage genomes obtained from the NCBI nr database (Figure 6), vB_VhaM_pir03 was shown to have the highest degree of genomic synteny with vB_BONAISHI. Both phages also shared six collinear blocks with similar length. The longest shared collinear block had a sequence length almost 50,000 bp. Despite similar genomic arrangements, the shared collinear blocks showed very low DNA sequence similarities between them. vB_VhaM_pir03 also shared a single collinear block (<3000 bp) with Salmonella phage SKML 39. Alignment with another four similar jumbo phages; vB_VmeM-Yong_MS32; vB_VmeM-Yong_XC31; vB_VmeM-Yong_XC32 and vB_VmeM-Yong_MS31; showed shared collinear blocks of different length with no synteny and very low sequence similarities.

Phylogenetic Analysis
Wide genome proteomic tree analysis (Figure 7) showed that vB_VhaM_pir03 belong to the Myoviridae taxonomic family however, it was observed that the position of vB_VhaM_pir03 was in a subcluster within the Siphoviridae family. In addition, vB_VhaM_pir03 was also determined to infect a host from the Gammaproteobacteria class which includes Vibrionaceae family. Figure 7. Determination of taxa and host group for vB_VhaM_pir03 by a proteomic tree using VIPTree. vB_VhaM_pir03 was determined to belong to Myoviridae family and infect Gammaproteobacteria group (red star and line). vB_VhaM_pir03 (asterisk) proteome was compared with 2688 dsDNA phages proteomes. The branch length scale was calculated as log values. The inner and outer ring indicate the taxonomic virus family and host group, respectively. Phylogeny using large terminase subunits of jumbo phages (Figure 8) showed that vB_VhaM_pir03 have a recent common ancestor with vB_VmeM-Yong_MS32, vB_VmeM-Yong_XC31, vB_VmeM-Yong_XC32 and vB_VmeM-Yong_MS31 although with the bootstrap value (0.536) for this inference is below the acceptable threshold of 70% for bootstrapping [34]. However, vB_VhaM_pir03 has a high bootstrap support (0.996) with vB_BONAISHI. In addition, the branch length indicated that both phages share similar number of amino acid substitutions in their large terminase subunit since diverging from their common ancestor. The large terminase subunits of jumbo phages were downloaded from NCBI database and aligned using MUSCLE and a maximum likelihood (bootstrap = 1000) phylogenetic tree was constructed using MEGA X. The tree was visualized using Interactive Tree of Life (ITOL). The bootstrap support value was denoted in each branch.

In Vivo Phage Therapy Trial with Artemia nauplii
At 24 h post infection, the survival of the Vibrio harveyi-infected Artemia was approximately 30% which was lower than the phage-treated groups (except for MOI 10) and comparable to the untreated control group (Figure 9). At 48 h post infection, survival of the phage-treated group was higher than the Vibrio harveyi-infected control group although this difference was not statistically significant. The delayed treatment resulted in similar survival to the Vibrio harveyi-infected control group both at 24 and 48 h post infection. Measurement of the presumptive Vibrio load at 24 and 48 h post infection showed no significant differences (two-way ANOVA, p < 0.05) between all treatments (data not shown). No colony forming units were observed in the control group at both times of measurement.

Discussion
Phage therapy is a very promising alternative to antibiotics. While scientific publications about isolation of phages have increased in the last decade, there is only a small group of phages that have been applied for commercial use [35] and only one for aquaculture which is CUSTUS ® YRS against Yersinia ruckeri [36]. One of the main reasons for the lack of commercialized phages is insufficient characterization which is a prerequisite for phage therapy [37]. These characterizations are necessary to identify and reduce risks associated with phage therapy which in turn, will support progress towards regulatory approval [28,38]. Therefore, these considerations were the highest priorities when performing characterization of bacteriophages in this study.
In the present study, we have isolated and characterized a novel jumbo bacteriophage, vB_VhaM_pir03 with broad host lytic activity against Vibrio harveyi type strain DSM19623 and analyzed its therapeutic potential for aquaculture. Transmission electron microscopy revealed that vB_VhaM_pir03 is related to the Myoviridae family based on the presence of an icosahedral head and long contractile tail [39]. In addition, vB_VhaM_pir03 also had relatively large structural dimensions compared to other Myoviruses infecting Vibrio spp. [40,41]. The main factor for large structural dimensions is still undetermined but it has been posited that a ratcheting mechanism is present for large phages especially jumbo phages for the accommodation of large genome sizes and potential acquisition of more genes [42]. In relation to plaque formation, it has also been suggested previously that large structural dimensions of phages have contributed to small plaque formations as observed for vB_VhaM_pir03 due to reduced diffusion capacity of the large virion particles [29].
Several factors affect the reproductivity and stability of phages. In our study, the significant reduction of vB_VhaM_pir03 titers was first observed at 40 • C which was lower than that of previously described Vibrio phages [24,25,43,44]. Poullain et al. [45] have shown that heat treatment caused damage to tailed phages such as detachment of head and tail, empty capsids, and aggregation of tails. However, since the vB_VhaM_pir03 thermal tolerance was lower than previous reports, we suggest that this phage was more sensitive due to its large genome. Previous reports [46,47] have shown that phages with large genomes have high internal capsid pressure due to dense genome packaging. These phages were demonstrated to eject their genomic material at 37 • C by the internal capsid pressures. Nevertheless, the thermal stability of vB_VhaM_pir03 do not hinder its potential direct application in aquaculture since rearing temperatures of aquatic organisms do not reach 37 • C. vB_VhaM_pir03 was also found to be completely inactivated by chloroform in the present study therefore, chloroform was not used in all assays conducted subsequently. Tailed dsDNA phages do not contain lipid membrane [48] and are not sensitive to chloroform. A recent study [49] however has reported that several tailed dsDNA phages including members of Myoviridae and Podoviridae families showed reduction in phage titer after exposure to chloroform. In addition, vB_VhaM_pir03 was also completely inactivated after exposure to organic solvents in this study except for benzalkonium chloride, BKC. This suggests that vB_VhaM_pir03 applications can be controlled for contaminations and unintentional transfers such as observed in inactivation of Lactobacillus phages during milk productions [50].
vB_VhaM_pir03 showed a rapid adsorption time to its host compared to previously reported Vibrio and jumbo phages [24,25,[51][52][53]. This indicated the efficiency of vB_VhaM_pir03 to locate and irreversibly bind to the host receptors under controlled conditions. Although the main factor determining adsorption time is the rate of phage-host encounters [54], application of rigorous aeration in an aquaculture system may increase the rate of phage-host encounters by reduction of phage and host sedimentations [55]. In one-step growth assay, vB_VhaM_pir03 displayed a short latent period and high burst size similar to previously reported jumbo phages [44,56,57]. A short latent period would result in low burst sizes of phages due to limited time for viral reproduction cycles [54]. Nonetheless, vB_VhaM_pir03 was shown in this study to have an efficient viral replication mechanism which allowed high virion productions in short latent periods. As a result, a minimal concentration of vB_VhaM_pir03 can be considered in therapeutic applications in aquaculture due to its rapid adsorption time and high multiplication rate. Finally, high multiplication rate of phages has been suggested as an evolutionary trade off with low phage survivability under stressful conditions [46] which may provide an insight to the low thermal stability of vB_VhaM_pir03.
In small phages (<200,000 bp), the arrangement of core genes that encode for head, tail, DNA replication and nucleotide metabolism proteins are conserved in a modular order for maintenance of functions throughout its replication cycles [58,59] however, we found the arrangement of the core genes in the vB_VhaM_pir03 genome was generally scattered and formed subclusters as previously described by Yuan and Gao [60]. Interestingly, we also found genes in vB_VhaM_pir03 that encode multisubunit RNA polymerases (RNAPs). RNAPs are found in jumbo phages and function to trigger early DNA transcription during infection without requiring the host's DNA machinery [61]. Furthermore, we also found genes which encode proteins that have no definitive described functions in phages. A very large gene, ORF 200 with a nucleotide length of 19,254 bp contained a protein domain termed PARCEL (Palindromic Amphipathic Repeat Coding Elements) which has not been described in phages up to date. These repeats have dyad symmetry and variable hydrophilic and conserved hydrophobic regions and have been found in bacteria, eukaryotes [62] and recently in giant viruses [63]. PARCEL protein is related to mobile elements and has been considered as products of horizontal gene transfer [64], however the GC content of ORF 200 is not different to the GC content of the phage genome, which contradicts this likelihood. It has also been suggested that PARCEL protein facilitate diversification of bacterial surface protein [62] and also play a role in the coevolution of viruses with their hosts [63]. In addition, we also found a Ro60 related protein which has been described to have Y RNAs bound to its structure. The exact functions of this protein are unknown, but several phages have been reported to have Yr1A, a protein module within Y RNA that can mimic the structure of tRNA [65]. This supports the suggestion on the diverse genetic resources contained within jumbo phage genomes as a result of the large number of gene acquisition [66].
Phages have been associated with risks of horizontal gene transfer [20] therefore, it is imperative that any phage considered for therapy are investigated for temperateness and transduction potential.
In our present study, we did not find any integrase, virulence or antibiotic resistance encoding genes in vB_VhaM_pir03 genome. Hence, this phage has an exclusive lytic lifestyle and does not risk antibiotic resistance and virulence gene transfers for applications. The vB_VhaM_pir03 genome is also absent of any termini, is circularly permuted and terminally redundant which suggest a headful packaging mechanism [67]. Based on the reads mapped to the vB_VhaM_pir03 genome, the potential host sequence was ≤0.09% which indicated that vB_VhaM_pir03 does not exhibit any transduction potential [68]. From whole genome sequence homolog search, a novel jumbo phage, vB_BONAISHI which infects Vibrio coralliilyticus, was determined to be the only similar phage to vB_VhaM_pir03 with an average ORF similarity of 52.3%. Analysis of genomic synteny in this study revealed that both phages shared similar genomic arrangements but with low nucleotide sequence similarities. In addition, phylogenetic analysis using large terminase subunits of vB_VhaM_pir03 and other described jumbo phages produced strong bootstrap support to the evolutionary relationship between vB_VhaM_pir03 and vB_BONAISHI. This suggested that both phages may have diverged from a common ancestor but have since undergone multiple nucleotide substitution events [61]. Since, vB_BONAISHI was previously described as a singleton phage in the jumbo phage phylogenetic tree [29], the phylogenetic and phylogenomic analyses indicated that vB_VhaM_pir03 is a novel phage.
vB_VhaM_pir03 was shown to have a broad host lytic activity against different Vibrio species within the Harveyi clade [69,70]. To our knowledge, only one other Vibrio phage, KVP40 was reported to infect multiple Vibrio species [71]. The ability of phages to infect different strains and species of bacteria has been suggested as an adaptation tool for survival [72]. For jumbo phages, their diversity of gene functions was suggested to be responsible for their broad host lytic activity [73,74]. In our study, the broad host lytic activity of vB_VhaM_pir03 can be considered analogous to broad-spectrum antibiotics [75] therefore in aquaculture, pathogenic Vibrio species are ubiquitous and diverse [76,77] thus a broad host phage such as vB_VhaM_pir03 would be advantageous in controlling the Vibrio population. It has been previously reported that there is an inverse relationship between bacterial antibiotic and phage resistance due to the high biological cost to maintain each resistance mechanism [78]. In this study, the susceptible host strains to vB_VhaM_pir03 were determined to be multiple antibiotic resistant bacteria. Therefore, the use vB_VhaM_pir03 would be a suitable approach against antibiotic resistant bacteria either as an antibiotic alternative or as a co-therapeutant.
In vitro lysis is typically carried out as an intermediate step to large scale applications. In this step, the therapeutic effects of phages at different MOIs and the host resistance development are measured concurrently against time [38]. In the present study, several strains including the host were determined to be susceptible to vB_VhaM_pir03 suggesting that when applied in vivo, pharmacokinetics and pharmacodynamics of the phage therapy are the major factors in the treatment efficacy [79]. Based on the in vitro lysis assays, host bacterial growth which were inhibited by vB_VhaM_pir03 at MOIs 10 and below would be ideal for immersion type delivery method since the concentration of phage required for application in an aquaculture setting would be practical. However, for hosts needing higher concentrations of vB_VhaM_pir03, further investigations are required to determine if site-specific delivery methods such as oral or intraperitoneal injection can provide effective therapeutic effect. Nonetheless, several host bacterial strains used in the in vitro lysis showed continuous growth even after inoculation of vB_VhaM_pir03. This suggested the development of resistant bacterial subpopulation due to selective pressure [80]. This supports further investigations into co-administration of vB_VhaM_pir03 with another phage as a cocktail in which each phage utilizes different infection mechanisms to overcome bacterial phage resistance mechanisms and avoid competitive phage infection [53]. In addition, the continuous bacterial growth even after inoculation of vB_VhaM_pir03 during the in vitro lysis assays may also suggest the development of a phage carrier state population in which the phage and host exist in equilibrium without any lysogenic activity or viral replication. In a phage carrier state, the phage may reside in its host cell without any lysogenic conversion, as a possible means of protection against environmental factors and avoidance of bacterial phage resistance mechanisms [81]. While the development of phage carrier state during vB_VhaM_pir03 infections was not examined, previous studies have reported the existence of phage carrier states resulting in reduction of virulence of Pseudomonas aeruginosa in biofilms and new colonization of Campylobacter jejunii in chickens [82,83].
In the in vivo trial with Artemia nauplii, we found that a single dosage of vB_VhaM_pir03 was effective in increasing survival of Artemia nauplii infected with Vibrio harveyi strain Vh5 at 24 h post infection even at MOI 0.1. This result was showed that vB_VhaM_pir03 performed slightly better than a previously reported phage therapy trial with Artemia spp. in which the survival was measured at 50% [26] However, it was observed that vB_VhaM_pir03 was unable to provide protection to the Artemia nauplii at 48 h post infection. Nonetheless, Artemia nauplii population still showed a higher percentage of survival to the untreated population which suggested a residual effect of protection. Despite the short protective period provided by vB_VhaM_pir03, a single dosage of vB_VhaM_pir03 can also be considered as a potential live feed disinfectant since Artemia nauplii are fed to fish within 24 h post hatch. For the infected Artemia nauplii that received a delayed treatment, vB_VhaM_pir03 was unable to provide protection which suggested that the damage caused by vibriosis was irreversible similarly to the reported results of Diaz et al. [84]. Phages are conventionally used in therapeutic purposes [27], however, for pathogens which cause irreversible damage, consideration may be given to use of phages as prophylaxis such as suggested by Silva et al. and Zaczek et al. [85,86].

Conclusions and Future Directions
In this study, we have provided a comprehensive biological and genomic characterization of vB_VhaM_pir03 as a candidate for phage therapy against Vibrio harveyi. The biological characterization of vB_VhaM_pir03 showed that it can rapidly locate and adsorb to a host and produce a high burst size within a short latent phase. Further characterization showed that vB_VhaM_pir03 has a broad lytic activity against thirty-one multiple antibiotic resistant strains of species belonging to the Harveyi clade. This is a unique ability of vB_VhaM_pir03 which has only been reported in only one other Vibrio phage, KVP40. Genomic analysis revealed a wealth of diverse gene functions that contributed to the efficacy of vB_VhaM_pir03. Furthermore, we also suggest that vB_VhaM_pir03 is not a temperate phage, does not harbor virulence or antibiotic resistance genes and does not exhibit transduction potential. Evaluation of the performance vB_VhaM_pir03 in vitro showed that it can inhibit several host bacterial growths at low MOI which supports its application in phage therapy. Finally, in the in vivo trial, vB_VhaM_pir03 was able to provide some protection to Artemia nauplii against vibriosis at 24 h post infection at all MOIs. Further characterization of its genome to understand its underlying mechanism is suggested to be carried out. In addition, we also suggest that large scale phage therapy trial with vB_VhaM_pir03 which includes investigations into different types of delivery methods. Finally, we would also like to emphasize that phage characterizations should be comprehensive to ensure a safe and practical phage therapy. This is very important towards the progress of phage therapy from the regulatory perspective.

Bacterial Strains Used in This Study
Thirty-one strains of Vibrio harveyi, V. alginolyticus, V. owensii, V. anguillarum, V. campbellii, V. parahaemolyticus and other Vibrio spp. (Table 3) used in this study were obtained from the bacterial collection of the Laboratory of Aquaculture Microbiology, Institute of Marine Biology, Biotechnology and Aquaculture (IMBBC), Hellenic Center for Marine Research (HCMR) in Heraklion, Crete. The bacterial strains were previously identified either through their NCBI or ENA accession numbers for the type strains, biochemical test (BIOLOG GENiii) and PCR method (16 s rRNA and toxR amplifications). In addition, unidentified Vibrio spp. isolated from the live feeds of HCMR were also used. All the bacterial strains were maintained in microbeads (MicroBank) at −80 • C and were grown in Lysogeny Broth (10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl, 1L deionized water, 0.75 g/L MgSO 4 , 1.5 g/L KCl, 0.73 g/L CaCl 2 ) at 25 • C when used.

Antibiotic Susceptibility Testing
Antibiotic susceptibility testing was carried out for the bacterial strains used in this study according to standard disk diffusion test [87]. Bacterial suspension of the 31 selected strains were diluted to obtain a 0.7 absorbance read at OD 600 . The diluted bacterial suspensions were then plated on Mueller-Hinton agar (Difco, Detroit, MI, USA) with 2% NaCl. Antimicrobial susceptibility disks (ThermoFisher Scientific, Waltham, MA, USA) ( Table 4) were placed on the agar plates and incubated at 25 • C (optimal temperature for the bacteria used) for 24 h. The recorded diameters were interpreted as susceptible, medium, or resistant according to the Clinical Laboratory Standards Institute (CLSI) guidelines CLSI M45-A2 [88] and CLSI M100-S25 [89] as in Table 4.

Isolation and Purification of Bacteriophages
Water samples were collected from three locations: (a) the Port of Piraeus, Athens, (b) the Karavolas beach, Heraklion, Crete and (c) a fish tank in the broodstock section of HCMR in Heraklion. 250 mL of the collected water samples were then enriched with 25 mL of concentrated LB and 2.5 mL of the host strain, Vibrio harveyi type strain DSM19623 liquid culture. The enriched water sample were incubated at 25 • C with a shaking speed of 70 rpm for 24 h. Following filtration through a 0.22 µm sterile filter (GVS Life Sciences, Sanford, ME, USA), 10 µL of each sample were spotted on bacterial lawns of the host strain. Following 24 h incubation at 25°C, the clearest plaque formations were collected. Serial propagations for phage purification were then made for the collected plaques against its host by double agar layer method according to Clokie et al. [90]. A single plaque was carefully collected, serially diluted, and propagated again in host bacterial lawn. This step was repeated at least five times for a phage to be considered purified. Phages that showed drastic decrease or loss in plaque formation during purification steps were discarded. One of the purified phages was selected for further characterization and was named vB_VhaM_pir03.

Transmission Electron Microscopy
For transmission electron microscopy, aliquot of vB_VhaM_pir03 suspension with a titer of~10 10 PFU/mL was prepared and negatively stained with 4% w/v uranyl acetate (pH 7.2). The phage was observed using a JEOL transmission electron microscopy operated at 80 kV at the Electron Microscopy Laboratory in the University of Crete. From the obtained digital micrographs, structural dimensions of individual phages were measured with ImageJ software version 1.52t [91] for capsid width, capsid length, tail length, baseplate length and baseplate width.

Host Range Test
For determination of host range for the purified phage, fresh cultures of bacterial strains used in this study (Table 1) were grown in LB at an approximate concentration of 10 7 CFU/mL and were then mixed with top molten LB agar (0.75% agar) and poured on bottom LB/2 agar (5 g/L tryptone, 2.5 g/L yeast extract, 10 g/L NaCl, 1L deionized water, 0.75 g/L MgSO 4 , 1.5 g/L KCl, 0.73 g/L CaCl 2 ) which only contained half of the tryptone and yeast content from the LB agar. After solidification of top agar, 10 µL of vB_VhaM_pir03 were spotted on the host lawn. The phage titer was determined after the agar plates were incubated at 25 • C for 24 h.

Efficiency of Plating (EOP)
Efficiency of plating (EOP) was performed in this study according to [90]. The phage was serially diluted to~10 0 , 10 −1 , 10 −2 , 10 −3 , 10 −4 and 10 −5 and spotted on the bacterial lawns of the 31 susceptible strains. The phage titer was determined after the agar plates were incubated at 25 • C for 24 h. The EOP was calculated as a percentage of the number of plaque-forming units formed on a bacterial strain against the number of plaque-forming units formed on the host DSM19623. EOP more than 10 was categorized as high, EOP between 9.9 and 0.5 was considered medium while EOP less than 0.5 was considered low.

Stability of Phage in Different Temperatures and Organic Solvents
Phage thermal stability was measured by exposing the phage aliquots at~10 7 PFU/mL to different temperatures (25,30,35,40,45,50 and 55 • C) for 1 h before being rested at 25 • C for 10 min. The aliquots were then serially diluted and spotted on host bacterial lawn. The phage titer was determined after the agar plates were incubated at 25 • C for 24 h. vB_VhaM_pir03 stored at 4 • C for 24 h was used as a control.
The sensitivity of vB_VhaM_pir03 to chloroform was determined by exposing~10 7 PFU/mL of the phage aliquots to 10% chloroform at 4 • C for 1 h while the stability of the vB_VhaM_pir03 against commonly used disinfectants in aquaculture was measured by exposing~10 7 PFU/mL of vB_VhaM_pir03 to 0.001% benzalkonium chloride, BKC; 3% hydrogen peroxide, H 2 O 2 ; 1% sodium hypochlorite, NaOCl; 70% ethanol, EtOH and; 1% formaldehyde, CH 2 O at 25 • C for 1 h. vB_VhaM_pir03 incubated at 25 • C for 1 h were used as control. Each treatment was serially diluted and spotted on host bacterial lawn. The phage titer was determined after the agar plates were incubated at 25 • C for 24 h. All assays were done with triplicates.

Adsorption Time and One-Step Growth
Adsorption time and one-step growth of vB_VhaM_pir03 was determined according to Kutter [92] with some modifications. Briefly, host culture in exponential phase (~10 8 CFU/mL) was infected with vB_VhaM_pir03 at multiplicity of infection (MOI) 0.01. Aliquots of the infected culture were then collected and transferred to chilled Eppendorf tubes at 0, 2, 4, 6, 10, 15, 20, 30 min, and kept in ice. The aliquots were centrifuged at 13,000 rpm for 3 min and supernatants were collected and serially diluted. The serial dilutions were then spotted on the host bacterial lawn on LB/2 agar plates. The phage titer was determined after the agar plates were incubated at 25 • C for 24 h.
For one-step growth, 1 mL of host culture in exponential phase (~10 8 CFU/mL) was centrifuged at 13,000 rpm for 3 min. The supernatant was then discarded and the pellet was washed and resuspended in 1 mL of SM buffer (5.8 g/L NaCl, 2 g/L MgSO4, 50 mL 1M Tris-Cl pH 7.5 and 2% gelatin, 1 L deionized H 2 O). This step was then repeated twice before the pellet was finally resuspended in 1 mL of LB. The fresh host culture was then infected with vB_VhaM_pir03 at MOI 0.01. After incubation for 10 min at 25 • C, the infected DSM19623 culture was then transferred to LB with the final volume of 30 mL. Afterwards, 1 mL aliquots were then collected from the infected host culture and immediately transferred to chilled Eppendorf tubes. The aliquots were then centrifuged for 13,000 rpm for 3 min. Subsequently, the supernatants were collected and serially diluted. The serial dilutions were then spotted on the host bacterial lawn on LB/2 agar plates. This step was repeated at 10 min intervals. The phage titer was determined after the agar plates were incubated at 25 • C for 24 h.

In Vitro Cell Lysis
The in vitro cell lysis of vB_VhaM_pir03 against DSM19623 was carried out by loading 180 µL of fresh host bacterial culture in each well of sterile 96-well plates. The plates were then read at OD 600 using TECAN microplate reader (Infinite PRO 200) at 25 • C with orbital shaking. A total of 20 µL of vB_VhaM_pir03 was then added at MOIs 0.1, 1 and 10 when host culture was at exponential phase (~10 8 CFU/mL). Phages added to LB without host bacteria served as control. The assay was also carried out for the remaining 30 susceptible hosts of vB_VhaM_pir03. The growth curves of the cultures were then measured every 10 min for 18 h. All assays were done in triplicates.

DNA Extraction and Purification
The DNA extraction of vB_VhaM_pir03 was carried out using the phenol-chloroform method according to Higuera et al. [51]. The extracted DNA was visualized for quality on 1% agarose gel electrophoresis at 80 kV for 1 h with a 50 kbp ladder. Milli-Q ® Reference Water (Merck KGaA, Darmstadt, Germany) was used as a negative control. The extracted DNA of vB_VhaM_pir03was then stored in −20 • C.

Genomic Analysis
The whole genome of vB_VhaM_pir03 was sequenced, assembled, and annotated previously as described in Misol et al. [93]. The genome sequence of phage vB_VhaM_pir03 is available in GenBank under accession number MT811961. The associated BioProject, SRA, and BioSample accession numbers are PRJNA665717, SRR12712979, and SAMN16261552, respectively. QUAST v4.6.3 [94] and BBMap v38.88 [95] were used to map the reads back to the assembled genome while PhageTerm was used to predict phage termini [68] through the Galaxy server [96]. Proteins of vB_VhaM_pir03 were automatically annotated in Blast2GO Suite [97] using (i) NCBI Basic Local Alignment Search Tool (BLAST) [98] adjusted at non-redundant (nr) protein database, (ii) Gene Ontology [99] and (iii) InterPro [100]. Predicted proteins of vB_VhaM_pir03 were also manually annotated with NCBI Conserved Domain Database (NCBI CDD) [101] and HHpred tool [102]. All ORF predictions and annotations were manually inspected. Integrase, virulence and antibiotic resistance-encoding genes in vB_VhaM_pir03 were searched using INTEGRALL Database webserver [103], Virulence Factor Database (VFDB) [104], VirulenceFinder and ResFinder webservers [105]. The host DSM19623 genome was analyzed for prophage-like sequences using Prophinder [106] and PHAge Search Tool Enhanced Release (PHASTER) [107]. For protein structural homologies, only probabilities above 90% were accepted for manual protein function assignment of the vB_VhaM_pir03 predicted ORFs. All hits in existing databases with expected value above 10 −3 . The genome of vB_VhaM_pir03 with annotated predicted ORFs was then visualized in a circular representation with Geneious software (Geneious v9.1, Biomatters, Auckland, http://www.geneious.com).

Genome Alignment and Phylogenetic Analysis of vB_VhaM_pir03
The whole proteome of vB_VhaM_pir03 was searched for similarity to other phages using the NCBI BLASTP nr protein database. The phage genomes with significant similarities were then downloaded and aligned with vB_VhaM_pir03 using the progressiveMauve: Multiple Genome Alignment [108] for analysis of the genomic synteny. The viral taxonomic family of vB_VhaM_pir03 and its host taxonomic group were determined using ViPTree: the viral proteomic tree server [99]. A six-frame translation proteome of vB_VhaM_pir03 was generated and compared to other six-frame translations of dSDNA phages in the NCBI database [109]. Phylogeny and molecular evolutionary analyses with other jumbo phages were conducted with Molecular Evolutionary Genetics Analysis (MEGA X) webserver [110]. Forty-nine large terminase subunits of described jumbo phages were downloaded from NCBI database and were aligned with the large terminase subunit of vB_VhaM_pir03 using MUSCLE algorithm [111]. Gaps in the amino acid sequence alignments were trimmed. A maximum likelihood phylogenetic tree was constructed using JTT matrix-based model [112] with bootstrap test = 1000. The tree was visualized using Interactive Tree of Life web server [113].

In Vivo Phage Therapy Trial Using Artemia nauplii
Six different treatments were investigated to assess the efficacy of vB_VhaM_pir03 in controlling pathogenic Vibrio harveyi strain VH5 in brine shrimp, Artemia salina nauplii. The first two treatments were control groups: a negative control containing Artemia nauplii only and a positive control of Artemia nauplii with V. harveyi. The other three treatments were single doses of vB_VhaM_pir03 at MOI 0.1, 1 and 10. The final treatment was a group that received a delayed single dose of vB_VhaM_pir03 at MOI 10 at 24 h post infection. The V. harveyi strain Vh5 was determined earlier to be the most pathogenic to Artemia nauplii (data not shown). Newly hatched Artemia nauplii were obtained from the live feed section of IMBBC, HCMR and were disinfected according to the protocol by Gomez-Gil et al. [114]. The Artemia nauplii were then washed three times with autoclaved and filtered borehole water (T: 25 • C). Afterwards, 50 Artemia nauplii were transferred to each well in Thermo Scientific™ Nunc™ Cell-Culture 6-well plates (Thermo Fisher Scientific 168 Third Avenue, Waltham, MA, USA) with 10 mL autoclaved and filtered borehole water. Aliquots from the washed Artemia nauplii were spotted on a bacterial lawn to observe presence of sodium hypochlorite residues. All treatments except for the negative control were inoculated with Vh5 at~10 −5 CFU/mL. At 2 h post infection, the single dose vB_VhaM_pir03 treatments were inoculated. For the delayed treatment, inoculation of vB_VhaM_pir03 was only carried out at 24 h post infection. The Artemia nauplii were fed with autoclaved Aeromonas hydrophila at 10 cells per individuals daily [115]. The sterility of the autoclaved A. hydrophila was tested earlier by transferring 10 µL from its suspension to LB and streaking on a LB agar. Individual counts of the Artemia nauplii were carried out by visual counting using a NIKON SMZ-800 Stereomicroscope (Nikon Instruments Inc., 1300 Walt Whitman Road Melville, NY 11747-3064, USA) at 24 and 48 h post infection to determine survival percentage. 100 µL of water from each treatment were taken and serially diluted before spotted on TCBS agar at 24 and 48 h post infection to determine the total Vibrio load. All treatments were done in triplicates at 25 • C.

Statistical Analysis
One-way ANOVA was performed for the thermal stability and effects of organic solvents assays. Two-way ANOVA was performed for calculation the survival of Artemia nauplii (factor A: hours post infection, factor B: treatments) and total Vibrio load (factor A: hours post infection, factor B: treatments). Tukey's HSD post hoc test [116] was used as a multiple comparison tool after ANOVA was performed. Standard error of the mean was displayed for n = 3. All statistical analyses were carried out with PAST: Paleontological Statistics Software Package for Education and Data Analysis version 4.03 [117].

Conflicts of Interest:
The author declares that there is no conflict of interest.