New insights into the biodiversity of coliphages in the intestine of poultry

Phage isolation

In this study, 38 coliphages were isolated from poultry faecal samples collected from 27 Belgian poultry farms located in five different regions, including West Flanders, East Flanders, Antwerp, and Limburg. Between one and seven phages were isolated from each farm using E. coli C600 or K514 as host strain.

Phage morphological analysis

Based on a sequencing cut-off value of ≤ 95% nucleotide similarity, 18 coliphages were selected and subjected to TEM to determine phage morphology and confirm phage classification. Based on the morphological features, the phages were classified into the Caudovirales order and either the Siphoviridae family or the Myoviridae family. Analysis of the isolated Siphoviridae phages showed a long flexible non-contractile tail with a length varying between ~ 100 and ~ 200 nm and icosahedral heads with widths ranging from ~ 52 to ~ 77 nm (Fig. 1a–h). Among the isolated Myoviridae phages a long straight contractile tail was observed with a tailed length varied between ~ 100 and ~ 120 nm, head widths ranging from ~ 65 to ~ 84 nm, and head lengths from ~ 60 to ~ 110 nm (Fig. 1i–r). Taxonomic classification of each of the coliphages is shown in Table 1.

Phage genome sequence analysis and annotation

All 38 coliphages isolated in this study were characterized based on WGS data. An overview of the genomic characteristics and properties are listed in Table 1. According to FastQC parameters, good quality of the raw sequence data for all phages was confirmed. The phage genomes ranged in size between 44,324 and 173,384 bp, with a G+C content between 35.5 and 46.4%. Genomes smaller than 90,000 bp had a G+C content between 38.9 and 46.4%, whereas the larger genomes had a G+C content of 35.5–38%. For each coliphage, 72–275 putative CDSs were identified using both automatic and manual annotation. CDSs encoding the phage terminase small subunit, the phage terminase large subunit, the phage portal protein, and phage capsid and scaffold proteins were identified within all 38 coliphage genome sequences. They presented the same conserved genome structure with a general gene order: the terminase small subunit upstream from the terminase large subunit, the phage portal protein and two genes encoding phage capsid and scaffold proteins. In general, one phage terminase small subunit, one phage portal protein, and up to four phage capsid and scaffold proteins were found within each of the phage genomes. Besides, phage exonucleases were identified in all phage genomes. For each phage, one to three CDSs for exonucleases were found. No gene encoding for an integrase was found, indicating that these phages are strictly virulent/lytic phages. No known acquired resistance or virulence genes were detected in any of the 38 phage genomes.

Phage phylogeny and taxonomy

Taxonomic classification of the 38 isolated coliphages was performed through multiple WGS genome comparisons. These coliphages included 27 (71%) Siphoviridae coliphages and 11 (29%) Myoviridae coliphages. The Siphoviridae phages were compared with 146 published phages from this family. The Myoviridae phages were compared with 171 published Myoviridae phages. According to ICTV guidelines, phage family, subfamily and genus were predicted based on genome similarity. Results are shown in Table 1. All Siphoviridae phages belonged to the Tunavirinae subfamily, except for Phage 61. This phage was predicted to belong to the Tequintavirus genus, which do not have any ICTV subfamily. The ten phages, Phage 8, 53, 54, 63, 65, 68, 69, 71, 72, and 75 all belonged to the Hanrivervirus genus. Three phages belonged to the Rtpvirus genus, including Phage 17, 70, and 73. No existing ICTV genus could be assigned to the remaining 13 coliphages. Phage 28, 56_1 and 76 could be assigned to the same unknown genus. Phage 58 and Phage 74 were found to be in the same genus. Phage 47, 48, 59, 64, and 77 were predicted to belong to the same genus. Phage 52, 56_2, and 80 were predicted to belong to the same genus. The 11 Myoviridae belonged either to the Tevenvirinae or the Ounavirinae subfamily. Tevenvirinae phages included the six phages: Phage 10, 11, 15, 18, 30 and 55. Phage 10, 11 and 55 belonged to the Tequatrovirus genus, and Phage 18 and 30 belonged to the Mosigvirus genus. Ounavirinae phages included the remaining five phages; Phage 60, 62, 66, 78 and 79. All phages belonged to the Felixounavirus genus.

Phage diversity

To investigate the diversity of the coliphages, phage genomes were first clustered based on whole genome sequence. A total of 173 Siphoviridae and 182 Myoviridae coliphage genomes were included. Characteristics of selected reference genomes are listed in Supplementary Table S1. Siphoviridae phages isolated in this study were found in five different (sub)clusters, cluster A1–3, B and C, with a cut-off value of 0.82 (Fig. 2). Cluster A was divided into three subclusters. Fourteen of our phages, formed subcluster A1 together with the three pSf-1-like reference phages from the NCBI database. Phage 80, 28, 56_1, and 76 formed subcluster A2 with the three Swan01-like reference phages. Phages 69, 52, and 56_2 formed subcluster A3 with phage Jahat_MG145. Phage 73, 70, 17, 58 and 74 formed cluster B without any known reference phages. Phage 61 was placed in cluster C with 13 T5-like reference phages. For the Myoviridae phages, the resulting phylogeny placed phages isolated in this study in three different clusters with a cut-off height of 0.52 (Fig. 3). Phages 62, 78, 66, 60 and 79 formed cluster D with Felix01-like reference phage Alf5. Phage 30, 15 and 18 formed cluster E with 19 T4-like reference phages (cut-off height of 0.36). Phages 55, 11 and 10 were placed in cluster F with 57 reference phages. At the cut-off height of 0.39 Phage 55 was found in a different subcluster than Phage 10 and 11.

Coliphages were further assessed based on the presence/absence of families of orthologues genes in their pan genome. Similar clusters were observed with only minor changes. For the Siphoviridae phage analysis, 5,227 gene groups were included (Supplementary Fig. S1). The resulting phylogenetic analysis placed phages isolated in this study in the same five clusters, cluster A1–3, B and C, with a cut-off height of 0.81 (Supplementary Fig. S2). One additional reference phage was found in cluster B and C, including the T1-like reference phage CEB_EC3a and the T5-like reference phage EPS7, respectively. For the Myoviridae phage analysis, 9,420 gene groups were included (Supplementary Fig. S3). The resulting phylogeny placed phages isolated in this study in the same three clusters, cluster D, E, and F, with a cut-off height of 0.58 (Supplementary Fig. S4). For cluster D, additionally 13 Felix01-like reference phages were found. In contrast to the WGS-based analysis, at a cut-off height of 0.39, all cluster F phages isolated in this study, were found in one subcluster with 10 T4-like reference phages. The degree of topological and branch length agreement between the different phylogenetic methods were compared (Supplementary Table S2).

The coliphage diversity was further assessed based on three phage marker genes: the terminase large subunit and phage portal protein, and the phage exonuclease. Selected gene sequences from known phages were included for reference. Results are summarized in Table 1. For all three marker genes, cluster formation was in accordance with resulting clusters of the pan genome- and WGS-based phylogeny, cluster A–F, only with minor differences. Results based on the terminase large subunit analysis are shown below (Fig. 4).

For cluster A, all coliphages isolated in this study were found within same subclusters as for the WGS-based phylogeny except for Phage 63, which was found in the A2 subcluster instead of A1. Analysis based on the phage portal protein resulted in the division of our A2 subcluster phages into two groups: Phage 56_1, 80 and 28 in one group and Phage 76 in the other group (Supplementary Fig. S5). Analysis based on the exonuclease resulted in multiple clusters of cluster C and F, as phages from these clusters encoded 2 or 2–3 exonuclease genes, respectively (Supplementary Fig. S6). Comparison of the cluster construction of the three single genes analysis showed only minor topological and branch length differences (Supplementary Table S2). Moreover, cluster construction was in accordance with phage subfamily defined based on the whole genome. Siphoviridae phages from cluster A and B belonged to the Tunavirinae subfamily, and Siphoviridae phages form cluster C had no defined ICTV subfamily. Myoviridae phages from cluster D belonged to the Ounavirinae subfamily, and Myoviridae phages from cluster E and F belonged to the Tevenvirinae subfamily.

Phage comparative genomics

Pan genome analysis of Siphoviridae and Myoviridae phages isolated in this study revealed that neither of the two groups had any core genes. Analysis of coliphage genomes from each of the six clusters, A–F, identified core genes (core and softcore) and accessory genes (shell and cloud). As cluster A phages had only five core genes (2% of the total genome), analysis of subclusters, A1, A2 and A3, were performed additionally. Results are summarized in Table 2. The pan genome included between 81 and 333 genes, and core genes constituted between 22 and 73% of the pan genome.

The level of synteny and genomic rearrangement within each cluster or subcluster of related phages was assessed by genome comparison. Results are summarized in Table 2. Eight comparisons were performed, corresponding to the eight (sub)clusters, A1, A2, A3, B, C, D, E, and F resulting from the phage diversity analysis above (Supplementary Fig. S7–S14). Genome comparison of the phages resulted in identification of local collinear blocks (LCBs), indicating homologues DNA regions shared by two or more genomes without sequence rearrangements. The LCBs comprised different modules of genes with different functions, including modules for DNA packaging, structural proteins, head and tail morphogenesis, and host cell lysis. Several modules comprised only hypothetical proteins with unknown function. The average level of conservation varied between the different type of genes.

Genes encoding the terminase large and small subunit, the major capsid protein, DNA primase, single-stranded protein, portal protein, recombinase, specific tail protein and holin were the most conserved genes between all phages, whereas genes with the lowest level of conservation included, specific tail fiber proteins, tail tape measure proteins and HNH homing endonucleases. Hypothetical proteins were found with large variation in level of conservation. Each phage genome comprised between four and 17 LCBs. Genome comparison of phages belonging to subcluster A1, A2 and A3 identified 16, seven and four LCBs, respectively. All phages in each cluster comprised all LCBs. All cluster B phages comprised all six LCBs. For the cluster C phages, between 6 and 10 LCBs were identified for each phage. Phage 61 isolated in this study comprised all nine regions. Variation in number of LCBs was due to a variable repeat region comprising multiple LCBs, which was found only in some of the cluster C phages. For the cluster D comparison, 14–17 different LCBs were identified for each phage. Variation in number of LCBs was due to four different small variable regions of which some of all were missing in the majority of the phages. Phages isolated in this study, including Phage 79, Phage 78, Phage 60, Phage 66, and Phage 62, comprised 17, 17, 16, 15, and 14 LCBs, respectively. Comparison of phages belonging to cluster E identified 18 LCBs. All phages lacked one or both of the same two LCBs. Phages isolated in this study, including Phage 30, Phage 15, and Phage 18, comprised 16, 17, and 17 LCBs, respectively. All 13 cluster F phages included in the comparison comprised all five LCBs. The comparison confirmed the presence of homologue regions between the phages within the clusters but also highlighted that re-arrangement and/or gain/loss of LCBs must have occurred at some point during the evolution of the phages. The region encoding the terminase large subunit and portal protein were present in a conserved region all genomes in all eight comparisons.

Source: Ecology - nature.com