Expanding the diversity of mycobacteriophages: insights into genome architecture and evolution

Abstract

Mycobacteriophages are viruses that infect mycobacterial hosts such as Mycobacterium smegmatis and Mycobacterium tuberculosis. All mycobacteriophages characterized to date are dsDNA tailed phages, and have either siphoviral or myoviral morphotypes. However, their genetic diversity is considerable, and although sixty-two genomes have been sequenced and comparatively analyzed, these likely represent only a small portion of the diversity of the mycobacteriophage population at large. Here we report the isolation, sequencing and comparative genomic analysis of 18 new mycobacteriophages isolated from geographically distinct locations within the United States. Although no clear correlation between location and genome type can be discerned, these genomes expand our knowledge of mycobacteriophage diversity and enhance our understanding of the roles of mobile elements in viral evolution. Expansion of the number of mycobacteriophages grouped within Cluster A provides insights into the basis of immune specificity in these temperate phages, and we also describe a novel example of apparent immunity theft. The isolation and genomic analysis of bacteriophages by freshman college students provides an example of an authentic research experience for novice scientists.

Figure 1. Nucleotide sequence comparison of 18 newly isolated mycobacteriophage genomes.

A concatenated file of all 18 newly sequenced genomes (horizontal axis) was compared against a concatenated file of a representative genome of each of the clusters, subclusters, and singleton genomes [as defined in [7]] (vertical axis) using Gepard . The representative genomes on the vertical axis are: A1 (Bethlehem), A2 (D29), B1 (Chah), B2 (Qyrzula), B3 (Phaedrus), B4 (Cooper), C1 (Bxz1), C2 (Myrna), D (Adjutor), E (244), F1 (Boomer), F2 (Che9d), G (Angel), H1 (Konstantine), H2 (Barnyard), I (Brujita), singletons shown are Corndog, Giles, Omega, TM4, and Wildcat.

Figure 2. Nucleotide comparison of Cluster I and Cluster K genomes.

The three genomes of Cluster I (A) and the three genomes of Cluster K (B) were concatenated and compared against themselves using Gepard . Each Cluster is subdivided into Subclusters (I1 and I2, and K1 and K2 respectively), as shown below the dotplots.

Figure 3. Genome organization of mycobacteriophage LeBron.

The LeBron genome is represented as a horizontal bar with markers, and the 132 predicted ORFs are shown as colored boxes either above (rightwards transcribed) or below (leftwards transcribed) the genome. Gene names are shown inside the boxes and the phams to which they belong are indicated above, with the total number of pham members shown in parentheses; white ORFs are orphams with, by definition, no close mycobacteriophage homologues. tRNA genes are shown as short black bars. Putative gene functions primarily identified through database searches are shown.

Figure 4. Geographical distributions of genomically characterized mycobacteriophages.

A–C. Geographical distribution of the isolation sites of 60 previously described sequenced mycobacteriophages according to cluster assignation: United States (A), India, Ireland, and Japan (B), and Pittsburgh, PA (C). Locations of newly isolated mycobacteriophages reported in this study colored according to cluster; locations of previously isolated genomes are shown in black (D).

Figure 5. Evolution of Cluster I genome anatomies.

A. Alignment of genome segments of phages Brujita, Island3, and Che9c. Genes are shown as colored boxes with gene names (a serial number based on that phage) inside the boxes and the pham number indicated above the box with the number of pham members in parentheses. Pairwise nucleotide sequence similarity is shown as colored areas between adjacent genomes, with strength of similarity according to the color spectrum, violet being the most similar, and red the least. B. Alignment of Brujita and Island3 shows that Brujita genes 64 and 66 are related to Island3 genes 65 and 67, respectively, whereas Brujita gene 65 and Island3 gene 66 are distinctly different. The sequences in common are shown bold, and the common genes are shaded dark blue. C. Island3 genes 67 and 68 share a common 60 bp sequence at their 3′ ends. Brujita contains only a single copy of this sequence which represents a recombinant version that matches the upstream part of Island3 67 and the downstream part of Island3 68. Che9c also shares the upstream sequence but is different downstream of gene 75 with sequence discontinuity close to the end of the gene.

Figure 6. Determinants of immunity specificity in Cluster A genomes.

A. Phylogenetic relationship of Cluster A repressors. The neighbor-joining (NJ) tree was drawn by NJPlot using output from an alignment in ClustalW; bootstrap values from 1000 iterations are shown. The repressor clades correspond closely to the subclustering of the genomes as indicated by color shading: Subcluster A1, red: A2, green, A3, yellow, A4, blue. B. The predicted helix-turn-helix motifs of the Cluster A repressor are aligned to show conserved and variant residues. The positions in the second recognition helix of the HTH motif are numbered. The Cluster assignation of the genome encoding the repressors is colored as in A. Bxz2 gp74 and KBG gp73 are included even though both contain frameshift mutations in the repressor gene. In KBG gp73 the mutation lies downstream of the HTH motif. There is a presumed single base deletion in Bxz2 at coordinate 44,987 upstream of the HTH motif, and we have used the ‘corrected’ sequence in the alignment that would results from inclusion of one additional bp at that position. C. Alignment of the consensus stoperator sites in Cluster A genomes. Consensus sequences were derived from alignments of putative operator and stoperator sites (shown in Fig. S4); mixed base consensus positions are indicated as W: A or T, D: A, G or T, V: C, G or T, M: A or C, N: A, C, G or T. Color shading indicates identical consensus positions.

Figure 7. Immunity theft in mycobacteriophage LRRHood?

A. LRRHood is a Cluster C1 phage that contains a ∼1.4 kbp insertion relative to other C1 phages such as Cali. The insertion contains gene 44 encoding a repressor with >99% identity to Subcluster A1 repressors such as Bxb1 gp69. LRRHood does not contain any copies of the repressor binding site, and we propose that its gene 44 functions to protect LRRHood-infected cells from superinfection by Subcluster A1 phages. B. LRRHood gene 44 is one of four genes (43–46) that are absent from the related phage Cali. The 1.35 kbp additional DNA in LRRHood is flanked by 29 bp direct repeats, of which there is only a single copy in Cali and other Cluster C1 phages. Presumably, either LRRHood acquired this 1.35 kbp segment from another phage by recombination within the 29 bp region, or it has been lost from each of the other Cluster C1 phages. The 29 bp repeat is shown in the red box and is present once in Cali at the extreme 3′ end of gene 46, and twice in LRRHood in the end of gene 42 and upstream of gene 47. The upstream and downstream regions common to LRRHood and Cali are shown in green and blue respectively.

Figure 8. Novel intein insertions in mycobacteriophage ET08.

ET08 is a Subcluster C1 phage and encodes two products, gp79 (A) and gp248 (B) that contain intein insertions that are absent from other C1 genomes such as Catera and Bxz1.

Figure 9. MPME1 insertion in mycobacteriophage Hope.

A. Alignment of the four closely related phages in Cluster G reveals insertions of MPME elements in BPs, Hope, and Halo. Hope contains an MPME1 insertion at a new site corresponding to Angel 56. B. Comparison of the Hope and Angel sequences reveals the pre-integration site; left and right inverted repeats (IR-L and IR-R) are shown in bold type. At the left end there is the presence of an atypical 6 bp insertion (shown in turquoise) between IR-L of MPME1 and the target.

Figure 10. Gene swapping in Cluster B1 genomes.

A. Alignment of the Cluster B1 genomes shows a swap of genes between Chah genes 63 and 73 and their homologues. Segments of genome maps of phages Chah, Colbert, UncleHowie and Puhltonio are shown with regions of nucleotide similarity identified by colored shading between them; shading reflects degrees of similarity with the color spectrum, such that violet is most similar and red the least similar. B. A reduction in GC% content in an intergenic region between Colbert genes 63 and 64 and UncleHowie genes 63 and 64 suggests the presence of regulatory elements notwithstanding the swapping of the downstream genes.

Figure 11. Acquisition of an HNH homing endonuclease in mycobacteriophage Pumpkin.

A. Alignment of the Cluster E genomes 244, Pumpkin, and Cjw1 reveals the presence of a Pumpkin gene (40) that is absent from other Cluster E genomes. B. Phamily circles of Phams 1706, 1707, 1567, and 2942, which include Pumpkin genes 37, 38, 39, and 40. Pham1706 and 1707 have members exclusively within Cluster E genomes. Pham 1567 is more widely distributed with members in Subcluster A1, some in A2, F2, and singleton genomes, and is functionally a glutaredoxin. Pham 2942 is a homing endonuclease and is broadly distributed among the mycobacteriophages. Each of the 80 genomes is shown on the circumference of each circle – arranged by cluster – with arcs indicating pairs of genomes containing a pham member; thicker arcs indicate closer similarity. Red and blue arcs show BlastP and ClustalW comparisons respectively. C. Phylogenetic reconstruction of Pham2942. All members are distantly related, and Pumpkin gp40 probably was acquired independently from other Cluster E acquisitions (Porky gp109, 244 gp112). Members are colored according to cluster of parent genome: A1, green; F1, purple; E, orange; I1, blue; B3, yellow; Omega is a singleton. The neighbor-joining (NJ) tree was drawn by NJPlot using output from an alignment in ClustalW; bootstrap values from 1000 iterations are shown.