Diversity-generating retroelements (DGRs) are genetic cassettes that selectively mutate target genes to produce hypervariable proteins. First characterized in Bordetella bacteriophage BPP-1, the DGR creates a hypervariable phage tail fiber that enables host tropism switching. Subsequent surveys for DGRs conclude that the majority identified to date are bacterial or archaeal in origin. This work examines bacteriophage and bacterial genomes for novel phage-encoded DGRs.

This survey discovered 92 DGRs that were only found in phages exhibiting a temperate lifestyle. The majority of phage-encoded DGRs were identified as prophages in bacterial hosts from the phyla Bacteroidetes, Proteobacteria, and Firmicutes. Sequence reads from these previously unidentified prophages were present in viral metagenomes (viromes), indicating these prophages can produce functional viruses. Five phages possessed hypervariable proteins with structural similarity to the tail fiber of BPP-1, whereas the functions of the remaining DGR target proteins were unknown. A novel temperate phage that harbors a DGR cassette targeting a protein of unknown function was induced from Bacteroides dorei. This phage, here named Bacteroides dorei Hankyphage, lysogenizes 13 different Bacteroides species and was present in 34% and 21% of whole-community metagenomes and human-associated viromes, respectively.

Here, the number of known DGR-containing phages is increased from four to 92. All of these phages exhibit a temperate lifestyle, including a cosmopolitan human-associated phage. Targeted hypervariation by temperate phages may be a ubiquitous mechanism underlying phage-bacteria interaction in the human microbiome.