Bacteriophages (phages) dramatically shape microbial community composition, redistribute nutrients via host lysis and drive evolution through horizontal gene transfer. Despite their importance, much remains to be learned about phages in the human microbiome. We investigated the gut microbiomes of humans from Bangladesh and Tanzania, two African baboon social groups and Danish pigs; many of these microbiomes contain phages belonging to a clade with genomes >540 kilobases in length, the largest yet reported in the human microbiome and close to the maximum size ever reported for phages. We refer to these as Lak phages. CRISPR spacer targeting indicates that Lak phages infect bacteria of the genus Prevotella. We manually curated to completion 15 distinct Lak phage genomes recovered from metagenomes. The genomes display several interesting features, including use of an alternative genetic code, large intergenic regions that are highly expressed and up to 35 putative transfer RNAs, some of which contain enigmatic introns. Different individuals have distinct phage genotypes, and shifts in variant frequencies over consecutive sampling days reflect changes in the relative abundance of phage subpopulations. Recent homologous recombination has resulted in extensive genome admixture of nine baboon Lak phage populations. We infer that Lak phages are widespread in gut communities that contain the Prevotella species, and conclude that megaphages, with fascinating and underexplored biology, may be common but largely overlooked components of human and animal gut microbiomes.
Bibliographical noteFunding Information:
Statement of ethics. The human faecal samples obtained were part of a clinical phase I/II study in rural Bangladesh entitled ‘Selenium and arsenic pharmacodynamics’ (SEASP) run by Graham George (University of Saskatchewan) and funded by the Canadian Federal Government, through a programme entitled Grand Challenges Canada-Stars in Global Health, with additional funds from the Global Institute for Water Security at the University of Saskatchewan. The SEASP trial was approved by the University of Saskatchewan Research Ethics Board (14–284) and the Bangladesh Medical Research Council (940,BMRC/NREC/2010-2013/291). Additional ethics approval was also obtained by UCL (7591/001).
We thank T. Lowe for helpful discussion related to tRNAs and introns and D. Paez-Espino and colleagues at the Joint Genome Institute for generating sequences in a meta-analysis related to IMG-VR that enabled identification of the baboon cohort as a potential source of Lak phage sequences. We thank S. Lei and K. Lane for ggKbase support, L. Barreiro for his contributions to the baboon study and Dr. R. Meheust for his assistance with the protein clustering analysis. Funding was provided by the National Institutes of Health (RAI092531A) and Sloan Foundation (G 2012-10-05) to J.F.B., Chan Zuckerberg Biohub funding to J.F.B., P.T. and K.S., and National Science Foundation grant no. IOS 1053461 to E.A.A. We acknowledge Rampelli et al.14, Ghosh et al.15, Thomas et al.14,1517, and Obregon-Tito et al.18 whose published research generated, respectively, the Hadza, Indian children, cow rumen and Peruvian gut read data sets used in this study. Faecal samples were collected from patients in the clinical phase I/II SEASP trial in Bangladesh that was jointly led by G. George and I. Pickering (University of Saskatchewan), with the assistance of the SEASP team (https://clinicaltrials.gov/ct2/show/NCT02377635), and funded by the Canadian Federal Government, through Grand Challenges Canada-Stars in Global Health and by the Global Institute for Water Security. We thank O. Ponomarenko and S. Shaha for organizing the transport of the faecal samples.
© 2019, The Author(s), under exclusive licence to Springer Nature Limited.