CRISPR-associated (Cas) transposases (CAST) are RNA-guided systems capable of programmable integration of large segments of DNA without creating double-strand breaks. Engineered Cascade CAST function in human cells but are challenging to deploy due to the complexity of the targeting components. Unlike Cascade, which require three Cas proteins, type V-K CAST require a single Cas12k effector for targeting. Here, we show that compact type V-K CAST from uncultivated microbes are repurposable for programmable DNA integration into the genome of human cells. Engineering for nuclear localization and function enables integration of a therapeutically relevant transgene at a safe-harbor site in multiple human cell types. Notably, off-targets are rare events reproducibly found in specific genomic regions. These CAST advancements are expected to accelerate applications of genome editing to therapeutic development, biotechnology, and synthetic biology.

Type II Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)-Cas9 nucleases have been extensively used in biotechnology and therapeutics. However, many applications are not possible owing to the size, targetability, and potential off-target effects associated with currently known systems. In this study, we identified thousands of CRISPR type II effectors by mining an extensive, genome-resolved metagenomics database encompassing hundreds of thousands of microbial genomes. We developed a high-throughput pipeline that enabled us to predict tracrRNA sequences, to design single guide RNAs, and to demonstrate nuclease activity in vitro for 41 newly described subgroups. Active systems represent an extensive diversity of protein sequences and guide RNA structures and require diverse protospacer adjacent motifs (PAMs) that collectively expand the known targeting capability of current systems. Several nucleases showed activity levels comparable to or significantly higher than SpCas9, despite being smaller in size. In addition, top systems exhibited low levels of off-target editing in mammalian cells, and PAM-interacting domain engineered chimeras further expanded their targetability. These newly discovered nucleases are attractive enzymes for translation into many applications, including therapeutics.

Programmable, RNA-guided nucleases are diverse enzymes that have been repurposed for biotechnological applications. However, to further expand the therapeutic application of these tools there is a need for targetable systems that are small enough to be delivered efficiently. Here, we mined an extensive genome-resolved metagenomics database and identified families of uncharacterized RNA-guided, compact nucleases (between 450 and 1,050 aa). We report that Cas9d, a new CRISPR type II subtype, contains Zinc-finger motifs and high arginine content, features that we also found in nucleases related to HEARO effectors. These enzymes exhibit diverse biochemical characteristics and are broadly targetable. We show that natural Cas9d enzymes are capable of genome editing in mammalian cells with >90% efficiency, and further engineered nickase variants into the smallest base editors active in E. coli and human cells. Their small size, broad targeting potential, and translatability suggest that Cas9d and HEARO systems will enable a variety of genome editing applications.

Lak phages with alternatively coded ~540 kbp genomes were recently reported to replicate in Prevotella in the gut microbiomes of humans that consume a non-western diet, baboons and some pigs. Here, we investigate the diversity and broader distribution of Lak phages in human and animal microbiomes using diagnostic PCR and genome-resolved metagenomics. Lak phages were detected in 13 different animal types and are particularly prevalent in pigs, with significant enrichment in the hindgut compared to foregut. We reconstructed 34 new Lak genomes, including six curated complete genomes, all of which are alternatively coded. The most deeply branched Lak is from a horse faecal sample and is the largest phage genome from an animal microbiome (~660 kbp). From the Lak genomes, we identified families of hypothetical proteins associated with specific animal types. Overall, we substantially expanded Lak phage diversity and demonstrate their occurrence in a variety of human and animal microbiomes.

Lak phages with alternatively coded ∼540 kbp genomes were recently reported to replicate in Prevotella in microbiomes of humans that consume a non-Western diet, baboons, and pigs. Here, we explore Lak phage diversity and broader distribution using diagnostic polymerase chain reaction and genome-resolved metagenomics. Lak phages were detected in 13 animal types, including reptiles, and are particularly prevalent in pigs. Tracking Lak through the pig gastrointestinal tract revealed significant enrichment in the hindgut compared to the foregut. We reconstructed 34 new Lak genomes, including six curated complete genomes, all of which are alternatively coded. An anomalously large (∼660 kbp) complete genome reconstructed for the most deeply branched Lak from a horse microbiome is also alternatively coded. From the Lak genomes, we identified proteins associated with specific animal species; notably, most have no functional predictions. The presence of closely related Lak phages in diverse animals indicates facile distribution coupled to host-specific adaptation.

Phage typically have small genomes and depend on their bacterial hosts for replication. DNA sequenced from many diverse ecosystems revealed hundreds of huge phage genomes, between 200 kbp and 716 kbp in length. Thirty-four genomes were manually curated to completion, including the largest phage genomes yet reported. Expanded genetic repertoires include diverse and new CRISPR-Cas systems, tRNAs, tRNA synthetases, tRNA modification enzymes, translation initiation and elongation factors, and ribosomal proteins. Phage CRISPR-Cas systems have the capacity to silence host transcription factors and translational genes, potentially as part of a larger interaction network that intercepts translation to redirect biosynthesis to phage-encoded functions. In addition, some phage may repurpose bacterial CRISPR-Cas systems to eliminate competing phage. We phylogenetically define major clades of huge phage from human and other animal microbiomes, oceans, lakes, sediments, soils and the built environment. We conclude that their large gene inventories reflect a conserved biological strategy, observed over a broad bacterial host range and across Earth’s ecosystems.

Cas12a enzymes are quickly being adopted for use in a variety of genome-editing applications. These programmable nucleases are part of adaptive microbial immune systems, the natural diversity of which has been largely unexplored. Here, we identified novel families of Type V-A CRISPR nucleases through a large-scale analysis of metagenomes collected from a variety of complex environments, and developed representatives of these systems into gene-editing platforms. The nucleases display extensive protein variation and can be programmed by a single-guide RNA with specific motifs. The majority of these enzymes are part of systems recovered from uncultivated organisms, some of which also encode a divergent Type V effector. Biochemical analysis uncovered unexpected protospacer adjacent motif diversity, indicating that these systems will facilitate a variety of genome-engineering applications. The simplicity of guide sequences and activity in human cell lines suggest utility in gene and cell therapies.

Bacteriophage (phage) 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 phage in the human microbiome. We investigated gut microbiomes of humans from Bangladesh and Tanzania, two African baboon social groups, and Danish pigs, and report that many contain phage belonging to a clade with genomes >540 kb in length, the largest yet reported in the human microbiome and close to the maximum size ever reported for phage. We refer to these as Lak phage. CRISPR spacer targeting indicates that the Lak phage 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 tRNAs, some of which contain enigmatic introns. Different individuals have distinct phage genotypes, and shifts in variant frequencies over consecutive sampling days reflect changes in relative abundance of phage sub-populations. Recent homologous recombination has resulted in extensive genome admixture of nine baboon Lak phage populations. We infer that Lak phage are widespread in gut communities that contain Prevotella species, especially in individuals in the developing world, and conclude that megaphage, with fascinating and underexplored biology, may be common but largely overlooked components of human and animal gut microbiomes.