Retrons and their applications in genome engineering

Abstract

Precision genome editing technologies have transformed modern biology. These technologies have arisen from the redirection of natural biological machinery, such as bacteriophage lambda proteins for recombineering and CRISPR nucleases for eliciting site-specific double-strand breaks. Less well-known is a widely distributed class of bacterial retroelements, retrons, that employ specialized reverse transcriptases to produce noncoding intracellular DNAs. Retrons' natural function and mechanism of genetic transmission have remained enigmatic. However, recent studies have harnessed their ability to produce DNA in situ for genome editing and evolution. This review describes retron biology and function in both natural and synthetic contexts. We also highlight areas that require further study to advance retron-based precision genome editing platforms.

Figure 1.

Retron structure and organization. (A) Retrons are encoded as a single polycistronic transcriptional cassette containing a promoter, an msr (blue) including a conserved priming guanosine residue (green), an msd (red), self-complementary regions (yellow), and a reverse transcriptase (purple, not to scale) (12). The black arrow shows the direction of the coding strand for the retron cassette; the red arrow shows the direction of reverse transcription. Inset: sequence logo of nucleotides adjacent to the priming guanosine of 15 msr-msd sequences. Retron-Sen1 (Se72) was excluded because its msr lacks an apparent priming guanosine (46). (B) Structure of a representative msr-msd transcript encoded by Retron-Eco1 (Ec86) (23). The transcript encoding the reverse transcriptase is cleaved from that of the msr-msd before folding (12). The cytosine at which reverse transcription begins is circled in purple. (C) Structure of the mature Retron-Eco1 (Ec86) msDNA. Inset: the msr and the 5′ end of the reverse-transcribed msd are covalently bonded to the priming guanosine via a 2′-5′ linkage.

Figure 2.

Retrons form a distinct class within the larger group of bacterial RTs. (A) Phylogenetic tree showing relationships between Group II intron maturase reverse transcriptases, bRT RT from the Bordetella bacteriophage Diversity-Generating Retroelement (DGR), and retron RTs. Group II intron maturase and DGR RTs are colored black; retron RTs are colored by phylum. Putative retrons (dashed) identified from Bacteroidetes, Betaproteobacteria, Cyanobacteria, Fusobacteria, Firmicutes, and Euryarchaeota genomes were employed as these groups lack experimentally validated retrons (39). (B) Sequence alignments of RT-Eco1 (Ec86), RT-Mxa1 (Mx162), RT-Sen2 (Se72), Roseburia intestinalis Group II intron maturase reverse transcriptase domain (‘Group II RT-RI’), and the Bordetella bacteriophage DGR reverse transcriptase bRT. Each retron RT sequence aligns well to the seven universal reverse transcriptase regions (boxed in black) and includes several interdomain regions, including the retron-specific regions X and Y (boxed in red) (12,40). Conserved, retron-specific residues in regions X and Y are boxed in orange; the catalytic XXDD core common to all known reverse transcriptases is boxed in purple. (C) Distant-level view of the alignment of RT sequences from all 16 fully characterized retrons. Sequences were aligned with the NCBI online tool COBALT (38); the alignment was then employed to construct a phylogenetic tree using the neighbor-joining tree build method in Geneious software under a Jukes-Cantor genetic distance model. Residues shaded in red, dark blue, light blue, and grey correspond to residues that are 100%, 80–100%, 60–80% and <60% homologous across the aligned sequences. RT-Mxa1 (Mx162), RT-Mxa2 (Mx65), RT-Sau1 (Sa163), and RT-Nex2 (Ne144) contain an extended N-terminus; RT-Eco2 (Ec67) contains both N- and C-terminal extensions (boxed in light red).

Figure 3.

Organization and genomic context of select retron cassettes. (A) Structures of representative validated retron cassette genes. The msr (blue), msd (red), complementary repeat regions (yellow), and RT (purple), and associated unknown ORFs (green) are indicated. Many retron cassettes (‘standard’) consist only of an msr and msd flanked by complementary regions upstream of an RT. Variations on this theme include ORFs either immediately downstream of the RT or between the msr-msd and the RT. Retron-Sen1 (Se72) contains an msd that is transcribed in both directions to produce double-stranded DNA. A hairpin-containing ORF flanked by repeat regions may function as an alternative msr. The cassette also includes an RT and a second unknown ORF with homology to cold shock proteins (46). (B) Genomic contexts of retron cassettes. Several validated retrons occur in putative prophage sites flanked by repeat regions (12). Retron-Sen2 (St85) occurs in what is thought to be an ancient prophage site downstream of the Sgi1 genomic island associated with pathogenesis. Retron-Eco5 (Ec107) replaces 34 bases of an intergenetic region with no apparent mobile element nearby. Retron-Sen1 (Se72) occurs on a plasmid. (C) Several different retrons, along with additional orfs, may occupy the same locus in different bacterial strains or populations. For example, the Retron Eco1-(Ec86) and an upstream ORF occupies the same genetic site as Retron-Eco2 (Ec67) in different E. coli strains (59,85). The Vibrio cholerae retrons Retron-Vch1 (Vc95), Retron-Vch2 (Vc81) and Retron-Vch3 (Vc137) also occupy the same genetic site.

Figure 4.

Biotechnological applications of retrons. (A) In retron-based bacterial gene editing, the msd is homologous to a segment of the target gene (pink). The desired mutation is also introduced into the msd (turquoise). msDNA (thin dashed/solid line) is produced and edits the targeted sequence via an undetermined mechanism. (B) Schematic of retron-mediated genomic recording via Synthetic Cellular Recorders Integrating Biological Events, e.g., ‘SCRIBE’ (13). A small-molecule inducible retron is designed to turn on or off a selectable marker. Switching of the selectable marker occurs only in the presence of the small molecule inducer (orange hexagon). The desired edit is produced in proportion to the duration and concentration of inducer, enabling ‘tape recording’ of the inducer's presence. (C) Retron-mediated mutagenesis and evolution (15). In this application, a retron homologous to the desired target gene is expressed under an error-prone T7 RNA polymerase (green). The resulting msDNAs have random mutations (stars) that are subsequently introduced into their target gene as in (A). Alternatively, these mutation-containing msDNAs can edit the parent retron sequence, enabling continuous accumulation of mutations and evolution. (D) Schematic for Cas9 Retron precISe Parallel Editing via homologY, e.g. ‘CRISPEY’ (14). A retron msr/msd sequence containing homology and a desired edit to a targeted gene is expressed as a fusion to a sgRNA (light blue) targeting the gene of interest. The retron RT (purple) and Cas9 (green) are expressed in trans. Cas9 generates a DNA break and the RT-generated DNA is used as a template by the host's DNA repair machinery. Genes are not drawn to scale.