In summary: IStrons cram three mutually competing functions into one sequence. Our library approach reveals the molecular rules governing this balance, and why ωRNA consistently comes out on top.
www.biorxiv.org/content/10.6...
Splicing has an Achilles' heel: it's exquisitely sensitive to 3' exon sequences. Each new transposition event generates a novel exon context — many incompatible with efficient splicing. This vulnerability may explain why ωRNA formation is so heavily favored.
A stably folded ωRNA sterically occludes cryptic splice sites embedded within its sequence, confining 3′ splice site selection to the native intron–exon boundary. Splicing fidelity is enforced not by mechanism, but by competition.
The key determinant is SL5, the sole stem-loop that overlaps group I intron structural elements. Increased base-pairing stability at the base of SL5 is sufficient to abolish splicing while preserving DNA cleavage — and this relationship holds in the IS607-family CseIStron-1.
What determines whether a given transcript becomes an ωRNA or gets spliced? The answer is hardwired into RNA structure — not dynamic switching. Multiple lines of evidence establish ωRNA formation as the dominant fate.
The terminal CGG is a functional linchpin. Serine recombinases use a GG dinucleotide for target-site selectivity — but our data reveal a critical third nucleotide. Single changes to this CGG cripple all three IStron activities: transposition, cleavage, and splicing.
To map these overlapping requirements at scale, we designed a pooled library of thousands of systematic RE variants and subjected it to all three functional assays in parallel. This let us score every nucleotide's contribution across every pathway, at once.
At the RNA level, ωRNA maturation and splicing compete for the same sequence — and only one can win!
IStrons are a remarkable class of transposons: a single element encodes three distinct biochemical activities: TnpA-mediated transposition, TnpB-mediated RNA-guided DNA cleavage, and group I intron self-splicing.
New preprint from the Sternberg lab!
We used pooled oligonucleotide library mutagenesis + high-throughput sequencing to systematically dissect the molecular determinants of IStron transposition, RNA-guided DNA cleavage, and self-splicing.
www.biorxiv.org/content/10.6...
Out now! In collaboration with Leifu Chang, we uncover the molecular and structural underpinnings of CRISPR-Cas12f-like RNA-guided transcription systems!
Links to the published articles:
tinyurl.com/55kpavet
tinyurl.com/sk6djwx3
Previous thread for the preprint:
bsky.app/profile/did:...
9/9 Great perspective by @amandinemaire.bsky.social and @dbikard.bsky.social :
www.science.org/doi/10.1126/...
8/9 Next steps include better delivery, reducing fitness costs, and taking MetaEdit into new ecosystems.
🙏 to all co-authors and to Ivo, @sternberglab.bsky.social, and @harriswang.bsky.social for incredible leadership.
7/9 MetaEdit opens a promising path for editing individual species inside complex microbiomes.
This approach enables new opportunities to probe microbial functions in vivo, elucidate unculturable taxa, and build microbiome-based therapies.
More to come on the horizon. ⚙️🦠🧬💊
6/9 MetaEdit also enabled the first genetic modification of SFB, a key host–immune microbe that is famously hard to culture.
By integrating a GFP payload, we visualized edited SFB filaments, a major step toward uncovering how they shape immune development in the small intestine.
5/9 The result was striking. Edited Bt could now digest inulin with little disruption to the community.
Feeding mice inulin expanded the edited Bt ~30-fold, and removing inulin brought it back down.
A fully reversible, diet-tunable way to control engineered bacteria in vivo.
4/9 With accurate editing in hand, we asked a simple question:
Can we control engineered gut bacteria using diet?
The answer: yes!
We next used MetaEdit to install a 7.5 kb polysaccharide-utilization locus into the genome of native murine Bt in vivo.
3/9 In the complex mouse microbiome, MetaEdit accurately edits Bacteroides species (Bt) with 99.8% specificity across gigabases.
This let us recover engineered strains straight from stool by integrating a selective payload & using only a tiny 32-base target sequence as the tag.
2/9 MetaEdit works by delivering mobile plasmids from a donor E. coli into diverse gut bacteria.
These plasmids encode CRISPR-associated transposases (CAST) to direct RNA-guided insertion of DNA payloads at programmable sites within Gram- and Gram+ members of the gut community.
1/9 Metagenomics lets us read microbiomes in nature without cultivation, but writing (editing) them in their native context is still a major challenge.
Meet MetaEdit: a platform for pathway-scale metagenomic editing inside the gut microbiome. science.org/doi/10.1126/...
Please RT. Post-doc opportunity alert! 💥 closing 10th December.. Come join our team (www.thelowlab.org) at Imperial, London, working on the structure and mechanism of bacterial secretion systems.
For more details and to apply please see
www.imperial.ac.uk/jobs/search-...
10/10 Many thanks to all co-authors for their collective efforts in bringing this story to fruition! And a special thanks to @shsternberg.bsky.social for continually fostering a spirit of curiosity, creativity, and rigor in all aspects of our work.
9/10 Collectively, our findings suggest that telomerase-like activity emerged in an ancient bacterial ancestor, and was co-opted in early organisms with linear genomes to set the stage for the evolution of modern eukaryotes.
8/10 Furthermore, when equipped with the template sequence from the telomerase RNA, DRT10 readily synthesized telomeric DNA repeats.
7/10 We teamed up with RT aficionado @pentamorfico.bsky.social from the Rafa Pinilla-Redondo lab to build a new phylogenetic tree of RTs across all domains of life.
Remarkably, this revealed that the DRT enzymes are bacterial homologs of TERT.
6/10 We found that DRT10 synthesizes tandem-repeat cDNAs through a mechanism strikingly reminiscent of DNA repeat addition by telomerase. But does this similarity represent convergent evolution or shared ancestry?
5/10 Following our recent work on the DRT2 and DRT9 antiviral immune systems in bacteria (linked below), which revealed intricate mechanisms of RNA-templated repetitive DNA synthesis, we began studying a new system, DRT10.
DRT2: science.org/doi/10.1126/...
DRT9: nature.com/articles/s41...
4/10 Telomerase is found in nearly all eukaryotes and acts as a critical safeguard against genome instability from progressive DNA loss. Its aberrant activation is also key to the growth of cancer cells.
And yet, the evolutionary origin of telomerase has long remained unresolved.
3/10 Forty years ago, Carol Greider and Elizabeth Blackburn discovered an enzyme that solves this problem. Telomerase, which comprises a reverse transcriptase (TERT) and RNA (TR), directly extends chromosome ends by adding DNA repeats templated by the TR.
2/10 Linear chromosomes shorten with each round of cell division. Famously known as the “end-replication problem,” this phenomenon eventually leads to cellular senescence and is one of the hallmarks of aging.