2026

Future work#

First steps#

Three steps precede the first evolution cycle. The reasoning is given under Results.

Confirm cage assembly. The His-tag is removed and the targeting peptide is kept. DLS on wild-type cages provides the reference for 42 nm. Assembly is confirmed by negative-stain TEM or cryo-EM, which does not depend on elution volume or native-gel migration. Expressing monomers with and without the external moiety separately and mixing them in defined ratios gives mosaic cages, a route to tolerating external fusions.

Complete the selection plasmids. Leaky dCas9 and sgRNA expression is the likely burden during assembly. Options are chemical repression of the promoters, a non-targeting guide as a proof of concept, and a tri-plasmid system that separates the selection, mutagenesis and repressor components so that the regulators are present before the selection components. Removing the GATA/GAAA overhang clash between junctions J4 and J5 is also required (Table S2).

Reduce leaky MutaT7 activity. Reversion in the stop-codon assay was similar with and without induction. An inducible sgRNA directed at the T7 promoter region would suppress mutagenesis outside the evolution windows.

The co-encapsulation assays then separate single-handle capture from true co-encapsulation: dual-stained native PAGE, RNase challenge and SEC co-elution of a fluorescent cargo (protocols).

Planned campaign#

Cultures carry both plasmids in the MutaT7 host under dual selection. Induction starts hypermutation of the T7-flanked cassette only. Glucose in the growth medium before induction suppresses leaky expression. The MutaT7 strain transforms poorly, so competent-cell aliquots are 200 µL instead of 100 µL. Selection stringency follows the two settings described in Mechanism: guide length for the baseline and kanamycin, from 50 µg/mL upward, for the stepwise increase. Nitrotetrazolium blue plates provide a colorimetric viability readout.

Culture is continuous in a turbidostat (Pioreactor), so selection acts on growth rate in every generation. At each passage:

  1. Sample and archive as a glycerol stock, so that any passage can be recovered.
  2. Sequence the mutagenised cassette to follow which mutations are enriched and when.
  3. Re-sequence the selection plasmid to detect escape: loss of function in dCas9, the sgRNA or their promoters restores resistance without improving encapsulation.
  4. Set the next stringency step from the growth rate at the current dose.

An enriched allele is a candidate. It is confirmed after re-cloning into the expression vector and independent characterisation: assembly by SEC, DLS and EM, capture by the co-encapsulation assays, and specificity by the single-handle controls.

Broader cargo range#

The λN·boxB handle is not specific to a guide RNA. Any transcript carrying a boxB hairpin can be captured, so a shell evolved for guide capture could be tested for mRNA packaging. The selection can be repeated with the hairpin moved from the sgRNA scaffold to an unrelated transcript, to test whether the evolved shell still captures it. Evolution under a single selection can find a solution specific to the geometry of the dCas9 complex, so transfer to other transcripts is not guaranteed.

Additional selections#

The OR gate makes the output ambiguous as to which handle is used. Two additions address this.

AND gate. Once shells exist that capture either component, the handles can be split across two selective agents so that both must be sequestered simultaneously.

Nuclease challenge. Tetter and co-workers increased stringency by shortening the nuclease and lengthening the exposure: benzonase, then RNase A, then RNase T1, from one to four hours. This selects for protection of the cargo, which is a stricter property than the binding measured by the current circuit.

Escape#

Enrichment of cells that restored resistance without improving encapsulation is the main failure mode. Two measures are already included: dCas9 and the sgRNA are kept off the mutagenised plasmid, and the selection plasmid is re-sequenced at each passage. A counter-selection that periodically requires a functional repressor would also detect populations that have lost it.

Delivery#

The aim is a shell that packages a ribonucleoprotein and delivers it to the cytosol of a mammalian cell. For this scaffold, cargo release in the endosome using a pH-sensitive intein and endosomal escape using a fusogenic peptide have been demonstrated separately. This is compatible with the shell evolved here, provided that the shell assembles and loads cargo.