Mechanism#

A cell survives in proportion to how well it packages a repressor into its own encapsulin shells. The rest of the circuit exists to make that true.

The circuit#

A catalytically dead Cas9 (dCas9) is directed by an sgRNA into the kanamycin-resistance gene on the selection plasmid. It obstructs rather than cleaves, and the cell is left unable to make enough resistance protein for the kanamycin in the medium.

Fig 1. dCas9·sgRNA binding the resistance gene on the selection plasmid. Drawn from 6NJ8 and 5F9R; the cell and plasmids are schematic, the molecules are to scale relative to each other.

Why the guide targets the gene body, not the promoter#

CRISPRi knockdowns conventionally target promoters, which represses harder. Harder repression is the wrong objective here. The two placements fail differently:¹

	Target in the promoter	Target inside the ORF
What blocks what	RNA polymerase cannot bind an occupied promoter	RNA polymerase collides with the R-loop and eventually displaces dCas9
Escape route	diffusion only	processive read-through
Repression	near-absolute	partial, "throttled"
Depends on [dCas9]?	yes	no, once the target is saturated
Tunable by guide length?	weakly	yes, continuously

Inside a gene body, complementarity between guide and target sets the probability that RNA polymerase kicks dCas9 off during a transcription attempt, while spontaneous unbinding is negligible. Expression is then

\[ c = c_0\left[1 - P(\text{stop})\,P(\text{bound})\right] \]

and once dCas9 saturates the site, \(P(\text{bound}) \to 1\) and residual output collapses to the passage probability

\[ r = 1 - P(\text{stop}) \]

which depends on guide–target complementarity and on nothing else. Reported values run from \(r = 0.026 \pm 0.003\) at full complementarity to \(0.056 \pm 0.001\) with six mismatches.

A graded readout#

Absolute repression gives a binary live/die outcome and no gradient for selection to climb. Under kinetic repression, marginal improvements in sequestration produce marginal improvements in growth rate, and continuous culture integrates those margins over many generations.

Noise#

Because repression at saturation is independent of dCas9 concentration, cell-to-cell variation in dCas9 level does not propagate to the output. The measured noise plateau — about 0.3, matching ordinary constitutive genes in wild-type E. coli — is flat across the whole knockdown range. An inducer-titrated circuit would be noisiest in exactly the intermediate regime this selection operates in.

The selection therefore reads encapsulation efficiency rather than expression noise. Without that, a quieter promoter would be the cheapest way to win.

Sequestration restores expression#

Fig 2. 240 subunits closing around the repressor and carrying it off the gene. Transcription resumes and the cell survives.

240 encapsulin subunits assemble into a T=4 icosahedral compartment 42 nm across — roughly twice the span of the complex it has to hold. Either half can be caught: a cargo-loading peptide on dCas9, or a boxB hairpin on the sgRNA. See Design for why the OR gate was chosen over demanding both.

The two stringency knobs#

The circuit has one set-point and one ramp.

Guide length sets the set-point. A truncation series — non-targeting, 10, 11, 14, 17 and 20 nt of complementarity — spans a range of passage probabilities. The working guide is the one whose unrescued residual resistance sits just below the survival threshold, so a modest improvement in capture decides whether a cell grows.

Kanamycin concentration is the ramp. Raising it between passages raises the resistance output a cell must reach, and so the fraction of repressor it must sequester. For finer adjustment, the guide is swapped for a longer one instead of, or alongside, raising the dose.

Constructs involved#

The circuit is split across two compatible plasmids — mutation plasmid carrying the MutaT7 machinery and the mutable encapsulin cassette, and selection plasmid carrying dCas9, the sgRNA and the resistance gene. See Plasmids for maps.

Mutation plasmidSelection plasmid

Carries the MutaT7 T7 RNA polymerase–deaminase fusion and the QtEncapsulin open reading frame flanked by a T7 promoter and a T7 terminator. That cassette is the only hypermutated sequence in the cell.

Carries dCas9, the sgRNA cassette and the kanamycin-resistance gene with its targeted site. Built on a pSC101 backbone, outside the T7 transcription unit.

Escape routes to watch

Any mutation that reduces dCas9 or sgRNA expression restores resistance without improving encapsulation, and will be enriched just as strongly. Both are therefore held on the non-mutable plasmid, outside the T7 transcription unit.²

Loss of the shell's own expression is the mirror-image failure, caught by the same monitoring: a population whose resistance has recovered while its encapsulin cassette has acquired a frameshift has escaped, not evolved.

Vigouroux, Oldewurtel, Cui, Bikard & van Teeffelen, Molecular Systems Biology 14, e7899 (2018), doi:10.15252/msb.20177899. The kick-out model, the mismatch titration and the noise measurements below are all theirs. ↩
Placing them outside the MutaT7 target region does not make them immune to host-polymerase errors, only to hypermutation. Periodic re-sequencing of the selection plasmid is part of the passaging protocol. ↩