Notebook#

Condensed from the lab journal kept from 13 May 2026 onward. Two threads run in parallel from late July: shell expression and purification, and construction of the selection circuit.

May — getting started#

Competent cells (DH10β and BL21 (DE3)) prepared, buffer stocks made up, first transformations of the expression backbone.

An early plating went onto kanamycin when the plasmid carried ampicillin resistance; nothing grew, for the correct reason. Gibson assemblies of the first shell constructs followed, and colonies were picked and sent for sequencing.

June — first expression and the first purification#

First large cultures. Induction at 1 mM IPTG, 18 °C overnight.

The Ni-NTA protocol had to be written from scratch: the available Cytiva protocol assumed pre-packed HiTrap columns rather than columns packed in-house, so a hybrid was assembled from that, the published methods section and local practice.

First SDS-PAGE gels showed protein of about the right mass.

July — the assembly problem emerges#

Size exclusion on the first purified material gave four peaks — a very shallow one around 8 mL and substantial material at 16, 18 and 20 mL. The shallow peak is where an assembled cage should elute. DLS on the pooled fractions found nothing of the expected size.

Changes made in response: induction dropped to 0.1 mM IPTG; a 100 kDa Amicon filter introduced to remove monomer and free reporter; aliquots taken at every purification step for gel analysis; clear-native PAGE added to look for the assembled species directly; anti-His immunoblotting added to test whether the tag was accessible at all.

Addgene bacterial stabs for the dCas9 and MutaT7 source plasmids arrived at the end of the month and were streaked out.

August — two threads#

Purification. PEG precipitation and heat precipitation run side by side on split aliquots. Step-by-step A280 and gels located most of the protein in the pellet after lysis. SEC on heat-precipitated material still showed no 8–10 mL peak. DLS of the pooled fractions read as polydisperse and dilute; some individual measurements suggested particles near the right size, but pooling all measurements did not support it.

Circuit construction. dCas9 amplified out of the Addgene plasmid. pSC101 linearised both by restriction digest and by PCR. Seven selection constructs assembled by Gibson — the full guide truncation series plus the non-targeting control.

Not everything worked the first time:

• Early Gibson reactions produced no colonies, while a re-transformation control grew — which located the fault in the assembly rather than the cells or the heat shock.
• The first s002_20nt clones sequenced back as backbone plus an unrelated gBlock.
• m004 and m005 sequenced back as empty backbone.
• MutaT7 competent cells were made and found to transform poorly; aliquot size was increased in response.

By the end of the month the shell expression constructs f008–f011 and the s002 guide series were sequence-confirmed.

September — where it stands#

PCR and gel extraction of the remaining fragments, with repeated gel smearing traced to the reaction volume and run time rather than the products themselves. Kanamycin plates prepared at 50 and 200 µg/mL, and with nitrotetrazolium blue for a colorimetric viability readout, in preparation for the selection screen.

The s002 sequences are confirmed. Assembly of the mutation plasmid series continues.