Context
The Nucleus PURE genes were synthesized into the pOpen backbone—an high-copy open-source cloning vector. However, cloning vectors are often not good protein expression vectors. High-copy increases the cell burden of protein expression, and pOpen does not contain other common elements of protein expression vectors, such as an on-board copy of the lacI repressor used to reduce expression pre-induction. pET28a is a widely-used protein expression vector, better suited to protein expression in common expression strains.
This page summarizes experiments testing in vivo expression in pOpen vs pET28a.
Takeaway: use pOpen to express genes in vitro; use pET28a to express genes in vivo.
Protein expression from pOpen is inconsistent in expression strains
We designed Nucleus v0.1.0 DNA with pOpen as our plasmid backbone of choice. pOpen is an open source plasmid that was developed by FreeGenes at Stanford University. The plasmid has a number of nice features: standard sequencing sites, dual transcriptional isolation, compatibility with MoClo cloning, and barcodes. Its high copy number makes cloning large quantities of DNA for distribution easier. Most importantly, pOpen is available under the OpenMTA and therefore can be used to distribute our open source DNA.
We adapted pOpen into an expression vector, adding a promoter (pT7) and translation initiation site (UTR1). For genes we expected to purify, we also added a lacO repression operator, opening the possibility of direct purification from the distribution plasmid. We’ve found this plasmid clones well, can be purified at high yield, and gives strong protein expression in PURE without linearization. However, as expected based on its adaptation for high-copy cloning, we’ve noticed that that direct protein expression in vivo is inconsistent in E. coli BL21(DE3) and lysY/laqIq expression strains. Some genes express remarkably well, some genes not at all. Some genes express in some experiments and not in others.
We ran a series of experiments to determine whether this expression was due to the high-copy origin in pOpen or whether it was related to the Nucleus transcription unit design. So far, we have determined that both have an effect on consistent expression. While work is underway on designing and building a more appropriate open-source protein expression plasmid, we set out to test expression of the Nucleus PURE genes in a strong expression vector.
We sub-cloned Nucleus coding sequences into pET28a
pET28a is an expression plasmid first developed in 1987 by Studier and colleagues. It has medium copy number per cell (~ 20), Kanamycin resistance, and a full T7 transcription unit (pT7::UTR1::[your gene here]::T7term). pET28a is widely used as a “sane default” expression plasmid.
We decided to clone our PURE genes (A1:D12, E1 on Nucleus v0.1.0 Distribution Plate) into pET28a to see if they expressed more consistently than in pOpen in vivo. We cloned the genes using the MoClo level 0 CDS compatibility of the Nucleus DNA design; building an adapted pET28a with appropriate Golden Gate entry sites and performing a golden gate assembly into the new vector. We expect to publish a writeup for this assembly soon; in the meantime please reach out for more information.
pET28a gives consistent protein expression
We prepared two (2) sets of plasmids (pOpen backbones, labeled “O”; pET28a backbones, labeled “E”), with forty-six (46) members of each set corresponding to PURE genes in the 
DNA Distribution). We transformed these plasmids into BL21(DE3) cells and plated on selective LB + Agar, allowing the transformants to outgrow overnight. The next day, we picked colonies and started overnight outgrowths of each strain in selective LB (500 uL).
The next morning, we backdiluted overnight cultures in fresh, selective LB (1:100; final vol ~500 uL), outgrew to approximately mid log phase (judged by eye; 2.5 hrs at 37C / 250 rpm), induced with IPTG (500 uM) for two hours (2 hrs) and ran those cultures on gels (E-PAGE 48 well 8% Agarose).
pET28a (even-numbered wells; labeled “E”) showed strong bands at the expected molecular weights for most strains tested (38 of 46). Moreover, many of the unclear bands are strong “maybes”, with either a size that makes them difficult to discriminate against the E. coli background or a weaker expression level.
pOpen was substantially less consistent (24 of 46). Qualitatively, pOpen showed weaker bands than pET28a across most conditions.
Use pET28a while we work on pOpen
As we work on sourcing and validating an effective open-source protein expression plasmid, we recommend subcloning the Nucleus PURE genes into pET28a for expression. Having a reliable, open source, OpenMTA expression plasmid would be a boon for everyone in the community. If you have any ideas or learnings, please shoot us an email at build@bnext.bio.
pOpen remains an effective for both cloning and (re)distributing the Nucleus DNA distribution, and as an in vitro expression plasmid.
References
FreeGenes
OpenMTA
pET28a
MoClo