At the moment most of the Reclone Open DNA collection parts (not expression cassettes) are stored in the FreeGenes pOpen_v3 (aka pOpenBuild) vector or pOpen_v4 (aka pInducible).
This doesn’t need to be the default going forward and indeed we know some people are already thinking about recloning (!) some parts into different standard backbones for various reasons.
@osn_scott has been thinking a lot about this and we were just geeking out on a Zoom call about it so I would kindly invite him to share his thoughts!
It would be great to also get feedback from the rest of the community on plans going forward, to what extent they are back-compatible with the current vectors and/or any compelling reasons to change direction now while the collections are relatively small.
The refactoring job will only get bigger the more we expand the collections, so it is good to have this discussion now.
We are trying to get a discussion going on what makes a good storage plasmid for Golden Gate parts. Some of you might have experience with other parts collections and their storage plasmids.
pOpen_v3 is the standard storage plasmid for all parts in FreeGene’s DNA collections including my Open Yeast. I have worked with many parts from the various FreeGenes collections and have talked with various people who have also been having problems with, in particular, plasmid yields of specific parts.
In addition, last summer I set about cloning into pOpen_v3, in addition to a bunch of other DNA fragments, the three CDSs for the BsaI restriction system - the restriction enzyme and it’s two methyltransferases as separate CDSs. Their synthesis was done at IDT via the iGEM Engineering committee. All parts (16 in total) appeared to clone okay except the M1 methyltransferase. They all were perfect by Whole Plasmid Sequencing (WPS) except: The first clone of the BsaI restriction enzyme came back with a single bp deletion resulting in a dead enzyme! I then sent the remaining 5 clones and each of them had different mutations! All but one were single bp deletions leading to reading frame errors. The missense residue put a proline near what appears to be the catalytic site in the alphafold model. There should be no promoter & RBS in pOpen_v3 that causes expression any of these parts. Of course, a restriction enzyme expressed in a cell, even at low levels, without it’s methyltransfrases is not a good thing. The genes ATG is proximal to the beta lactamase promoter in pOpen_v3 and there is a tonB terminator between the promoter and cargo area. This terminators is not particularly strong. My hypothesis is that there is a cryptic promoter driving low level transcription and E. coli is inducing/selecting for dead genes.
I am building a ccdB dropout construct that has pOpen_v3’s AarI sites to permit migration of parts away from pOpen_v3 but also to be used for new parts. I’ve added PacI & SpeI sites to allow cloning into any SEVA (has oriT) /BASIC_SEVA (no oriT) vector. These vectors have no MTA attached and have strong terminators flanking the cargo area. I’ve been using BASIC_SEVA as destination/receiving vectors that I developed for GGA with Open Yeast.
My question is what characteristics are needed for the best storage plasmids? I have not yet seen a publication just focusing on storage plasmids. Obviously, no transcription of the cargo area where the parts are located. High plasmid yields when extracting plasmid DNA is important. Anything else? Any incites from other (non-FreeGenes) storage plasmids that could be useful?
hello
in my case, which is extremely rare
we keep always 5 plasmid minipreps stored in different places
and we keep 3 strains containing the plasmids also in different places, so i think creating a network for storage will serve preserving the plasmids physically.
as for golden gate, i think putting multiple cloning sites with different restriction enzymes could serve as a safety precaution for having shifts when inserting a gene or a part in the plasmid.
i managed to find some resources that could be helpful i hope:
tl;dr for anyone who missed the first email:
all of our parts collections are currently in [pOpen_v3] (aka pBuild)(Freegenes > pBuild | BBF10K_003498) either as unflanked parts that need to be PCR’d out for your assembly method of choice, or with BsaI sites and the relevant Reclone syntax overhangs. The iGEM distribution is also now currently in pOpen_v3. Some users have reported issues with assembly efficiency, point mutations, inconsistent miniprep yields. pOpen_v4 appears to be a better choice in the hands of some people, but we don’t have much data.
Scott posed the question
My question is what characteristics are needed for the best storage plasmids? I have not yet seen a publication just focusing on storage plasmids. Obviously, no transcription of the cargo area where the parts are located. High plasmid yields when extracting plasmid DNA is important. Anything else? Any incites from other (non-FreeGenes) storage plasmids that could be useful?
One of the issues he raised is that in this and many vectors (both expression vectors needing tight control and storage vectors) there is potential for cryptic promoters. We have run all of the open vectors (pOpen_v3, pOpen_v4 and pTi) through some recent prediction systems and will post the results here.
So if you have used these vectors and have opinions, ideally backed by data, please post here so we can make improvements and decisions about next steps for the most recent rounds of synthesis and vector onboarding.
Perhaps not immediately super helpful, but I want to echo the issues with pOpen_v3 while at iGEM. I don’t have hard data on hand, but will go through any notes I have for specifics. I did leave notes behind that this backbone should be switched out in the future for the distribution. Similarly, I think FreeGenes was planning on moving away from pOpen_v3 to pOpen_v4, perhaps for similar issues.
For the distribution, anything synthesized by iGEM was in pSB1C3 or derivatives.
Happy to share the reasons for using pSB1C3 in more detail, but most importantly and broadly:
long history of use within the iGEM community
high copy
an antibiotic selection (chloramphenicol) that wouldn’t interfere with many of the other assembly plasmids we’d seen (often L1s - kanamycin, as an example)
This isn’t advocation to use this backbone, just the reasoning we made at the time.
pSB1C3 was reliable for basic parts, but it did pose issues for composite parts/transcriptional units due to its high copy (including an example attributed to high read through from a promoter/rbs combo)*.
Since iGEM moved towards a kit designed around basic parts I did not address storage of transcriptional units, and instead focused on providing assembly plasmids alternatives. That said the ideas that came up for a replacement or addition were reliable low/medium origin, a SEVA-based, and potentially inducible.
I think there are collection(s) of PhytoBrick parts that also use(d) a derivative of pSB1C3. So groups that use/maintain those may be good to talk with, on what they use now and why.
Edit: I think looking at cryptic promoters is a great idea. I didn’t, but recall a tool that did and looked at pSB1C3.
A small group of us (@av600@osn_scott, Keoni Gandall who put together the plasmid at Free Genes and I) met and discussed this, basic outcomes of the discussion:
The variable copy number yield and issues with cloning slightly toxic parts are likely the same issue with read through, if we see improvements after removing cryptic sites that will support this hypothesis.
No-one has done any benchmarking with pOpen_v3 against other vectors, so we don’t really know the extent of the issue.
Next steps:
@av600 has run the plasmid through a cryptic transcription site tool and is analysing the outputs, more to follow.
@av600 will synthesis versions without the strongest predicted start sites, to see if this offers some improvements (also for expression vectors pOBL and pTi).
we should attempt at least some basic benchmarking. We have pSB1C3, pOpen_v3, pOPen_v4 and a new plasmid design from @osn_scott in our sights. Please should if you have a favourite DNA part cloning/storage vector to include in the mix.
we don’t have a lot of time, so the current plan is to do the test with BsaI CDS which we have all had issues with before in pOpen_v3 and is demonstrably a toxic part, to see which of the unmodified and modified backbones performs best. Most likely this will involve attempting to insert the part into each backbone simultaneously, then testing assembly and transformation and mutations in assembled plasmids. Obviously a panel of genes would be better, but realistically I don’t think we have the bandwidth in my lab at the moment.
Is there anyone who has some time (or can volunteer someone in their lab) to help with either testing BsaI only or a small panel of parrt inserts in the next 6-8 weeks? We can send you DNA! @av600 has quite a lot to do in parallel so any support would be most appreciated!