We have seen this type of thing in my lab many times. Certain DNA sequences are toxic to E. coli, often due to spurious transcription in the bacteria. There is strong selection against clones that carry the correct insert, so you see small colonies and genetic instability.

I can recommend three approaches.

1) Instead of growing in liquid culture to make minipreps, resuspend a colony in a small amount of water and spread the cells on a plate. Then harvest the lawn of cells to make a miniprep. You might also want to try carbenicillin instead of ampicillin to help stabilize the plasmid.

2) To reduce toxicity, use vectors or E. coli strains that result in low copy number plasmids. For example, try the ABLE C and ABLE K strains from Agilent.

3) Check out vectors from Lucigen. I haven't used these vectors myself, but they are specifically designed for the propagation of toxic DNA sequences. You may need to modify these vectors for mammalian expression.

What temperature are you growing the cells? what competent cells are are you using?

We have seen this type of thing in my lab many times. Certain DNA sequences are toxic to E. coli, often due to spurious transcription in the bacteria. There is strong selection against clones that carry the correct insert, so you see small colonies and genetic instability.

I can recommend three approaches.

1) Instead of growing in liquid culture to make minipreps, resuspend a colony in a small amount of water and spread the cells on a plate. Then harvest the lawn of cells to make a miniprep. You might also want to try carbenicillin instead of ampicillin to help stabilize the plasmid.

2) To reduce toxicity, use vectors or E. coli strains that result in low copy number plasmids. For example, try the ABLE C and ABLE K strains from Agilent.

3) Check out vectors from Lucigen. I haven't used these vectors myself, but they are specifically designed for the propagation of toxic DNA sequences. You may need to modify these vectors for mammalian expression.

Sounds like toxicity indeed. Probably even the normal leaky expression is just too much for the cells to handle -- a potent protein.

The Lucigen High Control BL21 or 10G cells have been extremely effective for me in allowing cloning and expression of a highly toxic protein like that. Those cells have over 100 times more LacI repressor than normal, and can basically fully repress leaky expression when grown in the presence of 0.5% glucose. Even LysY carrying cells (inactive lysosyme, which inhibits T7) are not as effective for controlling the toxicity of my protein as the High Controls. Could be worth a shot, if there is a LacI operand in your vectors to allow repression by that method, and they are pretty cheap. Hope you get it solved quick!

Sounds all too familiar... Some inserts are just a pain. Another good reason to swap the Amp(R) cassette in often-used vectors, e.g. for Kan(R) or Zeocin. Eliminates satellite colonies after prolonged growth on plates, plus it keeps plasmids stable in high-density liquid cultures.

About picking colonies after 48-72-96 h (@ Drew): You're right absolutely, it is bizarre. But, you usually do it only when you're in trouble already. Sometimes it's just the only thing that works. We've had this scenario when we tried to clone an insert encoding a large protein with 8 transmembrane domains, into a lentiviral backbone. In DH5a, nothing. In TOP10, the first positive clones came up after 3 days, and were unstable in liquid culture. We could propagate these clones only by repeated streaking on plates, and incubating at room temp.

Miracle: Plasmid instability/toxicity practically disappeared when we used the same LV vector backbone, but with a Gateway destination cassette instead of MCS. Same vector, same insert. Just a little difference in insert-flanking sequences. No logical explanation, but that's how it was.

If nothing else works, think about using a Gateway vector.

Good luck!

I wish I had come here to ask the question before I had so much trouble and hard time. your answers pretty much solved the inexplicable.I will definitely try some of the approaches. Thanks.

BL-21 AI cells have tighter control of expression (Arabinose-induced).

Failing that you can always try a cell-free expression kit.

Does your gene of interest contain an intron? If so, I’d amplify from genomic DNA not cDNA using Pfu, you would then only get expression in mammalian cells not bacteria.

Hi, Xiarong. Did you solve your problem in the end? I am dealing with more or less the same mystery here, only that I did not even have a single colony.

Cloning other genes is perfectly fine in my hands, and cloning with parts of the gene is fine as well. Cloning the long full-length is hell, which is somewhat similar to your c-kit situation.

To Rachel,

I solved the problem by culturing my bacteria at room temperature on the LB plate or in liquid. It takes 2 day to see colonies on the plate or grown up in the liquid for a maxi or mini-prep.

Thanks Xiarong for your reply.

So simply by culturing your bacteria at room temperature solved all the problems? You did not have to change to a different bacterial strain or to a different vector?

To Rachel,

I tried different strains and vectors, which all did not work. Finally I stuck to TOP10 and merely cultured it at R.T. problems solved. It seems that toxicity is inhibited at lower temperature. good luck

Wow, fascinating! I had a similar problem earlier this year. I was working with an undergrad to simply compare two promoters, one containing an intron and one without, driving the same GFP sequence. We could not get one of them to produce colonies, but the other was just fine. After months, I gave up and we changed her project. I never had any problems cloning before and this honestly made me feel like quite a dope. Great info from everyone here! Thanks!

I also experienced problems with cloning of a yeast glucose transporter (HXT7) but not with other HXTs which all have high homology. We use homologous recombination in yeast to assemble plasmids in vivo and then transform yeast-DNA-preparations into E. coli DH5 for amplifikation. I even can select for functional plasmids when constructing the plasmid in a transporter-less yeast strain. But transformation of E. coli, still is a problem. I could solve the problem by incubating E. coli at roomtemperature, with all the above mentioned problems (small colonies after 2-3 days, low plasmid yields)

It would be interesting to know, what exactly the problem is. If its the DNA itself that is toxic, maybe the reason is the HXT7-promotor (I use in all of my constructs) in combination with its native gene? If its expression of the protein, maybe its because its a membrane protein? But then again as I said, other very similar genes are not as toxic. I will try another E. coli strain, because I encounter problems once again cloning HXT7 into another vector.

Best wishes and good luck with your cloning :D

Sounds like your HXT7 promoter is "leaky" in bacteria, I stand by my original advice which was try and include an intron. If your gene of interest’s coding region doesn't contain one, then think about adding a leader intron before the CDS. I find carbenicillin a better selection than amp, same resistance gene, use at half the strength of amp. Good luck.

Any gene that is cloned in a multicopy plasmid will have some low level of expression even in the absence of any promoter. Mark's suggestion of including an intron that prevents translation of the functional protein in E. coli is a good one. On the other hand if low temperature works for you, then stick with that.

I know it's a very old question, but maybe in future somebody else will approach the same problem. So I would suggest to give a try method developed by me: Article A novel method for cloning of coding sequences of highly tox...

Thank you very much!

Hi, I know this is a slightly naive question... but how long does a repeat sequence need to be to make it toxic? And how many repeats are we talking about? I'm using Top10s. Many thanks

Hi, I got exactly the same trouble... I'm trying to clone a protein for a Y2H and I'm trying to clone different truncated versions of this protein. The only one I got it's the shorter version. The four others remain impossible to clone.

The construction of the plasmid by ligation works perfectly (checked by PCR with a primer in my gene and the other in the plasmid) but after the TOP10 transformation, I got a lot of colonies (KanR) but most of the time, they lose my gene... The plasmid is well incorporated but the gene is not longer here OR there is a lot of mutation inside the gene sequence.

So I'll try the low temperature growth (at 30°C and RT) and I'll see if it works for me too.

I am experience same problem as well. it likes my yeast promoter being leaky. I added a GTG in front of start codon of one gene, it works. But now, I have other two genes showing this problem.

I counteract the problem by changing the plasmid used (pGBK for pGAD). Looks like the pGAD plasmid I got my gene at the first try.... Well, problem solved for me 🙂