First of all thanks for giving this a look and I'm incredible grateful for anyone trying to help me, since my future may depend on solving this problem.
In short my problem is that I'm trying to clone several 10+ kB constructs, but almost everytime after a transformation I only get plasmids that are too short and appear to be all the same.
It is most definately not a contimination since I repeated it many times and worked as throughly as possible. Experiments I done suggest that the problem is plasmid stability and that somehow specific parts of my plasmids get deleted.
So lets get into this in detail.
I have faced and investigated this problems while working with
1.) Normal cloning of a 3 kB and also a 5 kB Insert in the vector pUWL201 (~7 kB).
2.) Site-specific mutagenesis with phosphorylated primers via blunt end ligation. The construct to be mutated is slightly bigger than 10 kB.
Both constructs mentioned as 1.) were successfully cloned two weeks ago (after around 4 month of work) but I still want to include them since more of the like will follow in the future and I fear I will face similar problems.
Lets first talk about the experiments I've done for identifying the problems with 1.) and which method finally worked. This protocol was performed with both the 3 kB and 5 kB Insert accordingly, so I will just refer to an "insert" in general.:
- PCR worked fine and subsequently was purified via Gel-Electrophoresis and a QIAgen Gel-Extraction kit.
The excision was performed under UV-light. I know this can cause problems but the length of the exposition was very brief and since we don't have an alternative in our lab everyone is using it and it normally never makes any problems.
- Cloning in TopoXL was succesfully performed with the insert. This was done to get easy access to bigger amounts of DNA and also to make sure the enzymatic digestion works properly
- The Insert and also the pUWL201 vector were digested with HindIII and XhoI (buffer2+BSA). For this not the original vector was used but some vector already containing another insert (but still both the HindIII and XhoI cleavage site were present). This was done so one could see both the cut vector and the removed insert on a gel and to properly remove uncut vector (if present). Accordingly the digest of the Insert-TopoXL-construct was seperated on a gel large enough to clearly differentiate between the two sets of bands.
- The Insert was dephosphorylated with Antarctic Phosphatase for 1h at 37°C, following 25 min at 65°C to inactivate the Antarctic Phosphatase.
- As control only vector DNA was incubated with T4 DNA Ligase (and of course T4 buffer) overnight at 16°C and transformed afterwards. When the plates showed no or only 1-2 colonies the sample was used for the next step.
- In a total of 10 uL (containing 1.0 uL T4 buffer, 0.5 uL T4 DNA Ligase) 25 ng of cut vector DNA were ligated with the cut, dephosphorylated insert in a 2:1 ratio (insert to vector) at 16°C over night.
- Subsequently the ligation mixture was dialysed for 1 hour at room temperature.
- 5 uL of the dialysed sample were then transformed through electroporation using 40 uL E. coli TOP10 cells, minimizing physical stress for cells and DNA (e.g. no pipetting up and down to mix cells and DNA) and incubating the cells afterwards with 360 uL LB-Medium at 37°C for 1-1.5 h before putting 30 uL and 300 uL of the sample on different Amp100 LB-plates.
- The transformation yielded, after overnight incubation at 37°C, 10 colonies with the 5 kB Insert (one of which was positive and no cases of uncut vector) and 15 colonies with the 3 kB Insert (one of which was positive and another one which was uncut vector). All other colonies showed shorter plasmids as seen all the different tries before. Both constructs were subsequently sequenced and seem perfectly fine.
Ligation was tested through making 20 uL ligation sample accordingly to what was mentioned above. This was done once with insert+vector and once with insert only. This showed that in both cases the bands of the insert (as control the same amount of non-ligated, cut insert was put on the gel also) faded and a (a little bit smeared) new band was formed at the height of a higher kB number (1 kB marker was used as reference).
Based on this I concluded that my inserts formed dimers as side reactions, which was the reason for dephosphorylating them. Thus I hoped to achieve a higher efficiency in the ligation process.
An accordingly performed experiment with dephosphorylated, cut insert showed that the sample with insert only still featured an intense band of the original insert and almost no dimers. The sample with both insert and vector looked similar but formed a vague smear at a higher position (for this a 30 uL ligation mixture was prepared from which 20 uL were put on the gel, while the remaining 10 uL were used for the transformation which finally yielded both my constructs).
One has too remember that I tried this for around 4 months (while working on other things on the side of course) and each and every other try ended most of the times in plates with (far) more colonies than mentioned above, but yielded only plasmids at around 3 kB on the gel, which weren't cut with KpnI (which was the digestion enzyme I used to varify the identy of my targed construct). This led me at one point to suspect that something happens after the ligation and prior to the plasmid preparation.
When I retransform my construct (or any other construct for this matter) I get a huge number of colonies (which shows my cells are definately competent enough) and never got a single instant where I didn't get the plasmid I was looking for. Lets keep this in mind, I'll come back to this later on again.
Prior to the successful cloning I tried ratios from >6:1 to 1:1 and several different ligation conditions (NEB FAQ suggests lower ratios when trying to ligate bigger inserts and doing the ligation overnight at 16°C)
Now let's talk about what I've done concerning 2.)
- PCR was performed with Phusion-Polymerase, the diluted 10kB construct as template, adding the dephosphorylated primers (and anything else needed) and worked just fine. Purification was performed as mentioned (yielding a good amounts of DNA). After eluating with EB-buffer a 1 hour digest with DpnI was performed to remove remaining methylated template. The digest was performed at 37 °C in buffer 4 and was followed by incubating 25 min at 65°C to inactivate the DpnI.
- Ligation was performed with 8.5 uL DNA, 1.0 uL T4 DNA Buffer and 0.5 uL T4 DNA Ligase at 16°C overnight. (2h at 20°C was also tested at one point)
- Transformation was either performed directly afterwards (using 1 uL undialysed sample) or after dialysis with 5 uL sample (both lead to plates with a sufficiently number of colonies). Again electroporation was used (see above) and after 1-1.5 h of incubation at 37°C, 30 uL and 300uL were put on Chloramphenicol-17 LB-plates (17 instead of 34 because it is a low copy plasmid. The overnight cultures for plasmid preparation contain Cm34 and always grew nicely).
I did this often and prepared, until now, probably at least around 150 if not 200 plasmids. In probably 99% of the cases I get the same plasmid around 4 kB (once in a while I get plasmids which are heavier but still are too small).
The test digestion is performed with HindIII (Buffer2) which should lead to a band at 1273 b and ~9000 b. And here is the kicker which makes me pretty sure that my main problem is the plasmid stability directly after electroporation:
I ALWAYS find the 1273 band. I compared it with a digest of the template and the supposedly 1273 b bands come at EXACTLY the same spot. I tried to sequence it but apparantly the region where my primer should bind is missing.
I did an additional experiment where I took one ~4 kB plasmid, one ~5.5 kb plasmid and the template plasmid and digested it once with EcoRI and once with EcoRV. EcoRI cuts at 5, EcoRV at 4 sites which are predominantely located on the gencluster of this plasmid.
The 4 kB plasmid in both cases only gets linearized. The 5.5 kb plasmid gets linearized by EcoRV and is cut into two bands by EcoRI. One band at around 2250 b is exactly on the same spot as a band from the template-digest while the other one is slightly above 3.0 kB and slightly under a ~3.5 kB band of the template-digest.
In theory I can build a construct which can satisfy my data rather good in two ways.
a.) Just removing the gencluster of interest
b.) cutting it with EcoRI and religating the fragments with 3.5 kB and 2.2 kB. Thus I also eliminate the gencluster, while keeping the resistence
gen. This construct just has a few houndred bases too much in the 3.5 kB fragment (should be around 3.1-3.2 kB).
I took a look in some lab journals of my colleagues and could also see similar findings, but never to this extend (normally most of the times a few of the plasmids were the correct ones, which lead them too neglect any similarities between the correct ones and the wrong ones which also featured some similar bands). This this problem apparantly is nothing new... but if a cloning works nobody is caring about what caused the wrong plasmids prepared alongside the right ones. Only poor souls as myself which face the same disappointing gels again and again, start to look into what causes this problems in the first place. As mentioned I don't have that much of a problem with the efficiency of my transformations as rather some side reactions I can't put my head around. This leads me preparing huge amounts of plasmids like an imbecile, but almost always without any results.
Because of this I tried several changes in the transformation protocol, i.e.:
- Increasing the length of the incubation without antibiotics from 1 h to 1.5 h. This was done to give the cells more time to incorporate and multiply the target construct without any outward stress.
- Using XL1 Blue instead of TOP10 Cells.
All of which lead to the same result: Too small plasmids which all feature a 1273 b band when digested with HindIII.
Still I have no problems with retransforming constructs which are around this size.
I'm starting to get desperate since I have no more ideas of what I can change. I even start thinking that the cloning of 1.) succeeded out of mere coincidence and am afraid that this is the only way to go. Just relentlessy repeating the same experiment over and over again, until fortune shines upon me on one single day...
If any more information is needed about how an experiment is performed, I'll gladly add it. If anything written is incomprehensible, I'll gladly look for other words to describe it. If anyone here is able to help me solve this problem, I'll probably just be very glad myself.