I have been trying to insert a short sequence of DNA into a plasmid. I wanted to make two different versions, with the insert at different locations. I chose HinDIII and SalI sites approximately opposite each other on the plasmid and designed oligos with HinDIII or SalI sites, and of course extra bases on the ends to allow proper cutting.

I cut the plasmids and oligos and dephosphorylated the vector to help prevent empty vector ligations. I mixed the plasmid and oligos and ligated, then transformed into bacteria, all standard stuff.

The HinDIII version worked the first time, with few colonies on the empty vector control plate, and more colonies on the plates that received oligos during the ligation. I confirmed insertion by using the BpmI cut site the oligos just happened to carry.

However, the SalI version has failed every time I tried it. The empty vector control and all other plates are a just a carpet of cells. There are very few individual colonies, and the ones that exist don't look right, too small. I haven't even tried a miniprep, too likely to get an empty vector back. I did do a nontransformed control, where the cells were not given any DNA and were just directly plated. Those cells died, indicating it's not an issue with my ampicillin.

My hypothesis is that unligated empty vector outnumbers the correctly ligated plasmids, and that cells that pick up a copy of the unligated DNA somehow repair the cut, even though it's dephosphorylated. This is made more likely because I only used a single digest, so the plasmid has matching sticky ends, however the HinDIII version used the same strategy. I have been including a 1 hour 37 C SOC media incubation, which could allow the bacteria more time to carry out this repair. If I skip the incubation, maybe that would prevent unwanted repair.

I am testing this hypothesis with some additional controls, including transforming bacteria with empty vector that hasn't been treated with ligase to make sure the ligation reaction isn't the problem, as well as using the previously ligated DNA without the SOC incubation. I will report that data as it becomes available. I am short on competent cells, so I can't do much else right now.

Also, what do you think about treating the ligated DNA with klenow to blunt any remaining sticky ends before transformation?

EDIT: I have the data for the transformations that skipped the ligation and SOC media incubation. Skipping the incubation did not change anything, neither did skipping the ligation. All plates continue to show a carpet of cells.

  • 1
    $\begingroup$ So your insert has same sites on both ends— a non-directional cloning? Plus, do you see the linearized plasmid on the gel? Do you use gel purification? Also, how much plasmid did you use for digestion and ligation? $\endgroup$ – WYSIWYG Feb 24 '16 at 8:42
  • $\begingroup$ It is indeed nondirectional. I'm just inserting a zinc finger binding site, I don't care about direction, just need to make a zinc finger protein we already have bind to this plasmid. I have confirmed that SalI cuts the plasmid, but have not used gel purification. Since it's a single cut site there would not be extra DNA to remove. I have not run a confirmation gel every time I have tried the experiment either, so hypothetically the enzyme could have stopped working. The oligos I am adding are short, 56 bases, but I suppose a high density polyacrylamide gel might work to see cuts there. $\endgroup$ – user137 Feb 24 '16 at 9:06
  • 1
    $\begingroup$ But you can see the linearized plasmid when you do a gel purification. No need to look for the drop-out. How do you clean up your digestion reaction? Just simple column-purification? There is a likelihood that your enzyme did not cut properly. Bacteria can repair nicks but I am not sure if they preferentially phosphorylate one enzyme site. You can try blunting and see if the problem persists. $\endgroup$ – WYSIWYG Feb 24 '16 at 9:14
  • $\begingroup$ I do phenol chloroform purification and isopropanol precipitation after digestion and dephosphorylation for both vector and oligos. (I don't dephosphorylate the oligos obviously) And when I say I have confirmed the digestion, I mean I did the SalI digest and ran a gel, and it was cut. There was a section of the plasmid where we weren't sure of the sequence and I needed to make sure it didn't contain any SalI sites. I do not run the gel every time I attempt the cloning. $\endgroup$ – user137 Feb 24 '16 at 11:33
  • $\begingroup$ @WYSIWYG I think you were right. I finally got around to trying this again, using gel purification as recommended. The gel revealed that the SalI enzyme was not cutting the plasmid very well, with most of the DNA still in the supercoiled form, and a lot as nicked circles. I carefully cut out the linear DNA in the middle and the control plate had only 1 colony. The problem is that the experimental plates have 2 - 5 colonies each, I'll have to see what I get on miniprep. Also, my gel extractions have been poor. When starting with 4 ug of DNA, I get less than 100 ng of DNA after extraction. $\endgroup$ – user137 Mar 23 '16 at 12:37

As per WYSIWYG's suggestion, I tried the digest and dephosphorylation again with SalI and ran a gel. The gel showed that SalI was less than ideal at cutting this plasmid. enter image description here

As you can see, the most intense band is the bottom band, which is most likely supercoiled DNA, with the linear DNA I want appearing in the middle, and open circle DNA coming in third. I very carefully cut the middle band out and did a gel extraction. When I ligated this the control plate had 1 colony. Unfortunately, the experimental plates had between 2 and 6 colonies each. I am running minipreps on those now, will just have to see what I get.

This explains why the previous attempts produced so many colonies on the control plates, most of the DNA was uncut, so no amount of dephosphorylation would help it.

One thing that does bother me, is that my gel extractions seem like they have had really poor yields. I typically start a digest with 4 ug of plasmid. After phenol:chloroform extraction and isopropanol precipitation, I'll usually recover 2 - 3 or more ug of DNA. After gel extraction, I'm lucky to get 300 ng of DNA. To make things weirder, the liquid that comes off the spin column in the final elution step will gel up, like a low density agarose gel. The only explanation I can come up with is that residual agarose in the spin column is somehow forcing its way through the membrane and into the elution buffer. The goo clogs pipette tips and makes life difficult.

  • $\begingroup$ SalI is a fussy cutter; I avoid it for routine cloning where possible. Since it's cutting poorly you'll need to gel purify, as you found. Be sure you're getting anything cut at all, I'm not convinced by your gel that you are. Use an alternate enzyme if possible. If you're doubting your gel purification, use a commercial product e.g. one of Qiagen's columns (or competitors, I have no particular attachment to Qiagen); they are more expensive but are pretty reliable. Don't waste your time with transformations until you know you have clean linear DNA. Good luck. $\endgroup$ – iayork Apr 23 '16 at 12:28
  • $\begingroup$ This is why I often recommend running a negative control of uncut plasmid in parallel to verify that your identification of the supercoiled band is accurate. $\endgroup$ – March Ho Apr 23 '16 at 23:33
  • $\begingroup$ @iayork I just got my sequencing results back last week, and my "Successful" plasmids have absolutely no difference from the original plasmid. Now that I can order oligos again, I think I'll try a PCR based cloning without using any restriction sites. If it works, I can avoid the SalI issues and I'll have a method to insert the binding site sequence anywhere I want. I couldn't do it earlier due to weird rules with the budget, we couldn't spend any money during the last month of the fiscal year. Now it's a new fiscal year and I can order as many oligos as I want. $\endgroup$ – user137 Apr 24 '16 at 23:46
  • $\begingroup$ Budget issues are exasperating. Can't spend $50 on a primer, but it's OK to waste a week of your time ... $\endgroup$ – iayork Apr 25 '16 at 12:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.