seemingly easy ligations don't work - (Sep/19/2005 )

Hi everybody,
I'm trying to get some ligations working for some weeks now, but I don't get my hands onto it...
First one is a cosmid subcloning, first I had to enter my cosmid in other bacteria to get some digestion at all, but still I don't get a lot of material. So I cut the cosmid with NotI/SphI (1 hour than add some more SphI and digest for one more hour as the buffer is not the correct one for SphI, it is only 75% active) and purify the band I want on gel followed by qiagen gel extraction. The inserts size is 5.9kb. This has to go into pBS cut with NotI/NspI, also gel purified/qiagen extracted. The vectors size is 2.5 kb. Of course I have a lot more vector than insert, so I ligated them overnight with 80µl (!) of insert, 0.5µl vector, and 1µl T4 ligase and added the appropriate amount of buffer and also ATP (as the buffer is already quite old, but in aliquots so it hasn't undergone multiple freeze-thaw cycles). I then transformed 75µl of XL10 ultracompetent cells with 7.5µl mixture and all my clones seem to be empty, as I thought after colony PCR for the insert (which works on the cosmid on its own)... Are my volumes far too high ? What could have gone wrong?

Second construction should be very easy, two fragments cloned in a TA vector, sequenced, bring fragment B in vector with fragment A... with BsrGI and XhoI for the vector and BsrGI/SalI for the insert. Here, my insert is 13 times smaller than the vector, so again I use a "lot" of insert (20µl) against 1.5µl of vector and the same recipe for the ligation, o/n. Here, I got a PCR signal but a digest of the minipreps showns me not at all what I expect....

I tried to have 3 times more insert than vector, with the concentrations as I think they are by comparison with a ladder (that is something difficult ). I think I had roughly 100ng of insert + vector in the ligation mixture.

I'm getting very frustrated, the vectors are dephosphorylated with antarctic phosphatase and I digested the ligation product with an enzyme cutting the self-closed vector but not the wanted construct... (a tric I found on a forum). Somebody has another idea of getting a lot of cosmid material? Should I speedvac my inserts to get them more concentrated? Should I try a quick ligation kit or should I ligate for some days? You see, a lot of questions....

Thanks anyway for any help!!!
inroe

A lot of information for once- makes a nice change from "my experiment didn't work what went wrong?" posts. You can concentrate you insert by ETOH precitation using glycogen to prevent loss of the dilute DNA (use 1µl of 20mg/ml stock).

A couple of extra questions:

1. You mention the volumes you use. What is the concentration of the various DNAs?

2. How close together of the NotI and NspI sites on your vector. If they are close you wont be able to separate vector cut once from vector cut twice by gel purification.

Daniel

DNA sequencing

OK here we go
*vector(2.5kb): around 20ng/µl which has to get the insert of 5.9kb which has a concentration of roughly1ng/µl I guess (I am very bad at determining concentrations, I find it difficult to compare the intensity of different bands) so here I used 80µl of insert to 0.5µl of vector
*vector (5kb): around 15ng/ band and insert (400pb) around 2ng/µl, I used 20µl of insert and 1.5µL of vector

I tried as much as possible to get 3 times more molecules of the insert than of the vector, as it are sticky end clonings. I am anyway redoing the digests for the inserts so that I can elute the products in water instead of elution buffer, then I can speedvac them to concentrate (now I would get 10x Tris HCl too)... this way I don't loose any material which I would if just EtOH precipitating (before I always used water, I don't know why I changed to EB, it's complicating my life right now!)

=> I will look up the ethanol precipitation with glycogen, I don't know this method and hope somebody in the institute has glycogen

=> for the restriction sites for the first construction: they're far apart, it cuts out the LacZ promotor (and I didn't realize yet I can use this to select my colonies on LacZ plates, they should stay very very white )

=> for the restriction sites for the second construction: there in fact they are quite close, impossible to see the difference on the gel. But I cut the other plasmid with the same enzymes and they work there... I don't think steric hindrance matters as I did the digests sequentially. Maybe I should run the vector again next to uncut plasmid to see whether there is any difference, which one would expect (as I have enough of it, that's not a big thing to do)


Thanks for the help yet and I'll let you know how it goes further

If the cut sites are too close, then the second enzyme will not cut, because it will no longer have a DS overhang. There is a reason to use Tris buffer to elute your DNA -- it elutes better, and it protects your DNA. Most DI water is slightly acidic (dissolved CO2), and can cause problems. Ethanol precipitation is the way to go for concentrating your DNA, rather than a speedvac. You can mix the correct amounts of vector and insert and precipitate the mixture. Redissolve directly into the ligation buffer.

Pet peeve: the word is LOSE, not LOOSE. Loose means not tight. Lose means you can't find it. Sorry to be pedantic.

52 bases apart should be far enough so that both RE site cut, but what can happen is you get some vector that is only cut by one of the RE. As the vector cut with one RE is only 52 bases longer, you can't separate the two bands in the gel purification. The once-cut vector then religates and you end up with lots of empty clones. This is a particular problem when using "troublesome" RE like NotI.

The solution is to SAP treat your vector before gel purification. This prevents the once cut vector from religating.

Daniel

Save on BigDye mix

Hi there,

I totally understand your flustration. Though I am not the expert of clone, I 've overcome some similar problem before. So just try to give some suggestions:
1. comparing to the volume of ligation reaction, the insert:vector molar ratio (3:1) is much less important. I 'd rather use a (1:1) ratio in 20ul ligation volume
2. you mentioned that you 've gel-purifed your insert and vector, but have you heat-inactivated your enzyme after restriction digestion and dephosphorylation. Many reports demonstrated that gel-purification cannot remove your enzyme from the DNA which result in failure ligation. That 's actually a very common mistake but very crucial to your successful ligation.
3. for quantitating low DNA concentration solution, you may use pico-green fluorescence kit from molecular probe.
4. since you mentioned you used the XL-ultracomp cells, I will try using small amount of vector and insert (so that you don't need to concentrate your insert) with total amount of 10-20ng DNA in a ligation reaction.

wish it will help

OK here we go
*vector(2.5kb): around 20ng/µl which has to get the insert of 5.9kb which has a concentration of roughly1ng/µl I guess (I am very bad at determining concentrations, I find it difficult to compare the intensity of different bands) so here I used 80µl of insert to 0.5µl of vector
*vector (5kb): around 15ng/ band and insert (400pb) around 2ng/µl, I used 20µl of insert and 1.5µL of vector

I tried as much as possible to get 3 times more molecules of the insert than of the vector, as it are sticky end clonings. I am anyway redoing the digests for the inserts so that I can elute the products in water instead of elution buffer, then I can speedvac them to concentrate (now I would get 10x Tris HCl too)... this way I don't loose any material which I would if just EtOH precipitating (before I always used water, I don't know why I changed to EB, it's complicating my life right now!)

=> I will look up the ethanol precipitation with glycogen, I don't know this method and hope somebody in the institute has glycogen

=> for the restriction sites for the first construction: they're far apart, it cuts out the LacZ promotor (and I didn't realize yet I can use this to select my colonies on LacZ plates, they should stay very very white )

=> for the restriction sites for the second construction: there in fact they are quite close, impossible to see the difference on the gel. But I cut the other plasmid with the same enzymes and they work there... I don't think steric hindrance matters as I did the digests sequentially. Maybe I should run the vector again next to uncut plasmid to see whether there is any difference, which one would expect (as I have enough of it, that's not a big thing to do)


Thanks for the help yet and I'll let you know how it goes further

I'd be interested in published reports on the effect of purifying DNA from restriction enzymes on ligation efficiency. All I've heard is hearsay, but I've got an open mind. Is it clear that heat killing the enzymes is effective in solving this problem? How effective is gel purification in solving this?

Hi there,
still no results....
so I redid it with all the controls, even the pBS SK- double digest fragments, trying to ligate both fragments together.... nothing!
So yesterday it tested the new ligase we ordered on linearized vector (NotI). I used 100ng of vector in a 30µl volume, as said by R*oche in its manual. I transformed into DH5alfa but nothing, where my control transformation gives me plenty of colonies (with uncut pBS). Now I'm really running out of ideas. The only thing might be that the UV is really damaging my DNA very badly, but how to select the good bands then if I cannot use UV? Or is it time to try a non-UV method (I still have a sample somewhere in the fridge...)

Does somebody have an idea?

Thanks,
inroe

I've been reading through other posts. I use home-made competent cells, they work fine for the uncut vector (although the pfu is rather 10 +7 than 10 +8), and if I'm right if ligation works, the linearized re-ligated vector should give me nearly as many colonies as the uncut,
And 20ngs of ligated DNA should work, isn't it? The pfu determination was done with 10pg of pUC18....

inroe