• Create Account

Help with 3-way ligation

11 replies to this topic

#1 PhilS

PhilS

member

• Active Members
• 17 posts
0
Neutral

Posted 24 February 2009 - 12:34 PM

Hi,
I've spent far too long preparing 3 bits of DNA that I now want to combine together in a 3-way ligation.  I've tried this before and failed multiple times, so this time I've been extra careful with everything and now I'd like some help with the molar ratios before I continue.

Here are the three things that I want to ligate together:

1) 58bp annealed oligos insert with EcorI and SpeI sticky ends.  Concentration is  160ng/uL
2) 1700bp insert fragment with XbaI and PstI sticky ends.  Concentration is 8.4ng/uL
3) 2700bp vector fragment with EcorI and PstI sticky ends.  Concentration is 6.6ng/uL

So they should ligate and circularize together like this:  E-58--S--X-1700bp-P-P-2700bp-E (because X and S leave compatible sticky ends).

The concentrations of the larger insert and the vector are quite low, I only have 9uL of each.  In the past, on the advise of someone else who has done this, I've tried and failed by adding to my 20ul total ligation equal amounts of each of the three, to a final volume of 3ng/uL.  So 60ng of each.   If I do that again this time, I will use up all of my DNA in one ligation reaction and if this doesn't work then I'll have to start again.

So as I'm trying to be extra careful this time, I wanted to ask for some help calculating the molar ratios of the three.  I have the formula for calculating molar ratios for vector-insert 2-way ligations, but I can't work out what to do to use this formula for 3-way ligations.  I think that I want an equal number of free ends, or that I want a 3:1 ratio of large insert to vector, but what about the ratio for the 58bp annealed oligos?  Ideally I'd like to try a bunch of different ratios, but I'm worried that my low DNA concentration will only allow me to do one ligation mix.  Or can I get away with using less DNA in the ligation?  Or should I start again and get better at purifying my DNA to get higher concentrations?

Any help much appreciated.

Thanks,
Phil

#2 perneseblue

perneseblue

Unlimited ligation works!

• Global Moderators
• 530 posts
8
Neutral

Posted 24 February 2009 - 01:34 PM

ideally for 3-way ligation (and all other multiway ligation) you want a 1:1:1 molar ratio of your insert to your vector. The normal 3:1 insert:vector ratio has a negative impact on ligation efficiency on multiway ligation, getting significant once you start doing 4-way ligation and higher.

so according to my calculation
you will need to make a 1:1000 dilution of the 58bp (EcorI and SpeI)

as for the for ligation formulation
3.6 ul of 1700bp insert (XbaI / PstI) 8.4ng/uL
7.3 ul of 2700bp vector (EcoRI / PstI) 6.6ng/uL
6.5 ul of 58bp insert (EcoRI and SpeI)  0.16ng/ul - diluted sample
2 ul T4 ligase buffer
0.5ul T4 ligase
(Don't use quick ligase buffer)

To quickly calculate molar concentration I use pseudoM, where the MW of the DNA strand is replaced by bp.

1700bp at 8.4ng/ul
pseudoM
= (8.4 ng/ul) / 1700bp
= 4.94 x 10^-3  pseM

My only concern is 58bp insert.... in two way ligations there is a problem of getting short oligoes to ligate to a vector. I am uncertain if this should prove to be a problem.

Make more insert? Or cut the DNA usage per ligation mix in half?

I would.... not bet all my money on this one ligation. I would do this one ligation (overnight ligation at 16 C) and transform it into the cells. I would then look at number of colonies recovered. I would check them by colony PCR (about 96 colonies since I am uncertain about the 58bp insert). If the right clones is not recovered, then I would cut more vector.

Meanwhile, there are probably other things beside this plasmid that is in need of your attention. However if this plasmid is holding up the train... then I would cut more vector as a precaution.

Could you please tell me about the failed ligations. What kind of ratio where used? Aside from the equal mass ligation.

#3 PhilS

PhilS

member

• Active Members
• 17 posts
0
Neutral

Posted 24 February 2009 - 02:33 PM

Many thanks perneseblue,
But please could you explain what pseudoM is, google returns blank, is it a rough way to calculate the number of molecules?   I see your calculation, but what is the significance of the 4.94 x 10^-3 pseM?  What do you do with this figure, it doesn't match any of the amounts you suggest to add to the ligation reaction?

Also, would you mind telling me why quick ligase buffer is not good for multi way ligations please? I see you've suggested it before, but I can't see why, is it the PEG that inhibits ligation?

The ligations that I tried before failed due to excess background, I figured out that this was due to non-efficient cutting rather than ligation problems, this time I gel extracted the exact bands that I wanted to reduce background from uncut vector.  I only tried the equal mass ratio before.

thanks again,
Phil

#4 perneseblue

perneseblue

Unlimited ligation works!

• Global Moderators
• 530 posts
8
Neutral

Posted 24 February 2009 - 05:07 PM

MW of DNA strand 1700bp long assuming an AT content of 50% = 1700*607.4
607.4 is the average MW of a base pair.

This calculation is an extra step.  And when you are look at ratio, it doesn't not influence the ratio value. So to simplify thing, I use bp only to calculate the ratio. But the value I am using isn't actually Molarity, even though I treat it the same. So to remind myself I use the term pseudoM.

insert (XbaI / PstI) = (8.4ng/uL) / 1700bp = 4.94 x 10^-3 pseM
vector (EcoRI / PstI) = (6.6ng/uL) / 2700bp = 2.44 x 10^-3 pseM
insert (EcoRI and SpeI) = (160ng/ul) / 58bp = 2.76 pseM

insert (EcoRI and SpeI) is far too concentrated. A dilution of 1:1000 should bring it into the same order of magnitude with the others

insert (XbaI / PstI) = (8.4ng/uL) / 1700bp = 4.94 x 10^-3 pseM
vector (EcoRI / PstI) = (6.6ng/uL) / 2700bp = 2.44 x 10^-3 pseM
insert (EcoRI and SpeI) = (0.160ng/ul) / 58bp = 2.76 x 10^-3 pseM

identify the most concentrate solution, as you don't want to be making dilution as it is extra work, or trying to concentrate DNA solution. To get a 1:1 ratio you need x time more volume than the most concentrated solution.

insert (XbaI / PstI) = 4.94 x 10^-3 pseM / 4.94 x 10^-3 pseM = 1
vector (EcoRI / PstI) = 4.94 x 10^-3 pseM / 2.44 x 10^-3 pseM = 2.02
insert (EcoRI and SpeI) = 4.94 x 10^-3 pseM / 2.76 x 10^-3 pseM = 1.79

So 1:1:1 mol ratio, I need
1ul of insert (XbaI / PstI)
2.02 ul of vector (EcoRI / PstI)
1.79ul of insert (EcoRI and SpeI) at 1:1000 dilution.
= total volume = 4.81ul

total ligation volume = 20ul, where 2ul is buffer, 0.5ul T4 ligase. Volume for DNA fragments = 17.5ul
Thus the maximum volume of each component added = 17.5ul / 4.81ul = 3.6x

insert (XbaI / PstI) = 1ul * 3.6 = 3.6ul
vector (EcoRI / PstI) = 2.02ul * 3.6 ~  7.3 ul
insert (EcoRI and SpeI) = 1.79ul * 3.6 ~ 6.5 ul

PhilS, on Feb 24 2009, 02:33 PM, said:

Also, would you mind telling me why quick ligase buffer is not good for multi way ligations please? I see you've suggested it before, but I can't see why, is it the PEG that inhibits ligation?

I am sorry Phil, I don't know why. It was just a costly observation I made. I don't know the mechanism but I do know PEG is the cause.

Old papers I read speculates that PEG is crowding agent, which reduces the effective volume of the reaction. this increases the concentration of the DNA fragments to increase ligation efficiency. This speculation has since become the "standard explanation" used. I have not seen any paper that actually did work to see if this idea is true.

PEG does reduce total ligation efficiency, however its effect on 2 way ligation is slight. It is only noticeable when working with large BAC insert (~70kb) and even then it is acceptable for the increase in speed. But when it comes to multiway ligation (3, 4, 5) it is horrible. At 3 ways, failure to recover the desired colony in a 72 colony screen is about 50%. At 4 way, the efficiency is so low, failure to recover a the desired colony is the normal outcome.

Try this. See how it goes. I am just concern about the oligo.

#5 swanny

swanny

Veteran

• Active Members
• 367 posts
4
Neutral

Posted 24 February 2009 - 06:19 PM

About PEG as a crowding agent:
Has anyone done studies using BSA in ligations? I mean, it's often used in restriction digests, and the received wisdom is that it acts as a crowding agent.
Heart disease kills more women than breast cancer, but heart attack symptoms differ from men's symptoms. Get to know your heart... it could save your life.

#6 perneseblue

perneseblue

Unlimited ligation works!

• Global Moderators
• 530 posts
8
Neutral

Posted 25 February 2009 - 01:54 AM

swanny, on Feb 24 2009, 06:19 PM, said:

About PEG as a crowding agent:
Has anyone done studies using BSA in ligations? I mean, it's often used in restriction digests, and the received wisdom is that it acts as a crowding agent.

this is quite an interesting idea.

#7 PhilS

PhilS

member

• Active Members
• 17 posts
0
Neutral

Posted 26 February 2009 - 12:27 PM

Many thanks perneseblue,
I will let you know how it goes.  btw, why do you think I need to do colony PCR on 96 colonies, that seems like it might be overkill to me, as there -should- be only one way for the products to ligate together no?  I know that you are worried about multiple inserts of the small 58bp piece, but is a gel going to be able to differentiate between 58bp?  If you think yes then do you suggest doing the colony PCR's and then looking for the smallest of a group of bands that look about the right size and sequencing that one?
Thanks,
Phil

Edited by PhilS, 26 February 2009 - 12:27 PM.

#8 perneseblue

perneseblue

Unlimited ligation works!

• Global Moderators
• 530 posts
8
Neutral

Posted 26 February 2009 - 01:45 PM

PhilS, on Feb 26 2009, 12:27 PM, said:

Many thanks perneseblue,
I will let you know how it goes.  btw, why do you think I need to do colony PCR on 96 colonies, that seems like it might be overkill to me, as there -should- be only one way for the products to ligate together no?  I know that you are worried about multiple inserts of the small 58bp piece, but is a gel going to be able to differentiate between 58bp?  If you think yes then do you suggest doing the colony PCR's and then looking for the smallest of a group of bands that look about the right size and sequencing that one?
Thanks,
Phil

96... well... I use a multichannel pipette. Checking 12 colonies isn't much different from checking 96. It is a very handy thing to have. However if one did not have a multichannel pipette, the number would have be scaled down.

While there should only be one structure... this is biology. The ligase appears to be able to ligate complementary ends that have skipped a few bp (creating a lip).

As for the colony PCR, it is important that you amplify across the junction between insert and vector. Or in this case between two different inserts. PCR is sensitive enough to amplify left over DNA (from the ligation reaction) on the plate. While you are conduction a 3 way ligation, you only need to check one junction. If the junction exist, very often the rest of the plasmid is correct.

Thus I suggest that colony PCR be used, which amplifies across the junction between insert and vector. Should a positive colonies be detected by PCR. These colonies (at least 2) should be grown up and checked by other methods such as restriction digest and sequencing. You  need to check at least two  (I grow up 3) for the rare situation that one plasmid is structurally wrong or contains mutations.

#9 PhilS

PhilS

member

• Active Members
• 17 posts
0
Neutral

Posted 26 February 2009 - 08:09 PM

Many thanks, I'll let you know
Phil

#10 phage434

phage434

Veteran

• Global Moderators
• 1,806 posts
130
Excellent

Posted 27 February 2009 - 06:02 PM

I would phosphorylate the annealed fragment (or order the oligos phosphorylated).  In principle the ligation might work without this (repair happens in the cell), but I think you are asking for trouble relying on that.

#11 PhilS

PhilS

member

• Active Members
• 17 posts
0
Neutral

Posted 15 March 2009 - 06:08 PM

phage434, on Feb 27 2009, 07:02 PM, said:

I would phosphorylate the annealed fragment (or order the oligos phosphorylated).  In principle the ligation might work without this (repair happens in the cell), but I think you are asking for trouble relying on that.

sorry for delay...
Yes thanks for the tip, the smallest insert was made from annealed oligos and they were ordered with 5'phosphates on them.
phil

#12 PhilS

PhilS

member

• Active Members
• 17 posts
0
Neutral

Posted 15 March 2009 - 06:15 PM

So just an update on the 3-way ligation for perneseblue et al...
I didn't get any colonies at all after trying the 3-way with your suggested ratios.  So I went to plan B which was to cut more vector as you suggested.  It's taken so long because I had some problems with some old LB, and also realized that the plasmid is a low copy one, so I'm having to optimize the minipreps for this to increase the yield.  I'll try again with more vector once I have it.  And working on another experiment in parallel now..
cheers,
Phil

perneseblue, on Feb 26 2009, 02:45 PM, said:

PhilS, on Feb 26 2009, 12:27 PM, said:

Many thanks perneseblue,
I will let you know how it goes.  btw, why do you think I need to do colony PCR on 96 colonies, that seems like it might be overkill to me, as there -should- be only one way for the products to ligate together no?  I know that you are worried about multiple inserts of the small 58bp piece, but is a gel going to be able to differentiate between 58bp?  If you think yes then do you suggest doing the colony PCR's and then looking for the smallest of a group of bands that look about the right size and sequencing that one?
Thanks,
Phil

96... well... I use a multichannel pipette. Checking 12 colonies isn't much different from checking 96. It is a very handy thing to have. However if one did not have a multichannel pipette, the number would have be scaled down.

While there should only be one structure... this is biology. The ligase appears to be able to ligate complementary ends that have skipped a few bp (creating a lip).

As for the colony PCR, it is important that you amplify across the junction between insert and vector. Or in this case between two different inserts. PCR is sensitive enough to amplify left over DNA (from the ligation reaction) on the plate. While you are conduction a 3 way ligation, you only need to check one junction. If the junction exist, very often the rest of the plasmid is correct.

Thus I suggest that colony PCR be used, which amplifies across the junction between insert and vector. Should a positive colonies be detected by PCR. These colonies (at least 2) should be grown up and checked by other methods such as restriction digest and sequencing. You  need to check at least two  (I grow up 3) for the rare situation that one plasmid is structurally wrong or contains mutations.