Hi! I generate my insert and linearized my vector by PCR. For my insert I usedNEB builder tool to generate my oligo with overlaps and used it for my PCR. The overlap is 45 bp.

Hi Julie,

You mention that your total reaction volume is 6.2ul, what do you mean by this? You have 6.2ul of DNA fragments that you are then topping up with water and 10ul of Master Mix? Are you able to run an agarose gel on your Gibson Assembly reaction? to help determine whether the Gibson Assembly reaction or the transformation might be the problem?

I yes I have 6.2µL of DNA fragments and then I put 10 µL of mix and 3.8 µL of water. I run a gel at the moment and I'll tell you what the result is. But I run a PCR yesterday to see if there is an assembly and there is but maybe not enough for the transformation I don't know.

Hi Julie,

Which type of competent cells are you using?I have previously had toxicity issues with the Gibson Assembly Kit and some types of E.coli. If you did pcr to confirm an insert (I use this method as well), then what may be happening in your transformation is that the Gibson Assembly kit is killing all your cells. What is the positive control for the transformation that you are using?

I use NEB E Coli 5 alpha...And the positive control provided wth the kit was ok. I was wrong previously my oligo are about 45 bp but my overlap is between 15 and 20 bp. Is it enough?

Is the positive control puc19?

yes

Yes, 15-25 nt overlap is what the protocol calls for, or more exactly the Tm by 4xGC + 2xAT's should add up to at least 48C or more.

Your construct is big. Regardless of cloning method, it may be difficult to clone, so I think you should try to optimize the reaction and perhaps a different bacteria might help. There are stable bacteria from... (who are they now) Gibco/invitrogen/thermo?) and NEB's NEB10b is supposed to be good for large plasmids (it hasn't helped me yet, but may help?).

You used 0.015 pmol vector + .079 pmol insert, with a ratio of 1:5.3 (I calculate)

They recommend ratio of 1:2-3 and total of .02- .5 pmols. You are using toward the bottom of the range, and considering the construct may be difficult, perhaps more DNA would help. 660ng vector would be 0.1pmol, I think, and 0.2 pmol of insert would be 760 ng. I might aim for that, if possible (and I'd double check the math).

Did you purify the PCR products? That worked better for me (with a limited sample size). Did you verify that the PCR products are good? How are they quantified (some people think they can OD a not purified PCR product).

Thanks a lot for the answer Michael!!

I realize now that I was wrong calculating the ratio. But they said 50 to 100 ng vector isn't 660 ng too much?

I'll try like you said. Yes I did PCR Clean up and then after I quantified with nanodrop. The PCR products are good. Maybe I should also do the DpnI digestion...

Michael's suggestions make a lot of sense Julie, both regarding quantity of DNA and size of insert being transformed into the cells. When you are confirming that your insert assembled to your vector via PCR is one of you primers on your vector and one on your insert?

No they are on the vector but they can amplify only if it's assembled because during my PCR I removed the part between the primer on the vector. 

The dpn digestion is to remove template vector, and avoid extra colonies. That's not the problem (although it's easy enough to add some)

Your reaction should have worked. Your template is big, so 600 ng is like 2 or 300 ng of a smaller vector. You used a normal amount of vector, and the extra insert I don't think should really hurt. Because it's a big construct (and because it didn't work) I suggested using toward the upper end of the amounts of DNA. 

However, you say the gel showed the fragments not going together. I've never tried to look afterward. The insert shouldn't change much since you used a molar excess of that. The vector should have gotten less, with a new bigger band (unless the new circular DNA runs strangely?). So, this suggests the problem is the assembly. You could check all the overlaps by eye, and calculate the Tm of the overlaps. If the Tm is low, I wonder if there's a chance that a lower reaction temp would help? (Or, if there are no G or C near the end of the overlap.... but even if the end can't ligate well, the exonuclease should chew more away, and the polymerase should chase it until it reaches a better spot... so I maybe that's not an issue.) Check for palindromes or hairpins in the overlap. ...Or, if it doesn't work the second time, and you are able to, you might shift where the ends of the fragments are to create new overlaps, and avoid any potential problems with the old overlaps.

You might try to spike the positive control with your DNA to see if it is dirty.

Okay I'll follow your advices and I'll tell you if it worked! Thanks a lot!!

I'll begin using a larger amount of DNA like you said !

With a difficult cloning I do think it’s best to be generous with DNA input, but your gel should still show the fragment going together, so you should be double checking the overlaps also, or just try new ones.

I now realize that I was looking at the old Gibson protocol, not the new NEBuilder one. It changes from 0.02- 0.5pmol fragments at 2-3 fold excess insert to NEBuilder: 0.03 – 0.2pmol with 2 fold excess. Using the maximum amount with your 10kb vector comes to:

444ng vector (10kb, .067pmol) + 507nt insert ( 5.7kb, 0.133pmol). I’m very sorry, but I did say to double check me.  This is 2/3 the earlier calculation, which I think is close enough that I think it should still work well. (You can do ½ or 1/3 size reactions to save resources).

Don't be sorry!!! And thank you I didn't  know the new protocol!! I'll try it. And don't get the last part, what do you mean 1/2 or 1/3 size reactions? 

I just meant that I save money by doing 6ul reactions instead of 20 ul. 3ul DNA assembly + 3ul enzyme mix. (Or 5ul + 5ul if  you are 40% less confident about your pipeting).

Oh okay! I was wondering if the efficiency was ok by dividing this but apparently yes thank you.

If everything is the same, a scaled down reaction 'should' work the same (unless, perhaps, it's so small that there's some kind of issue with surface tension of water, I imagine). However, pipeting inaccuracies are magnified, and there will be less in case you want to run some on a gel. 

 Actually, you did say "Gibson assembly", and were unaware of a new protocol... so the maximmum amounts of DNA to use that I first gave were probably correct. NEB has a new "NEBuilder® HiFi DNA Assembly Master Mix" that's their new, improved Gibson assembly kit. My correction was to this, slightly different, protocol. The original Gibson protocol is unchanged, to my knowledge. I don't know if the new protocol is better for both kits, but companies don't like to switch protocols, which confuses clients and reduces confidence. Both protocols are very similar, but they say the new kit is more efficient and accurate.

 So, Peter, even within the range listed in the NEB protocol you wouldn't choose the maximum? You would choose a vector amount in the middle of the range? So, for NEBuilder they recommend 0.03–0.2 pmols (that means total "vector" + "insert"?), so maybe .03pmol vector + .06pmol insert =0.09mols totoal?  For the  "Gibson Assembly" kit, they recommend .02-.5 pmol. So, maybe 2 or 3 times the one for NEBuilder?

You're right this is really great to see people's answers! Actually I already used the NEB Builder to design my primers...

Maybe I should set a min of 20 bp overlap lenght...

Maybe.

Or, maybe you should try another method, if that's possible with your insert and vector.

If you design new new overlaps, I'd shift the location, if possible , in case there's a troublesome sequence now. If you want to show me the overlaps, I could look for problems.

If you can tolerate duplicating the overlap, it may be possible to take your existing insert, kinase + dATP it to add phosphate to the ends, and blunt end ligate into the vector. I don't know if it's worth the time, or not, but it may be quicker. If the problem was dirty DNA, that problem may be repeated.

As a further out there idea, maybe you could amplify the Gibson product by inverse PCR. Use back to back primers, facing in opposite orientations on different strands, to produce a linear PCR product the full length of the plasmid. This would assume that you have a small amount of good ligation from the Gibson reaction product to use as template. The PCR product is then checked on a gel, kinased (or primers with 5' phosphates used), ligated to circularize, and transformed. I don't know if it's worth trying, but if you want to try it with a pair of primers you already have of adjacent insert and vector, you might try it while you wait for new primers to arrive, and see if you see the product on a gel. If you use existing overlap primers, it will duplicate the overlap sequence, one from each primer. A gel that can distinguish DNA that big must be used, or the PCR product cut into smaller pieces, but not the original insert and vector sizes.  A unique restriction enzyme site in the overlap could be used instead of kinking the ends (if it exists, the palindrome may have contributed to the problem).

Thank you for those ideas I'll try the amplification of my assembly to see if there is something in here and then the ligation.

My primers are the following (designed with NEBbuilder):

Insert_OL_vector Fwd: TTGCATGCCTGCAGGATTTATTTCATATCTTCTTTTTGGGTTG

Insert_OL_vector Rev: CAAAAGTGGAAAAGCTAGATGAATGAGTAAAGGAGAAGAACT

Vector Fwd: ATGAGTAAAGGAGAAGAACTTTTCACTG

Vector Rev: AGCTTGCATGCCTGCAGG

I hope there is nothing bad in it!

It looks like you gave me the sense strand of all the primers, judging by the overlaps. Were the "Rev" primers ordered as the reverse/complement of the sequences given?

I suppose they must have been, or the PCR would have failed?

Your forward overlap has 2 restriction enzyme sites (palindromes). NEB recommends avoiding this, and avoiding placing the overlap in a Multiple Cloning Area in general, for this reason. I also see a hairpin near the end (which has to primer extend in the Gibson fill-in):

TTGCATGCCTGCAGG  This hairpins with the following:

GACGT...            The first end nt here is a GT mismatch (the most tolerated). Following this are 4 nt of perfect match. The end of the strand, here,  must anneal, and primer extend to reach the 5' end of the other DNA fragment exposed by the exonuclease, and then the two ligate. Could this hairpin, or the restriction enzyme sites, interfere with the overlaps annealing efficiently? I don't really know if this can happen at 50C. If Peter Weigele is still monitoring this his comment is welcome, or you might call NEB and ask them.

I suppose NEB's builder tool can control one end of the overlap, by controling the length or the overlap, but it can't control where we tell it to start and stop the vector from. Maybe this is a reason to let the tool put the overlap where it wishes, and not forbid it to put insert on the vector primer by selecting the vector as "backbone", as I usually do so that I may reuse the primers and PCR product for future cloning.

Looking on your Rev vector primer for dimers, IDT's tool finds a perfect 8bp match (6bp GC) at the right end of the primer. I assume you gave the reverse complement of this primer, so it didn't affect PCR (not being the 3' end), but if this is large enough (?) it could mean that the DNA fragment could assemble to itself using this 8bp overlap. If it did anneal, this could compete with the correct insert reducing the number of clones. However you should still get some good constructs, especially since you used a molar excess of insert to outcompete it. The left end of the same Rev vector primer seq (the 3' end? if rev complement) has a perfect 4bp dimer. This might create primer dimer during PCR; however you successfully completed the PCR, and probably removed the dimers when you purified the fragment. 

Again, Peter is welcome to comment. Or, you can ask NEB (but if you point stuff like this out on a problem experiment, of course they'll say that could be the problem.)

I agree. You can trouble shoot more, while you wait for primers to arrive. Or, try another cloning method in case this method doesn't like this sequence (If you can't change the overlaps especially).

While reordering primers, you can make an alternative set with restriction enzyme sites, for T4 ligase, as a backup.

If you did do a blunt end ligation, remember to check the orientation.

Hi! I'm sorry I was doing some experiments elsewhere, I just saw your comments. I'm sorry, indeed the primer wasn't rev complement. I think I'll design new primers trying to avoid what you notices (the restriction site and secondary structures). I did a PCR and it showed that there is an assembly, but maybe not enough...

Hi Julie!

I am also having hard times with the Gibson assembly. I am trying to insert a 1569 bp fragment into a 10535 bp vector. I am using the NEB kit for Gibson. For my last reaction, I used 100 ng of vector and 200 ng of insert, but I did not get colonies ( I am using stabl-3 cells). I will appreciate if you share with me the strategy you followed to solve your problem. Did you electroporate your bacterias?, I also want to ask you what should I see in an agarose gel if check the products of my reaction? Thank you in advance. 

Best!

Ivan

Hi Ivan,

Gobson Assemvly is incompatible with Stbl3 cells, there is a buffer difference such that the Stbl3 cells are unable to take up your assembled plasmid and therefore all die under antibiotic selection. I suggest either cleaning up your Gibson assembly reaction with ethanol precipitation, or switching to one of the competent cells NEB specifically recommends. Also, NEB has a newer version of Gibson Assembly called NEBuilder which in my experience increases efficiency greatly, though you will still need to change your competent cells or clean up the reaction. Hope this helps.

Hi Natalie!

Thank you so much for answering my question. I am actually using NEBuilder, but now I understand what might be wrong with my protocol. I will check my reaction in a gel before switching to a new competent cells. I also have DH-5 alpha cells, but I have heard they don't do well taken up large vectors. Do you think that the electroporation technique would work? Best!

Hi Ivan,

Sounds like a good idea to check your reaction before worrying about the competent cells, though if you are using Stbl3 cells you definitely won't get colonies, even if your assembly worked. I've never used the DH-5 alpha cells so I can't comment on them, not have I ever electroporated bacteria. NEB does sell competent cells that are supposedly good for large vectors or vector with repeated elements, though in my experience these cells are not nearly as good as they company says they are (not really that stable). If the efficiency of your assembly is ok, cleaning up with ethanol precipatipm before transforming into Stbl3 cells shouldn't be an issue. If you do this however, make sure you purify immediately after the assembly is complete, in my experience letting the reaction sit at room temperature or even freezing (which takes a little bit) decreases the number of successful assemblies. Likewise, follow the NEB protocol and set the reaction up on ice.

Hopefully this helps!

-Natalie