More often than desired (about 75 % of the time), when building plasmids via Gibson Assembly (~ 5 kb plasmids; from a maximum of 2 fragments) we obtain clones with several random point mutations all along the plasmid (from 1–3 point mutations to 6+ depending on the day, without specific region targetted).

Since we use the NEBuilder HiFi DNA Assembly kit and since there doesn't seem to be any specificity as of where those mutation are found, I suspect that they arise during the PCR reaction when producing the amplicons.

We use the Phusion High-Fidelity polymerase to generate the amplicons from ~1 pmol of template in 50 µL of PCR reaction (or split in 5 x 10 µL since working with reduced volumes seems to improve the efficiency of amplification) and 30 cycles.

  • Is this a normal/common issue or does this result indicate that something wrong is happening in our experiments?

  • And more importantly, any idea of how to prevent such problems?

I have in mind:

  • Reducing the number of cycles (how low can one go?) and/or increasing the concentration of template (related).
  • Changing polymerase (but for which one and shouldn't Phusion High-Fidelity be one of the best suited?).
  • Changing other parameters or conditions (but which ones?)

Any bullet-proof experience on that? Thanks


Changing the DNA polymerase from a Phusion HF to the Q5 suggested in the NEBuilder kit manual (suggestion that I had overlooked) completely solved the problem.

Rem 1: Mind that the elongation time must be slightly increased to amplify "long" fragments with Q5—probably a side effect of its higher proofreading ability. Whereas 20–30 s/kb is recommended, 120 s where not enough to amplify a fragment of 3.5 kb; an elongation time of 240 s—i.e. 60 s/kb—did work though.

Rem 2: In the past, amplification with the Phusion HF DNA polymerase did result in correct sequences in most of the cases, so I was surprised to have issues here. A possible factor that came to my mind afterwards and that could explain the apparition of mutations could be a low purification quality of the DNA preparation used as the template (leftover chemical contaminants). Although I have not (yet) tested and confirmed this possible explanation in my case.

I marked the answer from @markur below as accepted, as his/her suggestions (along with the accompanying discussion underneath) prompted me to go and read the original articles and then to read again more carefully the manual of the NEBuilder kit where the solution was.


1 Answer 1


Polymerases are never perfect. Mistakes can happen and are usually the reason why one always picks several clones and sends them off to sequencing. Phusion polymerase is advertised to contain 1 error in 6% of all PCR products (5000 bp, 30 cycles, HF buffer; thermofisher error calculator). Companies always underestimate their errors, but in your case, having 6 point mutations in each clone is reason enough to optimize your protocol.

Changing the Enzyme will definitely have the greatest impact on your problem. Especially if it also includes changing the buffer system.

Phusion has been around quite a while, I'm sure that some companies did apply some optimizations trying to beat the competition (Kapa).

Another thing that came to my mind is that high magnesium concentrations increase the amplification rate, hence reducing corrective activity (aatbio). You might want to review your plasmid purification protocol to see if there might be salt contaminants (e.g. after EtOH precipitation).

Decreasing cycles also makes sense, and is definitely worth a shot.

Other aspects to try, if you're desperate:

  • Get a longer primer (~30 bp) and use a higher melting temperature to enforce strict and correct primer binding.
  • Getting fresh primers is also a good idea to avoid freeze/thaw damages
  • Use a fresh dNTP solution
  • Adding BSA is always worth a shot with PCR Farell 2012. Not just during the reaction, but also during storage to protect nucleic acids from freeze/thaw cycles Svec et al. 2013!
  • Check out thermofishers troubleshooting guide, there are some more ideas that might help!

Important: Please verify that it's indeed the PCR reaction that's causing point mutations, and not the bacterial clones! Check that by re-amplifying the same plasmid with PCR without prior Gibson Assembly and sequencing the product like any other clone!

  • 1
    $\begingroup$ Thank you very much for your answer + the tips + the links. A question: do you have experience in Gibson Assembly? I am asking because I am also wondering how my experience compares with others (1st point in my question) regardless of theoretical error rates as advertised by companies. $\endgroup$
    – The Quark
    Commented Oct 20, 2022 at 15:29
  • 1
    $\begingroup$ I've done a lot of Gibson Assembly with a home made kit (preparation identical to the original publication), and although I don't have numbers, I can't recall random mutations being particularly common. Do you see any pattern to where the mutations most often appear? Is it anywhere in the PCR amplified fragments, or are they more common in the repair regions (fragment overlaps)? $\endgroup$
    – gaspanic
    Commented Oct 22, 2022 at 1:43
  • 1
    $\begingroup$ @TheQuark - I did it a few times but I will ask some technical assistants if 1-6 random point mutations per 5 kb for each clone is common (I don't think it is). One clone should be "perfect". $\endgroup$
    – markur
    Commented Oct 22, 2022 at 9:32
  • $\begingroup$ Following these exchanges, see the edit in my original post. $\endgroup$
    – The Quark
    Commented Dec 1, 2022 at 17:30

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .