Odd point mutations - (Aug/13/2009 )

fishdoc on Aug 27 2009, 05:32 PM said:

My first guess when reading this thread was UV light....but I wonder...you say you "got back the sequence"....did you look at the actual chromatograms in the area you see a mutation, or are you assuming the sequencing is correct? Computers can make errors in base calling and by looking at the chromatograms you can additionally get a feel for how well the sequencing reactions worked, that could also be an indicator of the quality of the template. One thing I would try and be certain of is that the sequence you are getting is the actual sequence -- if not all the other experiments are null.

Another thing you might try, since you are going to great lengths to solve this problem, is do the "opposite" experiment and try Taq and see what the error rate with it is compared to your high fidelity enzymes. There are also alot of tricks you can do to increase "apparent" fidelity, like reduce the cycles (as others have suggested) or start with more template. You could even use one of your "correct" clones as a template, for experimental purposes.

I'd be interested to hear how things turn out...Warren..

Warren on Aug 30 2009, 05:27 PM said:

I have looked at the chromatograms, and in every instance save one, the computer made the correct call, based on the peak anyway. The one time it was incorrect was when the person that does all the sequencing for us looked at the chromatogram and called a T instead of a G. The chromatogram showed a small T peak, and the following base was a G. That G actually had a little shoulder before it that was above the peak for the T, so I assumed it was a miscalled G. However, another possibility is that the template to be sequenced was a pretty even mix of G and T at that position, and that's why there was a small T peak and a G shoulder. I'm not familiar enough with sequencing to know if that is a valid thing to happen. As for the sequence of the original, it's all from the genome sequencing project of our bacterium. I think the entire genome is at least 8 fold coverage (maybe more), so I think the template (sequence) I'm basing it on should be correct.

As for UV, many of these constructs are not being gel purified, but rather purified directly from the PCR reaction, so there is no UV exposure.

As for "going to great lengths", I was to begin with, but it began eating up a lot of time and resources to do side-by-side reactions and sequencing, and everything else. While I'm not happy at all with the poor sequences I've been getting, I can work through it as long as I'm getting some correct that I can carry on to the next step, at least in the short term. If this becomes a chronic problem, then I (we) will delve into it further. Because there has been relatively little response to this problem, and because a few google searches here and there have revealed no other similar issues with high fidelity enzymes, I'm assuming it's something on our end. Either a bad batch of polymerase, someone leaving it sit out too long, or perhaps the -20 isn't actually keeping at -20 (we are in a pretty humid climate, so the -20 frosts up pretty quickly and doesn't seem to work efficiently if it gets too frosty).

I was talking to my PI the other day about it, and for every reason we think of that it could be happening, we can think of 30 other reasons those aren't logical. But since we ARE having the problems, one of those seemingly illogical explanations really is quite logical, we just don't know it.

In any event, I will keep updating this thread as I find out any more information (if I find out any more information). We received new Phusion and new Pfu Ultra II on Friday, so this week I will be doing some more reactions and cloning, so maybe I will have some info on the new enzymes sometime next week.

I would immediately try new dNTP solution, or use a master-mix. dNTPs can easily go bad, and could explain your poor PCR performance. Also, how certain are you that the "errors" you are seeing are not real mutations in the template sequence?

phage434 on Aug 30 2009, 07:13 PM said:

Yes, new dNTPs are on order, too. We usually order a few tubes and aliquot as needed to lessen freeze-thaws. I had used different tubes from the freezer, but perhaps they weren't too good.

I don't think there errors are in the templates sequence, because they're in different places in different clones of a single construct. If the sequenced clones had the same "mutations" in the same place, I'd consider that my sequence of the template was just wrong, but that hasn't been the case. Furthermore, some of clones do have the expected sequence (no mutations) when cloned, but if I sequence multiple clones, I've had 2/3 have the mutations (different mutations in the two), but the other is good. In one case, I had 5 clones of one construct. Each one had between 3 and 5 mutations, and each mutation was in a different position. The incorrect bases haven't been consistent, which is why I don't believe they're the actual sequence.

fishdoc on Aug 30 2009, 09:12 PM said:

I think bad dNTPs are FAR more likely than two bad batches of thermostable polymerases that just happen to affect fidelity. Those polymerases are TOUGH (try to heat inactivate one and I'd say if your freezer wasn't holding temp you'd be far more likely to notice a problem in a more sensitive enzyme (like T4 DNA ligase) but I doubt that is the problem. I'd bet dollars to donuts that if its a component of the reaction, its the dNTPs (plus your looking for something specifically affecting fidelity). Although it seems like a waste of time and money if you figure it out you will save alot of time and money in the future! It costs a lot of money to buy those high fidelity enzymes, then do all that sequencing to find a correct clone because of the errors -- might as well use Taq! Good luck..Warren..

Is your -20 freezer of the frost-free variety?

Homebrew: The freezer is not frost-free.

Warren: Agreed on the time and money. I will be starting a couple other constructs starting this week with new polymerase, buffers, and dNTPs. The only thing that will be the same are the primers and the template. Also will try lowering the cycles. If we continue to get bad sequence, we may end up doing more to determine what the problem is. Hopefully, by changing the above things out, we'll get back on track.

Alright, so here's an update.

Had new polymerase, buffer, and dNTPs in hand and tried to amplify a 4.7 kb product with both Phusion and Pfu Ultra II. Neither worked, and that was odd considering it had worked fine the week before, with both polymerases. Ran some of my genomic template on a gel, and there was nothing to be seen, so I figured it had gone bad.

Did a new genomic prep, repeated the PCR with Phusion and got excellent results. Gel purified the band, inserted it, sequenced it, and everything was correct (sequenced the entire 4.7 kb piece), no point mutations.

Also, some of the other constructs I had been struggling with came out clean as well using Pfu Ultra II instead of Phusion, and using the "old" genomic DNA as template.

In the end, I'm really not sure what the hell was wrong. New genomic DNA helped get the 4.7 kb product made with Phusion (newly ordered), and it did not have any point mutations in either of the clones I sequenced. An older stock of Pfu Ultra II did better than the older stock of Phusion for the other constructs I was making, but there was still a pretty high amount of point mutations (relatively speaking). For instance, in making 2 different constructs, I sent 3 clones of each, and in both cases, 2 of 3 clones for each construct had point mutations. Luckily, there was one for each that was clean and I could continue on with.

Perhaps it was the genomic DNA that was causing the problems. It was an older prep, stored in water at 4C. Perhaps something happened to it that randomly affected a base here or there in some of the strands, resulting in mutations by the polymerase (no fault of the polymerase).

As far as I can tell, everything is back to good for now in terms of fidelity. If anything further happens, I'll be sure to update this with the details.

fishdoc on Sep 17 2009, 06:31 PM said:

Thanks for the update -- it could be the template, but I would be surprised if you could get amplification from a template that was being degraded (mutated), I might also expect the same type of mutation to be found. I am still suspicious of your old dNTPs!!! Warren..

Hi,

I've been having the same problem: sequencing of my plasmid constructs show several abnormal point mutations (PCR: Pfu Ultra II Fusion HS DNA Polymerase; cells: E.coli DH5alpha).

Did you manage to find out the reason for the mutations?

many thanks,

Teresa