Primer as limiting reagent in PCR reaction - (May/07/2014 )

Hi, 

I'm working on a custom NGS application where for various reasons PCR-induced chimeras are completely unacceptable, so I need to do everything I can to avoid them.  From looking through literature, I've seen that the common suggestions include using betaine, increasing extension time, and decreasing the number of PCR cycles.  Betaine and extension time are easy to implement, however decreasing the number of PCR cycles could be a bit problematic, mainly because the mass of template going into the PCR reactions could vary across 2 orders of magnitude (depending on what we get from collaborators).  Simply "decreasing the number of cycles" could work perfectly for the higher concentration samples but the lower concentration samples may not amplify enough to be sequenced.  Unfortunately the dilute-everything-to-a-standard-concentration approach won't work for various reasons (basically some samples will contain more amplification targets than others and there's no way to know that a priori) and would induce an unwanted bottleneck for our higher mass samples.  I had another idea that I'd like feedback on before sinking too much time into testing it out:

Contrary to what I thought before looking into this, primer exhaustion is not the reason for PCR plateau (it's actually a buildup of dsDNA product that acts as a polymerase inhibitor - page 25).  In fact, my calculations show that there are many orders of magnitude more primers put into the reaction than strands of DNA coming out, even when reactions are run to plateau.  Would it be feasible to severely limit the concentration of primers and run the reaction for many cycles (40?) to ensure that even the low-concentration samples are amplified to a detectable range without over-amplifying the high-concentration samples and inducing PCR chimeras.  My calculations suggest that this strategy would require a 6nM final concentration of each primer.  Will this pose any problems?

My main concern with this strategy is that I'll either A) get no amplification due to primers never finding their landing sites, or b ) get really weird PCR artifacts as a result of running the reaction past the point of primer exhaustion (although I'm hoping that long extension times, 3min for a 1kb fragment, will reduce the presence of chimera artifacts).

If you've done a similar reaction or have experience running reactions with vanishingly small amounts of primers, I would greatly appreciate your feedback!

Thanks!

If you run the PCR in a real time cycler, you could stop the reaction when it plateaus.

phage434 on Wed May 7 23:45:45 2014 said:

If you run the PCR in a real time cycler, you could stop the reaction when it plateaus.

I had thought of this, however we frequently run an entire 96-well plate worth of samples at once, each with a different mass of template.  We could have some reactions starting with 2.5ng and some starting with 500ng on the same plate.  That would make it a bit difficult to stop each individual reaction at exactly the right time without disturbing all the other reactions.

This evening I put in a 50-cycle PCR with 6nM final concentration of each primer - tomorrow morning I'll see if I have the expected amount of product.  If I have the right amount of product I'll spike some into our next Illumina run to get a fine-grained assessment of any artifacts/chimeras.  Any suggestions/input in the meantime is welcomed.