By now, the polymerase chain reaction (PCR) is no longer the new big thing in molecular biology and diagnostics; it’s not only taken for granted as the underpinning method behind a vast number of clinical tests, it’s the sort of thing taught in high school biology. Few, if any, readers of this column won’t have at least a basic appreciation for the process (similar to that outlined in this space in the May 2013 edition, https://www.mlo-online.com/home/article/13005202/pcr-the-basics-of-the-polymerase-chain-reaction). Two primers, some buffer, and thermocycling including denaturation, annealing, and extension steps and you’re good to go.
In a lot of assays, that’s really about all there is to it, too. Aspects like making the PCR into a real-time assay involve complications with adding means for ongoing product detection, but those don’t change the mechanics of how the product is formed. In this month’s episode of The Primer, we’re going to peek under the hood at some of the alterations that can be done to the underlying core process, how these impact product formation, and why or where you might find these applied.
Touchdown PCR
Specificity in a PCR reaction relates mechanistically to ensuring that the forward and reverse primers anneal productively at their intended priming sites, and not at a bunch of other unwanted loci. Part of the key to this is the “uniqueness” of each primer sequence, or how many near-match binding sites there exist in a template population as compared to intended annealing site(s). Ideally, one would design primers that have absolutely no significant similarity to template sequence elements other than the intended site.
In reality, unavoidable constraints on amplicon design based on minimum and maximum product sizes for efficient amplification, limited areas of sufficient genetic conservation pressure to serve as reliable primer sites, target genome size and GC content, and other factors can sometime leave an assay stuck with primers that have more than passing homology to unwanted secondary binding sites. While these will be thermodynamically unfavored compared to the perfect match, annealing occurs along a Boltzmann distribution curve: put more simply, at least some annealing will occur at less-than-perfect binding sites; and if your primers are less than ideal from a “uniqueness” standpoint, this can be a significant proportion. What’s an assay to do in this situation?
One way to enhance specificity in this scenario is by what is called Touchdown PCR. In this, rather than use the predicted optimal primer-annealing temperature, one starts off the initial PCR cycles with annealing step temperatures well above that; often, as much as 10⁰C over predicted Tm. Over the next 10 to 20 cycles, the annealing temperature is decreased by ~0.5⁰ - 1.0⁰C per cycle. What happens – thinking back to our Boltzmann curve – is that at temperatures well above ideal annealing, only a very small fraction of primers will have low enough thermal energy to bind, and such binding will be biased towards the perfect match sites. Thus, a single copy target gets preferentially (but very weakly) amplified in the first cycle, slightly less preferentially (but a bit more strongly) amplified in the second cycle, and so on until “optimal” Tm is reached. By this point, ideally, the number of perfect match priming sites from early cycle amplicons helps compete on a number of copies basis against the un- (or at least, less-) amplified spurious sites.
On this now selectively enriched template, a further 20-30 cycles or so of PCR at more usual optimal Tm, with no further decreases per cycle in annealing temperature, proceeds to drive the bulk of the amplification. At the cost of nothing more than some extra thermal cycles (minutes of time), this technique can help maximize specificity for non-ideal primers.
Asymmetric PCR
Traditional PCR makes the two strands of product amplicon in equal numbers. If you’re going to detect product by gel or capillary electrophoresis, or by a double-stranded DNA (dsDNA) selective fluorescent dye, that’s good. If, however, you want to use a hybridization-based product detection –including either product hybridization to fixed capture oligos (2D or liquid phase arrays), or fluorescently labeled probes using fluorescence resonant energy transfer (FRET) in its various guises – it’s less ideal.
That’s because the capture or hybridization probe oligo has to compete for product binding with the amplicon’s complementary strand, and there are losses in detection efficiency arising from amplicon strand reannealing. If your detection method hinges on capturing just one strand of the PCR product, asymmetric PCR may be the solution you’re looking for. In its simplest form, it consists of just limiting the amount of one primer (the one complementary to the desired detection strand) relative to the reverse (detected strand) primer.
Traditional cycling conditions are employed, and in early cycles where numbers of both primers available vastly outnumber template molecules, nominal two-fold amplification per cycle proceeds as you’d expect. As amplification proceeds and numbers of amplicons exponentially increase, the less-abundant primer becomes scarce by comparison to the detected strand primer and an increasing number of single-stranded products are formed where only the detected strand primer is available to successfully anneal and extend.
Denaturation allows the non-detected amplicon strand to then be available as template again. A key result of this – other than creating an excess of single-strand product – is that once the limited primer becomes scarce, amplification ceases to be exponential and becomes a much slower, linear process. Careful balancing of the ratio and total amounts of the two primers, and a relatively narrow optimal window for starting template concentration, are needed to ensure that sufficient levels of amplification can occur before this gradual shift to single detected target strand production takes over. Done properly, the result can be increased hybridization-based detection efficiency over classical “symmetric” methods.
Variations on this exist, most notably something referred to as LATE-PCR (“Linear After The Exponential PCR”). Essentially, this alters the designed Tm of the limiting primer to take into account the thermodynamic reality that primer concentration impacts observed Tm – that is, the limiting primer behaves as if its Tm were lower than it would at concentrations equalling that of the opposing primer. This interferes with and reduces the real-life efficiency of the simple model described above. By intentionally raising the designed Tm of the limiting primer, this effect can be offset. For those interested in more on the topic, the original concept can be found in reference1.
Nested PCR
Receiver operating characteristic (ROC) curves teach us that assay conditions are a trade-off between sensitivity and specificity. If increasing sensitivity in a PCR reaction is your goal, there is one method that can do so dramatically but simultaneously act to retain specificity – at a cost. Known as nested PCR, in essence, the process consists of a first “normal” PCR reaction (most commonly at a somewhat reduced number of cycles, 20 to 30), and then use of a small portion of this product as template for a second PCR in which the primer sets are designed to anneal to positions inside – “nested” – the first amplicon.
This allows for astronomical levels of amplification, but by forcing the second amplification to be specific to only correct products of the first amplification (that is, bearing the internal second priming sites), specificity is, in theory, retained. Such an approach can be useful when absolute limits of sensitivity are required, and/ or when samples are expected to contain inhibitors. In this second scenario, while only limited amplification may occur in the first stage, the effective dilution of template-borne inhibitors by taking a small sample onward gives the second stage a cleaner environment to work in, plus at least some amplicons to work from. Sounds great, so what’s the cost?
Aside from small increased labor and reagents and consumables, the really significant cost is an operational one, which should be sounding alarms in every clinical laboratorian’s mind – contamination! The process described above requires opening a reaction tube post-amplification and liquid handling amplicon containing products. This really should not be considered acceptable practice; the avoidance of this was the impetus behind development of real-time methods capable of detecting product without opening the reaction vessel, and a strong early driver of their adoption. There is, however, a way to do a form of nested PCR without opening the reaction vessel – that is to design the second (inner) primer set with a much lower annealing temperature (i.e. shorter primers and /or lower GC content than outer primers), and put both primer sets in the reaction at time of setup. A first set of thermocycling can be performed with an annealing Tm such that only the outer primer set significantly functions, with the inner set left in solution. After a number of such cycles, the annealing temperature is dropped to match the inner primers. While the outer primers will, of course, continue to anneal (and now, likely, mis-anneal at incomplete matches) under these conditions, two factors help to supress spurious products and favor the intended nested product.
First, some amplification has occurred during the first stage, so there is a relatively large number of first-stage amplicons compared to original bulk template, providing a stochastic advantage to the proper priming. Secondly, if the nested product is significantly shorter than the first-stage product, the extension time during the second stage can be shortened to give the nested product a kinetic advantage over the longer first amplicon. Challenges to this are in finding a suitable target sequence with inner and outer priming sites meeting all of these requirements. While this method avoids the calamity of having to open the product tube, it does not gain the dilution of inhibitor possible with the first method. For all of this, it can still give some boost to lower limit of detection with marginal or no loss of specificity.
Conclusion
There are far more variations on classical PCR conditions – both in reaction composition and thermocycling parameters – than can be covered in so short a space, but the three discussed here are among the ones the reader is most likely to encounter in common lab assays and may go some way to answering “why does this use a strange-looking thermal cycling profile?”
Not all PCRs are created equal, and often that’s because someone is using clever tools to maximize reaction behavior against a particular task and detection method.
References:
- Sanchez JA, Pierce KE, Rice JE, Wangh LJ. Linear-After-The-Exponential (LATE)–PCR: An advanced method of asymmetric PCR and its uses in quantitative real-time analysis. Proc Natl Acad Sci U S A. 2004; 101(7): 1933–1938.
