PCR functions because of two essential components - a thermostable DNA polymerase, and a pair of DNA 'primers'. Primers are short, made to order, stretches of oligonucleotides ('oligos' - from Greek meaning scanty or few). Modern "oligos" can be synthesized in lengths >100nt however the behaviour of oligonucleotides longer than 20nt is different from that of shorter oligos and different calculations are employed to calculate their thermodynamic characteristics.

Primers, as their name suggests, prime the nucleic acid template for the attachment of the polymerase. This is the first step towards duplicating the template to which the primer hybridises or 'binds'. The primer directs the polymerase to move 5' to 3' because of the 'direction' of DNA (See Nucleic Acid Structure for more background).

Figure 1. DNA has direction.
The polynucleotide chain shown above is 'read' in a 5' to 3' direction by the polymerase. This would be from the top to the bottom or from the phospate group to the hydroxyl group.

Primers hybridise at a temperature that is affected by their sequence, concentration, length and ionic environment. This annealing temperature is usually referred to as the T_M (melting temperature) but is in fact 5–10°C below the T_M. The term T_M describes the temperature at which 50% of the oligonucleotide–target duplexes have formed.

Figure 2. Deoxynucleotide triphosphates.
Each of the five deoxynucleotides are shown, 2'-deoxycytidine-5'-triphosphate (dCTP; C₉H₁₆N₃O₁₃P₃, MW=467), 2'-deoxyguanosine-5'-triphosphate (dGTP; C₁₀H₁₆N₅O₁₃P₃, MW=507), 2'-deoxyadenosine-5'-triphosphate (dATP; C₁₀H₁₆N₅O₁₂P₃, MW=491), 2'-deoxythymidine-5'-triphosphate (dTTP; C₁₀H₁₇N₂O₁₄P₃, MW=482) and 2'-deoxyuridine-5'-triphosphate (dUTP; C₉H₁₅N₂O₁₄P₃, MW=468).

PCR gleans its extreme specificity from the primers. At each and every position of a new primer, we have 4 nucleotides to choose from, dATP, dCTP, dGTP and dTTP. So, if we design a sequence-specific primer of 20-30nt nucleotides in length ('20-30mer'), the chance that that exact sequence will occur randomly in nature will be 1/4 x 1/4 x 1/4 etc, 20 or 30 times i.e.

That means a 1 in 10¹² to 10¹⁸ chance of a 100% homologous match to a non intended target. Or to put that in perspective, there are 2.85 x 10⁹ basepaired nucleotides in the entire human genome. While that sounds all very convincing, in reality, primers designed to detect viruses often share significant amounts of homology with the human genome - sometimes resulting in false positive amplifications. Even when the homology is far from 100%, primers may still amplify an unintended target as shown below. This most likely reflects the co-evolution of many viruses with humans during which time they have "captured" bits of our genome and deposited bits of their own genome. Below are some of the problems we can encounter when using the PCR.

The first problem I'll discuss is the most common, the most difficult to avoid and often the most insignificant. When a small amplicon results from the extension of self-annealed primers, you get primer-dimer (PD) i.e. a dimer of one (self-annealing) or both primers resulting in a template capable of being extended by the polymerase. PD formation is highly efficent because the primers are in vast excess compared to the amount of template or even the number of amplicon molecules at the end of the PCR. This drivs the formation of PD.Two main concerns arise from PD formation. Firstly, because PD formation is so efficient, it rapidly consumes dNTPs and primers and generates amplification inhibiting pyrophosphates. All of which can prematurely plateau the exponential accumulation of product. Secondly, if we using a dsDNA-associating fluorescent molecule to follow the PCR's progress during real-time PCR, then PD will also show up, and, at least during the kinetic portion of the assay, will be undifferentiable from the signal of specific amplicon acucmulation.

Figure 3. Examples of how primer-dimer (PD) amplicon can be formed. Ten examples are shown of sense and antisense primer interactions resulting in amplicon. Note that different length amplicons can be formed so PD size can be characteristic of the primer pair used. The largest PD would result from the hybridisation of the smallest number of nucleotides and would approximate the length of the two primers added together.

Mispriming is the result of a primer binding to an unintended template resulting in amplification. The amplicon can sometimes be the same size as the intended product, but is usually a different size when viewed following agarose gel electrophoresis. Mispriming occurs because of poorly optimised conditions or because we haven't checked whether our sequence will inadvertently bind to an entirely different target entity e.g. a region of the human genome instead of the intended virus genome.

Mispriming can usually be avoided by more intensive comparison of the primer's sequence against the GenBank database using the Basic Local Alignment Search Tool (BLAST) at NCBI. Of course, a BLAST comparison will only find matches among those sequences housed in the database. When it comes to PCR where a single nucleotide mismatch can cause an amplifictaion to fail, or at least perform with reduced efficiency, BLAST'ing primers can lead to a feeling of security. In some instances the homology of the primer to its template may indicate a perfect match simply because viral variants have not yet been sequenced and submitted. Also, because there may be many undiscovered viruses and unsequenced non-viral genomes in the world which are not represented on GenBank, a specific match, or a "no match", does not mean that you have exhaustively searched for all potential homologues.

In this section, the problem results from the way we design our primers. I am excluding self-annealing and secondary structures from here because we will deal with them specifically in the next section.

Self-complementarity