The Chemistry Behind the Polymerase Chain Reaction

Jolon Dyer
University of Canterbury

A mystery unveiled

The year is 1990. the country the United States of America. A woman is brutally raped and murdered in her own home. Three men are linked to the crime through fraudulent use of the victims credit card the night before her death. Let's call these three men Mr. X, Mr. Y, and Mr. Z. One of these men, Mr. X confesses to the rape and murder but, strangely, his confession is inconsistent with the physical evidence of the crime. What is this evidence?

1. A bloody footprint is left at the scene of the crime. This is identified as belonging to Mr. Y.
2. A large number of hairs are also found at the scene of the crime.
3. Semen stains on the victim and her clothing.

The conventional way of analysing biological specimens is to type them using protein marker typing. However, in this case such analysis was unable to identify the donors of the samples. Is it possible to sort out the role each of these men played in this heinous crime? Can the source of the hair and the semen be established?

Just five years earlier these questions would have been difficult or impossible to answer. In fact, this case would have remained unsolved or ambiguous had it not been for the advent of a new forensic technique which enables even minute amounts of DNA, deoxyribose nucleic acid, to be typed and identified. This new technique is called the polymerase chain reaction (PCR).

The PCR is a method whereby 'a DNA fragment can be amplified a million-fold or more for unequivocal identification.' One PCR-based genetic marker system is commercially available and validated for forensic casework. This system determines the DQa type of DNA extracted from forensic samples. simply put, a system such as this one works on the principle that, while everyone's genetic code is unique, the population can be divided up into categories. In this case the category, or type, of each individual is represented by two numbers. Table 1 shows the DQa of the victim, her boyfriend, and the three suspects, as established by the polymerase chain reaction. 

To analyse the physical evidence, 15 hairs distinguishable from the victim's selected from those found at the crime scene. DNA was isolated from these hairs and amplified by PCR for DQa typing. 13 of these hairs typed as DQa 2,3 while the remaining 2 gave a non-result. In addition, semen samples were differentially extracted from a vaginal swab of the victim and from her clothing. Sperm DNA from both these sources was amplified by PCR and typed as DQa 2,3 also.

What was the result of this PCR-based investigation? Firstly, the man who confessed, Mr. X, can be eliminated as the donor of the hair and semen. Mr. Z and the victim's boyfriend can also be eliminated. Mr. Y, who matched the bloody footprint, cannot be eliminated as the source of the hair and semen. In fact, his particular DNA type occurs in only 3% of the population. The jury for this case convicted Mr. Y of rape and murder. Mr. X was convicted of the lesser charge of co-conspirator.

This intriguing case demonstrates the remarkable power of the PCR. But just what is this reaction? How does it work? It is the intention of this review to outline the basics of this remarkable and elegant technique and show that, while on the surface the technique appears to be strictly the domain of biology, its roots are deeply embedded in chemistry. To achieve this, firstly we are going to step back in time and explore some of the vital chemistry which led to the discovery of the PCR (Table 2). Then we will look at the reaction itself and how it works. Finally, we'll look at the diverse array of applications which PCR has already acquired.

The structure of DNA

In 1953 the complementary double stranded (duplex) structure of DNA was discovered by Watson and Crick. According to this model, DNA consists of two polynucleotide strands coiled around each other in a double helix, and held together by hydrogen bonds between specific pairs of bases. Part of the immense beauty of Watson and Crick's model was that it explained how DNA could carry information, via a code of four distinct nucleotides.If we look closer at these nucleotides, it can be seen that they are each made up of three basic components: 1) a phosphate group, 2) a sugar group, and 3) a base. While the phosphate and sugar groups remain the same for all four nucleotides, it is the attached base which is the distinguishing feature of each nucleotide. The structure of these bases is crucial to the DNA as a whole because it is the bases that actually link the two strands of DNA together, via hydrogen bonding.

There are two kinds of bases present in DNA (Figure 1), the pyrimidine bases, including cytosine, thymine, and uracil (which replaces thymine in RNA), and the purine bases, which include adenine and guanine. The key difference in these two kinds of bases is best shown when one considers the base pairing in DNA. Each base pair is made up of a pyrimidine and a purine base. Thymine (T) always binds to adenine (A), while cytosine (C) always binds to guanina (G). These two base pairs, AT and CG have a major difference in their binding. The AT base pair is held together by two hydrogen bonds, whereas the CG base pair is held together by three hydrogen bonds, making it a stronger linkage. This is an important feature to keep in mind.

In essence then, DNA could be summarised as two polymer chains, each made up of only four basic building blocks, which are held together by hydrogen bonding. Naturally, once something was known about the structure of DNA, people tried to synthesise DNA themselves, both enzymatically and chemically.

1. Firstly the template DNA is denatured (i.e. the strands are separated) by heat, and then it is cooled to a temperature at which a large molar excess of the primers anneal specifically to their target sequences.
2. Next these primers, now attached to the template DNA, are extended by DNA polymerase, using deoxynucleotide triphosphates (dNTP) as the building blocks.
3. Repetition of this cycle leads to an exponential increase in the desired fragment. This is because the products of each cycle serve as templates for the next cycle.

In addition, minor side products accumulate at a linear rate, but these are soon swamped by the exponential increase of the desired product. So the whole reaction is carried out simply by cycling the reaction temperature! Figure 3 shows a typical thermocycling profile for a PCR reaction.

Variables of the PCRA 

PCR contains many variables, which can de varied so as to optimise the desired amplification. The major variables involved are:

1. Temperature cycling parameters - By varying the temperature at which the melting process occurs, the specificity of the reaction can be altered. In addition the total number of cycles performed should be kept as low as possible so as to avoid spurious amplifications.
2. Nature and concentration of template DNA - Extensive DNA purification has been found to be advantageous for only the least efficient PCR reactions, so usually the template DNA requires little purification.
3. Nature and concentration of the primers - The primers for each PCR reaction are designed specifically to optimise the reaction. In fact a whole set of rules is usually followed in the design of primers.
4. Buffer conditions - In practice standard buffer conditions vary little, but the concentration of Mg²⁺ has been found to be important.
5. dNTP concentrations - If the concentration of these is too high, then the frequency of the incorrect building block attaching itself increases.
6. Enzyme considerations - The enzyme used for most PCR reactions is Taq DNA polymerase as this is stable at high temperatures. However this enzyme does not possess a proof-reading function.

So just what are the advantages and merits of PCR. Probably most importantly, the reaction is extremely sensitive. In fact PCR has been used to amplify gene products from merely a single cell!

Then there is its specificity. As we have just seen parameters can be optimised to achieve excellent specificity. As a rule, the more specific the binding of the primers, the more specific the products.

The problem of abundance is solved too with PCR. Typically 1pg of starting template, in which there is only a single copy of the desired product in each DNA piece, can be amplified in a 30 cycle PCR reaction to between 0.05 and 1.0μg which in this area of science is plenty for most purposes.

PCR is also automation friendly, that is, the whole process lends itself to automation. All that is required is a device which can cycle the reaction through a range of temperatures. This in turn means that the speed of a PCR reaction is very much faster than any related techniques, such as cloning.

In addition, we have seen that PCR is not too fussy. The starting DNA does not have to be totally pure or intact, as long as it does not contain any other contaminating DNA. This feature is extremely important in areas such as forensic work, where the sample DNA may have been exposed to the elements for long periods of time and be fairly broken down or contaminated with dirt.

And finally, yet another very useful feature of PCR is that modification of the DNA can be made. These modifications are able to be introduced into the target DNA place via the use of primers. The primers may have deliberate mismatches incorporated into them, or they may have additional segments of DNA attached to them (containing, for example, restriction sites).

As with any process however, the polymerase chain reaction has its limitations.

The first has to do with contamination. Because PCR is such an ultra-sensitive technique, any contaminating genetic material can seriously affect the outcome. For example, in the criminal case outlined at the start of this report, if any DNA from another person had contaminated the samples, then the results could have been completely different. As a consequence, strict controls must always be performed in conjunction with any PCR reaction, and the reactions repeated-where possible. Also, elaborate isolation procedures are standard, so as to minimise the risk of contamination.

Another limitation is that the enzyme most commonly used for PCR reactions, Taq polymerase€, lacks a proofreading function. It is estimated that this results in a 0.25% error, or misincorporation, rate. In the majority of cases however, this is not a problem. The problem itself will be totally eliminated when a heat-stable polymerase with an efficient proofreading function becomes commercially available.

Finally, what is known as the 'plateau effect' must be taken into consideration. In a PCR experiment the amplification is not infinite. Rather, after a certain number of cycles the desired fragment gradually stops accumulating exponentially and enters a linear, or stationary, phase. This is a limiting factor in all PCR experiments.

Overall though, there can be no doubt that the incredible power, and sheer usefulness of PCR vastly outweighs any limitations. The elegant simplicity and efficiency of this reaction has really revolutionised many areas of science, with further applications coming to light all the time.

However such highly conserved sequences have been identified. Also, primers and probes for these regions have been designed. Even so, standard PCR experiments lack the extreme sensitivity required. So a super-sensitive variation of PCR has been developed, and this is called "Nested PCR".

Nested PCR is performed in a two-step sequence. The first PCR reaction is performed with a pair of outer primers and the second with a pair of inner, or "nested" primers (Figure 4). A typical protocol for such an experiment is that after 24 cycles of the first PCR, one tenth of the resulting product is amplified for a further 30 cycles with the corresponding inner primers.

In practice, for HIV diagnosis several such primer sets are used simultaneously, and the results are excellent! This detection method has indeed prove extraordinarily sensitive and effective in diagnosing HIV infection.

Yet PCR offers even further benefits over other detection techniques, for example in the detection of mother-to-child HIV transmission. Traditionally, HIV infection has been diagnosed by detection of HIV antibodies in the bloodstream. An individual possessing such antibodies is termed 'seropositive'. However, in a newborn child, maternal HIV antibodies may circulate for up to 15 months after birth, making the child seropositive. The important question is whether the child is HIV infected or not, and the is where PCR comes in.

By using PCR the HIV virus can be detected directly. In fact nested PCR is able to detect one copy of HIV proviral DNA against a background of 105 peripheral blood cells. Hence HIV infection can be detected in children 4-9 weeks old, where previously no such detection would be possible until the child was much older. Such PCR based studies have recently shown that somewhere between 20-60% of children born from seropositive mothers are infected by the HIV virus.

Not only is sensitive detection available using PCR, but also quantitation of the HIV virus is also possible. PCR can be used to quantitate proviral DNA, that is, determine the number of infected cells in HIV positive individuals. In people classed as seropositive, about 1 in 10⁴ T-cells contain proviral DNA. In people with full-blown AIDS on the other hand, about 1 in 10² T-cells contain proviral DNA. So PCR-based methods can be used to determine the extent of infection. As a consequence PCR is currently being used as a tool to monitor the effects of potential anti-HIV drugs. One example of this is with the drug 2',3'-dideoxyinosine, where in one case a patient's proviral DNA was found, using PCR methods, to have decreased by about 75% over a 15 week treatment. Obviously PCR has proved, and will continue to prove itself, a mighty tool in the study of infectious diseases such as HIV.