Source: R.K. Saiki, The design and optimization of the PCR. In PCR Technology: Principles and Applications for DNA Amplification (H.A. Erlich, Ed.), Stockton Press, NY, pp. 7-16 (1989).
PCR - Introduction
PCR is an in vitro method for the enzymatic synthesis of specific DNA sequences, using two oligonucleotide primers that hybridize to opposite strands and flank the region of interest in the target DNA. A repetitive series of cycles involving template denaturation, primer annealing, and the extension of the annealed primers by DNA polymerase results in the exponential accumulation of a specific fragment whose termini are defined by the 5' ends of the primers. Because the primer extension products synthesized in one cycle can serve as a template in the next, the number of target DNA copies approximately doubles at every cycle. Thus, 20 cycles of PCR yields about a million-fold (220) amplification. (Consult figure in your manual, p. 200.)Initially, the PCR used the Klenow fragment of E. coli DNA polymerase I to extend the annealed primers. This enzyme was inactivated by the high temperature required to separate the two DNA strands at the outset of each PCR cycle. Consequently, fresh enzyme had to be added during every cycle. The introduction of the thermostable DNA polymerase (Taq polymerase) isolated from Thermus aquaticus transformed the PCR into a simple and robust reaction which could now be automated by a thermal cycling device. The reaction components (template, primers, Taq polymerase, dNTP's, and buffer) could all be assembled and the amplification reaction carried out by simply cycling the temperature within the reaction tube. Although, for any given pair of oligonucleotide primers, an optimal set of conditions can be established, there is no single set of conditions that will be optimal for all possible reactions.
This popularity of PCR is primarily due to its apparent simplicity and high probability of success. Reduced to its most basic terms, PCR merely involves combining a DNA sample with oligonucleotide primers, deoxynucleotide triphosphates, and the thermostable Taq DNA polymerase in a suitable buffer, then repetitively heating and cooling the mixture for several hours until the desired amount of amplification is achieved. In fact, the PCR is a relatively complicated and, as yet, incompletely understood biochemical brew where constantly changing kinetic interactions among the several components determine the quality of the products obtained. Although the results will be good in most cases, there are a number of parameters that can be explored if better results are required or if the reaction fails altogether.
The "standard" PCR reaction
Because of the wide variety of applications in which PCR is being used, it is probably impossible to describe a single set of conditions that will guarantee success in all situations. Nevertheless, the reaction outlined below should prove to be adequate for most amplifications and for those cases where it isn't, it at least defines a common starting point from which changes can be attempted.The standard PCR is typically done in a 50- or 100-ml volume and, in addition to the sample DNA, contains 50 mM KCl, 10 mM Tris-HCl (pH 8.4 at room temp.), 1.5 mM MgCl2, 100 mg/ml gelatin, 0.25 mM of each primer, 200 mM of each deoxynucleoside triphosphate (dATP, dCTP, dGTP and dTTP), and 2.5 units of Taq polymerase. The type of the DNA sample will be variable, of course, but it will usually have between 102 to l05, copies of template (e.g., 0.1 mg human genomic DNA). A few drops of mineral oil are often added to seal the reaction and prevent condensation, except when a heated lid is available.
The amplification can be conveniently performed in a DNA Thermal Cycler using the "Step-Cycle" program set to denature at 94oC for 20 sec, anneal at 55oC for 20 sec, and extend at 72oC for 30 sec for a total of 30 cycles. (The "Step-Cycle" program causes the instrument to heat and cool to the target temperatures as quickly as possible. This results in a heating rate of about 0.3oC per sec and a cooling rate of about 1oC per sec, for an overall single cycle time of approximately 3.75 min.) These conditions can be used to amplify a wide range of target sequences with excellent specificity. However, for those reactions where the conditions described above do not produce the desired results, the following sections describe some ways in which a PCR can be improved.
PCR primer selection
Unfortunately, the approach to the selection of efficient and specific primers remains somewhat empirical. There is no set of rules that will ensure the synthesis of an effective primer pair. Yet it is the primers more than anything else that determine the success or failure of an amplification reaction. Fortunately, the majority of primers can be made to work and the following guidelines will help in their design.
Most primers will be between 20 and 30 bases in length and the optimal amount to use in an amplification will vary. Longer primers may be synthesized but are seldom necessary. Sequences not complementary to the template can be added to the 5'-end of the primers. These exogenous sequences become incorporated into the double-stranded PCR product and provide a means of introducing restriction sites or regulatory elements (e.g., promoters) at the ends of the amplified target sequences. If required, shorter primers or degenerate primers can be used as long as the thermal profile of the reaction is adapted to reflect the lower stability of the primed template. (For highly degenerate primers, it is preferable that the most unambiguous sequence be situated at the 3'-end of the primer, even to the extent of synthesizing a multiple series, in which the various permutations of the 3' sequence are held constant.) In general, concentrations ranging from 0.05 to 0.5 mM of each oligonucleotide should be acceptable.
- Where possible, select primers with a random base distribution and with a GC content similar to that of the fragment being amplified. Try to avoid primers with stretches of polypurines, polypyrimidines, or other unusual sequences.
- Avoid sequences with significant secondary structure, particularly at the 3'-end of the primer. Computer programs are very useful for revealing these structures, as well as incompatible primers on the basis of disparate predicted Tm's.
- Check the primers against each other for complementarity. In particular, avoiding primers with 3' overlaps will reduce the incidence of "primer dimer".
"Primer dimer" is an amplification artifact often observed in the PCR product, especially when many cycles of amplification are performed on a sample containing very few initial copies of template. It is a double-stranded fragment whose length is very close to the sum of the two primers and appears to occur when one primer is extended by the polymerase over the other primer. The resulting concatenation is an extremely efficient PCR template that can, if it occurs at an early cycle, easily overwhelm a reaction and become the predominant product.
The exact mechanism by which primer dimer is formed is not completely clear. The observation that primers with complementary 3'-ends are predisposed to dimer formation suggests that transient interactions that bring the termini in close proximity are the initiating event. Several polymerases, including Taq, have been shown to have a weak non-template-directed polymerization activity which can attach additional bases to a blunt-ended duplex. If such an activity were also to occur on a single-stranded oligonucleotide, there is a good chance that the extension would form a short 3' overlap with the other primer sufficient to promote dimerization. In any event, if dimers present an obstacle, they can be reduced somewhat by using minimal concentrations of primers and enzyme.
The PCR buffer
Changes to the PCR reaction buffer will usually affect the outcome of the amplification. In particular, the concentration of MgCl2 can have a profound effect on the specificity and yield of an amplification. Concentrations of about 1.5 mM are usually optimal, but in some circumstances, different amounts of Mg++ may prove to be necessary. Generally, excess Mg++ will result in the accumulation of non-specific amplification products and insufficient Mg++ will reduce the yield. More recently, it has been shown that the reduction or elimination of KCl and gelatin can sometimes be beneficial. Some protocols include 10% dimethyl sulfoxide (DMSO) ostensibly to reduce the secondary structure of the target DNA; note that DMSO can be slightly inhibitory to Taq polymerase and may decrease the overall yield of the amplification product.The deoxynucleotide triphosphates (dATP, dCTP, dGTP, and dTTP ) are usually present at 50 to 200 mM of each. Higher concentrations may tend to promote misincorporations by the polymerase and should be avoided. At 50 and 200 mM, there is sufficient precursor to synthesize approximately 6.5 and 25 mg of DNA, respectively. As deoxynucleotide triphosphates appear to quantitatively bind Mg++, the amount of dNTPs present in a reaction will determine the amount of free magnesium available. In the standard reaction, all four triphosphates are added to a final concentration of 0.8 mM; this leaves 0.7 mM of the original 1.5 mM MgCl2 not complexed with dNTP. Consequently, if the concentration is changed significantly, a compensatory change in MgCl2 may be necessary. The photo below displays the quantitative and qualitative effects of Mg++ concentration (0.5 to 10 mM) on PCR product (1.8 kb fragment of the human b-globin gene).
The concentration of Taq polymerase typically used in PCR is about 2.5 units per 100 ml reaction. For amplification reactions involving DNA samples with high sequence complexity, such as genomic DNA, there is an optimum concentration of Taq polymerase, usually 1 to 4 units per l00 ml. Increasing the amount of enzyme beyond this level can result in greater production of non-specific PCR products and reduced yield of the desired target fragment. Gel Photo
Cycling parameters
PCR is performed by incubating the samples at three temperatures, corresponding to the three steps in a cycle of amplification: denaturation, annealing, and extension. This cycling can be accomplished either manually with pre-set water baths, or automatically with the DNA Thermal Cycler. In a typical reaction, the double-stranded DNA is denatured by briefly heating the sample to 90-95oC (denaturation), the primers are allowed to anneal to their complementary sequences by briefly cooling to 40-60oC (annealing), followed by heating to 70-75oC to extend the annealed primers with the Taq polymerase (extension).The time of incubation at 70-75oC (extension) varies according to the length of the target being amplified; allowing 1 min for each kilobase of sequence is almost certainly excessive, but it is a good place to begin. Shorter times should be tried once the other amplification conditions have been established. The extension step can be eliminated altogether if the target sequence is approximately 150 bases or less. The polymerase retains significant activity at lower temperatures and complete extension will occur during the thermal transition from annealing to denaturation.
The ramp time, or time taken to change from one temperature to another, depends on the type of equipment used. With some notable exceptions, this rate of temperature change is not important and the fastest ramps attainable are used to shorten the cycle time. However, in order to be certain that the samples reach the intended temperatures, the actual ramp times for a particular setup should be determined by measuring the sample temperature during a test amplification. Insufficient heating during the denaturation step is a common cause of failure in a PCR reaction. It is very important that the reaction reaches a temperature at which complete strand separation occurs. A temperature of about 94oC should be adequate in most cases. As soon as the sample reaches 94oC, it can be cooled to the annealing temperature. Extensive denaturation is probably unnecessary and limited exposure to elevated temperatures helps maintain maximum polymerase activity throughout the reaction.
The temperature at which annealing is done depends on the length and GC content of the primers. A temperature of 55oC is a good starting point for typical 20-base oligonucleotide primers with about 50% GC contents; even higher temperatures may be necessary to increase primer specificity. Because of the very large molar excess of primers present in the reaction mix, hybridization occurs almost instantaneously and long incubation at the annealing temperature is not required. (The 20 sec denaturation and annealing incubation times used with the Step-Cycle program on the Thermal Cycler is the amount of time it takes for a 100 ml reaction in a 0.5-ml microcentrifuge tube to equilibrate with the block temperature.)
In some cases, primers of only 12 to 15 bases are available and an annealing temperature around 40-45oC is needed. However, primers of that length are unlikely to remain annealed at the 72oC extension temperature. This problem can be overcome by taking advantage of the partial enzymatic activity of the polymerase at lower temperatures to extend the primers by several bases and stabilize them. This is accomplished either by an intermediate incubation at 50-60oC or by heating gradually from 40oC to 72oC. Degenerate primers will often have multiple mismatches with their target sequence and should be treated in a similar manner.
It is often possible to anneal and extend the primers at the same temperature. In addition to simplifying the procedure to a two-temperature cycle, simultaneously annealing and extending at a temperature greater than 55oC may further improve the specificity of the reaction.
Amplification plateau
The amplification reaction is not infinite. After a certain number of cycles the desired amplification fragment gradually stops accumulating exponentially and enters a linear or stationary phase. This second stage of the reaction is called the "plateau". The point at which a PCR reaction reaches its plateau depends primarily on the number of copies of target originally present in the sample and by the total amount of DNA synthesized.In addition to the trivial possibilities, such as exhaustion of primer or dNTP or inactivation of polymerase or dNTP, there remains three more significant causes of plateau; these include substrate excess conditions, competition by non-specific products, and product reassociation. Substrate excess is simply the result of having synthesized more DNA than the amount of Taq polymerase present in the reaction is capable of replicating in the allotted extension time. For a standard l00 ml reaction containing 2.5 units of Taq polymerase, substrate excess conditions begin to occur around 1 mg of DNA (3 nmol of deoxynucleotide monophosphate). It can be overcome by increasing the extension time and/or increasing the amount of enzyme in the reaction. This is usually not practical, however, because each succeeding cycle would require the doubling of extension time and/or polymerase to continue exponential growth.
Competition by non-specific amplification products is closely related to substrate excess conditions. In this case, the unwanted DNA fragments compete with the desired fragment for the attention of the limiting DNA polymerase. Clearly, this problem can be alleviated by increasing the specificity of the reaction so that the non-target sequences are not allowed to accumulate to any significant extent.
Further accumulation of product can also be attenuated by reassociation of the single-stranded PCR fragments before the annealed primers can be extended. Whether this inhibition is due to branch migration and displacement of the primer or by inefficient displacement synthesis on the part of the polymerase itself is unclear. This limitation usually becomes significant when the product concentration approaches 10 pmol per 100 ml and is difficult to avoid except by dilution of the reaction.
In most of these cases, the plateau is an unavoidable and inherent limitation of the PCR reaction. Luckily, however, by the time it occurs sufficient amounts of product will have accumulated for almost any purpose. For those few situations where even more material is needed, it is probably much easier to set up multiple reactions than to try to avoid the plateau.
RT-PCR
When the target sequence is RNA, it is necessary to prepare a single-stranded or double-stranded DNA molecule, one strand of which is complementary to a particular RNA molecule. This can be easily accomplished in the laboratory using the enzyme reverse transcriptase (RT). The figure below shows one of the several ways by which this enzyme can be used. The enzyme is incapable of initiating polymerization; it requires an oligonucleotide primer that is hydrogen-bonded to the RNA and that bears a 3'-OH group. When eukaryotic messenger RNA is used, priming is especially simple, since these RNA molecules are naturally terminated at the 3' end with a long sequence of A's (poly A), allowing the use of an oligo(dT) primer. With other RNA molecules a mixture of random oligonucleotides can be used, since in the mixture there will invariably be an oligonucleotide that can base-pair with the terminus of the RNA. Addition of reverse transcriptase to the primed RNA results in synthesis of a DNA-RNA hybrid. To prepare double-stranded DNA, the RNA strand must be removed from the hybrid; this can be done in one of two ways: (1) a ribonuclease activity (RNase H) of reverse transcriptase will under certain conditions remove the RNA; (2) the hybrid can be treated with alkali, which hydrolyzes the RNA and leaves the single-stranded DNA untouched. The DNA strand complementary to the newly synthesized strand is usually made in one of two ways; in both cases the E. coli DNA polymerizing enzyme DNA polymerase I, which also needs a primer, is used. The figure below shows a procedure in which priming is again accomplished by addition of a random collection of oligonucleotides. The second procedure (not shown) makes use of a poorly understood peculiarity of the polymerization reaction catalyzed by reverse transcriptase. If the initial synthesis of the first single strand of DNA is carried out for a longer time than is necessary to complete the strand, and if certain solution conditions are used, the DNA strand extends beyond the RNA template strand with a short segment whose sequence is complementary to that of the last few hydrogen-bonded bases. Once the RNA is removed, this extended segment of DNA folds back and forms a hairpinlike structure, thereby eliminating the need for an added primer. However, when polymerase I completes synthesis of the second DNA strand, the result is a full molecular hairpin. S1 nuclease, an enzyme specific for single-stranded DNA, or single-stranded segments of double-stranded DNA, can be used to cleave the hairpin; the result is a proper double-stranded DNA molecule.RT-PCR Schematic