Chapter 1: Nucleic Acid Amplification

Introduction

The polymerase chain reaction (PCR) is a relatively simple technique that amplifies a DNA template to produce specific DNA fragments in vitro. Traditional methods of cloning a DNA sequence into a vector and replicating it in a living cell often require days or weeks of work, but amplification of DNA sequences by PCR requires only hours. While most biochemical analyses, including nucleic acid detection with radioisotopes, require the input of significant amounts of biological material, the PCR process requires very little. Thus, PCR can achieve more sensitive detection and higher levels of amplification of specific sequences in less time than previously used methods. These features make the technique extremely useful, not only in basic research, but also in commercial uses, including genetic identity testing, forensics, industrial quality control and in vitro diagnostics. Basic PCR has become commonplace in many molecular biology labs where it is used to amplify DNA fragments and detect DNA or RNA sequences within a cell or environment. However, PCR has evolved far beyond simple amplification and detection, and many extensions of the original PCR method have been described. This chapter provides an overview of different types of PCR methods, applications and optimization. A detailed treatment of these methods is beyond the scope of this publication. However, an extensive bibliography is provided in the References section for researchers who require more comprehensive information.

Basic PCR

The PCR process was originally developed to amplify short segments of a longer DNA molecule (Saiki et al. 1985). A typical amplification reaction includes the target DNA, a thermostable DNA polymerase, two oligonucleotide primers, deoxynucleotide triphosphates (dNTPs), reaction buffer and magnesium. Once assembled, the reaction is placed in a thermal cycler, an instrument that subjects the reaction to a series of different temperatures for varying amounts of time. This series of temperature and time adjustments is referred to as one cycle of amplification. Each PCR cycle theoretically doubles the amount of targeted sequence (amplicon) in the reaction. Ten cycles theoretically multiply the amplicon by a factor of about one thousand; 20 cycles, by a factor of more than a million in a matter of hours.

Each cycle of PCR includes steps for template denaturation, primer annealing and primer extension (Figure 1.1). The initial step denatures the target DNA by heating it to 94°C or higher for 15 seconds to 2 minutes. In the denaturation process, the two intertwined strands of DNA separate from one another, producing the necessary single-stranded DNA template for replication by the thermostable DNA polymerase. In the next step of a cycle, the temperature is reduced to approximately 40–60°C. At this temperature, the oligonucleotide primers can form stable associations (anneal) with the denatured target DNA and serve as primers for the DNA polymerase. This step lasts approximately 15–60 seconds. Finally, the synthesis of new DNA begins as the reaction temperature is raised to the optimum for the DNA polymerase. For most thermostable DNA polymerases, this temperature is in the range of 70–74°C. The extension step lasts approximately 1–2 minutes. The next cycle begins with a return to 94°C for denaturation.

Each step of the cycle should be optimized for each template and primer pair combination. If the temperature during the annealing and extension steps are similar, these two steps can be combined into a single step in which both primer annealing and extension take place. After 20–40 cycles, the amplified product may then be analyzed for size, quantity, sequence, etc., or used in further experimental procedures.

An animated presentation illustrating the PCR process is available.

RT-PCR

The thermostable DNA polymerases used in the basic PCR process require a DNA template, and as such, the technique is limited to the analysis of DNA samples. Yet numerous instances exist in which the amplification of RNA would be preferred. This is particularly true in analyses involving the differential expression of genes in tissues during development or the cloning of cDNAs from rare messages.

In order to apply PCR to the study of RNA, the RNA sample must first be reverse transcribed to cDNA to provide the necessary DNA template for the thermostable polymerase (Figure 1.2). This process is called reverse transcription (RT), hence the name RT-PCR.

Avian myeloblastosis virus (AMV) or Moloney murine leukemia virus (M-MLV or MuLV) reverse transcriptases are generally used to produce a DNA copy of the RNA template using either random primers, an oligo(dT) primer or a sequence-specific primer. Alternatively, some thermostable DNA polymerases (e.g., Tth DNA polymerase) possess a reverse transcriptase activity, which can be activated under certain conditions, namely using manganese instead of magnesium as a cofactor (Myers and Gelfand, 1991). After this initial reverse transcription step has produced the cDNA template, basic PCR is carried out to amplify the target sequence.

The quality and purity of the starting RNA template is crucial to the success of RT-PCR. Either total RNA or poly(A)+ RNA can be used as the starting template, but both must be intact and free of contaminating genomic DNA. Specific capture of poly(A)+ RNA will enrich a targeted message so that less of the reverse transcription reaction is needed for the subsequent amplification. The efficiency of the first-strand synthesis reaction, which can be related to the quality of the RNA template, will also significantly impact the results of the subsequent amplification.

Long PCR

Amplification of long DNA fragments is desirable for numerous applications such as physical mapping applications (Rose, 1991) and direct cloning from genomes. While basic PCR works well when smaller fragments are amplified, the efficiency of amplification (and therefore the yield of amplified fragments) decreases significantly as the size of the amplicon increases over 5kb. This decrease in yield is attributable to the accumulation of truncated products, which are not suitable substrates for the subsequent cycles of amplification. This is evidenced by the presence of smeared, as opposed to discrete, bands on a gel.

In 1994, Wayne Barnes (Barnes, 1994) and other researchers (Cheng et al. 1994) analyzed the factors affecting polymerization across larger regions of DNA by thermostable DNA polymerases and identified key variables affecting the yield of longer PCR fragments. Most significant of these was an approach relying on a mixture of two thermostable polymerases. The first polymerase lacks a 3′→5′ exonuclease (proofreading) activity; the second enzyme, present at a reduced concentration, contains a potent proofreading activity. Presumably, when the nonproofreading DNA polymerase (e.g., Taq DNA polymerase) misincorporates a dNTP, subsequent extension of the newly synthesized DNA either proceeds very slowly or stops completely. The proofreading polymerase (e.g., Pfu DNA polymerase or Tli DNA polymerase) serves to remove the misincorporated nucleotide, allowing the DNA polymerases to continue extension of the new strand.

Although the use of two thermostable DNA polymerases can significantly increase yield, other conditions can have a significant impact on the yield of longer PCR products (Cheng et al. 1995). Logically, longer extension times can increase the yield of longer PCR products because fewer partial products are synthesized. Extension times of 10–20 minutes are common and depend on the length of the target. In addition, template quality is crucial. Depurination of the template, which is promoted at elevated temperatures and lower pH, will result in more partial products and decreased overall yield. In long PCR, the denaturation time is reduced to 2–10 seconds to decrease depurination of the template. Additives, such as glycerol and dimethyl sulfoxide (DMSO), also help lower the strand-separation and primer-annealing temperatures, alleviating some of the depurination effects of high temperatures. Cheng et al. (Cheng et al. 1995) also found that reducing potassium concentrations by 10–40% increased the efficiency of amplification of longer products.

Quantitative End-Point PCR

PCR and RT-PCR are generally used in a qualitative format for evaluating biological samples. However, a wide variety of applications, such as the determining viral load, measuring responses to therapeutic agents and characterizing gene expression, would be improved by quantitative determination of target abundance. Theoretically, this should be easy to achieve, given the exponential nature of PCR, because a linear relationship exists between the number of amplification cycles and the logarithm of the number of molecules. In practice, however, the efficiency of amplification is decreased because of contaminants (inhibitors), competitive reactions, substrate exhaustion, inactivation of the polymerase and target reannealing. As the number of cycles increases, the amplification efficiency decreases, eventually resulting in a plateau effect.

Normally, quantitative PCR requires the measurement to be taken before the plateau phase, so the relationship between the number of cycles and molecules is relatively linear. This point must be determined empirically for different reactions because of the numerous factors that can affect the amplification efficiency. Because the measurement is taken prior to the reaction plateau, quantitative PCR uses fewer amplification cycles than basic PCR. This can cause problems in the detection of the final product because there is less product to detect.

To monitor the efficiency of amplification, many applications are designed to include an internal standard in the PCR. One such approach includes a second primer pair that is specific for a “housekeeping” gene (i.e., a gene that has constant expression levels among the samples compared) in the reaction (Gaudette and Crain, 1991; Murphy et al. 1990). Amplification of housekeeping genes verifies that the target nucleic acid and reaction components were of acceptable quality but does not account for differences in amplification efficiencies due to differences in product size or primer annealing efficiency between the internal standard and target being quantified.

The concept of competitive PCR—a variation of quantitative PCR—is a response to this limitation. In competitive PCR, a known amount of a control template is added to the reaction. This template is amplified using the same primer pair as the experimental target molecule but yields a distinguishable product (e.g., different size, restriction digest pattern, etc.). The amounts of control and test product are compared after amplification. While these approaches control for the quality of the target nucleic acid, buffer components and primer annealing efficiencies, they have their own limitations (Siebert and Larrick, 1993; McCulloch et al. 1995), including the fact that many depend on final analysis by electrophoresis.

Numerous fluorescent solution and solid-phase assays have been described to measure the amount of amplification product generated in each reaction, but they can fail to discriminate amplified DNA of interest from nonspecific amplification products. Some of these analyses rely on blotting techniques, which introduce another variable due to nucleic acid transfer efficiencies, while other assays have been developed to eliminate the need for gel electrophoresis yet provide the requisite specificity. Real-time PCR, which provides the ability to view the results of each amplification cycle, is a popular way of overcoming the need for analysis by electrophoresis.

Quantitative Real-Time PCR

The use of fluorescently labeled oligonucleotide probes or primers or DNA-binding fluorescent dyes, such as SYBR^® Green, to detect and quantitate a PCR product allows quantitative PCR to be performed in real time. DNA-binding dyes are easy to use but do not differentiate between specific and nonspecific PCR products. Fluorescently labeled nucleic acid probes have the advantage that they react with only specific PCR products. These probes can also be used to detect single nucleotide polymorphisms (Lee et al. 1993; Bernard et al. 1998). However in many cases, these approaches are not conducive to multiplex reactions, and there is no convenient way of distinguishing specific and nonspecific PCR products. A more recent technology, the Plexor™ technology, requires only a single fluorescently labeled primer, is compatible with multiplex PCR and allows specific and nonspecific amplification products to be differentiated (Sherrill et al. 2004; Frackman et al. 2006).

Real-time PCR using labeled oligonucleotide probes or primers employs two different fluorescent reporters and relies on the transfer of energy from one reporter (the energy donor) to the second reporter (the energy acceptor) when the reporters are in close proximity. The second reporter can be a quencher or a fluor. If the second reporter is a quencher, the energy from the first reporter is absorbed but re-emitted as heat rather than light, leading to a decrease in the fluorescent signal. Alternatively, if the second reporter is a fluor, the energy can be absorbed and re-emitted at another wavelength through fluorescent resonance energy transfer (FRET, reviewed in Didenko, 2001), and the progress of the reaction can be monitored by the decrease in fluorescence of the energy donor or the increase in fluorescence of the energy acceptor. During the exponential phase of PCR, the change in fluorescence is proportional to the accumulation of PCR product. To simplify quantitation, specially designed instruments perform both the thermal cycling steps to amplify the target and the fluorescence detection to monitor the change in fluorescence in real time during each PCR cycle.

There are several general categories of real-time PCR probes, including hydrolysis, hairpin and simple hybridization probes. These probes contain a complementary sequence that allows the probe to anneal to the accumulating PCR product, but probes can differ in the number and location of the fluorescent reporters.

Hydrolysis probes are labeled with a fluor at the 5′-end and a quencher at the 3′-end, and because the two reporters are in close proximity, the fluorescent signal is quenched. During the annealing step, the probe hybridizes to the PCR product generated in previous amplification cycles. The resulting probe:target hybrid is a substrate for the 5′→3′ exonuclease activity of the DNA polymerase, which degrades the annealed probe and liberates the fluor (Holland et al. 1991). The fluor is freed from the effects of the energy-absorbing quencher, and the progress of the reaction and accumulation of PCR product is monitored by the resulting increase in fluorescence. With this approach, preliminary experiments must be performed prior to the quantitation experiments to show that the signal generated is proportional to the amount of the desired PCR product and that nonspecific amplification does not occur.

Hairpin probes, also known as molecular beacons, contain inverted repeats separated by a sequence complementary to the target DNA. The repeats anneal to form a hairpin structure, where the fluor at the 5′-end and a quencher at the 3′-end are in close proximity, resulting in little fluorescent signal. The hairpin probe is designed so that the probe binds preferentially to the target DNA rather than retains the hairpin structure. As the reaction progresses, increasing amounts of the probe anneal to the accumulating PCR product, and as a result, the fluor and quencher become physically separated. The fluor is no longer quenched, and the level of fluorescence increases. One advantage of this technique is that hairpin probes are less likely to mismatch than hydrolysis probes (Tyagi et al. 1998). However, preliminary experiments must be performed to show that the signal is specific for the desired PCR product and that nonspecific amplification does not occur.

The use of simple hybridization probes involves using two labeled probes or using one labeled probe and a labeled PCR primer. In the first approach, the energy emitted by the fluor on one probe is absorbed by a fluor on the second probe, which hybridizes nearby. In the second approach, the emitted energy is absorbed by a second fluor that has been incorporated into the PCR product as part of the PCR primer. Both of these approaches result in increased fluorescence of the energy acceptor and decreased fluorescence of the energy donor. The use of hybridization probes can be simplified even further so that only one labeled probe is required. In this approach, quenching of the fluor by deoxyguanosine is used to bring about a change in fluorescence (Crockett and Wittwer, 2001; Kurata et al. 2001). The labeled probe anneals so that the fluor is in close proximity to G residues within the target sequence, and as probe annealing increases, the level of fluorescence in the reaction decreases due to deoxyguanosine quenching. With this approach, the location of probe is limited because the probe must hybridize so the fluorescent dye is very near a G residue. The advantage of simple hybridization probes is their ability to be multiplexed more easily than hydrolysis and hairpin probes through the use of differently colored fluors and probes with different melting temperatures (reviewed in Wittwer et al. 2001).

The Plexor™ qPCR and qRT-PCR Systems require no probes, only two PCR primers, one of which is fluorescently labeled. These systems take advantage of the specific interaction between two modified nucleotides (Sherrill et al. 2004; Johnson et al. 2004; Moser and Prudent, 2003). The two novel bases, isoguanine (iso-dG) and 5′-methylisocytosine (iso-dC), form a unique base pair in double-stranded DNA ( Johnson et al. 2004). To perform fluorescent quantitative PCR using this new technology, one primer is synthesized with an iso-dC residue as the 5′-terminal nucleotide and a fluorescent label at the 5′-end; the second primer is unlabeled. During PCR, this labeled primer is annealed and extended, becoming part of the template used during subsequent rounds of amplification, and the complementary iso-dGTP, which is available in the nucleotide mix as dabcyl-iso-dGTP, pairs specifically with iso-dC. When the dabcyl-iso-dGTP is incorporated, the close proximity of dabcyl and the fluorescent label on the opposite strand effectively quenches the fluorescent signal. This process is illustrated in Figure 1.3. The initial fluorescence level of the labeled primers is high in Plexor™ System reactions. As amplification product accumulates, the signal decreases.

Quenching of the fluorescent label by dabcyl is a reversible process. Fluorescence is quenched when the product is double-stranded. Denaturing the product separates the label and quencher, resulting in an increased fluorescent signal. Consequently, thermal melt curves can be generated by allowing all product to form double-stranded DNA at a lower temperature (approximately 60°C) and slowly ramping the temperature to denaturing levels (approximately 95°C). The product length and sequence impact melting temperature (T_m), so the melt curve is used to characterize amplicon homogeneity. Nonspecific amplification can be identified by broad peaks in the melt curve or peaks with different T_m values. By distinguishing specific and nonspecific amplification products, the melt curve adds a quality control aspect during routine use. The generation of melting curves is not possible with technologies that rely on the 5′→3′ exonuclease activity of Taq DNA polymerase.

A benefit of the Plexor™ technology over detection using simple DNA-binding dyes, such as SYBR^® Green, is the capacity for multiplexing. The labeled primer can be tagged with one of many common fluorescent labels, allowing two- to four-color multiplexing, depending on the instrument used. The simplicity of primer design for the Plexor™ technology is a distinct advantage over probe-based quantitative PCR approaches. Also the Plexor™ technology does not rely on enzymatic cleavage to generate signal and does not have the complex hybridization kinetics that can be typical of other approaches to real-time PCR. The Plexor™ technology can also be used for quantitative RT-PCR.

Rapid Amplified Polymorphic DNA Analysis

Genetic analysis of organisms (animals, plants and bacteria) at the molecular level is an important and widely practiced area of genetic science. Several techniques developed over more than a decade offer the opportunity to identify each individual or type of individual in a species uniquely and unambiguously. PCR has impacted this area of analysis because of its ease of use and simplicity over traditional VNTR- and RFLP-based methods (Jeffreys et al. 1985; Sambrook and Russell, 2001).

One important PCR-based genetic analysis is random amplified polymorphic DNA analysis (RAPD; reviewed in McClelland and Welsh, 1994; Power, 1996; Black, 1993). RAPD uses small, nonspecific primers for the amplification of regions of genomic DNA. Successful primer pairs produce different banding profiles of PCR products between individuals, strains, cultivars or species when analyzed by gel electrophoresis.

Slight modifications to the basic PCR method are made for RAPD. First, the primers are approximately 10 bases in length compared to the 17- to 23-base primer length of normal PCR. Because the primers are shorter, the temperature of the annealing reaction is reduced to less than 40°C.

As with most PCR techniques, RAPD requires very little material for analysis and is relatively insensitive to the integrity of the material. No blotting techniques are required, thus eliminating the use of ³²P, bypassing probe generation and decreasing the amount of time required to obtain results.

Rapid Amplification of cDNA Ends (RACE)

Rapid amplification of cDNA ends (RACE) is a variation of RT-PCR that amplifies unknown cDNA sequences corresponding to the 3′- or 5′-end of the RNA. Numerous variations of the original protocols have been published (Troutt et al. 1992; Edwards et al. 1991; Edwards et al. 1993; Liu and Gorovsky, 1993; Fromont-Racine et al. 1993; reviewed in Schaefer, 1995) but will not be discussed in detail here.

Two general RACE strategies exist: one amplifies 5′ cDNA ends (5′ RACE) and the other captures 3′ cDNA end sequences (3′ RACE). In either strategy, the first step in the RACE reaction involves the conversion of RNA into single-stranded cDNA using a reverse transcriptase. For the subsequent amplification reaction, two PCR primers are designed to flank the unknown sequence. One PCR primer is complementary to known sequences within the gene, and a second primer is complementary to an “anchor” site (anchor primer). The anchor site may be present naturally, such as the poly(A) tail of most mRNAs, or can be added in vitro after completion of the reverse transcription step. The anchor primer can also carry adaptor sequences, such as restriction enzyme recognition sites, to facilitate subsequent cloning of the amplified product. Amplification using these two PCR primers results in a product that spans the unknown 5′ or 3′ cDNA sequence, and sequencing this product will reveal the unknown sequence. The information obtained from partial cDNA sequences can then be used to assemble the sequence of the full-length cDNA (Frohman et al. 1988; Loh et al. 1989; Ohara et al. 1989).

In 5′ RACE (Figure 1.4), the first-strand cDNA synthesis reaction is primed using an oligonucleotide complementary to a known sequence in the gene. After removing the RNA template, an anchor site at the 3′-end of the single-stranded cDNA is created using terminal deoxynucleotidyl transferase, which adds a nucleotide tail. A typical amplification reaction follows using an anchor primer complementary to the newly added tail and another primer complementary to a known sequence within the gene. Some variations to the original 5′ RACE procedure use different approaches to add an anchor site adjacent to the 5′-end sequences of the cDNA. One of these approaches, single-stranded ligation of cDNA (SLIC), uses RNA ligase to covalently join an oligonucleotide anchor site adjacent to the 3′-end of the single-stranded cDNA.

The 3′-RACE procedure (Figure 1.5) uses an oligo(dT) primer/adaptor as a primer for the reverse transcription reaction. The oligo(dT) primer anneals to the poly(A)+ tail of the mRNA. This oligo(dT) primer/adaptor is also used as the anchor primer in the subsequent amplifications along with a primer complementary to known sequences within the gene.

Differential Display PCR

Differential display PCR (DDPCR) is another variation of RT-PCR and is used to identify differences in mRNA expression patterns between two cell lines or populations. In one example of this procedure, cDNA synthesis is primed using a set of modified oligo(dT) primers, which anneal to the poly(A)+ tail of mRNA (Liang and Pardee, 1992). Each of the oligo(dT) primers carries an additional two nucleotides at the 3′-end. This ensures that extension only occurs if the primer anneals immediately adjacent to the junction between the poly(A)+ tail and 3′ end of the mRNA. Because the two additional nucleotides will only anneal to a subset of the mRNA molecules, this also reduces the complexity of the RNA population that is reverse transcribed. The RNA is first reverse transcribed with one of the modified oligo(dT) primers to synthesize first-strand cDNA, which is then amplified by PCR using two random 10mer primers. After amplification, the reaction products are visualized by gel electrophoresis, and comparisons of banding patterns are made between the two cell populations to identify differentially expressed cDNAs.

Another form of analyzing differences between complex genomes is representational difference analysis (RDA). This method combines “subtractive” library techniques (Lisitsyn et al. 1993) with PCR amplification to find differences in complex genomes. A variation of this is cDNA RDA, where total RNA from the cell populations is first converted into cDNA, subtractive techniques are performed and the products are amplified by PCR (Hubank and Schatz, 1994). By using cDNA, the complexity is significantly reduced, providing another method to analyze differences in expression between cell types or in response to various treatments. In this regard, cDNA RDA can be used as an alternative to DDPCR.

In situ PCR

In situ PCR, first described in 1990, combines the sensitivity of PCR or RT-PCR amplification with the cellular or histological localization associated with in situ hybridization techniques (Haase et al. 1990). These features make in situ PCR a powerful tool for detecting proviral DNA, oncogenesis and localization of rare messages.

The technique is amenable to analysis of fixed cells or tissue cross-sections. Detection of amplified products can be accomplished indirectly by subsequent hybridization using either radiolabeled, fluorescently labeled or biotin-labeled nucleic acid probes. PCR products can also be detected directly by the incorporation of a labeled nucleotide, although this method is subject to higher background levels.

The use of in situ PCR requires altering some of the reaction parameters typical of basic PCR (Nuovo et al. 1993; Thaker, 1999). For example, increased Mg²⁺ concentrations (approximately 4.5mM versus the normal 1.5–2.5mM) are used for in situ PCR. An increased amount of DNA polymerase is also required unless BSA is added to the reaction, presumably because the polymerase binds to the glass plate and coverslip.

Tissue preparation also plays a significant role in the success of in situ PCR. A strong relationship exists between the time of fixation and protease digestion and the intensity of PCR signal. Tissue preparation also impacts the amount of side reaction, resulting in primer-independent signals, which are not normally present in basic PCR. These primer-independent signals often arise from Taq DNA polymerase-mediated repair of single-stranded gaps in the genomic DNA.

As the use of the technique has spread, the process has been further optimized. Numerous publications (reviewed in Nuovo, 1995; Staskus et al. 1995) describe process improvements that increase sensitivity and decrease nonspecific amplification products. Additionally, several thermal cycler manufacturers have introduced instruments designed specifically for in situ amplification.

Hot-Start PCR

Hot-start PCR is a common technique to reduce nonspecific amplification due to the assembly of amplification reactions at room temperature or on ice. At these lower temperatures, PCR primers can anneal to template sequences that are not perfectly complementary. Since thermostable DNA polymerases have activity at these low temperatures (although in most cases the activity is less than 25%) the polymerase can extend misannealed primers. This newly synthesized region is perfectly complementary to the DNA template, allowing primer extension and the synthesis of undesired amplification products. However, if the reaction is heated to temperatures >60°C before polymerization begins, the stringency of primer annealing is increased, and the subsequent synthesis of undesired PCR products is avoided or reduced.

Hot-start PCR can also reduce the amount of primer-dimer synthesized by increasing the stringency of primer annealing. At lower temperatures, the PCR primers can anneal to each other via regions of complementarity, and the DNA polymerase can extend the annealed primers to produce primer dimer, which often appears as a diffuse band of approximately 50–100bp on an ethidium bromide-stained gel. The formation of nonspecific products and primer-dimer can compete for reagent availability with the amplification of the desired product. Thus, hot-start PCR can improve the yield of the specific PCR products.

To perform hot-start PCR, the reactions are assembled on ice or at room temperature, but one critical component is omitted until the reaction has been heated to 60–65°C, at which point the missing reagent is added. This omission prevents the polymerase from extending primers until the the critical component is added at the higher temperature where primer annealing is more stringent. However, this method is tedious and increases the risk of contamination. A second, less labor-intensive approach involves the reversible inactivation or physical separation of one or more critical components in the reaction. For example, the magnesium or DNA polymerase can be sequestered in a wax bead, which melts as the reaction is heated to 94°C during the denaturation step, releasing the component only at higher temperatures (Carothers et al. 1989; Krishnan et al. 1991; Clark, 1988). Alternatively, the DNA polymerase can be kept in an inactive state by binding to an oligonucleotide, also known as an aptamer (Lin and Jayasena, 1997; Dang and Jayasena, 1996) or an antibody (Scalice et al. 1994; Sharkey et al. 1994). This bond is disrupted at the higher temperatures, releasing the functional DNA polymerase.

High-Fidelity PCR

For some applications, such as gene expression, mutagenesis or cloning, the number of mutations introduced during PCR needs to be minimized. For these applications, we recommend using a proofreading polymerase. Proofreading DNA polymerases, such as Pfu and Tli DNA polymerases, have a 3′→5′ exonuclease activity, which can remove any misincorporated nucleotides, and so the error rate is relatively low. The accuracy of Pfu DNA polymerase is approximately twofold higher than that of Tli DNA polymerase and 6-fold higher than that of Taq DNA polymerase (Cline, 1996).

The most commonly used DNA polymerase for PCR is Taq DNA polymerase, which has an error rate of approximately 1 × 10^–5 errors per base. This error rate is relatively high due to the enzyme's lack of 3′→5′ exonuclease (proofreading) activity. The error rate of Tfl DNA polymerase, another nonproofreading polymerase, is similar to that of Taq DNA polymerase.

Reaction conditions can affect DNA polymerase fidelity, and DNA polymerases may be affected in different ways or to different degrees. In general, excess magnesium or the presence of manganese will cause the fidelity of DNA polymerases to be reduced (Eckert and Kunkel, 1990). Unequal nucleotide concentrations can also affect fidelity; nucleotides that are present at higher concentrations will be misincorporated at a higher frequency (Eckert and Kunkel, 1990). Reaction pH can also have a big effect on fidelity (Eckert and Kunkel, 1990; Eckert and Kunkel, 1991). For example, the fidelity of Taq DNA polymerase increases as pH decreases, with the lowest error rate occurring in the range of pH 5–6 (Eckert and Kunkel, 1990), but the opposite is true for Pfu DNA polymerase. Pfu DNA polymerase has higher fidelity at higher pH (Cline, 1996). Finally, exposing the DNA template to very high temperatures (i.e., 94°C) for extended periods of time can lead to DNA damage, specifically the release of bases from the phosphodiester backbone. The resulting abasic sites can cause some DNA polymerases to stall but can also result in a higher rate of mutations, most frequently transversions, as the DNA polymerase adds a random nucleotide at an abasic site (Eckert and Kunkel, 1991).

PCR and DNA Sequencing: Cycle Sequencing

The PCR process has also been applied to DNA sequencing in a technique called cycle sequencing (Murray, 1989; Saluz and Jost, 1989; Carothers et al. 1989; Krishnan et al. 1991). The main differences between a cycle sequencing reaction and a typical DNA sequencing reaction is the choice of polymerase and the incorporation of thermal cycling. Cycle sequencing reactions use thermostable polymerases, while the conventional reactions use thermolabile polymerases, such as modified T7 DNA polymerase or the Klenow fragment of E. coli DNA polymerase I.

Cycle sequencing reactions also differ from typical PCR amplification reactions in that they use only a single primer, resulting in a linear (as opposed to theoretically exponential) amplification of the target molecule. Other reaction components are comparable, and either radioactive or fluorescent labels may be incorporated for detection.

Cloning PCR Products

Amplification with a DNA polymerase lacking 3′→5′ (proofreading) exonuclease activity (e.g., Taq DNA polymerase) yields products that contain a single 3′-terminal nucleotide overhang, typically an A residue (Clark, 1988; Hu, 1993). Before this overhang was identified, blunt-end cloning of PCR products was inefficient at best. However, these PCR products can now be conveniently cloned into T-vectors, which contain a single T overhang (reviewed in Mezei and Storts, 1994; Guo and Bi, 2002).

DNA polymerases that possess proofreading activity (e.g., Tli DNA polymerase or Pfu DNA polymerase) generate blunt-ended PCR products. These products are compatible with standard blunt-end cloning strategies. Conversely, blunt-end PCR products can be tailed with Taq DNA polymerase and dATP prior to cloning into a T-vector (Zhou et al. 1995).

return to top of page

General Considerations for PCR Optimization

This discussion focuses on the use of Taq DNA polymerase in PCR, since this is the enzyme most commonly used in PCR. Many of these suggestions also apply when using other DNA polymerases.

Magnesium Concentration

Magnesium is a required cofactor for thermostable DNA polymerases, and magnesium concentration is a crucial factor that can affect the success of the amplification. Template DNA concentration, chelating agents present in the sample (e.g., EDTA or citrate), dNTP concentration and the presence of proteins can all affect the amount of free magnesium in the reaction. In the absence of adequate free magnesium, Taq DNA polymerase is inactive (Figure 1.6). Excess free magnesium reduces enzyme fidelity (Eckert and Kunkel, 1990) and may increase the level of nonspecific amplification (Williams, 1989; Ellsworth et al. 1993). For these reasons, researchers should empirically determine the optimal magnesium concentration for each reaction. To do so, perform a series of reactions containing 1.0–4.0mM Mg²⁺ in 0.5–1mM increments and visualize the results to determine which magnesium concentration produced the highest yield of product and the minimum amount of nonspecific product. The effect of magnesium concentration and the optimal concentration range can vary with the particular DNA polymerase. For example, the performance of Pfu DNA polymerase seems to be less dependent on magnesium concentration, but when optimization is required, the optimal concentration is usually in the range of 2–6mM.

Many DNA polymerases are supplied with a magnesium-free reaction buffer and a tube of 25mM MgCl₂, so you can adjust the Mg²⁺ concentration to the level that is optimal for each reaction. Before assembling the reactions, be sure to thaw the magnesium solution completely prior to use and vortex the magnesium solution for several seconds before pipetting. Magnesium chloride solutions can form concentration gradients as a result of multiple freeze-thaw cycles, and vortex mixing is required to obtain a uniform solution. These two steps, though seemingly simple, eliminate the cause of many failed experiments.

Some scientists prefer to use reaction buffers that already contain MgCl₂ at a final concentration of 1.5mM. It should be noted, however, that Hu et al. reported performance variability of reaction buffer solutions containing magnesium. The free magnesium changes of 0.6mM observed in their experiments dramatically affected amplification yields in an allele-specific manner. The authors found that heating the buffer at 90°C for 10 minutes restored the homogeneity of the solution. They postulated that magnesium chloride precipitates as a result of multiple freeze-thaw cycles (Hu et al. 1992).

Buffer Considerations

Most reaction buffers consist of a buffering agent, most often a Tris-based buffer, and salt, commonly KCl. The buffer regulates the pH of the reaction, which affects the DNA polymerase activity and fidelity. Modest concentrations of KCl will increase DNA polymerase activity by 50–60% over activities in the absence of KCl; 50mM KCl is considered optimal (Gelfand, 1989). GoTaq^® DNA Polymerase contains native Taq DNA polymerase in a proprietary formulation. It is supplied with 5X Green GoTaq^® Reaction Buffer and 5X Colorless GoTaq^® Reaction Buffer. The 5X Green GoTaq^® Reaction Buffer contains two dyes (blue and yellow) that separate during electrophoresis to monitor migration progress. The buffer also contains a compound that increases the density of the sample, so it will sink into the well of the agarose gel, allowing reactions to be directly loaded onto an agarose gel without the need for loading dye. The blue dye comigrates at the same rate as a 3–5kb DNA fragment in a 1% agarose gel. The yellow dye migrates at a rate faster than primers (<50bp) in a 1% agarose gel. The 5X Colorless GoTaq^® Reaction Buffer and the 5X Green GoTaq^® Reaction Buffer have the same formulation, except for the dyes. The 5X Colorless GoTaq^® Reaction Buffer is recommended for any applications where absorbance or fluorescence measurements will be taken of the PCR amplimer without prior clean-up. Both buffers are supplied at pH 8.5 and contain MgCl₂ at a concentration of 7.5mM for a final concentration of 1.5mM.

GoTaq^® Flexi DNA Polymerase is supplied with 5X Green GoTaq^® Flexi Reaction Buffer and 5X Colorless GoTaq^® Flexi Reaction Buffer. The compositions are identical to the 5X Green GoTaq^® Reaction Buffer and 5X Colorless GoTaq^® Reaction Buffer, except that the GoTaq^® Flexi reaction buffers do not contain MgCl₂. Instead, the GoTaq^® Flexi DNA Polymerase is supplied with a tube of 25mM MgCl₂, so reactions can be suppplemented with varying concentrations of magnesium.

Enzyme Concentration

We recommend using 1–1.25 units of Taq DNA polymerase in a 50μl amplification reaction. In most cases, this is an excess of enzyme, and adding more enzyme will not significantly increase product yield. In fact, increased amounts of enzyme increase the likelihood of generating artifacts associated with the intrinsic 5′→3′ exonuclease activity of Taq DNA polymerase, resulting in smeared bands in an agarose gel (Longley et al. 1990; Bell and DeMarini, 1991).

Pipetting errors are a frequent cause of excessive enzyme levels. Accurate dispensing of small volumes of enzyme solutions in 50% glycerol is difficult, so we strongly recommend preparing a reaction master mix, which requires a larger volume of each reagent, to reduce pipetting errors.

PCR Primer Design

PCR primers define the target region to be amplified and generally range in length from 15–30 bases. Ideally primers will have a GC-content of 40–60%. Avoid three G or C residues in a row near the 3′-end of the primer to minimize nonspecific primer annealing. Also, avoid primers with intra- or intermolecular complementary sequences to minimize the production of primer-dimer. Intramolecular regions of secondary structure can interfere with primer annealing to the template and should be avoided.

Ideally, the melting temperature (T_m), the temperature at which 50% of the primer molecules are annealed to the complementary sequence, of the two primers will be within 5°C, so the primers anneal efficiently at the same temperature. Primers can be designed to include sequences that can be useful for downstream applications. For example, restriction enzyme sites can be placed at the 5′-ends of the PCR primers to facilitate subsequent cloning of the PCR product, or a T7 RNA polymerase promoter can be added to allow in vitro transcription without the need to subclone the PCR product into a vector.

Template Quality

Successful amplification depends on DNA template quantity and quality. Reagents commonly used in the purification of nucleic acids (salts, guanidine, proteases, organic solvents and SDS) are potent inactivators of DNA polymerases. For example, 0.01% SDS will inhibit Taq DNA polymerase by 90%, while 0.1% SDS will inhibit Taq DNA polymerase by 99.9% (Konat et al. 1994). A few other examples of PCR inhibitors are phenol (Katcher and Schwartz, 1994), heparin (Beutler et al. 1990; Holodniy et al. 1991), xylene cyanol, bromophenol blue (Hoppe et al. 1992), plant polysaccharides (Demeke and Adams, 1992), and the polyamines spermine and spermidine (Ahokas and Erkkila, 1993). In some cases, the inhibitor is not introduced into the reaction with the nucleic acid template. A good example of this is an inhibitory substance that can be released from polystyrene or polypropylene upon exposure to ultraviolet light (Pao et al. 1993; Linquist et al. 1998).

If an amplification reaction fails and you suspect the DNA template is contaminated with an inhibitor, add a control DNA and primer pair that has amplified well in the past to the amplification reaction with the suspect DNA preparation. Failure to amplify the control DNA usually indicates the presence of an inhibitor. Additional steps to clean up the DNA preparation, such as phenol:chloroform extraction or ethanol precipitation, may be necessary.

Template Quantity

The amount of template required for successful amplification depends upon the complexity of the DNA sample. For example, of a 4kb plasmid containing a 1kb target sequence, 25% of the input DNA is the target of interest. Conversely, a 1kb target sequence in the human genome (3.3 × 10⁹bp) represents approximately 0.00003% of the input DNA. Thus, approximately 1,000,000-fold more human genomic DNA is required to maintain the same number of target copies per reaction. Common mistakes include using too much plasmid DNA, too much PCR product or too little genomic DNA as the template. Reactions with too little DNA template will have low yields, while reactions with too much DNA template can be plagued by nonspecific amplification. If possible, start with >10⁴ copies of the target sequence to obtain a signal in 25–30 cycles, but try to keep the final DNA concentration of the reaction ≤10ng/μl. When reamplifying a PCR product, the concentration of the specific PCR product is often not known. We recommend diluting the previous amplification reaction 1:10 to 1:10,000 before reamplification.

1μg of 1kb RNA = 1.77 × 10¹² molecules

1μg of 1kb dsDNA = 9.12 × 10¹¹ molecules

1μg of pGEM^® Vector DNA = 2.85 × 10¹¹ molecules

1μg of lambda DNA = 1.9 × 10¹⁰ molecules

1μg of E. coli genomic DNA = 2 × 10⁸ molecules

1μg of human genomic DNA = 3.04 × 10⁵ molecules

Cycling Parameters

The two most commonly altered cycling parameters are annealing temperature and extension time. The lengths and temperatures for the other steps of a PCR cycle do not usually vary significantly. However in some cases, the denaturation cycle can be shortened or a lower denaturation temperature used to reduce the number of depurination events, which can lead to mutations in the PCR products.

Primer sequence is a major factor that determines the optimal annealing temperature, which is often within 5°C of the melting temperature (T_m) of the primers. Using an annealing temperature slightly higher than the primer T_m will increase annealing stringency and can minimize nonspecific primer annealing, decreasing the amount of undesired products synthesized. However, using an annealing temperature lower than the primer T_m can result in higher yields, as the primers anneal more efficiently at the lower temperature. We recommend testing several annealing temperatures, starting approximately 5°C below the T_m, to determine the best annealing conditions. In many cases, nonspecific amplification and primer-dimer formation can be reduced through the optimization of annealing temperature, but if undesirable PCR products remain a problem, consider incorporating one of the many hot-start PCR methods.

Oligonucleotide synthesis facilities will often provide an estimate of the primer's T_m. The T_m can also be calculated using the Biomath Calculators. Numerous formulas exist to determine the theoretical T_m of nucleic acids (Baldino, Jr. et al. 1989; Rychlik et al. 1990). The formula below can be used to estimate the melting temperature for oligonucleotides:

T_m = 81.5 + 16.6 × (log₁₀[Na⁺]) + 0.41 × (%G+C) – 675/n

where [Na⁺] is the molar salt concentration, [K⁺] = [Na⁺] and n = number of bases in the oligonucleotide

Example:

To calculate the melting temperature of a 22mer oligonucleotide with 60% G+C in 50mM KCl:

T_m = 81.5 + 16.6 × (log₁₀[0.05]) + 0.41 × (60) – 675/22

= 81.5 + 16.6 × (–1.30) + 24.60 – 30.68 = 54°C

The length of the extension cycle, which may also need to be optimized, depends on the size of the PCR product and the DNA polymerase being used. In general, allow approximately 1 minute for every 1kb of amplicon (minimum extension time = 1 minute) for nonproofreading DNA polymerases and 2 minutes for every 1kb of amplicon for proofreading DNA polymerases. Avoid excessively long extension times, as they can increase the likelihood of generating artifacts associated with the intrinsic 5′→3′ exonuclease activity of Taq DNA polymerase (Longley et al. 1990; Bell and DeMarini, 1991).

PCR typically involves 25–35 cycles of amplification. The risk of undesirable PCR products appearing in the reaction increases as the number of cycles increases, so we recommend performing only enough cycles to synthesize the desired amount of product. If nonspecific amplification products accumulate before sufficient amounts of PCR product can be synthesized, consider diluting the products of the first reaction and performing a second amplification with the same primers or primers that anneal to sequences within the desired PCR product (nested primers).

PCR Enhancers and Additives

The addition of PCR-enhancing agents can increase yield of the desired PCR product or decrease the production of undesired products. There are many PCR enhancers, which can act through a number of different mechanisms. These reagents will not enhance all PCR reactions; the beneficial effects are often template- and primer-specific and will need to be determined empirically. Some of the more common enhancing agents are discussed below.

The addition of betaine, DMSO and formamide can be helpful when amplifying GC-rich templates and templates that form strong secondary structures, which can cause DNA polymerases to stall. GC-rich templates can be problematic due to inefficient separation of the two strands of DNA or the tendency for the complementary, GC-rich primers to form intermolecular secondary structures, which will compete with primer annealing to the template. Betaine reduces the amount of energy required to separate the strands of DNA templates (Rees et al. 1993). DMSO and formamide are thought to aid in amplification in a similar manner by interfering with the formation of hydrogen bonds between the two strands of DNA (Geiduschek and Herskovits, 1961).

Some reactions that amplify poorly in the absence of enhancers will give a higher yield of PCR product when betaine (1M), DMSO (1–10%) or formamide (1–10%) are added. Concentrations of DMSO greater than 10% and formamide greater than 5% can inhibit Taq DNA polymerase and presumably other DNA polymerases as well (Varadaraj and Skinner, 1994). Specific examples of the effects of DMSO and betaine on GoTaq^® DNA Polymerase have been published in Neural Notes (Knoche, K., 2002).

In some cases, general stabilizing agents such as BSA (0.1mg/ml), gelatin (0.1–1.0%) and nonionic detergents (0–0.5%) can overcome failures to amplify a region of DNA. These additives can increase DNA polymerase stability and reduce the loss of reagents through adsorption to the tube walls. BSA has also been shown to overcome the inhibitory effects of melanin on RT-PCR (Giambernardi et al. 1998). Nonionic detergents, such as Tween^®-20, NP-40, and Triton^® X-100, have the added benefit of overcoming the inhibitory effects of trace amounts of strong ionic detergents, such as 0.01% SDS (Gelfand and White, 1990). Ammonium ions can make an amplification reaction more tolerant of nonoptimal conditions. For this reason, some PCR reagents include 10–20mM (NH₄)₂SO₄. Other PCR enhancers include glycerol (5–20%), polyethylene glycol (5–15%) and tetramethyl ammonium chloride (60mM).

Nucleic Acid Cross-Contamination

It is important to minimize cross-contamination between samples and prevent carryover of RNA and DNA from one experiment to the next. Use separate work areas and pipettors for pre- and postamplification steps. Use positive displacement pipets or aerosol-resistant tips to reduce cross-contamination during pipetting. Wear gloves and change them often.

There are a number of techniques that can be used to prevent the amplification of DNA contaminants. PCR reagents can be treated with isopsoralen, a photo-activated, cross-linking reagent that intercalates into double-stranded DNA molecules and forms covalent, interstrand crosslinks, to prevent DNA denaturation and replication. These interstrand crosslinks effectively render contaminating DNA unamplifiable.

Treatment of the PCR reagents with uracil-N-glycosylase (UNG), a DNA repair enzyme that hydrolyzes the base-ribose bond at uracil residues, eliminates one of the most common sources of DNA contamination: previously amplified PCR products. UNG treatment prevents replication of uracil-containing DNA by causing the DNA polymerase to stall at the resulting abasic sites. For UNG to be an effective safeguard against contamination, the products of previous amplifications must have been synthesized in the presence of dUTP. This is easily accomplished by substituting dUTP for some or all of the dTTP in the reaction. Nonproofreading polymerases will readily incorporate dUTP into a PCR product, but proofreading polymerases incorporate dUTP much less efficiently (Slupphaug et al. 1993; Greagg et al. 1999; Lasken et al. 1996). Since the incorporation of dUTP has no noticeable effect on the intensity of ethidium bromide staining or on the electrophoretic mobility of the PCR product, the reactions can be analyzed by standard agarose gel electrophoresis. While both methods are effective (Rys and Persing, 1993), UNG treatment has the advantage that both single-stranded and double-stranded DNA templates will be rendered unamplifiable (Longo et al. 1990).

return to top of page

General Considerations for RT-PCR

Please also read General Considerations for PCR Optimization. Many of the important parameters discussed there also apply to RT-PCR.

Overview of the Access and AccessQuick™ RT-PCR Systems

The Access RT-PCR System and AccessQuick™ RT-PCR System are designed for the reverse transcription and amplification of a specific target RNA from either total RNA or mRNA (Miller and Storts, 1995; Knoche and Denhart, 2001). These one-tube, two-enzyme systems provide sensitive, quick and reproducible analysis of even rare RNAs (Miller and Storts, 1996). The systems use AMV Reverse Transcriptase for first-strand cDNA synthesis and the thermostable Tfl DNA Polymerase from Thermus flavus (Kaledin et al. 1981) for second-strand cDNA synthesis and DNA amplification. The systems include an optimized single-buffer system that permits extremely sensitive detection of RNA transcripts without the need for buffer additions between the reverse transcription and PCR amplification steps. This simplifies the procedure and reduces the potential for contaminating the samples. The improved performance of AMV reverse transcriptase at elevated temperatures (45°C) minimizes problems encountered with secondary structures in RNA (Brooks et al. 1995).

Template Considerations

For RT-PCR, successful reverse transcription depends on the integrity and purity of the RNA used as a template. Procedures for creating and maintaining an RNase-free environment are described in Blumberg, 1987. The use of an RNase inhibitor (e.g., Recombinant RNasin^® Ribonuclease Inhibitor) is strongly recommended. For optimal results, the RNA template, whether a total RNA preparation, an mRNA population or a synthesized RNA transcript, should be DNA-free. Tfl DNA Polymerase (supplied with the Access and AccessQuick™ RT-PCR Systems) has no reverse transcriptase activity under the standard reaction conditions (Miller and Storts, 1995), but amplification products will be generated if the template contains trace amounts of DNA with similar sequences. Excellent amplification results can be obtained with the Access and AccessQuick™ RT-PCR Systems using total RNA template levels in the range of 10pg–1μg per reaction (Figure 1.7) or poly(A)+ RNA template levels in the range of 1pg–100ng.

Reverse Transcription Primer Design

Selection of an appropriate primer for reverse transcription depends on target mRNA size and the presence of secondary structure. For example, a primer that anneals specifically to the 3′-end of the transcript (a sequence-specific primer or oligo(dT) primer) may be problematic when reverse transcribing the 5′-ends of long mRNAs or molecules that have significant secondary structure, which can cause the reverse transcriptase to stall during cDNA synthesis. Random hexamers prime reverse transcription at multiple points along the transcript. For this reason, they are useful for either long mRNAs or transcripts with significant secondary structure.

Whenever possible, we recommend using a primer that anneals only to defined sequences in particular RNAs (sequence-specific primers) rather than to the entire RNA population in the sample (e.g., random hexamers or oligo(dT) primer). To differentiate between amplification of cDNA and amplification of contaminating genomic DNA, design primers to anneal to sequences in exons on opposite sides of an intron, so any amplification product derived from genomic DNA will be much larger than the product amplified from the target cDNA. This size difference not only makes it possible to differentiate the two products by gel electrophoresis but also favors the synthesis of the smaller cDNA-derived product (PCR favors the amplification of smaller fragments).

Regardless of primer choice, the final concentration of the primer in the reaction is usually within the range of 0.1–1.0μM, but this may need to be optimized. We recommend using a final concentration of 1μM primer (50pmol in a 50μl reaction) as a starting point for optimization. More information on PCR primer design is provided in the PCR Primer Design section.

Cycle Parameters

Efficient first-strand cDNA synthesis can be accomplished in a 20- to 60-minute incubation at 37–45°C using AMV reverse transcriptase. We recommend using a sequence-specific primer and performing the reverse transcription reaction at 45°C for 45 minutes as a starting point. The higher temperature will minimize the effects of RNA secondary structure and encourage full-length cDNA synthesis. First-strand cDNA synthesis with random hexamers and oligo(dT) primer should be conducted at room temperature (20–25°C) and 37°C, respectively.

The Access and AccessQuick™ RT-PCR Systems do not require an RNA denaturation step prior to initiation of the reverse transcription reaction. If desired, however, a denaturation step may be incorporated by incubating a separate tube containing the primers and RNA template at 94°C for 2 minutes. Do not incubate AMV reverse transcriptase at 94°C; it will be inactivated. The template/primer mixture can then be cooled to 45°C and added to the RT-PCR reaction mix for the standard reverse transcription incubation at 45°C. Following the reverse transcription, we recommend a 2-minute incubation at 94°C to denature the RNA/cDNA hybrid, inactivate AMV reverse transcriptase and dissociate AMV RT from the cDNA. It has been reported that AMV reverse transcriptase must be inactivated to obtain high yields of amplification product using thermophilic DNA polymerases such as Tfl DNA polymerase (Sellner et al. 1992; Chumakov, 1994).

Most RNA samples can be detected using 30–40 cycles of amplification. If the target RNA is rare or if only a small amount of starting material is available, it may be necessary to increase the number of cycles to 45 or 50 or dilute the products of the first reaction and reamplify.

return to top of page

Thermostable DNA Polymerases

Prior to the use of thermostable DNA polymerases in PCR, researchers had to laboriously replenish the reaction with fresh enzyme (such as Klenow or T4 DNA polymerase) after each denaturation cycle. Thermostable DNA polymerases revolutionized and popularized PCR because of their ability to withstand the high denaturation temperatures. The use of thermostable DNA polymerases also allowed higher annealing temperatures, which improved the stringency of primer annealing.

Thermostable DNA polymerases can also be used for either one-enzyme or two-enzyme RT-PCR (Myers and Gelfand, 1991; Chiocchia and Smith, 1997). For example, Tth DNA polymerase can act as a reverse transcriptase in the presence of Mn²⁺ for one-enzyme RT-PCR (Myers and Gelfand, 1991). All of the DNA polymerases mentioned below can be used for amplification of the first-strand cDNA produced by a reverse transcriptase, such as AMV RT, in two-enzyme RT-PCR.

The thermostable DNA polymerases can be divided into two groups: those with a 3′→5′ exonuclease (proofreading) activity, such as Pfu DNA polymerase, and those without the proofreading function, such as Taq DNA polymerase. These two groups have some important differences. Proofreading DNA polymerases are more accurate than nonproofreading polymerases due to the 3′→5′ exonuclease activity, which can remove a misincorporated nucleotide from a growing chain of DNA. When the amplified product is to be cloned, expressed or used in mutation analysis, Pfu DNA polymerase is a much better choice due to its high fidelity. However, for routine PCR, where simple detection of an amplification product is the goal, Taq DNA polymerase is the most commonly used enzyme because yields tend to be higher with a nonproofreading DNA polymerase.

Amplification with nonproofreading DNA polymerases results in the template-independent addition of a single nucleotide to the 3′-end of the PCR product, whereas the use of proofreading DNA polymerases results in blunt-ended PCR products (Clark, 1988; Hu, 1993). The single-nucleotide overhang can simplify the cloning of PCR products.

Proofreading DNA polymerases are also used in blends with nonproofreading DNA polymerases, or amino-terminally truncated versions of Taq DNA polymerase, to amplify longer stretches of DNA with greater accuracy than the nonproofreading DNA polymerase alone (Barnes, 1994; Cline et al. 1996). See Long PCR.

Taq DNA Polymerase

Taq DNA polymerase is isolated from Thermus aquaticus and catalyzes the primer-dependent incorporation of nucleotides into duplex DNA in the 5′→3′ direction in the presence of Mg²⁺. The enzyme does not possess 3′→5′ exonuclease activity but has a 5′→3′ exonuclease activity.

Taq DNA polymerase is suitable for most PCR amplifications that do not require a high-fidelity enzyme, such as the detection of specific DNA or RNA sequences. The error rate of Taq DNA polymerase is approximately 1 × 10^–5 errors/base, although the fidelity does depend somewhat on the reaction conditions. The fidelity is slightly higher at lower pH, lower magnesium concentration and relatively low dNTP concentration (Eckert and Kunkel, 1990; Eckert and Kunkel, 1991). See High-Fidelity PCR.

Taq DNA polymerase is commonly used to amplify PCR products of 5kb or less. PCR products in the range of 5–10kb can be amplified with Taq DNA polymerase but often require more optimization than smaller PCR products. For products larger than approximately 10kb, we recommend an enzyme or enzyme mix and reaction conditions that are designed for long PCR.

Taq DNA polymerase is a processive enzyme with an extension rate of >60 nucleotides/second at 70°C (Innis et al. 1988), so an extension step of 1 minute per 1kb to be amplified should be sufficient to generate full-length PCR products. The enzyme has a half-life of 40 minutes at 95°C (Lawyer et al. 1993). Because Taq DNA polymerase is a nonproofreading polymerase, PCR products generated with Taq DNA polymerase will contain a single-nucleotide 3′ overhang, usually a 3′ A overhang.

Tfl DNA Polymerase

Tfl DNA polymerase catalyzes the primer-dependent polymerization of nucleotides into duplex DNA in the presence of Mg²⁺. In the presence of Mn²⁺, Tfl DNA polymerase catalyzes the polymerization of nucleotides into DNA, using RNA as a template. Tfl DNA polymerase exhibits a 5′→3′ exonuclease activity but lacks a 3′→5′ exonuclease activity. This enzyme is commonly used in PCR (Gaensslen et al. 1992), where its activity is similar to that of Taq DNA polymerase. Tfl DNA polymerase is used in the Access and AccessQuick™ RT-PCR Systems.

Tth DNA Polymerase

Tth DNA polymerase catalyzes the polymerization of nucleotides into duplex DNA in the 5′→3′ direction in the presence of MgCl₂. The enzyme is also capable of catalyzing the polymerization of DNA using an RNA template in the presence of MnCl₂ (Myers and Gelfand, 1991; Ruttimann et al. 1985). Tth DNA polymerase exhibits a 5′→3′ exonuclease activity but lacks detectable 3′→5′ exonuclease activity. The error rate of Tth DNA polymerase has been measured at 7.7 × 10^–5 errors/base (Arakawa et al. 1996). Tth DNA polymerase can amplify target DNA in the presence of phenol-saturated buffer (Katcher and Schwartz, 1994) and has been reported to be more resistant to inhibition by blood components than other thermostable polymerases (Ehrlich et al. 1991; Bej and Mahbubani, 1992).

Tth DNA polymerase is commonly used for PCR (Myers and Gelfand, 1991; Carballeira et al. 1990) and RT-PCR (Myers and Gelfand, 1991; Chiocchia and Smith, 1997). For primer extension, RT-PCR and cDNA synthesis using RNA templates with complex secondary structure, the high reaction temperature of Tth DNA polymerase may be an advantage over more commonly used reverse transcriptases, such as AMV and M-MLV reverse transcriptases. Recombinant Tth DNA polymerase has been shown to exhibit RNase H-like activity (Auer et al. 1995).

Tli DNA Polymerase

Tli DNA polymerase replicates DNA through the polymerization of nucleotides into duplex DNA in the 5′→3′ direction in the presence of Mg²⁺. This enzyme also contains a 3′→5′ exonuclease activity, which results in increased fidelity of nucleotide incorporation. There is no detectable reverse transcriptase activity or 5′→3′ exonuclease activity. Tli DNA polymerase will promote strand displacement at 72°C but will not displace DNA at 55°C (Kong et al. 1993). Greater than 95% of the amplified products will be blunt-ended.

Tli DNA polymerase is commonly used for PCR and RT-PCR, where its proofreading activity makes it useful for high-fidelity and long PCR (Keohavong et al. 1993). Due to the 3′→5′ exonuclease activity of Tli DNA polymerase, the enzyme can degrade the oligonucleotide primers used to initiate DNA synthesis. This exonucleolytic attack can be effectively prevented by using hot-start PCR or introducing a single phosphorothioate bond at the 3′ termini of the primer (Byrappa et al. 1995). Tli DNA polymerase can also be used for primer extension, where the high optimal temperature of the enzyme may be an advantage for templates with complex secondary structure.

Pfu DNA Polymerase

Pfu DNA polymerase has one of the lowest error rates of all known thermophilic DNA polymerases used for amplification due to the highly active 3′→5′ exonuclease activity (Cline et al. 1996; Andre et al. 1997). For cloning and expressing DNA after PCR, Pfu DNA polymerase is the enzyme of choice. Pfu DNA polymerase can be used alone for the amplification of DNA fragments up to 5kb by increasing the extension time to 2 minutes per kilobase. It is also used in blends with DNA polymerases lacking the proofreading function, such as Taq DNA polymerase, to achieve longer amplification products than with Pfu DNA polymerase alone (Barnes, 1994). However, the proofreading activity can shorten PCR primers, leading to decreased yield and increased nonspecific amplification. This exonucleolytic attack can be effectively prevented by initiating the reaction using hot-start PCR or by introducing a single phosphorothioate bond at the 3′-termini of the primers (Byrappa et al. 1995).

return to top of page

Reverse Transcriptases

The discovery of reverse transcriptases, or RNA-dependent DNA polymerases, and their role in retrovirus infection (Baltimore, 1970; Temin and Mizutani, 1970) altered molecular biology’s central dogma of DNA→RNA→protein. Reverse transcriptases use an RNA template to synthesize DNA and require a primer for synthesis, like other DNA polymerases. For in vitro applications, the primer can be either oligo(dT), which hybridizes to the poly(A)+ tails of eukaryotic mRNAs, random hexamers, which prime the synthesis from internal sites of the RNA template, or a sequence-specific primer, which hybridizes to a known sequence within the RNA template. Polymerization from a primer then proceeds as for DNA-dependent DNA polymerases. The commonly used reverse transcriptases, avian myeloblastosis virus reverse transcriptase (AMV RT), Moloney murine leukemia virus reverse transcriptase (M-MLV RT) and M-MLV reverse transcriptase, RNase H minus, perform the same reaction but at different optimum temperatures (AMV, 42°C; M-MLV, 37°C; and M-MLV RT, RNase H–, 42°C).

Some reverse transcriptases also possess intrinsic 3′- or 5′-exoribonuclease (RNase) activity, which is generally used to degrade the RNA template after the first strand of a cDNA is produced. Absence of the 5′-exoribonuclease (RNase H) activity may aid in the production of longer cDNAs (Berger et al. 1983).

Some DNA-dependent DNA polymerases also possess a reverse transcriptase activity, which can be favored under certain conditions. For example, the thermostable, DNA-dependent Tth DNA polymerase exhibits reverse transcriptase activity when Mn²⁺ is substituted for Mg²⁺ in a reaction.

AMV Reverse Transcriptase

AMV RT catalyzes the polymerization of DNA using template DNA, RNA or RNA:DNA hybrids (Houts et al. 1979). AMV reverse transcriptase is the preferred reverse transcriptase for templates with high secondary structure due to its higher reaction temperature (up to 58°C). AMV RT is used in a wide variety of applications including cDNA synthesis (Houts et al. 1979; Gubler and Hoffman, 1983), RT-PCR and rapid amplification of cDNA ends (RACE; Skinner et al. 1994). Although the high optimal temperature (42°C) makes it the enzyme of choice for cDNA synthesis using templates with complex secondary structure, its relatively high RNase H activity limits its usefulness for generation of long cDNAs (>5kb). For these templates, M-MLV RT, RNase H minus, may be a better choice.

M-MLV Reverse Transcriptase

M-MLV RT is a single-polypeptide, RNA-dependent DNA polymerase. The enzyme also has DNA-dependent DNA polymerase activity at high enzyme levels (Roth et al. 1985). M-MLV RT is used in a variety of applications, including cDNA synthesis, RT-PCR and RACE (Gerard, 1983). Its relatively low RNase H activity compared to AMV RT makes M-MLV RT the enzyme of choice for generating long cDNAs (>5kb) (Sambrook and Russell, 2001). However, for short templates with compl