1 Introduction

PCR has been used for almost every aspect of biotechnology. Today, PCR-based assays are recognized as reference standard assays for the sensitive and specific diagnosis of a number of pathogens. Moreover, in the routine diagnostic laboratory, PCR-based assays have been used for genetic analysis, including gene expression studies, mutation analysis, and pharmacogenomics. Confirmation of PCR amplification products, however, can involve laborious methods. Longer than 10 yr ago, kinetic PCR analysis by real-time monitoring of DNA amplification reactions was described (1). This technology, also called real-time PCR, has significantly simplified routine molecular diagnostics. Today, the number of commercially available assays based on real-time PCR is continuously increasing, but the majority of tests still involve in-house developed (“home-brewed”) assays. When establishing a home-brewed assay, several issues must be addressed. Issues include detection formats, introduction of an internal control (IC), quantitation, and evaluation of the assay.

2 General Considerations

PCR-based assays generally involve three major steps. In the first step, target nucleic acid is extracted, also called sample preparation. In the second step, the PCR, which includes a denaturation, an annealing, and an elongation step, is performed. Both of the primers must have a similar annealing temperature and a similar GC content. 3′-end GC clamps should be avoided by allowing a maximum of two G or C bases within the last five nucleotides at the 3′-end of the oligonucleotide. The third step includes hybridization and detection of amplification products. The introduction of probes prevents false-positive results caused by nonspecific amplification products and guarantees specificity of results. Therefore, hybridization must be included in molecular assays, to be applicable in the routine diagnostic laboratory.

Real-time PCR uses fluorescence to detect PCR amplification. The linear correlation between PCR product generation and fluorescence intensity can be monitored. For real-time PCR, instruments that combine a thermal cycler and fluorometer are used. These instruments include optics for fluorescence excitation and emission collection, as well as software that can process and analyze data. Rapid thermal cycling in real-time PCR instruments is made possible by rapid air exchange or rapid thermal conductivity through solid-phase material surrounding the reaction vessel and by a high surface-to-volume ratio of the PCR mix. The latter property is facilitated by narrow elongated reaction cuvets.

Compared with conventional PCR-based assays, real-time PCR offers several important advantages. Real-time PCR combines amplification of target DNA with detection of amplification products in the same closed vessel. Therefore, the potential for contamination is significantly reduced. With real-time PCR assays, the analytical turnaround time is significantly shorter than that required for conventional PCR assays. In contrast to conventional PCR, real-time PCR allows for log-phase analysis. Therefore, the quantitation range for real PCR assays is significantly greater (5–6 logs) than for conventional PCR assays (2–3 logs).

When establishing a home-brewed assay, it is advisable to use a primer pair that has already been published in a highly recognized journal. This helps to avoid extended specificity testing. The published sequences, however, should always be subjected to BLAST analysis to ensure that the correct sequence has been written.

3 Detection Formats

Detection formats include detection of double-stranded DNA and probe detection formats. The SYBR Green I dye intercalates into double-stranded DNA. A fluorescence signal is not only generated by amplification products, but also by primer dimers and other PCR artifacts. This technology is, thus, similar to detection of amplification products by gel electrophoresis, and is less common than probe formats in the routine diagnostic laboratory in the assays of sensitive and specific pathogen detection. It does, however, allow melting curve analysis (see Subheading 3.5. ), and may be useful for a first evaluation during early developmental steps.

For use in the routine diagnostic laboratory, introduction of probes is usually required for maximum specificity. To ensure strong binding of the probe(s) during annealing, the melting temperature (Tm) of the probe(s) should be 5 to 10°C higher than that of the primers. Consequently, probes are usually between 25 and 30 nucleotides long and their design may be sometimes difficult to optimize because of formation of secondary structure or because of sequence variability. To overcome these problems, locked nucleic acid (LNA) probes may be used. LNA is a new class of DNA analog whose incorporation into oligonucleotides results in a significant increase in the thermal stability of duplexes with complementary DNA (2). A general increase of 3 to 8°C per modified base may be expected. Consequently, shorter probes can be used for hybridization-based assays (3–7).

Probe detection formats, which have been most frequently adapted to real-time instruments, include hybridization probes, TaqMan® probes, molecular beacons, and single probes.

3.1 Hybridization Probes

The hybridization probes format uses two different fluorescence-labeled oligonucleotides. The donor probe carries a fluorescein label at its 3′-end, whereas the acceptor probe is labeled with a different fluorescein label at its 5′-end. When the fluorescein at the donor probe is excited, it emits fluorescent light at a certain wavelength. The sequences of the two probes are selected so that they hybridize to the amplified DNA fragment in a head-to-tail arrangement, thereby bringing the two fluorescent dyes into close proximity. When both of the dyes are in close proximity, the energy emitted excites the dye attached to the acceptor probe that subsequently emits fluorescent light at a longer wavelength. This energy transfer, referred to as fluorescent resonance energy transfer (FRET), is highly dependent on close proximity (between 1 and 5 nucleotides) of the oligonucleotides. The increasing amount of measured fluorescence is proportional to the increasing amount of DNA generated during the ongoing PCR process. Because the signal is only emitted when both oligonucleotides are hybridized, fluorescence is measured just after the annealing step. After annealing, the temperature is raised, and the hybridization probes are displaced by the polymerase. At the end of the elongation step, the amplification product is double-stranded and the probes are too far apart to allow FRET.

When using the hybridization probes format, it is essential that both of the probes cannot be elongated during PCR. For the donor probe, the fluorescein at the 3′-end blocks elongation, whereas, for the acceptor probe, a nonfluorescent blocking agent, typically, a phosphate group, is linked to the 3′-end. Sudden loss of probe fluorescence has been observed. Photobleaching and/or repeated freeze-thaw cycles have been correlated with the phenomenon, but an additional mechanism for FRET probe failure has been described: loss of the phosphate blocker from the 3′-end of the acceptor probe (8,9). If this occurs, then the acceptor probe can act as a PCR primer. To prevent this, the phosphate blocker may be replaced by a carbon spacer as a blocker of potentially enhanced stability (9).

Amplification products are designed not to be too long, because shorter amplification products amplify more efficiently than longer ones and are more tolerant to reaction conditions. For gene expression studies, the amplification product should be designed across an intron-exon boundary, so that complementary regions in any contaminating genomic sequence are not amplified and products represent PCR of the complementary DNA template. If using FRET hybridization probes, use of LNA may be of special importance. Introduction of LNA residues offers the advantage of modulating the required Tm of probes without any modification of the sequences, and has been shown to be very useful for single nucleotide polymorphism detection (10,12).

3.2 TaqMan Probes

The TaqMan probe format uses an oligonucleotide with a fluorescein label (reporter dye) at its 3′-end and a different fluorescein label (quencher dye) at its 5′-end. The TaqMan probe anneals to the target DNA. During elongation, the 5′ exonuclease activity of the polymerase excises the reporter dye. Because of this separation of the reporter dye from the quencher dye, the reporter dye emits fluorescent light at a certain wavelength. In contrast to all other detection formats, complete hydrolysis of the probes by the DNA polymerase is essential to yield precise results. In addition to choice of the adequate polymerase, factors that additionally influence hydrolysis include concentration of probes, the primer-probe distance, and avoidance of regions that would produce either primer:primer dimers or primer:probe dimers (13).

3.3 Molecular Beacons

The hairpin-shaped molecular beacon is nonfluorescent because of its stem hybrid that keeps the reporter dye close to the quencher dye. When the probe sequence in the loop hybridizes to its target, the quencher dye is separated from the reporter dye and the fluorescence is restored. Assays that use molecular beacons are difficult to optimize. The stem structure forms by an intramolecular hybridization event and the signal yield is very sensitive to hybridization conditions. Calculation of the assay kinetics may, thus, be rather complicated.

3.4 Single Probes

The single probe format is similar to the hybridization probes format, but it uses only a single probe. During the annealing step of PCR, the probe hybridizes specifically to the target sequence of interest. When hybridized, the probe emits a greater fluorescent signal than it does when it is not hybridized to its target. In addition to DNA probes, peptide nucleic acids may be used. With this format, changes in fluorescent signal depend solely on the hybridization status of the probe. The single-probe format may especially serve as tool for mutation detection, but is limited by unspecific background fluorescence (14,15).

3.5 Melting Curve Analysis

The temperature at which double-stranded DNA separates can vary greatly depending on the length, but mainly depends on the sequence (GC content). After completion of the amplification protocol, the temperature is steadily increased while the fluorescence is monitored. Fluorescence decreases as the temperature increases. At a certain temperature, an abrupt decrease of fluorescence can be observed because of the melting of a product. The Tm of a product is defined as the temperature at which half of the DNA is single stranded, and represents the steepest decrease of the fluorescent signal. This can be identified conveniently as the peak value in the negative derivative of the melting curve. Melting curve analysis can be performed for all detection formats, except for the TaqMan probe format, because signal generation depends on the hydrolysis of the probe.

The Tm profile can provide additional information, e.g., the genotype of the DNA product (16). Even single-base differences in heterozygous DNA can change its Tm. A heterozygous sample contains two DNA sequences, each of which melts at a different temperature, resulting in a two-peak curve. Unexpected melting peaks may indicate primer: primer or primer:probe dimers or sequence variants ( Fig. 1 ).

Fig. 1.
figure 1

Detection of herpes simplex virus (HSV) type 1 and type 2 DNA by real-time PCR. Melting curves of clinical samples (genital swabs). In sample 3, HSV-1 was detected; in samples 2 and 5, HSV-2 was found. The positive control contained both HSV-1 and HSV-2. Sample 4 shows an unexpected melting peak at 60.5°C. Sequence variation in the HSV DNA polymerase gene may produce melting peak values that differ from expected values for HSV type 1 or type 2. (Modified from ref. 16.)

Full size image

4 Internal Control

Amplification may fail in a reaction because of interference from PCR inhibitors. Consequently, an IC should be incorporated in every molecular assay to exclude false-negative results. To ensure an accurate control of the entire molecular assay, the IC should be added to the specimen before the start of the nucleic acid extraction procedure. This guarantees validation of the entire analytical testing process. No matter what IC is chosen, it must be ensured that the IC is added at a suitable concentration to prevent extreme competition with the target template for reagents (17).

4.1 Homologous IC

The homologous IC is a DNA sequence (for DNA amplification targets) or an in vitro transcript (for RNA targets) consisting of primer binding regions identical to those of the target sequence, a randomized internal sequence with a length and base composition similar to those of the target sequence, and a unique probe-binding region that differentiates the IC amplification product from the target amplification product. If using the hybridization probe format, the IC amplification product is detected by IC-specific hybridization probes labeled with a fluorescence dye different from those used for the detection of the target amplification product. It has recently been shown that IC DNA can conveniently be generated by applying composite primer technology (18,19). Because ICs consisting of bare, unprotected DNA may become degraded during storage or within the clinical specimen before nucleic acid extraction, introduction into λ-phage particles may be useful (20).

4.2 Heterologous IC

In contrast to a homologous IC, the heterologous IC presents a second amplification system within the same reaction vessel. The control must have the same or similar amplification efficiency (see Heading 5. ) as the target. Plasmids or housekeeping genes (e.g., β-microglobulin) can be used as heterologous ICs. IC-specific probes must be labeled with a fluorescence dye different from those used for the detection of the target amplification product.

5 Quantitation

In theory, PCR amplification results in an exponential increase of amplification products according to the formula N = N0 − 2n, where N is the number of amplified molecules; N0 is the initial number of molecules; and n is the number of amplification cycles. In reality, however, PCR amplification consists of three distinct phases. In the initial lag phase, no product formation can be detected. In the second phase, a more or less exponential increase of amplification products can be observed, whereas, in the third phase, the plateau effect occurs. The plateau effect may be caused by several factors, for example, product inhibition, decrease of enzyme stability and reagent concentration, and reassociation of amplification product followed by competition with primers. The result of PCR amplification is, thus, better expressed with the formula N = N0 − (Evar)n, where E is the amplification efficiency.

In contrast to conventional PCR, which only allows quantitation by end-point analysis, real-time PCR permits analysis of the complete amplification course. This, however, requires knowledge of two values. First, the crossing point (Cp), also called the threshold value, must be determined. The Cp is defined as the point at which the fluorescence curve correlated with accumulation of the amplification product intersects the background fluorescence line. This point may be between two successive cycles (i.e., it may be a fractional number). The Cp value depends on the background fluorescence, the signal noise, and the Cp calculation method (21). In optimized assays, standard errors of less than ±0.2 cycles can be achieved. If the amplification efficiency would be 2.0, the minimum relative error for quantitation would be 10 to 20%. Second, the amplification efficiency must be known. The maximum value for efficiency is 2.0, corresponding to an exponential increase of amplification products. In reality, efficiency values usually vary between 1.5 and 1.9. Because low efficiency may limit the sensitivity, assays must be optimized for maximum efficiency.

In routine real-time PCR assays, quantitation is usually performed with the use of external standards (22,23). From the data obtained by a dilution series of an external standard, the standard curve is generated. The concentration of the target, which is amplified in the same run but in a separate vessel, can be derived from the standard curve. A prerequisite for this quantitation approach is the assumption of identical amplification efficiencies in standards and in samples. This quantitation approach may, however, be impaired by inhibitors, which may be present in different concentrations in standards and in samples. To overcome this problem, an internal standard must be introduced. As described in Subheading 4.1. , an analytically distinguishable standard template is added to the sample and co-amplified in the same reaction (24,25). However, even this approach poses problems because of different fluorescence dyes used to distinguish the sequences of the standard template, and those of the sample that may be overcome by introduction of an algorithm for normalization of possible differences in amplification efficiencies (26). Together with internal standardization, quantitative results may be obtained by melting curve analysis. The area under the curve is proportional to the amount of the amplification product. Melting curve analysis, however, may be used only if the Tms of standard template and sample are significantly different (27).

6 Evaluation

Before use in the routine molecular laboratory, the newly established assay must be checked regarding its analytical and diagnostic accuracy. Analytical accuracy comprises analytical characteristics of the assay, whereas diagnostic accuracy refers to the ability of the assay to identify a condition of interest. In studies of diagnostic accuracy, the results of the assay are compared with those obtained by a reference assay (“gold standard”) in a group of patients suspected of having the condition of interest. The term “accuracy” in this context, thus, refers to the amount of agreement between the studied assay and the reference assay. To obtain such data, guidelines toward diagnostic accuracy of medical tests have recently been published (28,29). Suggested steps for evaluation of a newly established molecular assay for use in the routine diagnostic laboratory are shown in Table 1 .

Table 1 Suggested Steps for Evaluation of a Newly Established Molecular Assay
Full size table