Abstract
Background: α1-Antitrypsin is the major plasma serine protease inhibitor. Its deficiency is mainly associated with the alleles PI*S and PI*Z and can lead to obstructive lung disease in adults and to liver cirrhosis during childhood.
Methods: A multiplex PCR method has been established that uses two sets of primers to amplify the gene regions covering the PI*S or PI*Z mutations sites. Mutation detection was performed on the LightCycler by melting curve analysis of detection probes labeled with two different fluorescent dyes, LC-Red640 and LC-Red705.
Results: Unequivocal genotyping results were obtained for all investigated samples in an assay time of ∼30 min. The color compensation procedure greatly improved the readability of the resulting diagnostic melting curves.
Conclusions: To our knowledge, this is the first report of simultaneous detection of two mutations in a single tube by PCR of genomic DNA and the use of two different reporter dyes with the LightCycler color compensation feature. This approach is a rapid, convenient, and economic alternative to other methods described to date for the detection of α1-antitrypsin deficiency alleles.
The molecular bases of several inherited diseases are now well known. These include the deficiency of α1-antitrypsin (α1AT), the major plasma serine protease inhibitor. The protein deficiency is caused by improper intracellular protein aggregation and secretion or by increased intracellular protein degradation mainly attributable to two mutant alleles, designated PIZ and PIS (1). Compared with the wild-type PIM allele, the PIS allele is characterized by an A→T exchange in exon 3 that codes for a change of glutamine to valine at position 264 of the protein (2). The PIZ genotype is caused by a G→A mutation in exon 5 that codes for a glutamine-to-lysine exchange at position 342 of the protein (3). α1AT deficiency can lead to development of liver cirrhosis during childhood, or more often to development of obstructive lung disease in adults caused by an imbalance in the protease/protease inhibitor system with concurrent overactivity of neutrophil elastase. Patients with clinical manifestation of the disease usually carry the homozygous PIZZ genotype and have substantially reduced α1AT plasma concentrations. Patients with PISZ or PIMZ genotypes have less reduced α1AT concentrations and usually manifest obstructive lung disease only if they are smokers. The genetic variants of the α1AT protein have been identified by isoelectric focusing, a laborious method that often gives equivocal results and therefore requires repeated testing (4). Both the PIS and PIZ alleles are rare or absent in black or oriental populations but can be as frequent as 1.5–2.9% and 0.4–2.3%, respectively, in different white populations (4). Apart from these deficiency-associated alleles, other rare alleles, including null alleles, have been reported. Such alleles must be considered in patients with low concentrations of α1AT protein and absence of the PIS or PIZ alleles.
The throughput of genotyping methods can be substantially improved by multiplex PCR, with a concurrent reduction in the use of reagents and DNA. Detection of α1AT genotypes is ideally suited to such a procedure because both deficiency alleles, PIS and PIZ, must be taken into account. This has been accomplished previously by the use of conventional PCR methods (5)(6). The LightCycler (Roche Diagnostics) is a rapid PCR cycler with an integrated three-channel fluorescence photometer that eliminates the need of any postamplification sample processing. Oligonucleotide probes labeled with different dyes are used for mutation detection. Probes added to a PCR reaction mixture will specifically hybridize to their complementary strand. Mutations under the probe decrease the stability of this duplex and lead to decreased melting temperatures (7). This principle of mutation detection has already been used for factor V (8), methylenetetrahydrofolate reductase (9), prothrombin G20210A (10), and hemochromatosis (11) genotyping. When two probes in the assay hybridize adjacent to each other, fluorescence resonance energy transfer (FRET) occurs between the fluorescein on the detection probe and the LC-Red dye (Roche Biochemica) on the anchor probe, producing a specific fluorescence emission. FRET is stopped when the detection probe melts off the strand. A mutation under the probe causes a lower probe melting point. Primers internally labeled with the reporter dye can also serve as counterparts of the fluorescein detection probe. Although the emission spectra of most dyes overlap, the LightCycler can be calibrated in such a way that the software virtually eliminates the crossover of one dye into the detection wavelength of the other.
The aim of this study was to establish a multiplex PCR method for PIS and PIZ genotyping by use of two different reporter dyes and to show that the color compensation procedure can greatly improve the readability of the melting curve analysis.
Materials and Methods
primer and probe design
For the AAT-S PCR (Fig. 1⇓ A), four oligonucleotides were used in the assay, of which two were the labeled anchor and detection probes and two served as amplification primers. In the case of the AAT-Z PCR (Fig. 1B⇓ ), three oligonucleotides were used with an internally labeled primer and a labeled detection probe.
Schemes of the primer and probe designs for PI*S and PI*Z genotyping.
Both detection probes are complementary to the mutant genotype. The amplified products of the α1AT gene are shown schematically. (A), top, sequences of probes for the *S allele with the fluorophores attached; bottom, wild-type antisense strand. The site of the PI*S mutation is shown bold italic. When both probes are hybridized, FRET occurs over the 1-bp gap. (B), oligonucleotides designed for detecting the PI*Z genotype. A forward primer was chosen that had the LC-Red dye attached to a thymidine amino-modifier (T) in its sequence. The site of the PI*Z mutation is shown in bold italic. The corresponding detection probe is hybridized to this strand. FRET occurs over a 5-bp gap. FLU, fluorescein; PHO, phosphorylated end.
For the genotyping of the α1AT S and Z alleles (GenBank accession nos. K02212 and J02619), new primer sets were constructed:
For PIS genotyping, the primers and probes were as follows: the forward primer (AAT-S-for) was 5′-AGC GTT TAG GCA TGT TTA AC-3′, and the reverse primer (AAT-S-rev) was 5′-AAG TTT ATA CAG AGT AGC AGT G-3′. The mutation site was covered by a PIS genotype complementary detection probe: AAT-S-probe (5′-CAC CTG GTA AAT GAA CTC-3′-FLU) labeled with fluorescein as indicated (Fig. 1A⇑ ). The 3′ end of the AAT-S-anchor probe was phosphorylated to prevent probe elongation by the Taq polymerase. The adjacent anchor probe AAT-S-anchor (5′-LC-Red705-CCC ACG ATA TCA TCA CCA AGT TCC-3′-PHO) was 5′ labeled with the LC-Red705 dye, which exhibits an emission maximum at 705 nm.
For PIZ genotyping, the primers and probes were as follows: the forward primer (AAT-Z-for) was 5′-GTG CAT AAG GCT GTG CTG AC-3′, and the reverse primer (AAT-Z-rev) was 5′-GGT GGG ATT CAC CAC TTT TC-3′. A thymidine amino-modifier (dT aminomodifier; Glen Research) was incorporated at the underlined position in the AAT-Z-for primer. The dye LC-Red640 was linked to the amino-modifier by reaction with the LC-Red640-N-hydroxysuccinimide ester and purified according to the manufacturer’s instructions. The emission maximum wavelength of this dye is at 640 nm. The mutation site was covered by a PIZ genotype complementary detection probe, AAT-Z-probe (5′-GCT TCA GTC CCT TTC TTG TCG A-3′-FLU), labeled with fluorescein as indicated (Fig. 1B⇑ ).
color compensation procedure
The color compensation procedure was performed according to the manufacturer’s instructions (LightCycler color compensation set; Roche Biochemica). After activation of the software’s color compensation option, it was possible to read the results from two different analytical PCRs in one tube. In the PIZ PCR, the LC-Red640 dye was used and the result was detected in channel two, whereas in the PIS PCR, the fluorescence of the LC-Red705 dye was visible in channel three.
genomic dna extraction
Genomic DNA was extracted using a commercial resin-based method (QIAamp DNA blood kit; Qiagen) or by a simple method modified according to Rudbeck and Dissing (12), which takes <10 min. In brief, 5 μL of anticoagulated blood was added to 1 mL of PCR-grade water in a 1.5-mL microcentrifuge tube. The tube was immediately centrifuged (12 000g for 1 min), and the supernatant was removed. The remaining leukocyte pellet was lysed by the addition of 20 μL of 0.2 mol/L NaOH, vortex-mixed thoroughly, and incubated for 5 min. Neutral pH was restored by the addition of 180 μL of 0.04 mol/L Tris puffer, pH 7.5, containing 50 g/L Chelex-100 resin, biotechnology grade (Bio-Rad). The DNA solution can be stored at 4 °C for at least 6 months. Neither the choice of anticoagulant (EDTA, heparin, or citrate) nor the freezing of samples before DNA isolation affected the method or inhibited PCR amplification.
multiplex pcr protocol
PCR reactions were carried out in a final volume of 20 μL in LightCycler glass capillaries. The reaction mixture consisted of 1 μL of DNA solution, 1 U of Taq DNA polymerase (Boehringer Mannheim), 2 μL of 10× PCR buffer (Boehringer Mannheim), 0.2 mmol/L each dNTP (Boehringer Mannheim), 2.5 mmol/L MgCl2, 500 mg/L bovine serum albumin (New England BioLabs), and 50 mL/L dimethyl sulfoxide (Sigma). Amplification primers were added at different concentrations to account for the different PCR efficiencies and to ensure adequate product formation from both primer sets: for the AAT-S PCR, 0.25 μmol/L AAT-S-for, 0.25 μmol/L AAT-S-rev, 0.1 μmol/L AAT-S-probe, and 0.3 μmol/L AAT-S-anchor; for the AAT-Z PCR, 0.3 μmol/L AAT-Z-for, 0.3 μmol/L AAT-Z-rev, and 0.1 μmol/L AAT-Z-probe. PCR-grade water was added to the final volume of 20 μL.
Each set of PCRs included a heterozygous DNA control, a contamination control from the DNA preparation, and a water control. The contamination control was prepared by mixing all of the reagents for DNA isolation but excluding DNA, and would indicate contamination of reagents used for the DNA isolation in contrast to reagents used in the PCR itself, where possible contamination would show up in the water control. The genotyping of control DNA was performed by a restriction fragment length polymorphism-PCR (6). The fluorometer gain setting was 5 in channel 1, 10 in channel 2, and 30 in channel 3. The cycling program consisted of initial denaturation for 30 s at 95 °C, followed by 55 cycles of 95 °C for 0 s, 50 °C for 5 s, and 72 °C for 5 s, with maximum ramp rate. The program for analytical melting was 95 °C for 30 s, 40 °C for 30 s, and an increase to 70 °C at a 0.1 °C/s ramp rate. Amplification and detection occurred in the same closed tube in <40 min. The PCR conditions and cycler program were essentially the same as those used previously (10)(13), thereby fully integrating the α1AT genotyping into our single master mixture, single cycler program approach for diagnostic genotyping.
Results
We have used DNA, isolated with a commercially available kit, at concentrations of 15–100 ng of DNA template in the PCR and found no influence of the concentration on the shape and height of the resulting melting curves (data not shown). We, therefore, do not routinely quantify our DNA before use. Samples prepared with the rapid lysis protocol described above do not differ from DNA prepared with the commercial method with respect to amplification efficiency or melting curve shape. When the rapid DNA isolation protocol was used, an estimated 0.75 ng of DNA, assuming 5000 leukocytes/μL blood, was used in the PCR assay. This amount was sufficient for amplification with this method. It is also possible to successfully amplify DNA isolated with this method from samples with <1000 leukocytes/μL.
If on-line PCR monitoring is required, it is advisable to monitor the fluorescence in channel two. Because the melting point of the AAT-Z probe is higher than that of the AAT-S probe, this produces a better hybridization fluorescence signal at the acquisition temperature. During the analytical melting step, fluorescence is measured simultaneously in all three channels of the LightCycler. The concentrations of the primers for the multiplex PCR were optimized to give well-shaped melting curves for all investigated alleles. This ensures that both PCR reactions are amplified to a similar extend. The PCR products were also visualized by agarose gel electrophoresis after ethidium bromide staining. Whenever a good FRET signal occurred in both fluorescence channels, distinctive bands were observed by agarose gel electrophoresis.
Successful amplification was evident from the appearance of specific fluorescence and the display of derived melting curves. In the case of a homozygous mutation, there was no mismatch under the mutation-compatible probe and a single melting peak with a characteristic high temperature was seen. A single-base-pair-mismatched probe, i.e., wild-type DNA, caused strand instability and consecutive melting at a lower temperature. The result was a single melting peak at a characteristic lower temperature. Patients with heterozygous mutations accordingly showed two melting peaks. The empirical melting points determined by the LightCycler software (Software Package V3.1 Data Analysis Module) were ∼62.5 °C for PIZ alleles, 55.0 °C for PIS alleles, 59.5 °C for the wild type at the site of the PIZ allelic mutation, and 49.5 °C for the wild type at the site of the PIS allelic mutation. Typical melting curves resulting from genotyping with this method and active color compensation are illustrated in Fig. 2⇓ . It is noteworthy that the color compensation works independent of the type of assay (three vs four oligonucleotides) and coupling of the dyes. When the same samples were analyzed with a color compensation file that had been calibrated 8 weeks earlier, the emission crossover of LC-Red640 into channel 3 (705 nm) was not sufficiently suppressed (Fig. 3⇓ , A and B). If analysis was performed without color compensation (Fig. 3⇓ , C and D) the substantial crossover of LC-Red640 into channel 3 was even more evident. Reading of the melting curves may be misleading because of overlaps in both the melting points and the emission spectra.
Different genotyping results from the same run in a multiplex PCR with optimal color compensation.
The negative derivatives of fluorescence vs temperature [−d(F2)/dT and −d(F3)/dT] are shown. (A), LC-Red640 fluorescence is read in channel 2, and results for the PI*Z genotyping are shown. Melting points are 62.5 °C for the PI*Z allele and 59.5 °C for wild type. (B), LC-Red705 fluorescence is read in channel 3. Results from genotyping of the PI*S mutation are shown. Melting points are 55.0 °C for the PI*S allele and 49.5 °C for wild type. Contamination and water controls were negative in all examples (data not shown). Genotypes are coded as follows: (——–), PI*ZZ; (· · · · · · · ·), PI*MZ; (− · − · − · − · −), PI*MS; (− − − −), PI*MM.
Effect of insufficient or no color compensation.
Data are from the same run shown in Fig. 2⇑ . (A and B), analyses performed using a color compensation calibrated 2 months earlier; (C and D), software color compensation feature turned off. (A), PI*Z genotyping in channel 2 (LC-Red640 fluorescence) is not compromised. (B), insufficient suppression of the LC-Red640 dye used in channel 2. The melting points of the PI*Z alleles at 62.5 and 59.5 °C are superimposed on the results from PI*SS genotyping with LC-Red705 in channel 3. (C), results for PI*Z genotyping in channel 2 are unmistakable even if color compensation is turned off. (D), the results for PI*S genotyping cannot be read because of crossover of LC-Red640 dye fluorescence into channel 3. Genotypes are coded as follows: (——–), PI*ZZ; (· · · · · · · ·), PI*MZ; (− · − · − · − · −), PI*MS; (− − − −), PI*MM. −d(F2)/dT and −d(F3)/dT, negative derivatives of fluorescence vs temperature.
Discussion
The design of a hybridization probe assay demands careful primer and probe selection. The FRET system consists of two probes, one situated over the mutation site and another, adjacent anchor probe. The melting temperature of the anchor probe should be at least 5 °C higher than that of the detection probe. Fluorescence emission decreases with an increasing gap size between the dyes attached to the probes (data not shown) and is known to be inversely proportional to the sixth power of the distance (14). Nevertheless, a 5-bp gap still gives good results. The gap size, therefore, can be increased if the sequence demands, e.g., if several guanosine bases are present near the dye, which would quench its fluorescence (15). When an internally labeled primer is used, the dye is incorporated directly into the strand formed by Taq elongation. Therefore, only the melting point of the detection probe is of importance in this situation: it must hybridize with the strand that is generated by extension of the dye-labeled oligonucleotide to produce a FRET signal. A 5-bp gap has been reported to give an optimal fluorescence reading with this kind of probe placement (8) and has been used successfully in this assay.
Multiplex PCR with hybridization probes has been used before on the LightCycler for hemochromatosis genotyping (11), but it was performed with two probes labeled with the same dye. This is possible only if the different melting points of the two probes are always clearly discernible, no matter what genotype is present under each probe. A second reporter dye, LC-Red705, is now available for the LightCycler, and its fluorescence is read in a separate channel. It is a general feature of fluorescent dyes that they exhibit broad emission spectra. This leads to crossover of LC-Red640 emission into the detection wavelength of LC-Red705 (Fig. 3⇑ , B and D). The application of color compensation on the LightCycler to correct this situation was demonstrated recently using a model system with artificial templates (16). We showed here that the compensation is also possible in a system where a primer labeled via an amino-linker is used. The most important measure to reduce dye interference is proper calibration of the color compensation feature of the LightCycler software. From our experience, it is advisable to repeat the color compensation more often than every 6 months, as is recommended by the manufacturer (17). This may be dependent on the instrument run time. However, it cannot be excluded that calibration of the color compensation per se may give results that lead to insufficient compensation. Therefore, we recommend that the color compensation be recalibrated whenever a crossover problem is suspected. The color compensation is a procedure independent of the PCR run itself. If suboptimal color compensation is suspected, a new calibration run can be performed in <30 min, and the calibration can be applied to the already existent data set.
Our cycling program is the same that we used before (10)(13). This allows the PCR to be carried out in the same run together with factor V, prothrombin G20210A, methylenetetrahydrofolate reductase, and apolipoprotein B-3500 genotyping. This adds to the value of the LightCycler instrument for the detection of common disease-causing mutations. If this PCR is performed with reaction mixtures without dimethyl sulfoxide, the resulting melting points will be ∼3 °C higher. Dimethyl sulfoxide, which is present in our reaction mixture, is known to lower the DNA melting point by 0.6 °C for each 1% of dimethyl sulfoxide in the reaction mixture (18).
All mutation detection methods not based on sequencing may give misleading results if there are other base exchanges in the vicinity of the mutation of interest (13). With this hybridization-based method, any other imperfect match under the mutation-compatible probe will produce a mismatch with a typical melting curve. Such mutations are extremely rare, and it is unlikely that a melting curve of such a base exchange would be indistinguishable from that of the wild type (13). All cases found to have unusual melting curves with this approach should, therefore, be further clarified either by sequencing or by use of a wild-type-compatible probe.
To our knowledge, we report here for the first time the simultaneous detection of two mutations in a single tube by PCR of genomic DNA using two different reporter dyes with the LightCycler color compensation feature. This approach is a rapid, convenient, and economic alternative to other methods described to date for the detection of α1AT deficiency alleles.
Acknowledgments
We thank Prof. Victor W. Armstrong for helpful comments on this manuscript, and Sandra Hartung for skillful technical assistance. The probe labeled with LC-Red705 dye was a gift from Roche Diagnostics GmbH, Mannheim, Germany.
Footnotes
This work was presented in abstract form at the 1999 meeting of the “Deutsche Gesellschaft für Klinische Chemie” and “Deutsche Gesellschaft für Laboratoriumsmedizin” October 3–5, 1999, Regensburg, Germany.
- © 2000 The American Association for Clinical Chemistry
References
In this issue
