Abstract
Background: An accurate determination of the major HFE mutation (C282Y), which is associated with hereditary hemochromatosis, is important in diagnosis and risk assessment for this disease. We report a single-tube high-throughput PCR method for the detection of C282Y.
Methods: We combined three previously described principles: allele-specific PCR, mutagenically separated PCR, and amplicon identification by specific dissociation curves. PCR amplification was performed with fluorescence detection or conventional thermocycler using the same primers, reactant constituents, and cycling protocol. Primer cross-reactions were prevented by deliberate primer:primer and primer:template mismatches.
Results: PCR products were identified by their characteristic melting temperatures based on SYBR Green I fluorescence. For each of the 256 random and 17 known HFE C282Y samples, mutant homozygous, wild-type, and heterozygous samples were unequivocally distinguished.
Conclusions: This homogeneous assay is rapid, reproducible, does not require fluorescent oligonucleotide probes, and correctly identifies HFE genotypes.
Hereditary hemochromatosis (HH)1 is the most common genetic illness in the northern hemisphere, with a prevalence of ∼5 in 1000 persons (1). Between 64% and 100% of HH patients (depending on the population studied) are homozygous for a one-base difference at cDNA position 845 of the hemochromatosis gene (HFE) (2)(3)(4)(5)(6)(7). This G-to-A transition produces a cysteine-to-tyrosine change at amino acid 282 (C282Y) of the HFE protein (2). A considerable proportion of HH patients are wild-type/wild-type at the C282Y HFE locus, indicating that additional genetic defects may be responsible for this disease (8). A second HFE mutation occurs at nucleotide position 187 (C-to-G) and changes amino acid 63 from histidine to aspartic acid (H63D). H63D may contribute to increased hepatic iron concentrations but does not produce iron overload in the absence of C282Y. Recently, a new HFE polymorphism, S65C, which leads to a serine-to-cysteine substitution in exon 2 was identified (1). This variant is enriched in HH patients, but whether it contributes to iron overload in HH patients remains controversial (9).
The ability to carry out rapid DNA analysis to determine the mutational status or genotype of an individual has become an increasingly important task for the clinical diagnostic laboratory. Consequently, there is a need for high-throughput and automatable assays to identify clinically significant, single nucleotide polymorphisms (SNPs). Currently, a wide variety of methods exist for detecting single base changes in a DNA molecule. Classical techniques include restriction isotyping (10), single-strand conformation polymorphism analysis (11), oligonucleotide ligation assay (2), heteroduplex analysis (12), and allele-specific (AS) oligonucleotide hybridization probes (3). A simple and inexpensive way to determine the genetic status of an individual is by the use of AS-PCR. In this method, an oligonucleotide primer is specially designed to match one allele but mismatch the other allele at or near the 3′ end. If the DNA polymerase cannot extend a primer with a 3′ mismatch, one allele is preferentially amplified over the other. The specificity of the AS primers can be further enhanced by engineering a deliberate base change very close to their 3′ ends.
To identify a bi-allelic polymorphism, two physically separate PCR reactions are required for each analysis. In addition, a pair of control primers that amplifies an independent fragment usually is included in the reaction to ensure that the PCR reaction itself was successful (13). This method is known by a variety of names, allele-specific amplification, amplification refractory mutation system, and PCR amplification of specific alleles (14). Several methods based on this principle have been developed to detect the C282Y mutation in the HFE gene (15)(16)(17)(18)(19)(20)(21).
An enhanced approach known as PCR amplification of multiple specific alleles (22), or mutagenically separated PCR (23), allows both AS oligonucleotides to be coamplified and differentiated in a single PCR reaction. Cross-reactions between the different AS primers are avoided by the use of deliberate mismatches at or near the 3′ and 5′ ends of the primers. Compared with AS amplification, mutagenically separated PCR eliminates the need for an internal control primer set and reduces the cost and labor of the techniques by approximately one-half. Merryweather-Clarke et al. (24) have applied this method to the detection of the HFE C282Y genotype. However, this technique and the aforementioned methods are not ideally suited to large-scale analysis because they require a laborious post-PCR processing step.
The problems of low throughput and the requirement for postamplification manipulations have been overcome by the development of new PCR instruments that can monitor the PCR reaction in real time. These devices are composed of a thermal cycler coupled to a fluorescent detector and are capable of PCR amplification with simultaneous amplicon analysis (25)(26). A simple approach for concurrent DNA amplification and detection is the use of AS primers and a fluorescent double-stranded DNA-specific binding dye (SYBR Green I). Products are detected by their characteristic dissociation profiles. A product dissociation profile is generated after the PCR reaction by monitoring the fluorescence of the SYBR Green I dye as the temperature passes through the amplicon’s denaturation temperature. Dissociation profiles are dependent on the GC content and length and sequence of the PCR products (25).
Germer and Higuchi (27) recently applied this strategy to identify SNPs of the paraoxonase and apolipoprotein B genes. In a method the authors call Tm-shift genotyping, a 5′ GC-tail of 26 base pairs is attached to one of the AS primers. This functions by increasing the melting profile of that primer and thus allows it to be differentiated from the second AS primer.
We now report a real-time, single-tube, homogeneous SNP detection method that combines the principles of Newton et al. (13), Rust et al. (23), and Germer and Higuchi (27), and we show its applicability to the rapid detection of the three different HFE C282Y genotypes.
Materials and Methods
patients
Blood samples were collected from a group of 256 Finnish individuals. This group was composed of a random sample of local hospital patients for whom the blood count was routinely examined. Samples were anonymized by retaining only the birth year and sex. In addition, samples were obtained from 10 patients with clinically confirmed or suspected hemochromatosis. Informed consent was obtained from these patients for the genetic analysis. The study protocol was in accordance with the Helsinki Declaration of 1975, as revised in 1983.
dna extraction
Genomic DNA was extracted from 3 mL of EDTA-anticoagulated blood. The DNA isolation was carried out using a salting-out method (28). Blood cells were lysed with 7.5 mL of Tris buffer 1, pH 8.0 (13 mmol/L Tris base, 10 mmol/L KCl, 14.5 mmol/L MgCl2 · 6 H2O, 2 mmol/L EDTA, and 25 mL/L Triton X-100). After centrifugation (7 min at 2400g), the pellet was washed with Tris buffer 1 and centrifuged (7 min at 1200g). The pellet was then lysed with 660 μL of Tris buffer 2, pH 8.0 (13 mmol/L Tris base, 10 mmol/L KCl, 4.7 mmol/L MgCl2 · 6 H2O, 2 mmol/L EDTA, 0.4 mol/L NaCl, and 10 g/L sodium dodecyl sulfate), and incubated for 15 min at 56 °C. Cellular proteins were removed by precipitation with 300 μL of 5 mol/L NaCl and centrifugation (7 min at 560g). DNA was isolated by ethanol precipitation and incubated for 1 h at 4 °C in Tris-EDTA buffer, pH 8.0 (13 mmol/L Tris base, 1 mmol/L EDTA). The DNA concentration was then measured spectrophotometrically at 260 nm, and samples were diluted to a final concentration of 20 mg/L (20 ng/μL).
pcr primers
Three oligonucleotide primers were designed based on the National Center for Biotechnology Information GenBank HFE cDNA sequence (Accession No. U91328). A schematic representation of the different oligonucleotide primers used for genotyping the C282Y locus is shown in Fig. 1⇓ . The wild-type, mutant, and common primers (sequences) were HFEW2 (5′-GGG GGG CCC CGG GCC CAG ATC ACA ATG AGG GGC ACA TCC AGG CCT GGG TGC TCC ACC TCG C-3′), HFEM (5′-TGA TCC AGG CCT GGG TGC TCC ACC TGC T-3′), and HFECOM (5′-CAG GGC TGG ATA ACC TTG GCT GTA CC-3′), respectively. The primers were purchased from MedProbe.
Diagram of the AS-PCR primers used to detect the C282Y HFE gene mutation.
The position of the newly discovered 5569 G/A polymorphism in intron 4 is shown in relation to the C282Y HFE PCR primers. The binding sites of all C282Y HFE primers exclude this single nucleotide polymorphism. There are deliberate nucleotide differences between primers HFEW2 and HFEM. The first difference, which occurs at the 3′ nucleotide, is illustrated by a gray box with black lettering, whereas the other four differences between the primers are represented by black boxes with white lettering. These mismatches ensure that mispriming and cross-reactions between primers and template are prevented.
To allow product identification from the single reaction mixture, the AS primers were designed with different lengths according to Rust et al. (23). The two forward AS primers, HFEW2 and HFEM, were 61 and 28 bp long, respectively, and the complementary primer, HFECOM, was 26 bp in length. Mispriming and cross-reactions were prevented by the introduction of deliberate mismatches between primers and template (13)(23).
The first nucleotide difference (C or T) between AS primers HFEW2 and HFEM is located at the 3′ terminal base. To ensure the specificity of these primers, a DNA polymerase that lacks the 3′ exonuclease proofreading activity (DyNAzyme II) was used in the PCR reaction. The second primer base change (G to C) generates a purine/pyrimidine primer/template mismatch, and this prevents amplification of the nonmatching AS primer. This mismatch is located three bases from the 3′ end of HFEW2 and two bases from the 3′ end of HFEM. Two additional nucleotide changes (A and C) were made to the HFEW2 primer, located at the same positions as the last two 5′ nucleotides of the HFEM primer. These changes should prevent the generation of possible spurious products, which could otherwise occur by the annealing and extension of the HFEM primer to the first-round product of HFEW2. Finally, to facilitate the discrimination of the AS primers by dissociation curve analysis, a 13-bp GC tail was added to the 5′ end of the HFEW2 primer.
pcr amplification
PCR reactions were carried out using MicroAmp Optical TubesTM and MicroAmp Optical CapsTM (PE Biosystems). The PCR reaction mixture contained the following in a final volume of 25 μL: 50 ng of genomic DNA, PCR reaction buffer (10 mmol/L Tris-HCl, pH 8.8, 1.5 mmol/L MgCl2, 50 mmol/L KCl, and 1 mL/L Triton X-100; Finnzymes), 5 mmol/L dNTP, 1 U of DyNAzyme II DNA Polymerase (Finnzymes), 5 pmol of HFECOM and HFEW2, 20 pmol of HFEM, and 2.5 μL of SYBR Green I 1:10 000 (Molecular Probes). Negative control reactions containing water in place of DNA were included in each batch of PCR reactions to ensure that contamination was not a problem. The PCR amplification profile was as follows: initial denaturation at 95 °C for 4 min, followed by 32 cycles of denaturation at 96 °C for 30 s and combined annealing and extension at 71 °C for 30 s. To investigate the versatility of the method, we carried out PCR amplification in two different thermocyclers, the PTC-200 DNA engine (MJ Research) and the GeneAmp® 5700 Sequence Detection System (PE Biosystems)
product analysis
In the GeneAmp 5700, analysis of the real-time fluorescence signal from SYBR Green I bound to double-stranded DNA was performed by the GeneAmp 5700 software (PE Biosystems). A threshold cycle (Ct) was determined for each sample, using the exponential growth phase and baseline data of the fluorescent amplification plots. A sample was deemed positive if it had an increase in fluorescence above the Ct. The fluorescent amplification plots were converted to dissociation curves by plotting the negative derivative of fluorescence with respect to temperature. Dissociation curves were subsequently used to identify PCR products. For confirmatory electrophoretic analysis of PCR products, a 10-μL aliquot of each reaction was run on a 2% gel in Tris-borate-EDTA buffer.
dna sequencing
PCR products for sequencing were generated using the primer HFESEKV (5′-TTA CCT CCT CAG GCA CTC CTC TCA ACC-3′) and HFECOM primers. Twelve different samples were analyzed by automated “Dye Terminator” cycle sequencing using an ABI 377 automatic sequencer (PE Biosystems) with DNA cycle sequencing reagent set (BigDye Terminator Cycle Sequencing FS Ready Reaction Kit; PE Biosystems) according to the manufacturer’s protocol.
Results
assay optimization
The newly designed AS primers were tested with the GeneAmp 5700 Sequence Detection system. The multiplex-PCR reaction was optimized based on the minimization of the variation in Ct values for each of the respective genotypes (Fig. 2⇓ ). This was a highly effective way to quickly optimize the assay in contrast to endpoint analysis, which yielded considerable fluorescence variation, especially before the number of PCR cycles was optimized and reduced to 32 (data not shown).
Optimization of the C282Y HFE assay based on Ct values.
Amplification plots for two duplicate wild-type (- - - -), heterozygous (− - - − - - −), and mutant homozygous (——) samples and no-template controls (- - - - -) are shown to illustrate the process of PCR optimization. This was achieved by minimizing the variation between Ct values for the three different C282Y HFE genotypes. The average Ct value for the different genotypes is 27.52 with a mean CV of 0.25%. In contrast, even with good optimization based on Ct values, end point fluorescence still has considerable variation, with a mean value of 0.46 and a mean CV of 4.6%.
identification of wild-type, heterozygous, and homozygous samples
The results for the HFEW2, HFEM, and HFECOM primers with the GeneAmp 5700 Sequence Detection system are shown in Fig. 3⇓ A. For each C282Y sample, the AS primers accurately distinguished between mutant homozygote, wild type, and heterozygote. The melting of the sample homozygous for the 845 G allele showed a marked change (decrease) in fluorescence between 86 and 88 °C, with a maximum rate of change around 87 °C. In contrast, the sample homozygous for the 845 A allele showed a marked decrease in fluorescence between 82 and 84 °C, with a maximum rate of change around 83 °C. The heterozygous sample contained both fluorescent melting peaks because of the presence of amplicons derived from both alleles. Confirmatory analysis of the products by slab gel electrophoresis revealed that the wild-type samples generated the expected 113 bp with the HFEW2 primer and no product was amplified with the HFEM primer. Similarly, the HFEW primer generated no product with the mutant homozygous sample, but as expected, the HFEM primer generated a band of 80 bp. For the heterozygous sample, both the 80- and 113-bp products were visible on the gel.
Comparison of C282Y genotyping by the GeneAmp 9600 (A) and the PTC-200 DNA Engine (MJ Research; B).
PCR products from both thermocyclers were analyzed by a short dissociation protocol using the GeneAmp 5700 Sequence Detection System. Fluorescence amplification plots were converted to derivative dissociation curves by plotting the negative derivative of the fluorescence with respect to temperature (-dF/dT) against temperature. The derivative dissociation curves are shown for a HH sample (peak 845 A; top), a wild-type sample (peak 845 G; second from top), a heterozygous sample (peaks 845 G and 845 A; second from bottom), and a no-template control (bottom). The 845 G peak has a higher temperature value than the 845 A peak because of a greater GC content of the amplicon.
We next tested the possibility of using a standard thermocycler (PTC-200 DNA Engine; MJ Research) to amplify the C282Y locus. We then used a short 20-min dissociation protocol on the 5700 machine to analyze the PTC-200 products; the results are presented in Fig. 3B⇑ . These results show that the dissociation curves produced by the products of either thermocycler are nearly identical.
assay performance
To test the reproducibility of the fluorescence dissociation curves, 10 wild-type, 10 heterozygous, and 5 mutant homozygous samples were analyzed in duplicate (Fig. 4⇓ ). For each allele, the dissociation curve was highly reproducible: the sample-to-sample and within-sample variations of the dissociation curves were <0.5 °C. The robustness of the technique was evaluated by analyzing >200 DNA samples, and all samples tested gave an unambiguous C282Y HFE genotype.
Sample-to-sample and within-sample variation of the C282Y derivative melting peaks.
A total of 70 PCR reactions comprising 10 duplicate wild-type, 10 duplicate heterozygous, and 5 duplicate mutant homozygous samples and 10 duplicate no-template controls were analyzed using the GeneAmp 5700 Sequence Detection System. All 70 dissociation curves for each individual genotype are superimposed in the graphs.
assay validation
The validity of the method was verified independently by testing 17 samples comprising all three C282Y HFE genotypes. Twelve samples (4 mutant homozygous, 4 heterozygous, and 4 wild-type homozygous) were sequenced using HFESEKV. Five other samples (one mutant homozygous, two heterozygous, and two wild-type homozygous) were analyzed by an outside reference laboratory using a restriction digestion method (29). The genotype results obtained for all 17 samples were 100% concordant to those obtained by the current method.
genotype scoring
The SDS 5700 software allows the export of numeric dissociation curve data to other software. We exported the data into Microsoft Excel and designed a macro program that calculated the area under the dissociation curve. Subsequently, a scatter graph was generated in which the area under the dissociation curve between temperatures 81.8 and 83.6 °C was plotted on the x axis and the area under the dissociation curve between temperatures 85.8 and 87.3 °C was plotted on the y axis. This generated a graph in which the three C282Y genotypes and the no-template controls separated into four discrete clusters with definable limits (Fig. 5⇓ ). Using this customized Excel sheet, we were able to automate the process of genotype scoring. Of the 256 random-study subjects, there was 1 mutant homozygous and 19 heterozygous individuals. Thus, the overall allele frequency for HFE Tyr 282 was 4.1%, and the carrier frequency was 7.8%.
Scatter graph of the different replicate C282Y HFE genotypes and the no-template controls (NTC), for the same set of samples as in Fig. 4⇑ .
The scatter graph was generated by plotting the area under the dissociation curve between temperatures 81.8 and 83.6 °C (Peak 1) on the x axis. Similarly, the area under the dissociation curve between temperatures 85.8 and 87.3 °C (Peak 2) was plotted on the y axis. Using fixed cutoff limits for the area under peak 1 (vertical dashed line crossing the x axis at 2) and peak 2 (horizontal dashed line crossing y axis at 1.5), we could automatically score the three different genotypes and no-template controls.
Discussion
Prevention of manifest hemochromatosis would benefit greatly from a simple and inexpensive DNA screening test to identify carriers and affected individuals. The disease is characterized by a lifelong excessive accumulation of iron and has a high morbidity and mortality rate that result from damage to cardiac, hepatic, and endocrine tissues. This disease, however, is preventable if identified and treated early by simple phlebotomy, which removes excess iron (1). However, whether HFE genetic screening has any added value over transferrin iron saturation or plasma ferritin is still being debated. The recent finding that C282Y heterozygosity may be associated with an increased risk of cardiovascular death adds to its public health importance, but this finding also is currently a matter of scientific debate (30)(31).
Many individuals with hemochromatosis (>60%) are homozygous for the missense mutation C282Y. Compound heterozygosity or homozygosity for the H63D mutation are also associated with hemochromatosis, but with very low penetrance. No more than ∼1% of the compound heterozygotes will develop hemochromatosis. This value is even lower for H63D homozygotes. Thus, ∼99% of compound heterozygotes that would be found in either a subpopulation of relatives of C282Y homozygotes or in the general population would be false positives (1). Therefore, we believe that H63D genotyping is relevant only for C282Y heterozygotes and only in those cases where clinical suspicion of hemochromatosis remains, as assessed by biochemical tests such as plasma ferritin and transferrin saturation. For a clinical laboratory engaged in hemochromatosis screening, the first-line genetic test is C282Y genotyping.
In this study, we have demonstrated the feasibility of using real-time AS-PCR and DNA dissociation curves to genotype the C282Y HFE locus. Compared with the widely used C282Y restriction isotyping assay, our method is less time-consuming and labor-intensive because it does not require any post-PCR processing. In addition, reports by both Jeffrey et al. (32) and Somerville et al. (33) have shown that the restriction isotyping assay, when used with the primers of Feder et al. (2), may have the potential to incorrectly classify a C282Y heterozygote as a C282Y homozygote. This is caused by a newly identified SNP (5569 G/A) located in the binding region of the Feder et al. (2) antisense primer. The primers used in our method avoid this polymorphism (Fig. 1⇑ ).
Recently, several real-time HFE PCR genotyping methods based on the principle of fluorescence resonance energy transfer have been developed (34)(35)(36). These methods use fluorescently labeled oligonucleotide probes. In contrast, our real-time C282Y assay uses novel AS primers and the inexpensive SYBR Green I DNA-binding dye to generate AS dissociation curves. Our report demonstrates that this approach is both reliable and robust. This is illustrated by the high degree of reproducibility obtained for each different allele dissociation curve. These curves were generated using several different individuals (Fig. 4⇑ ). Previously, we have shown that DyNAzyme II, a Thermus brockianus enzyme, functions specifically in multiplex AS-PCR (37). This observation is confirmed by the present report.
The capital costs of the GeneAmp 5700 Sequence Detection System are high. Therefore, to maximize productivity, the machine should be used for a variety of different assays. Consequently, machine time available to individual users may become scarce. We have shown, however, that it is possible to perform the PCR step in a conventional thermocycler and that a subsequent 20-min dissociation curve analysis of the products by the GeneAmp 5700 system gives unequivocal results (Fig. 3B⇑ ). Thus, an added benefit of this genotyping technique is its versatility, which yields substantial savings in machine time.
A potential disadvantage of the current method is that SYBR Green I binds to all amplification products, including primer dimers, and this could cause difficulty in identifying the intended amplicon. Although primer dimer formation did occur during the initial optimization phase, the primer dimers were differentiated from the desired amplicons by differences in melting temperatures. Moreover, the formation of primer dimers was prevented by a combination of increased annealing temperature, alteration in the ratio of primers, and a reduction in the number of PCR cycles (Fig. 2⇑ ). We found that, as reported by Rust et al. (23), it was necessary to use the longer AS primer (HFEW2) at a lower concentration than the shorter primer (HFEM). This probably reflects the greater annealing advantages of longer primers at high temperatures.
This assay format should represent a straightforward way to genotype the C282Y HFE locus. Theoretically, a similar strategy could be applied to the analysis of other clinically established SNPs. However, as Germer and Higuchi (27) pointed out, it remains to be seen whether all SNPs can be genotyped using specially designed AS primers and SYBR Green I. The fact that some SNP targets are very GC rich may make the optimization and discrimination of their PCR products difficult, but not necessarily impossible. Thus, an obvious aid for designing good PCR primers for use in this system is a method that can accurately predict the melting temperature of the resulting PCR product. We calculated the theoretical melting temperature for the HFE products, using the equation of Howley et al. (38) as described by Bohling et al. (39). Comparison of the theoretical melting temperature with the measured melting temperature revealed an overestimation of 2.4 °C for the HFEM amplicon and 3.7 °C for the HFEW2 amplicon. This discrepancy, as Bohling et al. (39) have suggested, probably results from the addition of SYBR Green I, which has been shown to affect dissociation curves in a concentration-dependent fashion. In addition, the equation developed by Howley et al. (38) itself comes from low-resolution melting experiments in the 1960s and 1970s, and therefore is imprecise. Thus, both of these factors probably account for the discrepancy. Additional work will be needed to create an algorithm for accurate prediction of oligonucleotide melting temperatures in a buffer containing SYBR Green I.
In conclusion, a new single-tube HFE genotyping method that combines the principles of AS-PCR, mutagenically separated PCR, and amplicon identification by SYBR Green I dissociation curves has been developed. The method is accurate and reliable and does not require oligonucleotide probes. It represents a rapid approach for HFE genotype determination.
Acknowledgments
This study was supported in part by a grant from the Center for International Mobility (CIMO; The Finnish Ministry of Education). We thank Dr. Eleanor Ryan for outside validation of our assay, and Dr. Richard Owczarzy and Profs. Kojo Elenitoba-Johnson, Carl Wittwer, and James G. Wetmur for helpful discussions on DNA melting.
Footnotes
-
↵1 Nonstandard abbreviations: HH, hereditary hemochromatosis; SNP, single nucleotide polymorphism; AS, allele-specific; and Ct, threshold cycle.
- © 2000 The American Association for Clinical Chemistry
References
In this issue
