| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Bristol Genetic Epidemiology Laboratory and 2 Department of Social Medicine, University of Bristol, Bristol, United Kingdom; 3 Department of Epidemiology & Population Health, London School of Hygiene & Tropical Medicine, London, United Kingdom; 4 Human Genetics Division, Southampton General Hospital, Southampton, United Kingdom;
aaddress correspondence to this author at: Bristol Genetic Epidemiology Laboratory, University of Bristol, No. 24 Tyndall Ave., Bristol, United Kingdom BS8 1TQ; fax 44-117-331-1729, e-mail r.abdollahi{at}bristol.ac.uk and rabdollahi{at}gmail.com
Abstract
Background: Apolipoprotein E (APOE) is an important element of lipid metabolism and, hence, cardiovascular disorders. APOE has 3 main allelic variants: 
3, 4, and 2. Of these, 3 is the most common, followed by 4 and 2. The associations of these isoforms with cardiovascular disorders and Alzheimer disease have been widely studied in different populations. Most of the genotyping in these studies has been performed with gel-based methods, which have important limitations, particularly for large epidemiologic studies. We therefore developed an integrated "one-tube" liquid-phase assay.
Methods: To measure APOE isoforms, we developed an integrated single-label liquid-phase fluorescence assay containing 2 PCR primers, 2 probes, and 2 quencher oligonucleotides. We used a 384-well LightTyper, but the assay would be generically applicable for use with any fluorescence detector with thermal ramp control. We validated this method and applied it in the British Women’s Heart and Health Study.
Results: There were 4 melting peaks, at 41, 56, 61, and 69 °C, which generated 6 distinctive patterns representing genotypic combinations of 3, 4, and 2. The magnitude and direction of the associations found with total cholesterol, HDL-cholesterol, triglycerides, and estimated LDL-cholesterol were consistent with previous reports.
Conclusion: The one-tube LightTyper assay presented here enables accurate, convenient, and economical genotyping of APOE and can be used for large epidemiologic studies.
Apolipoprotein E (APOE) has a pivotal role in the metabolism of chylomicrons, chylomicron remnants, VLDL-cholesterol, and HDL-cholesterol (HDL-C), and as a ligand, it binds both LDL-cholesterol (LDL-C) and APOE receptors (1). The APOE gene is located on chromosome 19q13.2 and has 3 main allelic variants: 3, 4, and 2. 3 is the most common, followed by 4 and 2. These 3 allelic variants differ at 2 single-base variations located in exon 4 at codon positions 112 and 158. The T and C alleles of APOE 112T>C (rs429358) and APOE 158C>T (rs7412) encode Cys and Arg, respectively: 2 has a T allele at both positions 112 and 158; 3 has T and C alleles at positions 112 and 158, respectively; and 4 has C at both positions. These isoforms have been widely studied because of their possible associations with coronary heart disease, type III hyperlipoproteinemia, strokes, and Alzheimer disease [see, for example, the review by Eichner et al.(2)].
Genotyping methods for APOE have been mostly gel based (3)(4)(5), and the fact that both codons represent HhaI restriction fragment length variations has enabled integrated PCR-digest gel-based assays. Nevertheless, the high GC richness of APOE exon 4 and the requirement for typing 2 single-base variations have been problematic for many approaches. Furthermore, because of the greater convenience and cost-effectiveness of liquid-phase assays compared with gel-based assays, there is increasing interest in switching to liquid-phase assays for large-scale epidemiologic studies. Thus, it would be desirable to establish an integrated "one-tube" liquid-phase assay.
Here we report a 6-oligonucleotide single-reaction approach to APOE genotyping in a 384-well microplate LightTyper (http://www.roche-applied-science.com/sis/lighttyper/index.htm). This method uses melting profiles of probes designed for typing single-base variations. An oligonucleotide pair (a quencher and a fluorescent probe) for each single-base variation was also present during PCR: the single-base variation binding probe (sense) was derivatized with fluorescein at the 5' end and phosphate at the 3' end and was complementary to one of the alleles represented in its middle third, and the quencher oligonucleotide was derivatized with Dabcyl at its 3' end located 4 bp (for APOE 112) or 3 bp (for APOE 158) upstream of the probe for the respective single-base variation. The anchor quencher oligonucleotide quenches the probe for the single-base variation while in its vicinity. With increasing temperature, the probe dissociates from its target strand, thus releasing its fluorescein from the vicinity of the Dabcyl quencher with a consequent increase in fluorescence. The first derivative of the fluorescence curve thus shows 2 peaks at each locus when the single-base variations are in the heterozygous state; thus, an 4/2 heterozygote shows 4 peaks in total.
For this study, we used genomic DNA from 3271 white British women 60–79 years of age who were participants in the multicenter British Women’s Heart and Health Study (BWHHS); ethics consent and DNA bank set up were as described previously (6)(7). Venous blood was taken after a fast of at least 6 h and was used to determine total cholesterol, HDL-C, LDL-C, and triglycerides (TGs).
Total cholesterol, HDL-C, and TGs were measured on frozen serum samples (maximum time frozen, 6 weeks) with a Hitachi 757 analyzer (Roche Diagnostics) and standard reagents. LDL-C was estimated using the Friedewald equation [LDL-C = total cholesterol – (HDL-C + TGs x 0.45)] (8). For total cholesterol, the within-batch imprecision (CV) was 0.96% at 5.11 mmol/L and 0.81% at 7.21 mmol/L (both based on 20 replicates), and between-batch imprecision was 1.3% at 5.24 mmol/L and 2.1% at 7.21 mmol/L (both based on 13 replicates).
For the LightTyper assay, a 495-bp sequence (GenBank accession no. AF261279) of APOE spanning 112T>C and 158C>T was amplified with 2 primers (MWGBiotech): forward, 5'-GCCTACAAATCGGAACTGGA-3'; reverse, 5'-ACCTGCTCCTTCACCTCGT-3'. The probes for the APOE single-base variations 112T>C and 158C>T were as follows: APOE 112F, 5'-fluorescein-GGACGTGCGCGGC-phosphate-3'; APOE 158F, 5'-fluoresceinTGCAGAAGCGCCTGGCAGTGTACC-phosphate-3'. The quencher oligonucleotides were as follows: APOE 112D, 5'-GCGCAGGCCCGGCTGGGCGCGGACA-Dabcyl-3'; APOE 158D, 5'-GCGTAAGCGGCTCCTCCGCGATGCCGATG-Dabcyl-3'.
Templates contained 20 ng of genomic DNA and were plated on 384-well PCR plates [cat. no. TF-0384/W; ABgene (www.abgene.com)] in 2 µL of water and dried at 80 °C for 10 min for storage. The PCR mixture contained 0.5 µL of 10x PCR buffer, 0.05 mM deoxynucleotide triphosphates, 0.05 µM forward primer and 0.25 µM reverse primer (i.e., asymmetric PCR), 0.04 µM each probe, 1.3 M betaine, 2 mM MgCl2, 0.025 U/µL Taq DNA polymerase (Promega), and H2O to 5 µL. Thermal cycling, performed on a DNA Engine Tetrad® 2 (MJ Research), was as follows: 94 °C for 2 min, 94 °C for 45 s, 61 °C for 45 s, and 72 °C for 45 s (last 3 steps repeated for 90 cycles), followed by 72 °C for 2 min. PCR products were covered with 5 µL of Chillout 14TM liquid wax (MJ Research), and then centrifuged at 3000 rpm for 3 min before plate analysis in the LightTyper instrument (cat. no. 03 357414001; Roche Diagnostics GmbH).
Genotypes were determined automatically with LightTyper software, Ver. 1 (Roche Diagnostics GmbH), and validated by in-house software (9). Genotyping using HhaI restriction digestion(5) was also performed as a validation test on 96 random samples representing all 6 genotypes.
As shown in Fig. 1
, the 6 APOE genotypes had distinctive patterns. The peaks at 41 and 56 °C represented the T and C alleles of APOE 112T>C, respectively, and the peaks at 61 and 69 °C corresponded to the T and C alleles of APOE 158C>T, respectively. The characteristic peaks for the different genotypes were as follows: 3/3 gave 2 peaks at 41 and 69 °C; 4/4 gave 2 peaks at 56 and 69 °C; 2/2 gave 2 peaks at 41 and 61 °C; 3/4 gave 3 peaks at 41, 56, and 69 °C; 3/2 gave 3 peaks at 41, 61, and 69 °C; and 4/2 gave 4 peaks at 41, 56, 61, and 69 °C (Fig. 1
). The relative peak heights also showed characteristic quantitative differences according to genotype. The frequencies of 3, 4, and 2 were 0.78, 0.14, and 0.08, respectively. We genotyped 94% of samples successfully, and genotype frequencies were in Hardy–Weinberg equilibrium (
2 = 5.722; P = 0.13; df = 3).
|
As a further validation of this assay, we examined the association of APOE with lipids in the participants (see Table 1 in the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem.org/content/vol52/issue7/). Total cholesterol, HDL-C, TGs, and estimated LDL-C were all associated with APOE genotype. The magnitudes and directions of these associations were consistent with those reported in the literature for similar populations (10). Thus, total cholesterol and LDL-C were lower in individuals with the 2 allele and higher in those with the 4. For HDL-C, 2 allele associations were more notable, and for TGs, the effect of 4 homozygosity was most notable. For LDL-C and total cholesterol, but not for HDL-C, the effect of the 2 allele dominated that of the 4 allele in 2/4 heterozygotes.
We have established a one-tube LightTyper assay for genotyping APOE alleles 2, 3, and 4 that could be performed in a 384-well LightTyper instrument. This assay is suitable for epidemiologic studies. To maximize resolution of alleles carrying single-base variations, it is preferable to choose the most destabilizing base pair mismatch to maximize the difference in melting temperature. The difference in melting temperature between the quencher nucleotide and the probe for the single-base variation is also of major importance.
In our initial development (data not shown), we used both separate short amplicons spanning codons 112 and 158, respectively, and subsequently, separate binding assays on one larger amplicon. The former was used to genotype the same initial core plate of 360 samples. The peak heights were somewhat higher for the 112 assay when we used the separate short amplicons, but all genotype calls were the same with the 2-assay format, the integrated assay, and the gel-based assay.
Integrated assays of 2, 3, and 4 have included molecular haplotyping using single-strand conformational polymorphism analysis (11), denaturing gradient gel electrophoresis(12), multiple-primer single-base extension capillary electrophoresis Snapshot assays(13), and a commercial dual-wavelength fluorescence assay (Roche LightCycler). The first 3 are gel-based methods that are not economical for use in high-throughput assays, and the latter is a dual-label approach that incompletely differentiates the 4 relevant melt peaks by temperature and uses proprietary design and reagents. The combination of 2 probes for codons 112 and 158 that give 4 well-separated melting temperatures for respective mismatch and match binding in conjunction with simple end-point fluorescence reading using 1 label and wavelength has enabled the integrated liquid-phase assay design described here.
Studies of 2/4 heterozygotes have revealed that the E4 protein isoform is catabolized from all lipoprotein fractions 
3-fold faster than the E2 protein isoform (14), whereas E3 is catabolized at an intermediate rate(15)(16)(17). In 2/4 heterozygotes, E2 is found predominantly in HDL-C, whereas E4 is equally distributed in VLDL and HDL(14). The APOE isoforms differ in metabolic pathways through the classes and sizes of lipoproteins underpinning the observed associations with total cholesterol, LDL-C, HDL-C, and TGs. The consistency of associations observed for each genotype group with other studies further validates this new integrated APOE assay. The same oligonucleotide mixture and PCR conditions may be suitable for adaptation to other PCR instruments and fluorescence readers.
Acknowledgments
M.R.A. and P.A.I.G. were supported by the University of Bristol. BWHHS is funded by the UK Department of Health. Genetic studies in BWHHS have been supported by the British Heart Foundation (BHF). D.A.L. is funded by a UK Department of Health career scientist award.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
