This work was supported by NHS R&D, and the British Bone Marrow Donor Association (BBMDA). There are no conflicts of interest between the authors.
Molecular typing of HLA genes using whole genome amplified DNA
Article first published online: 14 OCT 2008
© 2008 American Association of Blood Banks
Volume 49, Issue 1, pages 57–63, January 2009
How to Cite
Creary, L. E., Girdlestone, J., Zamora, J., Brown, J. and Navarrete, C. V. (2009), Molecular typing of HLA genes using whole genome amplified DNA. Transfusion, 49: 57–63. doi: 10.1111/j.1537-2995.2008.01943.x
- Issue published online: 23 DEC 2008
- Article first published online: 14 OCT 2008
- Received for publication June 10, 2008; revision received August 11, 2008; and accepted August 14, 2008.
BACKGROUND: The outcome of clinical transplantation and a number of disease susceptibilities show very strong associations with genetic variants within the major histocompatibility complex, particularly in the human leukocyte antigen (HLA) genes. A problem with many association studies is the lack of sufficient DNA to perform multiple genetic analyses, particularly with transplantation outcomes where donor and recipient DNA are often in short supply. This study assesses whether a multiple-strand displacement whole genome amplification (WGA) method could generate sufficient template of high quality to perform unbiased amplification for analysis of the HLA-A, -B, -C, -DRB1, and -DQB1 genes.
STUDY DESIGN AND METHODS: A panel of DNA samples from various biological sources was subjected to WGA reaction using Φ29 DNA polymerase. The HLA genotypes were subsequently determined using standard polymerase chain reaction (PCR)-based methods including sequence-specific oligonucleotide probes (PCR-SSOP, Luminex, Luminex Corp.) and sequence-based typing (PCR-SBT). WGA products and original DNA samples were used to determine the sensitivity of the Luminex assay; in addition, reamplified WGA products were also genotyped.
RESULTS: The WGA templates, as well as serially amplified DNA for two successive rounds, yielded HLA genotypes fully concordant with those determined for the original DNA samples. WGA products and original DNA gave reproducible HLA-DQB1 genotypes with 100 to 10 ng of template. Purification of the WGA products was required for successful PCR-SBT, but not for the PCR-SSOP method.
CONCLUSION: Our study suggests that WGA can be a reliable method for generating unlimited DNA for medium- or high-resolution HLA typing using the techniques described above.
multiple displacement amplification
sequence-specific oligonucleotide probes
whole genome amplification.
The highly polymorphic nature of the major histocompatibility complex (MHC) requires that numerous genetic tests are carried out for transplantation and disease association studies. Generally, Class I and Class II genes of the human leukocyte antigen (HLA) system are typed first at low or medium resolution. Thereafter, high-resolution typing may be used to further dissect subgroups of alleles and also to confirm uncertain or ambiguous HLA types. Moreover, high-resolution typing techniques for both Class I and Class II HLA antigens are being increasingly used to identify optimal donors for allogeneic hematopoietic stem cell transplantation. Studies have shown that a higher degree of HLA compatibility at the allele level between the donor-patient pair is associated with a lower incidence of posttransplant complications such as lower risk of acute and chronic graft-versus-host disease, graft rejection, and better overall survival.1-3 However, a thorough evaluation of the HLA system (or better matching between donor-recipient pairs) may be compromised if DNA samples are in limited supply or degraded. In light of this, we explored whether a whole genome amplification (WGA) method would give complete coverage and unbiased amplification for standard typing of genetic markers in HLA-A, -B, -C, -DRB1, and -DQB1 loci along chromosome 6p21-p24. We used a commercially available illustra GenomiPhi HY DNA amplification kit (GE Healthcare, Buckinghamshire, UK), which exploits the unique properties of the DNA polymerase from bacteriophage Φ29 to replicate the entire human genome by multiple displacement amplification (MDA).4-6 In the presence of random hexamers and optimal buffer conditions Φ29 DNA polymerase can amplify the entire DNA isothermally at 30°C thereby eliminating biased genomic region amplification that can occur under thermal cycling conditions. Other favorable characteristics of Φ29 DNA polymerase include its high 3′-to-5′ exonuclease proofreading activity (error rate of only 1 in 106-107 bp)7 and the high processivity (adds approx. 70,000 nucleotides every time it binds to a primer) generating extremely long DNA products averaging approximately 12 kb in length and ranging from 10 to 100 kb. These properties of Φ29 DNA polymerase ensure that accurate replication of the complete genome can occur in a single reaction producing on the order of a thousandfold amplification of DNA from little starting material (approx. 10 ng).
In this study, we evaluated how well the MDA WGA method performs for increasing genomic DNA stocks from blood and cultured cells. Importantly, we also determined whether this material was suitable for high-resolution HLA typing. Specifically, our goals were to determine the amplification yield of DNA from various starting concentrations of genomic DNA and to compare the HLA genotyping success rates of nonamplified and WGA DNA using sequence-specific oligonucleotide probes (PCR-SSOP, Luminex, Luminex Corp., Austin, TX) and sequence-based typing (PCR-SBT).
MATERIALS AND METHODS
Genomic DNA was extracted from 18 fresh human buffy coats, 3 human mesenchymal cell lines derived from cord blood (approx. 1 × 106 cells), and 1 human fibroblast cell line (HS27; approx. 1 × 106 cells), from the American Type Culture Collection using a fully automated robotic workstation (GenoM-6 Genovision, Vienna, Austria; Qiagen, Hilden, Germany) according to the manufacturer's instructions. In brief, the automated extraction method consisted of cell lysis using chaotropic reagents, binding of the DNA to silica-coated magnetic particles, washing steps, and elution of pure nucleic acid samples in Tris-ethylenediaminetetraacetate buffer solution. DNA was quantified by ultraviolet (UV) spectrophotometry at 260 nm using a photometer (NanoDrop-1000, NanoDrop, Wilmington, NC). The presence of DNA in the samples was also confirmed by gel electrophoresis using a 1 percent agarose gel stained with ethidium bromide followed by illumination with ultraviolet light.
WGA was performed using the illustra GenomiPhi HY DNA midi amplification kit (GE Healthcare) following the manufacturer's protocol. In brief, DNA was diluted to 10 ng per µL in distilled water, and 1 µL was added to 22.5 µL of sample buffer. A DNA sample provided with the kit served as a positive control, and the negative control was a nontemplate (water only) sample. Samples were denatured for 3 minutes at 95°C, placed on ice, and then added to 22.5 µL of reaction buffer and 2.5 µL of enzyme mix. Incubations were carried out at 30°C for 4 hours using a DNA thermocycler (Phoenix, Helena Biosciences, Sunderland, UK).
A subset of samples was subjected to four successive rounds of WGA (Rounds R1 to R4). In brief, this consisted of dilution of the initial WGA product (denoted R1) to 10 ng per µL, based on spectrophotometric analysis at 260 nm and using it as a template for a second WGA reaction (R2). Amplification of 10 ng per µL WGA products from Round 2 resulted in R3 followed by amplification of R3 to give the final WGA product 4 (R4). WGA products used for successive rounds of WGA and sequence-based HLA genotyping were column purified before use by gel filtration cartridges (Performa DTR, EdgeBio, Gaithersburg, MD). Briefly, the gel filtration cartridges were centrifuged for 3 minutes at 840 × g to remove the packaging fluid. The cartridge was placed in a 1.5-mL collection tube and the WGA product was placed in the packed column. The column was centrifuged using the conditions described previously, and the eluate was retained. All WGA products were quantified and confirmed using the spectrophotemetric and electrophoretic techniques described for the original DNA (pre-WGA) samples.
HLA-A, -B, -C, -DRB1, and -DQB1 genotyping
Reverse-SSOP using the Luminex platform
Pre- and post-WGA samples were typed at low to medium resolution by PCR and SSOP using HLA kits (Lifecode, Tepnel Lifecodes Corp., Stamford, CT) and Luminex xMAP technology (Luminex Corp.). The assay was performed according to the manufacturer's instruction, except that the PCR volume used was 25 µL rather than 50 µL with the amounts of all reagents reduced proportionally. For all HLA loci, template was diluted to 50 ng per µL and used at a total amount of 200 ng per PCR.
A subset of pre- and post-WGA samples (n = 8 from each group) was typed at high resolution for all HLA loci, using previously described methods.8-10 A DNA template amount of 100 ng was used for HLA Class I PCR and 250 ng was used for HLA Class II PCR. HLA Class I loci SBT entailed PCR amplification of Exons 2 and 3, whereas for HLA Class II loci Exon 2 was amplified. PCR products were checked on a 2 percent agarose gel stained with ethidium bromide and treated with exonuclease I and shrimp alkaline phosphatase (ExoSAP-IT, GE Healthcare) to remove excess primers and dNTPs.
Sequencing was performed in both forward and reverse directions using a cycle sequencing kit (ABI BigDye Terminator v1.1, Applied Biosystems, Foster City, CA) as recommended by the manufacturer. Sequencing products were purified using the Performa DTR gel filtration cartridges and analyzed on a genetic analyzer (Model 3130, Applied Biosystems).
HLA alleles were assigned using computer software (Assign-SBT v3.5, Abbott Molecular, Wiesbaden, Germany), which compares the test sequence with the most recent reference sequence from the IMGT/HLA database (http://www.ebi.ac.uk/imgt/hla/).
All experiments were carried out in accredited diagnostic laboratories, following good laboratory practice, using separate facilities for DNA isolation, processing, amplification steps, and postamplification analysis. No crosscontamination was observed.
Analysis of WGA products
Genomic DNA samples extracted from various biological sources exhibiting a wide range of DNA yields were subjected to WGA. Analysis of the post-WGA products by agarose gel electrophoresis showed the presence of high-molecular-weight products ranging from 1 to more than 20 kb (Fig. 1A, Lanes 1-8). As forewarned by the GenomiPhi HY kit manufacturer, the negative nontemplate DNA control shows an electrophoretic profile similar to the genomic DNA test samples and the positive control. Also, spectrophotometric analysis of the negative control indicates a high yield of DNA (approx. 2550 ng/µL). These results are nonspecific and are due to Φ29 DNA polymerase using the random hexamers primers in the reaction mix as template for amplification. Therefore, spectrophotometric and gel electrophoretic analysis are not reliable methods to confirm successful WGA products. It is therefore imperative to use negative control samples in further downstream reactions to validate the performance of the WGA method. However, to standardize the use of WGA DNA for downstream applications, we used OD260nm absorbance to determine the concentration of product. By this measure, 10 ng of genomic DNA yielded a mean of 123.6 µg (±16 µg; range, 95.2 to 156.5 µg) when amplified with a “midi” kit, corresponding to a fold increase of 12,360×.
Some WGA samples were column purified prior to SBT and for reamplification of WGA products. Figure 1B shows a typical electrophoretic profile of purified WGA products; the mean DNA length is more than 12 kb, ranging from 10 to more than 20 kb. The profile does not differ greatly from the electrophoretic profile of unpurified WGA products except that the bands are of lesser intensity, and low-molecular-weight products in each sample are not visible on the gel. However, even after column purification a high-molecular DNA product is still present in the negative control corresponding to approximately 680 ng per µg (OD260nm), suggesting that the column purification step has primarily removed small-molecular products such as dNTPs and primers. The yields from reamplification of WGA products after purification are similar to those shown in Table 1 (data not shown).
|WGA product||Total DNA (µg)/50-µL reaction||Standard deviation||Range||Amplification fold|
|Unpurified (n = 22)||123.6||16||95.2-156.5||12,361×|
|Purified (n = 8)||31||7.2||21.2-40.4||3,107×|
HLA-A, -B, -C, -DRB1, and -DQB1 genotyping by SSOP Luminex technology
To determine whether the WGA products are accurate replicates of the original unamplified samples, 22 unamplified DNA samples and their corresponding WGA products were typed for HLA-A, -B, -C, DRB1, and -DQB1 by Luminex-based PCR-SSOP. All 22 WGA products generated genotypes fully concordant with unamplified DNA for all 5 HLA loci tested, with the exception of one sample that failed for HLA-C typing with both the original and the WGA templates.
Dilution experiments were performed to evaluate the sensitivity of the Luminex assay and to provide an empirical method for determining the effective concentration of WGA DNA. Four DNA samples extracted from buffy coats and their corresponding WGA products were used at a total DNA amount of 100, 10, 1, and 0.1 ng per PCR procedure, according to OD260nm, for DQB1 genotyping. Genotypes were generated in 3 pre- and post-WGA pairs used at 10 ng or more (S1, S2, S3; Table 2). In one sample (S4) a DNA amount of 1 ng was sufficient for successful typing in both post and pre-WGA DNA samples.
|Sample||Pre-WGA DQB1 genotype||Post-WGA DQB1 genotype|
|S1_100 ng||0201/02/04, 0301/09/19||0201/02/04, 0301/09/19|
|S1_10 ng||0201/02/04, 0301/09/19||0201/02/04, 0301/09/19|
|S1_1 ng||No typing||No typing|
|S1_0.1 ng||No typing||No typing|
|S2_100 ng||0201/02/04, 0301/09/19||0201/02/04, 0301/09/19|
|S2_10 ng||0201/02/04, 0301/09/19||0201/02/04, 0301/09/19|
|S2_1 ng||No typing||No typing|
|S2_0.1 ng||No typing||No typing|
|S3_100 ng||0201/02/03/04, 0201/02/03/04||0201/02/03/04, 0201/02/03/04|
|S3_10 ng||0201/02/03/04, 0201/02/03/04||0201/02/03/04, 0201/02/03/04|
|S3_1 ng||No typing||No typing|
|S3_0.1 ng||No typing||No typing|
|S4_100 ng||0602/03/11/14/16/19, 0603/11/14/16/19/28||0602/03/11/14/16/19, 0603/11/14/16/19/28|
|S4_10 ng||0602/03/11/14/16/19, 0603/11/14/16/19/28||0602/03/11/14/16/19, 0603/11/14/16/19/28|
|S4_1 ng||0602/03/11/14/16/19, 0603/11/14/16/19/28||0602/03/11/14/16/19, 0603/11/14/16/19/28|
|S4_0.1 ng||No typing||No typing|
To determine whether high-fidelity WGA template could be generated from a WGA product without loss of genetic information, six samples were subjected to four successive rounds of WGA (R1 to R4). The samples were derived from various biological sources: S5, S6, and S7 from human buffy coats; MC8 and MC9 from mesenchymal cells isolated from cord blood; and HS27 is a human fibroblast cell line. The samples were genotyped for genetic markers in all HLA genes. For two samples (MC8 and MC9) there was 100 percent concordance between the original unamplified DNA and their WGA products generated for all four successive rounds of amplification (Table 3). For Samples S7 R4 typed at the B locus, Samples S5 R3 and R4 and S6 R3 and R4 typed at the C locus genotypes failed to be generated. Interestingly, genotyping failed for WGA templates HS27-R2 typed at the HLA A gene and S6-R2 typed at HLA DRB1 but was successful for subsequent rounds.
|Sample||Concentration (ng/µL)||Amplification fold||HLA gene concordance rates (%)|
HLA-A, -B, -C, -DRB1, and -DQB1 genotyping by SBT
SBT provides the highest resolution typing of HLA alleles and is the only method that can type to the single allele level and define new alleles. On several independent occasions we attempted to sequence nonpurified WGA products, but failed to produce a PCR product (data not shown). Following the guidelines of the amplification kit manufacturer, WGA products were then column purified before SBT, and all samples amplified successfully. However, purification does lead to an apparent loss of DNA. Table 1 shows that after column purification the mean DNA yield decreased to 31.1 µg per 50-µL reaction volume, which equates to a fold increase of 3107× (±0.72; range, 2124 to 4041).
Eight column-purified WGA products (sourced from seven buffy coats and one mesenchymal cell line) along with their corresponding unpurified original DNA samples were HLA typed by SBT. All WGA products generated genotypes fully concordant with unamplified DNA for all 5 HLA loci tested. Figure 2 shows a partial electropherogram sequence of allele HLA-A*02010101 from the unamplified and WGA product; as can be seen there are no discernible differences between the two traces.
A major hindrance in genetic association studies is the limited source of good quality DNA. This problem is further underscored when the DNA samples are precious and it is not feasible to go back to the donor source for additional biological material, such as in the case of transplanted patients and their donors. Therefore, we explored whether WGA could be used to generate DNA of high yield and fidelity for future candidate gene association studies.
We chose to validate a non-PCR multiple-displacement WGA method because of its reported capability of accurate genome replication, which is imperative to generate reliable samples for many downstream genetic applications. The illustra GenomiPhi HY midi prep kit is based on MDA of linear genomic DNA and is reported to produce 40 to 50 µg from 10 ng of starting template DNA.
We confirmed that spectrophotometric and gel electrophoresis analysis were not reliable procedures for estimating the quality and quantity of WGA products as shown by the presence of a high-molecular-weight DNA product in the water no-template control. Therefore, it is important to validate the MDA WGA method empirically by further downstream genetic tests. All the 22 original samples, derived from whole blood as well as cultured cells, amplified successfully, and it is important to highlight that some DNA samples were more than 10 years old and had been stored between −20 and 80°C. Although this suggests that older and perhaps slightly more degraded samples are amenable to WGA, other studies have found that highly degraded DNA cannot be used for WGA.11
We found that a starting template of 10 ng of genomic DNA yielded a mean of 123.6 µg, which would allow for approximately 618 PCR procedures, using 200 ng per PCR procedure, for genotyping using Luminex technology. Column purification of WGA products was found to be necessary for successful SBT genotyping. Although this resulted in some loss of DNA, the mean yield of 31 µg per reaction, this amount was still sufficient for 310 HLA Class I (100 ng/PCR procedure) or 124 PCR procedures for HLA Class II (250 ng/PCR procedure) sequencing reactions. We found that the genotypes generated from all WGA products were fully concordant with unamplified DNA samples by SSOP Luminex based typing. Sequence integrity of the samples was further validated by SBT. We found that there was no difference in sequence or genotyping results between the unamplified and WGA products.
We also explored whether reamplification of WGA products would introduce mutations resulting in misgenotyping, or increased incidence of allele dropout. Samples were subjected to four successive rounds of reamplification using 10 ng of purified WGA product as the template. For two samples there were no discernible differences in genotyping results between WGA products from Amplification Rounds 1 to 4, suggesting that reamplified WGA products can be used for HLA genotyping. For the remaining four samples, genotyping success was sporadic: some samples failed to generate genotypes for Rounds 3 and 4, for example, S5 R3, R4 for HLA C, whereas others genotypes were produced for Rounds 1, 2, and 3 but failed for Round 2, for example, HS27 typed at the HLA A gene and S5 typed at DRB1 gene. This pattern suggests that genotyping failure may be due to other sources of error rather than the quality of the reamplified WGA product.
The validity of our findings is further supported by other studies that used various biological samples such as mouth swabs, blood, cell lines, and dried blood spots for successful HLA typing of WGA products.12-14 Our laboratory has recently shown that WGA of genomic DNA from neonatal samples is successful for human platelet alloantigen genotyping from patients with neonatal alloimmune thrombocytopenia (unpublished data). WGA could be used as a tool for enriching DNA from clinical samples with limited cell populations as seen in pancytopenic patients such as leukemia, aplastic anemia, and splenomegaly. Nagy and colleagues15 recently showed that WGA was useful for single-tandem repeats profiling for chimerism analysis after allogeneic hematopoietic stem cell transplantation.
Our study is the first to show the limits of sensitivity of the Luminex assay for HLA-DQB1 genotyping of WGA products and also that reamplified MDA WGA products can be used for Luminex genotyping of HLA-A, -B, -C, -DRB1, and -DQB1 loci. In addition we also demonstrated that MDA WGA products must be purified for successful HLA Class I and Class II loci typing using sequence-based methods. In conclusion, our results indicate that the GenomiPhi HY MDA WGA method is reliable and can generate accurate replicates for medium- to high-resolution HLA genotyping using SSOP Luminex-based technology and SBT on DNA extracted from fresh blood samples and cultured cell lines.
The authors thank NHSBT Histocompatibility and Immunogenetics Department, Colindale, UK.
- 11Exploring whole genome amplification as a DNA recovery tool for molecular genetic studies. J Biomol Tech 2005;16:125-33., , .