Clinical Chemistry 51: 35-39, 2005.
First published October 28, 2004; 10.1373/clinchem.2004.039164
(Clinical Chemistry. 2005;51:35-39.)
© 2005 American Association for Clinical Chemistry, Inc.
|
Molecular Diagnostics and Genetics |
Sequencing of Genomic DNA by Combined Amplification and Cycle Sequencing Reaction
Kathleen M. Murphy1,a,
Karin D. Berg1,2 and
James R. Eshleman1,2
Departments of1
Pathology and 2
Oncology, Johns Hopkins Medical Institutions, Baltimore, MD.
aAddress correspondence to this author at: Johns Hopkins School of Medicine, Carnegie Bldg., Room 358, 600 North Wolfe St., Baltimore, MD 21287. Fax 410-614-7440; e-mail kmurphy4{at}jhmi.edu.
 |
Abstract
|
|---|
Background: Despite considerable advances, DNA sequencing has remained somewhat time-consuming and expensive, requiring three separate steps to generate sequencing products from a template: amplification of the target sequence; purification of the amplified product; and a sequencing reaction. Our aim was to develop a method to routinely combine PCR amplification and cycle sequencing into one single reaction, enabling direct sequencing of genomic DNA.
Methods: Combined amplification and sequencing reactions were set up with Big DyeTM sequencing reagents (Applied Biosystems) supplemented with variable amounts of forward and reverse primers, deoxynucleotide triphosphates (dNTPs), and input DNA. Reactions were thermal-cycled for 35 or 45 cycles. Products were analyzed by capillary electrophoresis to detect sequencing products.
Results: Reactions using two oligonucleotide primers at a ratio of 5:1 (500 nM primer 1 and 100 nM primer 2), 125 µM supplemental dNTPs, and 35–45 thermal cycles optimally supported combined amplification and cycle sequencing reactions. Our results suggest that these reactions are dominated by PCR during early cycles and convert to cycle sequencing in later cycles. We used this technique for a variety of sequencing applications, including the identification of germline mutations/polymorphisms in the Factor V and BRCA2 genes, sequencing of tumor DNA to identify somatic mutations in the DPC4/SMADH4 and FLT3 genes, and sequencing of 16S ribosomal DNA for bacterial speciation.
Conclusions: PCR amplification and cycle sequencing can be combined into a single reaction using the conditions described. This technique allows direct sequencing of genomic DNA, decreasing the cost and labor involved in gene sequencing.
|
Introduction
|
|---|
DNA sequencing (1) is considered the gold standard for the detection and characterization of mutations. Major advances in this technology, such as the development of automated sequencers [reviewed in Ref. (2)], fluorescent terminator chemistry (3), and cycle sequencing (4), have made DNA sequencing easier to perform and more widely used. Cycle sequencing, similar to PCR amplification, uses a thermostable polymerase to generate sequencing products during repeated cycles of denaturation, annealing, and extension (4). Because multiple rounds of synthesis/termination are used during cycle sequencing, stronger sequencing signals are generated with less DNA template. Despite this, current DNA sequencing techniques still require amplification of the template (by PCR or plasmid cloning) before cycle sequencing.
Although PCR amplification and cycle sequencing share many common features, the two reactions are usually performed separately for several reasons. One reason is that PCR requires two primers for amplification of the target. Under routine conditions, simultaneous use of two primers in a sequencing reaction leads to two superimposed sequences (forward and reverse) that preclude interpretation. Another reason is that PCR amplification and cycle sequencing reactions have traditionally been carried out with two different DNA polymerases. In contrast to standard Taq DNA polymerases used for amplification, cycle sequencing is generally carried out with a Taq DNA polymerase that contains a point mutation, usually phenylalanine to tyrosine at position 667, in the active site of the enzyme. This mutation permits efficient incorporation of dideoxynucleotide triphosphates (ddNTPs), which usually are highly discriminated against by standard wild-type Taq DNA polymerase (5).
Currently, most sequencing protocols require at least three steps to generate sequencing products from a target: (a) amplification of the target; (b) purification of the resulting PCR amplicons to remove unincorporated deoxynucleotide triphosphates (dNTPs) and oligonucleotide primers; and (c) a sequencing reaction to generate sequencing products. Here we describe a sequencing method that reduces these three separate steps to a single reaction. This novel method uses two standard oligonucleotide primers at an unequal molar ratio, a single thermostable DNA polymerase, and a higher than usual concentration of dNTPs than is typical in sequencing reactions. We have experimentally determined a standard set of conditions that lead to robust combined amplification and sequencing reactions. Using the conditions we describe, the reaction is dominated by PCR during early cycles but converts to sequencing in later cycles, eliminating the need for additional reagents or manual manipulations. We demonstrate the utility of this technique for a variety of gene sequencing applications, including human genetics, tumor biology, and microbiology.
|
Materials and Methods
|
|---|
samples and dna isolation
DNA was isolated from human peripheral blood or xenografted tumors by use of a QIAamp DNA Isolation Kit (Qiagen) according to the manufacturer’s instructions. Exemption from review was obtained from the Johns Hopkins University School of Medicine Institutional Review Board.
For bacteria, several colonies of Escherichia. coli, Pseudomonas aeruginosa, Klebsiella pneumoniae, or Staphylococcus aureus were added to 200 µL of 200 g/L Chelex, vortex-mixed for 5–10 s, and incubated for 20–30 min at 56 °C. Samples were heated to 100 °C for 8 min and centrifuged at 10 000–15 000g for 2–3 min. We added 2 µL of the undiluted bacterial lysates or lysates diluted 1:10 to each AmpliSeq reaction.
AMPLISEQ reactions
All reactions were prepared with Big DyeTM v 2.0 or 3.0 sequencing reagents (Applied Biosystems) in a final volume of 20 µL. Primers used for AmpliSeq reactions were synthesized and purified by Oligos etc. as follows: Factor V, 5'-GCCCAGTGCTTAACAAGACCA-3' and 5'-AAGGTTACTTCAAGGACAAAATAC-3'; BRCA2, 5'-TATAGGGAAGCTTCATAAGTCAG-3' and 5'-ACCACCTTCAACATTTAAGTTATT-3'; FLT3, 5'-GCAATTTAGGTATGAAAGCCAGC-3' and 5'-CTTTCAGCATTTTGACGGCAACC-3'; MADH4, 5'-TGAAATCATAAGATGACATCTATGAATG-3' and 5'-TCCGGGATGGGGCGGCATAG-3'; 16S ribosomal DNA, 5'-CACAAGCGGTGGAGCATGTGG-3' and 5'-AGGCCCGGGAACGTATTCAC-3'.
We optimized the AmpliSeq reactions with use of the Factor V primers. Reactions consisted of 8 µL of Big Dye Ready Reaction Mix; two primers at unequal concentrations ranging from 5 to 500 nM; dNTPs at final supplemental concentration of 0, 12.5, or 125 µM; and 50–500 ng of DNA. The reactions were cycled at 95 °C for 30 s, followed by 35–45 cycles at 95 °C for 15 s, 50 °C for 15 s, and 60 °C for 4 min. Sequencing products were purified by use of spincolumns (Biomax) or ethanol precipitation, and automated sequencing was performed by capillary electrophoresis on an ABI3700 (Applied Biosystems).
After optimization, the standard conditions of a 20-µL reaction containing 8 µL of Big Dye Ready Reaction Mix, 500 nM forward primer, 100 nM reverse primer, and 125 µM supplemental dNTPs was applied to four other gene targets (Table 1
). Reactions were cycled as described above for 35 cycles for 500 ng of input DNA or for 40 or 45 cycles for 50 ng of input DNA. Products were purified and separated as described above.
To compare amplification and sequencing product production, we prepared multiple identical AmpliSeq reactions. At various cycle numbers, cycling was terminated for two reactions. To identify PCR amplification products, we added loading buffer to the products of one reaction and separated the products on a 10% polyacrylamide gel. To identify the cycle number at which readable sequencing products were generated, we purified the products from the duplicate reaction and subjected them to capillary electrophoresis as described above.
|
Results
|
|---|
optimization of reaction conditions
To combine PCR and dye-terminator chemistry cycle sequencing into a single reaction, we hypothesized that two modifications to a standard sequencing reaction would be necessary. We first would require a forward and a reverse primer complementary to the gene region of interest to enable PCR amplification. The two primers would need to be in unequal molar ratios so that sequence data would in effect be generated only from the primer in excess. The reaction would also require an increased starting ratio of dNTPs to ddNTPs to allow for full extension of PCR products by the polymerase (amplification) during early rounds of thermal cycling rather than termination (sequencing). Under these conditions, PCR would predominate during early cycles because the reaction would contain sufficient amounts of both forward and reverse primers and a relatively high concentration of dNTPs. In later cycles we theorized that the limiting primer would be depleted and the ratio of ddNTPs to dNTPs would increase because of consumption of dNTPs into PCR products during the early cycles. The change in relative reagent concentrations in the later cycles would then favor sequencing over amplification.
To determine the conditions that would support combined PCR and sequencing, we prepared reactions using genomic DNA as the template and BigDye v 2.0 or 3.0 (containing buffer, AmpliTaq DNA polymerase FS, dNTPs, and fluorescently labeled ddNTPs). We then added forward and reverse primers (specific to exon 10 of the human Factor V gene) at ratios of 1:1, 1:2, 1:5, 1:10, 1:20, 1:50, and 1:100. Additional dNTPs were added to the reactions at final supplemental concentrations of 0, 12.5, or 125 µM. Reactions were subjected to standard cycle sequencing and capillary electrophoresis (see Materials and Methods). Several of the conditions tested supported combined amplification and sequencing with various signal-to-noise ratios. We determined that the primer ratio of 5:1 (500 nM primer 1 and 100 nM primer 2) was optimal, although ratios of 2:1 and 10:1 also yielded sequencing products. Although the amount of supplemental dNTPs only slightly influenced the signal-to-noise ratio, we considered 125 µM supplemental dNTPs to be optimal. Combined PCR and sequencing results obtained with the optimized conditions for a Factor V wild-type and Leiden heterozygote are shown in Fig. 1
. The AmpliSeq results correlated with the results of the Mnl1 restriction enzyme digestion assay (data not shown) (6).
View larger version (17K):
[in this window]
[in a new window]
|
Figure 1. Sequencing results generated by AmpliSeq reactions.
Shown are a Factor V wild-type sequence and a sequence containing a Leiden heterozygous mutation. The red arrow indicates the Factor V Leiden mutation site.
|
|
application of standard conditions to additional gene targets
To test whether this methodology has broad applicability, we applied the optimized conditions to AmpliSeq reactions for several different gene targets. Different sources of DNA were used as template materials, and the reactions were designed to produce a variety of different lengths of PCR/sequencing products. All reactions were prepared using the standard conditions determined by optimization using Factor V as the target (see above), and the results are summarized in Table 1
. In addition to Factor V germline sequencing, we also performed AmpliSeq reactions for germline BRCA2 sequencing. We sequenced 
500 bases of exon 11 of the BRCA2 gene that contain the common 6174delT mutation site (see the Data Supplement that accompanies the online version of this article athttp://www.clinchem.org/content/vol51/issue1/). Although 500 bases is the largest PCR and sequencing product analyzed to date, distal sequencing is generally of adequate intensity with low background, and we anticipate that AmpliSeq will be capable of longer reads. The formal upper limit of sequence length possible is currently under investigation.
In addition to germline sequencing, we used this strategy to sequence somatic mutations that occur during tumorigenesis. These included mutations such as FLT3 internal tandem duplication mutations, which occur in 20% of acute myeloid leukemias, and point mutations in MADH4, which commonly occur in pancreatic cancer. For MADH4 analysis, we used DNA from xenografted tumors as the template. Because the quantity of xenografted tumor DNA was more limited than that in the previous genomic DNA samples, we used substantially less of it in reactions (50 ng) and increased the number of cycles to 45 (see Table 1
). We identified a nonsense mutation in MADH4, which was confirmed by traditional PCR followed by cycle sequencing reactions (see the online Data Supplement).
An additional application we have developed is rapid bacterial species identification by 16S ribosomal DNA sequencing. We prepared combined PCR and cycle sequencing reactions according to the procedure described in the Materials and Methods, using crude bacterial lysates as the template. The reactions were designed to sequence 400 bases of 16S ribosomal DNA. The portions of the sequencing products shown in Fig. 2
highlight differences in 16S sequence among the four bacterial species tested.
View larger version (29K):
[in this window]
[in a new window]
|
Figure 2. AmpliSeq results for 16S DNA from four bacterial species.
Each species is noted. Differences in 16S sequence that aid in species differentiation are highlighted with red arrows.
|
|
kinetics of AMPLISEQ reactions
To better understand the kinetics of the AmpliSeq reaction, we compared the relative amounts of amplification and sequencing products generated during cycling. Using the optimized standard conditions, we ran duplicate AmpliSeq reactions specific for exons 11 and 12 of FLT3 for various numbers of cycles. For each time point, one reaction was analyzed by polyacrylamide gel electrophoresis to determine whether a PCR product was present, and the second reaction was purified and separated by capillary electrophoresis to identify whether readable sequencing products were present. The results are shown in Fig. 3
. Results of polyacrylamide gel electrophoresis of samples taken after 5, 10, 15, 20, 25, 30, and 35 cycles are shown in Fig. 3A
. The corresponding sequencing electropherograms for cycles 20, 25, 30, and 35 are shown in Fig. 3B
. The 330-bp PCR product was first detected by polyacrylamide gel electrophoresis at 20 cycles, but at that point essentially no sequencing products were detected. At 25 cycles, the amount of PCR product was increased, and sequencing products could be identified, although at a relatively low fluorescence intensity, producing suboptimal sequencing. The background fluorescence signals observed did not appear to be specific signals from the reverse primer, but rather random noise. By 30 cycles, sufficient sequencing products had been generated to yield clean, readable sequence data. The fluorescence intensity of the sequencing peaks observed after 35 cycles was within the range of intensities generated during traditional PCR followed by cycle sequencing, which can be variable depending on factors such as the total volume of the PCR reaction (from 5 to 50 µL), whether the product was concentrated during the purification step that precedes cycle sequencing, and the actual volume of PCR product added to the cycle sequencing reaction. The results of this experiment support our experimental hypothesis that under our standard conditions, amplification dominates earlier cycles and sequencing follows in subsequent cycles.
View larger version (39K):
[in this window]
[in a new window]
|
Figure 3. Comparison of the production by AmpliSeq of PCR products and sequencing products after various cycle numbers.
Duplicate AmpliSeq reactions specific for exons 11 and 12 of FLT3 were prepared. For each duplicate, one reaction was subjected to polyacrylamide gel electrophoresis to detect PCR amplification products, and the other sample was subjected to capillary electrophoresis to detect sequencing products. (A), polyacrylamide gel (10%) of AmpliSeq products after various cycle numbers. The wild-type FLT3 PCR product is 330 bp in length. (B), sequencing data after various cycle numbers. RFU, relative fluorescence units. AmpliSeq sequencing data after 5, 10, and 15 cycles were similar to the data generated at 20 cycles (data not shown).
|
|
|
Discussion
|
|---|
In this report we have demonstrated a novel method to effectively combine PCR and cycle sequencing into one reaction using a single DNA polymerase, standard oligonucleotide primers, and standard sequencing reagents. The two critical elements of AmpliSeq reactions are the unequal molar ratio of the forward and reverse primers and the increased ratio of dNTPs to ddNTPs. The concentrations and ratio of the two primers must be able to support PCR during early cycles but yield detectable sequencing products from only the primer in excess during later cycles. Likewise, the dNTP/ddNTP ratio must allow full-length extension (PCR amplification) during early cycles, but termination (sequencing) in later cycles. We optimized the experimental conditions for one template (Factor V) and then successfully applied these standard conditions to four additional targets without need for additional optimization. An independent laboratory has tested these conditions on an additional 10 targets in the size range 250–530 bp and successfully sequenced 8 (Dr. Sandra Spurgeon, Applied Biosystems, personal communication). The reason(s) for the failure of two targets is currently unclear. From these data it would appear that AmpliSeq reactions using these conditions are fairly robust. In addition, we submitted samples that consisted of genomic DNA with added dNTPs and primers (to final concentrations described for standard AmpliSeq reactions) to a DNA sequencing core facility that used 35 cycles for sequencing. The resulting sequence data were comparable to those obtained when PCR products are submitted for standard cycle sequencing.
The AmpliSeq protocol has features in common with asymmetric PCR, in particular the use of unequal molar ratios of primers. Although asymmetric PCR was originally developed to generate single-stranded DNA that could be used for DNA sequencing reactions, it has not been widely used because the reaction can be difficult to optimize (7)(8). A recent publication defines conditions for primer design for asymmetric amplifications that overcome the optimization difficulties (9). Although the method described, Linear-After-The-Exponential (LATE)-PCR, is a real-time, quantitative method, their findings regarding primer design may to be applicable to AmpliSeq reactions because both use asymmetric amplification. The authors found that the efficiency of asymmetric PCR reactions was best when the concentration-adjusted melting temperature of the limiting primer was greater than that of the primer in excess (9). Interestingly, only two of the five primer sets used in the present study conform to the conditions of primer design described for LATE-PCR. Our success with these primer pairs may, at least in part, be a result of the fairly permissive annealing temperature used for AmpliSeq reactions (50 °C). The two primer sets that failed to yield AmpliSeq sequencing were tested under only a single set of conditions. Optimization of the primers in terms of melting temperature or primer ratio (perhaps reversing which primer is present in excess or in limited amounts) may lead to successful AmpliSeq conditions.
Optimization of AmpliSeq reactions may be necessary to improve either the quantity of amplification and/or the quality (specificity) or the reaction. The data demonstrating the kinetics of AmpliSeq reactions (Fig. 3
) and our experience suggest that insufficient amplification (low yield of PCR product) can lead to failed AmpliSeq reactions. We have not yet quantified the magnitude of amplification necessary for AmpliSeq reactions. Ensuring high specificity of the reaction is an additional consideration when designing AmpliSeq reactions. AmpliSeq reactions using primers that are not specific or are designed to generate multiple PCR products will yield ambiguous sequencing. In our experience, primers that yield consistently robust PCR amplification of a single product have worked well for AmpliSeq reactions.
The similarities between PCR amplification and cycle sequencing reactions and the potential benefits in terms of cost and time have led others to develop methods to combine the two reactions into a single reaction. One previously described method, coupled amplification and sequencing, used a two-stage protocol in which samples were manually partitioned after several cycles of amplification so that labeled primers and ddNTPs could be added to eight individual reactions (10)(11). A second previously described method, termed DEXAS (direct exponential amplification and sequencing), was based on dye-primer sequencing chemistry (12)(13). DEXAS required four separate sequencing reactions and used two separate DNA polymerases; one for amplification (incorporation of dNTPs) and one for sequencing (incorporation of ddNTPs) (14). Finally, we have recently described a combined PCR and sequencing method that, although useful for some applications, requires the use of a specially modified primer (containing an abasic region and a long thymidine tail) and is somewhat limited in the length of sequencing data that can be generated (15).
In contrast to previously described methods, AmpliSeq reactions use standard oligonucleotide primers, a single polymerase, and are carried out in a single reaction tube without the need for additional manipulations during cycling. AmpliSeq reactions can be performed with commonly available reagents and instrumentation; thus, any laboratory currently performing sequencing reactions should be able to use this methodology. We have demonstrated a wide range of applications using AmpliSeq, including the identification of germline variants, identification of somatic mutations in tumor tissue, and bacterial speciation. The advantages of AmpliSeq include decreased turnaround time, fewer sample manipulations, and decreased cost compared with traditional PCR amplification followed by cycle sequencing.
|
Acknowledgments
|
|---|
We appreciate helpful discussions with, and bacterial samples provided by, Dr. Jim Dick and Joan Valentine (Department of Pathology, Johns Hopkins Medical Institutions, Baltimore, MD). We also acknowledge helpful discussions and the technical assistance of Dr. Antony Parker, Jodi Franklin, Scott Donover, and Major David Kulesh. We would also like to thank Dr. Scott Kern (Department of Oncology, Johns Hopkins Medical Institutions, Baltimore, MD) for kindly providing MADH4 primers and xenografted tumor DNA. Dr. Eshleman is the recipient of a grant from the NIH (RO1-CA-81439).
|
References
|
|---|
- Sanger F, Nicklen S, Coulson AR. DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci U S A 1977;74:5463-5467.[Abstract/Free Full Text]
- Cathcart R. Advances in automated DNA sequencing. Nature 1990;347:310.[CrossRef][Medline]
[Order article via Infotrieve]
- Prober JM, Trainor GL, Dam RJ, Hobbs FW, Robertson CW, Zagursky RJ, et al. A system for rapid DNA sequencing with fluorescent chain-terminating dideoxynucleotides. Science 1987;238:336-341.[Abstract/Free Full Text]
- Murray V. Improved double-stranded DNA sequencing using the linear polymerase chain reaction. Nucleic Acids Res 1989;17:8889.[Free Full Text]
- Tabor S, Richardson CC. A single residue in DNA polymerases of the Escherichia coli DNA polymerase I family is critical for distinguishing between deoxy- and dideoxyribonucleotides. Proc Natl Acad Sci U S A 1995;92:6339-6343.[Abstract/Free Full Text]
- Bertina RM, Koeleman BP, Koster T, Rosendaal FR, Dirven RJ, de Ronde H, et al. Mutation in blood coagulation factor V associated with resistance to activated protein C. Nature 1994;369:64-67.[CrossRef][Medline]
[Order article via Infotrieve]
- Innis MA, Myambo KB, Gelfand DH, Brow MA. DNA sequencing with Thermus aquaticus DNA polymerase and direct sequencing of polymerase chain reaction-amplified DNA. Proc Natl Acad Sci U S A 1988;85:9436-9440.[Abstract/Free Full Text]
- Gyllensten UB, Erlich HA. Generation of single-stranded DNA by the polymerase chain reaction and its application to direct sequencing of the HLA-DQA locus. Proc Natl Acad Sci U S A 1988;85:7652-7656.[Abstract/Free Full Text]
- Sanchez JA, Pierce KE, Rice JE, Wangh LJ. Linear-after-the-exponential (LATE)-PCR: an advanced method of asymmetric PCR and its uses in quantitative real-time analysis. Proc Natl Acad Sci U S A 2004;101:1933-1938.[Abstract/Free Full Text]
- Ruano G, Kidd KK. Coupled amplification and sequencing of genomic DNA. Proc Natl Acad Sci U S A 1991;88:2815-2819.[Abstract/Free Full Text]
- Reynolds TR, Uliana SR, Floeter-Winter LM, Buck GA. Optimization of coupled PCR amplification and cycle sequencing of cloned and genomic DNA. Biotechniques 1993;15:462-464, 466–7.[ISI][Medline]
[Order article via Infotrieve]
- Kilger C, Paabo S. Direct exponential amplification and sequencing (DEXAS) of genomic DNA. Biol Chem 1997;378:99-105.[ISI][Medline]
[Order article via Infotrieve]
- Kilger C, Paabo S. Direct DNA sequence determination from total genomic DNA. Nucleic Acids Res 1997;25:2032-2034.[Abstract/Free Full Text]
- Reeve MA, Fuller CW. A novel thermostable polymerase for DNA sequencing. Nature 1995;376:796-797.[CrossRef][Medline]
[Order article via Infotrieve]
- Murphy KM, Eshleman JR. Simultaneous sequencing of multiple polymerase chain reaction products and combined polymerase chain reaction with cycle sequencing in single reactions. Am J Pathol 2002;161:27-33.[Abstract/Free Full Text]
Top
Abstract
Introduction
