A primerless molecular diagnostic: phosphorothioated-terminal hairpin formation and self-priming extension (PS-THSP)
- First Online:
- Cite this article as:
- Jung, C. & Ellington, A.D. Anal Bioanal Chem (2016) 408: 8583. doi:10.1007/s00216-016-9479-y
There are various ways that priming can occur in nucleic acid amplification reactions. While most reactions rely on a primer to initiate amplification, a mechanism for DNA amplification has been developed in which hairpin sequences at the 3’ terminus of a single-stranded oligonucleotide fold on themselves to initiate priming. Unfortunately, this method is less useful for diagnostic applications because the self-folding efficiency is low and only works over a narrow range of reaction temperatures. In order to adapt this strategy for analytical applications we have developed a variant that we term phosphorothioated-terminal hairpin formation and self-priming extension (PS-THSP). In PS-THSP a phosphorothioate (PS) modification is incorporated into the DNA backbone, leading to a reduction in the thermal stability of dsDNA and increased self-folding of terminal hairpins. By optimizing the number of PS linkages that are included in the initial template, we greatly increased self-folding efficiency and the range of reaction temperatures, ultimately achieving a detection limit of 1 pM. This improved method was readily adapted to the detection of single nucleotide polymorphisms and to the detection of non-nucleic acid analytes, such as alkaline phosphatase, which was quantitatively detected at a limit of 0.05 mU/mL, approximately 10-fold better than commercial assays.
KeywordsPS-THSP Self-folding Isothermal amplification Phosphorothioate
Molecular diagnostics is at root based on simple and robust molecular amplification reactions. While the polymerase chain reaction (PCR) has been the most extensively employed approach [1, 2], isothermal amplification methods may prove to be more tractable in many applications, including for point-of-care diagnostics, since they avoid the need for precise temperature cycling and complicated devices.
Isothermal amplification methods span a range of approaches, including (but not limited to) nucleic acid sequence-based amplification (NASBA) , single primer isothermal amplification (SPIA) , isothermal chimeric primer-initiated amplification (ICAN) , strand displacement amplification (SDA) [6, 7], helicase-dependent amplification (HDA) , recombinase polymerase amplification (RPA) , isothermal target and signaling probe amplification (iTPA) , rolling circle amplification (RCA) [11, 12], loop-mediated amplification (LAMP) , and SmartAmp . In the main these methods yield sensitive amplification similar to PCR by having mechanisms for strand invasion and/or primer binding that do not require temperature cycling.
In almost all of these reactions, though, the point is generally to amplify a natural template. Another option is to create a defined template that is amenable to amplification, and to detect analytes on the basis of the presence or absence of this template. To this end, we have expanded a technique for the amplification of very short oligonucleotides (28 based pairs, bp) that have defined hairpin structures at each end, termed terminal hairpin formation and self-priming (THSP) . Like RCA and LAMP, THSP results in linear amplification and long repeated products. Unlike RCA, THSP generates a dsDNA rather than ssDNA product and importantly does not require template circularization. Unlike LAMP, which typically uses 4, 5, or 6 primers [13, 16, 17], depending on the amplicon and detection method, it requires no primers. It is perhaps one of the most simple and self-contained reactions available for molecular diagnostics, and therefore has great promise for developing point-of-care assays.
Unfortunately, the current version of THSP reaction is greatly limited because the self-folding efficiency is relatively low and the range of amplification reaction temperatures is quite narrow. We have now improved the self-folding efficiency and broadened the range of reaction temperatures by incorporating phosphorothioates (PS) into the nucleic acid substrate. Boczkowska and co-workers have carried out thermodynamic studies on the stability of duplexes formed between PS-modified ssDNA (all-Rp, all-Sp, and mixed Rp/Sp) and complementary phosphodiester (PO)-modified ssDNA . They reported that any PS modifications substantially reduced the Tm of PS-PO dsDNA, allowing refolding and self-priming over a broader range of conditions . The improved phosphorothioated-terminal hairpin formation and self-priming extension has been termed PS-THSP, and it represents the first example of applying the broad biochemical principle of phosphorothioate destabilization to strand invasion for isothermal amplification.
By developing design rules for dumbbell templates with phosphorothioate termini for efficient amplification, we have now been able to apply PS-THSP to identifying single nucleotide polymorphisms (SNPs) and even to the sensitive detection of enzymes, such as alkaline phosphatase. Most importantly, though, since PS-THSP enables the facile amplification of short oligonucleotides it encourages very simple assay development via a very short list of reagents (oligonucleotide, polymerase, nucleotides). The PS-modified DNA should also display enhanced stability against degradation by various nucleases that may be present in biological samples (see Electronic Supplementary Material (ESM) Fig. S1) [20, 21].
Materials and methods
All chemicals were of analytical grade and were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise indicated. All oligonucleotides were ordered from Integrated DNA Technology (IDT; Coralville, IA) and the sequences are summarized in Table S1. Bst 2.0 WarmStart DNA polymerase, Bst 2.0 DNA polymerase, dNTPs (10 mM each), and Taq DNA ligase were purchased from New England Biolabs (NEB; Ipswitch, MA). Alkaline phosphatase (E. coli C75) was purchased from Clontech (Mountain View, CA) and EvaGreen dye was obtained from Biotium (Hayward, CA).
Phosphorothioated-terminal hairpin formation and self-priming (PS-THSP)
All basic PS-THSP reactions for end-point detection were performed as detailed here. Reactions were prepared with 1X isothermal amplification buffer (20 mM Tris-HCl, 10 mM (NH4)2SO4, 50 mM KCl, 2 mM MgSO4, 0.1 % Tween-20, pH 8.8), dNTPs (0.25 mM each), and various concentrations of PS-THSP templates. Reactions were heated to 95 °C for 5 min and cooled at 0.1 °C/s to 37 °C for 5 min. After this, samples were incubated at 4 °C for 2 min, followed by the addition of 4 U of Bst 2.0 WarmStart DNA polymerase to produce a final volume of 20 μl (final concentration 0.2 U/μl). Each PS-THSP reaction was then initiated at a specific temperature and continued for 1 h (on a T100™ thermal cycler; Bio-Rad; Hercules, CA). To measure fluorescence, 1 μl of 20X EvaGreen dye was added to each sample, incubated for 10 min at room temperature, and then 17 μl of the reaction was loaded on a 384-well plate. Fluorescence intensities were measured with an Infinite 200 PRO (Tecan) (Ex 488 nm, Em 532 nm).
Detection of SNPs with PS-THSP
Two steps were involved: ligation and PS-THSP. For ligation, 10 μl reactions were prepared with 1X Taq DNA ligase reaction buffer (20 mM Tris-HCl, 25 mM KAc, 10 mM MgAc, 1 mM NAD, 10 mM DTT, 0.1 % Triton X-100, pH 7.6), 10 nM mixed target (different ratios of wild-type/mutant), 20 nM oligonucleotide substrates, and 10 U of Taq DNA ligase (final concentration 1 U/μl). The samples were heated at 95 °C for 5 min and cooled to 65 °C, followed by incubation at 65 °C for 20 min. To initiate the PS-THSP reaction 0.5 μl of 10 mM each dNTPs, 4 U of Bst 2.0 WarmStart DNA polymerase, and 9 μl of distilled water were added to the ligated samples and incubated at 65 °C for 1 h (final concentration of the wild-type/mutant mixed target 5 nM).
Detection of alkaline phosphatase (AP) with PS-THSP
Two steps were involved: dephosphorylation and PS-THSP. For dephosphorylation, 10 μl reactions were prepared with 1X alkaline phosphatase buffer (50 mM Tris-HCl, 1 mM MgCl2, pH 9.0) and varying concentrations of both alkaline phosphatase and template 1-4P. Reactions were incubated at 55 °C for 1 h. Following dephosphorylation, the PS-THSP reaction was initiated by adding 1 μl of 10X isothermal amplification (200 mM Tris-HCl, 100 mM (NH4)2SO4, 500 mM KCl, 20 mM MgSO4, 1 % Tween-20, pH 8.8), 0.5 μl of each 10 mM dNTP, and 8 μl of distilled water. Next, the samples were heated at 95 °C for 5 min and cooled at 0.1 °C/s to 37 °C for 5 min. Some 4 U of Bst 2.0 WarmStart DNA polymerase was added on ice to produce a final volume of 20 μl (final concentration 0.2 U/μl). PS-THSP was carried out at 60 °C for 1 h.
A 10-μL aliquot of a PS-THSP reaction was mixed with 6X agarose gel loading buffer (NEB; Ipswitch, MA) and subjected to electrophoretic analysis on a 2 % agarose gel containing ethidium bromide. The amplification products were visualized using a UV transilluminator.
Results and discussion
Scheme for phosphorothioated-terminal hairpin formation and self-priming extension (PS-THSP)
In greater detail, we anticipated that the 3’ end of the oligonucleotide (D2*) would be extended by a DNA polymerase, leading to the formation of an extended hairpin (i.e., the seeding state). In this extended hairpin the phosphorothioate-containing D1*-L1-D1 would form a duplex with its complementary strand (D1*-L1*-D1), but the D1*-L1*-D1 strand could potentially more readily dissociate and fold back on itself, leading to the next extension. This extension would read through both D2*-L2-D2 and D1-L1-D1*, thus placing D1*-L1-D1 and D1*-L1*-D1 sequences at each end of the molecule. Additional cycles of foldback and extension would result in concatemerization of the sequences D2-L2-D2* proximal to D1*-L1-D1 (with a short AA sequence in between). Over multiple cycles of extension and refolding, concatemeric dsDNA grows such that there is a repeating unit of D1*-(L1 or L1*)-D1-D2-(L2 or L2*)-D2* flanked by a D1*-L1-D1 (PS)/D1*-L1*-D1 (PO) heteroduplex at the base of the ever elongating hairpin. As with many isothermal amplification schemes, amplification of the concatemer is linear, rather than exponential.
Introduction of phosphorothioate modifications to improve THSP
Kato and co-workers previously found that the optimum temperature for the replication of THSP templates was generally quite narrow (within 1–2 °C of an optimum temperature) . This limitation can be explained by examining the individual steps of the THSP mechanism. First, the D1*-L1*-D1 strand from the extended, duplex hairpin dissociates. The dissociated single strand then folds into the D1*-L1*-D1 hairpin and extension is initiated. The competition between hybridization with the opposing strand (to form a duplex) and self-folding (to form a hairpin) relies on interactions that are very similar, and this of necessity means that there is only a narrow temperature range that distinguishes the two structures. In order to make the Tm difference between the extended hairpin (seeding state) and the folded hairpin (folded state) greater, we introduced a chemical distinction between the two strands by incorporating phosphorothioates at the 5’ end of the template, since it has been reported that the Tm of duplexes formed between PS-modified ssDNA (all-Rp, all-Sp, and mixed Rp/Sp) and complementary phosphodiester (PO)-modified ssDNA is decreased compared with PO-DNA/PO-DNA duplexes . The Tm values of PO-DNA/PO-DNA duplexes could be reduced by up to 9 °C by introducing PS modifications into one strand of a stem. Thus, if PS modifications are introduced into the upper strand (D1*-L1-D1) of the extended hairpin structure, the Tm of extended hairpin structure should be greatly reduced, while the Tm of D1*-L1*-D1 hairpin structure for self-folding will not change.
We designed THSP amplicons to test whether PS modifications would improve overall amplification. Previous work  emphasized that polymerization efficiency for THSP depended on the relative stabilities of the hairpins at both ends of the dumbbell-shaped template. The hairpin at 3’ end should be stable in order to efficiently induce an initial extension, and we therefore included a stable hairpin (D2-L2-D2*) with an 8-bp stem (50 % GC content). The hairpin at the 5’ end (D1*-L1-D1) (Table S1) should be less stable, and was therefore designed to be the same as the best AT-rich palindromic sequence previously reported .
We then modified the 5’ end (D1*-L1-D1) with different numbers of PS modifications (0, 5, 10, and 20). In order to confirm whether the PS modifications affected Tm of the extended hairpin, melting temperatures of the extended templates with different numbers of PS modifications were measured (ESM Fig. S2). As expected, the melting temperatures decreased from 70.5 °C to 65.5 °C as the number of PS modifications increased from 0 to 20. Because the 5’ end serves as a competitor for extension in every round of concatemer formation, the same initial modifications impact the overall amplification.
The effects of PS modification on amplification were further probed by gel-electrophoresis (Fig. 2b). No bands were observed for 0 and 5 phosphorothioate modifications, but concatemers were seen with 10PS of 20PS large size of bands were observed. This is consistent with amplification reported by intercalating dyes.
Quantitative measurement of PS-THSP
This quantitative relationship suggested that it might be possible to carry out real-time monitoring of PS-THSP. We tested two different intercalating dyes, EvaGreen and SYTO 82, for their ability to detect amplification products. When EvaGreen was included in the amplification reaction the end-point fluorescence intensities at various temperatures from 54 °C to 72 °C were greatly reduced (ESM Fig. S5a). In addition, the optimum temperature for amplification increased from 60 °C to 68 °C, suggesting that the intercalating dye (EvaGreen) might have increased the melting temperature of the extended hairpin and therefore decreased the efficiency of self-folding. On the other hand, the SYTO 82 dye had less of an effect on Tm and amplification efficiency (ESM Fig. S5b). Therefore, the PS-THSP reaction could be monitored in real-time in the presence of SYTO 82, as shown in ESM Fig. S5c.
Detection of SNPs using PS-THSP
As an example, we designed a split PS-THSP template and a complementary DNA ligation splint for an SNP in the nitric oxide synthase 1 adaptor protein gene (NOS1AP, rs12122048, A to G). Ligation substrate 1 has a PS-THSP domain 1 at its 5’ end and 16 complementary nucleotides (including the SNP recognition site) at its 3’ end. Ligation substrate 2 has 20 complementary residues and a phosphate group for ligation at its 5’ end and a PS-THSP domain 2 at its 3’ end. The PS-THSP domain sequences (domain 1 and domain 2) of the ligation substrates had no internal complementarity to the target-binding domains (i.e., there were no predicted alternative stable secondary structures).
We assessed the specificity and sensitivity of our PS-THSP-mediated SNP detection method by preparing mixtures of mutant and wild-type templates (mutant/wild-type mole ratios of 0:5000, 5:4995, 50:4950, 500:4500, and 5000:0; constant total nucleic acid concentration was 5 nM). The PS-THSP substrates were incubated with Taq DNA ligase [28, 29] in the presence of the different template mixtures. Ligation and PS-THSP-based amplification were only seen when the mutant template was present. The presence of PS-THSP products was confirmed using gel electrophoresis as shown in ESM Fig. S6. The fluorescence intensity increased as the mole fraction of mutant template increased (Fig. 5b). Even a small change in the ratio (from 0:5000 to 5:4995) significantly increased fluorescence levels (two-sample t test p value = 0.008). The PS-THSP/SNP detection method showed a limit of detection of 5 pM, which is about 20 times better than previously reported methods that relied on ligation and rolling circle amplification . The impact of PS modifications for SNP detection was assessed using template 1 ligation substrates with and without PS modifications. As shown in ESM Fig. S7, in the absence of PS modifications (0PS) the signal increased very little, while in the presence of PS modifications (20PS) a substantial signal increase was observed, indicating that the PS modifications improved self-folding efficiency. The ligation method for SNP detection should be completely generalizable, since new target-specific domains can be easily designed and the same high efficiency PS-THSP domain sequences can be appended to these target-specific domains. Of course, the PS-THSP sequences of the ligation substrates should be carefully designed not to have internal complementarity with the target-binding domains.
Detection of alkaline phosphatase activity using PS-THSP
To quantify AP activity, four different concentrations of AP (0, 0.05, 0.5, and 5 mU/mL) were incubated with a 10-nM sample of template 1-4P. Again, PS-THSP reaction products were confirmed by gel electrophoresis (ESM Fig. S8). The limit of detection for AP was found to be 0.05 mU/mL, about 10 times better than current commercial assays.
In summary, primerless, isothermal amplification can now be robustly adapted to molecular diagnostic assays. We have greatly improved the initial terminal hairpin formation and self-priming extension (THSP) reaction by exploiting the biophysical observation that phosphorothioates destabilize duplexes. Terminal phosphorothioates enable the repetitive refolding of self-priming amplicons over a broad range of reaction temperatures from 54 °C to 66 °C (as opposed to the need for optimization within only a few degrees for a given THSP reaction). This is the first demonstration of the use of phosphorothioate destabilization for amplification reactions, and just as temperature (PCR, LAMP) and enzymes (helicases, recombinases) have been exploited for strand invasion, phosphorothioate inclusion may become a standard means of developing new amplification and detection mechanisms. The improved PS-THSP has a limit of detection around 1 pM. By coupling PS-THSP with template-directed ligation SNPs could be readily detected. Upon coupling with 3’ end generation, PS-THSP could be used to detect enzyme activity. The latter strategy may prove to be generalizable, leading to the detection of specific nucleases that generate a given 3’ end or of other enzymes, such as esterases, that act on synthetic, blocked substrates.
This work was supported by the National Institutes of Health [5 R01 AI092839, 5 R01 GM094933]. We sincerely thank Caitlin Sanford for her editing services.
Compliance with ethical standards
Conflict of interest
The authors declare that they have no competing interests.