Isothermal DNA Amplification with Gold Nanosphere-Based Visual Colorimetric Readout for Herpes Simplex Virus Detection

Eric Tan, Barbara Erwin, Shale Dames, Karl V. Voelkerding, Angelika Niemz

Clinical Chemistry · 2007 · 58 citations · 13 references

DOIFull text

Open access

Concepts

Abstract

There is a need for rapid nucleic acid amplification and detection technologies suitable for point-of-care settings. PCR-based molecular diagnostics, portable thermocyclers, and rapid microfluidic techniques are emerging technologies (1), but the use of isothermal nucleic acid amplification would make less demanding instrumentation feasible. Most isothermal methods reported to date(2)(3)(4)(5)(6)(7) require reaction times >30 min. In this study, we describe very rapid isothermal DNA amplification through EXPAR (exponential amplification reaction)(8)(9) coupled with visual colorimetric detection facilitated by aggregation of DNA-functionalized gold nanospheres(10)(11)(12) applied to a genomic sequence derived from herpes simplex virus (HSV)1. Rapid detection of HSV infections is important in HIV-positive and immunocompromised individuals (13)(14), pregnant women, and newborns(15). PCR-based assays are superior to viral cultures and immunoassays for the diagnosis of HSV infections from cerebrospinal fluid(16) and genital, oral, and topical swabs(17)(18)(19)(20). New rapid, low-cost molecular diagnostic tools can enable broad-based testing of at-risk populations, providing effective patient treatment and preventing further disease spread. EXPAR (8)(9) isothermally amplifies a short oligonucleotide trigger 106- to 109-fold in <10 min. EXPAR occurs at 55 °C, permitting activity and stability of the polymerase and nicking endonuclease involved in the reaction(9). Trigger X primes an amplification template containing 2 complimentary X′ sequences and enables generation of the nicking enzyme recognition and cleavage site (see Fig. S1 in the Data Supplement that accompanies the online version of this Abstracts of Oak Ridge Posters at http://www.clinchem.org/content/vol53/issue11). Trigger extension and single-strand nicking create another trigger oligonucleotide, which dissociates from the amplification template. Through repeating cycles of elongating and nicking, the trigger is linearly amplified. Newly formed triggers prime other amplification templates, causing exponential amplification. The progression of EXPAR can be monitored in real time using SYBR Green. As an indirect measure for trigger amplification, the sigmoidal increase in fluorescence intensity reflects conversion of template into the partially double-stranded trigger-producing form, (see Fig. S2 in the online Data Supplement). The time at which each amplification curve reaches its inflection point is linearly correlated to the logarithm of the starting trigger concentration (see Fig. S3 in the online Data Supplement), similar to the correlation between cycle threshold and target DNA concentration in real-time PCR. Despite the positive attributes of EXPAR, the detection limit of this method is currently high because of nonspecific background amplification. Investigations are ongoing to determine the nature of this amplification observed in the absence of trigger, and our data suggest that unconventional DNA synthesis may be involved (unpublished data). The timing of background amplification depends on the template sequence and mastermix composition. For suitable EXPAR templates, trigger concentrations of 10 fmol/L (1.8 × 105 copies in a 30-μL reaction volume) can be reproducibly differentiated from background, using as a metric a ≥10% difference between the inflection points of the real-time fluorescence curves for the trigger-containing (positive, P) and no trigger–containing (negative, N) samples [(N-P)/N in %]. This value equals approximately 60-s absolute difference in the inflection points of the real-time fluorescence curves. The timing of background amplification has a typical intraassay imprecision (CV) value of approximately 2%, and an interassay CV of approximately 10%. This interassay variability does not negatively affect the relative separation of trigger and no-trigger–containing samples within each experiment. We are optimizing EXPAR, and have occasionally been able to differentiate much lower trigger concentrations from background (1 amol/L, 18 copies in 30 μL; see Fig. S4 in the online Data Supplement). Although such low limits of detection are not routinely attainable, we are optimistic that through systematic assay optimization it will be possible to perform EXPAR in a robust manner with PCR-like sensitivity. To facilitate visual detection, we have developed a 2-stage EXPAR amplification reaction, coupled to aggregation of DNA-functionalized gold nanospheres through a bridging reporter sequence (9). DNA nanosphere aggregation causes a red-to-blue color change attributable to a shift in plasmon resonance(10)(11)(12). We previously reported that 2-stage EXPAR coupled to DNA-nanosphere aggregation enables detection of 100 fmol/L trigger oligonucleotide in <10 min. Through continuing optimization, we can now reproducibly detect 10 fmol/L trigger (60 000 copies in 10 μL) with amplification times of 3.5 or 4 min. We have occasionally achieved low copy number detection limits (1 amol/L, 6 copies in 10 μL; see Fig. S5 in the online Data Supplement), but variability in assay performance exists at starting trigger concentrations <10 fmol/L. To be useful for clinical diagnostics, this assay must be coupled with trigger generation from a genomic target DNA sequence. One approach for trigger generation, called fingerprinting, is based on adjacent nicking enzyme recognition sites within genomic DNA (Fig. 1A1 ). When these sites are oriented head-to-head, both strands are cut by the nicking enzyme, and the genomic DNA dissociates, creating templates for 2 complementary triggers that are linearly amplified through consecutive cycles of polymerase extension and single-strand nicking, (analogous to linear amplification within EXPAR; see Fig. S1 in the online Data Supplement). Fingerprinting can be performed isothermally at the same temperature and in the same mastermix as EXPAR, without additional reagents. Differential HSV diagnosis requires distinction from other pathogens and from human genomic DNA. On the basis of in silico sequence analysis, HSV1 fingerprinting produces 14 trigger oligonucleotides without overlap with the predicted trigger sequences of the other human herpes viruses (see Table S1 in the online Data Supplement), other potentially interfering pathogens, or of human genomic DNA. In this proof-of-principle study, we targeted a 28-mer fingerprinting site within the open reading frame of the nuclear protein UL3 (Human herpesvirus 1) gene UL3 containing 2 GAGTC sites oriented head-to-head. We have performed fingerprinting with a preliminary model system, a pUC19 vector with a 273-bp HSV1-derived insert that contains the targeted UL3 fingerprinting site (Table 11 , sequence 1) and enables generation of the complementary trigger oligonucleotides HSV1-1a and HSV1-1b (Table 11 , sequences 2 and 3). Of these 2 triggers, we have selected HSV1-1a for subsequent assay design. This vector model system (termed HSV1-vector1), which can be obtained inexpensively in pure form and at high concentrations, serves as a practical intermediate step in establishing the assay. The same reaction applies in principle to trigger generation from HSV1 genomic DNA. Through negative ion mode liquid chromatography/electrospray ionization TOF mass spectrometry, we verified that HSV1-1a is generated via fingerprinting in the presence of 1 nmol/L HSV1-vector1. No product peaks were observed in the absence of the vector, a finding that corroborates the specificity of the reaction (Fig. 1B1 ). Because fingerprinting involves linear amplification only, nanomolar starting concentrations of vector are required without coupling to EXPAR. To improve this detection limit, we coupled fingerprinting with EXPAR by including the appropriate X′-X′ template in the reaction mixture (Table 11 sequence 4, Fig. 1C1 ). According to the SYBR Green fluorescence data (Fig. 1D1 ), fingerprinting coupled with single-stage EXPAR can differentiate 1 pmol/L HSV1-vector1 from nonspecific background amplification, with marginal differentiation at the 100 fmol/L level, a 103- to 104-fold lower detection limit than fingerprinting alone. We coupled trigger generation and amplification with visual colorimetric detection (Table 11 sequence 6, Fig. 1E1 ) by including the vector and both templates (X′-X′ and X′-Y′, Table 11 sequences 4 and 5) in the reaction mixture. After amplification at 55 °C, the reaction was quenched by adding a DNA nanosphere detection reagent, which contains 2 sets of gold nanospheres bearing the 5′ and 3′ probes, respectively (Table 11 , sequences 7 and 8). During incubation at room temperature for 2 min, reporter Y generated through 2-stage EXPAR hybridizes to and aggregates the DNA nanospheres. After being spotted on a C18 reverse-phase silica thin-layer chromatography plate, aggregated nanospheres yield a blue spot (positive) and monodisperse nanospheres yield a red spot (negative). We are currently able to detect 1 pmol/L HSV1-vector1 (6 × 106 copies in a 10 μL-reaction volume) after 4-min amplification at 55 °C (Fig. 1F1 ) with good intra- and interassay reproducibility and 4-min amplification time at 55 °C. These results are based on 3 different experiments conducted on 3 different days, each in duplicate. Longer amplification times produce nonspecific background as displayed by the “No Vector” negative control. The 2 additional controls labeled positive and negative signify that DNA nanospheres are stable under the reaction conditions and aggregate in the appropriate manner. In conclusion, we have demonstrated detection of an HSV1-derived genomic sequence through a simple, rapid molecular diagnostic assay that can be performed in ≤10 min and requires only a heating block and minimal consumption of reagents and supplies, characteristics that make this method suitable for point-of-care and limited resource settings. We are performing systematic assay optimization to ensure sensitivity, reproducibility, and robustness required for clinical diagnostic applications, and are investigating the cause of nonspecific background amplification. At a minimum, we are targeting a limit of detection appropriate for diagnosis of HSV from swab samples of herpetic lesions, reported to contain on average 8 × 104 virus particles/mL lysis buffer (19). The proof-of-principle results presented were obtained using a plasmid vector model system with HSV1 insert. We are in the process of establishing the reaction with DNA isolated from clinical samples, and expanding the assay to HSV2. Fingerprinting is the simplest approach for trigger generation from genomic DNA, but has limitations. We are therefore also exploring alternative probe-based trigger generation reactions that can interrogate arbitrary genomic DNA or RNA sequences. (A), scheme of the fingerprinting reaction on the basis of nicking at adjacent “GAGTC” sites oriented head-to-head within genomic DNA, resulting in the linear amplification of 2 complimentary trigger oligonucleotides. (B), negative ion mode liquid chromatography/electrospray ionization TOF mass spectra of fingerprinting reactions performed without coupling to exponential amplifications, i.e., involving a mastermix containing enzymes, deoxyribonucleotide triphosphates (dNTPs), and other general reagents plus either 1 nmol/L HSV1-Vector1 (top) with a 273-bp insert derived from HSV1 genomic DNA (Table 11 , sequence 1) or no vector (bottom), but no externally added template or trigger oligonucleotides. In the presence of 1 nmol/L HSV1-vector1 approximately 80 nmol/L of trigger HSV1-1a (Table 11 , sequence 2) are generated after 20 min [expected m/z: 1560.5 (2−) and 1040.0 (3−)]. No oligonucleotides are generated in the absence of HSV1-Vector1. (C), scheme of exponential trigger amplification through the single-stage EXPAR reaction. (D), SYBR Green-based real-time fluorescence monitoring of trigger generation via fingerprinting from HSV1-vector1 coupled with trigger amplification via EXPAR, as a function of starting vector concentration, by isothermal amplification of a mastermix containing enzymes, dNTPs, and other general reagents, in addition to various concentrations of HSV1-Vector1 as indicated, plus 100 nmol/L template X′-X′ (Table 11 , sequence 4). Nonspecific background amplification is observed for the no vector negative control. (E), scheme of trigger conversion into reporter Y through the X′-Y′ template, followed by reporter Y-induced DNA nanosphere aggregation. (F), visual detection of HSV1-vector1 through the spot test (positive result: blue, negative result: red) after isothermal generation and amplification of trigger HSV1-1a (Table 11 , sequence 2), and conversion of this trigger sequence to reporter Y (Table 11 , sequence 6). A mastermix containing enzymes, dNTPs, and other general reagents, in addition to various concentrations of HSV1-Vector1, 100 nmol/L template X′-X′, and 400 nmol/L template X′-Y′ (Table 11 , sequences 4 and 5) was incubated at 55 °C for 4 min or 6 min. The reaction was quenched by the addition of a nanosphere detection reagent containing 2 sets of Au nanospheres, respectively, functionalized with the 5′ and 3′ nanosphere probe sequences (Table 11 , sequences 7 and 8). Reporter Y-induced DNA nanosphere aggregation results in a color change that is enhanced through spotting onto a C18-reversed-phase silica plate. The No Vector negative control turns positive after 6-min amplification time, signifying the interference of nonspecific background amplification. Amplification times are faster than in (D) because of the omission of SYBR Green from the mastermix. The controls labeled “Positive” and “Negative” contain no vector, and no template and trigger oligonucleotides. The control labeled Positive contains 1 μmol/L externally added reporter Y. Fingerprinting model system. pUC19 insert corresponding to nucleotides 7–285 of HSV1 gene UL3. Italicized portion corresponds to the 28-mer fingerprinting site (nucleotides 118–145), consisting of 2 GAGTC recognition sites oriented head-to-head (underlined). The origin of the resulting trigger sequence is shown in bold. Template sequence for amplification of trigger HSV1-1a. Template sequence for conversion of trigger HSV1-1a into reporter Y. Probe sequences covalently linked to 13-nm diameter gold nanospheres. Fingerprinting model system. pUC19 insert corresponding to nucleotides 7–285 of HSV1 gene UL3. Italicized portion corresponds to the 28-mer fingerprinting site (nucleotides 118–145), consisting of 2 GAGTC recognition sites oriented head-to-head (underlined). The origin of the resulting trigger sequence is shown in bold. Template sequence for amplification of trigger HSV1-1a. Template sequence for conversion of trigger HSV1-1a into reporter Y. Probe sequences covalently linked to 13-nm diameter gold nanospheres. Grant/funding support: Funding for this project was provided by National Institutes of Health-National Institute of Allergy and Infectious Disease award 1R21AI064804, by the Keck Graduate Institute, and by ARUP Laboratories. Financial disclosures: None declared. Acknowledgments: We thank Megan Buechel and Bruce Irvine for their work in investigating the performance of EXPAR.

References

13