E.
Papadopoulou
a,
N.
Gale
b,
S. A.
Goodchild
c,
D. W.
Cleary
c,
S. A.
Weller
c,
T.
Brown
d and
P. N.
Bartlett
*a
aChemistry, University of Southampton, Highfield, Southampton SO17 1BJ, UK. E-mail: pnb@soton.ac.uk
bATDBio Ltd, Chemistry, University of Southampton, Highfield, Southampton SO17 1BJ, UK
cDSTL, Wiltshire SP4 0JQ, Salisbury, Porton Down, UK
dDepartment of Chemistry, University of Oxford, Chemistry Research Laboratory, Oxford OX1 3TA, UK
First published on 7th January 2015
Strain discrimination within genetically highly similar bacteria is critical for epidemiological studies and forensic applications. An electrochemically driven melting curve analysis monitored by SERS has been utilised to reliably discriminate strains of the bacterial pathogen Yersinia pestis, the causative agent of plague. DNA amplicons containing Variable Number Tandem Repeats (VNTRs) were generated from three strains of Y. pestis: CO92, Harbin 35 and Kim. These amplicons contained a 10 base pair VNTR repeated 6, 5, and 4 times in CO92, Harbin 35 and Kim respectively. The assay also included a blocker oligonucleotide comprising 3 repeats of the 10-mer VNTR sequence. The use of the blocker reduced the effective length of the target sequence available to bind to the surface bound probe and significantly improved the sensitivity of the discrimination. The results were consistent during three replicates that were carried out on different days, using different batches of PCR product and different SERS sphere segment void (SSV) substrate. This methodology which combines low cost, speed and sensitivity is a promising alternative to the time consuming current electrophoretic methods.
The most widely used methods for VNTR analysis are based on electrophoretic separation that measures the size in different alleles such as Multiple Locus Variable number repeat Analysis (MLVA) which involves PCR amplification of the selected VNTR loci followed by capillary electrophoresis.13–16 This approach, despite being highly discriminating, is time consuming and requires specialised equipment and skilled technicians. Simpler methods utilise melting curve analysis, which discriminates the DNA fragments on the basis of the differences in their melting temperature. HyBeacons®, which are fluorescent oligonucleotide probes that emit greater amounts of fluorescence after hybridization, have been used for a fluorescence-based differentiation and identification of short tandem repeats (repeat units 3–9 bp).17,18
We have developed an electrochemically induced DNA denaturation approach that is monitored by SERS. Similar to fluorescence-based melting curve analysis, we can construct potential melting curves by measuring the SERS intensity as a function of applied potential. In this approach, the DNA target is labelled with a Raman fluorophore and then hybridized to a DNA probe that is immobilised on an ordered sphere segment void (SSV) gold surface via three dithiol linkers. It has been previously shown that SSV surfaces, which are made by electrodeposition of gold around closely packed monolayers of spheres, can provide large and reproducible SERS enhancements.19–22 Upon application of an increasing negative potential on the SSV surface, the dsDNA denatures and the denaturation can be monitored by following the decrease in intensity of the Raman signal from the labelled target DNA as it diffuses away from the surface. This SERS-based methodology has been successfully used for the discrimination of mutations in the cystic fibrosis trans-membrane regulator gene23,24 and short tandem repeats in the D16S539 locus of the human genome.25 In those cases the target DNA was amplified using synthetic templates.
We have recently extended the application of this technique for exploitation of molecular assays for discriminating between DNA from closely related microorganisms.30 The target organism chosen for study was Yersinia pestis, which is the aetiological agent of the systemic, invasive disease known as plague; an ancient disease of substantial historic importance resulting in the deaths of millions worldwide.26–29Y. pestis also remains of interest to the defence and security community due to the potential for use as a biological weapon. The rapid discrimination of different species is therefore important for epidemiological studies and for forensic applications of importance to national security.2 This organism represents a particular challenge for the generation of simple molecular assays as it shares a high degree of similarity to other Yersinia spp. Therefore assays which discriminate within the Yersinia genus such as Single Nucleotide Polymorphism (SNP) or VNTR provide highly resolved lineage information for both evolutionary and epidemiological analysis.
In a previous study we targeted discrimination of Y. pestis from the closely related Y. pseudotuberculosis. This was achieved on the basis of species specific SNPs present within long (>250 bp) DNA PCR amplicons.30 In this study, we further exemplify the strength of this technique in molecular typing of bacterial species to achieve strain differentiation for Y. pestis. We show the differentiation of three strains of Y. pestis by VNTR analysis based on different numbers of repeats of a VNTR sequence present within these strains. Their VNTRs can be differentiated on the basis of the differences in their melting potential. Similarly to our previous assays, the PCR amplicons were used unpurified. An outline of the strategy is illustrated in Fig. 1. Specifically, genomic DNA was extracted and amplified from the polymorphic bacterial strains CO92, Harbin 35 and Kim that had a 10 bp repeat unit, repeated 6, 5 and 4 times respectively, for each strain. To improve the discrimination efficiency of the three bacterial strains a blocker oligonucleotide was employed to reduce the effective length of the target sequence available to bind to the surface bound probe. Hence, an increase in the differences in melting potential between the amplicons was achieved. To our knowledge this is the only SERS-based sensor that can be applied in VNTR genotyping.
Fig. 1 Schematic flow diagram of the SERS detection assay illustrating each operational step involved. First, the DNA is extracted from bacterial cultures and the amplified DNA product is used as a template to generate single stranded labelled DNA during asymmetric PCR. The step (ii) could be omitted for future assays as it is possible to perform asymmetric PCR on genomic DNA directly.31 The labelled target is then hybridized to surface bound probes in the presence of a blocker oligonucleotide. Following electrochemical analysis monitored by SERS the pathogenic strains can be discriminated on the basis of the differences in their melting potential. |
(1) |
Assay combined with probe-1 | |
CO92 | CCTATACCGCTACGATCAGCCTCTATCGCCAATCACTATCATCAACCATCAAC-GACTATCACC-ACTATCACC-ACTATCACC-GACTATCACC-GACTATCACC-GACTATCACC--CCCCAAATATCA |
Harbin35 | CCGTTACCGCTACGATCAGCCTCTATCGCCAATCACTATCATCAACCATCAAC-GACTATCACC-ACTATCACC-ACTATCACC-GACTATCACC-GACTATCACC-CCA |
Kim | CTCGTTACCGCTACGATCAGCCTCTATCGCCAATCACTATCATCAACCATCAAC-GACTATCACC-ACTATCACC-ACTATCACC-GACTATCACC-CCCCAAATATCAGA |
Probe-1 | GGTGATAGTC-GGTGATAGTC-GGTGATAGTC-GGTGATAGT-GGTGATAGT-GGTGATAGTC-HXXX |
Assay combined with probe-2 | |
CO92 | TGATATTTGGGGCTACCAGTCGCAGGTAGTT-GGTGATAGTC-GGTGATAGTC-GGTGATAGTC-GGTGATAGT-GGTGATAGT-GGTGATAGTC-GTTGATGGTTGAGGCTGATCGTAGCGGTATAGG |
Harbin35 | TGGCTACCAGTCGCAGGTAGTT-GGTGATAGTC-GGTGATAGTC-GGTGATAGT-GGTGATAGT-GGTGATAGTC-GTTGATGGTTGAGGCTGATCGTAGCGGTAACGG |
Kim | TCTGATATTTGGGGCTACCAGTCGCAGGTAGTT-GGTGATAGTC-GGTGATAGT-GGTGATAGT-GGTGATAGTC-GTTGATGGTTGAGGCTGATCGTAGCGGTAACGAG |
Probe-2 | GACTATCACC-GACTATCACC-GACTATCACC-AACTAC-HXXX |
Blocker | TATCGCCAATCACTATCATCAACCATC-AACGACTATCACC-ACTATCACC-ACTATCACC |
Amplified DNA was used as a template for asymmetric PCR to generate ssDNA amplicons carrying a SERS active reporter chromophore (Texas Red) at the 5′ end. It is worth mentioning that asymmetric PCR can be performed on genomic DNA directly for future assays,31 however primary template amplicon is produced here for health and safety reasons. The labelled target DNA was hybridized to the surface bound DNA probe directly from the PCR buffer.
The initial hybridization probe (probe-1) was designed to comprise of 6 repeat units (Fig. 2(a)), where two of the units differ by one base (T instead of C) in order to be perfectly complementary to the targets. It was assumed that the difference of two nucleotides within the repeat units would be adequate to ensure specific hybridization of the PCR amplicons to the probes and prevent probe slippage along the tandem repeat targets. The probe was immobilised on the SSV Au surfaces via three dithiol linkers on the 3′ end to ensure that it is strongly bound to the surface. The surface was then passivated by treatment with mercaptohexanol to prevent the non-specific adsorption of DNA and allow the DNA probes to adopt perpendicular orientation on the surface.33 Target DNA labelled with Texas Red was hybridized with the immobilised probe. In this orientation the fluorophore would be situated close to the surface while having small overhanging sequences (∼20 bases) at both ends. A schematic of the design is shown in Fig. 2(a). Hybridization of probe-1 to the PCR amplicons yields a 60, 50 and 40 bp duplex for the CO92, Harbin 35 and Kim strain respectively.
Fig. 2 (a) Schematic representation of the PCR amplicons hybridised to the immobilised on the SSV substrate probe-1. The darker lines illustrate the position of the one base difference (a G-C bp is replaced by an A-T bp) in the two 10 bp repeat units compared to the rest of the units. (b) SERS of Texas Red for the PCR fragment from CO92 bacterial strain hybridised to probe-1. The spectra have been reported at different applied potentials vs. Ag/AgCl as shown in the figure. Spectra were acquired with 633 nm excitation laser and have been background subtracted and normalised with respect to the laser power and accumulation time. (c) E-melting curves from CO92 (6 repeat units), Harbin 35 (5 repeat units), and Kim (4 repeat units) hybridized to probe-1. Each set of data was fitted to eqn (1). The potential was swept at a scan rate of 0.7 mV s−1 in 10 mM Tris buffer (pH 7.2) containing 1 M NaCl. |
Good signal to noise spectra of the Texas Red labelled DNA was recorded after hybridization of the surface bound probes. Application of cathodic potentials from −0.4 V to −1.4 V vs. Ag/AgCl changed the intensity of the Texas Red Raman bands that initially increased and then decreased sharply (see ESI, Fig. S2†). The initial increase in intensity is reversible and is most likely due to a change in orientation of the label with respect to the SSV surface.25 The subsequent decrease in intensity beyond −0.7 V is due to the electrochemically induced denaturation of the DNA duplexes that allows the labelled DNA target to diffuse away from the surface (Fig. 2(b)) The raw spectra, without baseline correction, are shown in the ESI, Fig. S3.† The intensity of the most intense Texas Red band (1505 cm−1) was plotted against the applied potential for each PCR amplicon and the data were fitted to eqn (1) in order to define the melting potential (Em) (Fig. 2(c)). The Em value is the potential at which the intensity of the band at 1505 cm−1 is half the intensity value at the plateau of the sigmoidal curve. In these experiments the difference between the Em values was less than 10 mV, which is inadequate to secure reliable discrimination of the three bacterial strains. In addition, taking into account the 95% confidence intervals (Fig. S4,† the width of the confidence intervals is proportional to the standard error of the predicted absolute signal intensity (I) value), it is evident that the discrimination is not reliable.
It was considered likely that the weak discrimination observed using this probe design was due to a poorly controlled binding of the VNTR sequences along the immobilised probe. It is possible that the difference of two single nucleotides within the VNTR repeat units was not adequate to stabilise the position of the duplex and prevent probe slippage. The use of a short additional anchor sequence was thus considered. However, the use of an anchor sequence with the current design would increase the length of the duplexes, thus decreasing the difference in melting potential (ΔEm) between the three PCR fragments and rendering the discrimination more difficult.
Instead, a new design using a combination of both anchor sequence and a blocker oligonucleotide was employed to reduce the effective length of the target sequence available to hybridize to the surface bound probe. Thus, the use of a shorter probe was allowed onto which an anchor sequence was added to secure precise hybridization of the probe to the PCR amplicons. The use of a blocker oligonucleotide has been evaluated in assays utilising HyBeacon® probes (fluorescence probes) and fluorescence melting analysis. It has been shown to permit the reliable differentiation of high numbers of repeats for forensic loci by effectively increasing the difference in melting temperature (Tm) between STRs of similar length.17,34
Three sets of data were collected for the electrochemically driven melting of each amplicon. Each data set was collected on different days, using different batches of PCR products and SSV substrates. The three sets of data for each amplicon were analysed as described previously30 in order to produce a single common melting curve for each set of replicate experiments together with the corresponding 95% confidence interval. Briefly, each individual data set was normalised by its Imax value, which is the intensity value at the plateau of the sigmoidal curve. The normalised sets of data for each amplicon were fitted to eqn (1) using a weighted global fit, where each data set was weighted in inverse proportion of its particular value of Imax2. The E-melting curves, for the three amplicons, along with the 95% confidence intervals are shown in Fig. 3(c). It is clear; taking into account the confidence intervals, that reliable discrimination of the three bacterial strains has been achieved. The use of a blocker oligonucleotide resulted in an increased difference in the Em between the amplicons (Table 2) that improved the sensitivity of the assay significantly.
CO92 | Harbin 35 | Kim | |
---|---|---|---|
E m/V | −0.974 ± 0.002 | −0.944 ± 0.003 | −0.919 ± 0.003 |
dE/V | 0.039 ± 0.003 | 0.040 ± 0.002 | 0.033 ± 0.002 |
Utilising a SERS-based electrochemically driven melting analysis for the discrimination of tandem repeats is a promising alternative over the conventional methods that are based on electrophoretic separation that measures the size in different alleles. This methodology does not require expensive equipment and has the potential to be incorporated into portable technology. More importantly, SERS has the capability to provide high degree of multiplexing due to the sharper vibrational bands provided from the Raman labels which can be synthetically attached to the DNA.35,36 In addition the SERS enhancement is distance dependent, therefore only the SERS signal for the label attached to the DNA when is in close proximity to the surface can be detected. This is an advantage over fluorescence melting since we can perform the discrimination using the PCR buffer and the only melting peaks detected originate from the probe melting, which simplifies sample analysis. In more detail, the blocker oligonucleotide is bound to an unlabelled part of the amplicon, therefore its melting peak is not detected by SERS. In addition the labelled, full length duplex amplicons that exist in the PCR solution cannot be adsorbed on the surface due to passivation of the surface with mercaptohexanol after immobilisation of the thiolated probes, therefore their melting curves are not detected. Finally the method is highly sensitive, here we show discrimination of VNTRs from 4 μl starting solutions of 55–65 pM for each amplicon in a total time of ∼100 min for asymmetric PCR amplification, hybridization to the surface bound probe and electrochemically driven melting.
Footnote |
† Electronic supplementary information (ESI) available: Electrochemical melting curves along with 95% confidence intervals for the three PCR amplicons hybridized to probe-1, fluorescence melting curves for amplicons hybridized to probe-2 in the presence of a blocker oligonucleotide, structure of the thiol linker. See DOI: 10.1039/c4sc03084b |
This journal is © The Royal Society of Chemistry 2015 |