Simplified development of multiplex real-time PCR through master mix augmented by universal fluorogenic reporters

Multiplex real-time PCR has become a valuable tool for the discrimination and quantification of multiple nucleic acid sequences present in a sample, since the method helps save reagents and time (1,2). The most commonly used detection techniques for multiplex real-time PCR employ fluorogenic dual-labeled target-specific probes, such as hydrolysis probes (HPs) or molecular beacons (3,4). To detect different targets, each probe must be labeled with a different fluorophore—quencher combination, making the development of these probes tedious and expensive. First, fluorescence modification of probes results in high manufacturing costs (5). Second, the fluorescence properties of each target-specific probe must be tested and optimized during assay development. A compromise between high signal generation efficiencies and low fluorescence signal cross-talk between detection wavelengths must be achieved (4,6).

The disadvantages of dual-labeled, target-specific probes in real-time PCR have been circumvented by using universal sequence-dependent detection (USD) technologies (7). USD relies on fluorogenic oligonucleotides that can be used independently of the target sequence. Once the universal detection oligonucleotide is optimized with regard to the fluorescence signal generation properties, this optimized version can be used in different assays. Mediator probe (MP) PCR is an example of USD (8). Design rules have been published recently for both single and biplex MP PCR (9,10). These rules comprise the design of the label-free target-specific MPs, which are cleaved during primer extension, as well as the design of the universal reporter oligonucleotides (URs). The MP cleavage product is extended after binding to the 3′ end of the UR, which leads to a detectable fluorescence signal due to separation of the fluorophore from the quencher in the UR.

To enable simultaneous detection of more than two targets, the set of URs must be extended. Rules that apply to published UR designs for biplex MP PCR have to be adapted to URs with other fluorescent labels to span the range of available fluorescence detection channels in real-time PCR thermocyclers.

Here, we present a set of five URs with different fluorescent labels. The quality of the UR set was first verified for signal generation upon addition of one or more MPs. The set was then validated using two different multiplex real-time MP PCR assays (Figure 1). The first assay is based on a pentaplex HP PCR food panel that discriminates DNA sequences from different animal species potentially present in pâté, a speciality food product originally made from goose or duck (11). Instead of the five dual-labeled HPs of the original assay, five label-free MPs together with this UR set were used. The same adaptation was done for a second assay, a quadruplex HP PCR for the simultaneous detection of Staphylococcus aureus together with its potential methicillin-resistance genes (mecA and mecC) and a toxin gene (Panton—Valentine leucocidin, PVL) (12).

Materials and methods

Universal reporter development

A UR for multiplex MP PCR must have a high signal-to-noise ratio (SNR) upon MP addition and a high selectivity for binding to its reverse-complementary MP compared to other MPs.

To ensure high SNRs, the fluorescence background signal was reduced by placing the fluorophore and quencher moieties in close proximity with the help of a stem-loop structure (Figure 1). Unlike previously published URs, the fluorophores were linked to the nucleotide exactly complementary to the 5′ quencher-modified terminal nucleotide of each of the five URs.

For the FAM-labeled UR A, the effect of guanosine quenching was considered (13–15). Only thymidine and adenosine bases were positioned adjacent and, thus, also opposite the FAM-modified thymidine. Since reports on the quenching of fluorophores other than FAM by nucleotides are contradictory, the effect was not considered in the design of URs B–E.

To ensure binding selectivity toward only one MP, the MP hybridization sequences (MHS) of the URs were randomly generated and showed minimal sequence homology in a BLAST analysis performed with VisualOMP Version 7.8.42.0 (DNA Software, Ann Arbor, MI). A BLAST analysis was also performed for the UR sequences to ensure there was no binding of an extended MP to a UR other than its intended target.

To verify the SNR and selectivity, URs A–E were analyzed in the presence of specific and non-specific MPs according to a previously published procedure (10).

Mediator probe design

The sequences of primers and probes for food panel and methicillin-resistant Staphylococcus aureus (MRSA) panel detection are based on the literature (11,12). All oligonucleotides were purchased from biomers.net (Ulm, Germany) except for primers and probes for the HP PCR-based food detection panel, which were sourced from a commercial kit (Pentaplex Real-Time PCR AllPaté, Microsynth, Switzerland).

Label-free MPs based on the respective HP sequences were easily designed according to recently described guidelines (11,12). Briefly, the 5′-terminal fluorophore of each HP was replaced by a 17-nucleotide generic sequence (18 nucleotides in the case of MP_PVL) that is the reverse-complement to the MP hybridization sequence of one of the URs. Instead of the 3′-terminal quencher of the HP, the MP has a C3-spacer to block extension by a polymerase. All oligonucleotide sequences for MP and HP PCR are listed in Supplementary Table S1.

Use of universal reporter oligonucleotides in multiplex PCR

Food panel: Multiplex real-time MP PCR with the novel UR set was compared with reference HP PCR using mixes of DNA (Table 1) extracted from pure samples of turkey, pork, duck, chicken, and goose, or from two different meat samples purchased from a local supermarket (DNA extracted as described in Reference 11).

Table 1.  Serial dilution of target DNA for food panel mediator probe (MP) and hydrolysis probe (HP) PCR.

HP PCR was set up according to the Pentaplex Real-Time PCR AllPaté Kit manual from Microsynth using the QuantiFast Multiplex PCR +R master mix (Qiagen, Hilden, Germany). For MP PCR, primers (200 nM final concentration except for goose primer, which was 100 nM), MPs (80 nM), and universal reporter A-E (40 nM) were used with the same master mix. Five microliters of 1 of the 5 DNA target mixes from Table 1 or 5 µL of DNA extract from the 2 meat samples (1:10 dilution in DNA dilution buffer) was mixed with 20 µL HP or MP PCR reagents, respectively.

All reactions were run in four technical replicates. PCR thermocycling was performed in a Rotor-Gene Q real-time PCR thermocycler (Qiagen) with the following protocol: 5 min at 95°C for polymerase hot start, then 45 cycles of 5 s 95°C, 20 s 60°C. Readout of all five fluorescence channels was performed during the annealing step. The fluorescence readout gain for HP and UR detection was optimized automatically using the Rotor-Gene Q software to reach background fluorescence values of 5–10 RFU (relative fluorescence units) (Supplementary Material). Data analysis was based on C_q values (Rotor-Gene Q analysis software Q 2.1.0.9 with slope correction for normalized fluorescence data) determined at a fluorescence threshold that was selected with the second derivation maximum method (16). All C_q values >35 were excluded from the analysis. C_q values were plotted against corresponding input concentrations for all five target DNAs to generate standard curves using Origin 9.0 software (OriginLab Coop., Northampton, MA). For DNA quantification of the two meat samples based on standard curves generated in the reactions with the other target mixes (Table 1), Target Mix 1 was used as the inter-run calibrator.

MRSA panel: As an important performance characteristic of diagnostic assays, the analytical limits of detection (LoDs) of multiplex MP and HP PCR were analyzed using gBlock gene fragments (Integrated DNA Technologies Inc., Leuven, Belgium) (sequences listed in the Supplementary Material) diluted in DNA dilution buffer. QuantiFast Multiplex PCR +R master mix was supplemented with the nuc gene gBlock gene fragments at 100 copies per reaction, primers (final concentration of 500 nM except for the mecA primer, which was 700 nM), and HPs (200 nM) or MPs (200 nM) together with URs B–E (50 nM). One microliter of each single-target DNA or 1 µL of a mix of the mecA gene and the PVL gene gBlock gene fragment was used in 15 µL reactions at final concentrations of 400, 200, 100, 50, 10, 5, and 1 copies per reaction. All samples were run as six technical replicates. PCR thermocycling was performed in a Rotor-Gene Q with the following protocol: 5 min 95°C for polymerase hot start, then 45 cycles of 15 s 94°C, 40 s 58°C. Readout was performed as described for the food panel reactions. LoDs were determined using SPSS software version 19 probit analysis (IBM, Armonk, NY) (17) with conditions for positive amplifications as described before (9).

For verification of the assay with clinical samples, MRSA surveillance swabs were collected from anterior nares and processed at the local microbiology department in accordance with the Microbiology Procedures Quality Standards (MiQ) issued by the German Society for Hygiene and Microbiology. Conventional culture was performed on Columbia blood agar and ChromID MRSA (BioMérieux, Marcy-l'Étoile, France) as the commercially available chromogenic MRSA screening agar. For multiplex PCR, DNA was extracted using the QIAamp DNA mini kit (Qiagen) after resuspension of bacteria from the swab samples in 2 mL PBS. The kit was used according to the manual for isolation of genomic DNA from Gram-positive bacteria using lysozyme (Sigma-Aldrich, Munich, Germany) for lysis. Three microliters of DNA extracts from clinical samples were used in 15 µL reactions. The clinical samples were analyzed as four technical replicates. Data were analyzed on the basis of C_q values as described for the food panel above.

Results and discussion

Universal reporter development

Verification results are presented in the Supplementary Material. SNRs of the 5 URs (A–E) ranged from 21.4-fold to >28-fold signal increase upon MP addition in comparison to no-MP control reactions. A signal increase >10 is considered sufficient for signal generation in MP PCR. No nonspecific signal generation and no significant fluorescence cross-talk were detectable. Concerning fluorescence quenching, the UR set showed reduced background fluorescence of 81%–95% and 23%–99% compared with food panel HPs and MRSA HPs, respectively (Supplementary Material).

Universal reporter oligonucleotide set use in mediator probe PCR for detection of food and MRSA panels

For the food panel, most values of multiplex MP PCR compared with multiplex HP PCR from Microsynth are within the acceptable range according to FDA guidelines (www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm368107.pdf) as well as within the range of previously published data (11) (Table 2).

Table 2.  Performance of the food panel mediator probe (MP) and hydrolysis probe (HP) PCR.

Significant dif ferences between the two methods were observed for turkey and pork DNA detection. For the detection of turkey DNA, the MP PCR showed a lower correlation (R²) between the input DNA concentration and the C_q values analyzed as well as a higher maximum imprecision and inaccuracy compared with the HP PCR. Considering the standard curves (Supplementary Material), the relatively high imprecision of 31% and, thus, the low R² were due to a high variance of the C_q values between the 4 replicate reactions containing 16 ng turkey DNA. This variance might be primarily due to different levels of nonspecific side reactions in the different replicates.

In the case of pork DNA detection, the HP PCR performed significantly worse than the MP PCR with respect to R², the maximum imprecision, and inaccuracy. Here, the low coefficient of determination, R² of the standard curve, arose from the relatively high inaccuracy in the detection of 1 ng (inaccuracy: 21%), 4 ng (22%), and 7.3 ng (33%) DNA per reaction (see standard curve in the Supplementary Material). This inaccuracy was also the reason for the implausibly high PCR efficiency of 137%.

Apart from these differences, the novel MP PCR was found to perform essentially equally to the commercially available HP PCR.

Equal per formance of the two detection formats was also observed for the analysis of DNA extracted from meat samples. Both methods were able to detect different DNA sequences present in the different samples (Table 3). The amount of DNA of the different sequences detected is in the same range for both methods. Contradicting the vendors' declarations, significant amounts of pork were detected using both methods. It has been reported previously that the usage of relatively cheap meat such as pork is common in the manufacture of pâté (11); however, with the samples tested here, the relative amount of pork is even higher than the amount of meat the samples are declared to be manufactured from: duck, chicken, and pheasant (Sample 1) and goose (Sample 2), respectively.

Table 3.  Quantities of meat¹ in samples analyzed by food panel mediator probe (MP) and hydrolysis probe (HP) PCR.

For the detection of MRSA, the analytical LoDs of multiplex MP and HP PCR according to a probit analysis are shown in Table 4. MP PCR showed significantly lower LoDs than HP PCR (e.g., for the mecA resistance gene, the MP was 18 copies per reaction compared to 99 copies per reaction for HP). This is primarily due to a higher SNR when using the URs for signal generation compared with the dual-labeled HPs, as shown previously (9). The higher SNR is in turn a function of the 23%–99% reduced fluorescence background of URs compared with HPs (Supplementary Material).

Table 4.  Limit of detection (LoD)¹ of multiplex quadruplex methicillin-resistant Staphylococcus aureus (MRSA) mediator probe (MP) and hydrolysis probe (HP) PCR.

In analyses of DNA extracted from patient swab samples, the HP PCR failed to detect the mecA gene in six samples, whereas MP PCR could detect the mecA gene in all samples, with one exception (Sample 1, Table 5). Reproducible nuc gene detection failed in two out of eight samples (Samples 1 and 3) using MP PCR. In comparison, HP PCR could not reproducibly detect the nuc gene in four out of eight samples. This result reflects the better LoD of multiplex MP PCR discussed above. The mecC gene and the PVL gene were not detected in any of the samples.

Table 5.  C_q values from methicillin-resistant Staphylococcus aureus (MRSA) mediator probe (MP) and hydrolysis probe (HP) PCR for nuc and mecA gene detection of DNA extracted from MRSA patient samples.

Here, we present a novel set of URs for simplified set-up of multiplex real-time PCR. This set was validated for detection using a food panel and a MRSA panel, and compared with the respective HP PCR, which uses five (food panel) or four (MRSA) different dual-labeled HPs. The MP PCR performed equally well in the case of the food panel detection (except for turkey detection) and even better than the reference HP PCR in the case of the MRSA panel.

Multiplex real-time MP PCR is superior to HP PCR in terms of optimization of the fluorogenic molecules for signal generation since the URs do not need to be optimized for the detection of each target. This set of URs can now be used for several multiplex nucleic acid target MP PCRs. For those assays, label-free MPs can be designed either based on already established HP assays or according to the MP design guidelines published previously (10). The UR set could also be pre-stored as lyophilized PCR mixes or by drying the URs into microtiter plates or PCR tubes, similar to the method described in Reference 18. Such universal consumables could be manufactured large-scale, allowing for cost-efficient multiplex real-time PCR.