RPA assay development and optimization

The target genes for PCR detection of STHs [15,17,41] were used for RPA assay design. For each target species, at least two probes and about 20 to 30 primer pairs per probe were designed according to guidelines (www.twistdx.co.uk). For optimal RPA performance, candidate primers and probe pairings were screened for each STH species in duplicate reactions using either 50 or 20 plasmid DNA copies per reaction using TwistAmp exo kits (TwistDx, Cambridge, UK) following the manufacturer’s recommended protocols. We selected the primers and probe sets that produced the shortest amplification times, steep fluorescence increases when template was present, a flat fluorescence baseline in the negative control, and consistent fluorescent curves between replicate reactions. Table 1 lists the optimal primers and probe for each target.

Note: TwistAmp exo Probe contains an abasic nucleotide analogue (H), fluorophore (F) and a corresponding quencher (Q). F and Q replaced T residue found within the corresponding target sequence. All probes are blocked from any potential polymerase extension by a 3’–modification group (C3-spacer). The analogue, H, is the cleavage site of Exonuclease III present in the TwistAmp exo kit.

Each singleplex RPA reaction contained 0.48 μM each of forward and reverse primers, 0.12 μM probe, 150 ng of background stool DNA, 2 μL of template DNA, and 1X of supplied rehydration buffer. Negative control (NC) reactions were included using background stool DNA instead of the target DNA. A total of 47.5 μL of master mix was aliquoted into each reaction tube containing the freeze-dried pellet, and then 2.5 μL of 280 mM magnesium acetate (MgAc) was added to the lid of the tube. Tubes were spun down to mix the MgAc into the solution to initiate the reaction. RPA reactions were monitored in real-time via fluorescent detection for FAM and ROX using either Twista (TwistDx, Cambridge, UK) or an ISO-T8 device (Axxin, Victoria, Australia) for 15 min at 40°C with a 30 sec measurement interval and a manual mixing step at 4 min incubation.

Duplex/triplex RPA assays were customized at TwistDx (Cambridge, UK) using the optimized primers and probe set for each target species. The assays were further optimized by testing different primer biases to improve performance. Different RPA protein ratios were compared to see whether a custom formulation would improve the limit of detection and speed of the reactions. Finally, different ratios of the individual assays were combined to maximize their performance. For internal control (IC), a novel plasmid DNA containing a Bacillus subtilis DNA fragment was developed (Fig 1), and then lyophilized into the Ad-RPA/ Na-RPA reactions along with its corresponding probe that specifically target the cloned B. subtilis DNA.

A short section of Bacillus subtilis gDNA was amplified using primers tailed with Ad- or Na-RPA primer sequences, and the resulting PCR product was cloned into a plasmid. The plasmid construct (IC) contains the Na-F4 (dark blue) and Ad-R1c (purple) primer binding sites. BSRUVB probe designed to target the B. subtilis DNA fragment [42] is shown in yellow, while Na and Ad probes are in black. In RPA assay, the IC is amplified by the Na-F4 and Ad-R1c primers and the amplicon detected by BSRUV.

The customized, two-tube duplex/triplex RPA assay, referred to as Dx4STH-1 (duplex, for detecting A. lumbricoides and T. trichiura) and Dx4STH-2 (triplex, for A. duodenale–N. americanus–IC), contains a lyophilized pellet with all the proteins, dNTPs and oligonucleotides needed to run an RPA reaction. For testing, 2 μL of template DNA was added into 23 μL water and 25 μL supplied rehydration buffer with MgAc, and then aliquoted into each reaction tube containing the customized RPA pellet. RPA reactions were monitored in real-time using fluorescent detection as described above.

RPA assays were tested using DNA extracts from STH-positive stools and using genomic DNA from NTEPs at 1 ng per reaction to test their specificity. The analytical sensitivity/ limit of detection of each assay was assessed using four plasmid DNA (pAl, pTt, pAd and pNa), each harboring the target DNA region at levels ranging from 5 to 1000 target copies per reaction in 10 or 20 replicate reactions per level. The limit of detection was determined using probit analysis.

Raw fluorescence data from RPA assays were exported into Excel for analysis. Threshold time (T_t), which represents the time in minutes at which the fluorescence signal crosses the fluorescence threshold (F_t) and is therefore analogous to threshold cycle in real-time PCR, was manually determined from each positive amplification. F_t for each probe was defined as the maximum fluorescence signal (from time 0 to 15 minutes) in a NC reaction + zSD, where NC is the negative control, z is an arbitrary value set between 10 and 100, and SD is the standard deviation of the mean fluorescence signal for each probe. A sample was positive if the fluorescence generated within a reaction was significantly above the background fluorescence and crossed F_t.

To correlate amplification time with target concentration, RPA was run using plasmid DNA standards, after which T_t was determined and then plotted against plasmid concentration to construct a standard curve. The procedure was conducted four times to establish reproducibility of results, creating 12 sample data points per concentration.