Introduction
The glassy-winged sharpshooter (GWSS), Homalodisca vitripennis, (previously known as Homalodisca coagulata) (Takiya et al., Reference Takiya, McKamey and Cavichioli2006) is a member of the Cicadellidae family. H. vitripennis is native to the south-eastern United States, and north-eastern Mexico (Triapitsyn & Phillips, Reference Triapitsyn and Phillips2000), but is known for its spread into California in the 1980s (Sorensen & Gill, Reference Sorensen and Gill1996). H. vitripennis is highly mobile and feeds on the xylem of over 100 species of plants (Turner & Pollard, Reference Turner and Pollard1959; Anderson et al., Reference Anderson, Brodbeck and Mizell1989; Northfield et al., Reference Northfield, Mizel, Paini, Andersen, Brodbeck, Riddle and Hunter2009) of which it can process up to 300 times its body mass per day (Brodbeck et al., Reference Brodbeck, Mizell and Andersen1993). H. vitripennis is a vector of the pathogenic bacterium, Xylella fastidiosa (Turner & Pollard, Reference Turner and Pollard1959; Almeida & Purcell, Reference Almeida and Purcell2003), which is the causative agent of diseases in many plants, most notably Oleander Leaf Scorch disease, phony peach disease and Pierce's Disease (PD) in grapevine (Purcell et al., Reference Purcell, Saunders, Hendson, Grebus and Henry1999; Purcell & Feil, Reference Purcell and Feil2001; Hopkins & Purcell, Reference Hopkins and Purcell2002). H. vitripennis has been shown to penetrate deeper into vineyards than other sharpshooters which carry the disease (Blua & Morgan, Reference Blua and Morgan2003), allowing it to spread the disease further than other insect vectors. This attribute was linked to massive outbreaks of PD in California in the late 1990s (Almeida & Purcell, Reference Almeida and Purcell2003; Blua & Morgan, Reference Blua and Morgan2003).
The virulence factors of X. fastidiosa are understood (Roper et al., Reference Roper, Creve, Warren, Labavitch and Kirkpatrick2007; Pérez-Donoso et al., Reference Pérez-Donoso, Sun, Roper, Greve, Kirkpatrick and Labavitch2010; Sun et al., Reference Sun, Greve and Labavitch2011) and diagnostic techniques for the detection of X. fastidiosa are well established (Minsavage et al., Reference Minsavage, Thompson, Hopkins, Leite and Stall1994; Guan et al., Reference Guan, Shao, Singh, Davis, Zhao and Huang2013), yet treatment of the disease is difficult (Dandekar et al., Reference Dandekar, Ibáñez, Gouran, Phu, Rao, Chakraborty, Esser and Randhawa2012). Preventing the spread of PD and limiting the rate of infection by controlling H. vitripennis directly is the most efficient method of combating PD. Chemical and biological control of H. vitripennis has been tested using a variety of agents, including pesticides, predation, and parasitic wasps (Triapitsyn et al., Reference Triapitsyn, Mizell, Bossart and Carlton1998; Bethke et al., Reference Bethke, Blua and Redak2001; Grandgirard et al., Reference Grandgirard, Hoddle, Petit, Roderick and Davies2008; Guiterrez et al., Reference Guiterrez, Ponti, Hoddle, Almeida and Irvin2011), and efforts have even included attempts to inoculate the insect with benign strains of X. fastidiosa (Hopkins, Reference Hopkins2005). The economic burden of these efforts is significant, in California alone the annual cost of monitoring, control and research is approximately US$50 million per year, and despite this effort H. vitripennis causes direct loses of an estimated US$60 million per year (Alston et al., Reference Alston, Fuller, Kaplan and Tumber2013; Tumber et al., Reference Tumber, Alston and Fuller2014). In the early 2000s H. vitripennis spread into the Pacific Islands, with reports of incursions into French Polynesia in 1999, Hawai'i in 2004, the Easter Islands in 2005 and the Cook Islands (Rarotonga) in 2007 (Grandgirard et al., Reference Grandgirard, Hoddle, Roderick, Petit, Percy, Putoa, Garnier and Davies2006; Gunawardana et al., Reference Gunawardana, Ashcroft, Braithwaite and Poeschko2008; Petit et al., Reference Petit, Hoddle, Grandgirard, Roderick and Davies2008). In warmer climates, H. vitripennis flourishes, displaying an increased rate of feeding (Johnson et al., Reference Johnson, Daane, Groves and Backus2006) and mating more frequently (Blua et al., Reference Blua, Phillips and Redak1999; Grandgirard et al., Reference Grandgirard, Hoddle, Roderick, Petit, Percy, Putoa, Garnier and Davies2006), making it capable of achieving a much greater population density than that observed in California (Petit et al., Reference Petit, Hoddle, Grandgirard, Roderick and Davies2008; Wistrom et al., Reference Wistrom, Sisterson, Pryor, Hashim-Buckey and Daane2010). Environmental modelling has suggested that H. vitripennis is capable of surviving in any climate that supports grape production, including that of Australasia (Hoddle, Reference Hoddle2004; Rathé et al., Reference Rathé, Pilkington, Gurr, Hoddle, Daugherty, Constable, Luck, Powell, Fletcher and Edwards2012). It has been speculated that the accidental transport of plants carrying H. vitripennis eggs was the original source of its introduction into California, French Polynesia, and Rarotonga (Sorensen & Gill, Reference Sorensen and Gill1996; Grandgirard et al., Reference Grandgirard, Hoddle, Roderick, Petit, Percy, Putoa, Garnier and Davies2006; Gunawardana et al., Reference Gunawardana, Ashcroft, Braithwaite and Poeschko2008; Petit et al., Reference Petit, Hoddle, Grandgirard, Roderick and Davies2008), and this mode of spread is considered the likeliest source of an incursion into Australasia (Grandgirard et al., Reference Grandgirard, Hoddle, Roderick, Petit, Percy, Putoa, Garnier and Davies2006; Rathé et al., Reference Rathé, Pilkington, Gurr, Hoddle, Daugherty, Constable, Luck, Powell, Fletcher and Edwards2012).
New Zealand is currently free from H. vitripennis and X. fastidiosa, which are considered high-risk organisms for New Zealand's biosecurity sector. While H. vitripennis adults can be identified morphologically this approach does not scale well when many individuals require analysis. Further, when immature stages of H. vitripennis are intercepted, morphology is not capable of distinguishing between the eggs and nymphal stages of H. vitripennis and its close relative, such as the smoke-tree sharpshooter (Homalodisca. liturata) (S. Winterton, personal communication, 2007). Molecular and genetic identification tools have been developed for H. vitripennis, which overcome the morphology limitation, targeting protein and genetic markers (de León et al., Reference de León, Fournier, Hagler and Daane2006; Fournier et al., Reference Fournier, Hagler, Daane, de León, Groves, Costa and Henneberry2006), including a polymerase chain reaction (PCR) assay that targets the mitochondrial cytochrome c oxidase subunit 1 (COI) gene. The COI gene is a commonly used gene in molecular entomology as it can possess sufficient genetic resolution to distinguish between species and subspecies of organisms, as well as explore population genetics (Smith, Reference Smith2005; Boykin et al., Reference Boykin, Shatters, Rosell, McKenzie, Bagnall, De Barro and Frohlich2007; Malausa et al., Reference Malausa, Fenis, Warot, Garmain, Ris, Prado, Botton, Vanlerberghe-Masutti, Sforza, Cruaud, Couloux and Kreiter2011; Rakauskas et al., Reference Rakauskas, Turčinavičeine and Bašilova2011). The COI gene has previously been used to evaluate the phylogeny of several Hemiptera suborders (Park et al., Reference Park, Foottit and Hebert2011; Li et al., Reference Li, Tian, Zhao and Bu2012; Foottit et al., Reference Foottit, Maw and Hebert2014), and the common occurrence of this gene in publically accessible data repositories makes it an excellent starting point for assay design.
In order to augment the existing entomological tools for the identification of H. vitripennis, we sought to develop a TaqMan based quantitative PCR (qPCR) protocol to rapidly and accurately identify H. vitripennis from both adult and egg life stages. The assay was based on the COI gene as it is a commonly used genetic marker in molecular entomology and thus provided a good range of reference organisms to test for non-specific binding. The specificity of the assay was examined extensively in silico and further validated through blind panel testing. Assay sensitivity was evaluated using controlled quantities of template DNA. Compared with conventional PCR techniques this assay is rapid, taking approximately 45 min to complete (compared with 2–3 h for PCR, followed by gel electrophoresis). Furthermore, we predict this assay will prove more robust than conventional PCR assays, due to the required annealing of a third oligonucleotide sequence.
Materials and methods
Sample collection and DNA extraction
Samples of the glassy-winged sharpshooter (H. vitripennis) collected from the USA and Cook Islands, and smoketree sharpshooter (H. liturata), and blue-green sharpshooter (Graphocephala atropunctata) samples from the USA were used to develop the assay. H. vitripennis eggs were obtained from the USA. Additional cicadellid samples were obtained from the Plant Health and Environment Laboratory entomology reference collection (PANZ) for use in specificity testing. DNA was extracted from the leg or whole body of the insect (table 1). DNA extraction was performed using the DNeasy Blood and Tissue kit (Qiagen NV, Venlo, Netherlands) according to manufacturer instructions and stored at −80°C until required.
Cicadellidae samples obtained for blind panel testing.
DNA was extracted from small or nymph-stage samples by means of non-destructively processing the entire insect. For the larger insects a single leg was removed from the insect and DNA extracted.
Analysis of phylogenetic relationships
Sequence data obtained from the COI gene for 164 Leafhoppers (162 Cicadellidae, 2 Fulgorodoidea) were downloaded from the GenBank nr (non-redundant) database (online table S1). Multiple sequence alignment was performed using MUSCLE version 3.8.31 (Edgar, Reference Edgar2004) with default parameters. Following alignment, a 370 bp subsection of the alignment common to all sequences was identified and extracted from the global alignment. Phylogenetic analysis was performed by constructing a maximum likelihood tree using RAxML version 8 (Stamatakis, Reference Stamatakis2014). The COI sequences of four species of the order Orthoptera were used as outgroup (online Supplemental table S1). Analysis of the phylogenetic tree was used to identify the closest genetic relatives to H. vitripennis, which were then extracted from the initial data set for use as references during primer design.
qPCR design
The original COI sequences of H. vitripennis and H. liturata were re-aligned, and the primers and probe were designed using Beacon Designer version 8.01 (Premier Biosoft, Palo Alto, CA, USA). A reverse primer and probe were designed to work in conjunction with the forward primer of a previously published H. vitripennis PCR assay (table 2) (de León et al., Reference de León, Fournier, Hagler and Daane2006) to take advantage of the previously established specificity of this target region. An additional 15,505 Cicadellidae COI sequences, six of which belonged to H. vitripennis, were downloaded from the Barcode of Life Data Systems (BOLD) (Ratnasingham & Hebert, Reference Ratnasingham and Hebert2007) for use in silico primer testing. The TaqMan probe was synthesized with a 5′ reporter fluorophore (FAM) and 3′ quencher molecule (BHQ) (Biosearch Technologies Inc, Petaluma, CA, USA).
Primer sequences used in the final assay.
The primer HcCOI-F was obtained from the previously reported H. vitripennis assay (de León et al., Reference de León, Fournier, Hagler and Daane2006). Probe GWSS_P1 was synthesized with the reported fluorophore FAM and proprietary quencher BHQ. The minimum and average sequence differences between the appropriate regions of the leafhopper data set are reported.
qPCR optimization
Initial qPCR annealing conditions were tested in 10 µl reactions using the SsoFast Probes Supermix (BioRad, Hercules, CA, USA) with 0.8 µg µl−1 bovine serum albumin (Sigma-Aldrich Co., St Louis, MO, USA), 400 nM of each primer and 250 nM of probe. An annealing gradient of 55–65°C was used with the cycling conditions as follows; initial denaturing at 95°C for 2 min, followed by 40 cycles of 95°C for 20 s and annealing/extension for 30 s followed by a plate read step. Following the establishment of the optimal annealing temperature, the primer concentration was tested at 50, 100, 200 and 400 nM for each primer. For primer concentrations below 400 nM the probe concentration was reduced to 125 and 200 nM. Repeatability (intra-run variation) and reproducibility (inter-run variation) were tested with three qPCR master mix solutions; SsoFast Probes Supermix, SsoAdvanced Universal Probes mix (BioRad) and PerfectA qPCR Toughmix (Quanta Biosciences, Gaithersburg, MD, USA). All reactions were performed using a C1000 Touch™ thermocycler with CFX96 Optical Reaction Module™ (BioRad) and results were analyzed using CFX Manager version 3.1 (BioRad).
Assay repeatability and reproducibility
Eight samples of H. vitripennis DNA were tested in triplicate in two independent machine runs. Repeatability and reproducibility were quantified using the percentage coefficient of variation (%CV) of the detection threshold value (Cq). All statistical analysis and plotting were performed in the R software environment (R Core Team, 2013).
Assay sensitivity
Sensitivity of the assay was determined through the use of qPCR against a controlled number of target copies. A single-gene target sequence was constructed by amplifying a 485 bp fragment of the H. vitripennis COI gene using the primer pair C1-J-1718 (forward, 5′- GGA GGA TTT GGA AAT TGA TTA GTT CC -3′) and C1-J-2191 (reverse, 5′-CCC GGT AAA ATT AAA ATA TAA ACT TC-3′) (Simon et al., Reference Simon, Frati, Beckenbach, Crespi, Liu and Flook1994). Reactions were performed in 20 µl reactions using the GoTaq master mix (Promega, Madison, WI, USA) with 0.4 µg µl−1 bovine serum albumin (Sigma-Aldrich Co.) and 500 nM of each primer. Cycling conditions were as follows: initial denaturing at 94°C for 3 min followed by 40 cycles of denaturing at 94°C, annealing at 50°C and extension at 72°C for 45 s each, then a final extension at 72°C for 5 min. PCR product was evaluated by running on a 1% (w/v) agarose gel stained with Invitrogen SYBR Safe (Life Technologies, Auckland, New Zealand) and results were visualized using a GelDoc XR+ system (BioRad). PCR product was purified using the illustra Microspin column (GE Healthcare, Pittsburgh, PA, USA) kit according to manufacturer instructions and the purified product cloned using the Invitrogen TOPO TA vector cloning kit (Life Technologies). Successfully transformed clones were selected and the insert examined by amplifying with the M13 primer pair, designed to amplify the insert sequence (provided as part of the TOPO TA vector cloning kit). Inserts were sequenced by EcoGene® (Auckland, New Zealand) to confirm that the correct sequence had been cloned. A single clone was then grown in overnight culture and plasmid DNA extracted using the Wizard Plus SV Miniprep kit (Promega) and linearized by incubating overnight with Pst I (BioLab Inc, Lawrenceville, GA, USA). Linearized plasmid DNA was then quantified using NanoDrop (Thermo Fisher Scientific, Auckland, New Zealand) and normalized to a concentration of 109 copies µl−1. Triplicate qPCR reactions were performed against a range of target gene concentrations (107–10 copies per reaction) using the previously reported conditions. The detection threshold of each reaction was plotted against the log10 of the template count and linear regression performed, measuring the fit as r2. The amplification efficiency was calculated from the linear trend line using the equation E = 10|1/slope| and was converted to a percentage value using E % = (E−1) × 100.
Assay specificity
Specificity of the assay was tested using a blind panel procedure, whereby 24 samples were tested using the developed assay by a diagnostician. The sample collection contained a mixture of Cicadellidae species including those native to New Zealand and commonly intercepted at the borders, as well as Graphocephala actropunctata and H. liturata, with H. vitripennis DNA randomly interspersed. Assay results were interpreted without knowledge of the sample identity and results were recorded as ‘GWSS’ (positive identification) or ‘Not GWSS’.
Results
Analysis of phylogenetic relationships and qPCR design
Phylogenetic analysis of publically available COI sequence fragments confirmed that H. vitripennis forms a deep-branching lineage separate from other Cicadellidae. The closest neighbour to H. vitripennis was H. liturata and these two organisms formed a strongly supported clade apart from other species of Homalodisca (fig. 1, online Supplemental table S1). We therefore sought to maximize the differences between these organisms while keeping within the constraints of the TaqMan chemistry. Primers from an existing PCR assay of the H. vitripennis COI gene were used as starting points during the design process (primer pair HcCOI-F/HcCOI-R) (de León et al., Reference de León, Fournier, Hagler and Daane2006). A stable probe and reverse primer were predicted to function with the HcCOI-F primer while still providing diagnostic specificity for H. vitripennis (table 2), which amplified a 163 bp fragment. Primers and probe sequence were mapped against the BOLD reference data using default methods, yielding no hits for either primer or probe to sequences that did not belong to H. vitripennis. Optimal reaction conditions were identified as annealing/extension temperature of 62°C and primer/probe concentrations of 200 nM per reaction.
Phylogenetic analysis of Cicadellidae-derived COI gene sequences. COI sequences were obtained from GenBank and a 370 bp region of sequence, common to 33 close relatives of H. vitripennis, aligned. Asterisk (*) denotes species found in New Zealand. Branch lengths were calculated using the maximum-likelihood method RAxML using a general time-reversible model with gamma rate distribution. Node support was calculated based on 1000 bootstrap iterations. Solid junctions denote nodes with ≥85% support and hollow junctions ≥65% support.
Assay repeatability, reproducibility and sensitivity
Repeatability and reproducibility of the assay were high; with the lowest median %CV reported for the SsoAdvanced universal probes mix (fig. 2), which was used exclusively for the sensitivity and specificity. The limit of detection for the assay was identified as 102 copies/reaction with a linear dynamic range of 107–102 copies. The efficiency for the assay was 89.9%, with an r 2 of 0.99 (fig. 3). A template concentration of 10 copies/reaction was sporadically detected in the assay with an average Cq of ~36 cycles. Due to inconsistent detection at this concentration it was excluded from the linear dynamic range, but was used to set an upper limit on the maximum allowable Cq in the assay (36 cycles), with any signal detected after this point considered a late-amplification false positive and treated as a negative result.
Intra- and inter-run variation using three qPCR master-mix solutions. The median %CV for all solutions was low, with the inter-run %CV being slightly higher than intra-run for each solution. The SsoAdvanced solution was selected due to its low overall %CV, and narrower range of variance between runs.
Linear dynamic range of the qPCR assay. Efficiency of the assay was derived from the linear regression using standard methods. Slope of the regression curve yielded an efficiency of 89.9%, with a fit of r 2 = 0.99. Shading denotes the 95% confidence interval of regression.
Assay specificity
Twenty-four samples were tested with the assay, tested by a diagnostician who was unaware of the source of each DNA sample (table 3). A single H. vitripennis sample, extracted from a poorly preserved nymph cast skin, was not correctly identified (table 3, Sample 3). Further analysis revealed that this sample yielded an extremely low DNA quality. One additional sample was identified late in the reaction (table 3, Sample 21); this sample had been extracted from an old and poorly preserved specimen and the DNA obtained from this sample possessed an extremely low quality, with a 260/280 ratio of 0.98. Running the assay against DNA extracted from another individual, obtained from the same location but preserved in ethanol prior to extraction, provided a more conclusive result (median Cq = 22.3, data not shown). DNA obtained from H. vitripennis eggs were all identified correctly using the assay. DNA extracted from H. vitripennis egg masses was identified in all tests performed (median Cq = 26.4).
Results of the blind panel testing procedure.
Mean detection Cq values are reported, dash (-) indicates no signal was generated during the assay. H. vitripennis samples are shaded. A cut-off of 36 cycles was used when interpreting Cq, based on the sensitivity limit of the assay. Results were recorded simply as an H. vitripennis hit (GWSS) or no signal (Not GWSS). Sample 3 produced a false negative likely due to originating from poorly preserved tissue sample, as discussed in the main article.
Discussion
The spread of H. vitripennis in the Pacific in recent years has been attributed primarily to poor quarantine procedures that resulted in the transport of plants carrying H. vitripennis eggs between islands (Grandgirard et al., Reference Grandgirard, Hoddle, Roderick, Petit, Percy, Putoa, Garnier and Davies2006; Petit et al., Reference Petit, Hoddle, Grandgirard, Roderick and Davies2008). There is lack of pattern to the spread of H. vitripennis in the Pacific Islands, which is consistent with random spread due to human activity rather than a progressive migration of the species, and human activity is considered to be the likeliest source of an incursion into New Zealand (Grandgirard et al., Reference Grandgirard, Hoddle, Roderick, Petit, Percy, Putoa, Garnier and Davies2006; Rathé et al., Reference Rathé, Pilkington, Gurr, Hoddle, Daugherty, Constable, Luck, Powell, Fletcher and Edwards2012). It is therefore important to develop a rapid diagnostic tool for the identification of H. vitripennis to prevent the spread of the pests in New Zealand once it is found.
DNA-based identification via qPCR is a technique commonly performed in virology, mycology, and bacteriology and is increasingly becoming popular within entomology (Jones et al., Reference Jones, Gorman, Denholm and Williamson2008; Huang et al., Reference Huang, Lee, Yeh, Shen, Mei and Chang2010; Dhami & Kumarasinghe, Reference Dhami and Kumarasinghe2014; ven de Vossenberg & van der Straten, Reference ven de Vossenberg and van der Straten2014). Comparisons of an existing H. vitripennis assay based on enzyme-linked immunosorbent assay (ELISA) of egg proteins and the previously mentioned COI gene PCR assay have shown that the PCR assay is able to detect H. vitripennis tissue in a higher number of samples than the ELISA method, primarily due to the universal presence of DNA compared with the gender-specific expression of the protein target (Fournier et al., Reference Fournier, Hagler, Daane, de León and Groves2008; Hagler et al., Reference Hagler, Blackmer, Krugner, Groves, Morse and Johnson2013). Within DNA-based identification techniques, qPCR assays are significantly faster than their conventional counter parts. The qPCR assay designed in this study required approximately 45 min to complete, compared with 3 h for the conventional assay. Rapid identification is the cornerstone of biosecurity when dealing with invasive organisms as delays in identification, however minor, have drastic impacts on the success of quarantine and containment protocols.
The qPCR developed can be used to detect H. vitripennis of all life stages, however, as all the other tests there are limitations for this real-time PCR assay. A universal limitation to molecular diagnostic techniques is that there is always the risk that a genetic sequence not foreseen during the design process may react with the assay causing a false positive reaction. New Zealand possesses a diverse insect population, of which it is estimated half is currently undescribed (Cranston, Reference Cranston2010). Of the described insect fauna, New Zealand hosts over 544 species of Hemiptera, at least 444 of which are endemic, including 80 Cicadellidae (51 endemic) (Larivière, Reference Larivière2005). As these fauna are poorly represented in GenBank (online table S1) a variety of these insects were included in testing in order to ensure that false positive reactions would not be likely to occur with insects that are already present in New Zealand. Phylogenetic analysis of the short COI gene region was sufficient to demonstrate previously reported findings obtained from the complete mitochondrial genome; that Homalodisca forms a distinct clade apart from the other Cicadellidae (Song et al., Reference Song, Liang and Bu2012). This does not guarantee that no Cicadellidae COI gene is capable of generating a false positive result, it gives confidence that Cicadellidae COI genes sequenced in the future will fall further away from H. vitripennis.
On the basis that H. liturata was the closest relative over the sequence fragment we aimed to maximize the differences between these two organisms in the assay. H. liturata possessed fewer mismatches across its three binding sites than any other sequence, including those with sequence data that only accounted for two binding sites (online Supplemental table S1). Our results showed that the samples of H. liturata were tested negative in the qPCR assay developed in this study (tables 1 and 3). Samples of two additional related species in the genus, Homalodisca elongata and H. insolata were not obtained for testing the qPCR assay despite multiple attempts to obtain specimens. However the in silico analysis showed that the degree of mismatches in the primers and probe sites (online Supplemental table S1) make it extremely unlikely that the qPCR assay for H. vitripennis will cross react with the H. elongata and H. insolata.
Quantitative PCR assays-based are subject to specific parameters for the assay to perform optimally. Factors such as the melting temperature (Tm) profile and GC content of the primers and probe and, in the case of TaqMan, the relative positioning of the oligonucleotides on the target sequence dictate the design of qPCR assays. As a consequence, the functionality of a conventional PCR assay cannot be directly extrapolated to a qPCR counterpart. For example, the HcCOI-F/HcCOI-R assay developed by de León et al. (Reference de León, Fournier, Hagler and Daane2006) possessed an limit of detection (LOD) of 6 pg of genomic DNA, and the assay developed in this study possessed an LOD of 100 copies per reaction (using cloned plasmid DNA containing a COI insert). Since there are multiple copies of COI in genomic DNA and a single copy in each plasmid these numbers cannot be directly compared. The degree of sensitivity of de Leon's primers was tested as a real-time PCR using SYBR chemistry with the plasmid construct, and a lower detection sensitivity was observed than for conventional PCR (data not shown). In additional, testing the blind panel samples using the original HcCOI-F/HcCOI-R assay yielded a sporadic, weak false-positive reaction with the New Zealand endemic fulgoroid Zeoliarus oppositus (data not shown), which was not observed using the qPCR assay. This result is likely due to the additional H. vitripennis-specific mutations introduced by the TaqMan probe as well as the cut-off threshold for accepting a positive result, determined by the linear dynamic range of the assay. The testing regime used during the development of this assay ensures that our protocol is conformant with minimum information for publication of quantitative real-time PCR experiments (MIQE) guidelines for qualitative assays development (Bustin et al., Reference Bustin, Benes, Garson, Hellemans, Huggett, Kubista, Mueller, Nolan, Pfaffl, Shipley, Vandesompele and Wittwer2009).
Within blind panel testing, late-stage amplification was observed in several samples from organisms endemic to New Zealand (table 3, Edwardsiana cratagi, Erythroneura elegantula, Idiocerus distinguendus) yet these signals could be ruled out using the cut-off threshold. During blind panel testing, the qPCR assay was unable to correctly identify a single H. vitripennis sample (table 3, Sample 3). This sample was an aged cast skin sample and further testing showed that the negative results for the sample was due to the low quality of DNA extracted. We concluded that this false-negative result was an artefact of DNA extraction due to the sample storage and did not reflect a lack of sensitivity in the assay. Although this finding does not detract from the sensitivity of the assay it does highlight an important limitation of molecular diagnostic approaches.
In summary we have designed, optimized, and validated a real-time PCR assay for the rapid and accurate identification of H. vitripennis to be employed as part of New Zealand biosecurity practices. This assay is much faster to perform than conventional PCR equivalents. The primers and probe have been extensively validated in silico and tested against a range of closely and distantly related samples to simulate a real diagnostic scenario. This assay has proven to be accurate and sensitive, which is essential for the future diagnostic applications.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S000748531600095X.
Acknowledgements
We are grateful to Dr Mark Hoddle, Dr David Morgan, Dr Youngsoo Son, and Humesh Kumar (California Department of Food and Agriculture, USA), Dr Matt Daugherty (University of California, USA) and Dr Maja Poeschko (Ministry of Agriculture, Cook Islands) for providing sharpshooter samples for assay validation. This work was funded by the Ministry for Primary Industries Operational Research Programme, New Zealand.
