A new approach for the simultaneous identification of the viruses and vectors responsible for tomato yellow leaf curl disease (TYLCD) epidemics is presented. A panel of quantitative multiplexed real-time PCR assays was developed for the sensitive and reliable detection of Tomato yellow leaf curl virus-Israel (TYLCV-IL), Tomato leaf curl virus (ToLCV), Bemisia tabaci Middle East Asia Minor 1 species (MEAM1, B biotype) and B. tabaci Mediterranean species (MED, Q biotype) from either plant or whitefly samples. For quality-assurance purposes, two internal control assays were included in the assay panel for the co-amplification of solanaceous plant DNA or B. tabaci DNA. All assays were shown to be specific and reproducible. The multiplexed assays were able to reliably detect as few as 10 plasmid copies of TYLCV-IL, 100 plasmid copies of ToLCV, 500 fg B. tabaci MEAM1 and 300 fg B. tabaci MED DNA. Evaluated methods for routine testing of field-collected whiteflies are presented, including protocols for processing B. tabaci captured on yellow sticky traps and for bulking of multiple B. tabaci individuals prior to DNA extraction. This work assembles all of the essential features of a validated and quality-assured diagnostic method for the identification and discrimination of tomato-infecting begomovirus and B. tabaci vector species in Australia. This flexible panel of assays will facilitate improved quarantine, biosecurity and disease-management programmes both in Australia and worldwide.
Tomato yellow leaf curl disease (TYLCD) has emerged as one of the most economically important diseases of cultivated tomato (Solanum lycopersicum) worldwide (Navas-Castillo et al., 2011). It is caused by a complex of several distinct begomovirus species and their strains (family Geminiviridae), often referred to as TYLCV-like viruses (Brown et al., 2012). The most geographically widespread and well studied of these species is Tomato yellow leaf curl virus (TYLCV) (Scholthof et al., 2011). Begomoviruses are transmitted in a circulative and persistent manner by the highly polyphagous whitefly Bemisia tabaci (Hemiptera: Sternorrhyncha: Aleyrodoidea: Aleyrodidae) (Cohen & Nitzany, 1966). Bemisia tabaci is an invasive agricultural pest with a global distribution, causing crop damage directly through phloem feeding and honeydew secretion, and indirectly as the vector of plant viruses, predominantly begomoviruses (De Barro et al., 2011; Navas-Castillo et al., 2011). Recent research indicates that B. tabaci is a cryptic species complex composed of at least 28 morphologically indistinguishable species, commonly referred to as biotypes (Dinsdale et al., 2010; Hu et al., 2011). The emergence and global spread of TYLCV-like viruses is closely linked with the parallel dispersal and establishment of members of the cognate vector B. tabaci species complex, with both virus and vector dispersal concomitantly associated with increased international movement and trade of horticultural products (Polston & Anderson, 1997; De Barro et al., 2011).
In Australia, Tomato yellow leaf curl virus-Israel (TYLCV-IL) was first discovered in open-field tomato crops in Brisbane, southeast Queensland, and is now considered a significant constraint to tomato production in all major tomato-growing regions of Queensland. Prior to the introduction of TYLCV-IL, the only other begomovirus reported from Australia was the endemic Tomato leaf curl virus (ToLCV), which causes disease in minor production areas in northern Australia and probably predates European colonization (Stonor et al., 2003). At present the distributions of TYLCV and ToLCV do not overlap (the most southerly record of ToLCV is Mossman, about 70 km northeast of Mareeba, the most northerly record of TYLCV) and thus there has been no confusion as to which of the viruses has been causing TYLCD. However, it is likely that within a few years the geographical ranges of these two begomoviruses will overlap, and suitable diagnostic methods will be required for differential diagnosis.
The establishment of TYLCV-IL in Australia was aided by the presence of B. tabaci Middle East Asia Minor 1 (commonly referred to as biotype B, herein MEAM1), which was first identified in Australia in the early 1990s (Gunning et al., 1995). The introductions of both B. tabaci MEAM1 and TYLCV-IL highlight the need for improved detection and surveillance measures to limit the expansion of TYLCV-IL and to reduce the risk of future introductions of new species of begomovirus and B. tabaci. Of particular quarantine concern in Australia is the exotic B. tabaci Mediterranean (commonly referred to as biotype Q, herein MED), a globally invasive whitefly that is resistant to a wide range of insecticides (Horowitz et al., 2005). To date, a number of separate real-time PCR assays have been described for the detection of TYLCV-IL and for the differentiation of B. tabaci MEAM1 and MED (Jones et al., 2008; Papayiannis et al., 2009, 2010). However, these assays do not incorporate internal controls, an essential requirement for the exclusion of false negative results (Hoorfar et al., 2004). At present no tests provide the capability to simultaneously identify B. tabaci species and begomovirus species, nor to accurately quantify target species. In Australia, TYLCV-IL and ToLCV are identified using double-antibody sandwich enzyme-linked immunosorbent assay (DAS-ELISA) and conventional PCR (van Brunschot et al., 2010). To differentiate B. tabaci MEAM1 from the exotic B. tabaci MED and the indigenous B. tabaci Australia (AUS) and Australia/Indonesia (AUS/INDO) species, microsatellite-based PCR methods are used (De Barro et al., 2003). Although these methods are generally adequate for initial diagnosis, significant improvements in reliability, efficiency and throughput could be achieved using a multiplexed (simultaneous), internally controlled, nucleic-acid-based diagnostic approach.
In Australia, quarantine regulations aim to control the importation of potential hosts of exotic plant pathogens and pests. Interstate quarantine regulations control the movement of hosts of TYLCV-IL and B. tabaci within Australia to restrict their spread. Also, active monitoring and surveillance programmes for natural B. tabaci populations exhibiting high levels of insecticide resistance have been introduced as a means of early warning, should an incursion of B. tabaci MED occur. Therefore, the objectives of this study were to: (i) determine if a panel of real-time PCR assays could accurately detect TYLCV-IL, ToLCV, B. tabaci MEAM1 and B. tabaci MED, from either plant or B. tabaci samples; (ii) reduce the risk of false negative results via the incorporation of internal control assays for plant and B. tabaci sample types; (iii) determine if the specificity, sensitivity and reproducibility of the real-time PCR assays are adequate for routine diagnostic use; (iv) evaluate methods for testing large numbers of B. tabaci samples; and (v) assess the effectiveness of the developed assays using field-collected B. tabaci and tomato samples.
Materials and methods
Virus isolates and whitefly samples
Freeze-dried plant virus isolates used in this study (stored at −20°C) are listed in Table 1. Healthy tomato plants grown in an insect-proof glasshouse were used as negative controls. Collaborators provided adult B. tabaci, Bemisia afer and Trialeurodes vaporariorium individuals preserved in 95% ethanol (stored at −20°C), which were initially obtained from either wild populations or laboratory colonies (Table 2). Field-collected virus and whitefly samples are described in the sections on multiplexed analysis of B. tabaci and tomato samples.
Table 1. Characteristics of the virus isolates used in this study
Virus isolates stored in the Queensland Department of Agriculture, Fisheries and Forestry (QDAFF) Plant Pathology Herbarium and QDAFF Virology collection.
Genome components were confirmed by sequencing and/or conventional PCR specific for DNA-A, DNA-B and DNA-β satellite molecules.
Multiple GenBank accessions are available where multi-component virus genomes were characterized (either DNA-β satellite or DNA-B molecules).
Mean cycle threshold (Ct) values were computed for infected plant DNA extracts, with each sample tested in triplicate over separate runs. Ct values below 40 indicate a positive result; standard deviation is indicated in parentheses.
The species delimitation terminology of Dinsdale et al. (2010) was used for B. tabaci isolates.
Mean cycle threshold (Ct) values were computed for insect DNA extracts, with each sample tested in duplicate over separate runs. Ct values below 40 indicate a positive result; standard deviation is indicated in parentheses.
DNA was extracted from 0·01 g lyophilized leaf tissue using a TissueLyser bead mill (QIAGEN) and a BioSprint 15 DNA Plant Kit (QIAGEN), both used according to the manufacturer's instructions. DNA from individual and bulked B. tabaci was extracted using the method of De Barro & Driver (1997), except that 0·45% (v/v) Triton-X was used instead of 0·45% NP40 in the lysis buffer. Bulked whitefly samples (between five and 20 individuals) were homogenized in 50 μL lysis buffer. All DNA extracts were stored at −20°C.
Conventional PCR and sequencing
In order to accurately assess the specificity of the real-time PCR assays, reference isolates of begomoviruses and whiteflies from a variety of hosts and geographical locations were collected and sequenced. Partial or complete DNA-A, DNA-B and DNA-β satellite molecules were sequenced for begomoviruses, and the mitochondrial cytochrome oxidase subunit 1 (mtCOI) gene was sequenced for reference whitefly specimens. The full-length DNA-A molecules of begomovirus isolates BRIP5702, BRIP5703 and BRIP5701 were amplified by rolling-circle amplification (RCA) with random priming using Phi29 DNA polymerase (TempliPhi; GE Healthcare), and cloned and sequenced as described previously in van Brunschot et al. (2010). Briefly, RCA-amplified concatamers were digested with BamHI to yield c. 2·7- kb unit-length genomes. These were ligated into pBluescript II SK (+) (Stratagene) and sequenced by primer walking. Partial begomovirus DNA-A sequences (c. 1·1 kb) were amplified using primers PALIv1978/PAR1c496, and partial DNA-B sequences (c. 0·6 kb) were amplified using primers PBL1v2040/PCRc1 (Rojas et al., 1993). Complete DNA-β satellite sequences (1·35 kb) were amplified using outward-facing primers β 01/β 02 (Briddon et al., 2002). The mtCOI genes of B. tabaci and B. afer isolates were amplified using the PCR primers C1-J-2195/TL2-N-3014 (Simon et al., 1994). PCR products were purified using the E-Gel CloneWell system (Invitrogen) and directly sequenced using the corresponding amplification primers. All sequencing reactions were prepared using the BigDye Terminator v. 3.1 Cycle Sequencing Kit (Applied Biosystems) and processed by the Australian Genome Research Facility (AGRF, The University of Queensland). Sequence data from this study were deposited in the GenBank database of the National Center for Biotechnology Information (NCBI), with the accession numbers listed in Tables 1 and 2.
Real-time PCR assay design
In order to design the begomovirus and B. tabaci species-specific primers, multiple sequence alignments of sequence data available on GenBank and sequences generated in this study (Tables 1, 2) were compiled using clustalW (Thompson et al., 1994). Based on these multiple alignments, variable regions were identified to differentiate: (i) TYLCV-IL from other TYLCV strains including TYLCV-Mild and TYLCV-Gezira (Brown et al., 2012), and from related begomovirus species and strains (target C4 ORF); (ii) ToLCV from related begomovirus species and strains (target V1/V2 ORF); and (iii) B. tabaci MEAM1 and B. tabaci MED from each other and from all other cryptic species of B. tabaci (target mtCOI gene). For analysis of plant samples, an internal control assay (tomato internal control assay (TomIC)) was designed to a conserved region of the single copy tomato FLORICAULA/LEAFY gene homologue, for the co-amplification of solanaceous plant DNA. For analysis of B. tabaci samples, an internal control assay (B. tabaci internal control assay (BtabIC)), targeting a conserved region of the B. tabaci 18S ribosomal RNA gene, was developed.
Oligonucleotide primers and dual-labelled hydrolysis (TaqMan) probes were designed for each target-specific assay according to established guidelines (Dorak, 2006). All assays were designed, developed and validated in accordance with the accepted guidelines and requirements for clinical diagnostic assay development (Bustin et al., 2009). Primers (Geneworks), TaqMan MGB probes (Applied Biosystems) and dual-labelled fluorogenic probes (Biosearch Technologies) are listed in Table 3. Primer and probe characteristics were evaluated using beacon designer free online tools (Premier Biosoft International), with specificity checked against sequence data available on GenBank using blast (NCBI). All assays were designed to amplify a product shorter than 200 bp. Each of the four species-specific probes was labelled with a spectrally discrete fluorescent dye to minimize potential cross-talk between the channels of the real-time PCR instrument.
Table 3. Oligonucleotide primers and probes developed in this study
Melting temperature (Tm) determined by Beacon Designer free online tools (Premier Biosoft International).
MGB: minor groove binder; NFQ: non-fluorescent quencher; BHQ: black hole quencher.
Nucleotide positions of primer and probe sequences as compared with reference sequences: TYLCV-IL:X15656; ToLCV: AF084006; TomIC: AF197936; MEAM1: GU086340; MED: GU086329; BtabIC: EF405536.
Tomato yellow leaf curl virus-Israel (TYLCV-IL)
Tomato leaf curl virus (ToLCV)
Solanaceous plant internal control (TomIC)
FLORICAULA/LEAFY single copy gene
Bemisia tabaci MEAM1 (MEAM1)
Mitochondrial cytochrome oxidase 1 (mtCOI) gene
B. tabaci MED (MED)
CAL Fluor Red® 610
B. tabaci internal control (BtabIC)
18S rRNA gene/internal transcribed spacer
Real-time PCR conditions
Real-time PCR conditions were optimized in singleplex to maximize assay specificity, sensitivity and signal-to-noise ratio and then in multiplex (either in duplex using an appropriate internal control assay, or in triplex using an appropriate internal control in conjunction with two species-specific assays), conditional to the samples being tested. The final optimized reaction mix for all assays contained 1× RealMasterMix Probe without ROX (5′ PRIME), probes at 0·2 μm each, corresponding forward and reverse primers at 0·9 μm each and 1 μL DNA extract, to a final volume of 30 μL. Cycling parameters for all assays consisted of an initial denaturation step at 94°C for 2 min, followed by 40 cycles of denaturation at 94°C for 20 s, annealing at 62°C for 20 s and extension at 68°C for 20 s, with fluorescence acquisition at the end of each annealing step.
All real-time PCRs were performed using a Rotor-Gene 6000 instrument (Corbett Research) set with an autogain optimization for each channel, which was done before the first fluorescence acquisition. The cycle threshold value (Ct) for each reaction was determined using the rotor-gene 6000 software v. 1.7 (Corbett Research), with the threshold line set manually. All samples were tested in triplicate, with the mean Ct value and standard deviation calculated. A positive result was determined by identifying the Ct value at which normalized reporter dye emission rose above background noise (determined empirically). If the fluorescent signal did not exceed the threshold for positivity within 40 cycles, the sample was considered negative. Negative (no template) and positive control samples for each assay target were included in each run.
The specificities of the multiplexed TYLCV-IL, ToLCV and TomIC assays were assessed by testing for cross-reactivity using 22 sequence-characterized reference begomovirus species originating from nine host plant species, including Solanum lycopersicum, S. melongena, Capsicum annuum, Gossypium hirsutum, Euphorbia pulcherrima, Abutilon striatum, Lonicera japonica, Cucumis sp. and Cucurbita pepo (Table 1). The specificities of the multiplexed B. tabaci MEAM1, MED and BtabIC assays were assessed by testing for cross-reactivity using a global collection of 26 sequence-characterized B. tabaci MEAM1, B. tabaci MED, B. tabaci AUS, B. tabaci AUS/INDO, B. afer and T. vaporariorium samples (Table 2). Bemisia afer and T. vaporariorium were included in the specificity analysis as they are commonly encountered during phytosanitary inspections in Europe, and both Europe and Australia, respectively. Furthermore, these whitefly species can be difficult to differentiate from B. tabaci using morphological characteristics by non-experts, or when embedded in the adhesive of a yellow sticky trap, and so the capacity to discriminate these species would improve the utility of the test.
Sensitivity and amplification efficiency
To assess the analytical sensitivities of the TYLCV-IL and ToLCV assays in singleplex and multiplex, a plasmid dilution series was used. To generate recombinant plasmids containing the complete real-time PCR amplicon sequence, DNA from TYLCV-IL isolate BRIP49036 and ToLCV isolate BRIP57502 were amplified in separate conventional PCRs using the respective real-time PCR primers (Table 3) under standard conditions (annealing temperature of 60°C). Amplicons were purified using the E-Gel CloneWell system (Invitrogen), and cloned into the pCR2.1 TOPO vector using the TOPO TA Cloning Kit (Invitrogen) and transformed into One Shot Electrocomp cells (Invitrogen) according to the manufacturer's recommendations. Plasmid DNA was purified using the QIAprep Spin Miniprep Kit (QIAGEN) and sequenced to verify the accuracy of cloned inserts. Each plasmid preparation was quantified using a NanoDrop spectrophotometer ND-1000 (NanoDrop Technologies) and the number of plasmid copies per microlitre was calculated using the following equation:
Tenfold serial dilutions of each purified plasmid preparation, ranging from 1 to 109 copies μL−1, were prepared in a background of healthy tomato DNA (1 in 500 dilution of neat extract), with each dilution tested in triplicate to generate the standard curves used to determine the analytical sensitivity of the begomovirus-specific assays (in singleplex and multiplex). To evaluate the lower limit of detection (LOD) in a tomato sample naturally infected with TYLCV-IL, six tenfold serial dilutions of total DNA extracted from TYLCV-IL isolate BRIP49046 were prepared ranging from 200 ng μL−1 to 200 fg μL−1 in a background of healthy tomato DNA (1 in 500 dilution), and tested in singleplex. For comparison of LODs, the same dilution series was then tested using the conventional PCR assay specific for TYLCV-IL (van Brunschot et al., 2010). To determine the LOD for the B. tabaci MEAM1 and B. tabaci MED assays, tenfold serial dilutions of DNA extracts of each species (10 whiteflies per extract) were prepared ranging from 500 ng μL−1 to 500 pg μL−1 and 300 ng μL−1 to 300 pg μL−1, respectively, and tested in singleplex and multiplex.
Linear regression analysis was performed on each of the virus and B. tabaci species-specific calibration dilution curves where the threshold cycle (Ct) values for each input amount of template were plotted as a function of the log10 concentration of the input amounts and a linear regression trend line was fitted to the data. The resulting slope of the fitted line was used to derive the PCR amplification efficiency (E) as a percentage, using the equation E = 100 × (10[−1/slope]−1), with E = 100 showing perfect amplicon doubling during the exponential phase of the reaction. Robust real-time PCR assays were considered to have amplification efficiency values between 80 and 100% whilst showing linearity over the range of dilutions tested (R2 coefficient ≥ 0·99).
Inter- and intra-assay reproducibility
To determine the inter- and intra-assay reproducibility of the assays, six tenfold dilutions representing six orders of magnitude of concentration were tested in triplicate on separate days to determine the coefficient of variation (CV), as a percentage, for the Ct values calculated for each dilution in the series. The Ct CVs were calculated as CV = 100 × standard deviation/mean. Data were analysed using a one-way analysis of variance (anova) with data grouped by runs.
Multiplexed analysis of field-collected B. tabaci samples
For the routine testing of field-collected B. tabaci samples, the bulking of samples increases the efficiency of testing and decreases costs. To provide recommendations for future users on the maximum number of B. tabaci samples that can be bulked without compromising detection of targets (leading to false negative results), the LODs of virus and vector in a bulked sample were determined. Bulking experiments were undertaken by combining various numbers and combinations of viruliferous B. tabaci MEAM1, virus-free B. tabaci MEAM1 and B. tabaci MED.
To evaluate the performance of the multiplexed real-time PCR assays on field-collected B. tabaci samples, three yellow sticky traps (Integrated Pest Management) were placed in a tomato farm (Darwin, Northern Territory) for a period of 2 weeks in August 2010. To simulate the capture of B. tabaci MED, stored specimens collected from Morocco (GenBank accession EF080822) were placed on yellow sticky traps and kept at room temperature for 2 weeks in the laboratory prior to analysis. The B. tabaci specimens captured on the yellow sticky traps were identified using a stereomicroscope (×10 magnification), removed using a sterile micropipette tip and transferred to a sterile microfuge tube. Prior to DNA extraction, residual adhesive was removed using the organic solvent hexane (Boonham et al., 2002). Each whitefly was rinsed once with 50 μL hexane then removed using a pipette. This rinsing process was repeated using 100 μL 100% ethanol, then sterile distilled water. The whitefly was then processed using the standard DNA extraction protocol, as described previously. For comparison, B. tabaci MEAM1 and B. tabaci MED samples collected using an aspirator were extracted using standard methods prior to analysis. To evaluate the effect of sticky trap collection and hexane treatment on quality and/or quantity of DNA recovered, the Ct values of six B. tabaci MEAM1 and six B. tabaci MED extracted using both the standard and hexane-mediated methods were compared and analysed using a two-way anova, with the DNA extraction treatment and B. tabaci species as factors. Two different Bemisia tabaci species were included in the analysis to reduce residual variance, with the hypothesis tested being whether any difference existed between the means of the two DNA extraction treatments.
The performance of the multiplexed real-time PCRs was also evaluated on field-collected B. tabaci samples of unknown identity. One hundred and twenty-two adult B. tabaci were collected using an aspirator from cotton plants grown on commercial farms in Queensland and the Northern Territory during regular surveys for B. tabaci MED. Bemisia tabaci samples were characterized for ‘biotype’ using a microsatellite profiling method that is currently used as the diagnostic protocol for the surveys (De Barro et al., 2003). For confirmation of B. tabaci identity, the same DNA extracts were then tested by multiplexed real-time PCR for B. tabaci MEAM1, MED and BtabIC.
Multiplexed analysis of field-collected tomato samples
To validate the assays, field-grown tomatoes displaying symptoms of leaflet chlorosis and curling, suggestive of TYLCD, were collected from Queensland and New South Wales and tested. Fifty-nine tomato samples were indexed using the routine conventional TYLCV-IL PCR method (van Brunschot et al., 2010), with the same DNA extracts tested by multiplexed real-time PCR for TYLCV-IL, ToLCV and TomIC.
Specificity of multiplexed real-time PCR assays
For the multiplexed begomovirus-specific assays, six sequence-characterized isolates of TYLCV-IL from tomato (Queensland, Australia), TYLCV-IL from two field-collected viruliferous whiteflies (Florida, USA) and three isolates of ToLCV from tomato were correctly identified using the relevant assays. No cross-reactivity was detected with either the TYLCV-IL- or ToLCV-specific assays when screened against non-target DNA extracts, including 15 other non-target begomovirus species (including the closely related begomoviruses TYLCV-Mild and Tomato yellow leaf curl Sardinia virus). The TomIC assay gave positive results for S. lycopersicum, S. melongena and Capsicum annuum, and negative results for G. hirsutum, E. pulcherrima, A. striatum, L. japonica, Cucumis sp. and C. pepo, demonstrating the specificity of this assay for members of the Solanaceae.
For the multiplexed B. tabaci species-specific assays, eight sequence-characterized specimens of B. tabaci MED were correctly amplified using the multiplexed MED assay. Nine sequence-characterized specimens of B. tabaci MEAM1 were amplified using the B. tabaci MEAM1 assay. No cross-reactivity was observed for either the B. tabaci MED or B. tabaci MEAM1 assays when screened against B. tabaci MEAM1 and B. tabaci MED DNA extracts, respectively, nor when further screened against the non-target specimens of B. tabaci, B. afer and T. vaporariorium. The BtabIC assay correctly amplified all B. tabaci samples tested, and did not cross-react with B. afer or T. vaporariorium samples.
Sensitivity of multiplexed real-time PCR assays
The tenfold dilution series of DNA extracts from target organisms were tested to determine the sensitivity and amplification efficiency of each assay. Standard curves showed high correlation coefficient values (R2 > 0·99) for both the singleplex and multiplexed data (Figs 1, 2). When compared with the singleplex assays, the LODs for all assays were unaffected by multiplexing. In general, when the assays were multiplexed the amplification efficiency decreased as a result of delayed Ct values, but, importantly from a diagnostics perspective, the LOD remained unaffected.
The analytical LOD was 10 plasmid copies for the TYLCV-IL assay (Fig. 1) and 100 plasmid copies for the ToLCV assay. For naturally infected samples of TYLCV-IL diluted in a background of healthy tomato DNA, the LOD was 200 fg total DNA, which was c. 1000 times more sensitive than the conventional PCR (Fig. 3).
For the B. tabaci species-specific assays, the LODs for the B. tabaci MEAM1 and MED assays were equivalent to six tenfold dilutions of the initial DNA extract, or 500 and 300 fg total DNA, respectively (Fig. 2).
Reproducibility of multiplexed real-time PCR assays
To evaluate the reproducibility of the multiplexed assays over a wide range of template concentrations, the intra- and inter-assay coefficients of variation (CV) were calculated using five tenfold dilutions of each assay target. The intra-assay CVs varied between 0 and 2·2% for the begomovirus-specific assays, with inter-assay CVs varying between 0·4 and 3·5%. For the B. tabaci species-specific assays, the intra-assay CVs varied between 0 and 1·2%, with inter-assay CVs of 0·2–4·4%. For all multiplexed assays, one-way anova showed no significant difference in Ct values of samples run over separate days (at a significance level of 0·05, data not shown).
Multiplexed analysis of field collected B. tabaci samples
One B. tabaci MED was reproducibly detected in a bulked sample containing up to 19 B. tabaci MEAM1 specimens. TYLCV-IL was reliably detected when one viruliferous B. tabaci MEAM1 was mixed with up to 14 non-viruliferous B. tabaci MEAM1 (Table 4).
Table 4. Effect of bulking treatments and hexane pretreatment of Bemisia tabaci collected using yellow sticky traps on the sensitivity and reproducibility of multiplexed real-time PCR detection
(vir) indicates viruliferous specimens carrying Tomato yellow leaf curl virus-Israel, as confirmed by conventional TYLCV-IL PCR; (vf) indicates virus-free specimens that originated from a laboratory-reared B. tabaci colony; YST indicates sample collected using yellow sticky trap and treated with hexane prior to DNA extraction.
All samples were run in triplicate with data presented as mean Ct (standard deviation,% coefficient of variation); − denotes not tested; Ct values below 40 indicate a positive result.
1 Bemisia tabaci MED + 1 B. tabaci MEAM1
23·02 (0·09, 0·40)
19·64 (0·06, 0·32)
21·41 (3·07, 14·33)
1 B. tabaci MED + 4 B. tabaci MEAM1
15·69 (2·18, 13·88)
17·70 (0·04, 0·24)
20·06 (1·27, 6·34)
1 B. tabaci MED + 9 B. tabaci MEAM1
14·54 (0·8, 5·50)
17·38 (0·00, 0·00)
18·84 (2·04, 10·81)
1 B. tabaci MED + 14 B. tabaci MEAM1
14·49 (1·27, 8·78)
17·32 (0·04, 0·24)
17·93 (0·88, 4·89)
1 B. tabaci MED + 19 B. tabaci MEAM1
14·51 (1·65, 11·40)
17·52 (1·00, 5·69)
16·81 (0·57, 3·37)
1 B. tabaci MEAM1 (vir) + 1 B. tabaci MEAM1 (vf)
28·14 (0·10, 0·36)
1 B. tabaci MEAM1 (vir) + 4 B. tabaci MEAM1 (vf)
19·75 (0·30, 1·50)
1 B. tabaci MEAM1 (vir) + 9 B. tabaci MEAM1 (vf)
19·96 (1·67, 8·37)
1 B. tabaci MEAM1 (vir) + 14 B. tabaci MEAM1 (vf)
24·62 (0·94, 3·82)
1 B. tabaci MEAM1 (vir) + 19 B. tabaci MEAM1 (vf)
1 B. tabaci MED (aspirator)
17·41 (0·59, 3·41)
20·64 (0·84, 4·09)
1 B. tabaci MED (YST; hexane)
17·06 (1·35, 7·91)
20·63 (0·62, 3·02)
1 B. tabaci MEAM1 (aspirator)
16·71 (1·19, 7·12)
21·23 (1·24, 5·84)
1 B. tabaci MEAM1 (YST; hexane)
15·98 (0·51, 3·22)
21·06 (1·36, 6·44)
Within a two-way anova, a test for potential differences between the mean Ct values obtained from either B. tabaci MED or MEAM1 collected on yellow sticky traps and processed using hexane, as compared to aspirator-collected individuals, showed no significant differences (P = 0·19, F1,14 = 1·866). These results demonstrate that hexane treatment did not alter the sensitivity of detection. Moreover, reproducible multiplex real-time PCR results were obtained using samples extracted using both the standard extraction method and the hexane-mediated extraction method (all CV values <8·0%).
For the detection of whiteflies of unknown identity, 122 B. tabaci specimens from various cotton-growing regions in Australia were tested using the B. tabaci MEAM1, MED and BtabIC assays. The real-time PCR results, as well as the microsatellite identifications, are shown in Table 5. Of the 32 B. tabaci samples from the Northern Territory, nine were identified as B. tabaci MEAM1. The remaining 23 samples tested negative using both MED and MEAM1 assays, and tested positive using the BtabIC assay. However, five of these latter individuals were identified as B. tabaci MED using the microsatellite method. To resolve this discrepancy, confirmatory mtCOI sequence analysis of three representative samples from this group (GenBank accessions JX416166, JX416167 and KC109796) was performed. All three samples showed a high level of sequence similarity (>99%) to each other, but >12% divergence from B. tabaci AUS/INDO and averaged >3·6% divergence from B. tabaci AUS. These results suggest that this population may be a different putative species, referred to here as B. tabaci Australia II (AUSII). Furthermore, the sequence data confirmed the real-time PCR results. Of the 90 samples from Queensland and New South Wales, all samples tested positive in the B. tabaci MEAM1 assay, and there was complete agreement between the real-time PCR and microsatellite results.
Table 5. Comparison of multiplexed real-time PCR and microsatellite PCR typing methods for the identification of field-collected Bemisia tabaci and virus-infected tomato samples from Australia
Multiplexed analysis of field-collected tomato samples
Overall, there was 100% correlation of results for the detection of TYLCV-IL using the conventional and real-time PCR methods. Three tomato samples from Mareeba (Queensland) and two samples from Murwillumbah (New South Wales) tested positive in the TYLCV-IL and TomIC assays. Of the 54 samples from Bowen (Queensland), 20 of the samples tested positive for TYLCV-IL and all 54 were positive in the TomIC assay. For the remaining 34 samples that tested negative for TYLCV-IL, additional PCR analysis revealed the presence of Potato leafroll virus, which is known to occur in the same production area and displays similar symptoms to TYLCD (data not shown).
In this study, a panel of quantitative multiplexed real-time PCR assays including four species-specific assays (TYLCV-IL, ToLCV, B. tabaci MEAM1 and B. tabaci MED), in conjunction with two internal control assays for both plant (TomIC) and B. tabaci samples (BtabIC), were developed and evaluated for use in TYLCD epidemiological and disease management applications. The internal control assays co-amplify host sample DNA (either plant or whitefly), enabling the monitoring of the entire diagnostic process from extraction to real-time PCR result to exclude false negative results. The flexible panel of multiplexed real-time PCRs was designed to be used interchangeably, contingent upon the sample being tested, with up to three assays (triplex) capable of being combined in one tube.
The analytical specificity of the multiplexed real-time PCR assay panel was excellent, with no cross-reactions observed for any assay when screened against a large panel of sequence-characterized non-target DNA extracts. Furthermore, TYLCV-IL was readily detected in B. tabaci samples from Queensland (Australia), Guadeloupe (France) and Florida (USA).
For each assay it was shown that the sensitivity and LOD were not compromised by performing the assay in multiplex, a significant result considering the principal purpose of the assays is to determine the presence or absence of the target organism. The analytical LODs of the multiplexed TYLCV-IL and ToLCV assays (10 and 100 plasmid copies, respectively) were comparable to LODs reported for other plant pathogens based on absolute standard curves of plasmid DNA diluted in a background of healthy plant DNA (Chandelier et al., 2010). A direct comparison of the LOD of the TYLCV-IL real-time PCR and conventional PCR demonstrated the superior sensitivity of the real-time PCR method (1000 times more sensitive). The LOD of the multiplexed B. tabaci MEAM1 and MED assays (500 and 300 fg DNA, respectively) were similar to published reports for the real-time PCR identification of insects (Huang et al., 2010). These results demonstrate the usefulness of the multiplexed assays for the detection of targets in low abundance and will permit increased bulking of plant samples prior to DNA extraction and testing, to greatly improve the throughput of testing without the risk of false negative results.
The results clearly showed that the real-time PCR assays performed consistently and reproducibly over a large dynamic range of input template concentrations, even when multiplexed and performed on different days. High correlation coefficient values for all assays (R2 > 0·99 for all assays in singleplex and multiplex) indicated a linear response over all dilutions. These results verified the suitability of the assays for quantitative analysis and highlighted the excellent amplification efficiencies for all dilutions (even at very low template concentrations). However, for all assays except B. tabaci MEAM1, amplification efficiencies decreased (from 1·3 to 15·4%) when multiplexed. Also, for all assays, the inter- and intra-assay coefficients of variation increased when run in multiplex. Taken together, these results indicate that the reaction efficiencies were reduced when the assays were multiplexed, probably as a result of competition for reaction components when more than one target is amplified in a single tube. Therefore, it is recommended that assays should be performed in singleplex for quantitative studies, e.g. virus-resistance screening in plant breeding programmes, to ensure the accuracy of quantification.
New methods were developed and evaluated to improve the efficiency and throughput when testing large numbers of field-collected B. tabaci samples. Bulking treatments were investigated, where multiple B. tabaci samples were combined prior to DNA extraction. In view of the results, it is recommended that 20 B. tabaci at most can be bulked prior to DNA extraction when screening for B. tabaci identity, and up to 15 B. tabaci when screening for TYLCV-IL-carrying viruliferous B. tabaci. It is possible that the LOD could be improved if the methodology was modified to physically accommodate larger numbers of whiteflies. For whiteflies collected using yellow sticky traps, the modified DNA extraction protocol that incorporated a hexane pretreatment (for adhesive removal) prior to DNA extraction was shown not to alter the sensitivity of detection. This method should prove useful for testing B. tabaci samples collected using yellow sticky traps, a simple and common trapping method used in both open-field and glasshouse cropping.
The performance of the multiplexed real-time PCR panel was evaluated successfully in a ‘blind’ study using field-collected B. tabaci of unknown identity and suspected virus-infected plant specimens. Importantly, some B. tabaci AUSII samples from Katherine (Northern Territory) were incorrectly identified as B. tabaci MED using the microsatellite profiling method. These samples tested negative for B. tabaci MED using the real-time PCR method, with their identity resolved by mtCOI sequence analysis, confirming the real-time PCR results. It was later determined that the microsatellite method gave polymorphic results for B. tabaci AUSII populations and therefore an alternate method of identification was required. The multiplexed real-time PCR assays provided a reliable, independent method for differentiating B. tabaci MEAM1, MED and indigenous B. tabaci species. These results emphasize the importance of confirmatory independent testing when potential B. tabaci MED samples are identified in Australia.
This is apparently the first time that a diagnostic assay has been developed that allows the simultaneous identification of the viruses and vectors responsible for TYLCD epidemics in a region. The panel of assays provides the user with greater flexibility for choosing targets to test, and provides unambiguous results for the presence/absence of each target. The inclusion of internal controls for plant and whitefly sample substrates represents a significant improvement for the reliable detection of targets. Also, the closed-tube system decreases the risk of contamination by limiting post-PCR processing. Furthermore, this real-time PCR panel increases Australia's preparedness for detecting new incursions of exotic begomoviruses and B. tabaci species, facilitating confident testing for and exclusion of endemic tomato-infecting begomoviruses and also differentiating endemic B. tabaci from the exotic B. tabaci MED, thereby decreasing the risk of an undetected incursion. This flexible panel of assays will provide end users with a unique system for indexing/screening samples, amenable to high-throughput screening, to aid in quarantine surveillance and monitoring programmes. Finally, the quantitative virus-specific assays, particularly when applied in singleplex, are well suited for use in plant virus resistance breeding programmes. Future work could explore the potential for adapting the six assays described in this study to a true multiplex platform employing a highly paralleled PCR architecture, such as TaqMan array cards or PCR-Luminex bead arrays (Dunbar, 2006; Rachwal et al., 2012). These platforms could enable expansion of the current assay panel to incorporate tests for other plant-infecting viruses, whiteflies and a generic plant internal control for monitoring amplification from nonsolanaceous host plants.
The authors would like to acknowledge the support of the Cooperative Research Centre for National Plant Biosecurity, established and supported under the Australian Government's Cooperative Research Centres Program. We thank the following researchers for their kind provision of samples: Z. Hall, B. Conde, C. Pearce, P. Campbell, E. Moriones, P. Stephens, P. Chiemsombat, S. K. Green, K. Zanic, M. Cahill, A. Hanafi, M. Muniz, S. Abd-Rab and A. Hulthen. We thank D. Pagendam (CSIRO Mathematics, Informatics and Statistics) for statistical advice, and the anonymous reviewers for valuable comments and suggestions.