A diagnostic LAMP assay for rapid identification of an invasive plant pest, fall armyworm Spodoptera frugiperda (Lepidoptera: Noctuidae)

Fall armyworm (FAW), Spodoptera frugiperda (Lepidoptera: Noctuidae), is a highly polyphagous invasive plant pest that has expanded its global geographic distribution, including recently into much of Australia. Rapid diagnostic tests are required for identification of FAW to assist subsequent management and control. We developed a new loop-mediated isothermal amplification (LAMP) assay based on the mitochondrial cytochrome c oxidase subunit I (COI) gene for accurate and timely diagnosis of FAW in the field. The specificity of the new assay was tested against a broad panel of twenty non-target noctuids, including eight other Spodoptera species. Only S. frugiperda samples produced amplification within 20 min, with an anneal derivative temperature of 78.3 ± 0.3 °C. A gBlock dsDNA fragment was developed and trialled as a synthetic positive control, with a different anneal derivative of 81 °C. The new FAW LAMP assay was able to detect FAW DNA down to 2.4 pg, similar to an existing laboratory-based real-time PCR assay. We also trialled the new FAW assay with a colorimetric master mix and found it could successfully amplify positive FAW samples in half the time compared to an existing FAW colorimetric LAMP assay. Given the high sensitivity and rapid amplification time, we recommend the use of this newly developed FAW LAMP assay in a portable real-time fluorometer for in-field diagnosis of FAW.


Results
Molecular variation. The panel of species tested included eight non-target Spodoptera species which were between 4.3%, and 7.4% divergent (5′-COI, uncorrected p-distances) from FAW, while the other twelve Noctuidae species tested were from 9.3 to 11.2% divergent from FAW (Fig. 1). All new target and non-target new DNA barcode sequences obtained in the current study have been submitted to GenBank (OL539263 -OL539329). (Table 1) were developed to target a 249 bp portion of the FAW COI locus (3′ region) which has been shown to be highly variable in numerous Spodoptera species (Fig. 2). Ambiguous bases were added to primers (Fig. 2, Table 1) to account for genetic diversity present in the wider COI dataset of FAW individuals available on GenBank (accessed Dec 2020). Six primers were employed in the FAW LAMP assay, two inner primers (FIP and BIP) and two outer primers (F3 and B3).  (Table 3, Fig. 3). Amplification in less than 20 min was considered as positive. The specificity of the FAW LAMP assay was validated against a broad range of non-target taxa ( Table 2, Fig. 1), with no off-target amplification observed within 20 min (Table 3). A small degree of off-target amplification was observed when reactions were run past 20 min (Table 3), hence the recommended cut-off time for positive amplification is 20 min. LAMP amplification was not sensitive to the DNA extraction method employed, with in-field compatible DNA extractions providing results consistent with laboratory DNA extractions (Table 4).

FAW LAMP assay design and optimisation. LAMP primers
Kim et al. 37 LAMP assay results with published primers. The first set of primer master mix tested for this FAW LAMP assay on the Genie III using 1:6:3 ratio including all six primers (F3/B3, FIP/BIP and Floop/ Bloop) and 65 °C amplification temperature, amplified all DNA samples tested, including the negative control, within one minute confirming a strong primer dimer ( Supplementary Fig. 1a). The second set of primer master mix tested for this FAW LAMP assay on the Genie III using 1:8:2 ratio of five primers, as recommended by Kim, et al. 37 (F3/B3, FIP/BIP and Bloop no Floop), and an amplification temperature of 61 °C resulted in no amplification after a 25 min reaction time. (Supplementary Fig. 1b).
Detection sensitivity of gBlock DNA fragment. The detection sensitivity of the FAW 252 bp gBlock dsDNA fragment (Table 1) was tested using ten-fold dilutions ranging from ~ 100 million copies to ~ 10 copies www.nature.com/scientificreports/ in LAMP reactions (Fig. 4), with positive detections found as low as ~ 10 copies within 20 min (Fig. 4a). The gBlock anneal derivative peak occurred at 81 °C (Fig. 4b). From the amplification profile it was calculated that one hundred thousand copies (10 5 ) of gBlock DNA equates to less than 0.4 ng/µL of FAW DNA (Fig. 4c). The anneal derivative of LAMP amplicons exhibited two peaks, with the FAW DNA peak present at 78.5 °C and the gBlock DNA peak present at 81 °C (Fig. 4d). Alternatively, 10 6 gBlock could be used as positive control which was found to amplify within 10-11 min.   5) was found to take significantly longer than the standard OptiGene reagents (which amplify within 20 min). Of six species tested in the colorimetric assay for the new FAW LAMP assay, only S. frugiperda and gBlock dilution 10 6 produced a positive colour change in less than one hour, further demonstrating the robustness of our assay (Fig. 5). Optimal colorimetric results were achieved in 45 to 60 min for this assay. A small degree of non-target amplification was observed when the assay was run for longer (i.e., up to 90 min). Following the Kim, et al. 37 protocol, amplification using the colorimetric master mix (at 61 °C) was found to take approx. double the time to produce positive amplification (> 105 min). Of the six species tested in the colorimetric assay for the FAW LAMP assay, S. frugiperda produced a positive colour change in about 105 to 120 min ( Supplementary Fig. 2). A small degree of non-target amplification was observed when the assay was run for longer (i.e., up to 165 min).
Sensitivity of LAMP and real-time PCR assay. The sensitivity of the new FAW LAMP assay (Fig. 6a) and real-time PCR assay (Fig. 6b) were tested using a four-fold serial dilution of DNA for biological replicates of late-instar larval thoracic leg. A four-fold serial dilution of DNA for biological replicates of an adult moth leg was tested with the new FAW LAMP assay (Fig. 6c) and real-time PCR assay (Fig. 6d). The results from both assays were very similar, both proving to be very sensitive. Both LAMP and real-time PCR assays was able to detect FAW DNA from larvae down to the lowest dilution tested 2.4E−03 ng/µL (equal to 2.4 pg) (Fig. 6a,b). Both LAMP and real-time PCR assays was able to detect FAW DNA from adult moths down to 5 out of 8 dilutions only 3.9E−03 ng/µL (equal to 3.9 pg) (Fig. 6c,d). Both LAMP and real-time PCR results produced similar results as the starting DNA amount for a late-instar larval thoracic leg was found to be approx. 40 times higher than from an adult moth leg.

Discussion
This study reports on a LAMP assay for rapid and reliable in-field detection of fall armyworm (FAW, Spodoptera frugiperda), an invasive noctuid pest at both adult and larval stages. Our assay has been shown to be speciesspecific, when tested against a panel of twenty commonly encountered noctuids, including approx. a third of all known Spodoptera species (i.e. 9 of the 31 species 39 ). Our primers were capable of amplifying positive FAW DNA in under 10 min, with amplification within 20 min considered positive. We recommend that extended LAMP amplification not be performed, as a small degree of off-target amplification was observed when reactions were run past 20 min. We also designed and optimised a synthetic DNA positive control (gBlock dsDNA fragment) for use in our FAW LAMP assay. This gBlock is beneficial in providing a consistent control to allow tracking of the performance of LAMP assays across runs and provides confidence that positive amplification of samples is not due to contamination, as the gBlock DNA has a different anneal derivative temperature compared to FAW DNA. The new FAW LAMP assay described here has also been shown to perform well using the technologically simpler colorimetric approach, with amplification of positive FAW DNA occurring in less than an hour.
In its native New World geographic range, the FAW is widely considered to consist of either the corn or the rice host-preferred strains 20 . At the whole genome level however and based on the widely applied partial mitochondrial COI nucleotide distance estimates, these rice and corn-host strains could potentially be regarded as two closely related sister species 40,41 . However, whole genome analyses of invasive populations in the Old World Table 1. FAW COI LAMP primer and amplicon sequences (gBlock) and parameters. The F2 and B2 primer regions of FIP and BIP are underlined. Lowercase letters in the gBlock indicate extra "ccc" or "ggg" added between LAMP primer sites to increase the overall Tm of the amplicon.  TAG TTG CTC  ATTTCgggCAC TAT GTT TTA TCA ATA GGA  GCT GcccGCT ATT TTA GGT GGA TTT ATT  CAC TGgggCCA TTA TTT ACT GGA TTA TCT  TTA AATCCgggCCT TAT ATA TTA AAA ATT  CAA TTT TTT ATT ATA TTT ATCcccGGA  GTA AAT TTA ACT TTC TTC CCA gggTTT  AGG ATT AGC AGG TAT ACC TCG cccTGA  TTA TCC TGA TTC TTA TAT  showed FAW to consist of admixed genome signature overall, indicating that these were generally hybridised populations [42][43][44] . Our LAMP assay was developed to accurately identify all FAW, incorporating the range of COI DNA sequence variation known from S. frugiperda, sensu lato, regardless of their rice / corn host preferences, or if they represented hybridised individuals. While the FAW LAMP assay developed here showed rapid and robust confirmation of FAW regardless of host strains or hybrids, the DNA data generated also unexpectedly revealed incongruency in the taxonomic status of some non-target Noctuidae species. For example, our DNA sequencing of S. exigua revealed that the    Fig. 1) and likely represent two discrete cryptic species, despite both matching with 100% similarity to different reference sequences identified as S. exigua in the Barcode of Life DNA (BOLD) and GenBank databases. Likewise, the generic designation of Leucania loreyi, which is commonly erroneously placed within the genus Mythimna in many studies, is also unclear 45 . When the geographic distribution of both L. loreyi and M. loreyi are combined this species is very widespread (GBIF, accessed 20-Sept 2021), however confirmation of the taxonomic status for this widespread Old World pest is required to afford confidence in the LAMP assay which has recently been developed for M. loreyi 46 .
For both pre-border interception and in-field applications, all FAW life stages are expected to be encountered, including egg-masses and early-instar larvae. Laboratory testing of the new LAMP assay on such samples was not possible here, as they were not available. However, other molecular approaches, such as PCR-RFLP which has been used to differentiate Helicoverpa species 47 , have been shown to be effective on these early life stages. Additionally, LAMP assays for other insects have been shown to work on all lifestages 30,31 and given the sensitivity of our FAW LAMP assay (i.e., down to ~ 10 copies of gBlock DNA fragments and 2.4 pg of FAW DNA) it is anticipated that the assay outlined here will also be able to accurately identify these early life stages of FAW.
The comparison of the published Kim, et al. 37 LAMP assay revealed that our new assay is the most suitable for in-field use, being capable of producing results more rapidly in a portable real-time fluorometer, using appropriate commercially available reagents, or using the alternative colorimetric approach. As LAMP amplification times are influenced by both amplicon length and the presence of loop primers 48 , it is likely that our new assay produces such rapid amplification times (approx. 10 min) due to it consisting of a relatively small amplicon fragment and Table 3. Performance of the FAW LAMP assay using the optimised primer ratio 1:6:3.   38 also does not use loop primers, but was not directly compared in our study. This latter assay has also been shown to be capable of rapidly amplifying FAW DNA using the GenieIII, and has been tested against larvae of eleven non-target Noctuidae, including two Spodoptera species to-date 38 . Prevention of on-going introduction of novel economically significant traits in new invasive pest populations is a biosecurity priority 2, 49 . FAW has been shown to have diverse insecticide and Bt resistances 10,50,51 . Currently, there are no rapid in-field molecular tools available to screen for resistance in FAW, although these may be developed in the future. For now, our LAMP assay provides a new tool to aid in monitoring FAW incursions, through providing rapid, accurate species identification, using simple protocols for DNA extraction and LAMP amplification including a gBlock positive control, being capable of being performed on a portable real-time fluorometer in the field.

Materials and methods
Specimens examined. Adult and larval specimens of the FAW target species, Spodoptera frugiperda (n = 19 individuals), and non-target Noctuidae species (n = 20 species, n = 49 individuals), were examined in this study ( Table 2). All species identities were confirmed through DNA barcoding of the mitochondrial COI locus (5′-region) following standard laboratory procedures 52,53 . DNA extractions. DNA was extracted from single (thoracic) legs removed from adult and late-instar larval Noctuidae specimens (Table 2), in the laboratory using the DNeasy Blood and Tissue extraction kit (Qiagen, USA); the ISOLATE II Genomic DNA Kit (Bioline, UK); and the 5% Chelex 100 (BioRad, USA) extraction  In-field compatible extraction procedures employed were either the QuickExtract™ (QE) solution 1.0 (Epicentre Biotechnologies, USA) (as per Blacket, et al. 30 ), or the Xtract (Xt) DNA extraction solution (GeneWorks, Australia) as follows: 50 µL of extraction solution was pipetted into each well of an 8-well Genie strip (OptiGene, UK) with one leg of FAW adult moth (n = 3) and one leg of FAW larva (n = 3) and incubated in the Genie III at 65 °C for 6 min, followed by 2 min at 98 °C 30 , and then kept on ice for > 1 min. LAMP primers were manually designed by eye to target eight DNA regions from a COI reference alignment (3′-COI region), including twenty-three Spodoptera species (from Kergoat, et al. 55 Nagoshi, et al. 20 ). The reference alignment was compiled from existing DNA sequences of Spodoptera species available on GenBank (accessed Dec 2020), to the mitochondrial genome from Kim, et al. 37 . The 3′-COI region was used for primer design as it was found that there were a larger number of DNA sequences available for this region than the standard 5′-COI DNA barcoding region for FAW and related Spodoptera species. The 3′-COI region has also been previously used for identification of host-specific strains within FAW 20 . For all primers, the GC content (%), predicted melting temperature (Tm), and potential secondary structure formations (hairpins or dimers) were analysed using the Integrated DNA Technologies (IDT) online OligoAnalyzer tool (https:// sg. idtdna. com/ calc/ analy zer), using the qPCR parameter sets. Complete sets of LAMP primers were analysed together to detect potential primer dimer interactions using the Thermo Fisher Multiple Primer Analyzer tool (www. therm ofish er. com). Primers were synthesised by Sigma (Australia).
LAMP assay optimisation. Optimisation was performed following the protocols previously outlined in Blacket, et al. 30 , which include testing multiple primer ratios to obtain optimum amplification time and a consistent anneal derivative temperature. Primers F3 and B3 were used at 10 µM concentration, whilst FIP, BIP, Bloop and Floop were used at 100 µM concentration. The primer master mix was prepared to a ratio 1:6:3 by adding 10 µL of F3 and B3; 6 µL of FIP and BIP; 3 µL of Bloop and Floop; and 62 µL of ultrapure water, for a total volume of 100 µL. Each LAMP reaction mix was made by adding 10 µL of primer master mix to 14 µL of Isothermal Master Mix (ISO-004, OptiGene, UK) and 1 µL of template DNA into each well of the Genie strip (25 µL total reaction volume). Each run included a positive control (i.e., known FAW DNA, VAITC 10707), a no-template negative control, and six test samples.
LAMP assays were run in the Genie III at 65 °C for 25 min followed by an annealing curve analysis from 98 to 73 °C with ramping at 0.05 °C/s, visualised in the blue channel. LAMP runs taking approx. 35 min in total. The run date, Genie III serial number and the run number of each LAMP assay completed on the machine were recorded to allow run files to be transferred and analysed using a PC version of the software Genie Explorer version 2.0.7.11. The amplification and anneal derivative curves were visualised on the Genie III screen to ensure www.nature.com/scientificreports/ that amplification occurred as expected, with positive amplification plots showing an 'S' shaped sigmoid curve reflecting the increase in fluorescence detected, and negative results staying relatively flat. Positive results were further confirmed through performing the annealing step which results in a single product peak at a specific temperature.
Validation of Kim, et al. 37 primers on Genie III. The first set of primers tested were prepared in a master mix that included all six primers F3/B3, FIP/BIP and Floop/Bloop to a ratio of 1:6:3. The reaction was run on the Genie III using the OptiGene mastermix (ISO-004, OptiGene, UK) at 65 °C for an extended time of 35 min. The second primer set (colorimetric assay primers) tested consisted of only five primers F3/B3, FIP/BIP and Bloop (i.e. no Floop) at a ratio of 1:8:2 (as per Kim,et al. 37 ). The reaction was run on the Genie III using the OptiGene mastermix at 61 °C for 25 min. Both 5 and 6-primer sets were tested in two LAMP runs with OptiGene reagents as mentioned above using DNA templates in well 1 to 8 as listed sequentially. The first well containing DNA of the target species (1) S. frugiperda, (2) Spodoptera litura (Papua New Guinea, PNG), (3) Spodoptera exigua, (4) Helicoverpa armigera conferta, (5) Mythimna convecta, (6) Leucania loreyi, (7) no-template negative control and (8) FAW gBlock dilution 10 6 .
Evaluation and design of a gBlock dsDNA fragment for the FAW LAMP assay. We designed a gBlock dsDNA fragment (Integrated DNA Technologies, Iowa, USA) for use as a positive control for the FAW LAMP assay. This DNA fragment consisted solely of concatenated LAMP primers separated by runs of "ccc" and "ggg" to increase the overall Tm of the fragment, and therefore, a different annealing derivative when compared to positive FAW samples.
Sensitivity of the LAMP assay was tested using the serially diluted gBlock DNA. The copy number calculation and ten-fold serial dilution (1:10) of the gBlock was prepared as outlined in 31 . Serial dilutions ranged from ~ 100 million copies down to ~ 10 copies (10 8 copies to 10 copies) and were run in the Genie III following the same FAW LAMP protocol as previously mentioned. Following this, a second FAW LAMP run was conducted to determine the best dilution to be used as a positive control. The four-fold serial dilution of laboratory extracted FAW larva DNA (VAITC 10726) (40 ng/µL to 0.0391 ng/µL) was used as DNA template to compare with one hundred thousand copies (10 5 ) of gBlock DNA. The amount of FAW DNA was then equated from the amplification time of FAW gBlock dilution 10 5 .

Colorimetric FAW LAMP assay with new and published primers.
We also tested our FAW LAMP assay primers using an alternative colorimetric LAMP master mix (WarmStart Colorimetric LAMP 2 × master mix, DNA & RNA, New England Biolabs Inc.) following published protocols 32 .
The volume of primer master mix in each reaction mixture was initially optimised using 2.5, 5 and 10 µL of primer mix. Ten microlitres of primer mix added to the reaction was able to produce colour change from pink to yellow in the shortest time. The reactions were set up in a 25 µL reaction volume and included 12. We ran both assays side-by-side for comparison, showing a clear timeline for each test. The tubes were incubated on a heat block at 65 °C (our assay) and 61 °C (Kim,et al. 37 assay) and the colour change was monitored by photographing with a Canon 5D digital SLR camera every 15 min for 90 min (our assay) and 165 min (Kim,et al. 37

assay).
Analytical sensitivity of the FAW LAMP assay compared to real-time PCR 28 . A four-fold serial dilution (1:4) of biological replicates of FAW late-instar larval thoracic leg (reference specimens VAITC 10707 and 10726) and FAW adult moth leg (reference specimens VAITC 10728 and 10729) laboratory extracted DNA was prepared using Ultrapure water (Invitrogen, Life Technologies, Australia). Starting DNA concentration was quantified using a Qubit 2.0 Flourometer (Invitrogen, Life Technologies, Australia) following manufacturers protocol. The DNA sample was serially diluted from 40.0 to 2.441 × 10 -3 ng/µL (1:1 to 1:16,384) for the FAW larval DNA and 1.0 ng/µL to 6.1 × 10 -5 ng/µL (1:1 to 1:16,384) for the FAW adult DNA. Sensitivity of the LAMP assay was tested for each of the four samples using the eight serially diluted DNA samples in the Genie III, following the same assay conditions as described above. The time of amplification and anneal derivative temperature was recorded for all samples.
The same serial dilution of DNA from above was compared for FAW DNA sensitivity using a real-time PCR assay. The primers and probe set (Sigma) and cycling conditions used were as published in 28 except that the primer concentration was increased from 0.3 µM to 0.5 µM and probe concentration from 0.1 to 0.2 µM for optimum amplification. Real-time PCR was performed in QuantStudio 3 Real time PCR system (Thermo Fisher Scientific) in a total volume of 25 µL with technical replicates for each dilution. Each reaction mixture included 12.5 µL GoTaq Probe qPCR mastermix (Promega), 0.5 µM of each forward and reverse primers, 0.2 µM Taqman probe, 4 µL of template DNA and made up to 25 µL with RNA-free water. A non-template control with 4 µL of water instead of DNA was included in each run to check for reagent contamination. The PCR thermal cycling conditions consisted of a one-step denaturation: 2 min at 95 °C, followed by 40 cycles of amplification in www.nature.com/scientificreports/ a two-step procedure: 95 °C for 15 s and 60 °C for 1 min. The average Cq value (cycling quantification value) of the eight dilutions was recorded for comparison with the amplification time from the LAMP assay.

Data availability
GenBank, accession numbers OL539263 -OL539329. www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.