A loop-mediated isothermal amplification (LAMP) assay for the detection of Cryptotermes brevis West Indian drywood termite (Blattodea: Kalotermitidae)

Cryptotermes brevis is one of the most destructive invasive termites in the subtropics and tropics and is a common biosecurity intercept at the Australian border. Drywood termite species are cryptic and difficult to identify morphologically in situations when soldiers or imagos are unavailable. We developed a novel DNA based loop-mediated isothermal amplification (LAMP) assay to detect C. brevis and differentiate it from other drywood termites. Validated voucher specimens of 30 different drywood termite species were obtained from several insect collections from which DNA was extracted and amplified. The amplicons containing partial mitochondrial 16S rRNA were sequenced and a DNA database was created from which C. brevis LAMP primers were developed, optimized, and tested. The assay was assessed against a range of target and non-target species and found to be specific, successfully amplifying the target specimens of C. brevis in under 30 min. Amplification success was variable against C. brevis faecal pellets due to minute, unmeasurable or degraded DNA. This LAMP test is a new tool for the rapid detection of C. brevis that will enable faster and less destructive management of drywood termite infestations.

are sent for fumigation to prevent pest spread and establishment. There is a need for a fast accurate molecular technique to distinguish C. brevis from other non-quarantinable and quarantinable drywood termites.
To date, the only existing molecular tool for the identification of Cryptotermes species are polymerase chain reaction (PCR) and DNA sequencing methods. These rely on specimens from which good quality DNA can be extracted, and the availability of sequences on public databases such as GenBank and Barcode of Life Data (BOLD). Time to identification can be several days depending on commercial sequencing services. Loop-mediated isothermal amplification (LAMP) is an alternative molecular tool that can be species-specific and deliver a result withing 30-60 min. An additional potential benefit to LAMP-based techniques is the identification of termites from faecal pellets. Pellets may contain minute amounts of termite DNA, although extracting the DNA and subjecting it to PCR is problematic. This is due to a lack of sensitivity of current PCR methods to the amount of target DNA present in pellets, high levels of PCR inhibitors, and DNA contamination from micro-organisms present in pellets. LAMP is a DNA-based method using loop-mediated isothermal amplification, which is robust, portable, simple, rapid, taxon specific and less sensitive to contaminants. It is therefore suitable for pest detection in the field and does not require specialist molecular technicians to operate. The advantage of the LAMP method over PCR and DNA sequencing is its high specificity due to the use of six primers pairs, less sensitivity from sample inhibitors, and simplicity. LAMP tests have been used for the detection of invasive insect pests [14][15][16] and diseases [17][18][19] . The primary objective of this study was to develop a specific LAMP assay for the detection of C. brevis and their pellets.

Results
DNA extraction. The average yield of total DNA per individual termite was 3.96 ± 1.09 (SD) ng/µL, range 1.03-7.71 ng/µL. The DNA concentration from pellets was very low ranging from 0.07 to 0.3 ng/µL or was not measurable (Table 1).

Development and assessment of the C. brevis LAMP assay. DNA sequence analysis and LAMP
primer design. Novel LAMP primers ( Table 2, Fig. 1a, b) were developed for the detection of C. brevis based on the DNA sequences of the partial mitochondrial 16S rRNA gene. The gene was successfully amplified across all Cryptotermes species used in the study (Supplementary Table S1). Based on phylogenetic analysis, none of the Cryptotermes species were very close to C. brevis (Fig. 2). The DNA sequence similarity between C. brevis and other Cryptotermes species ranged from 79 to 88%. The DNA sequences of the three most common intercepted species associated with wood in service that are morphologically similar to C. brevis were different: C. domesticus (81% identity), C. dudleyi (82%), Incisitermes minor (75.6%) (Supplementary Table S2). This gives confidence in the identification of these species and in the species-specific LAMP primers developed. The C. brevis LAMP assay consists of six primers, the outer forward primer F3, the outer reverse primer B3, the inner forward primer FIP, the inner reverse primer BIP, the reverse loop primer LoopF and the forward loop primer LoopB ( Table 2). The optimal primers ratio was (F3/B3: FIP/BIP: LoopF/LoopB) 10: 1: 2. In addition, a 203 bp gBlock dsDNA fragment was created for use as synthetic DNA positive control for C. brevis LAMP assay (Table 2).
In-silico analysis. The phylogenetic analysis did not reveal any closely related Cryptotermes species to the target species C. brevis. In-silico analysis of the primer binding sites indicated specificity exclusively for C. brevis with none of the other 30 species tested indicating potential for binding. Therefore, the LAMP assay was tested in-vitro on the target species and three other morphologically similar and commonly intercepted species at the Australian border. The test included six representative termite specimens of C. brevis, two C. domesticus, two C. dudleyi, and one I. minor. The assay was also tested on the pellet samples of these species (Table 1).
Specificity and sensitivity of LAMP assay. The specificity of the LAMP assay was tested on six representatives of C. brevis (alates, soldiers and pseudergates) and two morphologically similar species (as pseudergates or imagos): two representatives of C. domesticus, two of C. dudleyi; and one representative of the herterogeneric I. minor. The primers were also tested on the pellets of all three species (four C. brevis including positive control, and one of each pellet of; C. domesticus, C. dudleyi, I. minor).
The C. brevis LAMP assay detected all six tested C. brevis insect samples. The assay produced amplification of the target species in less than ≤ 30 min (Mean value 15.0 min, Standard Deviation ± 4.0) at anneal derivate 79.9-80.9 °C (Mean value 80.6 °C, Standard Deviation ± 0.4) ( Table 1, Fig. 3a, b, Supplementary Table S3). Amplification peaks were tall and regular for four samples indicating efficient amplification. Two positive samples (HI 49.0-red peak and 10-001,243-dark green peak) had slightly irregular amplification peaks suggesting the possibility of DNA degradation (Fig. 3a). The C. brevis LAMP assay did not detect any cross-reactivity on the tested non-target species, only the positive controls (biological C. brevis and gBlock) produced an amplification peak after 09:15-10:15 min resulting in a negative LAMP test result (Table 1, Fig. 3a, b).
In testing pellet samples, the LAMP assay detected the positive control C. brevis pellet before time to positivity of 30 min (26.30 min). This was at annealing derivate 80.5 °C, with DNA concentration of 0.3 ng/µl. The three other C. brevis pellet samples (USA 1; USA 2, 10-001243) tested, showed late amplification. Because these samples amplified after 30 min (30.15, 36.09, 48.05) they were assigned as negative in this test. The DNA concentrations of those three samples were low (0.07 ng/µl) or unmeasurable. When compared to the positive control, the amplification rate peaks for the samples producing amplicons after 30 min were short and were irregular ( Table 1, Fig. 4a, b). Pellets of the non-target species did not produce any amplification peaks (Table 1, Fig. 4a www.nature.com/scientificreports/ The sensitivity of ten-fold serial dilutions of biological C. brevis positive control (USA1) showed positive amplification peaks after 15.00 to 25.45 min with annealing derivative temperature peak at 80.7 °C (Fig. 5a, b) and detection sensitivity of 0.0019 ng/µl (approximately 3 copies of DNA per µl) of DNA.
The performance of gBlock was tested using ten-fold dilutions DNA concentration ranging from 10 to 0.00001 ng/µl in LAMP reactions with positive detections found after 7.30-20.00 min. The detection sensitivity of gBlock was 0.00001 ng/µl (G6) of DNA within 20 min (Fig. 6a). The gBlock anneal derivate peak occurred at 80.7 °C (Fig. 6b). Based on amplification time, reproducibility, and consistency we conclude that gBlock G1 to G4 could be alternatively used as positive control.

Discussion
In this study, a novel C. brevis LAMP assay was developed and tested. The DNA sequences of the mt16S DNA gene were sufficiently variable to distinguish closely related Cryptotermes species including C. domesticus, C. dudleyi, and I. minor that are morphologically similar as pseudergates. Variability at this locus was sufficient to enable design of the C. brevis specific LAMP assay. The phylogeny of the mt16S gene revealed that the target termite, C. brevis is monophyletic, forming three sub-clades with slight sequence variation, all of which the primers were designed to detect. The LAMP primers were tested in-silico against all 30 species and in-vitro on close relatives (C. domesticus, C. dudleyi, and I. minor) that are difficult to distinguish as pseudergates and are most important for the purpose of Biosecurity decision making. The LAMP assay developed in this study is specific and able to detect the target species. No cross-reactions were observed in morphologically similar species, and this gives confidence in our C. brevis LAMP assay. Previously, the mitochondrial 16S gene has been successfully used in a LAMP assay for detection of pseudergates and faecal pellets of Incisitermes minor 14 . Although this gene www.nature.com/scientificreports/ is not the standard barcode for insects it was found to be the most informative for termites diagnostics in the present and previous studies 19,20 . We also designed a synthetic positive control "gBlock" that may be utilised in the C. brevis LAMP assay. This offers a reliable positive control that can be readily acquired from commercial gene fragment manufacturers without the challenges of obtaining biological DNA for positive control material. The synthetic positive control enables validation of each LAMP assay and gives confidence in the results obtained. A few species of Kalotermitidae, including C. brevis are pre-adapted to invade new locations via anthropogenic global trade and movement, due to their evolved strategy to disperse via pieces of wood 21,22 . First and foremost, mature colonies of pestiferous kalotermitids produce primary winged reproductives to found new colonies such as those from docked boats to lighted shore structures. Additionally, they produce secondary reproductive stages 23 , enabling residual colonies to establish in new locations 21 and presenting significant biosecurity risk. Because of the difficulty in extracting suitable termite life-stages to identify C. brevis morphologically, molecular tools are crucial for managing the biosecurity risk. Currently, the only molecular biology diagnostic method for C. brevis is Sanger DNA sequencing method 24 of the mt16S gene. This method requires significantly longer time and effort, up to 3-7 days, compared to 1-3 h for LAMP diagnostics. These time savings are critical in enabling a rapid response in managing Biosecurity risk. LAMP technology is becoming one of the most popular molecular methods for the detection of pests and diseases as it is a reliable, fast, and sensitive diagnostic tool. For instance, LAMP has been used in the early detection of pests and diseases such as Phytophthora ramorum 25 , Xylella fastidiosa 26 Bemisia tabaci, Thrips palmi 27 , Mythimna loreyi 28 . Currently, the only LAMP assay developed for drywood termite detection is for Incisitermes minor developed by Ide et al. 14 .
In many cases termite pellets are the only sample that can be obtained without destructive sampling of goods of high value, such as timber containing vessels. In this study, the LAMP assay detected C. brevis from pellet samples. However, in some cases the amplification peaks were produced after 30 min which exceeds the acceptable detection time, therefore the results were called negative. Pellets of non-target species did not produce peaks. Rizzo et al. 16 suggested that when developing an assay for pellet samples, three critical points must be considered: (1) insufficient amount of DNA in the sample; (2) the presence of inhibitors deriving from pellets and (3) the possibility of DNA degradation over the time. These factors are likely the cause of the late amplifications of C. brevis pellet samples. The DNA concentration was much lower in pellets (0.07-0.3 ng/µL) than in insect specimens (1.03-7.71 ng/µL) and some of the pellets were 20 years old. Also, pellets contain polysaccharides and other compounds known to inhibit PCR 29 . Ide et al. 14 developed a visual LAMP test for I. minor and concluded that detection in pellets older than 3 months is inconsistent. Obtaining fresh pellets was very difficult because C. brevis is absent from most of Australia. Therefore, most of the pellet samples used in this study were from insect collections more than a year old. Kyei-Poku et al. 2020 used a LAMP technique for the detection of the emerald ash borer from pellets, and also concluded the amount of extracted DNA from insects was more stable and higher than in their pellet samples 30 .
This is the first LAMP detection method developed for C. brevis, one of the most damaging drywood termite pests in the world, and a high priority drywood termite species of biosecurity concern to Australia. The LAMPbased assay optimized in this study only takes three hours to complete the full diagnostic (DNA extraction, amplification, and annealing). The C. brevis LAMP assay developed in this study can be used for the detection of C. brevis termite specimens; but is not reliable for their detection from pellet samples. This LAMP assay will assist in managing C. brevis threats across the biosecurity continuum; and allow for faster detection and mitigation of risks in a range of situations where drywood termites have infested timber.  AGT TTT TTA ATT GTG AAC  TGG TAT GAA TGG TTTG  ACG AGG CAT AAT CTG TCT TTA ATT  TGG ATT ATT ATT GAAT  TTA TTT TTT GGG TAA AAA TGC TCA  AAT TTT GTT ATG GGAC  GAG AAG ACC CTA TAG AGT TTT ATA  TAG TAC ACT TAT GAGT  ATT TGT TTT GTT GTG TTG GAG TGT  AAC TAA TAT TTT     www.nature.com/scientificreports/ netic analysis of the partial mitochondrial mt16S rRNA gene region. The species were selected to include examples of the target species C. brevis and the closest relatives for testing the specificity of the LAMP assay, as well as the three morphologically similar and commonly intercepted species at the Australian border: C. domesticus, C. dudleyi, I. minor (Table 1).    Table S1). This library was used for BLAST searches for this study. The consensus sequences were aligned using the MAFFT algorithm implemented in Geneious. Highly conserved regions within the sequence alignment targeting C. brevis were identified manually and assessed for their suitability for LAMP primer development (Fig. 1a). The LAMP primers were designed from a validated voucher specimen of C. brevis 10-001243 from which the mt16S rRNA sequences were obtained in this study (GenBank accession no. MT535992; Fig. 1b) using LAMP Designer software version 1.15 (Premier Biosoft International, USA) and confirmed by a local BLAST search to ensure their specificity (in-silico). The phylogeny of Cryptotermes species was estimated using Bayesian analysis which was performed with MrBayes 33 software implemented in Geneious Prime 2020 2.2 (https:// www. genei ous. com). Phylogenetic analysis was performed to estimate the relationship between closely related Cryptotermes species and C. brevis (Fig. 2).   Specificity and sensitivity of LAMP assay. The LAMP assay was tested against the target species and checked for cross reactivity on morphologically related species (non-target species) C. brevis and C. domesticus, C. dudleyi, and I. minor (Table 1). After the run completion, the amplification, amplification rate, and anneal derivative curves were visualised and assessed using Genie Explorer software version 1.15 (www. optig ene. co. uk/ suppo rt). The time of amplification (minutes) and anneal derivative temperature (°C) were compared against positive controls to confirm that no false positive or negative was amplified ( Table 1). The primer specificity was assessed following the manufacturer's recommendations assessing the amplification time and amplification peaks. The samples showing traces in less or equal to 30 (≤ 30) minutes with regular peaks were called positive and one above 30 (> 30) with flat peaks were called negative. Non-target species and the negative control are expected to have flat amplification lines. The detection sensitivity of C. brevis LAMP assay was assessed by preparing a ten-fold serial dilution of verified DNA of C. brevis termite (USA1). Starting DNA concentration of C. brevis was quantified using Qubit 3.0 fluorometer following the manufacturer protocol. The DNA sample was serially diluted at six concentration points ranging from 19.1 to 0.000191 ng/µl. The samples were run in the Genie II as described above. The initial www.nature.com/scientificreports/ quantification values from Qubit, were converted from nanogram per microlitre to copies per microlitre. Calculations were based on a C brevis genome size of 1.3 × 10 9 base pairs as per C. secundus 34 .

DNA extractions. Genomic
To test the performance of the synthetic positive control-gBlock, a ten-fold serial dilution of the gBlock DNA fragment was prepared using ultrapure water (Fisher Biotech, Australia). The DNA was diluted from 10 to 0.00010 ng/µl. The gBlock LAMP test was performed in the Genie II under the same LAMP assay conditions as previously described.