A Loop-mediated Isothermal Amplification (LAMP) Assay for Rapid Identification of Bemisia tabaci

The whitefly Bemisia tabaci (Gennadius) is an invasive pest of considerable importance, affecting the production of vegetable and ornamental crops in many countries around the world. Severe yield losses are caused by direct feeding, and even more importantly, also by the transmission of more than 100 harmful plant pathogenic viruses. As for other invasive pests, increased international trade facilitates the dispersal of B. tabaci to areas beyond its native range. Inspections of plant import products at points of entry such as seaports and airports are, therefore, seen as an important prevention measure. However, this last line of defense against pest invasions is only effective if rapid identification methods for suspicious insect specimens are readily available. Because the morphological differentiation between the regulated B. tabaci and close relatives without quarantine status is difficult for non-taxonomists, a rapid molecular identification assay based on the loop-mediated isothermal amplification (LAMP) technology has been developed. This publication reports the detailed protocol of the novel assay describing rapid DNA extraction, set-up of the LAMP reaction, as well as interpretation of its read-out, which allows identifying B. tabaci specimens within one hour. Compared to existing protocols for the detection of specific B. tabaci biotypes, the developed method targets the whole B. tabaci species complex in one assay. Moreover the assay is designed to be applied on-site by plant health inspectors with minimal laboratory training directly at points of entry. Thorough validation performed under laboratory and on-site conditions demonstrates that the reported LAMP assay is a rapid and reliable identification tool, improving the management of B. tabaci.


Introduction
The whitefly Bemisia tabaci (Gennadius) is an invasive insect pest affecting the yield of many economically important crops including ornamental plants, vegetables, grain legumes, and cotton 1,2 . Beside damage caused through direct phloem-feeding, the homopteran species harms plants indirectly by the excretion of large amounts of honeydew onto the surfaces of leaves and fruits, as well as by the transmission of numerous plant pathogenic viruses 1,3,4 . Recent genetic studies comparing DNA sequences of the mitochondrial gene cytochrome c oxidase 1 (COI) revealed that B. tabaci is a species complex of at least 34 morphocryptic species 3,4 . Two highly invasive and damaging members within this complex, biotype B originating from the Middle East and the Asian Minor region, as well as biotype Q originating from the Mediterranean region, have been dispersed globally through international trading activities with plant products, particularly by the transportation of ornamentals 1,5,6 . Due to its worldwide pest status, the International Union for the Conservation of Nature and Natural Resources (IUCN) listed B. tabaci as one of the "world's 100 worst invasive alien species" and members of the species complex are regulated organisms by many countries 1,3,4 .
In the European Union (EU), B. tabaci is listed in the Plant Health Directive 2000/29/EC Annex 1AI as a quarantine organism whose introduction from non-EU countries and its dissemination within the EU are banned 4 . An essential prevention measure against the spread of quarantine organisms is the inspection of plant shipments at points of entry (POEs) such as airports and seaports 7,8 . Especially the identification of immature life stages (e.g., eggs and larvae) without distinct morphological keys is virtually impossible for non-taxonomists 8,9,10 . Consequently, to enable implementation of quarantine measures with minimal delay, there is a need for alternative, rapid on-site identification assays 9 .
A candidate method is the loop-mediated isothermal DNA amplification (LAMP) technology that has recently been shown to be a suitable technology for the identification of plant pathogens 11,12,13 . LAMP is highly specific because the method uses at least two primer pairs recognizing six distinct DNA target sequences 14 . Due to the DNA strand displacement activity of the Bst DNA polymerase, LAMP reactions are performed under isothermal conditions 14 . Hence, in contrast to conventional polymerase chain reaction (PCR)-based assays there is no need for a thermal cycler 13,14 . Another advantage over PCR-based assays is its resilience against potential inhibitors in the DNA extract, circumventing the need for a DNA purification step 13 . Due to the protocol's speed and simplicity, LAMP may even be performed under on-site conditions using a portable, battery driven real-time detection device 8,15 .
A LAMP assay was designed in response to the demand for a rapid on-site identification method for B. tabaci 8 . The overarching aim was to develop a protocol that can be performed by plant health inspectors with limited laboratory training. A strong focus was, therefore, set on optimizing speed and simplicity of the protocol. While existing diagnostic tests have generally been developed for the identification of one or several biotypes of B. tabaci, the novel LAMP assay covers the whole B. tabaci species complex 8,16,17,18 . The problem of the pronounced genetic within-taxon diversity of the complex was solved by using combinations of different primer sets and the application of degenerate primers 8 . The novel B. tabaci LAMP assay is designed in such a way that the primers target a fragment at the 3' end of the mitochondrial COI gene 8 . This gene presents a suitable target for animal diagnostic assays because it harbors regions conserved enough to ensure diagnostic sensitivity for a specific species, while discriminating enough between closely related organisms 19,20 . Furthermore, the COI gene is often used as a genetic marker in population genetic studies and as a signature sequence in DNA barcoding analyses, resulting in numerous DNA sequence entries in open source databases such as GenBank and BOLD 21,22 . Beside the publicly available COI sequences from B. tabaci, COI sequences from closely related species (Aleurocanthus spp. Due to the accuracy of the method, its speed (<1 h) and the simplicity of the protocol, the assay has been shown to be suitable for on-site application when implemented as part of the import control procedure at a Swiss POE  With each kit, it is either possible to test two different specimens or to analyze the DNA extract of one specimen in duplicate. 2. Add 2.5 µL of sample DNA extract (prepared in step 2.1) into the tubes labeled "S1" and "S2" of the ready-to-use B. tabaci LAMP kit (Figure 1). 3. Add 2.5 µL of pure alkaline DNA extraction solution (prepared in section 1.1) into the tube labeled "NAC" for the negative amplification control (Figure 1). 4. Vortex the ready-to-use B. tabaci LAMP kit quickly and pulse centrifuge. 5. Insert the ready-to-use B. tabaci LAMP kits into the LAMP analysis device (with real-time fluorescence measurement) or a real-time PCR platform and perform an isothermal DNA amplification analysis at 65 °C for 60 min. 6. Measure the melting temperatures of DNA amplification products by heating up to 98 °C with a subsequent cooling step (ramp rate of 0.05 °C/s) to 75 °C, while measuring fluorescence in real-time.

Set up the PCR reaction as described in
3. LAMP assay read-out 1. Validate the LAMP read-out manually as follows.
1. If DNA amplifications were measured for the sample and the PAC, no DNA amplification was measured for the NAC, and the annealing temperature of the amplification products were between 80.0 and 85.5 °C, consider the LAMP results as POSITIVE (Figure 2). 2. If there is no DNA amplification for the samples (i.e., tubes labeled S1 and S2) but for PAC and NAC then consider the LAMP result as NEGATIVE (Figure 2). 3. If DNA amplification was measured for the samples, but the annealing temperatures of corresponding amplification products were outside the range 80.0 -85.5 °C, and/or PAC gave no DNA amplification, and/or NAC gave a DNA amplification, consider the LAMP result as INVALID (Figure 2).

2.
Optionally, validate the LAMP read-out using the LAMP validation application (Supplemental file 1). 1. Define target species and define the number of tested samples. Click the "Generate Report" button. 2. Transfer the read-out (DNA amplification yes/no, annealing temperature amplification product, results of PAC and NAC) from the on-site LAMP analysis device or real-time PCR platform to the corresponding input fields of the validation application. The result of the validation is immediately displayed after entering the data.

Representative Results
During the validation of the B. tabaci LAMP assay, insect specimens intercepted in the course of the regular Swiss import control process were analyzed 8 . The specimens originated from eight different countries (Canary Islands, Dominican Republic, Israel, Malaysia, Morocco, Singapore, Thailand, and Vietnam) and reflect the genetic diversity of B. tabaci found at European POEs 8 . All LAMP results were cross-validated by DNA barcoding 8 . From a total of 80 specimens analyzed by LAMP, 75 specimens (93.8%) were correctly identified as B. tabaci (true-positives), two specimens (2.5%) were correctly identified as not being B. tabaci (true-negatives), and three specimens (3.8%) were wrongly identified as not being B. tabaci (false-negatives) ( Table 2) 8 . The correct-negative results originated from two Trialeurodes vaporariorum specimens, a non-regulated species at high risk to be confused with B. tabaci at POEs for plant products 8 . Based on these results, the following measurements of diagnostic accuracy were calculated: test specificity (true-negative rate), 100%; test sensitivity (true-positive rate), 96.2%; test efficiency (percentage of correct test results), 96.3% ( Table 2) 8 . When assessing the analytical sensitivity (detection limit), the B. tabaci LAMP assay successfully amplified sample DNA diluted to 100 fg/µL across three technical replicates (  8 . When cross-validated in the reference laboratory, all results from on-site testing were found to be correct (test efficiency = 100%) 8 . Assessing the on-site LAMP assay performance, the average time to positive (time until a positive results was available) was 38.4 ± 10.3 min (mean ± standard deviation) 8 . A representative DNA amplification plot and the corresponding annealing derivative from a B. tabaci LAMP analysis performed under on-site conditions are shown in Figure 3A and B. In this example, sample one and two were correctly identified as B. tabaci indicated by DNA amplification after approximately 30 min ( Figure 3A) together with the expected annealing temperatures at approximately 82 °C (Figure 3B).

Discussion
The ability to accurately identify potentially harmful organisms without time delay represents a critical aspect for the management of pest species 9,10,26 . Besides being rapid, for plant import products, an ideal pest identification method should be simple to perform on-site at POEs 8,26 . This paper reports the protocol of a novel LAMP assay for the rapid identification of B. tabaci, a quarantine insect organism frequently intercepted at European borders (https://ec.europa.eu/food/sites/food/files/plant/docs/ph_biosec_europhyt_annual-report _2016.pdf).
The rationale behind the development of the diagnostic test was to design an easy-to-follow protocol which can be performed during the plant import control procedure by plant health inspectors with minimal laboratory training. In order to make on-site testing as rapid and simple as possible, the protocol is divided into two parts, the preparation of a ready-to-use kit and the actual performance of the LAMP assay. The first part may be done in an external laboratory so that the plant health inspector can perform the DNA extraction and LAMP assay on-site with only one pipetting step.
Though only one step, pipetting small amounts of liquid may be challenging for users with little or no laboratory experience. To address this issue, a dye (cresol red) is added to the extraction solution so that the operator can visually confirm the small amount (i.e., 2.5 µL) of DNA is correctly transferred to the respective tube. Another important simplification of the protocol is the validation application as it facilitates a reliable interpretation of the LAMP read-out (Supplemental file 1).
The novel B. tabaci LAMP assay has been validated under laboratory and on-site conditions by testing insect specimens intercepted during the regular import control process of Switzerland 8 . In total, 80 specimens from three continents, Africa, Eurasia, and North America, were analyzed by LAMP. Of the 80 specimens, only three (3.8%) were wrongly identified (false-negatives) 8 . When analyzing the primer target DNA sequences of the false-negative specimens, it was found that they were new B. tabaci haplotypes that have so far not been described 8 . Based on these results, the B. tabaci LAMP primer set has been modified and successfully re-validated 8 .
One major limitation of any DNA amplification-based method including LAMP is that they only identify pre-defined target DNA sequences 8,27 . A comprehensive knowledge of the genetic variation found in the primer target sequence is therefore crucial to ensure diagnostic accuracy 8,27 . However, such information is often very limited, especially in the case of newly emerging pest species 8 . Though rare, false-negative results caused by mutations in the target sequence are expected 8 . In the case of the present B. tabaci LAMP assay, a solution for this problem is the combination with a DNA barcoding-based technology, a strategy realized in the course of the implementation of this diagnostic test at the POE Zurich Airport 8 . Here, all LAMP-negative results were re-analyzed by DNA barcoding in an external laboratory 8 . In case a novel pest haplotype not yet described is encountered, the LAMP primers can be modified using the DNA sequence generated in the barcoding process 8 . Thereby, the resulting loss of speed in case of a negative LAMP result is compensated for the maximum diagnostic accuracy ensured in this two-stage process 8 .
The set-up costs for the current LAMP assay at a POE are approximately USD 25,000. With the increasing number of LAMP tests developed for plant pests (e.g., Erwinia amylovora, Flavescence dorée, and Guignardia citricarpa), such a one-time investment appears justified 13,15,28 . However, the protocol could potentially be modified to reduce these costs even further. For example, for the DNA extraction step at 95 °C the thermo mixer used here could be replaced by a less expensive water bath, or by performing this step directly in the real time LAMP device. Furthermore, the mixing steps on the vortex could probably be replaced by manually flicking the tubes, and in the DNA transfer step the pipettor might be replaced by sterile inoculation loops.
Future improvements for a rapid identification of B. tabaci and pest species in general could be an implementation of an on-site sequencing approach that would allow to perform DNA barcoding analyses at POEs. A promising candidate system for such an implementation is the nanopore sequencing technology. Indeed, the technology has recently been successfully implemented in an on-site DNA barcoding effort to assess the biodiversity of a rainforest 8,29,30 . An on-site DNA barcoding identification system can completely replace the need for the development  8 . Nevertheless, until novel sequencing technologies will be implemented routinely, the B. tabaci LAMP assay represents a rapid (<1 h) and accurate identification method.

Disclosures
The author Michael Andreou is a shareholder of OptiGene Limited that produces reagents and instruments used in this article. The other authors have nothing to disclose.