Molecular Identification of ten species of stored-product psocids through microarray method based on ITS2 rDNA

Stored-product psocids (Psocoptera: Liposcelididae) are cosmopolitan storage pests that can damage stored products and cause serious economic loss. However, because of the body size (~1 mm) of eggs, nymphs, and adults, morphological identification of most stored-product psocids is difficult and hampers effective identification. In this study, 10 economically important stored-product Liposcelis spp. psocids (Liposcelis brunnea, L. entomophila, L. decolor, L. pearmani, L. rufa, L.mendax, L. bostrychophila, L. corrodens, L. paeta, and L. tricolor) were collected from 25 geographic locations in 3 countries (China, Czech Republic, and the United States). Ten species-specific probes for identifying these 10 psocid species were designed based on ITS2 sequences. The microarray method and reaction system were optimized. Specificity of each of the ten probes was tested, and all probes were found suitable for use in identification of the respective10 Liposcelis spp. psocids at 66 °C. This method was also used to identify an unknown psocid species collected in Taian, China. This work has contributed to the development of a molecular identification method for stored-product psocids, and can provide technical support not only to facilitate identification of intercepted samples in relation to plant quarantine, but also for use in insect pest monitoring.

In order to overcome the limitations of morphological identification methods, many molecular technologies, including restriction fragment length polymorphism (PCR-RFLP) 3 , DNA barcoding 4 , species-specific primer PCR (SS-PCR) 7,8 , multiplex endpoint PCR and multiplex TaqMan qPCR 9 , have been developed for psocid identification. Molecular markers are the foundation for molecular identification, and both mtDNA (mitochondrial DNA) and rDNA (ribosomal DNA) are species-specific and can be used as molecular markers for species identification 10,11 . To date, many molecular makers, including cytochrome c oxidase subunit I (COI), 16S ribosomal RNA gene (16S rDNA), internal transcribed spacer2 (ITS2) and microsatellites have been confirmed effective for identifying stored-product psocids from different species 12 . Successful use of these molecular markers and methods has made stored-product psocid identification much quicker and more accurate, also laying a foundation for establishing a systemic and rapid molecular identification system for this small pest.
Among numerous molecular methods, the microarray methodology has received much attention; it has been extensively researched and has practical applications, stemming from its advantages of high throughput, high parallelism and high sensitivity. The solid gene chip, as a traditional microarray method, has been widely used in many areas including disease diagnosis and prediction, drug screening, agriculture, food hygiene and environmental monitoring 13,14 . In recent years, this method is beginning to be used in insect identification. Sequences of two mitochondrial genes (COI and ND2) and a ribosomal gene (ITS2) have been used to design species-specific probes, and a gene chip method thus developed was able to provide simultaneous identification of nine important medical and veterinary species, including immatures, from genera of Aedes, Anopheles, Armigeres, and Culex 15 . The gene chip helps overcome limitations caused by phenotypic variations in adults, lack of recognizable features in immatures, and the fragility of mosquitoes 15 . An innovative and accurate rapid high-throughput microfluidic dynamic array method based on COI DNA sequence has also been used for detecting 27 economically important tephritidae species in six genera (Anastrepha, Bactrocera, Carpomya, Ceratitis, Dacus and Rhagoletis) 16 . A gene chip based on the COI gene has been constructed for identifying 3 species of quarantine Ceratitis (C. capitata, C. cosyra and C. rosa), 3 species of Frankliniella, and 15 species of 5 genera in Culicidae [17][18][19] .
Although the gene chip has been used to identify some kinds of insects, there is currently no report of this method being applied in stored-product psocid identification. In this study, for the first time we designed 10 species-specific probes for identifying 10 stored-product psocid species, tested the specificity of these probes, established a microarray method based on the ITS2 rDNA sequence, and then optimized the method and reaction system. This method and the species-specific probes were also used to identify an unknown psocid species that was collected in China. Our results showed that the microarray method using species-specific probes based on ITS2 rDNA sequence can improve the molecular identification of stored product psocid species and can increase the likelihood for correctly identifying an intercepted sample in the field, in relation to plant quarantine. It can also be used to facilitate insect pest monitoring in the field of stored product protection.

Results
Probe sequences for ten species. All the probes used in this research were designed based on ITS2 sequences (Fig. S1). The species-specific probes that were used for the identification of ten booklice species (Table 1) are listed in Table 2. Quality-control probes, including the anchor position probe, positive probe, negative probe and blank control probe are listed in Table 3, all the probes were oligonucleotide fragments with no sequence homology to ITS2 sequences of the ten species. The anchor position probe was used for detecting whether there was a normal coupling between the probe and the chip, because there was a Cy5 label at the 5′ end of this kind of probe. In addition, this kind of probe can also be used as a reference for positioning other probes when reading the hybridization result. The positive probe was used as a positive control for detecting whether the hybridization was successful or not. The negative control can monitor if there is non-specific hybridization.
Microarray chip design. There were 12 of the same probe lattices, which distributed as 6 rows and 2 columns for each microarray chip. The detailed layout of the probes in each probe lattice is shown in Fig. 1. There were 8 rows and 10 columns for each probe lattice, and the first row, the first and second column were the anchor position probes. The other points were for the species-specific probes and there were four repeats for each probe. The space between columns was 0.4 mm, while the space between rows was 0.8 mm. This probe lattice can be used for identifying ten species of stored-product psocids.
Establishment of the reaction system. On the basis of optimizing primer concentrations, upstream and downstream primer proportion, detection sensitivity and hybridization temperature, one optimized reaction system was established for the microarray chip identification method. Asymmetric PCR was useful for amplifying enough ITS2 sequences, especially when the proportion of upstream and downstream primers is 1:20 (Fig. S2). The best result, comprising the highest hybridization efficiency, the best detection specificity and the weakest false positive can be obtained at the temperature of 66 °C (Fig. S3). The detection sensitivity showed that the signal was strongest when DNA without dilution was used as the template for ITS2 sequence amplification (Fig. S4).
Microarray scanning results for the 10 species of Liposcelis. The hybridized pattern was scanned at the excitation wavelength of 635 nm/532 nm, and the Cy5 displayed red color at this wavelength. The signal value for each hybridized site was obtained and used as the basis for deciding whether the hybridization was successful or not, whether the hybridization was positive or not, and whether there are some problems of hybridization. The value of F635 Mean-B635 (the mean spot pixel intensity at wavelength 635 nm/532 nm with the background subtracted) was used as the signal value. The signal value for the blank and negative control was the mean value of four repeated hybridized sites. If the signal value for one detecting site minus the mean signal value of blank control was higher than ten times the mean signal value of negative control, and also higher than 500, this detecting site was determined as positive, which meant that the probe for this detecting site hybridized with the provided DNA   Table 3. Quality-control probes of gene chip for 10 species of Liposcelis a The anchor point probe had -Cy5 added at the 5′ side and -NH 2 added at the 3′ side, while the other probes had -NH 2 -(T) 15 -added at the 5′ side.
successfully. When the mean signal value of four repeated detecting sites for each probe was positive, it means the hybridization result for this probe was positive. The result was inaccurate under certain circumstances -when the negative hybridization result for some or all the anchor point probes indicated the probe had not combined with the chip, a positive hybridization result for negative control indicated there was non-specific hybridization or there was pollution during the process of amplification and hybridization, while a negative hybridization result for positive control indicated a problem existing in the process of amplification or hybridization. When the hybridization result for the positive probe and one specific probe was positive, it implies the microarray worked well and showed there was successful hybridization between the specific probe and the corresponding Liposcelis species. According to this analysis method, the hybridization results for ten species-specific probes are shown in Fig. 2. Figure 2A shows a successful hybridization between probe Lbr and the ITS2 sequence from L. brunnea, which were collected from the United States (USA) or Prague in Czech Republic. Figure 2B shows the scanning result for hybridization between probe Len and the ITS2 sequence from L. entomophila, which were collected from Beijing, Chongqing, Wuhan (Hubei Province) in China, or Prague in Czech Republic. Figure 2C shows a successful hybridization between probe Lde and the ITS2 sequence from L. decolor, which were collected from Chongqing in China, Prague in Czech Republic or the USA. Figure 2D shows that probe Lpe can hybridize successfully with the ITS2 sequence from L. pearmani, which were collected from the USA. The successful hybridization between probe Lru and the ITS2 sequence from L. rufa (USA population) is shown in Fig. 2E. The scanning result for hybridization between probe Lme and the ITS2 sequence from L. mendax, which were collected from Jiangsu Province in China, is shown in Fig. 2F. As shown in Fig. 2G, probe Lbo can hybridize successfully with the ITS2 sequence from L. bostrychophila, which were collected from Beijing, Guangxi, Chongqing in China, Prague in Czech Republic, or Manhattan in the USA. L. corrodens ITS2 sequence from two populations, including Prague in Czech Republic and the USA, can hybridize successfully with probe Lco separately, and the scanning result is shown in Fig. 2H. Probe Lpa can hybridize successfully with the ITS2 sequence from L. paeta, which were collected from Shijiazhuang (Hebei Province), Zhejiang Province, Wuhan (Hubei Province) in China, Prague in Czech Republic, or the USA (Fig. 2I). Figure 2J shows a successful hybridization between probe Ltr and the ITS2 sequence from L. tricolor, which were collected from Heze (Shandong Province) in China.

Practice of the microarray chip method.
To examine the specificity of these 10 species-specific probes and the microarray method in practice, a sample of one Liposcelis species, which was collected from Taian, Shandong Province in China, was used for DNA extraction and ITS2 sequence amplification. As shown in Fig. 3, after being hybridized with the provided ITS2 sequence, the microarray showed positive signals at the position of probe Lpa and the positive probe, which is the same as in Fig. 2I. The scanning result indicated that the Liposcelis species from Taian, Shandong Province was L. paeta. The traditional morphological identification result also showed that the supplied sample had the typical morphological characteristics of L. paeta, including tawny body color, segmacoria at the posterior edge of the third and fourth abdomen tergum (link type abdomen), one pair of bristles at the latter part of antesternum, no furcation at the end of gonapophysis, short and tapered receptor on the maxillae palpus, three ommatidia and so on. ITS2 sequence of this unknown species was also amplified and sequenced. The sequence was submitted to NCBI and got a GenBank NO (MF434638). NCBI blast result also showed that the unknown species has a highest similarity with L. paeta. The microarray result was coincident with the morphological identification and NCBI blast result.

Discussion
In this study, 10 species-specific probes based on ITS2 rDNA were provided and an identification method using a microarray chip was developed. An optimized reaction system that comprised primer concentrations, upstream and downstream primer proportion, detecting sensitivity and hybridization temperature was established. This method was used to identify 10 economically important Liposcelis species from 25 geographical populations in one microarray chip. It was also used to identify an unknown psocid species that was collected from Taian, China. To the best of our knowledge, although ITS2 rDNA has been used as an effective molecular marker for identifying stored-product psocids for a long time, this is the first report of using species-specific probes based on ITS2 rDNA for psocid species identification.
For eukaryotes, there are two internal transcribed spacers (ITS1 and ITS2), highly repetitive sequences that are removed after DNA transcription 20,21 . ITS1 is located between 18S rDNA and 5.8S rDNA, whereas ITS2 is located between 5.8S rDNA and 28S rDNA 22,23 . As with 5.8S rDNA, 18S rDNA, and 28S rDNA, ITS2 displays high conservation within a species but significant variation among species; this characteristic makes this gene suitable for use as a molecular marker for identification and phylogenetic analysis in many species 22,24,25 . In recent years, ITS2 is increasingly being used in taxonomy, phylogenetic development, and investigating species evolution 26,27 . Since 2011, the ITS gene has been used to study the gene structure and phylogenetics of stored-product psocid species. Wei et al. 6  Because primer and probes specificity are the most important factors for species identification, it means higher annealing temperature can enhance the specificity. The microarray used a higher annealing temperature (66 °C) than the previous identification methods for psocid, especially the species-specific primer method published by Zhao et al. in 2016 (52 °C). Additionally, the microarray method can also improve the specificity by two steps, including asymmetric PCR and the subsequent hybridization. In published studies, it's rare to find a positive control in PCR-based methods for identification, because of how difficult it is for this to occur. However, in this study, we set a positive probe which can indicate normal hybridization and this become of obvious advantage for microarray. In the present study, 2.5 h were required to identify 12 psocid samples on one chip with high efficiency (excluding DNA extraction); the sequencing-based method used in the identification of 14 Noctuoidea species using a thin-film biosensor required 4 h to identify 14 Noctuoidea species 28 . The throughput can also be increased by dotting more probes in one layout. When identifying one unknown species using the PCR-based method, 10 reaction volumes (total 250 μL) for 10 pairs of species-specific primers were used. While the microarray-based method requires only one reaction volume (25 μL) for 10 pairs of species-specific probes. Thus, the conclusion is that this study developed a reagent-saving identification method with high specificity, accuracy and efficiency. This technique also offers great potential for detecting high species diversity, and can satisfy plant quarantine needs for rapid, accurate and effective identification.
Although the microarray method has some advantages, it sometimes produces false positive results. For further research, more species from more populations should be used to increase the specificity of probes. The optimal reaction volume and temperature, optimal distribution of probes in the microarray chip, optimal concentration of DNA and other protocol details need to be improved to avoid the possible false positive results. Additionally, only adults were used for this study, but future research needs to use other stages as well. DNA was extracted from a single adult using a DNA extracting Kit. However, typically, eggs, nymphs and body parts are intercepted and rarely are adult individuals collected. The DNA extraction method used in this study can also be applicable for eggs and nymphs. The microarray identification method developed was not used for non-adult samples hence this needs to be the subject of future studies.
As previously mentioned, stored-product psocids are now recognized as cosmopolitan pests of substance 6,29-31 . The economic losses, high interception frequency, and difficulty of morphological identification caused by their small body size necessitate the development of rapid molecular identification methods. Many evidences have shown that even closely related Liposcelis species that are hard to differentiate by morphological methods may respond differentially to various control methods and to cereal storage environmental conditions. For example, Guedes et al. (2008a) found that L. bostrychophila was more tolerant than L. entomophila in relation to response to surfaces treated using ß-cyfluthrin and chlorfenapyr insecticides; this was due to differential species-specific behavioural reaction to pesticide residues 32 . Guedes et al. (2008b) showed that L. entomophila was twice as tolerant to heat treatment as L. reticulatus 33 . In addition, various Liposcelis species may exhibit pose different levels of risk based on variation in how rapidly their populations grow after infestation under identical storage temperature and humidity [34][35][36] . Thus, correctly identifying psocid species responsible for infestations is important in selecting effective control methods; developing rapid, systematic molecular identification methods for stored-product psocid identification facilitates their control 12,37 . The microarray methodology, which is an accurate, high throughput, and convenient molecular identification method, has played an important part in improving the timeliness of psocid identification for the port and food sector in China. Future effort needs to develop commercial detection technology and products, partly through pilot promotion in different ports. Improving the specification and optimization of the selective reaction system can help this method (microarray method) play an increasingly important role in facilitating quarantine and stored-product protection in China and elsewhere.

Methods
Sample collection and morphological identification. Ten stored-product psocid species from 25 geographic populations were collected from locations in China, Czech Republic and the United States. All the 25 geographic populations were identified by two taxonomists Drs. Zhihong Li and Zuzana Kučerová, using morphological characteristics, and were preserved in absolute ethyl alcohol at the Plant Quarantine and Invasion Biology Laboratory of China Agricultural University (Table 1).

DNA extraction.
Total DNA of 10 psocid species was extracted from individual booklice (Table 1)  ITS2 sequence amplification. Partial sequence of human gene (KY962518.1, from 4019 to 4136) was used for positive control, and this sequence can be amplified by using one pair of primers, For2 and Rev2 (Table 4) via normal symmetric PCR.One pair of universal primers, For1 and Rev1 (Table 4), used by Zhao et al., were used for amplifying the ITS2 gene of 10 stored-product psocid species based on a screening process. Oligo synthesis was conducted by the Beijing Aoke Biotechnology Company. An asymmetric PCR was performed using the 5′ Cy5-labelled reverse primer (Cy5-Rev1) at a concentration 19 times higher than that of the forward primer (For1), which means the proportion of the two primers is 20:1. A 25 μL PCR system was used, which contained 1.5 μL of DNA template (psocid DNA or positive control amplification product, 25 ng/μL), 2.5 μL of 10× EX Taq  Probe design. ITS2 sequences of the 25 populations of 10 Liposcelis species used were retrieved from GenBank and were aligned using DNAMAN 6.0; species-specific regions suitable for probe design were selected for species-specific probe design. Primer premier 5.0 and DNASTAR Laser gene 7.0 were used to design the probes. Quality-control probes, including anchor point probe, positive probe, negative probe and blank control probe, were designed to confirm whether the PCR and the hybridization were normal.

Microarray chip construction. On a commercially purchased aldehyde modified glass slide (Capital Bio
Corporation, China), the probe solution (50 μM), prepared by mixing the 100 μM oligonucleotide probe and 2× probe solution, was spotted on the surface of polymer membrane using the DNA Microarray and Protein Microarray Spotter AD1500 (Biodot, the USA), and four spots were used for each booklice species. These spotted slides were then incubated at 37 °C in a humid chamber for ≥12 h. The slides were then submerged in wash buffer (300 mM Bicine, 300 mM NaCl, 0.1% SDS) for 30 min at 66 °C, then in sterile deionized water for 5 min and dried by centrifugation. Before hybridization, fencing was pasted to divide the glass slides into hybridization plots. And then the layout was covered to make it separate from the others, so that the reactions would not influence each other.
Microarray hybridization and scanning. Ten μL of mixture containing 4.5 μL of the labeled PCR product of booklice species, 1 μL of applied positive sequence, and 4.5 μL of hybridization buffer were added to the plots on the microarray, which was put in a hybridization box with water to create wet conditions. The microarray was then incubated at 66 °C in the oven with vibration for 2.5 h. After removing the fencing, the glass slides were washed in wash buffer I (2× SSC, 0.1% SDS) two times for 5 min, then in wash buffer II (0.2× SSC, 0.1% SDS) two times for 5 min, and the last washing was conducted once in wash buffer III (0.2× SSC). Slides were put in absolute ethyl alcohol for extraction and dried after washing. The hybridization pattern was detected using a Microarray Scanner InnoScan 900 (Innopsys, France), and the software Mapix 2.9.5 was used to get the fluorescence signal data. For the species with more than 2 geographic populations, at least 2 adults from each population were used for hybridization; while for the species with only one geographic population, at least 3 adults were used for probes detecting. At least 3 biological replicates were used for each probes detector. All the experiments were repeated for three times.  Table 4. Universal primers for ITS2 amplification and long primers for positive template amplification.