Development of a robust, field-deployable loop-mediated isothermal amplification (LAMP) assay for specific detection of potato pathogen Dickeya dianthicola targeting a unique genomic region

Destructive maceration, a wide host range, and longevity in non-plant substrates has established Dickeya dianthicola (blackleg of potato) as a significant threat to potato industries worldwide. To protect these businesses, a specific and sensitive point-of-care D. dianthicola detection tool is necessary. We have developed a loop-mediated isothermal amplification (LAMP) assay for specific, sensitive, and rapid detection of D. dianthicola, which can be streamlined for point-of-care use. The developed LAMP assay targets a unique gene, alcohol dehydrogenase, of D. dianthicola. Assay specificity was assessed using strains present in inclusivity (16 D. dianthicola strains) and exclusivity panels (56 closely related, potato pathogenic, and other bacterial strains). Amplification with strains of inclusivity panel occurred, and cross-reactivity with non-target DNA was not observed. The limit of detection (LOD) was 10 CFU/ml when dilutions were made before isolating the genomic DNA; however, LOD was determined as 1 pg using 10-fold serially diluted D. dianthicola genomic DNA. Similar LOD of 1 pg was observed when serially diluted target genomic DNA was mixed with host genomic DNA. LOD (1 pg) was also calculated with 10-fold serially diluted synthetic DNA fragments containing primer target sites. Naturally and artificially inoculated plant samples were used for field adaptability tests with the field-deployable Optigene Plant Material Lysis Kit and a heat block (65°C); the results were obtained within 20 minutes. Despite the lack of method precision, no false positives or false negatives were observed. Therefore, with prepared reactions and a steady heat source, this assay can be used for rapid point-of-care detection, which is imperative for quarantine, eradication, disease management, and border protection.

Introduction Potato (Solanum tuberosum) is the tenth most produced crop in the world [1]. High production may be attributed to its use in multiple industries and ability to grow in diverse climate conditions [2][3]. Under various climates, several bacterial diseases have been reported on potatoes [4][5]. Of the bacterial diseases, soft rot and blackleg of potato caused by Dickeya dianthicola is one of the most devastating diseases damaging economies worldwide [6][7][8].
D. dianthicola (previously Erwinia chrysanthemi pv. dianthicola; Pectobacterium chrysanthemi pv. dianthicola) is a seed borne phytopathogen [9] and has been detected in the European Union, South America, New Zealand, Western Australia, Japan, and the United States [7,10,11]. Throughout these countries, D. dianthicola has a wide host range [6,12,13], but potato is considered the main host [14,15]. Infected potato plants exhibit wilting, dwarfing, and wet, oozy, black stems [3]. The pathogen is highly virulent and requires less inoculum than Pectobacterium atrosepticum and P. carotovorum to cause severe symptoms in potato [6,14]. At high temperatures (between 25 and 30˚C), D. dianthicola produces noticeable destruction and decay, but at low temperatures can be visually undetectable. Furthermore, D. dianthicola can survive in soil and water for several months, increasing the probability of spread and contamination to potential hosts [6,7,16]. Consequently, it is important to develop efficient, sensitive, field-ready diagnostic tools for specific detection of D. dianthicola.
Dickeya species can be identified using a series of biochemical tests paired with semi-selective media as well as lateral flow immunoassay [17]; neither method is time efficient or accurate. In contrast, molecular techniques are specific and sensitive for detection and differentiation of Dickeya to species [18][19][20][21]. Conventional PCR is an important nucleic-acid technique but is not time efficient or field-deployable. The advancements of isothermal methods have provided rapid and sensitive techniques that can be used at point-of-care without the need of special equipment.
Loop-mediated isothermal amplification (LAMP) is a popular isothermal, nucleic acid amplification-based technique used for detection of several plant bacterial pathogens [22][23][24]. LAMP employs a strand displacing Bst polymerase for the million-folds amplification of the target DNA duplex [25]. The reaction uses 4-6 primers to specifically bind to 6-8 specific regions in the target genome [22,26].
For a LAMP reaction to be highly specific and exclusive, it is imperative to find a unique gene region for primer design [22,23]. The unique regions can be identified by comparative genomic analyses of different strains of the same species and other neighboring species/genera [22,[27][28][29]. Primer specificity and broad range detection capabilities can be tested against strains of inclusivity and exclusivity panels [30]. If a truly unique region is not used, then the diagnostic tool may produce non-specific results [27,31]. Overall, determining a sequence unique to a species is the key to developing a successful and robust LAMP diagnostic tool [22,23].
Numerous chemistries (pyrophosphate turbidity, fluorescence, gel electrophoresis) are available for observing positive LAMP amplification but require special equipment for visualization [32][33][34]. However, SYBR Green I stain has the advantage of producing a color change that is observable without equipment [22,23]. When added to a LAMP reaction, results are visible almost immediately, which is useful in time-sensitive situations. Although, a heat block is adequate, several battery-operated portable commercial real-time LAMP instruments, like the Genie II [35] or BioRanger, suitable for field application are available [23,36].
Currently, LAMP has been developed for the Dickeya genus [37]. Nonetheless, a swift, convenient, and reliable diagnostic method is needed for direct identification of the aggressive D. dianthicola. The purpose of this study was to develop a LAMP reaction for specific, accurate, and rapid detection of D. dianthicola from infected plant tissues. Potential applications include point-of-care plant disease diagnosis for disease management, field surveys, and biosecurity of agricultural crops.

Ethical statement
No permission was required from government agencies or regulatory bodies to include the infected samples in this study. Endangered or protected species were not collected or used in this study. No samples were collected from endangered or protected field sites.

Source isolates, plant inoculation and DNA isolation
Sixteen isolates of D. dianthicola and fifty-six isolates of closely related species and genera were selected for inclusivity and exclusivity panels, respectively (Tables 1 and 2). Strains from worldwide locations and hosts were chosen from the Pacific Bacterial Collection (University of Hawaii at Manoa). Selected bacterial strains were cultured on tetrazolium chloride media (TZC; 5 g peptone, 2.5 g dextrose, 8.5 g agar and 0.5 ml 1% TZC in 500 ml distilled water) at room temperature. Culture plates were incubated at 26 ± 2˚C and a single colony of each plate was re-cultured.
Healthy, greenhouse grown potato plants (~4 weeks old) were inoculated with four D. dianthicola isolates: PL22, PL24, PL25 and PL31. Stems of each seedling were stab inoculated with a sterile scalpel dipped in inoculum. Plants were kept in the greenhouse for three days; plants showing a black leg symptom on stems were collected for DNA extraction. Additional inoculations of tubers were completed with A5278 (P. carotovorum subsp. carotovorum) and A6152 (P. carotovorum subsp. brasiliensis). Briefly described, tubers were surface sterilized with 10% sodium hypochlorite (NaOCl) for three minutes, washed three times with sterile water, and then cut into slices. Slices were placed on filter paper (moistened with 5 ml sterile water) in petri dishes and stabbed with sterile toothpicks dipped in Pectobacterium sp. inoculum. Petri dishes were incubated at 28˚C for 24 hours. Tubers exhibiting maceration symptoms were selected for DNA extraction. DNA was extracted from pure bacterial colonies; healthy potato stems and tubers; naturally D. dianthicola infected potato plants; and artificially D. dianthicola and Pectobacterium sp. infected potato plants and tubers, respectively. Genomic DNA of all bacterial strains in the inclusivity and exclusivity panels were extracted using the DNeasy Blood and Tissue Kit (Qiagen, Germantown, MA) following the manufacturer's instructions. DNA from infected and non-infected potato plant tissues were extracted using the Wizard Genomic DNA Purification Kit (Promega, Madison, WI). The manufacturer's instructions were followed with an additional step of using a Mini-Bead Beater 16 Center Bolt (Biospec products, Bartlesville, OK) for one minute at maximum speed to thoroughly rupture cells. Following extraction, the Nano-Drop 2000/c Spectrophotometer (Thermo Fisher Scientific, Waltham, MA) was used to estimate the DNA concentration of all samples (pure cultures and tissue).

PCR and identity confirmation
dnaA and 16s rRNA gene regions were used for the identification of bacterial strains. 16s rRNA primers were selected from Dobhal et al. [22] to amplify 16s rRNA gene regions. Primers from Schneider et al. [38] were used to amplify the dnaA gene region. Dickeya sp. dnaA primer PCR conditions were: initial denaturation at 95˚C at 5 min followed by 35 cycles of denaturation at 95˚C at 20 sec, annealing 53˚C at 60 sec, extension 72˚C at 1 min, and final extension at 72˚C at 2 min. Pectobacterium sp. and R. solanacearum dnaA primer PCR conditions were: initial denaturation at 94˚C for 5 min followed by 35 cycles of denaturation at 94˚C for 30 sec, annealing 61˚C for 1 min, extension 72˚C for 30 seconds, and final extension at 72˚C for 4 min. Xanthomonas sp. dnaA primer conditions were: initial denaturation at 94˚C for 5 min followed by 35 cycles of denaturation at 94˚C for 30 sec, annealing at 61˚C for 1 min, extension at 72˚C for 30 sec, and final extension at 72˚C for 10 mins. The 16s rRNA PCR conditions were followed as described by Dobhal et al. [22]. All PCRs were performed in the BIORAD T100 Thermocycler (Bio-Rad, Hercules, California). PCR products were electrophoresed on 1.5% agarose gel and the bands were visualized under FOTO/UV 26 (US PAT-ENT 5347342) gel doc assembly. To clean the PCR products, 2 μl ExoSAP (Affymetrix Inc, Santa Clara, CA) was added with 5 μl of PCR product; incubated at 37˚C for 15 min followed by 80˚C for 15 min to deactivate the enzyme. The Sanger sequencing was performed at GENE-WIZ facility, La Jolla, CA. The forward and reverse sequences were aligned using Geneious 10.2.3 software and evaluated manually for errors. Consensus sequences were obtained, and the identity was confirmed by searching the NCBI GenBank nucleotide and genome databases using the BLASTn tool. Multiple alignments of consensus sequences from the exclusivity and inclusivity panels were performed using Geneious.

Target selection, primer design, and in silico validation
The alcohol dehydrogenase gene was determined as a unique region in D. dianthicola through genomic comparison of D. dianthicola (NZ_CM001838, NZ_CM001840, NZ_CM001841 and solani (NZ_CP009460, NZ_CP015137) and D. zeae (NZ_CP006929, NC_013592) (Dobhal and Arif, unpublished information). This unique gene was used to design LAMP primers for specific detection of D. dianthicola (Fig 1). Forward inner primer (FIP), forward outer primer (F3), backward inner primer (BIP), backward outer primer (B3), forward loop primer (LF) and backward loop primer (LB) were designed using PrimerExplorer V5 (https://primerexplorer.jp/e/). Specificity of each primer was verified by comparing the primer sequences against the NCBI GenBank nucleotide and genome databases using BLASTn tool. Primers were checked for possible secondary structures using MFOLD (http://unafold.rna.albany.edu/?q=mfold). Primer information is provided in Table 3.

LAMP assay specificity determination
The specificity of the developed LAMP primers was tested with a total of 16 D. dianthicola strains and 56 other strains included in inclusivity and exclusivity panels, respectively (Tables  1 and 2). DNA from soil and healthy plants were used as negative controls and sterile distilled water (molecular grade) served as non-template control (NTC). Three pairs of primer were used in the LAMP reaction, inner (FIP and BIP), outer (F3 and B3), and internal (LF and LB) primers, targeted the alcohol dehydrogenase gene (Table 3)

Limit of detection determination
Four tests were completed to determine the limit of detection: 10-fold serially diluted bacterial culture before DNA purification, 10-fold serially diluted purified genomic DNA, 10-fold serially diluted purified genomic DNA mixed with host DNA, and a 10-fold serially diluted synthetic DNA fragment. For 1 st LOD assay, overnight grown culture of D. dianthicola (1X10 9 CFU/ml) which was confirmed by plating the serial dilutions prepared in 0.1% (w/v) peptone water (BBL, Becton Dickinson, Sparks, MD) on TZC media [22] and incubating the plates at 28˚C for 12-18 h. The cell counts were recorded and calculated in terms of log10 CFU/ml. Serial dilutions were made from 10 8 CFU/ml to 1 CFU/ml. Each of this concentration was mixed with 100 mg of the healthy plant tissues (potato stem) and DNA extraction was performed using the Qiagen DNeasy Plant Mini Kit following the manufacturer's instructions.
The LAMP assay was performed as described previously. For 2 nd LOD assay, purified genomic DNA of D. dianthicola strain A5573 was 10-fold serially diluted with water from 1 ng to 1 fg and used to perform the LAMP assay. For 3rd LOD assay, healthy potato tuber genomic DNA was added to each 10-fold serially diluted D. dianthicola strain A5573 purified genomic DNA during reaction preparation. This is referred to as a spiked test and was completed to observe any cross reactions with host material. For 4 th LOD assay, a synthetic DNA fragment was developed from Genewiz to confirm the LOD with higher accuracy. The synthetic DNA fragment was designed as mentioned by Arif et al. [39]. As with the purified genomic DNA, the fragments were 10-fold serially diluted (10 9 to 10 1 copies) with water. Non-template controls were included in each LAMP assay.

Multi-operator validation tests
Robustness of the developed LAMP assay was validated by a multi-operator test. Three operators independently performed the assay following the developed protocol. Each operator completed a blind test with six samples (three D. dianthicola isolates (PL23, PL24 and PL25), and three isolates (A5582, A5150 and A6159) from exclusivity panel, and non-template control).

Primer design and in silico specificity
Six LAMP primers were designed with PrimerExplorer V5 using the alcohol dehydrogenase gene. Whole genomes of D. dianthicola and other closely related bacteria were aligned to identify alcohol dehydrogenase as the uniquely present genomic region in D. dianthicola (Dobhal and Arif, unpublished information). Using the NCBI GenBank BLASTn tool, primers showed 100% identity with 100% query coverage for D. dianthicola strains only (Table 3).

Isolate identity confirmation
The detailed description of 16 bacterial strains used in inclusivity panel and 56 strains used in the exclusivity panel for developing an accurate LAMP diagnostic for D. dianthicola is presented in Tables 1 and 2, respectively. Identity confirmation of D. dianthicola and other bacterial strains were done by sequencing the sense and antisense strands using forward and reverse primers. Isolate identities were confirmed using the NCBI BLASTn tool. The accession numbers, MK189263-MK189274, MK243480-MK243481, and MK202803 -MK202805, of the submitted sequences were obtained during this study and are presented in Tables 1 and 2. Other accession numbers presented in Tables 1 and 2 were obtained during the other studies in the lab.

LAMP assay specificity determination
Specificity of the designed primers was assessed using 16 different D. dianthicola strains isolated from distinct geographic locations (Table 1). Additionally, specificity was tested with the exclusivity panel consisting of 56 other bacterial strains of Dickeya sp., Pectobacterium sp., other potato pathogens and saprophytes, and distant relatives ( Table 2). All 16 D. dianthicola strains in the inclusivity panel were amplified by the LAMP assay (Table 1). Conversely, no amplification was observed for bacterial DNA from the exclusivity panel (Table 2). Results for inclusivity and exclusivity panel strains were confirmed using three different strategies (Fig 2). The first approach included a qPCR thermocycler-based fluorescence measurement and melting curve analysis. Fig 2A depicted  ) was observed, indicated that non-specific products were not present. Thus, the designed primers were highly specific and did not form non-specific products and/or primer-dimers. The second approach for amplification confirmation was a colorimetric based detection (Fig 2B). In this procedure, SYBR Green I was added to each tube after reaction completion. Only LAMP positive amplification turned bright green from orange and displayed fluorescence under UV light (Fig 2C). In contrast, non-amplified samples produced an orange color with no fluorescence. Finally, the third approach was electrophoresis; a 2% agarose gel stained with ethidium bromide was used to electrophorese the LAMP product. Positive amplification was indicated by the presence of a smear-like pattern ( Fig  2D). All strains of the exclusivity panel showed no smear or band pattern on the gel, indicating no amplification. During all three confirmation tests neither healthy potato leaf tissue DNA nor the non-template control (NTC; water control) exhibited positive amplification.

Specificity with naturally and artificially infected samples
The LAMP assay was evaluated using D. dianthicola DNA extracted from naturally and artificially infected potato plants. Additionally, P. carotovorum subsp. carotovorum and P. carotovorum subsp. brasiliense DNA isolated from artificially infected potato plants were also used. DNA extracted from D. dianthicola infected plants produced a melt curve and a color change from orange to bright green after the addition of SYBR Green I, and fluorescence under UV light. On the other hand, potato plants artificially infected with Pectobacterium sp. did not produce any positive results (Fig 3). No amplification was observed in non-template control and healthy potato DNA. No false positives or false negatives were observed.

Limit of detection determination
LOD or sensitivity of the developed assay was determined using a 10-fold serially diluted pure culture of D. dianthicola before DNA isolation; LOD was 10 CFU/ml (Fig 4). Other LOD experiments were performed from 10-fold serially diluted D. dianthicola genomic DNA or synthetic DNA; assays were performed three times; each time, the detection limit was 1 pg. Additionally, spiked assays were performed by adding 1 μl of healthy potato genomic DNA into the LAMP reaction containing 10-fold serially diluted D. dianthicola genomic DNA to confirm the inhibitory or background effect of the host genomic DNA. The spiked assays also detected 1 pg of D. dianthicola DNA; host DNA did not show any adverse effect on LAMP assay performance (Fig 5). LOD assay was also performed using the 10-fold serially diluted synthetic DNA fragment containing the primer target sites to confirm the detection limit; the assays detected down to 1,000 copies (S1 Fig). NTC was included in each sensitivity assay.

Field applicability
DNA was extracted from D. dianthicola infected, greenhouse-grown potato plants using a completely field-deployable plant material lysis kit. The LAMP reactions were incubated in a  heating block at 65˚C. The obtained results were in 100% agreement to the results observed using a real-time qPCR machine. Results were reproducible and obtained in 20 minutes ( Fig  6). Visualization of LAMP products with SYBR Green I revealed product color change to bright green (D. dianthicola; positive) or orange (negative). NTC was included in each run; no false positives or negatives were detected.

Multi-operator validation tests
Multi-operator validation tests were performed by three different and independent operators with a blind panel of six different samples to confirm the reproducibility and robustness of the developed LAMP assay. The samples consisted of genomic DNA from D. dianthicola strains (PL23, PL24 and PL25), D. solani (A5582), P. parmentieri (A6159) and, E. cloacae (A5150). Non-template control was included with the six samples as negative control. The obtained results from all three operators were in 100% agreement with previously obtained results. No false positives or false negatives were detected.

Discussion
We have developed a LAMP assay that is rapid, sensitive, and specific for detection of D. dianthicola. This phytopathogen is of great concern because it is highly quarantined against and causes the destructive disease, blackleg of potato. Other molecular techniques have been developed for specific detection of D. dianthicola, but lack specificity or portability [40, 41]. Our LAMP assay has shown to be not only in-field usable, but also rapid, which is important for produce and plants that are time-sensitive commodities. Specificity of the LAMP assay was tested using strains present in inclusivity (16 D. dianthicola) and exclusivity (56 other bacteria) panels. Analysis of the melt curve obtained in the qPCR displayed homogeneous melt peaks around 92.5˚C exclusively for D. dianthicola while no melting curves were observed for the non-target bacterial strains (Fig 2A). Additionally, no irregular curves formed below the mean temperature (92.5˚C) suggesting a lack of primerdimer formation or cross-reaction with other targets, demonstrating the high specificity of our primers. The LAMP assay amplified all D. dianthicola strains (Table 1) and detected no amplification of non-target bacteria (Table 2), indicating high assay specificity. Other molecular detection methods exist for Dickeya sp. [37,41], but either lack specificity or were not tested for field applicability. In another assay, developed primers detected all D. dianthicola strains, but only Dickeya isolates were used for testing [40]. We incorporated closely related genera such as Pectobacterium and other potato pathogens to ensure that developed LAMP assay was exclusive of bacteria with similar genes or genomes. Moreover, no field-deployable LAMP assay exists for the specific detection of D. dianthicola. Consequently, the inclusivity and exclusivity panels indirectly confirmed the target, alcohol dehydrogenase gene, as unique to D. dianthicola. This unique gene was identified by performing comparative whole genome analyses of D. dianthicola, Dickeya species, and other closely related genera (Dobhal and Arif, unpublished information). In TaqMan qPCR assay, the authors targeted the dnaX gene, but were not able to detect all target D. dianthicola strains [41]; also, dnaX gene is not completely specific to D. dianthicola. Additionally, a study developed a LAMP assay targeting a region of the mglC gene, but the assay was limited to detection of the genus Dickeya [37]. Identifying a unique target sequence is imperative to developing a robust and highly specific assay [22,23].
Field-testing was completed with field-deployable DNA extraction kits and a heat block (65˚C). When compared to LAMP reactions incubated in a qPCR machine under specificity panel conditions, results were 100% comparable within 20 minutes when SYBR Green I stain was added for visualization of amplification products (Fig 6). Additional visualization under UV light and through gel electrophoresis confirmed accurate amplification using the heat block. Consequently, we demonstrated that developed LAMP assay equipment could be simplified to a steady heat source and can be performed in field. Similarly, Larrea-Sarmiento et al. [23] reduced complexity and portability through use of the field-deployable, portable BioRanger that detected target bacteria within 15 minutes. In contrast, other developed molecular detection techniques [40,41] are time consuming and require complex tools. Reducing complexity and time is important for use by any operator at point-of-care sites. Moreover, simplifying the machinery adds to the cost-effectiveness of the protocol.
LOD of LAMP assay was confirmed by performing four independent sensitivity tests: 1) 10-fold serially diluted D. dianthicola pure culture; 2) 10-fold serially diluted D. dianthicola DNA; 2) 10-fold serially diluted D. dianthicola DNA spiked with host DNA; and 3) 10-fold serially diluted synthesized DNA fragment (Figs 4 and 5 and S1). The limit of detection for sensitivity and spiked sensitivity tests were consistent to 1 pg and for synthesized targets up to 1,000 copies (Figs 5 and S1). However, Yasuhara-Bell et al. [37] had a detection limit of 5 pg for Dickeya sp. and detection time varied depending on the type of sample DNA (purified, cultured, or crude). Detection using the 10-fold serially diluted cells followed by DNA isolation, showed high sensitivity (10 CFU/ml) compared to the sensitivity performed using 10-fold serially diluted genomic DNA-the lower LOD could be the result of quantification method used, that is NanoDrop. NanoDrop does not provide precise quantification of the double stranded DNA, and we have experienced this in our lab. But the method used to determine the CFU/ml was very accurate since colonies from each dilution were recalculated on media plate. Before DNA isolation from each dilution, 100 mg of plant tissues were added to mimic as the real infected sample. In our study, three operators independently performed the LAMP assay with unknown samples; all operators produced concordant results, confirming the high robustness of the developed LAMP assay.
LAMP assays are comparatively prone to cross contamination because of the high number of copies produced during amplification. However, contamination can be reduced by adding the detection dye before the reaction starts or devising a method to release the dye after reaction completion [42]. Consequently, including the detection dye in a prepared reaction tube reduces complexity as well as increases in-field usability and portability. Here we have demonstrated that LAMP assays can be simplified to 3 steps: DNA extraction with Optigene DNA purification kit, incubation in a heat block, and addition of detection dye for visualization ( Fig  6). This feature of LAMP is convenient in low-resource field situations where conventional DNA or RNA extraction prior to diagnostic testing is impossible. Our LAMP assay lays the groundwork for not only D. dianthicola diagnostics, but also for other pathogens. Ultimately, the developed detection assay can be incorporated in diagnostics for securing our borders against phytopathogens that threaten food security and economies worldwide.