Development of a reverse transcription recombinase polymerase based isothermal amplification coupled with lateral flow immunochromatographic assay (CTV-RT-RPA-LFICA) for rapid detection of Citrus tristeza virus

Tristeza is a highly destructive disease of citrus caused by the phloem-limited, flexuous filamentous Citrus tristeza virus (CTV) in the genus Closterovirus and the family Closteroviridae. It has been a major constraint for higher productivity and has destroyed millions of citrus trees globally. CTV is graft transmissible and spread through use of virus infected nursery plants. Therefore, virus detection by using specific and reliable diagnostic tools is very important to mitigate disease outbreaks. Currently, the standard molecular techniques for CTV detection include RT-PCR and RT-qPCR. These diagnostic methods are highly sensitive but time consuming, labor intensive and require sophisticated expensive instruments, thus not suitable for point-of-care use. In the present study, we report the development of a rapid, sensitive, robust, reliable, and highly specific reverse transcription-RPA technique coupled with a lateral flow immunochromatographic assay (CTV-RT-RPA-LFICA). RT-RPA technique was standardized to amplify the coat protein gene of CTV (CTV-p25) and detect double labeled amplicons on a sandwich immunoassay by designing specific labeled primer pair and probe combinations. The optimally performing primer set (CTRPA-F1/CTRPA-R9-Btn) and the corresponding TwistAmp nfo probe (CTRPA-Probe) was optimized for temperature and reaction time using purified cDNA and viral RNA as template. The sensitivity of the developed assay was compared with other detection techniques using in vitro-transcribed RNA. The efficacy and specificity of the assay was evaluated using CTV positive controls, healthy samples, field grown citrus plants of unknown status, and other virus and bacterial pathogens that infect citrus plants. The RT-RPA-LFICA was able to detect ≤ 141 fg of RNA when cDNA used as a template. The assay detected ≤ 0.23 ng/µl of CTV RNA when directly used as template without cross-reactivity with other citrus pathogens. Best results were achieved at the isothermal temperature of 40 °C within 15–20 min. The study demonstrated that RT-RPA-LFICA has potential to become an improved detection technique for end users in bud-wood certification and quarantine programs and a promising platform for rapid point-of-care diagnostics for citrus farmers and small nurseries in low resource settings.


Scientific Reports
| (2020) 10:20593 | https://doi.org/10.1038/s41598-020-77692-w www.nature.com/scientificreports/ most destructive epidemics occurred in Argentina, California, Brazil, Florida, Spain, Israel, and Venezuela 4 . Phloem-limited long flexuous filamentous virions of CTV (2000 × 11 nm) contained single stranded plus sense RNA of approximately 19.3 kb and organized into 12 ORFs (with UTRs at the 5′ and 3′ termini) that potentially encode at least 19 proteins [3][4][5][6] . Different aphid species, Aphis gossypii Glover, Aphis (Toxoptera) citricidus Kirkaldy and Aphis spiraecola Patch act as vectors and transmit CTV from infected to healthy plants in citrus groves in a semi-persistent manner 3,7 . Consequently, reduced production and fruit quality and increase in disease severity result in citrus decline. The symptoms of Tristeza (stem pitting, vein clearing, vein flecking, stunting, slow decline, and quick decline) often were mistaken with other diseases and nutrient deficiency 8 . CTV, similar to another major citrus pathogen 'Candidatus Liberibacter spp. ' (causal agent of citrus greening/HLB), is graft transmissible and spread to other areas by propagation of infected buds. Therefore, dispersal of this pathogen could be reduced by use of healthy propagation material. Indexing of mother-plants using specific and reliable molecular detection techniques in the nursery would be an important step to prevent its spread 9 .
In addition to biological indexing, many serological and molecular techniques have been employed for the detection of CTV such as, enzyme-linked immunosorbent assays (ELISA), dot immunobinding assays (DIBA), RT-PCR, RT-qPCR, electron microscopy, and Loop-mediated isothermal amplification (LAMP) [10][11][12][13][14][15][16] . These methods (excluding LAMP), have certain limitations inherently i.e. time consuming, labor intensive, require sophisticated expensive equipments, specific technical expertise and are not able to be used at point-of-care. Some disadvantages of LAMP include need for high temperature (65 °C), more amplification time, susceptible to carryover contaminations, and complex in primers design [17][18][19][20] . During development of any new detection method, concerns must account for sensitivity, specificity, simplicity, cost, robustness and rapidity 21 . The recombinase polymerase amplification (RPA) is a nucleic acid amplification technique that depends on the extension of primers induced by the recombination process and could be performed at isothermal temperature (37-42 °C) within 15-25 min 19,22 . RPA overcomes the limitations of many current methods as it does not need expensive instruments, provides rapid and reliable results under low-resource conditions and therefore one of the most promising new molecular diagnostic technologies 21 . The specific combination of enzymes involved in an isothermal RPA reaction are a recombinase, a strand displacing DNA polymerase and a single strand binding protein (SSB). The principle of the RPA technology mainly relies on the ATP dependent recombinase enzyme, which forms a nucleoprotein filament by binding the primers and scanning for the homologous double stranded DNA (dsDNA) sequence of the template to facilitate strand exchange at cognate sites. The SSB protein stabilizes the resulting D loop structure to perform the extension at the 3ˈ end of the invading primer using the complementary strand as a template by DNA polymerase I and accomplishes exponential amplification by cyclic repetition [23][24][25][26][27] . The amplified end product of the RPA reaction can be detected by different methods including agarose gel electrophoresis, colour-based visual detection, real-time fluorescence analysis with exo or fpg probes, and lateral flow immunochromatographic assays (LFICA). The LFICA-based detection system requires a Twist Amp nfo probe labeled with an antigenic molecule such as FAM/FITC/digoxigenin at the 5′ termini and labeled reverse primers with an antigenic biotin molecule at the 5′ termini. The amplified amplicons using these Twist Amp nfo probe and biotin labeled reverse primers can be detected on 'sandwich immunochromatographic' assays such as, the PCRD nucleic acid detector and Milenia GenLine HybriDetect strip (TwistDx Limited, Cambridge, UK) 27,28 . The different types of RPA-based assays have been reported to detect pathogens in animals [29][30][31] , humans 32 37 , Phytophthora spp. 38 and begomoviruses 22 . The aim of the present study was to develop a novel, rapid, robust and highly specific reverse transcription-RPA technique coupled with lateral flow immunochromatographic assay for CTV (working principle is illustrated in Fig. 1) and evaluate its efficacy in comparison with RT-PCR and RT-qPCR.

Results
Screening of CTV-RT-RPA-LFICA primers and specificity assessment. Forty different combinations of forward and reverse primers (4 forward and 15 reverse) targeting a portion of the coat protein gene of CTV were analyzed for specificity and cross reactivity using different strains of CTV and other citrus pathogens with primer-BLAST software. During in silico analysis it was observed that all primer sets were unable to find complementary regions against other citrus infecting pathogen except CTV. Among these, 26 combinations of primers were identified as capable of amplifying target CTV RNA by conventional RT-PCR (Fig. 2). Optimally working primer sets from the initial RPA assays (CTRPAF1/R1, CTRPAF2/R1 and CTRPAF3/R1) were selected for RT-RPA-LFICA. The CTRPA-R9-Btn reverse primer was selected as it showed the minimum complementary energy (ΔG) − 4.41 kcal/mole between reverse primer and probe. The CTRPA-F1/R9-Btn combination was identified as the most optimally performing primer set which consistently amplified a ~ 165 bp specific region of the CTV-p25 gene and was therefore used for further optimization. The CTV positive control and the negative sample used in the present study were further confirmed using conventional RT-PCR. An expected size of ~ 672 bp amplicons with CN150/CN151 primers and ~ 630 bp with P23RBP-F/R primers was observed in CTV infected samples (A1, A2, A3, A4, M1, M2, M3, M4, N1, N2, N3 and N4) and no band was observed in the negative controls (HA1, HA2, HM1, HM2, HN1, HN2). However, amplicons of only six representative samples (A1, A2, M1, M2, N1 and N2) along with positive and negative controls were separated on 1% agarose gel (Fig. 3A,B). The CTV positive and negative samples were used to evaluate the specificity and efficacy of the RT-RPA primers (CTRPA-F1/CTRPA-R9) by conventional PCR. The expected amplification product of ~ 165 bp was observed in CTV positive samples whereas no amplification was observed in the healthy samples and non-template control (Fig. 3C). The PCR products amplified by RT-RPA primers were purified by gel elution and sequenced. In silico analyses confirmed sequences were specific for CTV. Optimization of the CTV-RT-RPA-LFIC assay. The CTV-RT-RPA-LFIC assay was optimized using synthesized cDNA and the total RNA isolated from CTV positive citrus plants. A PCRD nucleic acid detector was used to capture and detect CTV specific amplified double labeled products generated by RT-RPA. Upon application of the amplified products on the sample port, the carbon conjugated anti-biotin antibodies react with the Dig/biotin labeled amplicons. The complex of carbon conjugated antibody-Dig/biotin amplicons get captured at the test line (T-lines) and control line (C-line) by manifestation of a coloured visible line. The appearance of both lines occurring simultaneously within 2 min after application of the amplified product was considered as a positive result for CTV whereas development of only the control line indicated negative results (Fig. 4). It was observed that the sample with higher CTV titer required minimum time (within 60 s) to develop a more intensevisible line (T-lines). It was also observed that RNA templates required more incubation time (20-25 min) than cDNA as initial template. The optimal results of the CTV-RT-RPA-LFIC assay were observed at 40-42 °C  sandwich immunochromatographic assay consistently detected target cDNA of CTV up to 10 -5 serial dilution synthesized from 141 ng of in vitro-transcribed RNA as the initial template. A faint band was observed for the 10 -6 serial dilution of target cDNA (corresponding to 3.77 × 10 5 RNA copies). The test (T-line) band intensity approximately correlated with initial template concentration. The detection limit of the CTV-RT-RPA-LFIC assay was ≤ 141 fg of RNA when converted into cDNA and used as a template (3.77 × 10 5 RNA copies) (Fig. 6A). However, the assay detected ≤ 0.23 ng/µl (6.288 × 10 8 RNA copies) when RNA was used directly as template. The detection limit for conventional RT-PCR was nearly 10 -5 serial dilution (3.77 × 10 6 RNA copies) of target cDNA ( Fig. 6B) and the detection limit of TaqMan real-time PCR assay was recorded near > 10 -8 serial dilution of target cDNA with Ct value > 33.3 (3.773 × 10 3 RNA copies) (Fig. 7). The primer pair was modified such that they contained unique nucleotide stretches and did not mix with the template before adding enzyme to avoid any non-specific binding of recombinase enzyme during the reaction. Specificity analysis showed that the developed CTV-RT-RPA-LFIC assay was highly specific to CTV and failed to detect any other non target citrus pathogens ( Fig. 8).  (Table 1) with the kappa value of 0.926. These results indicate that the developed assay demonstrated an excellent diagnostic agreement with RT-qPCR and would be effective for the detection of CTV in field samples.

Discussion
CTV is one of the most economically important pathogens of citrus and has destroyed millions of citrus trees worldwide 1 . The virus infection causes reduction in yield and quality of citrus fruits and induced stem pitting and devastating quick-decline symptoms. Citrus is mainly a vegetatively propagated crop and the major pathogens are transmitted through disease-infected buds and propagating planting material. In the field, the horizontal transmission is by aphid vectors in a semi-persistent manner. To prevent outbreaks of the disease, indexing of planting material is a crucial step that requires a specific and reliable virus detection techniques 9 . Numerous techniques have been developed for CTV detection but most of them have limitations viz., time consuming, www.nature.com/scientificreports/ requirement of expensive equipment and trained personnel. RT-PCR and RT-qPCR are the most acceptable techniques for CTV detection but require well setup laboratory, equipment and personel. Therefore, there is need to develop a rapid, robust and reliable on-site detection technique to assist in the certification of virus-free planting material. The RPA approach is an emerging isothermal, low cost, rapid, and point-of-care diagnostic tool 23 . It is a highly sensitive, reliable nucleic acid based method and has become a rapid detection tool for many pathogens including viruses 34 , bacteria 19,27 , and Phytophthora species 39 . It is also used for detection of RNA viruses without the need for a separate step to synthesize cDNA by RT-RPA 21,36,37,40 . This technology has the potential to be a promising alternative to RT-qPCR 41 . Enzymes (recombinase, SSB and strand displacement DNA polymerase) are required for exponential amplification of the target template 23 . To target the RNA template, extra reverse transcriptase enzyme need to be added to the reaction. Present study is first to report the development of a robust reverse transcription recombinase polymerase-based isothermal amplification technique coupled with lateral flow immunochromatographic assay (CTV-RT-RPA-LFICA) for the rapid detection of CTV.
The specificity of the optimized RT-RPA primers selected to amplify the CTV specific coat protein gene (CTV-p25) demonstrated by comparison with conventional RT-PCR and sequencing of PCR products as cognate gene specific. The analysis of the products of basic RPA by agarose electrophoresis requires an extra chloroform/ isoamyl alcohol purification step to reduce the crowding and complexity of the amplified RPA product on the   www.nature.com/scientificreports/ agarose gel for better band visibility compared to non-purified RT-RPA products (Fig. 3D). The specificity of the RT-RPA primers were also judged by in silico analysis (primer BLAST) that specifically detected the target pathogen and showed no cross reactivity with other major citrus pathogens. Hetero-dimer analysis using the Oligo Analyzer tool was performed between probe and reverse primers to obtain the realistic results on lateral flow immunochromatographic assay. Efforts were also made to obtain a minimum ∆G (free energy of the oligo sequence binding to its complement site) between probe and reverse primer. The ∆G value of − 4.41 kcal/mole for probe and CTRPA-R9 was observed as best to achieve reliable results owing to less complementarity between reverse primer and probe. CTV-RT-RPA-LFICA developed in the present investigation was rapid, needed 20-25 min for amplification and 5 min for visualization. Application of RT-RPA amplified products on the sample port, the test line (T-lines) and control line (C-line) were visible as coloured bands with CTV positive samples simultaneously within 2 min, whereas development of color with only the C-line indicated negative results (Fig. 4). The assay needs less RT-RPA product (0.5 µl) for end point detection, the visualization process on the PCRD detector, as compared with other techniques. The results also suggested that RNA as an initial template needed more reaction incubation time (20-25 min) compared to cDNA as template (15 min). Higher pathogen titer in the sample would require much less time (~ 60 s) to develop more intense color band in T-lines. The reaction works at low isothermal temperature of 40 °C in a simple dry bath without expensive thermal cyclers (Fig. 5A). The reaction also could  42 . The reaction needed a single set of primers, whereas LAMP needs as many as four to six primers 39 . The quick lateral flow immunochromatographic assay optimized using the Twist Amp nfo probe which saved an additional 60 to 80 min compared to RPA where products have to be visualized by agarose gel electrophoresis. The sensitivity of RT-RPA-LFICA was shown to be equal or better compared to conventional RT-PCR, but less sensitive compared to the TaqMan-RT-qPCR (Figs. 6, 7 and Supplementary Fig. S1). The ability of RT-RPA-LFICA to detect the CTV using genomic RNA as template with a limit up to 0.23 ng/µl of RNA was demonstrably a strong advantage compared to RT-PCR and TaqMan-qPCR. Specificity analysis showed that the developed CTV-RT-RPA-LFIC assay is highly specific to CTV and did not cross react with any other non-target citrus pathogens (Fig. 8). The specificity of the assay was further validated by testing seventy-four field grown CTVsuspected samples and compared with RT-PCR and RT-qPCR. These results confirm the robustness and specificity of the newly-designed RT-RPA-LFICA primers-probes and demonstrated excellent diagnostic agreement with RT-qPCR (kappa value = 0.926). The major advantage of this technique is that the result can be easily judged as positive or negative visually and thus a great potential to be used as a point-of-care diagnostic tool and a valuable tool for nursery citrus bud wood certification programs.

Design of CTV-specific primers and nfo-probe for RT-RPA-LFICA. The primers and probe used in
the CTV-RT-RPA-LFICA were designed using Twist Amp nfo assay design manual guidelines (www.twist dx.co. uk) to amplify a segment of the coat protein gene of CTV (CTV-p25). Number of coat protein gene sequences of CTV were retrieved from the GenBank and aligned using the software MEGA7. The highly conserved region of coat protein gene (CTV isolate MD; GenBank KY011909.1) was targeted and several sets of primer combinations (4 forward and 15 reverse) were designed and custom synthesized from Integrated DNA Technologies (IDT, Iowa, USA). The standard parameters of RPA primer design were taken into consideration and in silico specificity was considered using primer-BLAST software (www.ncbi.nlm.nih.gov). After initial screening by RT-    To confirm the presence of CTV, RT-PCR was conducted using a coat protein gene specific primer pair (CN150/CN151) as previously described by Warghane et al. 43 The PCR products were visualized by 1% agarose gel electrophoresis and the UV GelDoc system (G: Box Syngene). Another primer set, P23RBP-F/R, specific to the RNA binding protein (CTV-P23) gene located at the 3′-terminus and adjacent to the untranslated region of the RNA genome of CTV 44 was designed and custom synthesized (IDT, Iowa, USA). It was also used to validate the RT-RPA technique (Table 3). PCR amplification was performed using the P23RBP-F/R primer set according to Kokane et al. 45 .

Generation of in vitro-transcribed RNA standard for CTV-p25.
To assess the sensitivity of the developed RT-RPA, an in vitro-transcribed RNA standard was generated by using a MEGAscript T7 transcription kit (Invitrogen). Total RNA was isolated from CTV infected plants and converted into single stranded cDNA using coat protein gene specific reverse primer (RPARNA-P25R2) which contains the RNA polymerase T7 promoter sequence. The cDNA was used as template for PCR amplification of the 700 bp coat protein gene using RPARNA-P25F1/R2 primers ( Table 3). The PCR amplified products were checked on 1.5% agarose gel, eluted and sequenced. The sequence validated PCR product (120 ng) containing the T7 RNA polymerase promoter sequence was used as template for in vitro RNA synthesis. The in vitro RNA transcription was performed at 37 °C for 4 h in a mix consisting of 3 mM of each T7 NTPs and 2 µl Enzyme mix in 1 × T7 reaction buffer. After reaction completion, the RNA was treated with 2U TURBO DNase (2U/µl) and recovered by lithium chloride precipitation according to manufacturer's instructions. The copy number of in vitro synthesized RNA was determined using the formula, RNA copy number = Moles of ssRNA × Avogadro's number (6.022 × 10 23 ). The moles of ssRNA was calculated as mass of ssRNA (g)/[(number of ribonucleotides of ssRNA × average molecular weight of a ribonucleotide) + 18.02 g/mol].
Primer optimization and screening based on RT-RPA and RT-PCR. The designed sets of primer combinations (4 forward and 15 reverse) were optimized using conventional PCR and RT-RPA. The most optimally performing CTV specific primer sets were used for RT-RPA-LFICA optimization. For screening and vali- Table 3. Primer and probe sequences used for TaqMan-qPCR assay, conventional RT-PCR and generation of in vitro RNA standard in the present study.