Optimization of a magnetic capture RT-LAMP assay for fast and real-time detection of potato virus Y and differentiation of N and O serotypes

Potato virus Y (PVY) infection has been a global challenge for potato production and the leading cause of downgrading and rejection of seed crops for certification. Accurate and timely diagnosis is a key for effective disease control. Here, we have optimized a reverse transcription loop-mediated amplification (RT-LAMP) assay to differentiate the PVY O and N serotypes. The RT-LAMP assay is based on isothermal autocyclic strand displacement during DNA synthesis. The high specificity of this method relies heavily on the primer sets designed for the amplification of the targeted regions. We designed specific primer sets targeting a region within the coat protein gene that contains nucleotide signatures typical for O and N coat protein types, and these primers differ in their annealing temperature. Combining this assay with total RNA extraction by magnetic capture, we have established a highly sensitive, simplified and shortened RT-LAMP procedure as an alternative to conventional nucleic acid assays for diagnosis. This optimized procedure for virus detection may be used as a preliminary test for identifying the viral serotype prior to investing time and effort in multiplex RT-PCR tests when a specific strain is needed.


Introduction
Potato virus Y (PVY) is the type member of the genus Potyvirus, family Potyviridae [9]. It has recently been listed among the 10 most important plant viruses in the world [12] and recognized as the major virus infecting potato [8]. A major challenge in the control of PVY infection is the continual emergence of new recombinants derived mostly from the ordinary PVY O strain and the necrotic PVY N strain [2,5,7]. At least nine PVY strains are currently known and differ at the biological, serological, and molecular levels but can be grouped into two serotypes: the O serotype, which includes PVY O , PVY N:O , PVY N-Wi , and PVY C , and the N serotype, which includes PVY N , PVY E , PVY Z , PVY NA-N and PVY NE11 [8]. Due to their economic importance [2,8], the identification of PVY isolates by strain, or at least by serotype, is needed. Many procedures for the detection of PVY in potato leaves and tubers have been reported [3,6,[13][14][15][16]. Due to their easy adaptation to automation, their high specificity, and their ability to detect viral strains in a single reaction, real-time-RT-PCR-based assays appear to be ideal successors of the double sandwich enzyme-linked immunosorbent assay (DAS-ELISA) that is routinely used in the grow-out test of dormant tubers for the diagnosis of potato viruses for seed certification [3]. However, such methods have not yet been adapted for routine diagnostic schemes.
Here, we describe a streamlined magnetic capture version of the reverse transcription loop-mediated amplification (MC-RT-LAMP) assay that we optimized for fast detection Abstract Potato virus Y (PVY) infection has been a global challenge for potato production and the leading cause of downgrading and rejection of seed crops for certification. Accurate and timely diagnosis is a key for effective disease control. Here, we have optimized a reverse transcription loop-mediated amplification (RT-LAMP) assay to differentiate the PVY O and N serotypes. The RT-LAMP assay is based on isothermal autocyclic strand displacement during DNA synthesis. The high specificity of this method relies heavily on the primer sets designed for the amplification of the targeted regions. We designed specific primer sets targeting a region within the coat protein gene that contains nucleotide signatures typical for O and N coat protein types, and these primers differ in their annealing temperature. Combining this assay with total RNA extraction by magnetic capture, we have established a highly sensitive, simplified and shortened RT-LAMP procedure as an alternative to conventional nucleic acid assays for diagnosis. This 1 3 and discrimination of PVY by their serotypes. To contrast this assay with previously reported RT-LAMP variants for the detection of PVY [1,10,11], we designed a primer set that amplified a conserved region within the coat protein gene that contains two single and two double nucleotide signatures typical for O and N coat protein types, which differ in their annealing temperatures in the RT-LAMP assay depending on the serotype of the virus tested. This allowed us to rapidly distinguish the PVY O and PVY N types in one reaction. In addition, the optimization of the RNA isolation step by magnetic capture, both from potato leaves and tubers, shortened the procedure while retaining its high sensitivity. The present study offers a fast alternative to conventional nucleic-acid-based assays and to the routinely used DAS-ELISA for grow-out tests, which can be time consuming, for the direct identification of PVY isolates.

RNA isolation
Total RNA was extracted using different methods as indicated. The initial sap was extracted from approximately 100-200 mg of potato leaf or tuber tissue.

Silica capture procedure
Total nucleic acid was purified using the silica capture procedure as described by Zacharzewska et al. [16]. Nine Triton X-100) was added to 50 µl of silica beads and mixed with 100 µl of undiluted plant sap. After a 10-min incubation at room temperature with gentle shaking, the samples were centrifuged at 12,000g. The pellet was next washed twice with 1 ml of L2 buffer (5.25 M guanidinium thiocyanate, 50 mM Tris-HCl, pH 6.4), twice with ethanol, and once with acetone. The final pellet was dried, resuspended in 50 µl of RNase-free water, and incubated for 10 min at 56 °C. The samples were next quickly centrifuged, and the supernatants containing total nucleic acids were collected.

Manufacturer's magnetic capture procedure
Total RNA was purified from 100 µl of undiluted sap using a Novabeads Total RNA Purification Kit (Novazym Polska s. c.) according to the manufacturer's protocol.

Modified magnetic capture procedure
All steps and buffers were as described as above for the silica capture method except that the silica beads were replaced with 25 µl of magnetic particle suspensions from the Novabeads Total RNA Purification Kit (Novazym Polska s. c.). All centrifugation steps were replaced by the use of a magnetic stand.

Shortened magnetic capture procedure
This is a shortened version of the modified magnetic capture procedure. One hundred microliters of undiluted sap was added to 25 µl of Novabeads magnetic particles (Novazym Polska s. c.), which were resuspended in L6 buffer. The samples were incubated at room temperature with gentle shaking for 10 min. Next, the magnetic particles were captured on a magnetic stand, and the supernatants were removed. The beads were washed with L2 buffer and then with 70% ethanol. The beads were finally resuspended in 50 µl of RNasefree water and incubated for 5 min at 56 °C prior to collection of the supernatants. For RT-LAMP and real-time RT-PCR, total nucleic acid preparations were made without prior DNAse treatment. Each RNA isolation was conducted with at least two replicates and was repeated independently at least three times.

RT-LAMP assay
Different primer sets were tested in this study, including the N set developed by Nie [10], the Y5 set developed by Przewodowska et al. [11], and a new Y4 primer set that we designed using LAMP Designer software (Premier Biosoft) as described by Przewodowska et al. [11] (Table 2). All primer sets consisted of four (N set) or six (Y4 and Y5 sets) primers and targeted different regions of the PVY coat protein gene (Fig. 1). The RT-LAMP reaction mixture (10 μl) contained 0.375 µM each of outer primer (F3, B3), 1.5 µM each of inner primer (FIP, BIP), 0.75 µM each of loop primer (LF, LB), 1X Isothermal Master Mix containing proprietary fluorescent dye (Novazym Polska sc.), 0.25 U of AMV reverse transcriptase (Novazym Polska s. c.), and 100 pg of total RNA. The amplification was performed either with a Genie II Ultra rapid amplification instrument (OptiGene Ltd.) or with a CFX96 Touch™ Real-Time PCR Detection System (Bio-Rad Ltd). With the OptiGene instrument, the assay was conducted at 65 °C for 30 min. The annealing temperature (T a ) of the amplified products was analyzed in a slow annealing step (0.05 °C/s) for 5 min, starting at 95 °C and ending at 80 °C, with monitoring of the fluorescence. The thermal profile with the Bio-Rad instrument included 60 cycles of 30 s at 65 °C. The amplification was followed by the analysis of the melting temperature (65 °C to 98 °C, 0.5 °C/s).

Real-time reverse transcription PCR
Real-time RT-PCR was conducted according to the two-step protocol developed by Singh et al. [13]. Total RNA was isolated by short magnetic capture and was reverse transcribed as described above. The real-time PCR was carried out on a CFX96 Touch™ Real-Time PCR Detection System (Bio-Rad Ltd.) using an AmpliQ Real-Time PCR OptiProbe kit or an AmpiQ Real-Time PCR SYBR Green kit (Novazym Polska s. c.). The final reactions (10 µl) contained 2.5 µl of cDNA, 5 µl of AmpiQ Real-Time PCR reagent, 0.15 µl of a 20 µM stock of forward and reverse primers, and 0.15 µl of 10 µM probe in the probe-based assays, or nuclease-free water in the dye-based assays. Different sets of primers were tested, including (i) the PVY-specific primers and TaqMan probe (TM) designed by Singh et al. [13], (ii) the Y4 F3 and B3 primers from the Y4 primer set (TM), combined with the TaqMan probe [13], and (iii) the Y4 F3 and B3 primers from the Y4 primer set in the presence of SYBR Green (SG) (for sequences, see Table 1). The conventional RT-PCR was performed with the Y4 F3 and B3 primers from the Y4 primer set. The thermal cycling steps were 10 min at 95 °C followed by 60 cycles of 10 s at 95 °C and 20 s at 60 °C. Reactions containing only water (no-template control) or cDNA transcribed from total RNA purified from virus-free plants were included in each run as controls. The fluorescence was monitored on the FAM channel and the results were interpreted in terms of Ct (cycle threshold values) determined used default settings in CFX Manager Software (Bio-Rad Ltd.).

Sensitivity of the RT-LAMP assay
The sensitivity of the RT-LAMP assay was determined by two separate approaches. The first approach was based on the isolation of total RNA by silica capture. The total RNA concentration was adjusted to 100 ng/ml, and tenfold dilutions were prepared down to 1 fg/ml. One microliter of the RNA sample was then amplified by RT-LAMP using the Y4, Y5 and N primer sets. The second approach was based on the serial dilution of the sap extracted from PVY-infected and healthy potato plants. We investigated (i) the impact of the isolation method on the sensitivity of the detection of PVY by the Y4-based RT-LAMP and (ii) the sensitivity of the Y4-based RT-LAMP when compared to real-time RT-PCR. The sap dilutions ranged from tenfold (dilution factor = 1 × 10 1 ) to two-million-fold (dilution factor = 2 × 10 6 ). For comparison of the sensitivity of the tested methods, the RNA isolation techniques, or the amplification procedures, total RNA was isolated from all prepared sap dilutions by magnetic capture. These RNAs were subsequently tested by RT-LAMP alone or also by real-time RT-PCR according to the procedures described above.

Comparison of detection methods using field samples
Thirty tubers from 30 offspring plants were harvested after planting secondarily PVY-infected and healthy tubers. The tubers were tested by the Y4-based RT-LAMP method as well as by probe-based and dye-based real-time PCR as described above using the magnetic III protocol for RNA extraction. For the standard grow-out (GO) tests, eye cores were cut, sprouted and grown in a glasshouse to produce 4-to 6 week-old offspring plants. Leaves from these plants were then tested for PVY using the double sandwich enzyme-linked immuno-absorbent assay (DAS-ELISA) according to the manufacturer's protocol (Neogen).

Statistical analysis
Each experiment was conducted with at least two replicates and was repeated independently at least three times. A nontemplate control (sterile water), a negative control (total RNA extracted from a virus-free plant), and a positive control (total RNA extracted from a PVY-infected plant) were included for each LAMP/qPCR run. The time-to-positive (Tp) or Ct values were calculated using the manual baseline settings in Genie II (OptiGene Ltd) or CFX Manager Software (Bio-Rad Laboratories). Mean values and standard deviations were calculated, and the statistical analysis was carried out using GraphPad Prism 6.04 for Windows (GraphPad Software, La Jolla, California, USA, http://www. graphpad.com). An unpaired t-test was used to compare the mean annealing temperature (T a ) values recorded using a

Performance of the RT-LAMP primer sets
The high specificity of the RT-LAMP method relies heavily on the primer sets designed for the amplification of the targeted region. We thus performed a direct comparison of three RT-LAMP primer sets for PVY detection. These included two primer sets (N and Y5) previously reported in the literature [10,11] and the newly designed Y4 set ( Table 2 and Fig. 1), all of which targeted the coat protein region. While all primer sets successfully amplified the PVY targeted region as expected, their reaction rates were significantly different (Fig. 2a). The slowest amplification was observed for the N set, with a time-to-positive (Tp) value of 23 min. The Y5 primers had a Tp value equal to 17 min. The Y4 set gave the fastest PVY detection with a Tp value of 8 min (Fig. 2a). The specificity of all primer sets was confirmed by the lack of amplification from total RNA isolated from a mock-inoculated plant or a no-template control (Fig. 2a).
To compare the detection sensitivity achieved with the different primer sets, we performed a serial dilution of the total RNA, with the final amount of RNA ranging between 100 pg to 1 fg per reaction. We plotted the time-to-positive values for each primer set as a function of the logarithm of the RNA concentration (Fig. 2b), which showed a linear relationship for all primer sets. The highest sensitivity was obtained with the Y4 set, which successfully detected PVY with as little as 0.01 pg of RNA. Lower sensitivities by one (0.1 pg) and ten orders of magnitude (1 pg) were recorded for the Y5 and N sets, respectively.
The specificity of the Y4 primer set against PVY was confirmed by testing total RNA isolated from leaves infected with other viruses commonly infecting potato, including potato virus A (PVA), potato virus M (PVM), potato virus S (PVS), and potato virus X (PVX), along with PVY. As expected, only PVY was detected (Fig. 2c). No peak was detectable for any other viruses. The annealing temperature of the amplified RT-LAMP product was 84.4 °C (Figure 2d).
We concluded that the newly developed Y4 primer set specifically amplified the targeted region of PVY and that it yielded higher sensitivity than the other tested primer sets, including the previously described Y5 set, which had already been reported to be tenfold more sensitive than the RT-PCR and 1000-fold more sensitive than the enzyme immunoassays [11].

Discrimination of the PVY O and N types with the Y4 primer set
The ability to discriminate PVY isolates by strain or at least by serotype is a useful tool in disease diagnosis. We thus tested whether the Y4-based RT-LAMP assay could differentiate the PVY isolates that were positive in the multiplex RT-PCR (Fig. 3). These included different PVY O , PVY N-Wi , PVY N and PVY NTN isolates. We performed the RT-LAMP assay using a Genie II rapid amplification system and the Y4 and the Y5 primer sets (Fig. 4). While no amplification was observed when using total RNA isolated from virus-free plants, strong amplification was detected when using total RNA from all of the PVY-infected plants using both primer sets ( Fig. 4a and b). We next assessed strain specificity by comparing the T a values of each of the amplified products (Fig. 4c and e). The resulting curves revealed that the products amplified from PVY isolates with an O serotype, which included PVY N-Wi and PVY O , had approximately a 0.46 °C variation in their T a values when compared to the amplified products from the N type isolates, which included PVY N and PVY NTN , but this was only observable with the Y4 primer set (p < 0.0001, Table 4). The average T a value was about 84.37 °C for the N-type and 84.83 °C for the O-type (Table 3, Fig. 4c). Such significant temperature variation was not observed with the Y5 primer set (difference, 0.14 °C, p = 0.2191, Table 3, Fig. 4d and f). The average T a value for all amplified products was 86.63 °C. Thus, the Y5 set detected PVY, but in contrast to the Y4 set, it failed to differentiate the PVY isolates (Table 3, Fig. 4d and f).
To determine whether the observed difference in T a values with the Y4 primers ( Fig. 4c and e) correlated with a difference in melting temperature (T m ), all PVY strains were tested using a real-time thermal cycler (Fig. 4g). Our data revealed that the PVY strains with an N-type coat protein sequence had a T m value of 83.65 °C, while strains with an O-type coat protein sequence had a T m value of 84.21 °C (difference, 0.56 °C, p < 0.0001, Table 3, Fig. 4g).
We expanded the RT-LAMP assays with the Y4 primers to include North American PVY O , PVYN N:O , PVY N-NA and PVY NTN isolates (Fig. 4h). The melting curves revealed that the products of the RT-LAMP amplification obtained for PVY O and PVYN N:O with the O-type coat protein gene had a T a value (85.16 °C) that was 0.47 °C higher than that of the products obtained for PVY N-NA and PVY NTN (84.69 °C) (p = 0.0005, Table 3, Fig. 4h).
In summary, the Y4-based RT-LAMP set allowed us to discriminate PVY O and PVY N types in one reaction. Sequence analysis (Fig. 1) showed that the amplified region within the coat protein gene contains two single and two double nucleotide signatures typical for O and N coat protein types, which may be sufficient to contribute in the differences in annealing/melting temperatures that we observed between the N and O types ( Table 3, Fig. 4).

Optimization of the RNA isolation procedure
Our next step was to streamline the RT-LAMP procedure. We thus compared different RNA isolation procedures, including magnetic capture and RNA silica capture, that we previously optimized for RT-PCR [13]. We first compared three versions of the magnetic-capture-based RNA isolation procedure as described in Materials and methods. These included (i) the magnetic capture following the manufacturer's procedure, (ii) a modified magnetic capture with buffers adapted for silica capture [13], and (iii) a shortened modified magnetic capture with fewer washes and incubation steps than the modified capture procedure. We observed that all these procedures resulted in nearly identical amplification profiles in RT-LAMP (Fig. 5a). However, we shortened the procedure from start to completion from 30 and 60 min with the original and modified method, respectively, to 20 min with the shortened isolation approach.
To test whether the technique of RNA isolation would influence the sensitivity of PVY detection by RT-LAMP, we compared the original magnetic capture method and the shorter magnetic capture method with the silica capture Fig. 4 Differentiation of PVY types N and O by the Y4-primer-setbased RT-LAMP. Amplification plots recorded for RT-LAMP reactions with the Y4 (a) and Y5 (b) primer sets and total RNA isolated from plants infected with the indicated European PVY isolates, with total RNA from a virus-free plant (nc), or with water instead of RNA (ntc). Second derivatives of annealing curves recorded for amplicons amplified using the Y4 (c) and Y5 (d) primer sets in reactions described in panels a and b, respectively. Annealing temperature (T a ) ranges of the RT-LAMP products obtained on European PVY strains by amplification with the Y4 (e) and Y5 (f) primer sets. The annealing temperatures were recorded using a Genie II apparatus (g and h). Melting temperature (T m ) ranges of the RT-LAMP products obtained using European (g) or North American (h) PVY strains by amplification with the Y4 primer set. The melting temperatures were measured using the Bio-Rad apparatus. PVY strains with coat protein O or N types are indicated by symbols with a black or white background, respectively. The data are averages from at least two independent experiments performed in triplicate. Error bars indicate standard deviation ◂ Table 3 Statistical analysis of differences between mean annealing (T a ) and melting temperatures (T m ) recorded for O-and N-type PVY isolates by Y5 and Y4 RT-LAMP assays. T a values were determined using a Genie II  method, which resulted in high sensitivity in the RT-PCR assay [13] (Fig. 5b). Our data showed that all isolation methods had the same detection limit (a 1 × 10 5 -fold dilution of sap), but the time-to-positive values were slightly higher with RNA isolated by magnetic capture than with RNA isolated by the silica-based RNA capture procedure (Fig. 5b).
In summary, we reduced the time required for the magnetic capture procedure for PVY detection by RT-LAMP from RNA isolation to assay to less than one hour without compromising the detection sensitivity.

Comparison of the sensitivity of PVY detection by RT-LAMP and by conventional and real-time RT-PCR
Next, we compared the sensitivity of the Y4-based RT-LAMP for PVY detection with that of real-time RT-PCR and conventional RT-PCR using serial dilutions of sap. The data revealed that the RT-LAMP was as sensitive as all investigated variants of the real-time RT-PCR, with the detection of PVY in sap diluted up to 2 × 10 6 fold in fewer than 40 cycles (Fig. 6a). However, the detection of PVY at the highest RNA concentration (1 × 10 2 -fold dilution) required more than 25 cycles for all tested real-time RT-PCR variants but only 19 cycles by RT-LAMP. To contrast, the conventional RT-PCR amplified a PCR product of the expected size in samples diluted only up to 1 × 10 6 -fold (Fig. 6b).
While the RT-LAMP assay shares similar sensitivity with the standard real-time RT-PCR assay, it is clearly faster. The detection of PVY in the highest dilution of sap can be accomplished in 15-60 min by RT-LAMP with its 30-s cycles. In contrast, the completion of the real-time RT-PCR reaction with the same sample requires 2-3 h.

Reliability of PVY detection in dormant tubers
To estimate the reliability of the RT-LAMP for detection of PVY in dormant tubers, tubers collected from secondarily Comparison of the effect of the different RNA isolation methods on sensitivity. Sap from plants infected with PVY was serially diluted up to 2 × 10 6 times, and total RNA was purified from each dilution by the magnetic capture procedure recommended by the manufacturer, by the shortened magnetic capture procedure, and by silica capture. Time-to-positive results are shown as a function of the logarithm of the dilution factor. The data are averages from at least two independent experiments performed in triplicate. Error bars indicate standard deviation  (Table 4). Our data suggest that while the grow-out test remains a reliable method for the detection of PVY in dormant tubers, the RT-LAMP assay clearly outperformed the real-time RT-PCR assay. This may be linked to the requirement for high-quality RNA preparations for PCR-based amplification, which can be a challenge with the low virus titer in dormant tubers [3]. It is worth noting that some differences were observed between the real-time PCR variants depending on whether the detection was dye-or probe-based (Table 4).