Pepino mosaic virus (PepMV) belongs to the genus Potexvirus
and has virions that are non-enveloped flexuous rods of 508 nm in length [1
]. The PepMV genome is a 6410-nucleotide-long single strand of RNA [2
]. It was first isolated in 1974 in Peru from pepino plants (Solanum muricatum
], and was reported as a pathogen of tomato (Lycopersicon esculentum
) in 1999 in greenhouses in The Netherlands [3
]. Shortly afterwards, PepMV became a major pathogen for tomato worldwide, with the first reports of its epidemic expansion in Europe, and later in North and South America [4
]. PepMV can cause a wide range of symptoms that reduce the economic value of tomato crops. The symptoms typically include fruit marbling and leaf or stem necrosis [5
]. The development and intensity of the symptoms are affected by various factors, including virus genotype, climate and tomato cultivar [6
Pepino mosaic virus isolates show considerable genomic diversity, which appears to be linked to the geographic origins, specific hosts and/or the high mutation rates of RNA viruses. These differences appear to increase adaptation of PepMV to new environments [9
]. There are five main PepMV genotypes currently recognized: Peruvian (including the original pepino isolate); European (Eur); American (US1, US2); Chilean (Ch2); and south Peruvian [10
]. For several European countries and in North America, population genetic studies have revealed a shift in the dominant PepMV genotype from Eur to Ch2 [13
]. Aggressive and highly virulent isolates that can cause severe symptoms and significant damage are distinguished from mild isolates by single nucleotide variations [8
]. In addition, mixed infections of different PepMV genotypes have been reported [15
]. Mixed infections can have greater detrimental effects on tomato crops due to symptom expression [15
]. In mixed infections, there is also a high probability of recombination between the PepMV genotypes [15
Pepino mosaic virus is highly infectious, and it is readily transmitted through mechanical means. It can be spread easily by plant-to-plant contact, and during routine cultivation procedures, while seed transmission might have a role in its long-distance spread [19
]. Water has also been confirmed as a source of PepMV infection [20
]. For the prevention of PepMV disease, there is the need for strict hygiene measures, which are currently the main control strategy for the protection of tomato production. However, severe outbreaks of PepMV continue in greenhouse tomato production in many European countries, and thus there is the need for further development of plant protection products. The first vaccine that was developed, PVM-01, contained a well-defined and characterised stable Ch2 mild isolate, and was registered in the European Union under Regulation 1107/2009 [21
The PVM-01 vaccine was designed on the principle of cross-protection, which uses a mild isolate as a protective agent against more aggressive strains. Such cross-protection depends strongly on the genotype, and it is effective only under well-defined conditions [19
]. For the vaccine, there is the need for accurate determination of the concentration of the PepMV. Therefore, reliable detection, quantification and genotype characterisation are the most important tools for disease control. In addition, accurate quantification of individual PepMV genotypes can be used for a variety of other applications, such as for evaluation of the resistance of plants to PepMV infection, of the variability of PepMV concentrations in different plant organs, and of the seasonal variations in virus concentrations, and also for estimation of the concentrations of different PepMV genotypes in a mixed infection.
Reverse-transcription real-time quantitative polymerase chain reaction (RT-qPCR) is at present still the method of choice for quantitative detection of the PepMV genotypes [22
]. This quantification is based on standard curves, which are ideally prepared from certified reference material. As no reference material is available for PepMV, only relative quantification (i.e., between samples) is possible using RT-qPCR. Additionally, due to the high variability of the PepMV isolates, mismatches in primer or probe sequences can influence this quantification. Moreover, reverse transcription and PCR inhibitors can have major effects on PCR based assays.
In contrast to real-time quantitative PCR (qPCR), digital PCR (dPCR) can provide absolute quantification of target sequences without reliance on standard curves. By partitioning the samples into thousands of reactions before amplification, the copy number in the initial sample can be determined based on the number of positive and negative partitions after end-point PCR amplification, using Poisson’s statistics [23
]. As well as eliminating the need for standard curves, dPCR has been shown to be more resilient to inhibitors [24
]. In qPCR, a less-efficient amplification moves the cycle of quantification (Cq) to a higher value, while in dPCR, positive and negative partitions can be separated even with less efficient amplification. Thus, data from dPCR have lower sensitivity to variations in PCR amplification efficiencies. Furthermore, on the basis of this high-level sample partitioning, dPCR can provide very precise and accurate data even at particularly low target copy numbers [26
Today there are several dPCR platforms available, which can be divided into two general groups: emulsion-based or droplet dPCR; and chip-based dPCR [27
]. For the droplet dPCR platforms, the reaction mixture is divided into several individual droplets, while for chip-based dPCR, the reaction is divided into several chambers on a single plate or array. A combination of droplet-based and chip-based technology platforms is provided by the Naica system (Stilla Technologies) for crystal dPCR.
The aim of the present study was to evaluate two different dPCR platforms for quantification of PepMV RNA: QX100 or QX200 Droplet Digital PCR from Bio-Rad (henceforth as ‘QX100/200’); and Naica Crystal dPCR from Stilla Technologies (henceforth as ‘Naica’). We took the well-characterised assays for detection and genotyping of PepMV by qPCR, developed dPCR assays, and evaluated their performances. Despite some differences observed between these two dPCR platforms, which are indicated and discussed here, the newly developed reverse-transcription dPCR (RT-dPCR) assays could be potential candidate for higher order methods for absolute quantification of PepMV.
3. Materials and Methods
3.1. Samples and Sample Preparation
The PepMV genotypes used in this study were Eur, Ch2 and US1. PepMV-Eur (isolate 99901066) was kindly provided by Dr J.Th.J. Verhoeven (NVWA, Netherlands), and PepMV-Ch2 (isolate 1906) and PepMV-US1 (isolate ST 08/008) were kindly provided by Dr Inge Hanssen (Scientia Terrae Research Institute, Belgium). PepMV isolates were propagated on Lycopersicon esculentum cv. ‘Moneymaker’.
Total RNA extraction was carried out for infected and healthy plant leaf material with RNeasy Plant mini kits (Qiagen, Hilden, Germany), according to the manufacturer recommendations, with some minor modifications. Briefly, 200 mg plant material was homogenised in 900 µL lysis buffer (RLT buffer without mercaptoethanol; Qiagen), then 600 µL of homogenate was loaded onto RNeasy Mini Spin columns. RNA elution from these RNeasy Mini Spin columns was carried out as two consecutive applications of 50 µL (i.e., 100 µL total) of RNase-free warm water (65 °C). Buffer controls were included with all isolations (i.e., for negative isolation control), to monitor potential contamination through the extraction procedures.
3.2. Assays Used
The RNA of PepMV was amplified using the PepMV-universal primers and probe described by Ling et al. [16
], and by three PepMV-genotype-specific RT-qPCR assays described by Gutierrez-Aguirre et al. [22
]: (i) Eur-rep (assay for specific detection of Eur and Peruvian genotypes; PepMV-Eur RT-qPCR); (ii) Ch2 and US2-Cp (assay for specific detection of Ch2 and US2 genotype; PepMV-Ch2 RT-qPCR); and (iii) US1-Cp (assay for specific detection of US1 genotype; PepMV-US1 RT-qPCR). All of the probes were labelled with FAM as the reporter and carboxy tetramethylrhodamine (TAMRA) as the quencher. There were exceptions for the probes used for the Naica platform, where double-quenched ZEN probes (Integrated DNA Technologies) were used. The probes were labelled as follows: (i) probe for PepMV-universal and specific probe for PepMV-Eur with FAM-ZEN; (ii) specific probe for PepMV-Ch2 with HEX-ZEN; (iii) specific probe for PepMV-US1 with Cy5-ZEN.
The RT-qPCR assays were carried out with AgPath-ID One-Step RT-qPCR mix (Ambion, USA). The final reaction volume (10 µL) included 2 µL RNA sample, 900 nM primers and 200 nM probe for the specific PepMV-Ch2, PepMV-Eur and PepMV-US1 assays, or 200 nM primers and 400 nM probe for the PepMV-universal assay. Each RT-qPCR run included a non-template control and a negative isolation control. RT-qPCR was performed with 384-well plates (Applied Biosystems, Foster City, CA, USA), with reactions run as at least two technical repeats on an ABI Prism 7900HT Fast Detection system (Applied Biosystems, Foster City, CA, USA). The condition for the cycling were 10 min at 48 °C, 10 min at 95 °C, and 45 cycles of 15 s at 95 °C, with 1 min at 60 °C. Reactions were considered positive if an exponential amplification curve was produced that could be distinguished from the negative controls; with no production of an exponential amplification curve, the reaction was considered negative. Fluorescence acquisition and of Cq determination was carried out with the SDS 2.4 software (Applied Biosystems, Foster City, CA, USA). Automatic setting of the baseline was used, with manual setting of the fluorescence threshold at 0.3. This represented a level above the baseline but low enough to remain within the exponential increase region of the amplification curve.
3.4. Optimisation of RT-dPCR: QX100/200
The QX100/200 RT-dPCR assays were carried out with One-Step RT-ddPCR Advanced kits for probes (Bio-Rad, CA, USA). The samples were each analysed as at least two technical repeats, with each run including a non-template control. The final volume of the reaction mixtures was 20 µL, which included 5 µL Supermix (Bio-Rad, CA, USA), 2 µL reverse transcriptase (final concentration, 20 U/µL; Bio-Rad, CA, USA), 1 µL 300 mM dithiothreitol (Bio-Rad, CA, USA), forward and reverse primers, and the probe, with molecular grade RNAse-free water, and finally 4 µL RNA sample. Different final concentrations of primers and probe were compared to optimise the RT-dPCR reactions: (i) the same as for the RT-qPCR assays described above; (ii) 900 nM primers (specific PepMV-Ch2, PepMV-Eur, PepMV-US1 assays)/450 nM primers (PepMV-universal assay) and 250 nM probe; (iii) 300 nM primers and 100 nM probe (PepMV-universal assay).
Preparation of the reaction mixtures was carried out in 96-well polypropylene plates (Eppendorf, Hamburg, Germany). These were then heat-sealed using pierceable foil, and the droplets were generated with an automated droplet generator (DG32; Bio-Rad, CA, USA) using automated droplet generation oil for probes (Bio-Rad, CA, USA) and a cartridge for automated generation of the droplets (DG32; Bio-Rad, CA, USA). Droplets suspensions were collected in separate 96-well polypropylene plates. After the heat sealing of these plates with droplets with pierceable foil, they were transferred to a thermal cycler (T100 or C1000 Touch; Bio-Rad, CA, USA). The conditions for the thermal cycling included 60 min reverse transcription at 50 °C and 10 min enzyme activation at 95 °C, which were followed by 40 cycles using a two-step thermal profile, of 30 s denaturation at 95 °C and 1 min annealing–elongation. Finally, this was followed by 10 min enzyme deactivation at 98 °C. For best possible separation of clusters of positive and negative droplets, annealing–elongation temperatures were tested between 55 °C and 65 °C. The ramp rate was set as 2.5 °C/s. Following the thermal cycling, plates were transferred into a droplet reader (QX100 or QX200; Bio-Rad, CA, USA). A fluorescence amplitude threshold was used to discriminate the positive droplets containing the amplification products from the negative droplets, using the QuantaSoft software, version 1.7.4 (Bio-Rad, CA, USA). The manual setting of the threshold was to just above the negative droplet cluster, as visualised using the fluorescence amplitude versus event number in the FAM channel (Figure 1
). Samples with up to three positive droplets with fluorescence intensities that were very low (i.e., <1000 above the threshold) were considered negative. Rejection from subsequent analysis of the data generated by the QuantaSoft software was based on analysis of <9000 droplets per 20 µL reaction (i.e., <7.65 µL effective reaction size), or also when there were >99.99% positive droplets (which indicated reaction saturation) (Table S1
3.5. Optimisation of RT-dPCR: Naica
The Naica RT-dPCR assays were carried out with qScript XLT One-Step RT-qPCR ToughMix (Quanta BioSciences, Beverly, MA, USA). Each sample was analysed as at least two technical repeats, with inclusion of a non-template control for each run. The final volume of the reaction mixtures was 25 µL, which included 12.5 µL ToughMix, 2.5 µL fluorescein (FITC; final concentration, 100 nM), forward and reverse primers, and the probe, with molecular grade RNAse-free water, and 4 µL RNA sample. Simplex, duplex (PepMV-universal + specific PepMV-US1, and specific PepMV-Eur + PepMV-Ch2) and triplex (PepMV-Ch2 + PepMV-Eur + PepMV-US1) reactions were tested and compared. The final concentration of each primers and each probes used for the Naica platform in simplex, duplex and triplex reactions were the same as for the RT-qPCR assays described above.
The reaction mixtures were loaded into 4-well Naica sapphire chips (Stilla Technologies, Villejuif, France), and the chips were transferred into the Naica Geode (Stilla Technologies, Villejuif, France) for droplet generation and PCR cycling. The droplets were generated at 40 °C, followed by thermal cycling of 10 min reverse transcription at 50 °C, 1 min enzyme activation at 95 °C, then 45 cycles of a two-step thermal profile with 30 s denaturation at 95 °C, and the 15 s annealing–elongation at 56 °C. Following thermal cycling, the chips were scanned using a Crystal reader (Stilla Technologies, Villejuif, France) for the following exposure times: (i) 70 ms for FAM (blue channel); (ii) 125 ms for HEX (green channel); (iii) 25 ms for Cy5 (red channel). Data analysis used the Cristal Miner software, version 1.4.3 (Stilla Technologies, Villejuif, France). After application of spillover compensation to the raw fluorescence data [35
], a threshold was applied to discriminate the positive droplets from the negative droplets, using the automated tool supplied by the software (Figure 2
). Samples with up to three positive droplets with fluorescence intensities that was very low (i.e., <1000 above the threshold) were considered negative. Rejection from subsequent analysis of the data generated by the Cristal Miner software was based on analysis of <15,000 droplets per 25 µL reaction (i.e., <6.45 µL effective reaction size), or also when there were >99.99% positive droplets (which indicated reaction saturation).
3.6. Validation of RT-dPCR Assays for Quantification of Different PepMV Genotypes
The optimised RT-dPCR assays underwent validation based on the recommendations of the European and Mediterranean Plant Protection Organisation [36
]. The RT-dPCR assay specificities were determined experimentally through two (QX100/200) or three (Naica) technical replicate measurements of RNA isolated from leaf material infected with PepMV-Ch2, PepMV-Eur and PepMV-US1. To determine the reproducibility, two dilutions of each RNA sample, one with medium and the other with low PepMV RNA concentrations were prepared and stored below −15 °C (Table 2
). Samples without PepMV RNA were used as negative controls. Three runs were carried out for each experiment (on three different days), with each run including sample measurements as two or three technical repeats. The technical repeats were carried out under repeatability conditions within each single run: the analyst, reaction mix and instruments were the same, and the chemicals and cartridges were from the same batch. At least one additional condition was changed between the QX100/200 runs, in order to contrast: two analysts, two thermal cyclers for amplification (T100, C1000 Touch; Bio-Rad, CA, USA) and two droplet readers (QX200, QX100; Bio-Rad, CA, USA).
The sensitivities and linear ranges of the PepMV-genotype-specific RT-dPCR assays were tested on a dilution series (10-fold) of PepMV-positive RNA samples (in RNAse-free water), over the RNA dilution range of 10−3 to 10−8. RNA was extracted from test plants infected with PepMV-Ch2, PepMV-Eur and PepMV-US1. The results were compared to the RT-qPCR results. For PepMV-universal assays, only the sensitivity was tested, using a dilution series (10-fold; from 10−3 to 10−8) of a mix of PepMV-Ch2, PepMV-Eur and PepMV-US1 RNA-positive samples (the mix contained the same volume of RNA of each PepMV genotype). Additionally, a dilution series (10-fold; from 10−4 to 10−9) of PepMV-Ch2 in RNA isolated from leaf material of healthy tomato plants was prepared and tested. Each dilution was analysed as three technical repeats for each assay. The same RNA dilutions were used for each assay.
3.7. Data Analysis
Relative quantification was carried out using the standard curve method [37
]. For comparisons of output data from the RT-qPCR and RT-dPCR, Cq was used to define the RNA copy numbers. For this conversion, actual copy numbers were initially determined with RT-dPCR QX100/200, as the calibrator [30
]. Cq conversion was defined by the most accurately calculated PepMV RNA concentration from RT-dPCR (i.e., which showed the lowest CV across technical repetitions). This PepMV RNA concentration was then used for the back calculation of those in the other dilutions. Using linear regression, the equation for calculation of the number of detected targets using the Cq of the samples was defined, which was based on the copy numbers determined [38
The aim of the present study was for the development of an accurate method to provide absolute quantification of different PepMV genotypes without the need for standard curves. This was achieved by the transfer of PepMV-genotype-specific RT-qPCR assays to two dPCR platforms. In comparison with RT-qPCR, the PepMV-genotype-specific RT-dPCR assays showed sensitivities that were similar or were up to 10-fold greater, with better overall repeatability, for both of the tested platforms. With the aim of designing an assay that can detect all PepMV genotypes by dPCR, the PepMV-universal RT-qPCR assay also underwent transfer to the two RT-dPCR platforms. The PepMV-universal assay showed similar performance in dPCR and qPCR; however, the validation of the PepMV-universal assay was only carried out for the purpose of PepMV detection, and not for quantification.
The high accuracy of this newly developed PepMV-genotype-specific RT-dPCR will provide for greater accuracy for PepMV RNA quantification in studies that are needed to support disease control, and for more precise quantification of PepMV in vaccines used for cross-protection. Moreover, the RT-dPCR assays can also be used for characterisation of the calibrators needed for standard curves in RT-qPCR. However, to more accurately determine the copy numbers, further studies are necessary to determine a more accurate droplet volume for the One-Step RT-ddPCR Advanced kits for probes used in combination with the QX100/200 platform, and to determine a more accurate droplet volume for the Naica system. The general unavailability of standardised reference materials represents a problem for detection and quantification of plant pathogens, and so the present study represents the first step in the development of calibrators or higher order quantitative methods for application to not only PepMV, but also to further plant viruses and further plant pathogens.