A reproducible methodology for absolute viral quantification and viability determination in mechanical inoculations of wheat streak mosaic virus

Wheat streak mosaic virus (WSMV) is a common wheat virus causing economic losses to production in the Great Plains of North America. Reproducible inoculation of WSMV by mechanical methods is essential to evaluate the resistance in breeding lines and relies on successful inoculation and infectivity of the virus particles. We used reverse transcription-quantitative PCR (RT-qPCR) for absolute quantification of viral genome copy numbers in both WSMV inoculum and in infected wheat leaves. A time-course study was designed to determine the viability of WSMV in inoculum over time as well as the copy number related to the phenotypic rating scale. In the phosphate inoculation buffer, WSMV was stable with average viral genome copy number 1.86 × 106 ± 4.85 × 105. Plants inoculated with this inoculation buffer using finger rub mechanical inoculation contained WSMV genome copy numbers in the infected leaves ranging between 2.66 × 104 and 4.69 × 106 at 21 to 28 days post-inoculation. Viral copy numbers were statistically similar between leaves inoculated immediately and those inoculated at later time points. There was a weak linear relationship between phenotypic rating score and copy number in infected leaves with the linear model explaining 40% of the variability (R2 = 0.40) indicating the difficulty in disease assessment based solely on phenotypic symptoms. This work describes an accurate methodology to quantify virus concentration in the inoculum and infected plants, as well as emphasizes the demand for absolute measurement of virus load to validate the subjective assessment for unbiased viral disease assessment.


Introduction
Wheat streak mosaic virus (Family: Potyviridae; Genus: Tritimovirus) is one of the most common cereal viruses infecting wheat (Triticum aestivum L.) in the North American Great Plains (Burrows et al. 2009). Wheat streak mosaic virus (WSMV) has also been reported across the globe, causing sporadic epidemics on wheat (Ellis et al. 2003;French and Stenger 2003;Navia et al. 2013). In the Great Plains, WSMV infections can cause wheat yield losses that range from 1 to 5% annually; however, total loss in localized fields is common (Burrows et al. 2009;Appel et al. 2015;Ranabhat, N. Kansas State University, USA, 2019 personal observation). In 2017, Kansas wheat producers lost 19.2 million bushels of wheat worth $76.8 million due to WSMV (KSwheat 2017). WSMV symptomology includes yellow streaked leaves with a mosaic pattern, stunted growth, reduced root mass, and decreased yield (Rahman et al. 1974;Price et al. 2010b).
WSMV, along with two other viruses triticum mosaic virus (Family: Potyviridae; Genus: Poacevirus) and High Plains wheat mosaic emaravirus (Family: Fimoviridae; Genus: Emaravirus) constitutes the wheat streak mosaic (WSM) complex. All three viruses are primarily transmitted by the eriophyid wheat curl mite (Aceria tosichella Kiefer, (Slykhuis 1973;Seifers et al. 1997;Tatineni et al. 2009). WSMV and Triticum mosaic virus (TriMV) can also be transmitted mechanically (Byamukama et al. 2014;Wosula et al. 2018). Commercial wheat cultivars with genetic 1 3 resistance to WSMV are available; however, resistance is limited to three major resistance genes; Wsm1, Wsm2, and Wsm3 (Graybosch et al. 2009;Friebe et al. 2011;Lu et al. 2011). Therefore, reliable identification of new germplasm with WSMV resistance and the subsequent development of new resistant cultivars through breeding programs are essential for the sustainable management of WSMV.
Plant cultivars that are screened for virus resistance are most often inoculated mechanically (Seifers et al. 2006(Seifers et al. , 2007Lu et al. 2011;Wosula et al. 2018) except those viruses which cannot be mechanically inoculated. Reproducible inoculation of the virus by mechanical methods is necessary to evaluate the level of resistance in breeding lines. Inoculation requires appropriate delivery of the virus to wheat and success largely depends on the infectivity of the virus particle. The effectiveness of the inoculum depends on concentration and its viability to systemically infect plants. Traditionally, the resistance level of cultivars inoculated with the virus is based on phenotypic symptom assessment (DeWolf et al. 2019;Johnson et al. 2019;Marburger et al. 2018;Rupp et al. 2014). There are currently no recognized standard area diagrams for the assessment of wheat streak mosaic, adding greater subjectivity phenotypic scoring. This is further complicated by the difficulty of accurately capturing symptoms to provide a reference. However, rating based only on symptom assessment during virus resistance screening cannot provide an accurate assessment as mild or no symptoms can have a relatively higher virus titer or vice versa (Ranieri et al. 1993;Lecoq et al. 2004).
Previous work has shown that WSMV has differing infection rates (Tatineni et al. 2010;Byamukama et al. 2012;Oliveira-Hofman et al. 2015). The relative level of viral genomes was calculated by comparing the threshold cycle (Ct) values of virus-infected samples with the normalized expression of Ct values of wheat 18S RNA (Tatineni et al. 2010(Tatineni et al. , 2019. Absolute viral genome copy number is more accurate quantification of viral load in both viral inoculum and infected tissue. Proper quantification of viral load provides actual host-virus interaction and helps in making an unbiased decision during the breeding cycle. The following work was designed with the overall objective of developing a reproducible method of absolute quantification of WSMV and determination of the inoculum viability and infectivity over time. We hypothesized that the initial viral genome copy number in inoculum influences the final virus titer in planta and the virus titer in inoculum is unaffected by the time measuring the titer immediately and after a few hours post inoculum preparation. In this study, we used sensitive and robust SYBR green RT-qPCR assay for absolute quantification of viral genome copy number in the inoculum and inoculated wheat, as well as its relationship to the phenotypic rating scale. Accurate quantification of virus load will provide an unbiased plant-virus interaction evaluation for those breeding lines having no or mild symptoms, but having higher virus titer as these lines contribute to virus spread in the field.
Virus inoculum was prepared by macerating 0.45 g tissue per 15 ml ice-cold 0.01 M sodium phosphate buffer, pH 7.2 (1:33.3 [wt/vol] tissue: buffer) using a chilled ceramic mortar and pestle. Inoculum viability was tested in a timecourse study where plants were inoculated at 0, 2, 6, 10, 24, 48, 72, and 96 h after inoculum preparation. The inoculum was stored at 4 °C after preparation for successive inoculation at different times. Wheat plants inoculated after 10 h were of different ages by day(s). Forty plants were arranged in a completely randomized design with four subsamples (individual plants) per treatment per biological replication. Each time point consisted of four inoculated plants and one mock-inoculated (buffer only) plant as a negative control. The entire experiment was repeated twice for a sum total of 80 plants in two independent biological replicates. In each biological replicate, the inoculum was prepared independently as described. For each time point, the second leaf of plants at the three leaf-stage was dusted lightly with carborundum (Fisher Scientific, MA, USA) and 40 µl of inoculum was placed over the carborundum. The leaf was pinched between the thumb and forefinger and the virus inoculum was pulled down the length of the leaf three to four times to increase the chances of infection (Rupp et al. 2014). For each time point, 500 µl of inoculum was flash-frozen in liquid nitrogen for reserve transcription-quantitative PCR (RT-qPCR). The youngest Feekes stage 4 leaf was sampled from each plant after 21 days of inoculation for RT-qPCR analysis and phenotypic rating. Leaf samples were flashfrozen in liquid nitrogen and stored at − 80 °C until RNA extraction. The phenotypic rating score of the sample was 1 3 done following the 1-9 Kansas State University rating scale (Online Resource 1) (Rupp 2015).

RNA extraction and reserve transcription
Leaf tissue samples from each plant (total 4 plants per biological replication) were pooled and the pooled sample was considered one sample per time-point for each biological replication. That means four plants of each treatment per replication were pooled for RT-qPCR analysis (n = 8 per treatment, two such pools for each treatment). The leaf sample was subjected to RNA extraction. For RNA extraction from inoculum, 50 µl of inoculum solution was used as one sample per time-point for each biological replication (two replications per treatment). Total RNA from each sample was extracted with the mirVana RNA extraction kit (Ambion Catalog number: AM1560, Thermo Fisher, MA, USA) according to the manufacturer's instructions for total RNA. The RNA concentration was measured spectrophotometrically by a NanoDrop spectrophotometer (NanoDrop Technologies, Rockland, DE, USA). Extracted total RNA (7 µg) was cleaned with the Turbo DNase-Free™ kit (AM 1907, Ambion®, Thermo Fisher, MA, USA) according to the manufacturer's instructions in a 50 µL reaction.
DNase treated total RNA (1 µg) was used to synthesize the first-strand cDNA in 20 µL reaction volume by using a 5 × iScript supermix (BioRad, La Jolla, CA, USA) according to the manufacturer's instructions and reaction conditions. Primers for RT-qPCR were designed from the gene sequence of WSMV KSMHK (GenBank MK318280.1) using Integrated DNA Technologies, Primer Quest (www. idtdna. com) with the parameters of Tm 58 °C, 18 to 22 nt, and 100-200 bp amplicon size. Primer efficiencies were tested before use and primers with efficiencies between 90 and 110% were used in the experiment (Table 2) on a fourpoint dilution series.

Cloning and sequencing of WSMV genome fragment
The first-strand cDNA obtained from RNA extracted from symptomatic leaf sample was used as a template for PCR. To obtain a large amplicon of the WSMV genome, WSMV-MHK3-F (forward) and WSMV-MHK1-R (reverse) primers were used to produce a 2,264 bp product covering the CI and VPg-NIa (3398 to 5644 bp) region of the WSMV genome (Table 1). PCR was performed using a 2x PCR master mix (Promega Corp., Madison, WI, USA) in a 25 µl of volume according to the manufacturer's instructions. The temperature profile for PCR included an initial denaturation at 94 °C for 3 min followed by 35 cycles of 1 min denaturation at 94 °C, 30 s annealing at 55 °C, and 2 min extension at 72 °C with final extension for 7 min at 72 °C. Amplified products were analyzed by agarose gel electrophoresis.
The PCR product was purified using a QIAquick gel extraction kit (QIAGEN, USA) according to the manufacturer's instructions. The purified amplicon was cloned into a topoisomerase-activated vector, pCR® 2.1-TOPO (Invitrogen, Life Technologies, CA, USA) according to the manufacturer's instructions. Transformation of the vector was performed using TOP10 competent cells (One Shot®, Invitrogen, CA, USA). Plasmid DNA was extracted from cultures using a plasmid purification kit (QIAGEN Spin Miniprep Kit, QIAGEN, USA) according to the company's instructions. The plasmid was sequenced in MCLAB, CA, USA, by using universal sequencing primer M13 forward (-20) and M13 reverse. Plasmid sequenced was blasted by using the NCBI BLAST tool (Altschul et al. 1990) after being obtained from MCLAB, CA.

Plasmid dilution, standard curve, and reverse transcription qPCR
A standard curve method was used to determine the absolute WSMV copy number in samples. A serial dilution of plasmid DNA carrying the previously cloned WSMV fragment was used as a template in a 5 point, tenfold dilution series. The mass of plasmid and concentration of the plasmid solution was used to calculate the desired copy number and the dilution series was calculated as described in Table 2. The plasmid solution with calculated copy number 3,000,000 was serially diluted (Table 2) and used as a template to establish a standard curve. The RT-qPCR was performed in a BioRad CFX96 Real-Time System (BioRad, La Jolla, CA, USA) with a total reaction volume of 25 µl. Each reaction included 12.5 µl  of iQ SYBR Green Supermix (BioRad, La Jolla, CA, USA), 50 ng of cDNA, 10 pmol of MHK3-F and MHK3-R primer (Table 1), and water added to a total volume of 25 µl according to Neugebauer et al. (2018). Technical replicates were prepared in triplicate. The thermocycler conditions were 95 °C for 3 min for initial denaturation, followed by 35 cycles of 10 s at 95 °C denaturation, and 30 s anneal/extend at 60 °C. The run was completed with a melt curve of 65 to 95 °C heating for 0.5 °C increments for 5 s. For each biological replication of the experiment, the plasmid dilution series reactions were loaded on the same plate to build the appropriate standard curve. Data of each reaction was analyzed by BioRad Real-Time System (CFX96, version 3.01215.0601) to determine the quantification cycle (Cq) values, PCR efficiency, and the standard curve.

Statistical analysis
The WSMV copy number in each sample was calculated by plotting the quantification cycle (Cq) values of the samples into the standard curve generated from the plasmid dilution series by BioRad CFX96 software. Plotted copy numbers of samples within the range of the dilution series standard curve descending from 3 × 10 6 to 3 × 10 2 were included for further analysis. In this experiment, data below the lower boundary of the standard curve were excluded from the analysis to avoid extrapolation. Data of mock-inoculated control Cq values that were below the lower boundary of the standard curve were excluded from the analysis. The absolute copy number of the samples from two biological replicates was analyzed using the SAS PROC mixed procedure (SAS Institute Inc., Cary, NC, USA). Treatment was used as a fixed effect and replication was used as a random effect. Tukey's multiple comparison test (Tukey 1949) was used to compare the mean difference in the WSMV copy number in inoculum among treatments. The mean comparison of WSMV copy number in infected wheat leaves was done by using Dunnett's method (Dunnett 1955) to compare the treatment means against treatment; 0 h was used as a baseline group. The average phenotypic rating scores from each treatment were regressed with the viral copy number of the infected leaf. Log transformed copy number data and square root transformed rating data were used for analysis to satisfy the constant variance and normality. An influence observation was detected from Cook's distance and removed from the analysis (Ramsey and Schafer 2012).

Plasmid and standard curve determination
Cloned plasmid fragments were verified as the sequence aligned to the genome sequence of WSMV (isolates KSMHK, Accession number MK318280.1) of the desired WSMV-CI-NIa fragment in the plasmid (3483-5573 nt in the reference sequence) (Online Resource 2). The standard curve was prepared using a tenfold serial dilution of plasmid copy number descending from 3 × 10 6 to 3 × 10 2 molecules and it showed a strong linear relationship with the value of R 2 0.999. The amplification efficiencies were ranging from 96.9 to 103.8% between replicates (Online Resource 3A, B). While the standard curve of each assay was used to calculate the copy number of each sample, the standard curve data from different assays were also plotted to calculate the overall R 2 of 0.982, indicating good reproducibility of the plasmid, primer, and the technique (Online Resource 4). A single sharp peak at the melting temperature of 80.0 °C (Online Resource 3C, D) shows the specific PCR products amplified with primer set MHK1. Similar results were observed in each replication.

Assessment of the WSMV copy number in the inoculum
The WSMV inoculum genomic RNA copy numbers in the inoculum ranged between 1.15 × 10 6 and 3.28 × 10 6 and varied with time post preparation (F = 4.36, P = 0.001). RNA copy number at 0 h was similar to numbers at 2 h, 6 h, 24 h, and 48 h (P > 0.05, Table 3) but it was higher than the RNA copy number of 10 h, 72 h, and 96 h (P < 0.05, Table 3).

Estimation of the absolute WSMV copy number in inoculated wheat leaves
The absolute genomic RNA copy numbers in the WSMV infected leaves inoculated with inoculum prepared at different times ranged between 2.66 × 10 4 and 4.69 × 10 6 . Dunnett's test showed that the WSMV RNA copy numbers in wheat leaves inoculated at different time points were statistically similar to the leaf inoculated at 0 h (P > 0.05, Table 3). As expected, Cq values of mock-inoculated control were outside of the lowest point in the dilution series; therefore, those data were not included in the table.

The relation between absolute WSMV copy number and the phenotypic rating score
In order to establish the relationship between viral copy number and phenotypic rating score, a linear regression of the average rating scores with the average viral genome copy number of infected leaves at each time point of inoculation post preparation was performed. There was a weak linear trend of phenotypic rating score and WMSV copy number (Fig. 1). The value of the regression coefficient of determination R 2 = 0.40 (Fig. 1). The typical WSMV symptoms on infected leaves varied with the time of inoculation (Fig. 2).

Discussion
The results presented in this study describe quantifiable methodologies that can be applied to plant-virus resistance evaluation studies and highlights the necessity of using quantitative methods to facilitate accurate disease assessment in plant breeding programs especially to those breeding line having no or mild symptoms with higher virus titer or vice-versa. Accurate measurement of viral load provides the quantitative host-virus interaction and helps to make the accurate decision of varietal selection. The Table 3 Absolute quantification of genomic RNA copies in wheat streak mosaic virus (WSMV) inoculum and the leaf of wheat (Tomahawk) infected with that inoculum during the different time after inoculum prepared obtained by SYBR green quantitative (RT-qPCR) using the standard curve of plasmid DNA * Time after inoculation preparation † Average copy number in inoculum from two replicates. Different letter within copy number of inoculum column indicates significantly different groups of means based on Tukey's HSD test (P < 0.05) † † Average copy number and standard error in infected wheat leaves from two biological replicates (Four plants of each treatment per biological replication were pooled for RT-Qpcr analysis, therefore a total of eight plants were used per treatment). Same letter within the column of copy number in infected leaves indicates no significant difference between 0 h to other times based on Dunnett's method (P < 0.05) ‡ Baseline control in Dunnett's method of multiple comparisons ‡ ‡ Average phenotypic score obtained by following the Kansas State University standard scoring protocol (Rupp 2015) TAIP * Copy number in inoculum † Copy number in infected leaf † † Average phenotypic score ‡ ‡ 0 h 3.28 × 10 6 ± 4.85 × 10 5 a 1.71 × 10 5 ‡ ± 1.09 × 10 5 5.25 ± 5.3 2 h 2.21 × 10 6 ± 4.85 × 10 5 ab 4.69 × 10 5 ± 1.09 × 10 5 a 7.25 ± 2.4 6 h 1.81 × 10 6 ± 4.85 × 10 5 ab 1.34 × 10 5 ± 1.09 × 10 5 a 4.62 ± 5.1 10 h 1.15 × 10 6 ± 4.85 × 10 5 b 2.66 × 10 4 ± 1.09 × 10 5 a 4.25 ± 4.5 24 h 1.88 × 10 6 ± 4.85 × 10 5 ab 3.68 × 10 4 ± 1.09 × 10 5 a 1.25 ± 0.3 48 h 1.91 × 10 6 ± 4.85 × 10 5 ab 4.46 × 10 5 ± 1.09 × 10 5 a 2.37 ± 1.9 72 h 1.34 × 10 6 ± 4.85 × 10 5 b 1.57 × 10 5 ± 1.09 × 10 5 a 2.75 ± 0.3 96 h 1.30 × 10 6 ± 4.85 × 10 5 b 2.10 × 10 5 ± 1.09 × 10 5 a 2.00 ± 1.4 Fig. 1 Regression of Wheat streak mosaic virus (WSMV) log copy number in inoculated leaves and phenotypic rating score. Blue shading around the fitted line represents 95% confidence limits. Sqrt, square root information on virus stability helps to make the decision of inoculum preparation time while screening the breeding nursery with mechanical inoculation of plant viruses. Viral genome copy numbers were determined using RT-qPCR. The RT-qPCR has previously been used for WSMV detection (Price et al. 2010a) and examined the relative quantification of WSMV compared with wheat reference genes (Tatineni et al. 2010(Tatineni et al. , 2019 but not the absolute viral copy number. Here, a quantitative method was used to calculate the absolute viral copy number using a WSMV genome fragment-containing plasmid to simulate viral RNA strands of a known copy number. The dilution series determined that RT-qPCR could detect the plasmid at dilutions of 3 × 10 2 to 3 × 10 6 particles, covering a wide dynamic range of concentrations (4 orders of magnitude). A strong linear relationship with the value of coefficients of determination, R 2 = 0.999, and the presence of a single fluorescence peak in the melting curve analysis support the primer specificity and indicating good reproducibility of the plasmid as a standard and the RT-qPCR technique to determine viral copy number in the inoculum and in planta.
In mechanical infections, viral movement and titer are affected by what cells are infected, how many cells are infected, and how many virions are active. The knowledge of the viability of the virus particles in the homogenized buffer over time will particularly assist researchers and breeders during field or other large-scale inoculations. Generally, it is assumed that inoculum has a viability limit. Our data revealed that WSMV was stable in the phosphate buffer for hours after post inoculum preparation as the inoculum contained more than a million viral genome copies per 50 µl of inoculum solution at each time point tested. Virus copy number at 10 h showed deviation from the tendency, which might need further inquiry, but data were consistent over all the subsamples within the replicate and no outlier was detected statistically. This finding indicates that inoculum can be made in the laboratory at least 6 h from preparation to inoculation with no loss of viability when considering both viral copy number and adequate ability to phenotype.
The number of virus particles in the initial inoculum influences the final virus titer in inoculated wheat leaves. Our data showed that the viral copy number in wheat leaves inoculated immediately after inoculum preparation was statistically similar to the viral copy number at other time points. This result further supports the stable viability of WSMV KSMHK in the described buffer. The variability of the end viral titer in the inoculated leaves depends on several factors such as inoculation efficiency, the number of wounds made by carborundum during inoculation, and physiological conditions of the plant during infection (Roenhorst et al. 1988). The success of mechanical inoculation is further impeded by heavily damaged cells during inoculation resulting in inducing stress response proteins and release of vacuole contents, further producing antagonists for virus infection mediated by plant hormones (Savatin et al. 2014).
Phenotyping is a common screening method used by plant breeders to develop new plant cultivars. It is timesensitive and considered a bottleneck for a breeding program (Furbank and Tester 2011). Visual rating scales are subjective and require highly skilled workers. Our data revealed that the linear model robustness calculated by the coefficient of determination was only 40%. This implies that only 40% of the sample variability of copy numbers and phenotypic rating scores was explained by the estimated regression model. This implies that the relationship between viral copy numbers and visual rating has a weak linear relationship. R 2 provides an estimate of the proportion of variability between copy number and rating score and explains the dynamics of the relationship between them.
Although analysis might warrant a larger data set to claim the non-linear relationship between copy number and rating scale, it demands the quantitative methods of viral titer measurement to enable the accurate screening of breeding material. Plants infected with pathogens can show inconsistent in QTL-type resistance phenotypes that need careful phenotypic analysis (Poland et al. 2009). Roossinck (2012) found that plants infected in a natural ecological setting often lacked symptoms and titer levels didn't correlate. Scoring based only on phenotypic symptom expression could select the lines having high virus titer and low phenotypic symptom expression. Plants with low or no symptoms but having high titer can serve as disease reservoirs and are epidemiologically significant. Therefore, it is important to validate phenotypic assessment with absolute quantification of viral load for at least selected lines during the development of virus-resistant germplasm through a breeding program.
Copy number (Table 3) showed comparatively high virus accumulation (high mean copy number) in wheat leaves inoculated after 48 h of post inoculum preparation but did not display a "severe" phenotype (Fig. 2). This inconsistency between viral load and symptom severity might be due to several molecular interactions between host and virus, and one possible reason could be that after 10 h of post inoculum preparation, the aggressiveness of the virus particles might decrease. However, in-depth molecular analysis on the interaction between viral proteins and host gene expression might add a clearer picture. The phenomenon of weak symptoms and high virus accumulation and vice-versa is common in plant viruses and viroids (Flores et al. 2016). Therefore, using only phenotypic assessment of cultivars for breeding programs might increase the likelihood of misleading results as high virus titer correlates with high yield loss and vice-versa.
This work is the first to examine the effect of the initial viability (stability of virus in buffer) and the concentration of WSMV particles used in inoculated studies. Using the highly sensitive absolute measurement of viral copy numbers to validate the visual rating score will provide an accurate disease severity assessment thus sets an improved standard in virus resistance breeding.