Seasonal fluctuation and host species affect Tobacco ringspot virus detection


 Tobacco ringspot virus (TRSV) is a pathogen with a wide host range that can cause disease in many perennial species. Yet detection of this virus can be insistent and unreliable, complicating disease management. In this study we asked whether seasonal effects changed the titer and distribution of the virus in Vitis vinifera, Malus domestica, and Prunus armeniaca. We developed an RT-qPCR assay and used it to monitor TRSV accumulation in leaves, stems, and roots over the course of a year. TRSV titer, and the tissues in which it accumulated, changed throughout the year in a host-specific manner which may be due to host plant growth patterns and temperature. These data will aid in determining optimal sampling time, and which tissues to collect to have confidence in diagnostic results.


Introduction
Management of viruses in perennials relies on sensitive, accurate detection, particularly when the viruses in question either do not produce signi cant symptoms or take many years to do so. However, the e cacy of virus detection is often impacted by factors including tissue tropism of the infected host [13,29] and the seasonal variability in viral titer [17].
One such case is Tobacco ringspot virus (TRSV). This nematode-transmitted subgroup-A nepovirus infects a wide host range of eld crops, ornamentals, weeds, and perennials [23,25,20]. TRSV causes a number of diseases of economic importance including bud blight disease of soybean [23] blueberry necrotic ringspot disease on blueberry [20] and decline on grapevines [28]. This virus has also been associated with foliar chlorotic ringspots or lesions on some stone fruit cultivars, and union necrosis in grafted apples, although infection is often latent and asymptomatic in both pome and stone fruit [32,18,Harper unpublished). Because of its wide host range and the potential for disease, TRSV is regarded as a pathogen of concern for the certi cation, distribution, and maintenance of "virus-tested" material [10].
Yet TRSV detection is complicated by the virus being unevenly distributed throughout the host plant [8], and as is common with many viruses, uctuating due to seasonal or environmental effects [20]. This in turn reduces con dence in diagnostic results, leading us to ask whether a negative plant is truly negative, or did sampling occur at the wrong time, or was the wrong tissue sampled? Therefore, in this study we aimed to answer this question for three major perennial host genera: Vitis, Malus, and Prunus, for host-speci c effects on virus titer are poorly studied within single host genera [13], let alone between plant genera or families. This particularly is true of perennial species that are not considered as the major or primary host of a given virus but are to some degree impacted by it. Here we developed and optimized RT-qPCR assay for the detection of TRSV and used it to follow TRSV titer in plants of these three genera over the course of one year. While we found notable differences in virus titer and distribution between the three genera, there was a general pattern of virus titer over time that suggested seasonal effects play a major role in how and where TRSV accumulates.

Assay development and optimization
To study the effect of seasonal uctuation of TRSV titer and distribution it was rst necessary to select an assay suitable for the project. As there was, at the time the project commenced, a paucity of real-time RT-qPCR assays available for this virus, we elected to design a new assay, as described below.
An alignment (Supplementary data 1) was made from near-complete sequences of TRSV capsid protein present in the NCBI database using ClustalW [30], and primers and Taqman probe (Table 1) were designed in silico using the Biosearch Real-Time Design online software program (LGC Biosearch Technologies, Novato, CA) with default parameters. The designed primers and probes were then analyzed for predicted melting temperature, self-dimer and heterodimer formation using the Oligo Analyzer (IDT Technologies, Coralville, IA) online software, and BLASTn searches conducted to determine that they did not match non-target viruses or potential host genomic sequences.
The assay was optimized against TRSV isolates from Malus, Vitis, and Prunus cultivars held in the Clean Plant Center Northwest's positive control collection at Washington State University, Prosser, WA. Total RNA was extracted from each source plant as per [3]. Reaction conditions for this assay were optimized, using Invitrogen SuperScript™ III Platinum™ One-Step qRT-PCR reagents (Thermo-Fisher Scienti c, Waltham, MA), across the following ranges: primer concentration (400 to 600 nM), probe concentration (100 to 300 nM), MgSO 4 concentration (3 to 6 mM), and annealing/extension temperature (57°C to 70°C) on a BioRad CFX96 real-time thermocycler (BioRad Laboratories, Hercules, CA). Conditions that gave the lowest Cq values across three technical replicates for each parameter on the test samples were selected.
Once optimized, the assay was assessed for exclusivity against: 1) potential host species using TRSV-free samples of The sensitivity of the designed assay was examined using an arti cial RNA standard of TRSV target region designed and produced as per [3]. Serial dilutions containing between 10 10 through 10 1 copies of transcript per reaction, diluted in TRSV-free total RNA extract from Vitis vinifera to account for the effect of plant inhibitors, were created, tested in triplicate, and the resulting data used to determine detection limit of the assay. The reaction e ciency of the assay was calculated from this dilution series using the CFX96 manager software (Bio-Rad).
Finally, the performance and inclusivity of the newly designed assay was compared to two RT-PCR assays, performed as described in the respective publications (15,8], against a panel of 7 Malus, 1 Cydonia (quince), 3 Vitis, 5 Prunus and 2 Vaccinium samples positive for TRSV. The ampli cation products from the two RT-PCR assays were visualized using capillary gel electrophoresis on a QIAxcel Advanced system (Qiagen, Hilden, Germany). Given the difference in numbers of samples detected by the assays, a serial 10-fold dilution series of a TRSV positive extract from Prunus armeniaca was created down to 1:10,000,000 and compared across all three assays as above.

Seasonal TRSV accumulation in perennial hosts
The RT-qPCR assay developed above was used to study the seasonal accumulation of TRSV in three different perennial hosts: Vitis vinifera, Malus domestica, and Prunus armeniaca. A total of two plants per species, naturally infected with TRSV that had been identi ed by RT-PCR using the method of [8] and con rmed by sanger sequencing (Genewiz Inc., South Plain eld, NJ) (NCBI numbers: MW574103, MW574104, BMW574105) were used for this study. These plants were maintained in 2-gallon root-training pots, in contained screenhouse conditions and subject to ambient temperature uctuations throughout the course of the observation period (August 2019 through August 2020). 400 mg samples of leaf midrib and lamina, current-year growth stem, and secondary root tissues were collected from each of the plants on a monthly basis, and total RNA extracted as above. TRSV titer present in each of the tissues was determined using the optimized TRSV-speci c RT-qPCR assay developed in this study, performed as above with 2 µl of total RNA for each sample. Normalization of samples was carried out using a combination of NAD5 [21,5], with EF1 [26] for V. vinifera, with Tubulin [34] for M. domestica, and with TEF2 [34] for P. armeniaca, ampli ed as described in the respective references. All samples were tested in technical replicates of three, and e ciency values calculated from a 10fold dilution series for each reference gene (Supplementary Figure 1).
Prior to quanti cation of the experimental samples, the standard deviation of the respective reference genes in the three tissue types was calculated. All showed values of less than 1.0 (Supplementary Table 1), indicating stability and suitability for use to normalize TRSV expression [6]. Relative quanti cation (RQ) values for TRSV titer at each timepoint and sample were calculated using the formula described by [24], then log-transformed (base 10) to normalize the nonparametric values. All statistical analysis was performed using the R project for statistical computing (R core team, 2017).

Assay development and optimization
To study the effect of seasonal uctuation of TRSV titer and distribution we rst developed and optimized a real-time RT-qPCR assay. Optimized conditions were 400 nM each of the forward and reverse primers, 100 nM of the FAM-labelled probe, 3 mM MgSO 4 , 0.4 µl Superscript III / Platinum Taq enzyme mix, and 2 µl of total RNA in a nal reaction volume of 20 µl. One-step thermocycling conditions were 50 °C for 10 minutes and 94 °C for 2 minutes, followed by 40 cycles of 94°C for 10 seconds, and 58 °C for 45 seconds.
The optimized assay was capable of detecting as low as 10 1 copies of TRSV RNA transcript target with an e ciency of 132.4% (Supplementary Figure 2) and demonstrated exclusivity against non-target host and potential related and/or coinfecting virus species ( Table 2). Comparison of the assay against two conventional RT-PCR assays [15,8] using a panel of 18 TRSV isolates from different host species and sources showed that the newly developed real-time RT-qPCR assay could detect TRSV at a lower titer than the conventional assays (Table 3). This was further con rmed by serial dilution of a high concentration sample in water, where the real-time RT-qPCR assay could amplify down to a dilution of 1:10 7 versus 1:10 4 for the two conventional assays (Supplementary Figure 3).

Seasonal TRSV accumulation in perennial hosts
Having developed a reliable RT-qPCR assay we next turned to an examination of the seasonal variation of TRSV titer in three perennial plant hosts, V. vinifera, M. domestica, P. domestica, and P.armeniaca. Leaf, stem, and secondary root tissue was collected from each of the monitored plants on a monthly basis, where available, and TRSV titer estimated by relative quanti cation against host-speci c reference genes ( Figure 1). TRSV titer was generally low for all three host species throughout the experiment, with virus titer dropping below the limit of detection at several timepoints for speci c tissues.
Plotting quanti ed TRSV titer as a timecourse from the beginning of the observation period (August 2019) to the end (August 2020) revealed considerable uctuation in virus titer in all three species over time and in different tissues ( Figure   1). Examining each in turn, we found that in grapevine, TRSV titer was low throughout much of the late summer through to winter dormancy where it became undetectable in all three tissue types, with a peak in spring that then declined by 1-2 orders of magnitude (log 10) during early summer. Titer in stem tissue showed the greatest degree of uctuation, followed by leaf (when tissue was available), and root tissues (Figure 1a). A similar pattern was observed in apple, although root tissue consistently contained higher relative titer than stem. Interestingly, titer in leaf tissue remained relatively constant throughout the observation period, while roots and stem showed uctuations of 3-4 orders of magnitude (log10) ( Figure   1b). Finally, TRSV in plum was broadly comparable to the pattern observed in apple, though both stem and root tissue were much more comparable across the seasons (Figure 1c).
Interestingly, when these data were compared to the average monthly temperatures for Prosser, WA, during the observation period, we noted that in all hosts TRSV titer was highest during late spring to early summer temperatures (~12-17 °C, Figure 1), and lower when outside of this range. This may be related to seasonal growth patterns for all three hosts, where in January 2020 mild winter conditions led to early bud break followed by an unexpected cold weather event in February which led to some dieback of emerging tissues, evidenced by the rise then decline in TRSV titer in all three species examined.

Discussion
In this study we aimed to determine whether there were seasonal uctuations in TRSV titer, driven in small part by frequent observations of individual plants testing positive one season, and negative the next (Beaver-Kanuya, unpublished). Fluctuation of virus titer over time, and in different tissues is a common phenomenon of plant viruses from many different genera, and host species [33,29,11,14]. What drives this phenomenon is speculative, with virus-host interactions, plant growth state, and temperature, being proposed as potential causes.
How, and the extent to which a virus interacts with its host, is the result of a complex interplay of host structures and defenses such as silencing on one hand, and whether a virus possess compatible gene products, how adapted they are to the given host and whether the virus can evade or suppress the host's defenses [22,4]. Nepoviruses such as TRSV are impacted by host hypersensitive or SA-mediated responses as well as RNA silencing, leading in some cases, to symptom recovery but not clearance of the virus [9]. It is unclear whether nepoviruses possess viral-derived suppressors of silencing, although the ToRSV coat protein has been demonstrated to interact with AGO1 proteins to potentially prevent translation repression [16].
Beyond overall tness in a given host, a virus may not be able to infect all tissues equally. Here we observed that while TRSV consistently accumulated to higher relative titer in grapevine stem and leaf tissues, a phenomenon reported for other nepoviruses in grapevine [27,11], it accumulated to higher relative titer in the roots of both apple and plum. Higher virus titer, or more frequent detection of viruses in roots, has been frequently reported [2,19,13], including for TRSV in blueberry [8] which may indicate that roots are more permissive to viral infection. This may be due to reduced silencing [1] or more effective suppression thereof by the virus [2,19]. Whether this is the cause for our observations in these two species remains speculative, although a root-biased tropism would be advantageous for a nematode transmitted virus [9] such as TRSV.
Fluctuations in titer may be correlated with changing patterns of plant growth, with replication and accumulation slowed during dormancy, and accelerated as the plant produces seasonal ush growth [14]. Here we observed signi cant decreases in relative TRSV titer in all three host species tested throughout autumn and winter, rising as dormancy broke in early spring then falling again as growth slowed in summer. This pattern has been similarly observed for a diverse range of viruses including ilarviruses in pome and stone fruit [31] potyviruses in garlic [7], Turnip mosaic virus in Arabidopsis [14], and Little cherry virus-2 in sweet cherry [34].
However, temperature may also play a signi cant role in seasonal uctuation of TRSV titer in the three species. Here we observed that TRSV titers were lower during months with average temperatures below 4°C and above 20°C, suggesting an optimum range for accumulation of this virus. Nepoviruses have been reported to be sensitive to higher temperatures where RNAi is more active [12], while cold temperatures can reduce virus titer signi cantly [14], perhaps due to reduced replicase activity. As roots are below-ground and hence insulated from the extremes of temperature uctuation, as well as being cooler on average than aerial parts of the plant it is not unexpected that virus titer would remain higher than in other tissues, as we observed here.
This study demonstrates the need for empirical examination of virus accumulation and tissue speci city prior to conducting diagnostic surveys, for we found that relative TRSV titer uctuated markedly through one year of observation and was undetectable in some tissues during dormancy. It also differed between host species, with the distribution in grapevine differing markedly from the pattern observed in both apple and plum. It is to be hoped that this work will increase diagnostic accuracy and con dence by identifying optimal times and tissues in these three different perennial genera.

Declarations Acknowledgements
The authors wish to acknowledge the assistance of C. Beaver on performing statistical analysis.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.