The cap‐snatching frequency of a plant bunyavirus from nonsense mRNAs is low but is increased by silencing of UPF1 or SMG7

Abstract Bunyaviruses cleave host cellular mRNAs to acquire cap structures for their own mRNAs in a process called cap‐snatching. How bunyaviruses interact with cellular mRNA surveillance pathways such as nonsense‐mediated decay (NMD) during cap‐snatching remains poorly understood, especially in plants. Rice stripe virus (RSV) is a plant bunyavirus threatening rice production in East Asia. Here, with a newly developed system allowing us to present defined mRNAs to RSV in Nicotiana benthamiana, we found that the frequency of RSV to target nonsense mRNAs (nsRNAs) during cap‐snatching was much lower than its frequency to target normal mRNAs. The frequency of RSV to target nsRNAs was increased by virus‐induced gene silencing of UPF1 or SMG7, each encoding a protein component involved in early steps of NMD (in an rdr6 RNAi background). Coincidently, RSV accumulation was increased in the UPF1‐ or SMG7‐silenced plants. These data indicated that the frequency of RSV to target nsRNAs during cap‐snatching is restricted by NMD. By restricting the frequency of RSV to target nsRNAs, NMD may impose a constraint to the overall cap‐snatching efficiency of RSV. Besides a deeper understanding for the cap‐snatching of RSV, these findings point to a novel role of NMD in plant–bunyavirus interactions.

Because cap-snatching and NMD may co-occur in PBs, nsRNAs (and other NMD-regulated transcripts) are ideal cap-snatching targets for bunyaviruses. However, the degradation of nsRNAs may be much faster than that of mRNAs transported to PBs via NMDunrelated pathways. This suggests that the chance for a bunyavirus to target nsRNAs is low. The two contradictory predictions are reconciled in a finding of Mir et al. (2008), which showed that a hantavirus seems to have a mechanism to inhibit later stages of NMD: the nucleocapsid protein (NP) of the hantavirus has a cap-binding activity. By binding to the cap structure of a nsRNA in PBs, it protects the degradation of the nsRNA from the 5′ terminus. With this mechanism, the hantavirus was shown to target a PTC-containing GFP mRNA more than twofold more frequently than a normal GFP mRNA during cap-snatching (Mir et al., 2008). Such a mechanism predicts that deficiencies in early steps of NMD, which reduce nsRNA accumulation in PBs, are detrimental to the hantavirus. This idea was not tested but was supported by a recent report that showed that Arabidopsis with a deficiency in an early step of NMD shows increased resistance to a plant bunyavirus named tomato spotted wilt virus (TSWV; Ma et al., 2019). However, whether TSWV frequently targets nsRNAs during cap-snatching remains unknown.
Rice stripe tenuivirus is a species of genus Tenuivirus in the family Phenuiviridae of the order Bunyavirales (Xu et al., 2021). By infecting plants of the Gramineae family, particularly rice, rice stripe virus (RSV) poses a serious threat to crop production in some countries of East Asia (Falk & Tsai, 1998). Although many aspects of RSV have been studied intensively in the past decade (Xu et al., 2021), our understanding of the cap-snatching of RSV remains poor (Kormelink et al., 2021;Liu et al., , 2018Yao et al., 2012). We recently showed that RSV can target mRNAs transiently expressed in Nicotiana benthamiana using agro-infiltration (Lin et al., 2020). This provided us with a system to present defined mRNAs to RSV to investigate its cap-snatching in planta. With the availability of this system and the background described above, we decided to investigate (a) whether RSV targets nsRNAs more frequently in comparison to normal mRNAs during cap-snatching; (b) how deficiencies in early steps of NMD influence the cap-snatching of RSV from nsRNAs; and (c) how deficiencies in early steps of NMD influence the infection of RSV.
A competition assay was used to investigate whether RSV targets nsRNAs more frequently in comparison to normal mRNAs.
In this assay, Agrobacterium tumefaciens cell cultures carrying the plasmid pCHF 3 -C 12 , which expresses a normal green fluorescent protein gene (GFP) mRNA (GFP-n), or pCHF 3 -C 11 -PTC, which expresses a mutant GFP mRNA with a PTC at codon 3 (GFP-m), were mixed at a ratio of 1:1 (see File S1 for experimental procedures). The mixture was infiltrated into a leaf of RSV-infected N. benthamiana ( Figure 1a). The infiltrated leaf patch was collected at 3 days postagro-infiltration (dpi) and RSV NP mRNA in the collected leaf patch was deep sequenced with a method briefly indicated in Figure 1b.
The NP mRNA sequences are highly heterogenous with respect to their CRLs because RSV targets a great diversity of cellular mRNAs during cap-snatching (Lin et al., 2017;Liu et al., 2018). For simplicity, RSV NP mRNA sequences with their CRLs acquired from GFP-m and GFP-n are called GFP-m-NP and GFP-n-NP, respectively. GFP-m-NP and GFP-n-NP can be distinguished from each other because RSV cleaves GFP-m and GFP-n at C 12 and C 11 (Figure 1a), respectively, acquiring two CRLs differing in length (Lin et al., 2020).
The assay was done in triplicate. The total number of NP mRNA sequences obtained from each replicate was 80,664, 99,962, and 79,889. The numbers of GFP-m-NP in the three replicates were 319, 148, and 194, whereas those of GFP-n-NP were 677, 466, and 662. Thus, the accumulation of GFP-m-NP relative to that of GFPn-NP (GFP-m-NP/GFP-n-NP) in each replicate deviated slightly from a mean value of 0.36:1 (Figure 1c). This indicated that GFP-m had been targeted nearly threefold less frequently than GFP-n by the cap-snatching of RSV.
As the CRL donated by GFP-n is one nucleotide longer than that donated by GFP-m, this observation can be explained by a preference of RSV for longer CRLs. A reciprocal experiment was done to rule out this possibility. In this experiment, GFP-n donated an 11-nt CRL, whereas GFP-m donated a 12-nt CRL to RSV. The same result was obtained: GFP-m was much less frequently targeted than was GFP-n (see File S2 for the raw data). To investigate how deficiencies in early steps of NMD influence the cap-snatching of RSV from nsRNAs, the competition assays described above were carried out in N. benthamiana whose UPF1 or SMG7, each encoding a protein component involved in early steps of NMD, had been silenced using virus-induced gene silencing (VIGS).
To do this, a cDNA fragment of UPF1 or SMG7 was cloned into the tobacco rattle virus (TRV)-based VIGS vector pTRV2 (Liu et al., 2002).
A. tumefaciens cell cultures containing pTRV2-UPF1, pTRV2-SMG7, or pTRV2 carrying a fragment of luciferase (pTRV2-LUC), which was used as a control, were each mixed with cultures containing pTRV1 before being infiltrated to leaves of N. benthamiana. At 10 dpi, when UPF1 and SMG7 had been silenced by about 68% and 76% (data not shown), respectively, the N. benthamiana was rub-inoculated with RSV. Twenty days after the rub-inoculation, agro-infiltration, sample collection, and deep sequencing of RSV NP mRNA were performed as done above. Because reducing UPF1 or SMG7 expression may enhance the activity of RDR6-mediated gene silencing, which may influence data interpretation, all these assays were done with an rdr6 RNAi line of N. benthamiana (Liu & Chen, 2016;Moreno et al., 2013;Qu et al., 2005).
Before deep sequencing of RSV NP mRNAs, we investigated how the relative accumulation of GFP-m or PHA-m was influenced in UPF1-or SMG7-silenced plants. The accumulation of PHA-m relative to that of PHA-n was detected with reverse transcriptionquantitative PCR (RT-qPCR). As shown in Figure 2a, the relative accumulation of PHA-m was increased 2.2-fold in UPF1-silenced plants in comparison to control plants (plants preinfected by TRV-LUC). In contrast, its relative accumulation was unchanged in SMG7silenced plants. The relative accumulation of GFP-m and GFP-n was studied with a different approach. In this approach, total RNA F I G U R E 1 The low frequency of RSV to target nonsense mRNAs during cap-snatching. (a) A diagrammatic sketch of GFP-n and GFP-m and the way by which they were expressed in RSV-infected Nicotiana benthamiana. The GFP amplified region (underlined) was used to evaluate the relative accumulation of GFP-n and GFP-m by deep sequencing. (b) Deep sequencing of RSV NP mRNA. The experiment was performed as described previously (Lin et al., 2017;Liu et al., 2018). Briefly, the NP mRNA was ligated to an adaptor after decapping (with total RNA as the starting material). The oligo-tagged NP mRNA was deep sequenced after reverse transcription-PCR and library construction. The green-coloured region of the mRNA (in black) indicates capped RNA leader. The red-coloured fragment indicates the adaptor ligated to the mRNA. The purple-coloured region indicates the adaptor ligated to the PCR amplicon during library construction. (c) The relative accumulation of GFP-m-NP to GFP-n-NP. The value of GFP-n-NP was set to be 1. (d) A diagrammatic sketch of PHA-n and PHA-m and the way by which they were expressed in RSV-infected N. benthamiana. (e) The relative accumulation of PHA-n-NP to GFP-NP and PHA-m-NP to GFP-NP. The value of GFP-NP was set to be 1. Each value is the mean (±SEM) of three biological replicates. Different letters indicate statistically significant differences as determined by one-way analysis of variance with Tukey's post hoc test (p < 0.05) The observation that silencing of UPF1 or SMG7 each influenced the relative accumulation of only one nsRNA is unexpected.
However, this is consistent with a recent report showing that UPF1 or SMG7 may each regulate an overlapping but different set of cellular transcripts (Raxwal et al., 2020). Alternatively, the residual UPF1/  In all, by using a recently established system that allows us to artificially present defined mRNAs to RSV, this study for the first time investigated the cap-snatching of a plant bunyavirus from nsRNAs.
In contrast to a previous report for a hantavirus (Mir et al., 2008), RSV targets nsRNAs much less frequently than it targets normal mRNAs. The frequency of RSV to target nsRNAs was increased in UPF1-or SMG7-silenced plants, indicating that NMD is responsible for the low frequency of RSV to target nsRNAs.
Assuming that nsRNAs are transported to PBs at later stages of NMD in plants, our findings can be interpreted in two different ways.
First, RSV performs cap-snatching mainly in the diffuse cytoplasm.
Secondly, RSV performs cap-snatching in PBs but lacks a mechanism to cope with the high rate of nsRNA degradation in PBs. Given that diverse bunyaviruses including one belonging to the same family as RSV have been suggested to use PBs as important sites for capsnatching (Hopkins et al., 2013), we hypothesize that the second explanation is likely to be correct. However, our data suggest that PBs are not the sole sites for the cap-snatching of RSV, otherwise it will be ) difficult to explain the increased frequency of RSV to target nsRNAs in UPF1-or SMG7-silenced plants. Is it possible that neither of GFP-m and PHA-m goes to PBs at later stages of NMD? We cannot rule out this possibility. If this is true, the interpretation of our data becomes a little more complex. However, our conclusion that NMD restricts the frequency of RSV to target nsRNAs seems to be inarguable.
Given the complex interactions between viruses and NMD of their host cells, the mechanisms underlying the increased accumulation of RSV in UPF1-or SMG7-silenced N. benthamiana are uncertain at present (Balistreri et al., 2017;Li & Wang, 2019). A plausible explanation, however, is that nsRNAs as well as other transcripts regulated by NMD were accumulated in these plants. This leads to a larger mRNA pool that is available for RSV to perform capsnatching. If this explanation is true, our finding points to a novel role of NMD in plant-bunyavirus interactions, that is, NMD may limit the infection of bunyaviruses by posing a constraint to their cap-snatching.

ACK N OWLED G EM ENTS
This work was supported by grants from the National Natural Science

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.