Identification of the Matrix Protein as a Conserved and Central Determinant of Superinfection Exclusion in Plant Rhabdoviruses

ABSTRACT

Superinfection exclusion (SIE) is a finely tuned virus–virus interaction mechanism closely linked to the viral infection cycle. However, the mechanistic basis of SIE remains incompletely understood in plant viruses, particularly among negative‐sense, single‐stranded RNA viruses. In this study, we first describe the development of an efficient reverse genetics system for the plant nucleorhabdovirus Physostegia chlorotic mottle virus (PhCMoV) by codon optimisation of the large polymerase coding sequence. Using fluorescently tagged variants of PhCMoV, as well as three additional closely or distantly related plant rhabdoviruses, we found that each rhabdovirus displayed homotypic SIE. Moreover, two closely related alphanucleorhabdoviruses, PhCMoV and eggplant mottled dwarf virus, also exhibited mutual exclusion. Loss‐ and gain‐of‐function reverse genetics analyses identified the rhabdovirus matrix (M) protein as the central SIE effector: M‐deficient mutant viruses lost exclusion capacity, whereas ectopically expressed heterologous M proteins conferred SIE against otherwise compatible, distantly related rhabdoviruses. Additional functional assays demonstrated that the ability of rhabdovirus M proteins to suppress cognate and noncognate viral RNA synthesis correlated with the intra‐ and interspecies SIE capacity. The widespread occurrence of SIE across distinct plant rhabdoviruses underscores its importance for understanding the viral replication cycle and highlights its practical relevance for the development of novel virus control strategies.

Superinfection exclusion occurs between variants of the same plant rhabdoviruses or between closely related viruses, which is mediated through repression of viral RNA synthesis by the matrix protein.

1. Introduction

Numerous viral species have evolved sophisticated mechanisms to inhibit secondary infections of the same host cells by closely related viruses, but not distantly related or unrelated viruses. This selective exclusion phenomenon, known as superinfection exclusion (SIE) or homologous interference, was initially described in plant viruses as cross‐protection (Folimonova et al. 2020; Salaman 1933; Zhang et al. 2018; Ziebell and Carr 2010). Similar mechanisms were also found across diverse bacteriophages, animal viruses and human viruses (Bratt and Rubin 1968; Johnston et al. 1974; Sauri and Earhart 1971). SIE represents a sophisticated viral strategy to optimise replication by reducing competition for limited host resources. From an evolutionary perspective, SIE constrains genetic diversity by preventing similar strains from co‐infection, thereby reducing opportunities for recombination or reassortment (Berngruber et al. 2010; Ganti et al. 2022; Gutiérrez et al. 2015). In mixed infections, exclusion mechanisms can influence viral population dynamics and determine which strains persist within host populations (Amato et al. 2022; Sims et al. 2023; Zhang et al. 2015). In addition to targeting co‐infecting or superinfecting viruses, SIE may also act on progeny viruses, preventing them from replicating in the same cells, thereby enforcing a self‐imposed population bottleneck (Ren et al. 2023; Zhang et al. 2018). It has been proposed that SIE functions as a filter that preserves beneficial mutations while purging deleterious ones, thus maintaining population fitness (Perdoncini Carvalho et al. 2022; Qu et al. 2020).

Studies using fluorescent protein‐tagged variants of several positive‐sense, single‐stranded (ss)RNA plant viruses have revealed spatial mutual exclusion between virus variants in co‐infected host tissues (Dietrich and Maiss 2003; González‐Jara et al. 2009; Julve et al. 2013; Miyashita and Kishino 2010; Takahashi et al. 2007; Takahashi and Yoshikawa 2008). Moreover, sequential plant inoculation—mimicking cross‐protection—completely blocked secondary infections (Bergua et al. 2014, 2016; Folimonova et al. 2010; Hall et al. 2001; Zhang et al. 2015), suggesting that SIE underlies the cross‐protection phenomenon. Further investigations have shown that SIE is a protein‐driven, active process rather than a passive consequence of resource depletion or host immunity induction (Folimonova 2012; Folimonova et al. 2014; Nunna et al. 2023; Ziebell et al. 2007; Ziebell and Carr 2009). Identified examples of viral proteins mediating SIE include the replication protein p28 of turnip crinkle virus (TCV) (Zhang et al. 2017), the pathogenicity protein p33 and leader proteases (L1‐L2) of citrus tristeza virus (CTV) (Atallah et al. 2016; Bergua et al. 2014; Folimonova 2012), the coat protein (CP) and NIa protease (NIa‐Pro) of wheat streak mosaic virus and Triticum mosaic virus (Tatineni and French 2016), and the P3 and NIa proteins of turnip mosaic virus (Nunna et al. 2023). The functional diversity of these SIE elicitors suggests multiple modes of action, although in most cases the exact mechanisms remain unclear.

Rhabdoviruses are negative‐sense ssRNA viruses with genomes ranging from 10 to 16 kb. Non‐segmented, plant‐infecting rhabdoviruses include nucleorhabdoviruses and cytorhabdoviruses, which replicate in the nucleus and cytoplasm, respectively (Jackson et al. 2005). Nucleorhabdoviruses are assigned to four genera: Alphanucleorhabdovirus, Betanucleorhabdovirus, Gammanucleorhabdovirus and Deltanucleorhabdovirus (Simmonds et al. 2024). Plant rhabdoviruses have a modular genome structure typically encoding at least six canonical genes arranged as: 5′‐N (nucleocapsid), P (phosphoprotein), P3 (movement protein), M (matrix protein), G (glycoprotein) and L (RNA‐dependent RNA polymerase)‐3′. Each gene is separated by gene junctions containing conserved sequence elements responsible for directing mRNA transcription. The N protein functions to encapsidate the viral genome, and together with the P‐L RNA polymerase complex, forms a ribonucleoprotein complex (RNP) essential for genome replication and mRNA transcription. During late stages of infection, RNPs accumulate in viroplasms and associate with M proteins to form a dense helical M‐RNP complex. Driven by the budding activity of the M protein, the M‐RNP complex buds through host membranes enriched with G protein, leading to the formation of enveloped virions (Jackson et al. 2005; Jayakar et al. 2004; Jackson and Li 2016).

Despite extensive characterisation of SIE in positive‐sense ssRNA plant viruses and its practical application in crop protection (Ziebell and Carr 2010), little is known about SIE in plant‐infecting negative‐sense ssRNA viruses. Among these viruses, SIE was first reported in Sonchus yellow net virus (SYNV), a member of the Betanucleorhabdovirus genus, in which the M protein was identified as a key SIE elicitor (Zhou et al. 2019). Additionally, two related cytorhabdoviruses, barley yellow striate mosaic virus (BYSMV) and northern cereal mosaic virus (NCMV), exhibit mutual exclusion in co‐infected barley plants, although the specific mediators remain unidentified (Fang et al. 2022).

In this study, we report the development of a reverse genetics system for Physostegia chlorotic mottle virus (PhCMoV), a member of Alphanucleorhabdovirus (Temple et al. 2022). Leveraging this system, alongside previously established infectious clones of SYNV (Wang et al. 2015), eggplant mottle dwarf virus (EMDV) (Wang et al. 2024), and the cytorhabdovirus tomato yellow mottle‐associated virus (TYMaV) (Liang et al. 2023), we demonstrate that SIE occurs between fluorescently tagged variants of the same viruses and between closely related viruses (PhCMoV and EMDV), but not between distantly related rhabdoviruses. Furthermore, gain‐ and loss‐of‐function analyses identify the M protein as a conserved SIE elicitor across these viruses. These findings advance our understanding of SIE in negative‐sense ssRNA viruses and offer a potential strategy for engineering cross‐protection in agricultural systems.

2. Results

2.1. Development of PhCMoV Minireplicon Systems

We first attempted to develop a minireplicon (MR) reverse genetics system for PhCMoV by transiently co‐expressing an antigenomic sense MR (agMR) together with the N, P and L core proteins required for the assembly of biologically active RNPs. Additionally, three viral suppressors of RNA silencing (VSRs), that is, tomato bushy stunt virus p19, tobacco etch virus Hc‐Pro and barley stripe mosaic virus γb, were co‐expressed to enhance transient expression. The MR construct contained green and red fluorescent protein (GFP and RFP) genes flanked by the viral leader and trailer sequences and interspaced by the N/X gene junction sequence (Figure 1a), which are required for viral replication and mRNA transcription (Jackson et al. 2005). Discrete fluorescent foci were observed in agroinfiltrated leaf areas of Nicotiana benthamiana plants, but the number of foci, especially for RFP, appeared to be small (Figure 1b). Nevertheless, PhCMoV agMR reporter gene expression required the co‐expressed N, P and L proteins, as well as the VSRs, indicating authentic viral replication (Figure S1). Interestingly, co‐expression of N, P and L proteins of EMDV, an alphanucleorhabdovirus closely related to PhCMoV, supported significantly higher MR activity than the cognate PhCMoV core proteins. In contrast, core proteins from SYNV (Betanucleorhabdovirus) or TYMaV (Cytorhabdovirus) failed to support PhCMoV MR reporter gene expression (Figure 1b). Reporter gene expression was also confirmed by measuring the mRNA and protein levels of GFP and RFP in the infiltrated tissues (Figure 1c,d).

FIGURE 1.

Physostegia chlorotic mottle virus (PhCMoV) minireplicon activities supported by cognate and noncognate rhabdoviral core proteins. (a) Schematic diagram of the PhCMoV antigenome and the antigenomic minireplicon (agMR) construct. 35S, cauliflower mosaic virus 35S promoter; le, leader; N/X J, N/X gene junction; NOS, nopaline synthase terminator; RZ, hepatitis delta virus ribozyme; tr, trailer. (b) GFP and RFP expressed from agMR supported by cognate and noncognate core proteins N, P and L. Nicotiana benthamiana leaf zones infiltrated with corresponding constructs, along with three viral suppressors of RNA silencing (VSRs; p19, Hc‐Pro and γb), were photographed at 8 dpi. Scale bar = 1 mm. (c) Reverse transcription‐quantitative PCR analysis of the levels of GFP and RFP mRNAs transcribed from agMR. NbEF1α mRNA serves as the internal reference. Values are means ± SD (n = 3 biological replicates), and statistical differences are determined by Student's t test. *p < 0.05; **p < 0.01. (d) Western blot analysis of GFP and RFP proteins expressed from MR shown in (b). (e) Analysis of the expression level of the FLAG‐tagged core proteins. RuBisCO large subunit (Rub L) serves as a protein loading control. Lopt, codon optimised L coding sequence. (f) agMR activities supported with various combinations of PhCMoV and eggplant mottle dwarf virus (EMDV) core proteins at 8 days post‐inoculation. Scale bar = 1 mm.

We suspected that one or more PhCMoV core proteins might not have been expressed at sufficient levels or in a functional form. To test this, we fused a FLAG tag to the carboxyl terminus of each PhCMoV and EMDV core protein and compared their expression levels and integrity following transient expression in N. benthamiana leaves. Western blot analyses revealed that the N and P proteins of both viruses accumulated to similarly high levels. While the PhCMoV L protein was barely detectable, a faint but visible protein band was detected for EMDV L expression (Figure 1e). In MR assays involving all possible combinations of N, P and L proteins from PhCMoV and EMDV, the EMDV L consistently supported overall higher reporter gene expression than PhCMoV L. However, in two combinations, that is, EMDV L + PhCMoV P + PhCMoV N and EMDV L + EMDV P + PhCMoV N, the improvements were less pronounced, likely due to compatibility issues (Figure 1f).

The large size of the L coding sequence (5841 nucleotides), exceeding that of most protein‐coding genes, may harbour cryptic splicing sites and/or transcriptional termination signals that could affect the level and integrity of the mRNA transcripts. To address this, we chemically synthesised a codon‐optimised version of the PhCMoV L coding sequence (Lopt), designed to remove potential cryptic splicing sites and polyadenylation signals. Indeed, expression of Lopt was dramatically improved (Figure 1e), as was the MR activity it supported, regardless of the origin of the paired N and P proteins (Figure 1f). Additionally, we performed a similar experiment using a genomic‐sense MR (gMR) of PhCMoV and obtained comparable results (Figure S2). These findings suggest that the PhCMoV MR system can be efficiently supported by the core proteins of the closely related EMDV, and that codon optimisation is essential for achieving optimal expression and function of the PhCMoV L protein.

2.2. Recovery of Recombinant PhCMoV From Cloned cDNAs

To generate recombinant PhCMoV, we constructed an antigenomic sense, full‐length PhCMoV cDNA clone (PhC‐RFP) containing an RFP reporter gene inserted between the N and X genes, directed by a duplicated N/X gene junction. Additionally, we substituted the L gene in PhC‐RFP with either the codon‐optimised version (Lopt) or the EMDV L gene, generating PhC‐RFP‐Lopt and PhC‐RFP‐EMD L, respectively (Figure 2a). Each of these clones was delivered into N. benthamiana leaves via agroinfiltration, together with binary vectors expressing a set of supporting core proteins—either PhCMoV N, P and Lopt or EMDV N, P and L. In leaf zones infiltrated with PhC‐RFP, only a few RFP‐positive foci were observed, regardless of which set of core proteins was co‐expressed. Under the same conditions, however, a greater number of infection foci were observed in PhC‐RFP‐EMD L‐infiltrated leaves, and the highest number of RFP foci appeared in leaves infiltrated with PhC‐RFP‐Lopt (Figure 2b).

FIGURE 2.

Recovery of recombinant Physostegia chlorotic mottle virus (PhCMoV) in Nicotiana benthamiana. (a) Schematic diagram of the full‐length antigenomic clones of PhCMoV‐RFP and derivatives. PhC, PhCMoV; EMD, eggplant mottle dwarf virus (EMDV). (b) Local infection foci in leaves infiltrated with the antigenomic clones and binary constructs expressing EMDV or PhCMoV core proteins. Images were taken at 12 days post‐inoculation. Scale bar = 1 mm. (c) Percentage of infected plants at various days post‐agroinoculation. Data represent mean ± SEM based on three independent replicates (n = 3) and 12 plants per replicate. Different letters above the bars indicate statistically significant differences determined using two‐way ANOVA followed by post hoc Tukey's test (p < 0.05). (d) Symptoms and RFP fluorescence in N. benthamiana plants systemically infected with PhCMoV‐RFP or PhCMoV‐RFP‐Lopt. Infected whole plants and systemic leaves with RFP fluorescence were photographed at 22 days after mechanical inoculation. Scale bar = 1 mm. (e) Electron micrographs showing nuclei infected with PhCMoV‐RFP and PhCMoV‐RFP‐Lopt. N, nucleoplasm; V, virus particles. Scale bar = 1 μm. (f) Percentage of infected N. benthamiana plants over time following mechanical inoculation. Data represent mean ± SEM based on three independent replicates (n = 3) and 12 plants per replicate. Statistical analysis was performed with two‐tailed Student t tests. ns, p > 0.05.

Consistent with the low numbers of foci in the leaves infiltrated with PhC‐RFP, only 10.7% of plants developed systemic infections at 27 days post‐inoculation (dpi). In contrast, up to 100% of plants infiltrated with PhC‐RFP‐Lopt developed systemic infections by 27 dpi (Figure 2c). Notably, the recovery efficiency of PhC‐RFP was even lower when using a genomic sense full‐length cDNA clone (Figure S3), which has been shown to be more effective than the antigenomic clone for SYNV recovery (Ma and Li 2020). PhC‐RFP‐EMD L was also not recovered from systemic leaves of infiltrated plants despite producing more local RFP foci than PhC‐RFP, suggesting that the chimeric virus is not viable. Neither PhC‐RFP nor PhC‐RFP‐Lopt could be recovered when binary vectors expressing PhCMoV N, P and the wild‐type L were used.

Nevertheless, recombinant PhC‐RFP and PhC‐RFP‐Lopt induced similar symptoms in systemically infected plants. Transmission electron microscopy (TEM) analysis revealed indistinguishable cytopathology and virion morphology between PhC‐RFP and PhC‐RFP‐Lopt (Figure 2d,e). Upon mechanical passage, the two viruses also exhibited similar infection dynamics (Figure 2f). The codon‐optimised Lopt sequence was stably maintained in progeny virus genomes after passages, as validated by restriction digestion and Sanger sequencing (Figure S4). In subsequent experiments, we used the codon‐optimised PhCMoV cDNA clone due to its high infectivity and, for simplicity, referred to it as PhC.

Several conclusions can be drawn from these data: (i) the EMDV core proteins were as effective in supporting PhCMoV recovery as the corresponding PhCMoV homologues; (ii) optimal PhCMoV recovery required codon optimisation of the L sequence in both the full‐length PhCMoV cDNA clone and the supporting L binary vector; (iii) although the ectopically expressed EMDV L protein could support PhCMoV replication and mRNA transcription, it failed to complement the full functions of PhCMoV L when engineered into the virus genome.

2.3. SIE Between Fluorescent Protein‐Tagged Variants of the Same Plant Rhabdoviruses

Using the efficient reverse genetics systems described above for PhCMoV and for other plant rhabdoviruses established previously (Liang et al. 2023; Wang et al. 2024), we next engineered GFP‐ and RFP‐tagged versions of PhCMoV, TYMaV and EMDV clones (abbreviated PhC, TYM and EMD, respectively) by inserting the reporter gene downstream of the N gene in each viral genome. To investigate the SIE phenomenon in these different rhabdoviruses, we co‐inoculated the GFP‐ and RFP‐tagged variants of the same virus into N. benthamiana leaves using Agrobacterium‐mediated infiltration. Fluorescence distributions were monitored in systemically infected leaves at 21 dpi. Fluorescence stereomicroscopy showed that both fluorescent variants of each virus could establish systemic co‐infection in agroinoculated N. benthamiana plants. However, for all three viruses, their GFP and RFP fluorescence exhibited spatially separated distributions (Figure 3a). Confocal microscopy further confirmed the spatial separation of the GFP‐ and RFP‐expressing cells. At the boundaries between the GFP‐ and RFP‐expressing leaf patches, only a few cells co‐expressed both fluorescent proteins, whereas most cells displayed mutual exclusion of GFP and RFP signals (Figure 3b).

FIGURE 3.

Mutual exclusion of fluorescent protein‐tagged variants of the same plant rhabdoviruses. (a) Fluorescence distribution patterns in leaf tissues of Nicotiana benthamiana plants co‐inoculated with sap mixtures of GFP‐ and RFP‐tagged variants of Physostegia chlorotic mottle virus (PhCMoV), eggplant mottle dwarf virus (EMDV) or tomato yellow mottle‐associated virus (TYMaV). Upper non‐inoculated leaves of infected plants were photographed with a fluorescence stereomicroscope at 22 days post‐inoculation. Scale bar = 1 mm. (b) Confocal micrographs showing the cellular distribution of GFP and RFP. Scale bar = 300 μm. (c) Fluorescence in leaf tissues of N. benthamiana plants sequentially inoculated with RFP‐ and GFP‐tagged variants of PhCMoV, EMDV or TYMaV. The interval between primary (1st) and secondary (2nd) inoculation was 12 days. Scale bar =1 mm.

Next, we performed challenge inoculation experiments to further validate the observed exclusion. Leaves were first mechanically inoculated with an RFP‐tagged variant, followed by challenge inoculation with a GFP‐tagged variant of the same species at 12 days post‐primary inoculation. For all three viruses, that is, PhCMoV, EMDV and TYMaV, only RFP fluorescence appeared in the upper non‐inoculated leaves, indicating that the primary RFP‐tagged viruses completely blocked subsequent infection by the GFP‐tagged variants (Figure 3c). In parallel control experiments involving mock primary inoculation, challenge inoculation with the GFP‐tagged viruses resulted in robust green fluorescence in systemic leaves (Figure 3c). These data, together with the previous findings with SYNV (Zhou et al. 2019), establish that SIE is a conserved mechanism among these plant rhabdoviruses.

2.4. SIE Between Different Plant Rhabdoviruses Tagged With Fluorescent Proteins

To investigate SIE between different plant rhabdoviruses, we generated GFP‐ and/or RFP‐tagged PhCMoV, EMDV, SYNV and TYMaV, and recovered individual recombinant viruses through agroinoculation. Leaf saps extracted from a GFP‐tagged virus and an RFP‐tagged virus were mixed in equal volumes and co‐inoculated onto N. benthamiana plants. The distribution of GFP and RFP fluorescence in systemic leaf tissues was then analysed using fluorescence microscopy (Figure 4a) and confocal microscopy (Figure 4b). The results showed that the cytorhabdovirus TYMaV could coexist with any of the three nucleorhabdoviruses, and that SYNV also did not exclude the two alphanucleorhabdoviruses, PhCMoV and EMDV. By contrast, clear mutual exclusion was observed between PhCMoV and EMDV, as evidenced by the spatial separation of GFP and RFP fluorescence (Figure 4a,b). These findings are consistent with the phylogenetic relatedness of EMDV and PhCMoV and the PhCMoV MR assays (Figure 1), underscoring that SIE occurs primarily between the same or closely related rhabdoviruses.

FIGURE 4.

Differential exclusion of different plant rhabdoviruses tagged with fluorescent proteins. PhC, Physostegia chlorotic mottle virus (PhCMoV); TYM, tomato yellow mottle‐associated virus (TYMaV); SYN, Sonchus yellow net virus (SYNV); EMD, eggplant mottle dwarf virus (EMDV). (a) Fluorescence micrographs showing the GFP and RFP distribution patterns in upper leaves of Nicotiana benthamiana plants co‐inoculated with fluorescently tagged viruses. Images were taken at 21 days post‐inoculation. Scale bar = 1 mm. (b) Confocal micrographs depicting the cellular distribution of GFP and RFP. Scale bar = 300 μm.

2.5. Compromised SIE in Rhabdovirus M Deletion Mutants

A previous study implicated SYNV M as a key SIE mediator (Zhou et al. 2019). To extend this finding, we generated M‐deletion mutants of PhCMoV and EMDV (PhC‐RFP‐ΔM and EMD‐GFP‐ΔM, respectively) by replacing their M coding regions with a fragment of the β‐glucuronidase (GUS) gene of similar size. These mutants were analysed for their ability to mediate SIE. For this purpose, N. benthamiana leaves were agroinoculated with bacterial mixtures containing either the M‐deletion mutant together with a wild‐type or M mutant clone of EMDV or PhCMoV carrying a different fluorescent marker. Additionally, the mixtures also contained the supporting plasmids for expressing EMDV core proteins (N, P and L), which efficiently supported the recovery of both viruses as described above. Because both M‐deficient viruses were unable to infect N. benthamiana systemically, SIE was assessed in the infiltrated leaves at 14 dpi, when extensive virus replication and cell‐to‐cell movement had occurred.

In control experiments involving co‐inoculation of PhC‐RFP with PhC‐GFP or EMD‐GFP, apparent mutual exclusion of the two viruses resulted in only 0.20% and 0.45% of co‐infected cells displaying yellow fluorescence (Figure 5a). By contrast, significantly higher proportions of co‐infected cells were observed when PhC‐RFP‐ΔM was co‐inoculated with PhC‐RFP (4.07%) or EMD‐GFP (5.56%), or when EMD‐GFP‐ΔM was co‐inoculated with PhC‐RFP (3.29%). Notably, co‐infiltration with both M‐deficient viruses (PhC‐RFP‐ΔM and EMD‐GFP‐ΔM) resulted in 17.17% of cells exhibiting yellow fluorescence (Figure 5a), indicating that SIE is a ‘mutual’ exclusion phenomenon.

FIGURE 5.

Analysis of the superinfection exclusion capacity of rhabdovirus M deletion mutants. (a) GFP and RFP fluorescence in Nicotiana benthamiana leaves co‐infiltrated with various combinations of wild‐type or M deletion mutants of Physostegia chlorotic mottle virus (PhCMoV; PhC) and eggplant mottle dwarf virus (EMDV; EMD) at 12 days post‐agroinfiltration. Numbers on the right indicate the percentages of co‐infected cells showing yellow fluorescence (white arrows). Data are presented as mean ± SD. Superscript letters denote statistically significant differences determined using one‐way ANOVA followed by post hoc Tukey's test. Scale bar = 500 μm. (b) Confocal micrographs showing the cellular distribution of GFP and RFP. Scale bar = 150 μm.

Confocal microscopy further confirmed these findings by revealing cellular‐level exclusion patterns, with co‐expression of both fluorescent proteins only occurring when one or both viruses lacked functional M proteins (Figure 5b). Together, these data strongly support the conclusion that the rhabdovirus M proteins are essential contributors to SIE.

2.6. Gain of SIE by Rhabdoviruses Expressing Heterologous M Proteins

Using PhCMoV and TYMaV as models, we next investigated whether reciprocal expression of heterologous M proteins is sufficient to confer SIE against an otherwise non‐excluding virus. To this end, we inserted a transcription cassette encoding FLAG‐tagged TYMaV or PhCMoV M proteins between the M/G gene in the PhC‐RFP and TYM‐RFP genomes, generating PhC‐GFP‐TYM M and TYM‐RFP‐PhC M clones, respectively (Figure 6a). These two recombinant viruses were recovered through agroinoculation, and western blot analysis confirmed successful expression of the exogenous M proteins (Figure 6b).

FIGURE 6.

Heterologous expression of the M protein is sufficient to confer superinfection exclusion against non‐excluding virus. (a) Schematic representation of Physostegia chlorotic mottle virus (PhCMoV; PhC) and tomato yellow mottle‐associated virus (TYMaV; TYM) constructs expressing a heterologous M protein. (b) Western blot analysis of exogenous M protein expression using anti‐FLAG antibody. Lane 1: PhC‐GFP‐TYM M; Lane 2: PhC‐GFP; Lane 3: TYM RFP‐PhC M; Lane 4: TYM RFP; Lane 5: Healthy control. Ponceau S staining of RuBisCO large subunit (Rub L) serves as a protein loading control. (c) Fluorescence micrographs showing the GFP and RFP fluorescence distribution in upper leaves of Nicotiana benthamiana plants co‐infected with two viruses at 21 days post‐inoculation. Scale bar = 1 mm. (d) Confocal micrographs showing the cellular distribution of GFP and RFP. Scale bar = 300 μm.

Co‐infection assays revealed that PhC‐GFP‐TYM M and TYM‐RFP could coexist in the inoculated plants but exhibited clear mutual exclusion in systemically infected leaf tissues. Similarly, TYM‐RFP‐PhC M gained the ability to exclude PhC‐RFP in upper co‐infected leaf tissues. To rule out the possibility that this gained exclusion resulted from the induction of sequence homology‐dependent RNA silencing response, we constructed a control virus, TYM‐RFP‐PhC m, in which the PhCMoV M coding sequence was rendered untranslatable by mutating the AUG start codon to a UAA stop codon. Unlike TYM‐RFP‐PhC M, the TYM‐RFP‐PhC m mutant lost the ability to exclude PhC‐RFP (Figure S5a,b), demonstrating that SIE is mediated by M at the protein level.

2.7. Correlating M‐Mediated Inhibition of Viral RNA Synthesis With SIE

Previous studies demonstrated that the ability of SYNV M protein to suppress viral replication/transcription is essential for mediating SIE (Zhou et al. 2019). We further investigated the effect of expressing different rhabdoviral M proteins on viral RNA synthesis using the PhCMoV, EMDV and TYMaV gMR assays. Encapsidated gMRs undergo direct transcription by viral RNA polymerase to synthesise the reporter mRNAs (Zhou et al. 2019). As shown in Figure 7, both PhCMoV and EMDV M proteins considerably decreased the number of GFP and RFP foci derived from PhCMoV or EMDV gMR, but not from TYMaV MR. In contrast, TYMaV M inhibited only its cognate gMR activity, whereas SYNV M exhibited either no or moderate inhibitory effect on the three heterologous gMRs. Again, using untranslatable M mutants, we confirmed that the inhibitory effects were due to the expressed M proteins rather than their RNA transcripts (Figure S5c–e). Our data therefore correlate the inter‐ and intraspecies SIE effects with the ability of M to inhibit viral RNA synthesis.

FIGURE 7.

Effect of expressing cognate and noncognate rhabdoviral M protein on genomic‐sense minireplicon (gMR) reporter gene expression. GFP and RFP fluorescence in Nicotiana benthamiana leaves co‐infiltrated with tomato yellow mottle‐associated virus (TYMaV; TYM), eggplant mottle dwarf virus (EMDV; EMD) or Physostegia chlorotic mottle virus (PhCMoV; PhC) gMR construct, along with binary plasmids expressing the M proteins or an empty vector (EV) control. Images were captured at 8 days post‐inoculation. Scale bar = 1 mm.

3. Discussion

We first established an efficient reverse genetics system for PhCMoV to facilitate SIE studies in plant rhabdoviruses. Negative‐sense ssRNA viruses, including rhabdoviruses, encode an exceptionally large RNA polymerase (L), with coding regions ranging from approximately 6–9 kb, making its expression from plasmids technically challenging. In several segmented negative‐sense ssRNA viruses, codon optimisation of the L gene has proven essential for successful recovery of recombinant viruses or MRs (Bergeron et al. 2015; Feng et al. 2020, 2021, 2023; Liu et al. 2023; Zhang et al. 2021). These cytoplasmic viruses lack a nuclear replication phase during natural infections and therefore may not be adapted to the nuclear environment encountered during plasmid‐based expression systems. In contrast, nucleorhabdoviruses, such as PhCMoV, replicate in the nucleus and are generally considered invulnerable to the nuclear pre‐mRNA splicing machinery. Surprisingly, we found that efficient expression of the PhCMoV L protein and successful rescue of recombinant virus also required codon optimisation of the L coding sequence. One possibility is that cryptic splicing sites within PhCMoV L mRNAs and antigenomic RNAs are efficiently recognised and spliced only when transcribed by host RNA polymerase II (Pol II) from plasmids, but not during viral replication and transcription mediated by the viral polymerase. Indeed, pre‐mRNA splicing is tightly coupled with Pol II transcription, and splicing is more efficient when the machinery is tethered to the transcription elongation complex (Shenasa and Bentley 2023). Alternatively, potential polyadenylation signals (AAUAAA or variants) present in the native L sequence, while ignored by viral polymerase during RNA synthesis, may cause Pol II‐mediated premature transcription termination. Nevertheless, our study provides a useful example for rescuing nucleus‐adapted negative‐sense ssRNA viruses through codon optimisation.

Using distinct plant rhabdoviruses tagged with GFP and RFP, we demonstrate that SIE is a conserved mechanism among plant rhabdoviruses. Moreover, the two closely related alphanucleorhabdoviruses, PhCMoV and EMDV, also exhibited mutual exclusion. Cross‐species SIE has previously been observed between two cytorhabdoviruses, NCMV and BYSMV, which share 62% amino acid identity in the L protein (Fang et al. 2022). This contrasts with SIE seen in several positive‐sense ssRNA plant viruses, in which cross‐protection is generally effective only between different isolates within the same strain. Historically, this strain‐specific SIE was used as a criterion to distinguish virus ‘isolates’ from ‘strains’ (Salaman 1933). For instance, secondary infections with the T36 strain of CTV are blocked by a primary infection with the same strain but not by isolates belonging to the other four strains (Folimonova et al. 2010). This strain‐specific SIE likely explains why multiple heterologous CTV strains can coexist within individual perennial citrus trees in the field (Bergua et al. 2016; Folimonova 2013; Folimonova et al. 2020). Similarly, strain‐specific cross‐protection has been reported for papaya ringspot virus (Tennant et al. 1994; Yeh and Gonsalves 1984) and pepino mosaic virus (Agüero et al. 2018; Hanssen et al. 2010).

Reverse genetics approaches also facilitate the identification of viral SIE effectors and the elucidation of underlying mechanisms. SIE effectors that are not essential for infection could be identified through analysing virus deletion mutants (Atallah et al. 2016; Bergua et al. 2014; Folimonova 2012). For those SIE‐enforcing proteins that are essential for viral replication, gain‐of‐function approaches have been employed to identify the SIE elicitors of tobacco mosaic virus (Lu et al. 1998), wheat streak mosaic virus (Tatineni and French 2016) and turnip mosaic virus (Nunna et al. 2023). Here, our loss‐of‐function analyses using EMDV and PhCMoV deletion mutants reveal that the M protein is required for effective SIE. Additionally, the data emphasise the ‘mutual’ nature of SIE, as two mutant viruses lacking the M gene could co‐infect significantly more cells compared to infections involving one wild‐type and one mutant virus. Furthermore, our gain‐of‐function approaches demonstrated that reciprocal expression of the heterologous M protein from two otherwise compatible rhabdoviruses (PhCMoV and TYMaV) was sufficient to confer mutual exclusion. These findings, together with previous studies on SYNV (Zhou et al. 2019), provide compelling evidence to support the notion that plant rhabdovirus M proteins play a conserved and central role in mediating SIE.

During rhabdovirus assembly, the M protein condenses the viral RNP (nucleocapsid) into a tightly coiled, helical M‐RNP complex, thereby suppressing the transcriptional activity of the RNP (Jayakar et al. 2004). SYNV M mutants deficient in nuclear interaction with the nucleocapsid protein exhibit reduced transcriptional suppressive activity and compromised SIE capacity. It has therefore been proposed that M proteins present during later stages of primary SYNV infection impose SIE by suppressing transcription of superinfecting viruses (Zhou et al. 2019). After analysing the SIE behaviours of distinct plant rhabdoviruses, we further correlated SIE capacity with the ability of their M proteins to suppress transcription of cognate or noncognate MRs. These findings extend the existing hypothesis and provide a mechanistic explanation for the observed cross‐species SIE.

Purified M proteins of rhabdoviruses undergo polymerisation in vitro to form globular aggregates or long fibres (Gaudin et al. 1995, 1997; Graham et al. 2008). During virus infection, however, M protein is initially present in a soluble form (McCreedy Jr. et al. 1990) and at relatively low abundance due to the polar transcription of rhabdovirus mRNAs (Whelan et al. 2004). As virus replication proceeds, large amounts of RNP and M protein accumulate and begin to assemble into budding virions. This transition of M protein molecules from the soluble form into virion assemblage likely renders infected cells non‐permissive for further replication and thus refractory to superinfection. This model is reminiscent of that proposed for TCV SIE, in which p28 undergoes a concentration‐dependent transition from a replication‐active, soluble state to a replication‐suppressive, aggregated state (Zhang et al. 2017). Given the functional similarities between the rhabdovirus M proteins and the CPs of non‐enveloped viruses, for example, oligomerisation, late accumulation, replication suppression and virion assembly, a similar mode‐of‐action may underlie the CP‐mediated SIE documented in several positive‐sense ssRNA viruses (Lu et al. 1998; Tatineni and French 2016; Valkonen et al. 2002).

The identification of M as a conserved SIE effector in plant rhabdoviruses, along with the finding that its heterologous expression confers effective SIE, opens new avenues for engineering ‘viral vaccines’ to combat viral diseases. For example, attenuated strains, or even unrelated viruses that are already approved for safe field use, could be engineered to express a potent M effector protein, thereby providing cross‐protection against rhabdovirus infections.

4. Experimental Procedures

4.1. Virus and Plant Materials

PhCMoV was originated from Leibniz Institute DSMZ (stock number: PV1182). SYNV, EMDV and TYMaV infectious clones and binary constructs encoding MRs and core proteins have been previously described (Wang et al. 2015, 2024; Liang et al. 2023). N. benthamiana plants were maintained in insect‐proof growth chambers under controlled conditions (22°C, 60% RH, 16/8 h light/dark cycle).

4.2. Reverse Transcription‐PCR, Determination of PhCMoV Sequence and Real‐Time Quantitative PCR

Total RNA was isolated from leaf tissues infected with PhCMoV using RNAiso plus reagent (TaKaRa). The PhCMoV reference genome sequence (GenBank accession number: KX636164) was retrieved from NCBI. Terminal sequences of the PhCMoV genome were verified using modified rapid amplification of cDNA ends (RACE) protocols (Wang et al. 2024). Briefly, for 5′ RACE, first‐strand cDNA was synthesised using a virus‐gene‐specific primer (PhC/5′GSP1) (Table S1). End tailing was performed with poly(dA) or poly(dC) for the plus‐strand cDNA using terminal transferase (Thermo Fisher Scientific). Nested PCR amplification was carried out using Adapter‐dT/dG primers and PhC/5′GSP2. For 3′ RACE, viral genomic RNA was tailed with poly(A) or poly(U) using poly(U) polymerase (NEB). First‐strand cDNA was synthesised by reverse transcription (RT) using the primers Adapter‐dT/dA, which served as a template in a first round of PCR using the primers Adapter + PhC/3′GSP1 and then a second round of nested PCR using the primers Adapter + PhC/3′GSP2. PCR products were ligated into the pLB blunt‐end cloning vector (Tiangen), and 15 colonies per construct were randomly selected for Sanger sequencing (Tsingke Biotechnology, China). After determining the sequences of viral ends, full‐length viral cDNA was reverse transcribed and amplified by PCR using specific primers and high‐fidelity DNA polymerase KOD One (TOYOBO, Japan).

In RT‐real‐time quantitative PCR (RT‐qPCR) assays, reverse transcription was performed using the HiScript II Reverse Transcriptase and oligo(dT) primers (Vazyme), which synthesised the first strand of cDNA from polyadenylated viral mRNAs but not from the non‐polyadenylated antigenomic RNAs. Real‐time qPCR was performed using the ChamQ SYBR Colour qPCR Master Mix (Vazyme) on a CFX real‐time machine (Roche). The expression level of NbEF1α was used as the internal reference in the RT‐qPCR assay. Primers used in this study are listed in Table S1.

4.3. Plasmid Construction

To generate binary constructs for expressing PhCMoV proteins, the coding sequences of N, P, M and L were amplified from the full‐length cDNA clone using gene‐specific primers. The amplified N, P and M fragments were individually inserted into the pGD vector via BamHI restriction digestion. The L coding sequence was inserted into the pCB301 vector using In‐Fusion HD Cloning Kit (Vazyme). The codon‐optimised L coding sequence (Table S2) was chemically synthesised by GenScript (Nanjing) and similarly inserted into pCB301.

The PhCMoV agMR construct consisted of five fragments with 15 bp overlapping ends: Le‐N, GFP, N/X gene junction, RFP and L‐Tr (designed). These fragments were amplified by PCR using primer pairs PhC/Le/F + PhC/Le/R (fragment 1), PhC/NXj/F + PhC/NXj/R (fragment 2), RFP/F + RFP/R (fragment 3), PhC/Tr/F + PhC/Tr/R (fragment 4). The amplified fragments were assembled with the linearised pCB301 backbone using In‐Fusion cloning reagents. After verifying the correct assembly by sequencing, the entire insert was amplified and cloned in the reverse orientation to generate the gMR construct.

The antigenomic PhCMoV‐RFP construct consisted of four fragments with 15 bp overlapping ends: Le‐N/X gene junction, RFP, N/X gene junction‐G, G‐Trailer (designed). These fragments were amplified by PCR using primer pairs PhC/Le/F + PhC/N‐RFP‐X/R (fragment 1), RFP/F + RFP/R (fragment 2), PhC/N‐RFP‐X/F + PhC/G‐KpnI/R (fragment 3), PhC/G‐KpnI/F + PhC/Tr/R (fragment 4). The amplified fragments were assembled with the linearised pCB301 backbone using In‐Fusion cloning reagents. After verifying the correct assembly by sequencing, the entire insert was amplified and cloned in the reverse orientation to generate the genomic PhCMoV‐RFP construct.

To generate PhCMoV‐RFP‐Lopt and PhCMoV‐RFP‐EMD L, we first constructed an intermediate vector PhCMoV‐RFP‐L:SalI, in which the native L sequence was replaced with a SalI restriction site to facilitate cloning. This was achieved by amplification of the fragment 1 from pPhCMoV‐RFP using the primers PhC/G/KpnI/F and PhC/L:SalI/R and fragment 2 using the primers PhC/L:SalI/F and pCB/SacI/R. The two fragments were assembled with the PhCMoV‐RFP construct linearized via KpnI and SacI using In‐Fusion cloning reagents. The Lopt and EMDV L sequences were amplified using the primer pairs PhC/Lopt/F + PhC/Lopt/R and PhC/EMD L/F + PhC/EMD L/R, respectively, and then assembled with the SalI‐linearised PhCMoV‐RFP‐L:SalI product.

The PhCMoV‐RFP‐ΔM construct consisted of three fragments with 15 bp overlapping ends: X‐Y/M gene junction, Gus‐FLAG, M/G gene junction‐Lopt (designed). These fragments were amplified by PCR using primer pairs PhC/ClaI/X/F + PhC/YMj/R (fragment 1), GUS/F + Flag/C/R (fragment 2), PhC/MGj/F + PhC/ClaI/Lopt/R (fragment 3). The amplified fragments were assembled with the linearised PhCMoV‐RFP‐Lopt backbone by In‐Fusion cloning reagents.

The PhCMoV‐RFP‐TYM M construct consisted of three fragments with 15 bp overlapping ends: LB‐Y/M gene junction, TYMaV M, M/G gene junction. These fragments were amplified by PCR using primer pairs NotI/F + PhC/MGj‐C/R (fragment 1), TYM M/PhC/F + TYM M/PhC/R (fragment 2), PhC/MGj‐N/F + SmaI/R (fragment 3). The amplified fragments were assembled with the linearised PhCMoV‐RFP‐Lopt backbone using In‐Fusion cloning reagents. The PhCMoV‐RFP‐TYM m construct was generated using a similar procedure.

4.4. Agrobacterium Infiltration

Recombinant binary plasmids were electroporated into Agrobacterium tumefaciens EHA105. Bacterial cultures were grown at 28°C with shaking at 220 rpm to an OD₆₀₀ of approximately 1.0, harvested by centrifugation, and resuspended in infiltration buffer (10 mM MES pH 5.6, 10 mM MgCl₂, 100 μM acetosyringone). For initiation of recombinant virus infections or MR reporter assays, bacterial resuspensions containing the full‐length cDNA clone or the MR construct were mixed with suspensions harbouring binary vectors expressing the core proteins (N, P and L) and three viral suppressors of RNA silencing (pGD‐p19, pGD‐γb and pGD‐HcPro), as described in previous studies (Ganesan et al. 2013; Wang et al. 2015). The final OD₆₀₀ values were adjusted to 0.23 for the infectious clones or MR constructs and to 0.1 for the pGD‐N/P/L/p19/γb/HcPro. For co‐inoculation of two viruses, bacterial suspensions required for rescuing each virus, as mentioned above, were mixed in equal volumes, except in the case of co‐inoculation of PhCMoV and EMDV derivatives, where only the EMDV core protein constructs were included. To assess the effects of M protein on MR activity, bacterial cultures carrying the pGD‐M constructs were added to the mixtures at a final OD₆₀₀ of 0.1. The resulting bacterial mixtures were infiltrated into the leaves of 4‐ to 5‐week‐old N. benthamiana plants.

4.5. Mechanical Inoculation

Leaf tissues (1–2 g) were harvested from systemically infected N. benthamiana plants and homogenised in precooled mortars containing 3–5 mL of Paul's formulation buffer (0.05 M potassium phosphate pH 7.0 supplemented with 5 mM diethyldithiocarbamic acid [DIECA], 1 mM EDTA, 5 mM sodium thioglycolate, 2% abrasive particles [Celite] and activated charcoal powder). The resulting sap was gently rubbed onto the abaxial surface of 4‐week‐old healthy N. benthamiana leaves. For co‐inoculation experiments, equal quantities of infected leaf material from each virus source were combined and homogenised in the same buffer, and the homogenate was used for inoculation as described above. Following inoculation, residual abrasive particles were removed by gently rinsing the leaves with water, and the plants were transferred to controlled‐environment chambers for further observation.

4.6. Immunoblotting

Cryogenically pulverised leaf tissues (0.1 g) were homogenised in 200 μL of lysis buffer (50 mM Tris–HCl pH 6.8, 4.5% wt/vol SDS, 9 M urea and 7.5% vol/vol β‐mercaptoethanol) and centrifuged to collect soluble proteins. Protein samples were separated by 12.5% SDS‐PAGE and electroblotted onto nitrocellulose membranes. The membranes were probed with primary antibodies specific for GFP (Abcam), RFP (MBL) and the FLAG epitope (Merck). Horseradish peroxidase (HRP)‐conjugated secondary antibodies (HUABIO) were used for enhanced chemiluminescence detection. RuBisCO large subunits were stained as a loading control using either Ponceau S or 0.1% wt/vol Coomassie Brilliant Blue R‐250.

4.7. Electron Microscopy

Leaf tissues (1–3 mm²) of N. benthamiana plants systemically infected with PhCMoV were excised and immediately fixed in a solution of 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.2). Following primary fixation, samples were rinsed three times with the same phosphate buffer and post‐fixed in 1% osmium tetroxide. Dehydration was performed using a graded ethanol–acetone series followed by pure acetone. Samples were embedded in Spurr's resin, polymerised at 70°C for 48 h. Ultrathin sections (∼70–90 nm) were stained with uranyl acetate and lead citrate, and examined under a transmission electron microscope (H‐7650; Hitachi).

4.8. Fluorescence and Confocal Microscopy

Leaf tissues were examined with a Zeiss SteREO Lumar V12 imaging system. GFP and RFP emissions were captured using filter sets 38 (470/40 nm excitation, 525/50 nm collection) and 31 (565/30 nm excitation, 620/60 nm collection), respectively. Image analysis was conducted using ZEN 2.3 software. Cellular localisation of fluorescence in epidermal cells was visualised with an Olympus FV3000 confocal microscope, with GFP and RFP signals acquired through sequential excitation wavelengths at 488 and 561 nm. Raw data were processed using Olympus FV31S‐SW software.

Author Contributions

Junyun Jiang: methodology, software, data curation, investigation, formal analysis, writing – original draft, writing – review and editing. Shuang Ni: conceptualization, methodology, investigation, visualization, writing – original draft. Shuo Wang: methodology, investigation. Li Xie: methodology. Zhenghe Li: conceptualization, supervision, resources, funding acquisition, project administration, writing – original draft, writing – review and editing.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Requirements of core proteins and viral suppressors of RNA silencing for PhCMoV MR activity. Nicotiana benthamiana leaves were infiltrated with agrobacterial mixtures containing binary constructs for expressing PhCMoV agMR, N, P, L and viral suppressors of RNA silencing (VSRs; p19, Hc‐Pro and γb), or mixtures lacking N, P, L or VSRs. The infiltrated leaves were photographed at 8 days post‐infiltration with a fluorescence microscope. Scale bars = 1 mm.

Figure S2: Improved PhCMoV genomic minireplicon (MR) activities with codon‐optimised L construct. (a) Schematic diagram of the genomic MR (gMR) of PhCMoV. 35S, cauliflower mosaic virus 35S promoter; tr, trailer; le, leader; N/X J, N/X gene junction; RZ, hepatitis delta virus ribozyme; NOS, nopaline synthase terminator. (b) GFP and RFP foci in Nicotiana benthamiana infiltrated with PhCMoV gMR supported by various combinations of EMDV and PhCMoV core proteins. Images were acquired at 8 days post‐infiltration with a fluorescence microscope. Scale bars = 1 mm. (c) Western blot analysis of the GFP and RFP levels in leaf samples shown in (b). Protein samples from leaves with PhCMoV Lopt or EMDV L were diluted tenfold prior to loading.

Figure S3: Comparison of antigenomic and genomic sense constructs for PhCMoV recovery. (a) Schematic diagrams of the antigenomic (ag) and genomic (g) PhCMoV cDNA clones. (b) Local infection foci of PhCMoV‐RFP in infiltrated Nicotiana benthamiana leaves. Scale bar = 1 mm. (c) Symptoms and RFP fluorescence in N. benthamiana plants systemically infected with PhCMoV‐RFP. Scale bar = 1 mm. (d) Time course of infection rates for PhC‐RFP(ag) and PhC‐RFP(g). Data represent mean ± SEM from three independent biological replicates (n = 3), with 48 plants per virus per replicate. Statistical significance between viruses at individual time points was assessed by unpaired two‐tailed Student's t test. ns, p > 0.05; ****p < 0.0001.

Figure S4: Stability of recombinant PhCMoV with codon‐optimised L sequence. (a) PhCMoV‐RFP and PhCMoV‐RFP‐Lopt were mechanically passaged in Nicotiana benthamiana plants. The L cDNA fragments were amplified from total RNA extracted from systemically infected leaf tissues and from full‐length cDNA constructs (pPhC‐RFP and pPhc‐RFP‐Lopt) as controls, and then subjected to restriction digestion with PmlI or EcoRI, which cleaved the original L sequence but not Lopt sequence. (b) Sanger sequencing validation of the Lopt sequence in passaged PhCMoV‐RFP‐Lopt. The introduced synonymous mutations in Lopt, denoted by lowercase letters, were maintained in progeny virus genomes, as shown in three representative DNA sequencing chromotograms.

Figure S5: M mediates superinfection exclusion (SIE) at the protein level. (a, b) Heterologous expression of M protein but not RNA confers SIE. Nicotiana benthamiana plants were co‐inoculated with PhC‐RFP and TYM‐GFP expressing translatable PhCMoV M or a untranslatable mutant (m). GFP and RFP fluorescence in upper non‐inoculated leaves were photographed with a fluorescence stereo microscope (a) or confocal microscope (b) at 21 days post‐inoculation (dpi). Scale bar = 1 mm (a) or 300 μm (b). (c, d) Expression of M protein but not RNA suppresses viral transcription. PhCMoV or TYMaV minireplicon (MR) assays were performed through agroinfiltration of N. benthamiana leaves. Fluorescent foci were photographed at 8 dpi, and the levels of GFP, RFP, and M proteins were analysed by western blots. Ponceau S staining of RuBisCO large subunit (Rub L) serves as a protein loading control. Scale bar = 1 mm.

Table S1: List of primers used in this study.

Table S2: Codon‐optimised PhCMoV L sequence.

Jiang, J. , Ni S., Wang S., Xie L., and Li Z.. 2025. “Identification of the Matrix Protein as a Conserved and Central Determinant of Superinfection Exclusion in Plant Rhabdoviruses.” Molecular Plant Pathology 26, no. 9: e70152. 10.1111/mpp.70152.

Data Availability Statement

Data supporting the findings of this study are available within the paper and Data S1.