Palmitoylation of γb protein directs a dynamic switch between Barley stripe mosaic virus replication and movement

Abstract

Viral replication and movement are intimately linked; however, the molecular mechanisms regulating the transition between replication and subsequent movement remain largely unknown. We previously demonstrated that the Barley stripe mosaic virus (BSMV) γb protein promotes viral replication and movement by interacting with the αa replicase and TGB1 movement proteins. Here, we found that γb is palmitoylated at Cys‐10, Cys‐19, and Cys‐60 in Nicotiana benthamiana, which supports BSMV infection. Intriguingly, non‐palmitoylated γb is anchored to chloroplast replication sites and enhances BSMV replication, whereas palmitoylated γb protein recruits TGB1 to the chloroplasts and forms viral replication‐movement intermediate complexes. At the late stages of replication, γb interacts with NbPAT15 and NbPAT21 and is palmitoylated at the chloroplast periphery, thereby shifting viral replication to intracellular and intercellular movement. We also show that palmitoylated γb promotes virus cell‐to‐cell movement by interacting with NbREM1 to inhibit callose deposition at the plasmodesmata. Altogether, our experiments reveal a model whereby palmitoylation of γb directs a dynamic switch between BSMV replication and movement events during infection.

The transition from plant hordeivirus replication at the chloroplasts to its intercellular movement is induced by palmitoylation of the multifunctional viral γb protein.

Introduction

Positive‐stranded RNA virus infections exhibit complicated but well‐organized processes. Usually, virus infections can be divided into penetration, uncoating, viral protein translation, viral genome replication, virion assembly, intracellular and intercellular movement, and long‐distance trafficking, each of which is orchestrated by versatile virus‐encoded proteins (Saxena & Lomonossoff, 2014; Wang, 2015). To establish infections, plant RNA viruses usually hijack the host endomembrane system and replicate their genomes within viral replication complexes (VRCs) (Wang, 2015; Medina‐Puche & Lozano‐Duran, 2019; Hyodo & Okuno, 2020). Subsequently, progeny viral genomes assemble into viral ribonucleoprotein complexes (vRNP) or into virions for intra‐ and intercellular movement, which are regulated by numerous viral and host proteins (Hyodo et al, 2014; Schoelz et al, 2016; Jin et al, 2018a; Levy & Tilsner, 2020). Replication and movement of plant viruses are often coupled by recruiting viral movement proteins (MPs) to the periphery of VRCs to guide genome trafficking into adjacent cells (Wu & Cheng, 2020). Although several hypotheses about MP recognition and translocation of viral RNA from VRCs to neighboring cells have been proposed (Tilsner & Oparka, 2012; Heinlein, 2015; Jiang & Laliberté, 2016; Pitzalis & Heinlein, 2017), the molecular mechanisms underlying the switches between viral replication and movement remain elusive.

Intracellular proteins often undergo various post‐translational modifications (PTM) such as phosphorylation, ubiquitination, SUMOylation, glycosylation, and persulfidation to exert multiple functions (Millar et al, 2019; Hyodo & Okuno, 2020; Liu et al, 2021a). Palmitoylation, commonly referred to as S‐acylation, is a PTM in which palmitic acid is attached to cysteine residues by a thioester linkage to provide a dynamic and reversible post‐translational process in eukaryotes (Hemsley, 2020). Palmitoylation is catalyzed by palmitoyl acyltransferase (PAT) family proteins (Li & Qi, 2017; Bürger, 2021) and functions in regulation, subcellular localization, trafficking, stress responses, and cytoskeletal organization of proteins (Hemsley, 2017, 2020; Li & Qi, 2017; Bürger, 2021).

Palmitoylation of the envelope glycoproteins of Sindbis virus and Vesicular stomatitis virus (VSV) was first reported in 1979 (Schmidt & Schlesinger, 1979; Schmidt et al, 1979). Subsequently, accumulating evidence showed that palmitoylation of viral proteins has numerous important roles in infection. For example, the palmitoylation of the influenza virus HA protein is required for fusion of the virus to endosomal membranes and the budding process is facilitated by palmitoylated M2 protein (Sobocinska et al, 2017). In addition, palmitoyl acyltransferase DHHC‐20 specifically palmitoylates the HIV‐1 Tat protein to increase Tat affinity to PI(4,5)P2, thereby allowing the viral toxin to interfere with PI(4,5)P2‐dependent membrane trafficking (Chopard et al, 2018). Moreover, palmitoylation of Hepatitis C virus (HCV) nonstructural protein 2 (NS2) facilitates HCV RNA replication by promoting NS2‐NS3 autoprocessing (Wu et al, 2019).

In plant viruses, palmitoylation modifications have only been found in geminiviruses, which are ssDNA viruses that cause plant diseases worldwide (Gong et al, 2021; Medina‐Puche et al, 2021). For example, palmitoylation of Mungbean yellow mosaic virus (MYMV) AC4 protein is required for plasma membrane (PM) localization. In this case, palmitoylated AC4 strongly suppresses systemic silencing, whereas interference with PM localization impairs viral suppressor of RNA silencing (VSR) activity upon mutation of the AC4 palmitoylated site (Carluccio et al, 2018). Palmitoylation of Beet severe curly top virus (BSCTV) C4 protein is also critical for membrane localization and regulation of shoot symptoms by facilitating interactions with an important receptor kinase CLAVATA 1 (Li et al, 2018a). Interestingly, some plant viruses have been reported to use different mechanisms to facilitate virus infection by hijacking host palmitoylated proteins to facilitate virus infection (Hemsley, 2020). Rice stripe virus (RSV) NScv4 interacts with Nicotiana benthamiana Remorin 1 (NbREM1) and interferes with palmitoylation of NbREM1, which impedes NbREM1 membrane anchorage and reduces the negative effects of callose deposition on virus movement (Fu et al, 2018). Likewise, Potato mop‐top virus (PMTV) TGB1 interacts with the N. benthamiana heavy metal‐associated isoprenylated plant protein 26 (NbHIPP26) to interfere with the palmitoylation of NbHIPP26, which facilitates translocation of NbHIPP26 from the PM to the nucleus and enhances PTMV long‐distance movement (Cowan et al, 2018). However, the effects of palmitoylation of proteins encoded by a plant RNA virus on infections remain poorly understood. Moreover, PATs responsible for the palmitoylation of plant virus proteins remain to be determined (Hemsley, 2020).

Barley stripe mosaic virus (BSMV) is a positive single‐stranded RNA virus with tripartite genomic RNAs designated α, β, and γ. The methyltransferase/helicase subunit (αa) and the polymerase subunit (γa) of the RNA‐dependent RNA polymerase (RdRp) complex are encoded by RNAα and RNAγ, respectively. Both the αa and γa proteins are essential for chloroplast‐targeted replication of BSMV (Zhang et al, 2017; Jin et al, 2018b). The first RNAβ ORF encodes the coat protein (CP) which is required for viral RNA encapsidation, and this ORF is followed by overlapping triple gene block proteins (TGB1, TGB2, and TGB3), which collectively contribute to viral cell‐to‐cell and systemic movement (Jackson et al, 2009; Li et al, 2018b). Subgenomic RNAγ encodes the multifunctional 17 kDa cysteine‐rich γb protein, which has a pivotal role in viral pathogenesis (Jackson et al, 2009; Jiang et al, 2021). An increasing body of evidence indicates that the γb protein counteracts host defense responses by interacting with distinct host factors, including the autophagy‐related gene 7 (ATG7) (Yang et al, 2018b), chloroplast NADPH‐dependent thioredoxin reductase C (NTRC) (Wang et al, 2021), thioredoxin h‐type 1 (NbTRXh1) (Jiang et al, 2022a), and peroxisomal glycolate oxidase (GOX) (Yang et al, 2018a). Our recent studies also indicate that the PKA and serine/threonine/tyrosine protein kinase STY46 function oppositely in BSMV pathogenesis by phosphorylating the γb protein at different stages of infection (Zhang et al, 2018, 2021).

Previously, we found that during the early stages of BSMV infection, γb proteins are recruited to the chloroplast outer membrane by interacting with the αa replicase subunit to serve as an enhancer of helicase activity and promote virus replication (Zhang et al, 2017). In addition, γb also functions in assembly of viral ribonucleoprotein movement complexes by interacting directly with TGB1 to enhance its ATPase activity (Jiang et al, 2020). These observations raise intriguing questions about the role of γb during the dynamic switch between BSMV replication and movement. In this study, we have extended our previous investigations by demonstrating that in N. benthamiana, the γb protein is palmitoylated by NbPAT15 and NbPAT21 at Cys‐10, Cys‐19, and Cys‐60 sites. At the late stages of virus replication, palmitoylated γb proteins recruit TGB1 to chloroplast replication sites and subsequently shift replication to a cell‐to‐cell movement phase. Moreover, palmitoylated γb also interacts with NbREM1 at the PM to inhibit callose deposition at plasmodesmata (PD) and facilitate virus cell‐to‐cell movement. These results reveal a novel role of palmitoylation in directing a dynamic switch between replication and movement processes during a plant virus infection.

Results

γb protein is palmitoylated in vivo

BSMV γb is a cysteine‐rich protein (CRP) containing 11 cysteines, and previous studies have indicated that some of these cysteines are critical for viral pathogenesis (Donald & Jackson, 1994; Bragg et al, 2004). To determine whether the cysteines in γb are palmitoylated, putative palmitoylation sites were predicted with the CSS‐Palm 4.0 online server (http://csspalm.biocuckoo.org/online.php) (Ren et al, 2008). The results indicated that cysteines at 7, 9, 10, 19, 60, and 144 are possible palmitoylation sites (Appendix Fig S1A). To verify this prediction, the immunoprecipitated γb‐3xFlag proteins from BSMV_γb‐3xFlag‐infected N. benthamiana leaves were analyzed by Q‐exactive liquid chromatography‐tandem mass spectrometry (LC‐MS/MS) (Zhang et al, 2018), which identified cysteine palmitoylation at γb positions 7, 9, 10, and 60 (Appendix Fig S1B).

To provide independent evidence to verify γb palmitoylation, Agrobacterium tumefaciens containing the γb‐3xFlag plasmid was infiltrated into N. benthamiana, and the leaves were harvested at 3 days post infiltration (dpi), followed by performing a biotin‐switch assay (Hemsley et al, 2008). As shown in Fig 1A, strong γb palmitoylation signals were detected for γb in the presence of hydroxylamine (Hyd), an agent that cleaves palmitate from cysteines, and bands corresponding to the γb palmitoylation signals were undetectable after treatment with 400 μM 2‐bromopalmitate (2‐BP), a specific inhibitor of protein palmitoylation (Hemsley et al, 2008) (Fig 1A). Furthermore, based on online predictions and LC‐MS/MS results, we individually or simultaneously substituted several cysteines within γb with serines in the γb‐3xFlag plasmid, and the palmitoylation levels of the mutants were analyzed by the biotin‐switch assay. The results showed that a triple γb mutant with cysteine substitutions at positions 10, 19, and 60 (designated γb_3CS) abolished palmitoylation of γb as evidenced by the absence of palmitoylation signals on the membrane (Fig 1B, Appendix Fig S2). Notably, mutation of these residues within the γb zinc‐finger motif could result in conformational changes that could result in protein mobility variations within the immunoblots (Bragg et al, 2004). In contrast, 2‐BP inhibitor treatment had only a marginal effect on gel mobility of the γb protein (Fig 1A). Taken together, these results indicate that γb is palmitoylated in vivo at Cys‐10, Cys‐19, and Cys‐60 sites.

Figure 1. Identification of palmitoylation sites within γb protein.

• A, B
  Detection of γb palmitoylation through the biotin‐switch assay in N. benthamiana. The γb‐3xFlag and γb_3CS‐3xFlag proteins were treated with (Hyd+) or without (Hyd‐) hydroxylamine, the thioester cleavage reagent. Lanes labeled “Palmitoylation” show γb‐3xFlag or γb_3CS‐3xFlag amounts recovered from the neutravidin beads. The loading controls indicate sample loading onto the neutravidin beads.
• C
  Systemic symptoms of N. benthamiana plants inoculated with BSMV and BSMV_3CS at 10 dpi.
• D
  Western blot detection of viral accumulation in BSMV‐ and BSMV_3CS‐infected N. benthamiana plants with anti‐TGB1 or anti‐γb antibodies. SL8, eighth systemically infected leaf; SL9, ninth systemically infected leaf. Ino, agroinfiltrated leaves. Rubisco large subunit (RbcL) was used as a loading control.
• E
  Symptoms in the first (SL1) and second (SL2) systemically infected leaves of barley inoculated with BSMV or BSMV_3CS at 12 dpi.
• F
  Western blot analysis of protein accumulation in BSMV‐ and BSMV_3CS‐ systemically infected barley leaves with anti‐TGB1 or anti‐γb antibodies. RbcL was used as a loading control.

Data information: In (A, B, D, and F), representative data are shown and three biological replicates had similar results.

Source data are available online for this figure.

To investigate the biological functions of γb protein palmitoylation during BSMV infections, we generated a virus mutant (BSMV_3CS) with serine substitutions for cysteines at sites 10, 19, and 60 within γb protein in the BSMV infectious clone (Hu et al, 2019), N. benthamiana leaves inoculated with BSMV_3CS developed mild symptoms compared with the wild type BSMV, failed to spread from the infiltrated leaves into newly emerging leaves by 10 dpi (Fig 1C), and Western blot analyses failed to reveal BSMV_3CS proteins in the emerging leaves (Fig 1C and D). In additional experiments, the two‐leaf stages of barley were inoculated with sap from the BSMV‐ and BSMV_3CS‐infiltrated N. benthamiana leaves (Yuan et al, 2011). The results showed that, compared with the wild type BSMV, BSMV_3CS elicited mild symptoms in barley along with compromised systemic virus infectivity (Fig 1E and F). These results indicate that palmitoylation of γb modulates BSMV infections in both monocot and dicot hosts.

Non‐palmitoylated γb enhances BSMV replication

Because γb affects BSMV replication (Zhang et al, 2017), movement‐deficient BSMV mutant (RNAα + RNAγ or RNAγ_3CS) plasmids were agroinfiltrated into N. benthamiana leaves to investigate whether palmitoylation of the γb protein affects virus replication. Surprisingly, Western and Northern blot analyses revealed that both the protein and viral RNAs in RNAα + RNAγ_3CS‐infiltrated leaves were significantly higher than those of the RNAα + RNAγ infiltrations (Fig 2A). We next investigated the molecular mechanisms underlying the enhancement of non‐palmitoylated γb on BSMV replication. Given that γb acts as an enhancer of αa helicase activity during BSMV replication at chloroplast sites (Zhang et al, 2017), we isolated chloroplasts from BSMV_γb‐3xFlag‐ or BSMV_3CS‐3xFlag‐infected N. benthamiana leaves. The results revealed a significantly higher proportion of the non‐palmitoylated γb_3CS mutant proteins in the chloroplast fraction than wild type γb (Fig 2B, top panel), but only a small amount of TGB1 proteins was associated with chloroplasts in the BSMV_3CS‐3xFlag infections (Fig 2B, second panel), whereas transiently expressed αa‐GFP protein was present in approximately equal amounts in the chloroplasts of both the BSMV_γb‐3xFlag and BSMV_3CS‐3xFlag infections (Fig 2B, third panel).

Figure 2. Non‐palmitoylated γb enhancement of BSMV replication by localization in close proximity to chloroplasts.

1. Effects of γb palmitoylation on BSMV replication. Movement‐deficient BSMV (RNAα + RNAγ/RNAγ_3CS) was agroinfiltrated into N. benthamiana leaves. At 3 dpi, total protein and RNA were extracted for Western and Northern blot analyses. Antibodies and probes used for molecular analyses are indicated on the right of each panel. Actin was used to monitor equal protein loading, and methylene blue‐stained rRNAs served as RNA loading controls.
2. Western blot analyses of chloroplast localization of γb and γb_3CS. Transiently expressed αa‐GFP in intact chloroplasts isolated from BSMV_γb‐3xFlag‐ or BSMV_3CS‐3xFlag‐infected N. benthamiana leaves at 4 dpi. Detection of γb‐3xFlag and γb_3CS‐3xFlag with an anti‐Flag antibody; TGB1 analysis with an anti‐TGB1 antibody; αa‐GFP detection with an anti‐GFP antibody. NTRC was a loading control for intact chloroplasts (Wang et al, 2021). Phosphoenolpyruvate carboxylase (PEPC) (Kai et al, 2003) was used to assess cytosolic contamination of isolated chloroplasts.
3. Subcellular localization of γb‐GFP and γb_3CS‐GFP in N. benthamiana epidermal leaf cells. The first and second rows show subcellular localization of γb‐GFP and γb_3CS‐GFP transient expression in leaves. The third and fourth rows show localization of γb‐GFP and γb_3CS‐GFP in BSMV_γb‐GFP‐ and BSMV_3CS‐GFP‐infected leaves at 3 dpi. Figures on the right indicate the normalized fluorescence intensities of the GFP (green) or the chloroplast autofluorescence (red) channels along the dashed white lines shown in the merged images of the left panel. Scale bars, 20 μm.
4. Subcellular localization of γb‐GFP or γb_3CS‐GFP in isolated N. benthamiana protoplasts harvested at 3 dpi. The fifth and sixth columns show γb localization in BSMV_γb‐GFP‐ (movement‐deficient) N. benthamiana protoplasts after treatment with 200 μM 2‐BP inhibitor. The released protoplasts were collected after 48 h incubation in DMSO or 2‐BP followed by confocal microscopy visualization. Figures at the bottom indicate the normalized fluorescence intensities of the GFP (green) and the chloroplast autofluorescence (red) channels along the dashed white lines in the merged images above. Scale bars, 20 μm.

Data information: In (A‐D), representative data are shown and three biological replicates had similar results.

Source data are available online for this figure.

To further determine the translocation of non‐palmitoylated γb, confocal analyses were performed to observe the subcellular localization of palmitoylated and non‐palmitoylated γb. The results revealed that transient expression of γb_3CS‐GFP alone resulted in fluorescent puncta adjacent to the chloroplasts that were distinct from the γb‐GFP puncta that were mainly present in the cytoplasm (Fig 2C), which is consistent with our previous results (Zhang et al, 2017). Furthermore, we also used confocal microscopy to investigate localization of γb_3CS‐GFP or γb‐GFP in the context of the BSMV_γb‐GFP (RNAα + RNAβ + RNAγ_γb‐GFP) and BSMV_3CS‐GFP (RNAα + RNAβ + RNAγ_3CS‐GFP) infections. Consistent with the transient expression assays, the BSMV_3CS‐GFP‐infected N. benthamiana leaves had large amounts of fluorescent puncta in close proximity to the chloroplasts and differed from those of the control BSMV_γb‐GFP‐infected leaves at 3 dpi (Fig 2C). Moreover, results similar to the leaf infections were also observed in N. benthamiana protoplasts (Fig 2D, left four panels).

To further determine whether palmitoylation changes the subcellular localization of γb, movement‐deficient BSMV mutant (RNAα + RNAγ_γb‐GFP)‐infected N. benthamiana protoplast cells were treated with 200 μM 2‐BP inhibitor for 48 h before confocal microscopy analyses. The results indicated that 2‐BP treatment significantly intensified the association of γb with the chloroplasts compared with DMSO‐treated control cells (Fig 2D). To further determine whether non‐palmitoylated γb_3CS has a strong affinity with the αa replicase, αa‐GFP was transiently expressed in BSMV_γb‐3xFlag‐ or BSMV_3CS‐3xFlag‐infected N. benthamiana leaves. Co‐IP analyses showed that more αa‐GFP protein co‐precipitated with γb_3CS‐3xFlag than with γb‐3xFlag (Appendix Fig S3B), which is consistent with the BiFC results (Appendix Fig S3A). Together, these results indicate stronger affinities of the αa replicase protein for non‐palmitoylated γb_3CS than for palmitoylated γb; elimination of γb palmitoylation during BSMV replication enhances localization of γb proteins to the chloroplasts.

Palmitoylated γb facilitates BSMV cell‐to‐cell movement

Because palmitoylated γb promotes BSMV replication, but impairs systemic infection (Fig 1C and E), and previous studies have demonstrated an important role of γb proteins in cell‐to‐cell movement (Jiang et al, 2020), we used the BSMV duplex fluorescence (dfBSMV) reporter system (Li et al, 2018b) to evaluate the non‐palmitoylated γb mutant (dfBSMV_3CS) infections and to investigate the effects of palmitoylated γb on BSMV intercellular movement. At 3 dpi, the results showed that the cell‐to‐cell movement of dfBSMV_3CS was only 25% of dfBSMV, and only slightly higher than the dfBSMV_mγb mutant with an AUG→UUG substitution of the γb initiation codon (Jiang et al, 2020) (Fig 3A and B). These results suggest that palmitoylation of γb is important for BSMV cell‐to‐cell movement.

Figure 3. Palmitoylation of γb facilitates BSMV cell‐to‐cell movement.

• A
  Effects of γb palmitoylation on virus cell‐to‐cell movement at 3 dpi with the BSMV duplex fluorescence (dfBSMV) reporter system (Li et al, 2018b). The dfBSMV_mγb mutant control contains an ATG substitution of the γb initiation codon (Jiang et al, 2020). Red fluorescence produced from the mCherry expression cassette illustrates the primary infection foci, whereas GFP fluorescence outside the red region shows invaded cells surrounding the primary infection foci. Fluorescence was evaluated by confocal microscopy. Scale bars, 100 μm.
• B
  Quantification of virus movement in Panel A. To determine the movement capacity of the viruses depicted in each panel, the mCherry and GFP fluorescent intensities were first determined by ImageJ software. Then, the intensities within the secondarily invaded cells were divided by the fluorescent mCherry plus the GFP intensities in the primary infection foci (n = 10). The Y‐axis indicates the relative sizes of the green areas in comparison to that of the red‐colored areas.
• C
  Confocal microscopy analysis of BSMV movement at different timepoints. The dfBSMV derivatives were infiltrated into N. benthamiana leaves, which were harvested at 48 hpi, 60 hpi, and 72 hpi, and analyzed by confocal microscopy. Scale bars, 100 μm.
• D
  Biotin‐switch assay to access palmitoylation of γb in BSMV‐infected N. benthamiana leaves at different timepoints. BSMV_γb‐3xFlag was treated with the hydroxylamine (Hyd⁺) thioester cleavage reagent. Equal sample amounts used for palmitoylation analyses were monitored by Western blot analysis with an anti‐Flag antibody.
• E
  Subcellular localization of BSMV_γb‐GFP and BSMV_3CS‐GFP in N. benthamiana leaf epidermal cells. The mCherry‐HDEL and mTalin‐RFP reporters were used as markers to indicate the ER and actin localization (Jiang et al, 2020). Figures at the bottom indicate the normalized fluorescence intensities of the GFP and mCherry channels along the white dashed lines in the merged images above. Scale bars, 10 μm.
• F, G
  Western blot analysis of subcellular fractions extracted from N. benthamiana leaf tissues at 3 dpi. S3 indicates the supernatant separated by centrifugation at 3,000 g; and S30 is the 30,000 g supernatant; P30 is the 30,000 g pellet. The association of γb with membranes was determined by 1% Triton X‐100 extraction.

Data information: In (A, C, D, and E), representative data are shown and three biological replicates had similar results. In (B), the box boundaries indicate the upper (25^th percentile) and lower (75^th percentile) quartiles, whiskers indicate minimum and maximum values, central band indicates the median. Letters in the chart denote statistically significant differences among groups according to the Duncan’s multiple range test (P < 0.05).

Source data are available online for this figure.

The replication and movement of plant viruses are often coupled, and since the presence or absence of palmitoylation of γb had opposite effects on BSMV movement and replication, we conducted time‐course confocal observations to investigate the dynamics of BSMV cell‐to‐cell movement. The results revealed BSMV replication by 48 hpi in N. benthamiana and suggested that cell‐to‐cell movement was initiated at approximately 60 hpi (Fig 3C, Appendix Fig S4A). Intriguingly, we also detected palmitoylation of γb at 60 hpi and found that palmitoylated γb increased significantly by 72 hpi (Fig 3D) in parallel with BSMV cell‐to‐cell movement. Furthermore, to evaluate temporal localization of the γb protein at different cellular sites, A. tumefaciens containing BSMV_γb‐GFP was agroinfiltrated into N. benthamiana leaves followed by confocal analyses at 36, 48, 60, and 72 hpi. The results showed that by 36 hpi during the early stages of infection, γb‐GFP fluorescence was readily observed at the chloroplast periphery. Subsequently, increased γb‐GFP fluorescence appeared in the cytoplasm (Appendix Fig S4B), which is consistent with our previous study (Jiang et al, 2020). These results thus imply that γb palmitoylation is central to conversion of the BSMV replication phase to cell‐to‐cell movement.

To determine the molecular mechanisms whereby palmitoylated γb enhances BSMV cell‐to‐cell movement, A. tumefaciens containing BSMV_γb‐GFP or BSMV_3CS‐GFP plasmids were infiltrated into N. benthamiana along with the mCherry‐HDEL (ER marker) (Nelson et al, 2007) or mTalin‐RFP (actin marker) (Lim et al, 2009) expression vectors. Consistent with our previous study (Jiang et al, 2020), γb‐GFP punctate foci were detected by 48 hpi within the dense ER network of some cells. By 60 hpi, the γb‐GFP foci co‐localized with thickened RFP‐labeled actin filament bundles, along with some punctate bodies. In contrast, non‐palmitoylated γb_3CS failed to localize to the ER or the actin network (Fig 3E). To further examine the membrane affinities of γb and γb_3CS, N. benthamiana leaves transiently expressing γb‐3xFlag or γb_3CS‐3xFlag were harvested at 3 dpi. Protein extracts from these leaves were separated into soluble cytoplasm and insoluble microsomal membrane fractions by centrifugation. The γb proteins in these fractionations were predominant in the membrane‐enriched fraction (P30), as only 32% of total γb proteins (S30 + P30) were released into the soluble fraction (S30) after 1% Triton X‐100 treatment. This is significantly lower than the 66% of the γb_3CS proteins that were released into the S30 fraction after Triton X‐100 treatment (Fig 3F). Similar results were also observed in BSMV_γb‐GFP‐infected N. benthamiana leaves, in which approximate 55 and 30% of γb‐GFP was released into the S30 fraction with or without 2‐BP inhibitor treatment (Fig 3G). These results indicate that palmitoylation of γb is required for targeting to the ER/actin network prior to cell‐to‐cell movement.

Palmitoylated γb recruits TGB1 to chloroplast replication sites for formation of viral replication‐movement intermediate complexes

BSMV replicates at the chloroplast outer membrane (Zhang et al, 2017) and the vRNP movement complexes initially assemble at the chloroplast periphery (Jiang et al, 2020); however, the TGB1 protein fails to target to chloroplasts when expressed alone (Li et al, 2018b). Thus, we hypothesized that palmitoylation of γb might be important for the dynamic switch between BSMV replication and movement. To test this hypothesis, A. tumefaciens containing αa‐YFPc and TGB1‐YFPn were co‐infiltrated into N. benthamiana together with γb‐3xFlag, γb_3CS‐3xFlag, or Tomato bushy stunt virus P19 (as a control). The results showed that γb, in contrast to the γb_3CS and the P19 proteins, significantly increased the reconstituted YFP signals around the chloroplasts (Fig 4A and B, Appendix Fig S5). Considering that γb can interact directly with both αa and TGB1 (Zhang et al, 2017; Jiang et al, 2020), it is reasonable to propose that γb serves as a bridge to promote αa‐TGB1 associations at the chloroplast replication sites.

Figure 4. Palmitoylated γb recruitment of TGB1 to chloroplast replication sites to form intermediate viral replication and movement complexes.

1. BiFC co‐localization analyses of αa and TGB1 in N. benthamiana leaf cells at 3 dpi. YFPn and YFPc was fused to the TGB1 and the αa proteins, respectively. Combinations of γb, γb_3CS, and the TBSV P19 control are shown above the micrographs. YFP signals were visualized by confocal microscopy at 3 dpi and depicted as a false‐green color. Scale bars, 30 μm.
2. Quantification of the numbers of YFP foci per visual field (n = 10).
3. Co‐IP assays to determine in vivo interactions between TGB1 and γb or γb_3CS. N. benthamiana leaf tissues co‐agroinfiltrated with RNAα, RNAγ_γb‐GFP, or RNAγ_3CS‐GFP, and TGB1‐3xFlag plasmids were harvested at 3 dpi. Total proteins were immunoprecipitated with anti‐Flag beads. Input and immunoprecipitated proteins (IP) were analyzed by Western blotting with anti‐GFP or anti‐Flag antibodies.
4. Confocal microscopy analyses showing RFP‐TGB1 localization at the chloroplast periphery in N. benthamiana leaves agroinfiltrated with αa‐GFP and CFP‐γb plasmids. CFP‐γb or CFP‐γb_3CS were co‐expressed with RFP‐TGB1 and αa‐GFP in N. benthamiana leaves. Chloroplast autofluorescence was displayed in false pink color. The third and fourth rows show leaves treated with 400 μM 2‐BP inhibitor or the DMSO control at 1 dpi. Two days later, confocal microscopy observations of RFP‐TGB1, αa‐GFP, and CFP‐γb were performed. Figures on the right indicate the normalized fluorescence intensity of GFP, RFP, and CFP channels along the white dashed line in the merged confocal images. Scale bars, 10 μm.
5. Statistical analyses of the GFP and RFP co‐localization in panel D. Pearson’s correlation coefficient (Pearson’s r) of GFP and RFP at the periphery of the chloroplasts was measured with the ImageJ PSC co‐localization tool (n = 8).
6. Relative percentages of RFP‐TGB1 foci that co‐localized with αa‐GFP and CFP‐γb at the periphery of the chloroplast images shown in panel D (n = 8).
7. Chloroplast extraction experiment for detection of TGB1 chloroplast localization. Plasmids harboring αa‐GFP and RFP‐TGB1 were co‐infiltrated into N. benthamiana along with CFP‐γb or CFP‐γb_3CS. Intact chloroplast preparations isolated from infiltrated N. benthamiana leaves at 4 dpi were subjected to Western blot analyses. RFP‐TGB1 was analyzed with an anti‐TGB1 antibody. The αa‐GFP, CFP‐γb, and CFP‐γb_3CS proteins were detected with an anti‐GFP antibody. NTRC was the loading control for the chloroplast preparations (Wang et al, 2021). Phosphoenolpyruvate carboxylase (PEPC) served as a control to assess cytosolic contamination of isolated chloroplasts.

Data information: In (A, D, and G), representative data are shown and three biological replicates had similar results. In (B, E, and F), the box boundaries indicate the upper (25^th percentile) and lower (75^th percentile) quartiles, whiskers indicate minimum and maximum values, central band indicates the median. Letters in the chart denote statistically significant differences among groups according to the Duncan’s multiple range test (P < 0.05).

Source data are available online for this figure.

To test the effects of palmitoylation on interactions between γb and TGB1 during BSMV replication, we performed a co‐IP assay of extracts from N. benthamiana leaves that had been infiltrated with A. tumefaciens mixtures harboring TGB1‐3xFlag, RNAα, and RNAγ_γb‐GFP or RNAγ_3CS‐GFP constructs. Total protein extracts recovered at 72 hpi were subjected to anti‐Flag affinity chromatography, followed by Western blot analyses. Co‐IP analysis showed that the TGB1‐3xFlag protein co‐precipitated with γb‐GFP but not with γb_3CS‐GFP (Fig 4C). Moreover, to further examine whether palmitoylation of γb is essential for redistribution of RFP‐TGB1 from the cytoplasm to the chloroplasts, αa‐GFP and RFP‐TGB1 were transiently co‐expressed in N. benthamiana leaves along with CFP‐γb or CFP‐γb_3CS. Confocal microscope observations showed that CFP‐γb was closely associated with αa‐GFP and RFP‐TGB1 at the chloroplasts, but in the presence of CFP‐γb_3CS, RFP‐TGB1 occurred primarily in the cytoplasm (Fig 4D, top two panels). In addition, when 400 μM 2‐BP or DMSO was infiltrated into leaves expressing αa‐GFP, CFP‐γb, and RFP‐TGB1 for 24 h, confocal microscopy revealed that TGB1 localization at the chloroplasts decreased significantly when palmitoylation of γb was inhibited by 2‐BP. In contrast, control experiments showed that most of the TGB1 co‐localized with αa‐GFP and CFP‐γb after treatment with DMSO (Fig 4D, bottom two panels). Statistical analysis of RFP‐TGB1 co‐localization with αa‐GFP and CFP‐γb near the chloroplasts also confirmed the fluorescence results (Fig 4E and F). To further verify the chloroplast localization of TGB1, A. tumefaciens containing αa‐GFP and RFP‐TGB1 constructs along with CFP‐γb or CFP‐γb_3CS were infiltrated into N. benthamiana leaves, and intact chloroplasts were centrifuged at 4 dpi. Immunoblots showed that substantial amounts of RFP‐TGB1 proteins were present in chloroplasts of CFP‐γb expressing samples compared with minor amounts in the CFP‐γb_3CS samples (Fig 4G) and, consistent with Fig 2B, a significantly higher proportion of the non‐palmitoylated γb_3CS mutant proteins in the chloroplast fraction than wide type γb. Together, these results revealed that palmitoylation of γb is required for TGB1 interactions, thereby leading to recruitment of TGB1 to chloroplast replication sites and consequent formation of viral replication‐movement intermediate complexes.

γb interacts with and is palmitoylated by NbPAT15 and NbPAT21

To investigate palmitoyl acyltransferase (PAT)‐catalyzed palmitoylation of γb, we performed RNA sequencing (RNA‐seq) analyses of the empty vector (EV)‐, BSMV‐, and BSMV_mγb (γb‐deficient mutant, AUG→UUG)‐inoculated N. benthamiana plants to identify differentially expressed genes during BSMV infection (Datasets [Link], [Link], [Link]). Several PATs varied in expression during infection (Fig 5A), and these were screened by intron‐spliced hairpin RNA‐mediated RNA interference (RNAi) approaches (Gao et al, 2022) to identify PATs that might function in γb palmitoylation. The results showed that compared with the EV control, NbPAT15‐RNAi and NbPAT21‐RNAi significantly reduced γb palmitoylation (Fig 5B). In addition, decreases in mRNA levels corresponding to the NbPAT15 and NbPAT21 reductions were confirmed by RT–qPCR (Appendix Fig S6). These results suggest that at least two N. benthamiana PATs are involved in γb palmitoylation during BSMV infection.

Figure 5. Palmitoylation of γb by NbPAT15 and NbPAT21.

1. Transcriptome analysis of NbPATs in BSMV‐ or BSMV_mγb‐infected N. benthamiana plants. The BSMV_mγb mutant has a UUG substitution for the AUG initiation codon (AUG→UUG) that prevents translation of the γb protein (Zhang et al, 2017; Jiang et al, 2020). The heat map shows the differential expression of NbPATs during BSMV or BSMV_mγb infections compared to plants infiltrated with the empty vector (EV) control. The numbers shown in the graph represent adjusted log₂ transformed transcription levels. The three rows of values represent different expression amounts of the NbPATs in two‐by‐two comparisons of the EV, BSMV, and BSMV_mγb. Minima and maxima as error bar.
2. Analysis of γb palmitoylation in NbPAT15‐ or NbPAT21‐silenced N. benthamiana plants. NbPAT15 and NbPAT21 were silenced by intron‐spliced hairpin RNA‐mediated RNAi approaches (Gao et al, 2022). The γb‐3xFlag plasmid was agroinfiltrated into the leaves, followed by biotin‐switch assays at 3 dpi.
3. Effects of γb palmitoylation by NbPAT15 or NbPAT21 on BSMV cell‐to‐cell movement. Infiltrated leaves were observed by confocal microscopy at 3 dpi. Red fluorescence produced from the mCherry expression cassette shows the primary infection foci, and GFP fluorescence outside the red regions shows virus spread into secondary infected cells. Scale bars, 100 μm.
4. Quantification of BSMV movement. The green and red fluorescent areas were measured by ImageJ software (n = 10).
5. RT–PCR analysis of BSMV RNA accumulation in NbPAT15‐ and NbPAT21‐silenced N. benthamiana leaves. EV‐treated leaves served as the control.

Data information: In (B and C), representative data are shown and three biological replicates had similar results. In (D), the box boundaries indicate the upper (25^th percentile) and lower (75^th percentile) quartiles, whiskers indicate minimum and maximum values, central band indicates the median. Letters in the chart denote statistically significant differences among groups according to the Duncan’s multiple range test (P < 0.05). In (E), error bars indicate mean ± SEM from three independent experiments, the letters indicate statistically significant differences (P < 0.05) from Student’s t‐test.

Source data are available online for this figure.

To follow up on the PAT expression experiments, NbPAT15‐RNAi, NbPAT21‐RNAi, and the EV were agroinfiltrated into N. benthamiana leaves, and after 24 h, these leaves were also agroinfiltrated with the dfBSMV reporter system (Li et al, 2018b). Confocal analyses revealed that cell‐to‐cell movement of BSMV was significantly reduced in NbPAT15‐RNAi and NbPAT21‐RNAi plants compared with EV‐infiltrated plants (Fig 5C and D). In addition, we agroinoculated N. benthamiana leaves with RNAα + RNAγ, and used the RT–qPCR assay to evaluate BSMV replication in unsilenced EV plants and the NbPAT15‐RNAi and NbPAT21‐RNAi treated plants. The results showed that NbPAT15‐RNAi and NbPAT21‐RNAi plants had higher BSMV RNA accumulation than in the EV‐agroinoculated plants (Fig 5E). Together, these results indicate that the NbPAT15 and NbPAT21 host proteins are responsible for γb palmitoylation and hence participate in BSMV replication and movement.

NbPAT15‐ and NbPAT21‐catalyzed palmitoylation of γb shifts BSMV replication to movement

Although previous studies showed that the Arabidopsis thaliana AtPAT15 is mainly ER targeted and that AtPAT21 localizes at the PM (Appendix Figs S7 and S8) (Batistic, 2012), the subcellular distributions of NbPAT15 and NbPAT21 have not been determined. However, online predictions by the BUSCA server (http://busca.biocomp.unibo.it/) suggest that NbPAT15 and NbPAT21 and their Arabidopsis orthologs have similar localization patterns (Appendix Fig S9A). To verify this prediction, we conducted a confocal analysis which confirmed that NbPAT15‐GFP localizes at the ER and that NbPAT21‐GFP targets the PM (Appendix Fig S9B and C). Nevertheless, when mixtures of A. tumefaciens containing BSMV_γb‐GFP and NbPAT15‐RFP or NbPAT21‐RFP were infiltrated into N. benthamiana leaves, confocal microscopy observations revealed that within 3 days after BSMV infection, both NbPAT15‐RFP and NbPAT21‐RFP formed fluorescent puncta in close proximity to the chloroplasts (Appendix Fig S9D). Therefore, we used the BSMV‐based BiFC assay (Jiang et al, 2020) to observe the association of NbPAT15 and NbPAT21 with γb during BSMV infection. The results showed that co‐expression of NbPAT15‐YFPn or NbPAT21‐YFPn with BSMV_γb‐YFPc resulted in reconstitution of YFP signals at the periphery of the chloroplasts and at the ER (Fig 6A and B, Appendix Fig S10). These results suggest that NbPAT15 and NbPAT21 are able to catalyze γb palmitoylation during switches between viral replication and movement as evidenced by NbPAT15 and NbPAT21 localization at the chloroplast replication sites and the ER.

Figure 6. NbPAT15‐ and NbPAT21‐mediated palmitoylation of γb to convert BSMV replication to cell‐to‐cell movement.

• A, B
  BiFC assays to identify interactions of γb with NbPAT15 and NbPAT21 during BSMV infection in N. benthamiana at 3 dpi. PAT15‐YFPn and PAT21‐YFPn were co‐expressed with BSMV_γb‐YFPc. Chloroplast autofluorescence is depicted as a false red color and mCherry‐HDEL provides the ER marker. Figures on the right show the normalized fluorescence intensities of GFP, mCherry, and the chloroplast autofluorescence channels along the white dashed lines in the merged images. Scale bars, 20 μm.
• C
  Co‐localization analyses of αa‐GFP, RFP‐TGB1, and CFP‐γb in N. benthamiana leaves at 3 dpi. Mixtures of A. tumefaciens containing αa‐GFP, RFP‐TGB1, and CFP‐γb were infiltrated into NbPATs‐silenced N. benthamiana leaves or the non‐silenced TRV:GUS control. Chloroplast autofluorescence is displayed as a false pink color. Figures on the right show the normalized fluorescence intensities of GFP, RFP, CFP, and the chloroplast autofluorescence channels along the white dashed lines in the merged images. Scale bars, 10 μm.
• D
  Co‐IP assays of interactions between γb and TGB1 in NbPAT15‐ and NbPAT21‐silenced N. benthamiana leaves; TRV:GUS and TRV:GUS combined with the 2‐BP inhibitor provided positive and negative controls, respectively. The silenced leaves were inoculated with BSMV_{3xFlag‐TGB1/γb‐GFP} mixtures (RNAα + RNAβ_{3xFlag‐TGB1} + RNAγ_γb‐GFP) and harvested at 3 dpi. Total proteins were immunoprecipitated with anti‐Flag beads. Input and IP proteins were analyzed by Western blotting with anti‐GFP or anti‐Flag antibodies.
• E
  Average amounts of RFP‐TGB1 foci co‐localizing with αa‐GFP and CFP‐γb at the chloroplast peripheries shown in panel C (n = 12).

Data information: In (A, B, C, and E), representative data are shown and three biological replicates had similar results. In (E), the box boundaries indicate the upper (25^th percentile) and lower (75^th percentile) quartiles, whiskers indicate minimum and maximum values, central band indicates the median. Letters in the chart denote statistically significant differences among groups according to the Duncan’s multiple range test (P < 0.05).

Source data are available online for this figure.

To further define the PAT interactions, we next down‐regulated NbPAT15 and NbPAT21 mRNA levels in N. benthamiana by using TRV‐based VIGS (Liu et al, 2002). In these experiments, αa‐GFP, CFP‐γb, and RFP‐TGB1 were co‐expressed in PAT‐silenced plants followed by confocal assays at 3 dpi. The results showed that αa‐γb‐TGB1 intermediate complexes were significantly lower in NbPAT15‐ or NbPAT21‐silenced N. benthamiana plants compared with the TRV:GUS negative control (Fig 6C and E). Decreases in NbPAT15 and NbPAT21 mRNA levels were confirmed by RT–qPCR (Appendix Fig S6), and co‐IP analyses consistently revealed that the γb‐TGB1 interaction was impaired to varying degrees in NbPAT15‐ and NbPAT21‐silenced plants (Fig 6D). We also used the 2‐BP inhibitor to block PAT activities in N. benthamiana leaves, and test the γb‐TGB1 interaction during virus infection. The results showed that the binding of γb‐GFP to 3xFlag‐TGB1 was reduced enormously when the palmitoylation was inhibited by the 2‐BP inhibitor (Fig 6D). These results indicate that PAT15/21‐mediated palmitoylation of γb is required for formation of αa‐γb‐TGB1 intermediate complexes during virus infection.

Taken together, these results show that NbPAT15 and NbPAT21 catalyze palmitoylation of γb protein, thereby shifting the viral replication phase to subsequent intra‐ and intercellular movement activities.

Palmitoylation enhances PM localization of γb and BSMV cell‐to‐cell movement

One of the distinctive features of palmitoylation of proteins is increased hydrophobicity and alterations of subcellular membrane localization patterns. Our previous study revealed the presence of γb in membrane‐enriched fractions during BSMV infection (Jiang et al, 2020). To determine whether palmitoylated γb is targeted to the PM during BSMV infection, A. tumefaciens harboring BSMV_γb‐GFP was infiltrated into N. benthamiana leaves. After plasmolysis, we observed that GFP fluorescence resulting from γb‐GFP was retracted along with the PM, and Hechtian strands (white arrow heads) connecting the PM to the CW (Fig 7A), suggesting that γb associates with the PM during BSMV infection.

Figure 7. Palmitoylation of γb facilitation of PM localization and BSMV cell‐to‐cell movement.

1. Plasma membrane (PM) targeting of γb in BSMV_γb‐GFP‐infected N. benthamiana leaves plasmolyzed at 72 hpi. Infiltrated leaf sections were cut into small strips and soaked in 5% NaCl for 5–10 min prior to confocal microscopy observations. The dotted lines in the merged panel highlight the cell wall (CW) after plasmolysis. GFP fluorescence was retracted along with the PM to reveal Hechtian strands connecting the PM to the CW (white arrow heads). Bars, 20 µm.
2. Fractionation assay to detect γb membrane associations. Separation of total protein (T) extracted from BSMV‐infected leaves into soluble (S) and membrane (M) fractions by ultracentrifugation at 100,000 g. PEPC and BiP, respectively, provide soluble and membrane fraction markers.
3. Aqueous two‐phase partitioning of γb PM localization. The microsomal pellet from BSMV_γb‐GFP‐infected leaves was obtained by ultracentrifugation at 100,000 g, and the two‐phase fractionation procedure was used to separate the plasma membrane (upper phase) from other cellular membranes (lower phase). H⁺‐ATPase is a PM marker.
4. PM associations of γb‐GFP and γb_3CS‐GFP in N. benthamiana protoplasts. The first and second columns show localization of γb at 3 dpi after infiltrations for transient expression of γb‐GFP and γb_3CS‐GFP. The third and fourth columns show γb localization in BSMV_γb‐GFP‐ and BSMV_3CS‐GFP‐infected leaves. PIP2A provided a PM marker (Nelson et al, 2007). Figures below the confocal images show the normalized fluorescence intensities of the GFP and mCherry proteins along the white dashed lines in the merged images. Scale bars, 10 μm.
5. Western blot analysis of subcellular fractions of BSMV‐infected N. benthamiana leaves at 3 dpi. The S3 and S30 fractions are the 3,000 and 30,000 g supernatants, and P30 is the 30,000 g pellet. Membrane‐associated γb was released by 1% Triton X‐100 treatment.
6. Western blot detection of BSMV accumulation in N. benthamiana plants. A. tumefaciens mixtures harboring RNAα and RNAβ_mTGB2, along with RNAγ, RNAγ_γb‐ROP, or RNAγ_γb‐mROP, were agroinfiltrated into N. benthamiana leaves. Western blotting to detect virus accumulation at 3 dpi used anti‐TGB1 or anti‐CP antibodies. Actin provided loading controls.
7. Analysis of the effects of enhanced PM association of γb on cell‐to‐cell virus movement with the BSMV duplex fluorescence (dfBSMV) reporter system. Mixtures of A. tumefaciens containing dfBSMV, dfBSMV_{γb‐GFP‐ROP}, and dfBSMV_{γb‐GFP‐mROP} were infiltrated into N. benthamiana leaves. Infiltrated leaves were observed by confocal microscope at 3 dpi. Scale bars, 100 μm. The determination of the movement capacity of the viruses depicted in each panel as descripted in Fig 3B (n = 6).

Data information: In (A, D, and G), representative data are shown and three biological replicates had similar results. In (G), the box boundaries indicate the upper (25^th percentile) and lower (75^th percentile) quartiles, whiskers indicate minimum and maximum values, central band indicates the median. Letters in the chart denote statistically significant differences among groups according to the Duncan’s multiple range test (P < 0.05).

Source data are available online for this figure.

To determine whether palmitoylation of γb contributes to PM localization, we performed cell fractionation analyses. The results showed that γb was present in both the soluble cytoplasmic fraction and the microsomal membrane fraction (Fig 7B). To extend this observation, we performed an aqueous two‐phase partitioning approach to separate the PM from the microsomal membranes (Chen et al, 2017; Xue et al, 2018). We found that γb‐GFP was present in both the upper phase (PM) fraction and lower phase (other microsomal membrane) fraction (Fig 7C). As shown in Fig 7D, γb‐GFP localized to the PM in N. benthamiana protoplasts in the presence or absence of BSMV infection. However, the non‐palmitoylated γb_3CS‐GFP mutant was only present in the cytoplasm and failed to associate with the PM (Fig 7D). These results indicate that palmitoylation is required for γb localization at the PM.

AtROP10 is a small GTP‐binding protein that functions as a molecular switch protein and localizes at the PM after palmitoylation. Only 25 C‐terminal amino acid residues are sufficient for PM anchoring, although substitution of 5 nonpolar residues (REDER) in the GC‐CG box (mROP) can disrupt PM associations (Lavy & Yalovsky, 2006; Chen et al, 2017). To investigate γb targeting to the PM, we fused a ROP tag to the C‐terminus of γb‐GFP to generate RNAγ_{γb‐GFP‐ROP}, and then extracted organelles from BSMV_γb‐GFP‐ or BSMV_{γb‐GFP‐ROP}‐infected N. benthamiana leaves. Immunoblotting of the S30 and P30 fractions showed that γb‐GFP was mainly present in the P30 membrane fraction (Fig 7E). However, when the P30 fraction was treated with the non‐ionic detergent Triton X‐100, approximately half of the γb‐GFP in the P30 fraction of BSMV_γb‐GFP‐infected N. benthamiana leaves was released into the S30 fraction. In contrast, only 8% of the γb‐GFP in the P30 fraction of BSMV_{γb‐GFP‐ROP}‐infected N. benthamiana leaves was released into the S30 fraction after Triton X‐100 treatment, whereas large amounts of γb‐GFP‐ROP remained in the P30 fraction. These results demonstrate that the C‐terminal domain of AtROP10 can be used to enhance the γb PM associations.

To determine whether the PM localization of γb affects BSMV infection, A. tumefaciens mixtures harboring RNAα and RNAβ_mTGB2, a premature translation termination mutant (Jiang et al, 2020), were mixed separately with RNAγ, RNAγ_γb‐ROP, or RNAγ_γb‐mROP (as a control) and agroinfiltrated into N. benthamiana leaves. Immunoblotting showed that the movement‐deficient virus encoding the γb‐ROP fusion protein significantly reduced TGB1 and CP accumulation, whereas leaves expressing γb‐mROP had no obvious effect on virus accumulation (Fig 7F), suggesting that disrupting of γb to the PM does not reduce BSMV replication. We also investigated whether enhanced PM localization γb affected BSMV cell‐to‐cell movement. For this purpose, mixtures of A. tumefaciens containing wild type dfBSMV, dfBSMV_{γb‐GFP‐ROP}, and dfBSMV_{γb‐GFP‐mROP} were infiltrated into N. benthamiana leaves followed by confocal microscopy analyses. At 60 hpi, the PM‐targeted γb‐GFP‐ROP resulted in significantly enhanced cell‐to‐cell movement of the dfBSMV_{γb‐GFP‐ROP} virus in the leaves, whereas the γb‐GFP‐mROP mutant that disrupted γb PM localization only slightly reduced intercellular virus movement (Fig 7G). These results show that PM association of palmitoylated γb promotes BSMV cell‐to‐cell movement.

Palmitoylated γb inhibits callose accumulation by NbREM1 and facilitates BSMV cell‐to‐cell movement

Due to the importance of palmitoylation for anchoring the γb protein to the PM and facilitating virus movement, we hypothesized that palmitoylated γb could also promote BSMV cell‐to‐cell movement by interacting with PM‐localized host proteins. To test this notion, we selected the N. benthamiana remorin 1 (NbREM1) protein, which has been reported to participate in various biotic and abiotic stress responses and to inhibit Potato virus X (PVX) and RSV cell‐to‐cell movement by enhancing callose deposition (Fu et al, 2018; Perraki et al, 2018). BiFC assays revealed that co‐expression of either the γb‐YFPn/NbREM1‐YFPc or γb‐YFPc/NbREM1‐YFPn combinations can lead to reconstitution of YFP fluorescence, but that the γb_3CS‐NbREM1 combination exhibited no obvious YFP fluorescence (Fig 8A, Appendix Fig S11A). Additional co‐IP assays also consistently confirmed interactions between γb and NbREM1 (Fig 8B). Together, these results show that γb, but not γb_3CS, interacts with NbREM1 at the PM in a palmitoylation‐dependent manner. In addition, we used the dfBSMV reporter system to monitor the effects of NbREM1 overexpression on viral movement, and the results showed that excess NbREM1 significantly inhibited cell‐to‐cell movement of dfBSMV. In contrast, the NbREM1_C206A dominant‐negative mutant (Fu et al, 2018) had little effect on BSMV cell‐to‐cell movement (Fig 8C, Appendix Fig S11B). These results show that NbREM1 expression reduces BSMV cell‐to‐cell movement.

Figure 8. Inhibition of callose accumulation during interactions of palmitoylated γb with NbREM1.

1. BiFC analyses of interactions between NbREM1 and γb or γb_3CS. Combinations of different constructs are indicated on the left. Scale bars, 20 μm.
2. Co‐IP analyses to analyze interactions of NbREM1 with γb or γb_3CS in N. benthamiana leaf tissues infiltrated with A. tumefaciens mixtures containing movement‐deficient BSMV plasmids (RNAα + RNAγ_γb‐GFP or RNAγ_3CS‐GFP) and NbREM1‐3xFlag. Total proteins were extracted at 3 dpi and immunoprecipitated with anti‐Flag beads. Input and immunoprecipitated proteins (IP) were analyzed by Western blot analyses with an anti‐GFP or anti‐Flag antibodies.
3. NbREM1 effects on BSMV cell‐to‐cell movement. A. tumefaciens containing NbREM1 or NbREM1 mutants were first agroinfiltrated into N. benthamiana leaves. After an additional 24 h, the dfBSMV reporter system was agroinfiltrated into the preinfiltrated regions of the leaves, and confocal microscopy was carried out 3 days later. Scale bars, 100 μm.
4. Analysis of the cleaved 5(6)‐carboxyfluorescein (CF) movement in N. benthamiana leaves. RNAα + RNAγ/RNAγ_3CS were co‐infiltrated into transiently overexpressed NbREM1 N. benthamiana leaves, and 1 μl CFDA was loaded onto the adaxial leaf blade surface and incubated for 5 min followed by confocal analysis at 3 dpi. The areas of dye diffusion are marked by dashed white circles. Scale bars, 100 µm.
5. Quantification of CF movement. Dye diffusion areas identified by the dashed white circles in panel D were measured by image J software (n = 6).
6. Aniline blue staining of callose in N. benthamiana leaf tissues. Leaves were infiltrated with mixtures of different A. tumefaciens constructs indicated beside the panels. At 3 dpi, the 0.1% aniline blue fluorochrome was infiltrated into N. benthamiana leaves. After incubation at 5 min in the dark, the leaves were observed by confocal microscopy at 405 nm. Scale bars, 20 μm.
7. Quantification of callose staining intensity as shown in panel F. The relative callose staining intensity was set to 1 in the REM samples (n = 10).
8. Aniline blue staining of callose in leaves with different treatments above the panel in NbPAT15 and NbPAT21 RNAi plants. Scale bars, 20 μm.
9. Quantification of callose staining intensity as shown in panel H. The relative callose staining intensity was set as 1 in the samples of EV (n = 8).

Data information: In (A, B, D, F, and H), representative data are shown and three biological replicates had similar results. In (E, G, and I), the box boundaries indicate the upper (25^th percentile) and lower (75^th percentile) quartiles, whiskers indicate minimum and maximum values, central band indicates the median. Letters in the chart denote statistically significant differences among groups according to the Duncan’s multiple range test (P < 0.05).

Source data are available online for this figure.

Several studies have found that NbREM1 is involved in callose deposition during virus infections (Raffaele et al, 2009; Fu et al, 2018; Perraki et al, 2018). Therefore, to determine whether γb affects the PD size exclusion limit (SEL) via hijacking NbREM1 in N. benthamiana, we performed a drop‐and‐see assay (DANS) (Gui et al, 2014) by using 5(6)‐carboxy fluorescein diacetate (CFDA), which is a symplastic tracer used to evaluate the PD SEL. We agroinfiltrated movement‐deficient BSMV (RNAα + RNAγ or RNAγ_3CS) with NbREM1 into N. benthamiana, and compared this combination with leaves agroinfiltrated with NbREM1 alone. The confocal results showed that cleaved carboxyfluorescein (CF) movement increased in the RNAα + RNAγ‐infected leaves, but was significantly restricted in the RNAα + RNAγ_3CS‐infected leaves (Fig 8D and E). Furthermore, aniline blue staining was performed followed by confocal microscopy quantification to evaluate callose distribution. The results showed that callose deposition increased in N. benthamiana agroinfiltrated with NbREM1 compared with mock‐infiltrated plants. Intriguingly, the γb protein significantly inhibited NbREM1‐mediated callose deposition in N. benthamiana epidermal cells, whereas γb_3CS did not (Fig 8F and G). Together, our data demonstrate that palmitoylated γb interacts with NbREM1 at the PM to manipulate callose deposition at the PD.

To further investigate whether NbPAT15 and NbPAT21 contribute to γb‐mediated suppression of callose deposition, RNAα + RNAγ were inoculated into leaves of NbPAT15‐RNAi and NbPAT21‐RNAi plants. The leaves were harvested at 3 dpi, stained with aniline blue, and observed by confocal microscopy. The results showed that callose accumulation increased in the NbPAT15‐RNAi or NbPAT21‐RNAi plants compared with the unsilenced plants, indicating that NbPAT15 and NbPAT21 are required for γb suppression of callose deposition (Fig 8H and I). As a control, the knock‐down of PATs in the absence of BSMV infection does not have a discernible effect on callose in the PD (Appendix Fig S14). Together, these results show that palmitoylated γb inhibits callose deposition at the PD to promote BSMV cell‐to‐cell movement.

Discussion

Although viral replication and movement processes are thought to be tightly coupled, the detailed regulatory mechanisms underlying changes from the replication to movement phases of plant virus infections are obscure (Heinlein, 2015). Our study of BSMV infection reveals that palmitoylation of the γb protein acts as a molecular switch to convert replication to the movement phases: we show that (i) non‐palmitoylated γb predominantly associates with the chloroplasts and promotes BSMV replication, (ii) palmitoylated γb protein recruits TGB1 to the chloroplasts and functions in BSMV cell‐to‐cell movement, (iii) γb interacts with NbPAT15 and NbPAT21 at the chloroplast periphery and is transferred to the ER and the PM as movement processes are initiated, and (iv) palmitoylated γb subsequently binds NbREM1 at the PM to enlarge PD SEL and facilitates cell‐to‐cell movement.

The dynamic switching of plant viruses from the replication to movement phases largely depends on recruitment of MPs to VRCs (Tilsner & Oparka, 2012). For example, the tobacco mosaic virus (TMV) 30k MP binds to viral RNAs in VRCs to link replication with movement (Liu et al, 2005; Tilsner & Oparka, 2012; Heinlein, 2015; Jiang & Laliberté, 2016; Pitzalis & Heinlein, 2017). Red clover necrotic mosaic virus (RCNMV) RNA1 also recruits MPs to VRCs, a process facilitated by glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH) (Kaido et al, 2009, 2014; Tilsner & Oparka, 2012). BSMV TGB1 localizes to the cytoplasm and nuclei when expressed alone, but mainly accumulates at the PD during BSMV infection (Hu et al, 2015; Li et al, 2018b). TGB1 is a core component of BSMV RNP movement complexes (Jackson et al, 2009; Li et al, 2018b); however, the mechanisms whereby TGB1 associates with BSMV VRCs have not been determined. Previously, we found that γb acts as a multifunctional protein by participating in both BSMV replication and cell‐to‐cell movement (Zhang et al, 2017; Jiang et al, 2020). Here, we have extended our previous BSMV studies by demonstrating that γb palmitoylation results in recruitment of TGB1 to the VRCs. These findings provide a novel regulation strategy in which palmitoylation has an essential role in the recruitment of MPs to VRCs.

The multifunctionality of virus‐encoded proteins is closely related to PTMs that affect diverse subcellular locations of virus components. For example, tomato leaf curl Yunnan virus (TLCYnV) C4 is phosphorylated by NbSKη in the nucleus, which in turn promotes myristoylation of C4 and directs relocalization to the PM (Mei et al, 2018). The BSMV γb protein, serine residue at site 96 (Ser‐96), can be phosphorylated by a PKA‐like protein kinase, which in turn functions to suppress host RNA silencing and cell death responses (Zhang et al, 2018). Moreover, the cytosolic kinase, NbSTY46, also phosphorylates the γb protein to negatively regulate BSMV infection (Zhang et al, 2021). These results suggest that phosphorylation modifications of γb have integral regulatory roles in molecular interplays between BSMV and the host plant. Recent studies have suggested that S‐adenosylmethionine decarboxylase 3 (SAMDC3) negatively regulates BSMV infection by promoting the ubiquitination and 26S proteasome‐mediated degradation of the γb protein (Li et al, 2022). However, it was not known previously whether other PTMs are involved in regulation of BSMV infections. Here, we found that palmitoylation of the BSMV γb protein affects the subcellular distribution of γb and permits γb to execute various functions at specific sites. Since γb can be phosphorylated, palmitoylated, and ubiquitinated at different sites, potential crosstalk among these PTMs is worth investigating in future studies.

Palmitoylation is a reversible PTM that is catalyzed by protein palmitoyl acyltransferases (PATs) and acyl‐protein thioesterases (APTs) to direct two opposing processes (Hemsley, 2020; Liu et al, 2021b). Palmitoylation is a PTM occurring in all eukaryotes, but little is known about the enzymology of palmitoylation in planta. A total of 24 PATs containing a conserved DHHC motif have been identified in A. thaliana, and it is predicted that each PAT may affect more than 100 substrates that undergo palmitoylation modifications in different temporal and spatial contexts (Hemsley, 2020). This complexity creates enormous difficulties in determining the functions of specific PATs by traditional genetic methods. Here, we have used transcriptome sequencing and RNA silencing approaches to demonstrate that γb is a substrate for NbPAT15 and NbPAT21. Generally, PATs are thought to localize on the membrane surfaces of individual organelles and to be responsible for palmitoylation of various proteins (Hemsley, 2020). Although AtPAT15 and AtPAT21 of A. thaliana were predicted to locate at the ER and PM, respectively (Batistic, 2012), we found that γb can interact with NbPAT15 and NbPAT21 at the chloroplast periphery, suggesting that γb might be involved in their recruitment from ER/PM sites to VRCs. It is conceivable that the relocalization of NbPAT15 and NbPAT21 could also affect the palmitoylation landscape of various host proteins affecting BSMV infection, albeit whether such modifications provide strategies for viruses to escape from host surveilance needs to be investigated in future studies.

Plant viruses usually enlarge PD apertures in two broad ways. TMV and cucumber mosaic virus (CMV) MPs can destroy PD‐associated actin directly to enlarge the PD SEL (Su et al, 2010). A more common approach employed by PVX TGB2 (Fridborg et al, 2003) and the TMV MP (Ueki et al, 2010) is to decrease callose deposition at cell walls. More recently, remorins have been shown to anchor PM nanodomains and to regulate PD SEL, and these proteins are hijacked by several virus proteins including PVX TGB1 (Perraki et al, 2018), RSV NSvc4 (Fu et al, 2018), and turnip mosaic virus (TuMV) VPg (Cheng et al, 2020) to facilitate cell‐to‐cell movement. In our current study, we found that NbREM1 is manipulated by BSMV γb to facilitate cell‐to‐cell movement. In this case, palmitoylated γb, but not TGB1 and TGB3 (Appendix Fig S12), interacts with NbREM1 to inhibit callose accumulation at the PD, whereas the non‐palmitoylated γb_3CS mutant impairs associations with NbREM1 in vivo (Fig 8). These results suggest that palmitoylation of γb not only modulates the switch between virus replication and subcellular movement, but also facilitates intercellular virus movement through the PD. These observations further illustrate the intricate interplay occurring during different stages of BSMV infection.

Interestingly, cysteine‐rich proteins (CRPs) are encoded by numerous rod‐shaped and filamentous viruses in diverse genera, including Hordei‐, Furo‐, Gora‐, Peclu‐, Pomo‐, and Tobravirus in the family Virgaviridae, as well as Benyvirus in the family Benyviridae, and Carlavirus and Vitivirus in the family Betaflexiviridae, aside from Allexivirus and Mandarivirus in the family Alphaflexiviridae (https://talk.ictvonline.org/taxonomy/). Although CRPs of different virus genera generally have low amino acid sequence similarity, they are usually characterized by the presence of multiple cysteine residues and zinc finger‐like motifs (Adams et al, 2017; Fujita et al, 2018), and importantly, the Cys‐10, Cys‐19, and Cys‐60 residues of hordeivirus γb proteins are highly conserved (Appendix Fig13) (Li et al, 2020; Reynard et al, 2021). To investigate the extent to which the BSMV γb functions are conserved, another chloroplast‐replicating hordeivirus, lychnis ringspot hordeivirus (LRSV), was selected for further experiments. The cysteines at sites 13, 24, and 65 of LRSV γb were substituted with serines to generate the LRSV_3CS mutant (Appendix Fig S13), and the mutant virus was agroinfiltrated into N. benthamiana plants. At 10 dpi, LRSV_3CS‐infected upper leaves failed to develop visible symptoms compared with those of leaves inoculated with LRSV. Moreover, Western blots revealed that the movement‐deficient LRSV_3CS (RNAα + RNAγ_3CS) accumulated significantly higher γb protein than the wild type (RNAα + RNAγ) in N. benthamiana leaves (Appendix Fig S15). However, whether CRPs of more diverse viruses are palmitoylated and have roles as viral infection switches is an open question that deserves to be investigated in the future.

One of our previous studies demonstrated that BSMV γb proteins are recruited to VRCs to enhance BSMV replication (Zhang et al, 2017) and showed that γb enhances the assembly of vRNP movement complexes at the periphery of chloroplasts to promote cell‐to‐cell movement (Jiang et al, 2020). Here, we extended that study by finding that palmitoylation of γb acts as a molecular switch to activate virus movement (Fig 9). The non‐palmitoylated γb protein is a component of the VRC that functions to promote BSMV replication. During the late stages of replication, γb interacts with NbPAT15 and NbPAT21 at the periphery of the chloroplasts and is palmitoylated at Cys‐10, Cys‐19, and Cys‐60, leading to recruitment of the TGB1 movement protein to the chloroplasts to initiate intracellular movement. Moreover, palmitoylated γb is targeted to the PM to direct interactions with host factors such as NbREM1 that result in increased PD SEL and facilitate intercellular spread of the virus. In summary, our study, for the first time, identifies a novel function whereby palmitoylation of the cysteine‐rich γb protein initiates a dynamic replication‐movement switch during infection.

Figure 9. A proposed model to illustrate the role of γb palmitoylation in BSMV infection.

Upon virus entry into the cell, BSMV RNAs are released and translated. The αa and γa replicase proteins initially recruit various host factors (HF) and γb protein to assemble viral replication complexes (VRCs) at the chloroplast outer membranes. At the later stages of replication, NbPAT15 and NbPAT21 interact with the γb protein at the chloroplast periphery and palmitoylate the Cys‐10, Cys‐19, and Cys‐60 sites. Then, the palmitoylated γb recruits TGB1 to chloroplast replication sites to form intermediate replication‐movement complexes. The viral RNAs bind to the palmitoylated γb protein and to the TGB1‐TGB2‐TGB3 complex to form stable viral ribonucleoprotein (vRNP) movement complexes that move along the ER/actin network. Palmitoylated γb also interacts with NbREM1 at the plasma membrane (PM) to inhibit callose deposition and increase PD size exclusion limits (SEL) to facilitate virus cell‐to‐cell movement.

Materials and Methods

Plant growth conditions

Nicotiana benthamiana plants were grown in a climate chamber at 23–24°C with a 15/9 h light/dark photoperiod as described previously (Yuan et al, 2011). Barley (Yangfu 4056) was grown in the greenhouse until the one‐leaf stage, and then transferred to a climate chamber as described previously (Lee et al, 2012; Hu et al, 2015).

Plasmid construction

Generation of the BSMV ND18 infectious cDNA clones, pCB301‐α, pCB301‐β, and pCB301‐γ, has been described in a previous study (Hu et al, 2019). The γb_3CS (C10S, C19S, and C60S) mutant and other mutant derivatives were constructed in pCB301‐γ by using reverse‐PCR methods.

For the biotin‐switch experiments, the γb mutants were constructed based on the pGD‐γb‐3xFlag (Zhang et al, 2018) by using reverse‐PCR.

For subcellular localization assays, γb and mutant derivatives were cloned into the SalI sites of the pGDGm and pGD‐CFP vectors (Goodin et al, 2002), respectively.

The BiFC assays used γb_3CS and other γb derivatives that had been inserted into the pSPYNE‐35S or pSPYCE‐35S BamHI sites (Walter et al, 2004).

To analyze virus cell‐to‐cell movement, dfBSMV_3CS and other BSMV derivatives were constructed in the pCB301‐γ_dupflu vector using the QuikChange‐PCR method (Li et al, 2018b). Rop or mRop tag derivatives (Chen et al, 2017) were amplified by reverse‐PCR, digested, and inserted into the corresponding plasmids.

All the primers used for constructing these plasmids are listed in Appendix Table S1, and all mutants used in this study were confirmed by DNA sequencing.

Biotin‐switch technique

Biotin‐switch experiments were performed as described previously with minor modifications (Hemsley et al, 2008). Briefly, A. tumefaciens harboring pGD‐γb‐3xFlag and other derivatives were agroinfiltrated into N. benthamiana leaves. Leaf tissues (250–500 mg) were harvested at 3 dpi, ground in liquid nitrogen, and suspended in 500 µL of freshly prepared lysis buffer [1 × PBS pH 7.4, protease inhibitors, 1 mM EDTA, 1% Triton X‐100, 25 mM N‐ethylmaleimide (NEM) (Thermo Scientific, E3876‐5G)]. The mixtures were stirred for 1 h at 4°C on a rocker platform and centrifuged at 4°C for 10 min at 500 g. The supernatants were then incubated overnight in 25 mM NEM at 4°C on a rocker platform, and the proteins were precipitated at room temperature with a methanol/chloroform extraction method (Hemsley et al, 2008). Briefly, four volumes of methanol were mixed with the precipitates followed by incubation at −20°C for 20 min. After centrifugation at 4°C, 10,000 g for 20 min, the pellet was resuspended by sonication in 200 μl resuspension buffer (1 × PBS, pH 7.4, 8 M Urea, 2% SDS) and divided into two equal aliquots. One aliquot was combined with 800 μl of freshly prepared 1 M hydroxylamine solution containing 1mM EDTA, protease inhibitors, and 100 μl fresh 4 mM biotin‐HPDP (Thermo Scientific) dissolved in DMF and gently mixed for 1 h at room temperature. The remaining aliquot was treated similarly except that the hydroxylamine was replaced with 50 mM Tris–HCl (pH 7.4). The treated proteins were precipitated again at room temperature with methanol/chloroform, and the pellets were resuspended in 200 μl resuspension buffer and mixed with 900 μl PBS containing 0.2% Triton X‐100. The final supernatants were incubated with neutravidin‐agarose beads (Thermo Scientific) for 1 h at room temperature, and the beads containing the immunoabsorbed proteins were pelleted and washed twice with wash buffer (1 × PBS, pH 7.4, 500 mM NaCl, 0.1% SDS) and twice with 1 × PBS buffer (pH 7.4) to release the proteins. The final protein extracts were analyzed by Western blotting with an anti‐Flag antibody (at 1:5,000 dilution; Sigma‐Aldrich, Cat. #A2220).

Northern blot

Northern blotting was performed as described previously (Zhang et al, 2017). Briefly, total RNAs were quantified by a NanoDrop ND‐1000 procedure (Thermo Fisher Scientific). To detected BSMV RNAs, 3 μg of total RNAs were separated in 1.2% agarose gels containing 1.1% formaldehyde. After electrophoresis, the total RNAs were transferred onto Hybond‐N⁺ nylon membranes (GE Healthcare), followed by UV cross‐linking, and stained with methylene blue solution (0.04% methylene blue, 500 mM NaOAc). After hybridizing with [γ‐³²P] UTP‐labeled BSMV 3′‐UTR transcript probes at 65°C overnight, the RNAs were sequentially incubated with 2×, 1×, and 0.5× SSC buffer to reduce nonspecific binding of the probe, and images of the hybridized RNAs were visualized with a Typhoon 9400 PhosphorImager (GE Healthcare).

Subcellular localization assays and confocal microscopy

Nicotiana benthamiana leaves were observed with Leica SP8 or Zeiss LSM880 confocal microscopes. GFP and YFP fluorophores were excited with a 488 nm laser. CFP, mCherry, and chlorophyll autofluorescences were visualized at 440, 543, and 633 nm, respectively. Overlapping fluorescence spectra were processed by line scan analysis using the pixel‐based method with the ImageJ plot profile tool. Pearson’s correlation coefficients (Pearson’s r) of GFP and RFP were measured with the ImageJ PSC co‐localization tool.

Isolation of chloroplasts

Isolation of N. benthamiana chloroplasts was performed as described previously with minor modifications (Ling & Jarvis, 2015). Briefly, infiltrated N. benthamiana was placed in a dark chamber for 10 h to reduce the starch content of the chloroplasts before harvesting the leaf tissues. Approximately, 10 g leaves were ground in a mortar and pestle with 60 ml of pre‐cooled 1× buffer I [0.15 M sorbitol, 2.5 mM MgCl₂, 2.5 mM EDTA, 2.5 mM EGTA, 5 mM NaHCO₃, 10 mM HEPES‐KOH (pH 8.0), 10 mM DTT and 0.1% BSA]. The tissue brei was removed by filtration through Miracloth (Sigma‐Aldrich), and the pellet resulting after centrifugation at 1,000 g for 10 min was resuspended in 10 ml of buffer II [0.3 M sorbitol, 1 mM MgCl₂, 2 mM EDTA, 50 mM HEPES‐KOH (pH 8.0)]. After centrifugation at 4°C, 1,000 g for 5 min, the pellet was resuspended in 8 ml Buffer II. The exact was layered over a Percoll step gradient (40 and 80% Percoll in buffer II), and centrifuged at 13,000 g for 20 min with deceleration at the lowest setting. Intact chloroplasts at the interface between the 40 and 80% Percoll layers were washed twice in buffer II at 4°C by pelleting at 1,000 g for 5 min and analyzed by Western blotting.

Nicotiana benthamiana protoplast preparation

Mesophyll protoplasts of N. benthamiana were prepared from A. tumefaciens infiltrated leaves as described previously with minor modifications. Briefly, approximately 1 g N. benthamiana leaves were harvested and cut into small pieces with a sharp razor blade at 3 dpi, followed by incubation in 15 ml enzyme solution [7.28% D‐mannitol, 1.5% (wt/vol) cellulase R‐10 (Yakult, Japan, MX7352), 0.3% (wt/vol) macerozyme R‐10 (Yakult, Japan, MX7351), 20 mM KCl, 20 mM morpholineethanesulfonic acid (MES), 10 mM CaCl₂, and 0.1% bovine serum albumin (BSA)] for 3 h at room temperature in the dark on a gentle shaker at 60 rpm. The extracts were then filtrated through a 125 μm nylon mesh screen and centrifuged for 3 min at 100 g. The pelleted protoplasts were resuspended in equal volumes of W5 buffer (131 mM NaCl, 125 mM, 5 mM KCl, 2 mM MES, pH 5.6), followed by centrifugation at 100 g for 3 min. Each pellet was then resuspended with 2 ml W5 buffer, followed by observations with a confocal microscope as described above.

Plasmolysis

Plasmolysis of leaves was performed as described previously with minor modifications (Lim et al, 2009). Plasmolysis was conducted at 72 h post infiltration by cutting and soaking the infiltrated leaf tissues in 5% NaCl for 5–10 min prior to visualization under a confocal microscope.

Membrane flotation assays

Membrane flotation assays were performed as described previously with minor modifications (Han & Sanfacon, 2003; Cao et al, 2015). A. tumefaciens containing various expression plasmids were co‐infiltrated into N. benthamiana leaves. The leaves were harvested at 3 dpi and ground in liquid nitrogen followed by suspension in homogenization buffer [50 mM Tris–HCl, pH 8.0, 10 mM KCl, 3 mM MgCl₂, 1 mM EDTA, 1 mM dithiothreitol (DTT), 0.1% BSA, 0.3% dextran, 13% sucrose, 1% protease inhibitor cocktail (Roche, Cat. #45364400)]. The extracts were then extruded through two layers of Miracloth followed by centrifugation at 3,000 g for 10 min. The supernatant (S3) was centrifuged at 30,000 g at 4°C for 1 h to obtain the soluble (S30) pellet (P30) fractions.

Triton X‐100 treatment was performed as described previously with minor modification (Han & Sanfacon, 2003). P30 pellets were resuspended in homogenization buffer containing 1% Triton X‐100 and incubated on ice for 1 h. After centrifugation at 30,000 g for 1 h, the resulting pellets were resuspended in equal volumes of homogenization buffer to obtain the soluble S30 fractions and pellet P30 fractions again. The percentages of γb solubilized by 1% Triton treatment was calculated by (S30 / S30 + P30) × 100%.

The aqueous two‐phase partitioning procedure was performed as described previously (Chen et al, 2017; Xue et al, 2018). Briefly, the BSMV_γb‐GFP‐infected N. benthamiana leaves were harvested at 3 dpi and ground in liquid nitrogen followed by suspended in isolation buffer (0.33 M sucrose, 0.6% polyvinylpyrrolidone, 0.2% bovine serum albumin, 10% glycerol, 5 mM DTT, 5 mM EDTA, 5 mM potassium phosphate, and 50 mM HEPES‐KOH, pH 7.5), and centrifuged at 4°C for 10 min at 8,000 g. The extracts were next passed through two layers of Miracloth followed by centrifugation at 100,000 g for 1 h. The resulting pellets were resuspended in buffer I (0.33 M sucrose, 3 mM KCl, 0.1 mM EDTA, and 5 mM potassium phosphate, pH 7.5) and centrifuged at 4°C for 5 min at 1,000 g into a two‐phase solution (6.2% Dextran T‐500, 0.33 M sucrose, 5 mM potassium phosphate, and 6.2% polyethylene glycol 3350, pH 7.5) to separate the PM (up phase) from other membranes (down phase). The upper phase was diluted with buffer II (0.33 M sucrose, 0.1 mM EDTA, and 20 mM HEPES‐KOH, pH 7.5), and then centrifuged at 100,000 g at 4°C for 1 h. The resulting pellet was resuspended with 1 × PBS buffer (pH 7.4) and analyzed by Western blotting.

Co‐IP

Co‐IP assays were performed as described previously (Jiang et al, 2022b). Briefly, A. tumefaciens derivatives harboring various expression plasmids were co‐infiltrated into N. benthamiana leaves. Total proteins were extracted from the leaves at 3 dpi by grinding in GTEN buffer (25 mM Tris–HCl, pH 7.5, 1 mM EDTA, 150 mM NaCl, 10% glycerol) with 0.2% NP40, 2% PVP40, 10 mM DTT, and protease inhibitor cocktail. The extracts were incubated on ice for 30–60 min followed by centrifugation at 8,000 g, 4°C for 15 min. After an additional centrifugation to remove residual insoluble contaminants, the supernatants were incubated with anti‐FLAG M2 magnetic beads (Millipore, M8823) for 4 h at 4°C on a rocker platform. The immunoprecipitated proteins were then washed six times with IP buffer (GTEN buffer containing 0.2% NP‐40) and analyzed by Western blotting with an anti‐Flag antibody (at 1:5,000 dilution; Sigma‐Aldrich, Cat. #A2220) or an anti‐GFP antibody.

PD size exclusion limit analysis and callose staining

PD SEL analysis and callose staining assays were performed as described previously (Lee et al, 2011). A 1 μl drop of 1 mM 5‐carboxyfluorescein diacetate (CFDA) (Sigma‐Aldrich, 79955‐27‐4) was loaded on the adaxial side of N. benthamiana leaves, and after incubating for 5 min, the tissues were cut into small strips and visualized at 488 nm with a Leica SP8 confocal microscope.

For callose staining assays, 0.1% aniline blue fluorochrome was infiltrated into N. benthamiana leaves. After incubation for 5 min in the dark, the leaves were observed by confocal microscopy at 405 nm (Gui et al, 2014). The size and intensity of the aniline blue fluorescence puncta were measured using ImageJ software (Bio‐Rad).

RNA extraction and RNA sequencing analysis

Nicotiana benthamiana plants were infiltrated with BSMV, BSMV_mγb, and the empty vector, which was used as a negative control. Total RNA was extracted from infiltrated leaves at 3 dpi using Trizol reagent (Invitrogen, CA, USA), and subjected to transcriptome sequencing at the Hengchuangjiyin Company (Shenzhen, China). A cDNA library was created with the Illumina sequencing platform, and paired‐end sequencing was conducted with the Hiseq PE150 method. About 40 million valid reads were obtained for each sample, and the clean reads were aligned with the Solanaceae Genomics Network (https://solgenomics.net/). Differentially expressed genes were chosen (log₂ (fold change) > 1 or log₂ (fold change) < −1, with statistical significance (P value < 0.05).

Statistical analyses

The data shown in this study were expressed as mean ± SEM from three independent experiments. The protein band signals were quantified with ImageJ software (Bio‐Rad), and statistical analyses were performed with the Student’s t‐test or one‐way ANOVA.

Accession numbers

Sequence information used in this article is available at Sol Genomics Network (https://solgenomics.net) and Nicotiana benthamiana Genome & Transcriptome database (http://benthgenome.qut.edu.au/) with the following accession numbers: PAT15 (Nbv5.1tr6213778), PAT19 (Nbv5.1tr6206560), PAT14 (Nbv5.1tr6224935), PAT11 (Nbv5.1tr6205253), PAT21 (Niben101Scf00712g06009.1), and PAT‐like (Nbv5.1tr6246553).

Author contributions

Ning Yue: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Writing—original draft; Writing—review & editing. Zhihao Jiang: Data curation; Formal analysis; Investigation; Methodology; Writing—original draft; Writing—review & editing. Xuan Zhang: Investigation; Methodology. Zhenggang Li: Investigation; Methodology. Xueting Wang: Investigation; Methodology. Zhiyan Wen: Investigation; Methodology. Zongyu Gao: Investigation; Methodology. Qinglin Pi: Investigation; Methodology. Yongliang Zhang: Resources; Data curation; Formal analysis; Funding acquisition; Methodology; Writing—original draft; Writing—review & editing. Xian‐Bing Wang: Resources; Data curation; Formal analysis; Methodology. Chenggui Han: Resources; Data curation; Formal analysis; Methodology. Jialin Yu: Resources; Data curation; Formal analysis; Methodology. Dawei Li: Conceptualization; Resources; Data curation; Formal analysis; Supervision; Funding acquisition; Methodology; Writing—original draft; Project administration; Writing—review & editing.

In addition to the CRediT author contributions listed above, the contributions in detail are:

DL and NY conceived and designed the experiments. NY performed the experiments with the help of ZJ, XZ, ZL, XW, ZW, ZG, and QP. YZ, X‐BW, CH, and JY discussed and interpreted the data. NY, ZJ, DL, and YZ wrote the manuscript.

Disclosure and competing interests statement

The authors declare that they have no conflict of interest.

Supporting information

Appendix

Dataset EV1

Dataset EV2

Dataset EV3

Source Data for Figure 1

Source Data for Figure 2

Source Data for Figure 3

Source Data for Figure 4

Source Data for Figure 5

Source Data for Figure 6

Source Data for Figure 7

Source Data for Figure 8

The EMBO Journal (2022) 41: e110060.

Data availability

This study includes no newly generated data deposited in external repositories.