Systemic infection of tobacco mosaic virus is limited by upward transport and phloem unloading without coat protein

SUMMARY

Coat protein (CP) is widely viewed as essential for systemic infection of tobacco mosaic virus (TMV), yet the mechanistic basis of this requirement remains incompletely resolved. In this study, we show that TMV can still achieve long‐distance vascular transport in the absence of CP, but systemic infection becomes intermittent because the processes underlying “systemic movement” can fail at multiple, separable steps. By resolving movement into phloem loading, vascular translocation, and phloem exit, we show that CP deficiency does not completely block entry into or transport within the vasculature. Instead, CP loss imposes a strong, route‐specific defect in bicollateral phloem, disproportionately reducing use of the adaxial/internal pathway associated with efficient upward spread, while downward movement through abaxial/external phloem remains comparatively permissive. In systemic leaves, CP‐deficient TMV frequently produces vein/phloem‐associated reporter patterns rather than broad lamina invasion, indicating that phloem exit/unloading is an additional, independent bottleneck. Enhancing RNA silencing suppression markedly enhances systemic viral accumulation and symptoms. However, it does not proportionally increase lamina invasion. This finding further supports the idea that exit and establishment are the limiting steps. Finally, among several helper viruses tested, only cucumber mosaic virus consistently alleviates this exit barrier, enabling CP‐deficient TMV to escape the vasculature and establish lamina infection.

Significance Statement

Tobacco mosaic virus has long been thought to require coat protein for systemic infection, but the affected step in plant‐wide spread has remained unclear. This study shows that, while coat protein promotes efficient upward transport and escape from veins, it is not an absolute requirement for vascular movement. This reframes systemic infection as a stepwise process with separable bottlenecks.

INTRODUCTION

Plant viruses must move beyond initially infected cells to establish systemic infection and, ultimately, transmission. After entry and replication in the inoculated tissue, most viruses spread locally through plasmodesmata (PD) via virus‐encoded movement proteins (MPs) that remodel PD permeability and traffic into neighboring cells (Heinlein, 2015). Long‐distance movement typically requires access to the phloem translocation stream, where viruses travel rapidly across the plant body, followed by unloading and invasion of non‐vascular tissues in distal sinks (Lezzhov et al., 2021). Importantly, these steps are not equivalent. Phloem entry (“loading”), translocation within sieve elements, and exit (“unloading”) are governed by distinct anatomical interfaces and host control points and can therefore be genetically and mechanistically separable (Folimonova & Tilsner, 2018).

For many plant viruses, the coat protein (CP) contributes to one or more of these transport steps. CP can be required to form virions or other transport‐competent assemblies that stabilize and protect the genome, and in some viruses, CP also contributes directly to efficient cell‐to‐cell movement (Hipper et al., 2013). Yet the degree to which encapsidation is necessary for systemic infection varies across virus families. This leads to a question of to what extent can long‐distance movement proceed without CP, and what precisely fails when CP is missing?

The tobacco mosaic virus (TMV; family Virgaviridae) is a promising model for investigating these principles, as its genome organization, replication, and movement processes have been studied in great detail (Carr, 2004). TMV has a genome of approximately 6.4 kb that encodes at least four proteins, including two replication‐associated proteins (126 and 183 kDa), a 30‐kDa MP, and a 17.5‐kDa CP (Scholthof, 2004). The possible functions of two additional open reading frames (ORFs) that encode a 54‐kDa protein and 4–6‐kDa proteins, designated as P6 proteins, have also been proposed (Palukaitis et al., 2024). The MP is sufficient for robust cell‐to‐cell movement, whereas the CP has been classically associated with efficient systemic movement and symptom development (Heinlein, 2015). CP‐defective mutants often exhibit reduced systemic infection, and systemic movement can be restored in transgenic plants expressing TMV CP (Osbourn et al., 1990). In line with this, CP‐deletion TMV replicons are widely used as high‐expression vectors that are expected to remain largely confined to inoculated tissues (Lindbo, 2007).

At the same time, older work on naturally arising or engineered defective TMV variants suggested that systemic infection might still occur, albeit inefficiently, without a fully functional CP gene (Dawson et al., 1988; Siegel et al., 1962). More recently, this topic has re‐emerged with the discovery that gain‐of‐function MP mutants with C‐terminal truncations were reported to restore systemic transport of a CP‐defective TMV (Tran et al., 2022). In addition, TMV CP can influence systemic infection through suppressing salicylic acid (SA) signaling, and downregulating the pathway, without complementing CP, restored systemic movement (Venturuzzi et al., 2021). These findings sharpen the hypothesis that, under particular genetic configurations, TMV RNA can traverse long distances without canonical encapsidation, while still leaving open which step(s) of systemic spread remain limiting without CP.

A useful framework is to treat systemic infection as a multi‐step logistics problem. First, viruses must reach vascular‐associated cells in source leaves and access sieve elements/companion cell complexes. Movement across these heterogeneous cell interfaces can impose rate‐limiting steps for long‐distance movement (Thompson & García‐Arenal, 1998; Vuorinen et al., 2011). In TMV, CP‐deficient mutants accumulate poorly in vascular elements, suggesting that CP is involved in phloem loading and/or stability during transport (Ding et al., 1996). Once inside the phloem, viral RNA must remain protected and functional. Although RNase activity is not usually detected in phloem exudates (Kehr & Kragler, 2018), phloem mRNAs can be regulated and degraded (Banerjee et al., 2006; Xia et al., 2018), implying additional layers of RNA quality control. Antiviral RNA silencing intersects with this stage by shaping the persistence of viral RNA and the ability to establish infection at endpoints. A clear example of functional decoupling is provided by turnip crinkle virus (TCV): Systemic spread depends on both CP‐associated silencing‐suppressor activity and CP‐dependent particle assembly. These processes map to different systemic barriers in silencing‐competent hosts (Cao et al., 2010). Silencing suppressors can more broadly affect systemic outcomes disproportionately relative to local accumulation, implicating the defense state as a limiter of successful long‐distance transport and/or establishment (Deleris et al., 2006).

Lastly, viruses need to exit into sink tissues and expand away from veins. In Nicotiana tabacum, silencing pectin methylesterase has been shown to prevent TMV from exiting the phloem (Chen & Citovsky, 2003). Several viruses are naturally phloem‐limited, such as luteoviruses and geminiviruses, though co‐infection with a helper virus can overcome this barrier (Pohl & Wege, 2007; Savenkov & Valkonen, 2001).

In many solanaceous hosts, additional complexity arises from compartmentalization within the vasculature, including internal (adaxial) and external (abaxial) phloem systems that can exhibit selective macromolecule trafficking properties (Turgeon, 2016). Thus, directional biases (basipetal vs. acropetal transport) and vascular‐domain boundaries may strongly shape the apparent systemic movement of viruses.

Here, we revisit TMV systemic movement in N. benthamiana using a CP‐deficient TMV (TMVΔCP) reporter system. We investigated whether a TMVΔCP replicon can enter the phloem and spread systemically, and do so with directional preferences consistent with the TMV transport route. We also examined what constrains its ability to invade distal tissues beyond the veins. Finally, since mixed infections can relax transport bottlenecks for phloem‐limited viruses, we examined whether co‐infection with other viruses could enable steps that are inefficient for TMVΔCP alone. Together, these studies reframe CP‐independent TMV movement as a stepwise phenotype rather than a binary one. Although TMVΔCP can achieve long‐distance transport signatures, its efficient exit from veins remains restricted.

RESULTS

CP‐deficient TMV occasionally appears on distal leaves

To monitor long‐distance movement of CP‐free TMV, we used an infectious TMV clone in which GFP replaces the CP open reading frame (pTMVΔCP‐GFP). As a CP‐positive control, we used a full‐length TMV‐GFP clone in which GFP is followed by a sgp and the intact CP coding sequence (pTMV‐GFP). Additionally, we generated both MP‐ and CP‐deletion mutants that are movement defective (pTMVΔMPΔCP‐GFP) (Figure 1A). In inoculated leaves, TMVΔCP‐GFP produced strong GFP fluorescence and accumulated viral RNA to levels comparable to the full‐length TMV‐GFP clone, with faster induction kinetics (Figure 1B). This was also evident in whole‐leaf imaging of the inoculated sites. By contrast, TMVΔMPΔCP‐GFP produced weak GFP and was therefore supplemented with P19, yet fluorescence remained restricted to the agroinfiltrated area with no spread into adjacent vein regions. By comparison, TMVΔCP‐GFP and TMV‐GFP expanded from the inoculation zone, filled vein‐associated regions, and further spread beyond the initially infected area (Figure 1C).

Figure 1.

Coat protein (CP)‐deficient tobacco mosaic virus (TMV) produces systemic infection foci.

(A) Schematic representation of virus constructs. The clones were derived from the TMVΔCP‐GFP backbone (Addgene #80083).

(B) Time‐course quantification of viral RNA accumulation in inoculated leaves (0–6 days postinoculation, dpi) by quantitative reverse transcription polymerase chain reaction (RT‐qPCR). Statistical comparisons were performed using two‐tailed t‐tests (P < 0.05).

(C) Representative fluorescence outputs from TMV vectors in inoculated leaves at 3 dpi. Intensity is displayed as pseudo‐color to facilitate comparison of spatial distribution. P19 was co‐expressed with TMVΔMPΔCP‐GFP to match the expression level.

(D) Green fluorescent protein (GFP) imaging of TMV infectious clones at 9 dpi. The bright‐field and GFP‐filtered images (lower right) were merged to generate the composite panel. Fluorescence is visible in the inoculated leaf and sporadically in systemic leaves (red arrows). IL, inoculated leaf.

Although CP is classically required for systemic TMV infection, we repeatedly detected GFP in upper systemic leaves following TMVΔCP‐GFP inoculation (Figure 1D; Figure S1A). Across more than 10 independent experiments, systemic GFP was observed in ~20–30% of plants. Two recurrent systemic patterns were observed: Discrete puncta within the lamina and vein‐associated fluorescence concentrated near the midrib (Figure 1D, arrows). RT‐PCR performed on distal GFP‐positive regions detected TMVΔCP sequences, arguing against contamination (Figure S1B). Sequencing of the MP coding region did not reveal consistent mutations across systemic events. Lastly, similar systemic GFP emergence was also observed when the Agrobacterium suspension was diluted to an OD₆₀₀ of 0.001 or when TMVΔCP‐GFP was delivered via plasmid rub inoculation. This indicates that this phenotype is not limited to high‐density agroinfiltration (Figure S1C). Together, these data show that TMVΔCP‐GFP can reproducibly reach distal tissues without an obligate, shared MP adaptation, while systemic signal remains weak and variable among plants.

TMVΔCP‐GFP rapidly traces vascular bundle and moves into roots

To distinguish local cell‐to‐cell expansion from vascular‐associated long‐distance movement, we used low‐density agroinfiltration to initiate infection from limited foci. The GFP expansion pattern between 7 and 8 dpi revealed two distinct movement behaviors: (i) slow isotropic enlargement from individual infection foci, consistent with cell‐to‐cell movement through plasmodesmata, and (ii) much faster signal extension along veins and into the petiole (Figure 2A). This difference in expansion rate suggested that the vein‐associated signal reflected vascular transport rather than simple local spread. We then examined transverse sections of the petiole. The GFP signal was enriched in the abaxial phloem‐associated region. Upstream–downstream comparisons along the petiole axis showed that the GFP signal first appeared in the phloem and then appeared in adjacent tissues. This is consistent with phloem‐associated transport followed by release into surrounding cells (Figure 2B). To test whether TMVΔCP‐GFP entered the phloem‐associated transport stream, we collected EDTA‐facilitated phloem exudates from infected plants. Since tobacco yields only small quantities of phloem sap, we assayed the viral RNA by nested RT‐PCR. TMVΔCP‐GFP RNA was detected in exudate samples, along with the phloem‐associated reference transcript PP2 (Figure 2C). This result supports phloem‐associated movement in the absence of CP. Notably, viral RNA was detectable in distal vascular samples before visible GFP accumulation, suggesting that vascular transport can precede detectable reporter expression, unloading, and local amplification. With longer incubation to 18 dpi, a discernable GFP signal appeared in the roots, though a comparable signal was not observed in the upper shoots (Figure 2D). Immunoblot analysis further supported GFP accumulation in roots (Figure 2E). Together, these observations document a downward (basipetal) long‐distance transport pattern through the abaxial phloem.

Figure 2.

CP‐deficient TMV accesses the vasculature and is transported efficiently to root tissues.

(A) Sequential GFP images show the rapid expansion along the veins. Red arrows indicate the initial infection focus. White arrows indicate the petiole positions for transverse sections.

(B) A petiole cross‐section at 8 dpi indicates phloem loading of the CP‐deletion clone. The vascular bundle is in the center, with xylem separating the adaxial and abaxial phloem. Chlorophyll autofluorescence is shown in red. AbP, abaxial phloem; AdP, adaxial phloem; X, xylem. Scale bar, 500 μm.

(C) Viral RNA detection in EDTA‐facilitated phloem exudates by nested RT‐PCR using primers targeting a conserved region in TMV ORF1. Primer sequences are listed in Table S1. PP2 served as a phloem‐associated marker confirming phloem exudate recovery. Exudate from the inoculated leaf serves as a positive control (PC).

(D) GFP imaging at 18 dpi highlights the robust downward flow of TMVΔCP‐GFP. The signal intensity is pseudo‐colored to distinguish root autofluorescence from GFP. Leaves were trimmed for visualization. IL, inoculated leaf.

(E) Immunoblot detection of GFP in inoculated leaves and roots supports true reporter accumulation in distal tissues. Ponceau S staining (Rubisco large subunit) is shown as a loading/transfer reference.

TMVΔCP‐GFP shows limited access to adaxial/internal phloem domains in transverse sections

To determine how CP deficiency alters long‐distance transport, we examined transverse sections along the movement pathway (Figure 3). To aid anatomical interpretation, we included a TBO‐stained reference section as an annotation channel independent of the infected samples. This reference was used to visualize vascular‐bundle organization and the relative positions of xylem and phloem domains. In petioles from inoculated leaves, both TMV‐GFP and TMVΔCP‐GFP were detected near the abaxial phloem region, consistent with vascular entry from the abaxial side (Figure 3A). However, whereas TMV‐GFP signal extended across the vascular bundle toward the adaxial phloem region, TMVΔCP‐GFP showed reduced continuity through the bundle and instead accumulated more prominently around the vascular bundle and in adjacent tissues.

Figure 3.

The access to the acropetal transport route is hindered for TMVΔCP.

Toluidine blue O (TBO) staining visualizes cell wall‐rich vascular anatomy, including xylem vessels and the relative positions of bicollateral phloem domains. Confocal imaging of transverse sections sampled along the long‐distance movement path. Chlorophyll autofluorescence is shown in red.

(A) Petiole sections from inoculated leaves at 7 dpi show vascular‐domain GFP distribution in TMV‐GFP‐ and TMVΔCP‐GFP‐infected plants.

(B) Stem sections below the inoculated leaf at 14 dpi show contrasting patterns of internal phloem‐associated signal between TMVΔCP‐GFP and TMV‐GFP.

(C) Microscopic imaging of transverse root sections. GFP fluorescence, autofluorescence, and differential interference contrast (DIC) are shown as merged.

(D) Stem transverse sections above the inoculated leaf at 14 dpi (TMVΔCP‐GFP) and at 9 dpi (TMV‐GFP). AbP, abaxial phloem; AdP, adaxial phloem; EP, external phloem; IP, internal phloem; X, xylem. Scale bars, 500 μm, if not stated.

In tissues along the downward movement route, TMVΔCP‐GFP remained readily detectable. In lower stem sections, TMVΔCP‐GFP was mainly associated with the external phloem and surrounding peripheral tissues, while TMV‐GFP was detected in both external and internal phloem‐associated regions (Figure 3B). TMVΔCP‐GFP was also readily detected in roots (Figure 3C). These observations indicate that CP deficiency does not prevent vascular access or basipetal movement toward belowground tissues. By contrast, TMVΔCP‐GFP showed a clear defect along the upward movement route. In upper stem sections, TMV‐GFP exhibited a distinct internal phloem‐associated signal, whereas TMVΔCP‐GFP showed no comparable signal even at a later time point (Figure 3D). These section‐based comparisons suggest that the abaxial and adaxial phloem domains act as distinct interfaces in this system and that TMVΔCP‐GFP inefficiently accesses the adaxial/internal phloem domains associated with upward acropetal movement.

Mechanical disturbance of the upward movement path alters the timing, leaf position, and vascular domain of systemic GFP emergence

Systemic TMV infection is often first observed in roots and young sink tissues, followed by progressively older leaves. However, it is not straightforward to resolve from whole‐plant imaging alone whether this progression reflects truly bidirectional transport or secondary redistribution after basipetal movement. To investigate the route of early acropetal movement, we mechanically damaged either the abaxial or adaxial side of the petiole of inoculated leaves in plants infected with the pTMV‐GFP clone. Then, we tracked the first appearance of GFP in the upper leaves (Figure 4A). In undamaged plants, systemic GFP was typically detected around 6 dpi. Abaxial‐side damage produced a similar timing distribution, whereas adaxial‐side damage yielded a subset of plants in which systemic GFP appearance was delayed by approximately 2–3 days (Figure 4A). Since precisely excising the phloem side is technically challenging in a living petiole, we observed two outcomes: Overly deep excision could prevent systemic GFP detection, and incomplete excision could allow GFP to appear around the cut site, indicating residual connectivity.

Figure 4.

Disrupting the upward transport route shifts WT TMV to a TMVΔCP‐like pattern.

The adaxial or abaxial side of the inoculated‐leaf petiole was mechanically damaged, and systemic outcomes were tracked.

(A) Distribution of plants by the day of first detectable systemic GFP in upper leaves. ND, not detected.

(B) Stem transverse sections (9 dpi) for corresponding damaged routes reveal complementary movement.

(C) Representative images of upper leaves showing the early systemic GFP expression (red arrows). Adaxially damaged plants with delayed emergence cases showed first systemic GFP detection in older systemic leaves (5/5 plants).

(D) Transverse sections from TMVΔCP‐GFP plants with systemic GFP spots, showing upper stems and petioles at 11 dpi. AbP, abaxial phloem; AdP, adaxial phloem; EP, external phloem; IP, internal phloem; X, xylem. Scale bars, 500 μm.

To evaluate the changes in transport patterns after perturbation, we examined transverse sections of stems from damaged plants (Figure 4B). In a subset of abaxially damaged plants, fluorescence in the lower stem sections was associated with the internal phloem. By contrast, in adaxially damaged plants that showed delayed systemic GFP emergence, fluorescence remained apparent in the external phloem of the upper stem sections (Figure 4B). These patterns suggest compensation occurs between phloem domains following successful localized disruption. The appearance of GFP in the upper leaves also differed by treatment (Figure 4C). In undamaged and abaxially damaged plants, the earliest systemic GFP was typically detected in younger, developing apical leaves. By contrast, in the adaxially damaged plants with delayed systemic GFP emergence, the first detectable signal was observed in the older leaves closer to the sink–source transition zone in all five examined cases (Figure 4C). Interestingly, this distribution was similar to the recurrent leaf‐position pattern observed for CP‐deficient TMV systemic GFP events (Figure 1D; Figure S1A). In the upper stem and petiole sections of discrete GFP spots, fluorescence was often aligned with the external phloem and the adjacent domains (Figure 4D). In some samples, the signal was detected across both phloem domains, indicating direct contact with the internal phloem (Figure 4D).

CP deficiency impairs vascular‐bundle transit toward the adaxial phloem

To determine whether the impaired upward movement of TMVΔCP‐GFP was associated with altered vascular‐domain usage, we examined transverse sections of inoculated‐leaf petioles over time (Figure 5A–C). In the early stages, both TMV‐GFP and TMVΔCP‐GFP were predominantly detected near the abaxial phloem region, suggesting that the initial vascular entry occurred mainly from the abaxial side (Figure 5A). However, their subsequent distributions differed markedly. In TMV‐GFP‐infected petioles, the GFP signal remained largely associated with the vascular bundle and progressively extended toward the adaxial phloem region, consistent with efficient continuity across the bundle. By contrast, TMVΔCP‐GFP signal showed weaker continuity through the central vascular‐bundle region. Instead, it expanded more prominently around the bundle and into adjacent parenchyma tissues (Figure 5B,C).

Figure 5.

Petiole transverse sections showing vascular‐domain GFP distribution.

(A–C) Representative petiole transverse sections from TMV‐GFP‐ and TMVΔCP‐GFP‐infected leaves at 6, 8, and 9 dpi. For each sample, the GFP channel alone and a merged image with the autofluorescence channel are shown. White arrows in the GFP‐channel images indicate the central region of the vascular bundle used to assess GFP signal continuity across the bundle. Scale bars, 100 μm.

(D, E) Quantification of GFP signal intensity in the central vascular‐bundle region of petiole sections collected at 8–9 dpi. GFP intensity was normalized to either the autofluorescence level measured in the same region (D) or the total GFP intensity measured across the corresponding petiole section (E). Ten petioles were analyzed for each virus construct in each of 3 independent experiments. Data were used to compare vascular‐bundle signal continuity between TMV‐GFP (WT) and TMVΔCP‐GFP (ΔCP). Statistical comparisons were performed using two‐tailed t‐tests (*P < 0.05, ***P < 0.001).

This pattern resembled the altered systemic movement observed when the adaxial/internal transport route was mechanically perturbed, suggesting that CP deficiency may impair efficient access to the adaxial phloem‐associated route required for upward movement. Quantification of normalized GFP intensity in the central vascular‐bundle region confirmed significantly reduced signal continuity in TMVΔCP‐GFP sections compared with TMV‐GFP sections (Figure 5D,E). Thus, although TMVΔCP‐GFP can enter vascular‐associated tissues, its movement through the bundle toward the adaxial phloem appears inefficient. The stronger accumulation in surrounding parenchyma may reflect leakage or redirection into a less efficient route, rather than productive passage through the acropetal vascular pathway.

TMVΔCP reaches upper vascular tissues but inefficiently invades the lamina

Although TMVΔCP‐GFP exhibited inefficient systemic GFP expression in the upper shoots, several observations suggested that the CP‐deficient viral genomes could still reach the upper tissues. Viral RNA was detected in phloem exudates from both lower stem and upper‐leaf samples (Figure 2C). In addition, transverse sections of inoculated‐leaf petioles showed a viral signal in the adaxial phloem region. This finding suggests that TMVΔCP‐GFP can access vascular domains associated with upward movement (Figure 3A).

To increase the sensitivity of histochemical detection relative to GFP imaging, we replaced GFP with GUS and evaluated distal tissues. In roots, TMVΔCP‐GUS produced broad staining similar to that observed with GFP imaging (Figure 6A). GUS signal was also detected in upper systemic leaves, but staining was largely restricted to vascular regions and aligned with the vein network (Figure 6B,C). In transverse stem sections upstream of the inoculated leaf, the signal was observed in association with the internal phloem (Figure 6D). Interestingly, prolonged incubation did not lead to progressive lamina invasion. Instead, staining was most readily detected around 10 dpi and became difficult to detect at later time points. To quantify this transient systemic presence, we sampled phloem‐enriched tissues (veins and petioles) and normalized viral RNA levels using the phloem marker PP2. TMVΔCP RNA levels peaked around 10 dpi and declined to near the detection limit after approximately 2 weeks (Figure 6E). A subset of plants retained higher levels of viral RNA at later time points, which is consistent with the idea that the virus was able to establish itself locally in a rare and successful manner rather than accumulating stably and systemically (Table S2). Together, these results suggest that although TMVΔCP can reach upper leaves through the phloem, it mostly fails to persist or expand efficiently after arrival.

Figure 6.

TMVΔCP utilizes internal phloem, but inefficient unloading limits lamina colonization.

(A) GUS staining at 14 dpi shows reporter activity in roots.

(B) GUS staining of upper systemic leaves at 10 dpi.

(C) Enlarged view of systemic leaves highlights reporter distribution relative to vein order (2°/3°/4°).

(D) Stem transverse section upstream of the inoculated leaf. EP, external phloem; IP, internal phloem; X, xylem. Scale bars, 500 μm.

(E) RT‐qPCR quantification of viral RNA in upper‐leaf midribs over time shows a declining trend despite detectable transport. PP2 is used as a phloem reference. Individual values (Table S2) are plotted with a median trend and error bars.

Suppressing RNA silencing increases systemic detection but GFP remains vein‐associated

RNA silencing is a major antiviral defense, and many studies have linked the strength of silencing to the efficiency of systemic infection. During TMVΔCP‐GFP infection, the co‐delivery of P19, a strong silencing suppressor from tomato bushy stunt virus, increased the frequency of GFP detection in upper leaves from ~20 to >80% (Figure 7A). In addition, systemic GFP foci expanded more noticeably over time than with TMVΔCP‐GFP alone (Figure 7B). However, trans‐supplementation could increase overall viral accumulation and indirectly affect the systemic outcome (Figure S2A–D). We therefore tested whether spatially separated P19 expression could enhance TMVΔCP‐GFP invasion in systemic tissues. However, no clear enhancement was observed, and GFP signals did not preferentially emerge in the P19‐supplemented regions (Figure S2E–H). Since Agrobacterium‐mediated expression in systemic leaves may not efficiently target the vascular‐associated cells, which are the first point of arrival for TMVΔCP‐GFP, silencing suppressors were instead expressed in cis within the TMVΔCP‐GFP backbone as GFP fusion constructs (Figure 7C). This panel included TRV 16 K, a relatively mild suppressor, as well as CMV 2b and PepMoV P1::HC‐Pro, which are stronger suppressors. Despite its larger size, P1::HC‐Pro was included to capture potential trade‐offs imposed by genome architecture, including insert length and positional constraints near the 3′ end.

Figure 7.

RNA silencing pressure contributes to the systemic bottleneck and can be relieved by VSR supplementation.

(A) Systemic GFP detection frequency in plants with or without P19 trans‐supplementation.

(B) Representative whole‐plant images of P19‐supplemented plants at 7 and 11 dpi show systemic GFP regions over time.

(C) Schematic of TMVΔCP‐GFP‐derived constructs encoding VSR candidates. 1 kb size bar is shown.

(D) Representative TMVΔCP‐2b::GFP infection at 12 dpi shows upper‐leaf symptoms with GFP largely confined to vascular territories. IL, inoculated leaf.

(E, F) Viral RNA accumulation in inoculated leaves and upper systemic leaves measured by RT‐qPCR. “Free” denotes TMVΔCP‐GFP without suppressor insertion. Systemic‐leaf titers show high variability and are presented with individual values overlaid. Statistical significance was assessed by one‐way ANOVA followed by Tukey's HSD test. Different letters indicate significant differences between groups (P < 0.05).

All VSR‐containing TMVΔCP constructs produced readily detectable GFP in inoculated leaves (Figure S3). However, viral RNA levels in inoculated leaves were not consistently higher than those of the “free” TMVΔCP‐GFP control (Figure 7E). In systemic leaves, however, these constructs induced visible symptoms (Figure 7D). The symptoms were vein‐centered, with vein‐associated contraction and downward curling, while the lamina remained comparatively intact. Quantification of viral RNA in upper systemic leaves showed that TMVΔCP constructs containing VSR accumulated at significantly higher levels than the VSR‐free control, though these levels were variable (Figure 7F). Yet, GFP fluorescence remained weak and was largely confined to vein‐associated regions even after symptom development (Figure 7D). Notably, the TMVΔCP‐P1::HC‐Pro::GFP construct exhibited reduced accumulation in inoculated leaves, yet comparatively robust accumulation and symptom development in systemic leaves (Figure 7E,F). Together with the P19 trans‐supplementation results, these patterns suggest that increased systemic titers indicate active infection in distal tissues rather than passive carryover. Meanwhile, vein‐limited reporter patterns persist despite enhanced silencing suppression.

Co‐infection with CMV restores systemic GFP detection in developing leaves

Phloem‐limited infection can be alleviated by viral co‐infection in certain systems. To test this theory with TMVΔCP‐GFP, co‐inoculations were performed with a panel of N. benthamiana‐infecting viruses, including TRV, PepMoV, CMV, TSWV, TYLCV, and PVY. While some combinations of helper viruses occasionally increased systemic GFP emergence, most produced sparse GFP spots in older leaves, resembling the pattern of TMVΔCP‐GFP alone. However, CMV consistently induced numerous GFP foci in young upper leaves and restored a qualitatively distinct systemic pattern in most examined plants (Figure 8A and Figures S4 and S5). The CMV‐associated recovery persisted when CMV and TMVΔCP‐GFP were inoculated onto different leaves (Figure S5D), which indicates that TMVΔCP‐GFP can reach the upper leaves independently and that the CMV effect does not require colocalization at the primary inoculation site. Systemic GFP recovery was also observed when CMV was co‐inoculated with VSR‐fused TMVΔCP derivatives (Figure S5E–G).

Figure 8.

Mixed infection overcomes unloading bottleneck without restoring CP.

Co‐infection with CMV enables TMVΔCP‐GFP to expand beyond vein‐associated domains into the lamina of upper systemic leaves. Images were collected at 10 dpi. (A) Overlay image was pseudo‐colored to improve visualization of weak lamina fluorescence.

(B) White light image.

(C) Green fluorescent filtered image.

DISCUSSION

The prevailing notion in TMV biology is that CP is essential for long‐distance movement. This notion is based on classic CP/assembly‐origin mutant analyses and host‐dependent systemic movement phenotypes (Hilf & Dawson, 1993; Saito et al., 1990). However, early studies of defective TMV isolates and CP‐compromised variants reported that systemic spread can occasionally occur, albeit inefficiently and inconsistently (Dawson et al., 1988; Siegel et al., 1962). More recently, several studies have demonstrated that systemic movement can be restored in CP‐defective TMV backgrounds via functional compensation (Tran et al., 2022; Venturuzzi et al., 2021). We reinterpret this long‐standing controversy by decomposing “systemic infection” into three steps: phloem loading, vascular translocation, and phloem exit. Each step can be hindered independently. Under this framework, CP deficiency may reduce the probability of productive systemic invasion; however, our data indicate that long‐distance vascular dissemination can occur even in the absence of CP (Figure 9).

Figure 9.

Schematic illustration of systemic infection by TMV in Nicotiana benthamiana.

The plant vascular system is depicted as a bicollateral phloem network, with internal and external phloem strands shown as continuous conduits through the stem and into the leaf traces and roots. Viral distribution is indicated by the violet color gradient, with darker shading representing higher local virus accumulation. The image was polished using figurelabs.ai.

Many economically important species in the Solanaceae and Cucurbitaceae families possess bicollateral vascular bundles. In these bundles, the external (abaxial) and internal (adaxial) phloem are separated by xylem (Slewinski et al., 2013). In these plants, the two phloem domains may become functionally specialized. For example, in tomato, export from source leaves uses external phloem preferentially, while import into sink leaves uses the internal phloem preferentially (Slewinski et al., 2013). This anatomy is consistent with the influential paradigm that TMV largely follows photoassimilate source‐to‐sink flow, with export via abaxial/external phloem and import via adaxial/internal phloem (Cheng et al., 2000). However, our localization data indicate that wild‐type TMV can access both phloem domains already in the inoculated leaf (Figure 3A). Phloem‐disruption experiments further suggest that TMV can travel directly from the adaxial phloem of the inoculated leaf rather than routing through belowground tissues (Figure 4). Similar observations in Plantago asiatica mosaic virus (PlAMV) support the notion that vascular entry and subsequent trafficking can involve multiple vascular domains (Novianti et al., 2021). More broadly, evidence that TMV can undergo vascular transport under conditions that block replication is consistent with direct acropetal dissemination (Susi, 1999).

Within this anatomical framework, CP loss did not weaken movement uniformly. Rather, it created a defect that was directional and domain‐specific. TMVΔCP exhibited a significant decrease in the adaxial/internal (acropetal) pathway, while the abaxial/external (basipetal) pathway remained relatively intact. This asymmetry is supported by cross‐sectional localization at vascular interfaces (Figures 3 and 5) and the resulting topography of systemic infection in distal tissues (Figure 2D). A similar internal phloem gate has been described for PepMoV in a virus‐resistant pepper cultivar. In this cultivar, the virus accumulates in the external phloem and moves downward, but it is restricted from the internal phloem. This restricts its ability to invade young tissues (Guerini & Murphy, 1999). Similarly, the induction of SA‐like defenses with acibenzolar‐S‐methyl delays systemic infection by PlAMV; however, detection in belowground tissues is far less affected than detection in aboveground tissues (Novianti et al., 2021). These findings suggest that restricting upward movement alone can be sufficient to confer resistance. Whether the directional/domain asymmetry we observed is merely a byproduct of functional partitioning in bicollateral phloem, or instead reflects an actively enforced selective immune barrier, remains unresolved. Nonetheless, our data suggest that the upward transport defect of the CP‐deficient mutant is associated primarily with inefficient entry into the adaxial/internal phloem route (Figure 5). Future work could address this distinction by combining domain‐resolved virus localization with markers of phloem identity and defense activation such as callose deposition.

If CP plays a broadly conserved role in long‐distance movement despite differences in particle structure and transport form, then genome protection is likely a shared function. RNA silencing is a plausible surveillance pressure during systemic movement, as mobile antiviral silencing signals often first appear around veins in systemic leaves (Mermigka et al., 2016). Without CP‐mediated encapsidation or genome stabilization, TMVΔCP RNA may be more vulnerable during vascular transport. Consistent with this idea, systemic TMVΔCP accumulation declined over time in GUS‐based detection and in the quantification of phloem‐enriched tissues (Figure 6E), suggesting that the persistence of CP‐deficient genomes is challenged in phloem‐associated tissues. This view is also consistent with the notion that phloem transport acts as a surveillance checkpoint (Folimonova & Tilsner, 2018) and with previous studies showing that silencing suppression allows for vascular arrival without efficient escape into surrounding tissues (Cao et al., 2010). VSR experiments further support this interpretation. P19 trans‐supplementation increased systemic TMVΔCP‐GFP detection (Figure 7A). Since P19 binds to siRNA duplexes and inhibits the spread of systemic silencing signals (Silhavy et al., 2002), this effect may reflect not only improved viral RNA stability but also altered timing or distribution of mobile antiviral silencing. Delayed systemic silencing could provide CP‐deficient genomes with a longer window for postarrival establishment. This would allow intermittent GFP‐positive islands to expand before restriction becomes effective. The spatial and temporal relationship between antiviral siRNA movement and TMVΔCP‐GFP emergence remains an important question for future research. The strongest symptom development and viral RNA recovery occurred when VSR activity was supplied in cis and directly linked to the transported viral genome (Figure 7D,F). This effect was most evident in systemic tissues rather than inoculated leaves, suggesting that RNA silencing imposes a stronger constraint during long‐distance movement and/or early systemic establishment.

Phloem exit represents an additional bottleneck. In upper systemic leaves infected with TMVΔCP, we observed a vein/phloem‐associated GUS pattern rather than widespread lamina invasion (Figure 6C). This pattern persisted even when VSR activity increased viral RNA accumulation and symptom severity in systemic tissues (Figure 7D,F). This decoupling—stronger systemic amplification without proportional spread into the lamina—suggests that phloem exit is an additional, independent bottleneck that is separable from both loading and long‐distance translocation. This finding is consistent with earlier imaging studies showing that systemic arrival is followed by selective unloading from specific vein classes, with mesophyll invasion occurring later (Roberts et al., 1997). Thus “exit” is a discrete step rather than an automatic consequence of transport. The specificity of our co‐infection results further suggests that this barrier can be addressed mechanistically. Among the multiple helper viruses tested, only CMV consistently enabled TMVΔCP to escape the vascular‐associated pattern and establish lamina infection (Figure 7; Figure S4). Notably, when TMVΔCP and CMV were inoculated into separate leaves, minimizing local synergism, the timing of the first appearance of systemic GFP was largely unchanged. This indicates that TMVΔCP can reach upper leaves autonomously. It also supports the idea that CMV primarily acts by unlocking the phloem‐exit interface, rather than by accelerating initial long‐distance transport (Figure S5).

Together, these findings provide a coherent explanation for the long‐distance movement pattern of TMVΔCP‐GFP (Figure 9). TMVΔCP moved efficiently toward roots but appeared only intermittently in upper systemic leaves. When upper GFP signals were detected, they were mainly observed in partially expanded leaves rather than in the youngest sink leaves. These leaves are developmentally close to the sink–source transition, although phloem‐domain behavior at this stage is relatively unexplored. Nevertheless, they undergo a developmental shift from import to export as source activity develops (Slewinski et al., 2013). This shift creates a vascular transport context that is distinct from that of fully sink or fully source leaves. Consistent with this interpretation, transverse sections of upper GFP‐positive tissues showed external phloem‐associated signals. A similar leaf‐position pattern was reproduced when the internal/acropetal route of WT TMV was physically disrupted. Therefore, the external‐domain‐associated systemic GFP events observed here do not necessarily contradict the predominant basipetal role of the external stream. Rather, they suggest that this domain is less restrictive to CP‐deficient TMV than the internal/acropetal pathway. Nevertheless, visible GFP emergence was less frequent than the phloem‐associated detection in upper tissues (Figure 2C), indicating that vascular arrival does not always lead to productive infection. Thus, phloem exit and postphloem establishment represent additional barriers downstream of long‐distance transport.

A central question, then, is how viruses move without CP. One plausible explanation lies in the evolutionary and structural features of the 30K MP superfamily. These MPs contain a conserved single jelly‐roll (SJR) domain, a fold also found in many icosahedral capsid proteins (Mushegian & Elena, 2015). Recent analyses based on structure prediction suggest that some MPs may have originated from ancestral CPs that subsequently became functional to support cell‐to‐cell movement and were later adopted by diverse virus lineages (Butkovic et al., 2023). Notably, most well‐characterized cases of CP‐independent systemic movement encode 30K‐SJR MPs (Jung & Kim, 2026). It is also intriguing that CMV, which rescued TMVΔCP unloading in our system, encodes both a 30K‐type MP and an SJR‐type CP. This raises the possibility that CMV infection promotes a trafficking‐permissive cellular state compatible with RNP‐based movement. Consistent with this view, the systemic movement of TMV in the absence of CP can be restored through MP modulation (Tran et al., 2022) or by replacing it with another 30K‐type MP (Lewandowski & Adkins, 2005).

RNA signals may further modulate CP‐independent movement. In plants, endogenous mobile mRNAs often require tRNA‐like structures (TLSs) or related motifs to gain phloem mobility (Thieme et al., 2015). Some viruses, including TMV, that are reported to move systemically without CP carry TLS in their 3′ UTRs, whereas others lack obvious TLS motifs or even use fundamentally different genome architectures, such as the circular single‐stranded DNA genomes of begomoviruses (Jung & Kim, 2026). Thus, TLSs are not universally required, though they may enhance systemic movement in specific contexts. For example, restoring an engineered TLS enabled systemic infection by a CP‐defective TCV (Li et al., 2009). Whether the TMV 3′ UTR similarly contributes to the transport competence of CP‐deficient TMV in our system remains an open, testable possibility.

MATERIALS AND METHODS

Plant material and growth conditions

N. benthamiana plants were grown in a controlled growth chamber under a 16 h light/8 h dark photoperiod, with day/night temperatures of 24–25°C/22°C and 60% relative humidity. Plants were used at 4–5 weeks after germination.

Plasmid construction

The TMV ΔCP vector pTMVΔCP‐GFP was derived from pJL‐TRBO‐G (Addgene #80083). pTMVΔMPΔCP‐GFP was generated by whole‐plasmid polymerase chain reaction (PCR) using primer pairs flanking the intended deletion with ~20‐bp overlaps. PCR products were treated with DpnI and transformed into Escherichia coli for in vivo circularization. Full‐length TMV infectious clones (pTMV‐GFP or pTMV‐GUS) were produced by inserting the TMV CP ORF downstream of the heterologous reporter, that is, green fluorescent protein (GFP) or beta‐glucuronidase (GUS). Because the native TMV CP subgenomic promoter (sgp) is occupied by the reporter, a heterologous CP sgp from tomato mild green mosaic virus (TMGMV) was duplicated immediately upstream of CP. The TMGMV‐sgp‐CP cassette was synthesized de novo (Twist Bioscience).

For viral suppressor of RNA silencing (VSR) fusions, VSR::GFP cassettes were cloned into pTMVΔCP‐GFP after PacI/NotI digestion. VSRs including tobacco rattle virus (TRV) 16 K, cucumber mosaic virus (CMV) 2b, and pepper mottle virus (PepMoV) P1::HC‐Pro were fused to GFP by overlap PCR using fragments with 20–25 bp overlaps and a GSGSG linker. PCR amplification used Phanta Max DNA Polymerase (Vazyme), assemblies used ClonExpress Ultra One‐Step (Vazyme), and restriction digests used FastDigest enzymes (Thermo Fisher Scientific). All constructs were verified by Sanger sequencing across assembly junctions, reporter ORFs, and engineered promoter regions. Primer sequences are listed in Table S1.

Agrobacterium culture and plant inoculation

A. tumefaciens GV3101 was transformed with each binary vector and selected on yeast extract peptone (YEP) agar containing rifampicin and kanamycin. Single colonies were grown overnight in Luria‐Bertani (LB) medium at 28°C with appropriate antibiotics. Cells were pelleted at 4000× g for 5 min and resuspended in infiltration buffer (MMA: 10 mM MgCl₂, 10 mM MES pH 5.6, 100 μM acetosyringone). Suspensions were incubated at room temperature for 2 h and adjusted to OD₆₀₀ = 0.2. Leaves were infiltrated from the abaxial side using a needleless syringe. For co‐infection assays, helper viruses were delivered using infectious cDNA clones and co‐inoculated onto the same leaf together with TMV constructs. Helper viruses used were CMV, TRV, PepMoV, tomato yellow leaf curl virus (TYLCV), tomato spotted wilt virus (TSWV), and potato virus Y (PVY) (clone details specified in figure legends). When delivered by agroinfiltration, Agrobacterium cultures were mixed at a 1:1 ratio by OD immediately before infiltration (final OD₆₀₀ maintained at 0.2). For low‐density agroinfiltration, the Agrobacterium suspensions were diluted to an OD₆₀₀ of 0.001 prior to infiltration. For plasmid DNA rub inoculation, 10 μg of purified plasmid DNA containing the TMV infectious clone were diluted in a 50 mM potassium phosphate buffer (pH 7.4) and applied mechanically to the adaxial leaf surface that had been lightly dusted with carborundum. Leaves were then gently rubbed to produce minor abrasion to the surface, rinsed with water after inoculation, and monitored for GFP fluorescence.

Plant imaging

GFP was imaged in a custom light‐tight chamber equipped with blue excitation (peak ~470 nm) and a 500‐nm long‐pass emission filter. Bright‐field and GFP images were acquired sequentially and merged. Images were flat‐field corrected, and raw intensities were mapped to a pseudo‐color gradient. Acquisition parameters were kept constant within each experiment. Confocal imaging was performed on a Leica SP8 X microscope. GFP was excited at 488 nm and emission collected at 500–550 nm. Chlorophyll autofluorescence was collected at 650–750 nm. For vascular localization, transverse sections of midrib and petiole were prepared manually using a razor blade, mounted in phosphate‐buffered saline/glycerol (9:1), and imaged immediately. Adaxial and abaxial phloem were distinguished using anatomical landmarks. Systemic infection was scored at approximately 10 days postinoculation (dpi) based on whole‐plant GFP imaging. A plant was counted as systemically positive if any discrete GFP spot was detected in upper, non‐inoculated leaves.

Sampling and RNA extraction

Leaf discs (30–50 mg), whole‐leaf tissue, or vein‐enriched tissue were harvested, flash‐frozen, and ground in liquid nitrogen. For vein‐enriched samples, interveinal lamina tissue was removed and the remaining midrib and primary veins were collected. Total RNA was extracted using TRIzol‐based reagent (RNAiso, Takara) and assessed by agarose gel electrophoresis and A₂₆₀/A₂₈₀ ratios. For phloem exudate extraction, at 10 dpi, petioles or stems were recut under 20 mM EDTA (pH 7.0) and immediately submerged in fresh EDTA solution. Exudation proceeded for 1 h at room temperature. Plants were irrigated 1 day prior to harvest. RNA was extracted from exudates using TRIzol‐based reagent. Due to low exudation in N. benthamiana, viral RNA and reference transcripts were detected using nested reverse transcription (RT)‐PCR.

PCR detection and statistics

One‐step RT‐PCR (tissue RNA) or two‐step RT‐PCR (phloem exudate RNA) was performed using gene‐specific primers (Table S1). For nested assays, 1–2 μl of the first‐round product was used as template for the second round. Typical cycling conditions were: 50°C for 30 min (RT), 95°C for 2 min; 35 cycles of 95°C for 15 sec, 58–60°C for 20 sec, and 72°C for 20–30 sec; followed by 72°C for 2 min. For quantitative RT‐PCR (RT‐qPCR), cDNA was synthesized from 0.5 to 1.0 μg total RNA using oligo(dT)₁₆ or virus‐specific primers. qPCR was performed with SYBR Green master mix (BioFact) in 10–20 μl reactions on a QuantStudio instrument (Applied Biosystems). Each biological replicate was measured in three technical replicates. Relative accumulation was calculated using the 2^−ΔΔCt method. For vein/phloem‐focused measurements, Ct values were normalized to PP2. For whole‐leaf assays, EF1α and Actin were used as reference genes. Experiments were performed with a minimum of three biological replicates. Statistical significance was assessed using two‐tailed Student's t‐tests for pairwise comparisons and one‐way ANOVA followed by the Tukey's HSD test for multiple comparisons, as indicated in the figure legends.

Protein analysis and histochemistry

Proteins were extracted from ~100 mg tissue using GTEN buffer (10% glycerol, 100 mM Tris–HCl pH 7.5, 1 mM EDTA, 150 mM NaCl) (Bozkurt, 2012), resolved by 10% SDS‐PAGE, transferred to PVDF (polyvinylidene difluoride), and probed with anti‐GFP (1:10 000; MA5‐15256, Invitrogen) followed by HRP‐conjugated secondary antibody. Signal was detected by chemiluminescence, and Rubisco large subunit was used as a loading/transfer control. For GUS staining, tissues were vacuum‐infiltrated with X‐Gluc staining buffer (100 mM sodium phosphate pH 7.0, 10 mM EDTA, 0.1% Triton X‐100, 0.5 mM K₃[Fe(CN)₆], 0.5 mM K₄[Fe(CN)₆], 0.5–1.0 mg mL⁻¹ X‐Gluc) at −70 to −90 kPa for 3–5 min, incubated at 37°C for 18 h, and cleared through an ethanol series (30–70%) (modified from Dedow et al., 2022). For histological sections, stained tissues were fixed in 4% paraformaldehyde for 30–60 min, dehydrated, and sectioned at 10 μm thickness.

For toluidine blue O (TBO) staining, transverse sections were immersed in 0.05% TBO solution for 30–60 sec, rinsed briefly in distilled water, mounted in water or 50% glycerol, and imaged immediately. TBO staining was used as an anatomical reference to visualize vascular organization, including xylem vessels, bundle structure, and the relative positions of external/internal phloem domains.

Petiole phloem‐disruption assay

Leaves were agroinfiltrated, and the petiole was partially incised immediately after inoculation at 0 dpi. Using a heated blade, a partial transverse cut was made from either the adaxial or abaxial side of the middle region of the petiole, being careful not to completely sever it. Since precisely localizing phloem domains in intact, living petioles is technically challenging, complete disruption of the targeted vascular side was prioritized over strict anatomical specificity. Thus, the cut was made deep enough to reliably disrupt the targeted side, even if it caused minor damage to the opposite side. We interpreted the assay based on the timing and position of systemic GFP emergence across biological replicates rather than on direct anatomical verification of phloem disruption in each petiole.

AUTHOR CONTRIBUTIONS

MJ and K‐HK designed research; MJ performed the research; MJ and K‐HK analyzed data. MJ and K‐HK wrote the manuscript with contributions from all authors. All authors read and approved the final manuscript.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

Supporting information

Figure S1. Additional systemic green fluorescent protein (GFP) events and validation of TMVΔCP identity. (A) Representative whole‐plant images showing systemic GFP emergence after TMVΔCP‐GFP inoculation. Each plant is from an independent biological replicate (#1~#4). (B) RT‐PCR detection of viral RNA from the upper‐leaf GFP‐positive foci. The amplicon spans from upstream of the GFP insertion site to the 3′ UTR. Sample #0 corresponds to a plant without detectable systemic GFP. Expected amplicon sizes for constructs with or without the subgenomic promoter–CP region are indicated. (C) Whole‐plant GFP imaging of TMVΔCP‐GFP infection under alternative inoculation conditions. Systemic GFP emergence was observed after low‐density agroinfiltration (OD₆₀₀ = 0.001; left panel) and after plasmid rub inoculation (right panel). IL—inoculated leaf.

Figure S2. Effects of P19 trans‐supplementation on TMV‐GFP and TMVΔCP‐GFP infections. Representative whole‐plant GFP images comparing same leaf and spatially separated P19 supplementation at 12 dpi. (A) TMV‐GFP alone. (B) TMV‐GFP co‐infiltrated with P19. (C) TMVΔCP‐GFP alone. (D) TMVΔCP‐GFP co‐infiltrated with P19 on the same inoculated leaf. (E) TMVΔCP‐GFP inoculated on one leaf, with P19 infiltrated on a different leaf. (F–H) TMVΔCP‐GFP plants with P19 infiltrated into upper systemic leaves at 4 dpi (F), 6 dpi (G), or 8 dpi (H). Images are representative of the observed patterns under comparable GFP imaging conditions. Arrows indicate P19 supplemented leaves.

Figure S3. Disease development of TMVΔCP clones with VSRs. Representative images of plants inoculated with VSR‐containing TMVΔCP derivatives. Inoculated leaves were imaged at 3 dpi (A) and upper systemic leaves at 12 dpi (B). (C) RT‐PCR amplifying the heterologous site. Samples were harvested at 9 dpi when the first symptomatic leaves were shown. IL—inoculated leaf.

Figure S4. Virus co‐infection assays. Plants were monitored until 18 dpi, and representative images were taken at time points corresponding to the development of symptoms for each virus. Strong necrosis or chlorosis were noted as potential sources of autofluorescence. (A) Healthy control at 9 dpi. (B) TMVΔCP‐GFP alone at 9 dpi. (C) TMVΔCP‐GFP + pepper mottle virus (PepMoV) at 9 dpi. (D) TMVΔCP‐GFP + cucumber mosaic virus (CMV) at 9 dpi. (E) TMVΔCP‐GFP + tomato spotted wilt virus (TSWV) at 9 dpi. (F) TMVΔCP‐GFP + potato virus Y (PVY) at 12 dpi. (G) TMVΔCP‐GFP + tobacco rattle virus (TRV) at 12 dpi. (H) TMVΔCP‐GFP + tomato yellow leaf curl virus (TYLCV) at 12 dpi.

Figure S5. CMV co‐infection with TMV constructs. Representative images at 9 dpi. TMVΔCP constructs were co‐inoculated on the same leaf as CMV except for (D), which TMVΔCP‐GFP was inoculated on a lower/older leaf to minimize direct synergism at the inoculation site. (A) CMV single infection. (B) CMV + TMV‐GFP. IL—inoculated leaf. (C) CMV + TMVΔCP‐GFP. (D) CMV (upper leaf) + TMVΔCP‐GFP (lower leaf). (E) CMV + TMVΔCP‐16 K::GFP. (F) CMV + TMVΔCP‐2b::GFP. (G) CMV + TMVΔCP‐P1::HC‐Pro::GFP.

Table S1. Primer sets used for cloning and for viral/host detection.

Table S2. Quantitative reverse transcription polymerase chain reaction (RT‐qPCR) dataset for upper systemic TMVΔCP‐GFP accumulation. Individual RT‐qPCR values underlying Figure 6E are listed. Upper systemic vein‐ and petiole‐enriched tissues were sampled over time, and TMVΔCP‐GFP RNA levels were normalized to the phloem‐associated reference transcript PP2. Samples deviating by >1 log from cohort values (presumed leakage into mesophyll) are flagged in the table and were excluded from the plotted summary.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available within the article and Supplementary Material.