Mixed viral infections uncover multifaceted RNA silencing strategies in plant antiviral defense

Abstract

Background

Alphanecrovirus tessellati and Alphanecrovirus oleae are closely related alphanecroviruses, sharing over 74% genomic identity and frequently co-infecting plants under natural conditions. Despite their frequent coexistence, the molecular consequences of single versus mixed infections, particularly concerning host RNA silencing responses, remain poorly understood. Here, we investigated the antiviral defense responses of Nicotiana benthamiana to single and mixed infections with both viruses, focusing on symptom development, viral accumulation, Dicer-like (DCL) gene expression, and virus-derived small interfering RNA (vsiRNA) profiles.

Results

Single infections resulted in contrasting disease phenotypes: A. oleae induced severe systemic necrosis and high viral accumulation, whereas A. tessellati caused transient infections with mild symptoms. In mixed infections, A. tessellati accumulated to significantly higher levels and was associated with enhanced disease severity, indicating that the interaction between the two viruses altered infection dynamics and favored A. tessellati accumulation. In contrast, A. oleae showed reduced relative genomic RNA accumulation in mixed infections despite the persistence of disease symptoms. Gene expression analyses showed consistent upregulation of DCL4 and DCL2 during mixed infection, supporting their involvement in antiviral RNA silencing under co-infection conditions. Small RNA profiling revealed infection-dependent differences in vsiRNA accumulation patterns. Notably, mixed infection was associated with a change in the predominant vsiRNA size class for A. oleae, alongside a shift in vsiRNAs genomic targeting from the RdRp gene during single infection to the suppressor-associated p6 gene during mixed infection, indicating a reprogramming of host silencing responses.

Conclusions

Our results demonstrate that closely related viruses can differentially modulate host RNA silencing responses depending on the infection context. These findings highlight the plasticity of antiviral RNA silencing during viral co-infections and provide new insights into how mixed infections reshape host-virus interactions at the molecular level.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12864-026-13141-8.

Background

Alphanecrovirus oleae and Alphanecrovirus tessellati, formerly called Olive latent virus 1 (OLV-1) and Olive mild mosaic virus (OMMV), respectively, belong to the genus Alphanecrovirus, family Tombusviridae. A. tessellati and A. oleae present a positive-sense single-stranded RNA (∼3.7 kb) encoding five proteins: p23 (a pre-readthrough protein); p82 (the RNA-dependent RNA polymerase, RdRp); p8 and p6 (small proteins); and the coat protein (CP); p29 in A. tessellati and p30 in A. oleae [1, 2]. Most of the protein functions have been characterized for both viruses: p23 and RdRp are involved in viral RNA replication; p8 and p6 are predicted to participate in virus movement, and CP is involved in capsid assembly, systemic movement, and vector transmission [3–6]. In A. tessellati, p6 and CP have been reported to function as silencing suppressors [2]. No viral suppressors have been described for A. oleae. The highest sequence similarity between A. tessellati and A. oleae lies in the p6, p8, and RdRp regions, 100%, 87.6% and 92.1%, respectively [7]. A. tessellati and A. oleae were first identified in olive trees (Olea europaea L.) [4, 8] where they frequently appear in mixed infections [9, 10]. In fact, the origin of A. tessellati has been suggested as a result from recombination events involving A. oleae and the Tombusvirus Betanecrovirus nicotianae (former TNV-D) during co-infection [9, 11].

Viral mixed infections can significantly change viral population dynamics, host immune responses and the course of disease progression in plants and are widespread in both agricultural and wild systems, although in agricultural systems the mixed infections can significantly decrease the crop yields [12, 13]. Depending on virus–virus and virus–host interactions, outcomes may be synergistic (e.g., enhanced symptoms or viral load), antagonistic (mutual suppression), or neutral [14, 15]. Synergy may arise through molecular complementation or one virus benefiting from another's VSRs, whereas antagonism may result from resource competition, spatial exclusion, or activation of divergent silencing pathways [16].

In olive, A. oleae and A. tessellati typically cause mild or no apparent symptoms; however, mixed infections have been associated with an intensification of symptoms and an increase in transmission efficiency [9, 10, 17, 18]. A. oleae is easily acquired by roots of N. benthamiana, whereas A. tessellati soil transmission is less efficient, and in nature, it must rely on its biological vectors Olpidium brassicae and Olpidium virulentus [6, 18]. In the reported study, A. tessellati was able to invade plant roots, in the presence of A. oleae and in the absence of a biological vector, suggesting a facilitative interaction [17, 18].

In addition, recent HTS data indicated that the simultaneous presence of multiple viruses (e.g., Olive leaf yellowing-associated virus (OLYaV) and Olive leaf mottling virus (OLMV)) may be associated with complex symptoms such as mosaic, chlorosis, and defoliative shedding [19]. These Alphanecroviruses have also been detected in non-olive hosts; A. oleae has been detected in citrus trees [20] and tulips [21], the latter where A. tessellati has also been identified as part of the viral complex associated with the Augusta disease, previously ascribed to TNV [22]. As for economically important horticultural crops, A. oleae has been reported in tomato [23, 24] and A. tessellati in spinach [25] and corn salad [26]. Both viruses are mechanically transmitted to leaves of N. benthamiana, where they induce chlorotic and necrotic lesions, often resulting in plant death [17, 20]. Although natural co-infection of these Alphanecrovirus species in horticultural crops has not yet been reported, it is anticipated, given their biological similarities and overlapping host potential.

From the host side, a key component of the plant immune system against RNA viruses is post-transcriptional gene silencing (PTGS), also referred to as RNA silencing [27]. This defense mechanism relies on the recognition and processing of viral double-stranded RNA (dsRNA) intermediates by Dicer-like (DCL) proteins [28]. The DCLs -RNase III enzymes process double-stranded viral RNA into small interfering RNAs (siRNAs), referred to as virus-derived small interfering RNAs (vsiRNAs), which are then incorporated into Argonaute (AGO) proteins, forming RNA-induced silencing complexes (RISCs) that degrade homologous guided by the vsiRNA [28, 29]. The DCL family in plants typically includes DCL1–DCL4, each with distinct but sometimes overlapping roles. DCL4 is considered the most potent class for antiviral defense and primarily generates 21-nt siRNAs, while DCL2 produces 22-nt siRNAs, especially when DCL4 is absent or overwhelmed [30, 31]. DCL3 is mainly associated with 24-nt siRNAs involved in transcriptional gene silencing via DNA methylation, but it can be engaged during viral infections as well [32]. DCL1 primarily processes microRNAs (miRNAs) and has a minor role in antiviral defense.

To overcome plant defense mechanisms and ensure successful replication and disease development, viruses have evolved several strategies. Among these, viral suppressors of RNA silencing (VSRs) represent one of the most efficient mechanisms and involve virus-encoded proteins that inhibit the plant RNA silencing pathway [33, 34]. The suppressor proteins interfere in different stages of the RNA silencing pathways, being highly divergent and appearing to have evolved independently in the different viruses. These suppressors bind dsRNAs inhibiting their processing into siRNAs; in addition, they are capable of sequestering vsiRNAs, further inhibiting their incorporation into RISC or directly interfering with the recognition of viral RNA, dicing, and RISC assembly [35]. In mixed infections, interactions between different VSRs and their combined effects on the host silencing system can result in distinct disease outcomes, ranging from enhanced resistance to systemic viral synergy, depending on the specific host-virus combination and environmental conditions [36]. Studies in N. benthamiana show that most RNA viruses produce abundant 21-nt or mixed 21/22-nt vsiRNAs [37, 38]. Exceptions include Tomato spotted wilt virus (TSWV) and Rice stripe virus (RSV), which show notable 24-nt dominance [39, 40]. A few reports describe increased 22-nt vsiRNAs or overall abundance during mixed infections involving satellite RNAs [41–44], however, to date, no study has reported a qualitative shift in vsiRNA genomic targeting between single and mixed infections for RNA viruses, namely a change in the predominant targeted viral genes or the emergence of new vsiRNA hotspots within the same gene.

Many synergistic interactions have been recognized to contribute to increased pathogenicity and breakdown of host resistance [14, 45–48]. Yet, the molecular underpinnings of this synergism, especially regarding silencing suppression and vsiRNA biogenesis, remain unexplored. In this study, we hypothesize that N. benthamiana mounts distinct antiviral responses against A. tessellati and A. oleae during single and mixed infections. We evaluated viral accumulation and symptom progression across time points, examined DCL gene expression, and profiled the biogenesis of vsiRNAs via small RNA deep sequencing. Our goal was to determine whether viral mixed-infection reshapes host RNA silencing responses, alters vsiRNA profiles, and redirects targeting patterns across viral genomes.

Results

Mixed-infection increases symptom severity in N. benthamiana

Disease progression in N. benthamiana was assessed using a 0–4 Disease Severity Index (DSI) scale (Fig. 1a, b). Mock-inoculated plants remained asymptomatic throughout the experiment. All virus-inoculated plants developed visible symptoms between 1 and 2 dpi, with clear differences in symptom progression depending on the infection type. Single infection with A. tessellati resulted in mild and transient symptoms, characterized by local chlorotic mosaic lesions appearing at 2 dpi (DSI: 4%). Disease severity peaked at 5 dpi (7%) and subsequently regressed, with full recovery observed by 12 dpi. In contrast, A. oleae single infection caused rapid disease progression, with symptoms appearing at 1 dpi and evolving from local chlorotic lesions to severe necrosis. Systemic symptoms became evident by 5dpi, culminating in complete disease severity by 12 dpi and substantial plant mortality thereafter. Mixed infection accelerated symptom development and increased disease severity. Chlorotic lesions appeared at 1 dpi with higher initial DSI values than in single infections, followed by a rapid progression to maximal disease severity by 8 dpi. Mixed infections were associated with higher mortality rates compared to single infections. These results demonstrate that mixed infection both accelerates disease development and exacerbates symptom severity, leading to earlier plant death.

Fig. 1.

Visual observation of leaves from N. benthamiana plants. a single infected with A. tessellati, A. oleae and A. tessellati and A. oleae mixed-infected at 1 dpi, 5 dpi, and 12 dpi. Control plants were inoculated with sodium phosphate buffer without the viruses presence. b Disease Severity Index at 0 (beginning of the assay), 12 h and 1 to 12 days after single and mixed-infection of A. tessellati and A. oleae in N. benthamiana plants

Enhanced accumulation of A. tessellati and reduced relative abundance of A. oleae in mixed infections

Relative viral genomic RNA accumulation was quantified by qRT-PCR. In wild type plants, both viruses reached peak accumulation at 5 dpi in single and mixed infections. In single infections, A. tessellati showed transient accumulation peaking at 5 dpi, whereas A. oleae maintained high accumulation at both 5 and 12 dpi. In mixed infections, A. tessellati accumulated at significantly higher levels than in single infection at all evaluated time points, although its accumulation within mixed infection peaked at 5 dpi and subsequently decreased at 12 dpi. Conversely, A. oleae showed consistently lower relative accumulation in mixed infections compared to single infections, despite sustained disease symptoms. At 5 dpi, both viruses accumulated to comparable levels during mixed infection. Together, these data indicate that mixed infection altered the relative accumulation patterns of both viruses, favoring the accumulation of A. tessellati, whereas A. oleae showed comparatively lower accumulation levels under mixed-infection conditions. Despite this, the presence of both viruses was associated with accelerated symptom development and increased disease severity in co-infected plants (Fig. 2; Table 1).

Fig. 2.

Mean relative abundance ± standard error (SE) of single A. tessellati-infection, single A. oleae-infection A. tessellati and A. oleae mixed-infections at each time points (1dpi, 5dpi and 12 dpi) in Wild-type N. benthamiana plants and dcl2 and dcl4-suppressed plants. Additionally, detailed results from complementary PERMANOVA statistical analyses are presented in Table 1

Table 1.

Pairwise PERMANOVA comparisons of viral relative abundance in wild-type N. benthamiana plants

a. Temporal comparisons within treatments (p-values)
	1 dpi vs 5 dpi	1 dpi vs 12 dpi	5 dpi vs 12 dpi
N. benthamiana Wild-type	p values
single-A. tessellati	0.0001	0.0001	0.0001
single-A. oleae	0.0002	0.0001	0.1017
mixed-A. tessellati	0.0051	0.0013	0.0001
mixed-A. oleae	0.0005	0.0001	0.0001

b. Comparisons among treatments at each time point (p-values)
	1 dpi	5 dpi	12 dpi
N. benthamiana Wild-type	p values
single-A. tessellati vs mixed-A. tessellati	0.0001	0.0046	0.0001
single-A. tessellati vs single-A. oleae	0.0001	0.0001	0.0002
single-A. tessellati vs mixed-A. oleae	0.0001	0.0001	0.0001
mixed-A. tessellati vs single-A. oleae	0.0003	0.0001	0.0001
mixed-A. tessellati vs mixed-A. oleae	0.0001	0.0001	0.0001
single-A. oleae vs mixed-A. oleae	0.0001	0.378	0.0001

c. Temporal comparisons within treatments (p-values)
	1 dpi vs 5 dpi	1 dpi vs 12 dpi	5 dpi vs 12 dpi
N. benthamiana DCL2 supressed	p values
single-A. tessellati	0,0002	0,0341	0,0003
single-A. oleae	0,0001	0,0001	0,0001
mixed-A. tessellati	0,0001	0,0001	0,0009
mixed-A. oleae	0,0001	0,0001	0,0001

d. Comparisons among treatments at each time point (p-values)
	1 dpi	5 dpi	12 dpi
N. benthamiana DCL2 supressed	p values
single-A. tessellati vs mixed-A. tessellati	0,0023	0,0001	0,0001
single-A. tessellati vs single-A. oleae	0,0001	0,0001	0,0001
single-A. tessellati vs mixed-A. oleae	0,0001	0,0001	0,0001
mixed-A. tessellati vs single-A. oleae	0,0001	0,0002	0,0001
mixed-A. tessellati vs mixed-A. oleae	0,0001	0,0001	0,0001
single-A. oleae vs mixed-A. oleae	0,0002	0,0001	0,0001

e. Temporal comparisons within treatments (p-values)
	1 dpi vs 5 dpi	1 dpi vs 12 dpi	5 dpi vs 12 dpi
N. benthamiana DCL4 supressed	p values
single-A. tessellati	0,0001	0,0001	0,0001
single-A. oleae	0,0001	0,0001	0,0001
mixed-A. tessellati	0,0001	0,0001	0,0001
mixed-A. oleae	0,0015	0,0021	0,0005

f. Comparisons among treatments at each time point (p-values)
	1 dpi	5 dpi	12 dpi
N. benthamiana DCL4 supressed	p values
single-A. tessellati vs mixed-A. tessellati	0,0001	0,0001	0,0003
single-A. tessellati vs single-A. oleae	0,0001	0,0001	0,0001
single-A. tessellati vs mixed-A. oleae	0,0008	0,4807	0,0001
mixed-A. tessellati vs single-A. oleae	0,0001	0,0001	0,0001
mixed-A. tessellati vs mixed-A. oleae	0,0021	0,0001	0,0001
single-A. oleae vs mixed-A. oleae	0,0001	0,0001	0,0001

a. Temporal comparisons (1 vs 5 dpi, 1 vs 12 dpi, and 5 vs 12 dpi) within treatments (single A. tessellati-infection; single A. oleae-infection; mixed A. tessellati-infection; and mixed A. oleae-infection); b. Comparisons among treatments (single-A. tessellati vs mixed-A. tessellati; single-A. tessellati vs single-A. oleae; single-A. tessellati vs mixed-A. oleae; mixed-A. tessellati vs single-A. oleae; mixed-A. tessellati vs mixed-A. oleae; and single-A. oleae vs mixed-A. oleae) at each time point (1, 5, and 12 dpi). Pairwise PERMANOVA comparisons of relative viral abundance in dcl2-suppressed plants: c Temporal comparisons (1 vs 5 dpi, 1 vs 12 dpi, and 5 vs 12 dpi) within treatments (single A. tessellati-infection; single A. oleae-infection; mixed A. tessellati-infection; and mixed A. oleae-infection); d Comparisons among treatments (single-A. tessellati vs mixed-A. tessellati; single-A. tessellati vs single-A. oleae; single-A. tessellati vs mixed-A. oleae; mixed-A. tessellati vs single-A. oleae; mixed-A. tessellati vs mixed-A. oleae; and single-A. oleae vs mixed-A. oleae) at each time point (1, 5, and 12 dpi). Pairwise PERMANOVA comparisons of relative viral abundance in dcl4-suppressed plants: e Temporal comparisons (1 vs 5 dpi, 1 vs 12 dpi, and 5 vs 12 dpi) within treatments (single A. tessellati-infection; single A. oleae-infection; mixed A. tessellati-infection; and mixed A. oleae-infection); f Comparisons among treatments (single-A. tessellati vs mixed-A. tessellati; single-A. tessellati vs single-A. oleae; single-A. tessellati vs mixed-A. oleae; mixed-A. tessellati vs single-A. oleae; mixed-A. tessellati vs mixed-A. oleae; and single-A. oleae vs mixed-A. oleae) at each time point (1, 5, and 12 dpi)

In dcl2-suppressed plants, A. tessellati accumulation remained low across infection types, indicating that suppression of DCL2 alone is insufficient to support robust systemic infection by this virus. In contrast, A. oleae accumulated efficiently in single infections and dominated mixed infections, establishing systemic infection in all cases. In dcl4-suppressed plants, A. tessellati accumulated at all time points in single infections and became the dominant virus during mixed infection, surpassing A. oleae across all sampling points. These results indicate that DCL4 plays a key role in restricting A. tessellati accumulation and that suppression of DCL4 shifts viral dominance during mixed infection (Fig. 2; Table 1).

Dynamic regulation of DCL gene expression across infection contexts

In wild-type plants, DCL4 expression peaked at 5 dpi across all infection conditions and was consistently higher than DCL2. Both genes were upregulated in infected plants relative to mock controls, with mixed infection inducing particularly strong responses at 5 dpi (Fig. 3a; Table 2). DCL2 expression also peaked at 5 dpi in most infection contexts. Mixed infection promoted significantly higher DCL2 expression compared to single infections, especially relative to A. oleae single infection and early stages of A. tessellati infection. Notably, DCL4 expression in mixed infection was lower than that observed during single A. tessellati infection at 5 dpi, suggesting that co-infection may alter the A. tessellati-associated DCL4 response. Overall, these results indicate that mixed infection modulates antiviral RNA silencing responses, characterized by elevated DCL4 expression and increased DCL2 responsiveness under co-infection conditions.

Fig. 3.

Mean relative expression ± standard error (SE) of DCLs (DCL1–DCL4) genes at each type of infection (mock-inoculated, single A. tessellati-infection, single A. oleae-infection A. tessellati and A. oleae mixed-infection) and time points (1dpi, 5dpi and 12 dpi) in Wild-type N. benthamiana (a.) plants and dcl2-suppressed plants (b.) and dcl4-suppressed plants (c.). Additionally, detailed results from complementary PERMANOVA statistical analyses are presented in Table 2 (Wild-type N. benthamiana) and Table S1 (dcl2 and dcl4-suppressed plants)

Table 2.

The two-way PERMANOVA test conducted between; “Treatment” (mock-inoculated, single A. tessellati-infection, single A. oleae-infection A. tessellati and A. oleae mixed-infection) and “Time points” (1dpi, 5dpi and 12 dpi) using pairwise comparisons performed with 9999 permutations under a reduced model to assess differences between DCLs (DCL1–DCL4) genes in Wild-type N. benthamiana plants

Treatments	DCL1			DCL2
	1dpi	5dpi	12 dpi	1dpi	5dpi	12 dpi
	p values			p values
mock-inoculated vs single-A. tessellati	0.6515	0.0013	0.4969	0.0006	0.0002	0.0001
mock-inoculated vs single-A. oleae	0.7955	0.5543	0.0104	0.001	0.0002	0.0065
mock-inoculated vs mixed infection	0.0302	0.0143	0.0876	0.0004	0.0001	0.0001
single-A. tessellati vs single-A. oleae	0.2487	0.0001	0.0001	0.0512	0.0012	0.004
single-A. tessellati vs mixed infection	0.008	0.0023	0.0016	0.0015	0.0003	0.6696
single-A. oleae vs mixed infection	0.0022	0.0032	0.0003	0.0004	0.0001	0.0039
Treatments	DCL3			DCL4
	1dpi	5dpi	12 dpi	1dpi	5dpi	12 dpi
	p values			p values
mock-inoculated vs single-A. tessellati	0.0109	0.0198	0.0102	0.0001	0.0001	0.0002
mock-inoculated vs single-A. oleae	0.0798	0.0916	0.0003	0.0001	0.0001	0.0002
mock-inoculated vs mixed infection	0.0169	0.0003	0.0137	0.0001	0.0001	0.0002
single-A. tessellati vs single-A. oleae	0.0016	0.479	0.0013	0.0001	0.0001	0.0056
single-A. tessellati vs mixed infection	0.1817	0.0003	0.0017	0.0002	0.0001	0.0017
single-A. oleae vs mixed infection	0.002	0.0004	0.0001	0.0002	0.0013	0.0003
Treatments	DCL1			DCL2
	1 dpi vs 5 dpi	1 dpi vs 12 dpi	5 dpi vs 12 dpi	1 dpi vs 5 dpi	1 dpi vs 12 dpi	5 dpi vs 12 dpi
	p values			p values
mock-inoculated	0.1719	0.3298	0.9248	0.0408	0.0158	0.2056
single-A. tessellati	0.0032	0.0477	0.0001	0.0064	0.1723	0.0091
single-A. oleae	0.0627	0.0008	0.0001	0.3399	0.2323	0.0789
mixed infection	0.0117	0.0084	0.874	0.0004	0.0006	0.0002
Treatments	DCL3			DCL4
	1 dpi vs 5 dpi	1 dpi vs 12 dpi	5 dpi vs 12 dpi	1 dpi vs 5 dpi	1 dpi vs 12 dpi	5 dpi vs 12 dpi
	p values			p values
mock-inoculated	0.4692	0.8447	0.3667	0.1351	0.3495	0.7103
single-A. tessellati	0.0232	0.0008	0.0009	0.0001	0.0003	0.0001
single-A. oleae	0.0037	0.0005	0.0003	0.0004	0.0071	0.0029
mixed infection	0.0004	0.2692	0.0004	0.0008	0.0011	0.0073

In dcl2-suppressed plants, DCL4 remained the most highly expressed DCL during viral infection, particularly in single A. tessellati and mixed infections, whereas DCL3 showed lower expression levels and a more limited response. In contrast, in dcl4-suppressed plants, DCL2 exhibited the highest expression levels across infection types, while DCL3 displayed moderate expression. However, the expression levels of DCL4 in the dcl2 mutant background and DCL2 in the dcl4 mutant background were generally lower than those observed in wild-type plants. Overall, these results suggest distinct patterns of DCL expression depending on the genetic background, with DCL2 and DCL4 remaining the predominant DCL transcripts under their respective silencing conditions (Fig. 3b, c; Table S1).

Mixed infection alters vsiRNAs accumulation and size profiles

Genome comparison performed by Job Dispatcher analysis showed 74.7% similarity between A. tessellati and A. oleae. Small RNA analysis was conducted at 5 dpi, corresponding to peak symptom severity, viral accumulation, and DCL expression. The total number of trimmed sequence reads recovered was 24,961,203 for single A. tessellati infection, 11,593,019 for single A. oleae infection, and 13,617,985 for mixed infection. Following the assembly of non-host (viral) reads, the total number of vsiRNAs was 5,830,169 for the single A. tessellati infection, compared to 2,519,107 for the mixed infection, representing an approximate 50% decrease in vsiRNA abundance. Similarly, vsiRNA counts for A. oleae dropped from 1,291,579 in the single infection to 76,791 in the mixed infection, reflecting a nearly 17-fold decrease (Table 3).

Table 3.

Summary of RNA sequencing performed using the Illumina NovaSeq platform, presenting the total number of reads (18–25 nt) for each infection type (single A. tessellati infection, single A. oleae infection, and mixed infection with A. tessellati and A. oleae)

Type of infection		Raw reads	Virus reads	Host reads	vsiRNAs orientation (%)
					Forward	Reverse
Single	A. tessellati	24 961 203	5 830 169	19 131 034	99.7	0.3
	A. oleae	11 593 019	1 291 579	10 301 440	99.7	0.3
Mixed	A. tessellati	11 099 051	2 519 107	8 580 116	99.7	0.3
	A. oleae	13 541 195	76 791	13 464 404	96.3	3.7

The table includes raw read counts, as well as the number of reads mapped to viral and host genomes. Additionally, the percentage distribution of vsiRNA orientations (forward and reverse) is provided for each infection type

The siRNA orientation profile showed a strong predominance of forward-strand reads (> 99%) in most conditions, indicating clear strand bias. However, A. oleae vsiRNAs in mixed infection showed an increased proportion of reverse-strand reads, indicating altered processing dynamics (Table 3). In addition, the number of vsiRNA contact points along the viral genomes decreased during mixed infection, particularly for A. oleae, suggesting reduced targeting breadth.

In single infections, vsiRNAs from both viruses were dominated by 20-nt class (30.5% to A. tessellati and 31.5% to A. oleae), followed by 21-nt (24.3% to A. tessellati and 25.8% to A. oleae) and 19-nt (18.4% to A. tessellati and 18.5% to A. oleae) classes (Fig. 4a). A similar pattern was observed for A. tessellati during mixed infection to 20-nt, 21nt and 19-nt classes (31.8%, 24.5% and 17.9%, respectively). In contrast, A. oleae showed a shift in vsiRNA size distribution under mixed infection, with 21-nt vsiRNAs becoming the most abundant class (33.9%), surpassing 20-nt reads (26.3%) (Fig. 4b).

Fig. 4.

Percentage distribution of vsiRNA size classes (18–25-nt) under different infection conditions: (a.) in the single infection of A. tessellati and of A. oleae; (b.) in the mixed-infection of A. tessellati and of A. oleae

Mixed infection redirects vsiRNA genomic targeting

In A. tessellati, vsiRNA targeting during single infection was predominantly concentrated in the RdRp/p23 region, which represented the major hotspot with 178,755 reads. This hotspot remained conserved during mixed infection, although the overall abundance was reduced to 60,033 reads. Additional targeting was observed in the CP gene under both infection conditions (Fig. 5). In A. oleae, single infection showed two dominant vsiRNA targeting within the RdRp/p23 region with 22,161 reads and 17,532 reads. In contrast, mixed infection caused a pronounced shift in vsiRNA targeting toward the suppressor-associated p6 gene with 14,130 reads, which became the primary hotspot. Secondary targeting was also detected in the p8 gene with 4,798 reads. These results demonstrate a qualitative reprogramming of vsiRNA targeting in A. oleae during mixed infection (Fig. 5; Tables 4, S2).

Fig. 5.

The nucleotides sequences alignment in the higher clived portions of the RdRp, p8, p6 and CP genes of A. oleae (DQ083996.1) and A. tessellati (HQ651834.1) retrieved from the GenBank database. Blue shaded nucleotides are highlighting the high clived vsiRNAs in the A. tessellati genome positions, inluding the number of the reads. Green shaded nucleotides are highlighting the high clived vsiRNAs in the A. oleae genome positions, inluding the number of the reads. Gray shaded nucleotides are highlighting all expanded positions in terms of cliveting

Table 4.

Summary of the total number of reads and their respective percentages in the coding genes (p23: readthrough; RdRp: RNA-dependent RNA polymerase; coat protein; p8: movement protein; p6: movement protein and silence suppressor) at each type of infection (single A. tessellati-infection, single A. oleae-infection, A. tessellati and A. oleae mixed-infection)

Type of	Gene	p23	RdRp	p8	p6	CP	Total reads	Total reads
infection	Position in genome	60–668	60–2234	2218–2439	2443–2613	2636–3445	within genes
A. tessellati	Nº of reads	1 567 790	3 606 927	380 539	309 656	1 046 308	5 319 304	5 830 169
	%	26.9	61.9	6.5	5.3	17.9	91.2	100
A. tessellati-mixed	Nº of reads	582 294	1 554 372	185 913	142 135	433 385	2 305 621	2 518 935
	%	23.1	61.7	7.4	5.6	17.2	91.5	100
	Gene	p23	RdRp	p8	p6	CP	genes reads	total reads
	Gene position	61–669	61–2235	2219–2440	2443–2613	2635–3447
A. oleae	Nº of reads	358 344	760 820	75 330	109 385	250 202	1 191 057	1 291 579
	%	27.7	58.9	5.8	8.5	19.4	92.2	100
A. oleae-mixed	Nº of reads	10 930	21 301	8754	34 724	5074	69 742	76 791
	%	14.2	27.7	11.4	45.2	6.6	90.8	100

Analysis of the 5′-terminal nucleotide composition revealed that A. tessellati-derived vsiRNAs showed similar nucleotide preferences in single and mixed infections, with a predominance of C (32–34.3%), followed by U (30.4–31.4%) and A (25.2–25.5%). Since vsiRNAs carrying a 5′ C, U, and A are preferentially loaded into AGO5, AGO1, and AGO2/AGO4 complexes, respectively, these results suggest that AGO recruitment remained largely stable during mixed infection. In contrast, A. oleae-derived vsiRNAs displayed altered nucleotide composition under mixed infection, characterized by an increase in 5′ C (from 30.3% to 50.4%) and a decrease in 5′ A (from 22.7% to 8.7%), suggesting potential shifts in AGO loading preferences and changes in host antiviral RNA silencing responses (Table 5).

Table 5.

Percentages of nucleotide 5’-terminal composition of the vsiRNAs and their associated Argonauts. C nucleotides at 5’-terminal preferentially load into the AGO5 complexes; U nucleotides into the AGO1 complexes and A nucleotides into the AGO2 and AGO4 complexes

Argonauts	Single-infection		Mixed-infection
	A. tessellati	A. oleae	A. tessellati	A. oleae
	Percentage (%)		Percentage (%)
AGO-1	31.4	38.2	30.4	33.1
AGO-2 and AGO-4	25.5	22.7	25.2	8.7
AGO-5	32.0	30.3	34.3	50.4

Discussion

Mixed infections between Alphanecroviruses are common in nature [9] and can increase disease severity, transmission, infection, and crop loss. Our study demonstrates that A. oleae, a highly virulent virus, facilitates the systemic infection of A. tessellati, a less virulent species, by overcoming post-transcriptional gene silencing (PTGS) in N. benthamiana [49]. To our knowledge, this is the first report revealing that two closely related viruses, sharing over 74% genome identity, elicit distinct silencing responses during single and mixed infections, including shifts in Dicer-dependent vsiRNAs processing and viral targeting. This finding indicates that genomic similarity alone does not predict host silencing outcomes in mixed-infection contexts.

Single infections showed contrasting phenotypes. A. oleae caused rapid and severe systemic symptoms in N. benthamiana, affecting all leaves and plants by 10 dpi and leading to high mortality, whereas A. tessellati induced only mild, transient symptoms, with plants fully recovering by 8 dpi. In mixed infection, symptom development was accelerated, and disease severity increased compared to single infections. Interestingly, viral accumulation showed divergent trends: A. tessellati levels increased significantly under mixed infection, while A. oleae abundance declined relative to single infections. These results suggest that A. oleae acts as a silencing suppressor, functionally complementing A. tessellati and enabling its systemic spread, thereby creating an asymmetric interaction [50–52]. Our findings also reinforce that symptom severity is not always correlated with the abundance of both viruses. This suggests that the two viruses may converge upon host responses or amplify disease via shared signaling pathways, including hormonal or oxidative stress responses. This pattern contrasts with unrelated viral pairs, which often show simultaneous synergistic accumulation, and aligns with observations that related viruses may compete or interact asymmetrically [15, 53–56]. DCL analyses are consistent with a higher DCL4 relative expression in antiviral responses, while DCL2 may function as a compensatory component [30, 31, 38, 57]. However, in mixed infections, both DCL4 and DCL2 were strongly upregulated, suggesting cooperative activation rather than a strict hierarchy. DCL2 expression exceeded levels observed in single infections, challenging the view of DCL2 redundancy in the presence of DCL4. Qin et al. [58] proposed that DCL4 and DCL2 may operate as a dual protective frontline during intra- and intercellular silencing in N. benthamiana. Consistently, Deleris et al. [28] demonstrated that increased viral dosage can overcome DCL2-mediated effects, although DCL4 activity does not appear to increase proportionally with viral load. Together, these results suggest regulatory flexibility in N. benthamiana, where DCL2 may act not only as a compensatory factor but also as an active co-participant in antiviral defense when multiple viruses challenge the silencing machinery, particularly in responses to Alphanecroviruses.

DCL3 was transiently induced early in infection, particularly in A. tessellati infections, and partially compensated in dcl2- or dcl4-suppressed plants, reflecting the flexible interplay among DCLs to reinforce antiviral silencing under increased viral challenge.

Deep sequencing analyses revealed further complexities in vsiRNA biogenesis. A. tessellati generated the highest vsiRNAs levels in single infection, while in the mixed-infection vsiRNA levels decreased by ∼50%. A. oleae vsiRNAs dropped nearly 17-fold in mixed infection despite still being present at moderate abundance, indicating that vsiRNA accumulation is not strictly proportional to viral load. This decoupling suggests that virus–virus interactions and modulation of host Dicer processing influence vsiRNA biogenesis.

Both viruses exhibited a strong positive-sense strand bias in vsiRNAs, consistent with previous reports for + ssRNA viruses [59–63], though A. oleae showed a slight increase in negative-strand vsiRNAs during mixed infection, possibly reflecting altered replication dynamics or silencing competition.

VsiRNA size profiling revealed a novel shift in A. oleae from 20-nt dominance in single infection to 21-nt in mixed infection, a phenomenon not previously reported. This observation may present a novel concept that vsiRNAs length profiles are dynamic and context-dependent, potentially shaped by competitive interactions among DCLs, viral suppressors, or structural features of the viral RNA altered in mixed infection. Since 21-nt vsiRNAs are typically associated with DCL4 from the same Tombusviridae family [43], at first sight this change might indicate a higher efficiency in silencing, however, the response to this better efficiency would be the accumulation of vsiRNAs, but again, the results of A. oleae vsiRNAs accumulation in the mixed infection showed the opposite. This shift from 20-nt to 21-nt could also suggest a reconfiguration of DCL activity, possibly a reallocation from DCLs or, based on the results, a shift in substrate preference due to viral suppressor interference seems the most plausible. Such plasticity in the silencing machinery adds a new layer of complexity to RNAi-based defenses in mixed infections.

20-nt vsiRNAs, consistently abundant across conditions and reported for other viruses [64, 65] may represent non-canonical cleavage products or DCL isoform activity, emphasizing gaps in our understanding of DCL processing specificity. Hotspot analyses further support this notion of dynamic antiviral response. In both single and mixed A. tessellati infections, vsiRNAs predominantly targeted the RdRp region establishing a stable hotspot. In contrast, A. oleae vsiRNA targeting shifted from the RdRp region in single infection to the suppressor-associated p6 gene in mixed infection. This suggests a host-directed redirection of silencing in response to co-infection or potential viral modulation to protect essential replication regions. A. oleae p6 shares sequence identity with A. tessellati p6, previously identified as a silencing suppressor, reinforcing the idea that virus–virus interactions can reshape host antiviral targeting.

5’ terminal nucleotide analysis indicated preferential AGO loading, with A. tesselati vsiRNAs favoring C and U, and A. oleae vsiRNAs showing increased C under mixed infection, suggesting selective AGO recruitment influenced by viral context. Other studies on RNA viruses observed a higher vsiRNAs with A and U in the 5`-terminal end [38, 39, 66]. Overall, these data reveal that mixed infections alter not only vsiRNA abundance and size but also genomic targeting and potential AGO engagement, reflecting a highly plastic and virus-specific silencing response. Our results have practical implications for RNAi-based crop protection. Stable vsiRNA hotspots, such as the RdRp region in A. tessellati, may serve as targets for transgenic hairpin constructs, topical dsRNA sprays, or CRISPR/Cas13d applications [67]. Understanding how mixed infections reshape silencing efficiency and vsiRNA distribution is essential for predicting antiviral outcomes and designing effective control strategies. In conclusion, this study presents novel insights into how N. benthamiana, reconfigures its RNA silencing machinery in response to single and mixed infections by closely related Alphanecroviruses. Mixed infection enhances DCL4 and DCL2 expression, modifies vsiRNA biogenesis, and redirects silencing targets, creating asymmetric and dynamic virus–host interactions. Taken together, our findings refine understanding of viral synergism and antagonism, highlight the plasticity of plant RNA silencing, and inform strategies for managing mixed viral infections in crops. To further build upon these insights, future high-throughput sequencing studies encompassing expanded biological datasets will be valuable to reinforce the statistical resolution and generalize the host response patterns observed here. Importantly, these findings highlight that subtle shifts in vsiRNA size distribution, target site selection, and DCL activity may substantially influence antiviral efficiency, a concept that may extend to other agriculturally important hosts potentially affected by A. tessellati and A. oleae.

Material and methods

Plant growth, virus infections and sampling protocol

Wild-type N. benthamiana plants and N. benthamiana transgenic lines suppressed for dcl2 and dcl4 genes [68, 69] were grown in controlled growth chambers at 25 °C under a 16 h light/8 h dark photoperiod. DCL-suppressed lines were used exclusively for virus accumulation and DCL gene expression analyses. Viral inocula consisted of purified preparations of A. tessellati (GenBank: HQ651834.1) and A. oleae isolate GM6 (GenBank: DQ083996.1). For single infections, purified virus preparations were diluted 1:1 (v/v) in 0.05 M sodium phosphate buffer prior to mechanical inoculation. For mixed infections, equal volumes of each purified virus preparation were combined 1:1 (v/v) immediately before inoculation, without further dilution. This approach ensured that each virus was delivered at the same final concentration and individual dose as in the respective single infections, resulting in a higher cumulative viral load in the mixed inoculum. Virus purification was performed as described previously [18]. Briefly, infected leaves were homogenized in cold 0.1 M sodium phosphate buffer (1:3, w/v), filtered, clarified in the presence of organic solvents, concentrated by differential centrifugation, and further purified by ultracentrifugation through sucrose density gradient columns. Virus concentration was determined by measuring absorbance at 260 nm using a Quawell Q9000 micro spectrophotometer (Quawell Technology, Beijing, China) applying an extinction coefficient E^1% ₂₆₀ = 5.0. Each plant (one leaf per plant) was mechanically inoculated with 6 µg of purified virus. Four experimental groups were established with 12 biological replicates each (total n = 48): (1) mock-inoculated (buffer only), (2) A. tessellati-infected, (3) A. oleae-infected, and (4) A. tessellati + A. oleae mixed-infected. Samples were collected at one day post-inoculation (1dpi; inoculated leaves) and at 5 and 12 dpi from newly emerged, systemically infected leaves located above the inoculated leaves.

At each time point, three independent plants per treatment were sampled, and plants were not resampled at subsequent time points. For each biological replicate, 100 mg of leaf tissue was ground in liquid nitrogen and used for total RNA extraction. Symptom development was monitored daily for up to 30 days post-inoculation.

Evaluation of disease symptoms

To quantify disease progression, a four-level symptom severity scale was used: 0 = asymptomatic; 1 = mild chlorotic mosaic; 2 = pronounced chlorotic mosaic; 3 = necrotic mosaic; and 4 = extensive leaf necrosis with deformation. Symptom evaluation was performed exclusively on wild-type N. benthamiana plants. Infected plants, either singly infected with A. tessellati or A. oleae, or mixed-infected (both), were observed daily for 30 days. As described by Varanda et al. [2] the disease severity was quantified using the Disease Severity Index (DSI), calculated as: DSI (%) = [Σ (individual disease scores)/(total observations × maximum score)] × 100.

RNA extraction, RT-qPCR and vsiRNA deep sequencing

Total RNA was extracted from leaf samples (three biological replicates per condition) at 1 dpi, 5 dpi, and 12 dpi using RNeasy Mini kit (Qiagen) according to the manufacturer’s instructions. Purified RNA was resuspended in RNase-free water and immediately stored at − 80 °C until further use. RNA integrity and concentration were assessed using a Quawell 9000 microvolume spectrophotometer.

Specific primers for real-time PCR were designed targeting the coat protein (CP) regions of A. tessellati and A. oleae, selected based on their sequence divergence as described by Varanda et al. [9], to enable accurate differentiation between the two viruses. Primer design was performed using Primer Express 3.0 (Applied Biosystems, Foster City, CA, USA), with default parameters, except for the A. tessellatid_3 primer, which was adopted from the same previous study. As endogenous reference genes, phosphatase 2 A (PP2) and F-box protein (F-box) were employed (Table 6). Additionally, to investigate the potential involvement of the RNA silencing pathway in viral replication and infectivity, expression levels of DCL1 to DCL4 genes were quantified by RT-qPCR in wild-type N. benthamiana plants and in transgenic lines suppressed for each DCL gene. All primer sequences used in RT-qPCR are listed in Table 1. Reverse transcription was carried out using 1 µg of total RNA in a final volume of 20 µL, employing the Maxima First Strand cDNA Synthesis Kit for RT-qPCR (Thermo Scientific), following the protocol recommended by the manufacturer.

Table 6 .

Primers used in this study for RT–qPCR analysis targeting the coat protein (CP) gene of A. tessellati and A. oleae in N. benthamiana infections

Primer name	Sequence (5’ – 3’)	Position in genome	Region	Reference
CP_A. tessellati_5’	CCAAAACATGCCTAAGAGAG	2629—2648	CP	This study
CP A. tessellatid_3’	CCTAGATCTTCTGGGCTAAGC	2791—2771	CP	Varanda et al. 2010 [9]
CP_A. oleae_5’	TATGAAAGCGGATTTCACCC	2708—2728	CP	This study
CP_A. oleae_3’	CTGCTCATTAGTAGTAGTAAT	2877—2857	CP	This study
PP2ArtFW	GACCCTGATGTTGATGTTCGCT	-	-	Liu et al. 2012
PP2ArtREV	GAGGGATTTGAAGAGAGATTTC	-	-	Liu et al. 2012
FBOXrtFW	GGCACTCACAAACGTCTATTTC	-	-	Liu et al. 2012
FBOXrtREV	ACCTGGGAGGCATCCTGCTTAT	-	-	Liu et al. 2012
qNbDCL1-F	AAAAGAATGAGATGGTATTTCGG	-	-	Dadami et al., 2013 [68]
qNbDCL1-R	TTCTTTCTGGCATGCTCAA	-	-	Dadami et al. 2013 [68]
qNbDCL2-F	GAAGAACCACTTCTTAGGGGAAA	-	-	Dadami et al. 2013 [68]
qNbDCL2-R	GGCCATAACAAGGACTCAA	-	-	Dadami et al. 2013 [68]
qNbDCL3-F	AGGACTGCAGCGTTTATGGT	-	-	Katsarou et al. 2016
qNbDCL3-R	GGCCACAACAACCATGTCCT	-	-	Katsarou et al. 2016
DCL4_For b	GCACTTAACTACAGAGAAATGCAATG	-	-	Kotakis et al. 2020
DCL4_Rev b	ACAATGTTTGAGCGCCTTCT	-	-	Kotakis et al. 2010

Quantitative real-time PCR was performed using the Line Gene 9600 Fluorescent Detection System (Bioer). Each reaction contained 12.5 µL of 2 × SYBR Green PCR Master Mix, 0.3 µM of forward and reverse primers, and 12.5 ng of cDNA in a total volume prepared in 96-well plates. Thermal cycling conditions included an initial denaturation at 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min. Negative controls included no-template reactions and cDNA from uninfected plants. All samples were analyzed in triplicate as technical replicates. Ct values were calculated using the fit-point method with fluorescence thresholds manually adjusted above background using the Line Gene 9600 software. Melt curve analysis was conducted at the end of each run to verify primer specificity. Relative expression levels were estimated using standard curves automatically generated by the system, and results were normalized against PP2 and F-box reference genes.

For vsiRNA profiling, only wild-type N. benthamiana plants were used. Total RNA from three biological replicates per infection type was pooled prior to small RNA sequencing, precluding independent statistical analysis. Therefore, the vsiRNA profiling results are descriptive. Small RNA sequencing was performed at 5 dpi, a time point that showed optimal viral load and DCL expression, without severe leaf necrosis or deformation. Total small RNA sequencing quality and concentration were assessed by determining the RNA Integrity Number (RIN). Small RNA libraries were constructed using the NEBNext Multiplex Small RNA Library Prep Kit (New England Biolabs), following the manufacturer’s instructions. Sequencing was performed on the Illumina NovaSeq platform, generating 18–36 bp single-end reads. Raw sequencing data corresponding to viral small interfering RNAs (vsiRNAs) were processed and quality-trimmed using CLC Genomics Workbench v12.0.3 (Qiagen, 2019). High-quality reads were mapped against the complete genomes of A. tessellati and A. oleae (GenBank accessions HQ651834.1 and DQ083996.1, respectively), using stringent parameters (length and similarity fractions set to 1.0) to ensure 100% sequence alignment. Only small RNAs with read counts ≥ 100 were retained for downstream analyses. Annotation was performed using a curated, non-redundant plant virus database (vrl_plant v232; [70]) and the miRBase v22.1 database [71] to classify known and novel small RNAs.

Statistical data analysis

Pairwise sequence identities of the full genomes of the two Alphanecroviruses representatives used in this work were determined using the Job Dispatcher bioinformatics tool at EMBL-EBI [72]. For nucleotide sequence alignment of both viruses, the Clustal Omega alignment program (European Bioinformatics Institute, EMBL-EBI) was used. PRIMER v6 software [73] and PERMANOVA add-on package [74] were used to perform univariate and multivariate analyses with the permutational analysis of variance to detect significant differences (p < 0.05): in the viruses' relative abundance and DCLs relative expression levels between: “Treatment” single A. tessellati-infection, single A. oleae-infection and A. tessellati and A. oleae mixed-infection and “Time points” 1 dpi, 5 dpi, and 12 dpi. To test the null hypotheses that no significant differences existed, a two-way PERMANOVA was applied between; “Treatment” and “Time points”. PERMANOVA analyses were carried out with the following two-factor design: “Treatment” single A. tessellati-infection, single A. oleae-infection and A. tessellati and A. oleae mixed-infection (three levels, fixed); and “Time points”, 1 dpi, 5 dpi, and 12 dpi (three levels, random). In addition, in the DCLs analysis it was included uninfected (only buffer inoculated) plants at the “Treatment” level. The data were square root transformed in order to scale down the importance of the high values of relative viruses and DCLs levels. PERMANOVA was applied using a Bray–Curtis similarity matrix, following the method of Clarke and Green [74]. When the number of unique permutations was lower than 150, p-values were adjusted using the Monte Carlo test. In cases where significant interaction effects were identified, pairwise comparisons were subsequently performed with 9999 permutations under a reduced model to assess differences between groups [75].

Supplementary Information

Below is the link to the electronic supplementary material.

Authors’ contributions

PM, MRF and CV designed the experiments; PM wrote the manuscript; JR, AA and CM contributed to the material planting and sample collection; PM, JR, AA and CM also analyzed the RNA-seq data; PM, MRF and CV revised the manuscript; PM and CV designed, led and coordinated the overall study. All authors read and approved the final manuscript.

Funding

This work was funded by the project “VIRUSREST – Genetic engineering of tomato plants for virus resistance, following EU new regulations” (Project No. 15821; COMPETE2030-FEDER-00693100), supported by FEDER through COMPETE 2030 – Programa de Inovação e Transição Digital (PITD).

Patrick Materatski’s contract was supported by Portuguese National Funds through FCT/MCTES, under the CEEC (https://doi.org/10.54499/2021.01553.CEECIND/CP1670/CT0003). J.R. was supported by Portuguese National Funds through FCT/MCTES, under the PhD scholarship 2022.13638.BD. The authors also thank the support of CERNAS—UIDB/00681/2025, MED—UID/05183/2025 ( https://doi.org/10.54499/UID/05183/2025) and CHANGE (https://doi.org/10.54499/LA/P/0121/2020).

Data availability

The vsiRNA datasets generated and analyzed during the current study are available in the NCBI Sequence Read Archive (SRA) repository under BioProject accession PRJNA1401704 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1401704). The datasets correspond to vsiRNAs obtained from Nicotiana benthamiana plants subjected to single infection with A. tessellati (OMMV) (BioSample SAMN54572195: https://www.ncbi.nlm.nih.gov/sra/SRX31778995 [accn]), single infection with A. oleae (OLV-1) (BioSample SAMN54572196: https://www.ncbi.nlm.nih.gov/sra/SRX31778996 [accn]), mixed infection with A. tessellati and A. oleae (OMMV and OLV-1) (BioSample SAMN54572197: https://www.ncbi.nlm.nih.gov/sra/SRX31778997 [accn]), and virus-free control plants (BioSample SAMN54572198: https://www.ncbi.nlm.nih.gov/sra/SRX31778998 [accn]).

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.