Evaluating the Negative‐Strand Coding Potential in Plum Pox Virus

ABSTRACT

Recent studies proposed that the negative strands of some single‐stranded, positive‐sense RNA viruses contain reverse open reading frames (rORFs) that encode functional peptides. Highly conserved rORFs were found in the negative strand of viruses within the Potyviridae family, including turnip mosaic virus (TuMV), where rORF2 was shown to be essential for infection and survival. In this study, we analysed the existence and potential functionality of rORFs in plum pox virus (PPV) using mass spectrometry and mutagenesis. Unbiased mass spectrometry analyses failed to identify any rORF‐encoded peptides and PPV mutants abolishing the expression of individual rORF‐derived peptides showed no significant difference in infection rates compared to the wild‐type virus in the experimental host Nicotiana benthamiana or in the natural host Prunus persica . Furthermore, competitive infection experiments revealed that neither the wild type nor the rORF2 knockout mutant, which had proven lethal when tested in TuMV, exhibited any competitive advantage over the other. Taken together, these findings suggest that rORFs in PPV are either not produced or are functionally irrelevant for infection, and they highlight the need to evaluate the role of rORFs in other potyviruses on a case‐by‐case basis.

Plum pox virus produces two polyproteins from ORFs in the positive strand. This study explores the possible expression of peptides from negative‐strand ORFs (rORFs).

Plant viruses employ numerous strategies to express their proteins. Beyond the canonical translation at an AUG codon in the main open reading frame (ORF), they utilise alternative mechanisms such as internal ribosome entry, ribosome shunting, leaky scanning, non‐AUG initiation, reinitiation, readthrough, StopGo, polymerase slippage or partial processing (Firth and Brierley 2012; Sorokin et al. 2021).

The Potyviridae family is the largest group of plant RNA viruses. Its members are widely distributed worldwide, affecting nearly every economically important crop (Revers and García 2015; Yang et al. 2021). According to the International Committee on Taxonomy of Viruses (ICTV) at May 2025, the family is currently composed of 12 genera and 285 species, with 236 of these belonging to the Potyvirus genus (Inoue‐Nagata et al. 2022). Like viruses from most genera in the family, potyviruses are monopartite, expressing all known proteins from a single‐stranded, positive‐sense RNA. This strand encodes a main ORF that produces a polyprotein that is further processed by viral endopeptidases to give rise to 10 proteins: P1, HCPro, P3, 6K1, CI, 6K2, VPg, NIapro, NIb and CP (Rodamilans et al. 2018; Sanfaçon 2025). Partial processing products like P3‐6K1 of unknown function, or VPg‐NIapro with a role in protease activity modulation (Mathur et al. 2012; Sabharwal et al. 2018), are also produced (García et al. 1992; Schaad et al. 1996). All potyviruses can modify their genome through polymerase slippage, usually at a conserved A₆ motif in the P3 coding sequences, generating alternative genomes with modified ORFs encoding P3N‐PIPO and/or P3N‐ALT (Chung et al. 2008; Olspert et al. 2015; Rodamilans et al. 2015; Valli et al. 2024). The A₆ motif is not restricted to the P3 coding sequence. Some potyviruses infecting sweet potato also present this motif in the P1 cistron, producing an alternative ORF that encodes the protein P1N‐PISPO, with RNA silencing suppression activity, among other possible roles (Mingot et al. 2016; Untiveros et al. 2016). Other potyviruses produce alternative peptides or proteins, but they do it using the main ORF, probably after an exaptation process (Koonin et al. 2022). Known examples include endive necrotic mosaic virus, which encodes an AlkB motif within the P1 cistron conferring additional functions to an already existing protein (Pasin et al. 2022); and euphorbia ringspot virus, which infects plants from the Euphorbiaceae family (Knierim et al. 2017) and encodes a Maf1‐HAM1‐like protein between NIb and CP, linked to host adaptation (Valli et al. 2022).

A recent publication on the Potyvirus rapae (turnip mosaic virus, TuMV) (Gong et al. 2023) revealed that this virus uses the negative strand for the production of small peptides that are critical for viral infection. This uncommon strategy to generate alternative proteins was described previously in narnaviruses (Dinan et al. 2020), and more recently in tobamoviruses (Gong et al. 2025). For TuMV, the authors explored the level of conservation of the possible ORFs found in the negative strand (termed rORFs) of the virus compared to the viral genomes of other potyviruses. They selected four different rORFs with high conservation (> 50%) and disrupted them by introducing silent mutations in the main ORF. Remarkably, the resulting mutated variants showed a defective infection capacity compared to the wild type. The most dramatic case was the rORF2 knockout, which completely lost its infection ability. Interestingly, infectivity was restored when the rORF2‐encoded protein was supplied in trans.

This remarkable finding led us to investigate how general this mechanism is among potyviruses. We focused our efforts on Potyvirus plumpoxi (plum pox virus, PPV), one of the best characterised viruses in the Potyviridae family, with abundant information regarding different strains and complete genome sequences (García et al. 2024). An in silico search with the bioinformatic tool ORFfinder from the NCBI, using PPV‐D (GenBank accession EF569214) as a reference, identified 81 possible rORFs when considering AUG as the starting codon and a minimum size for the coding peptides of 10 amino acids. We used this database of theoretical viral proteins as a reference for the identification of tryptic peptides following mass spectrometry (MS) analysis, which allows us to identify proteins present in low amounts in a certain tissue, being very useful in finding unknown or uncharacterised viral peptides without any bias for a specific sequence (van Bergen et al. 2022). Indeed, Gong et al. (2023), in their analysis of a TuMV‐infected sample of Nicotiana benthamiana, were able to identify one of the peptides corresponding to the coding sequence of TuMV rORF2, indicating that this protein was expressed during plant infection.

Thus, we used MS analysis as a first step in the identification of rORF‐derived peptides in PPV. We inoculated two plants of N. benthamiana with PPVr (Rodamilans et al. 2021), a PPV‐D isolate that shares more than 99% identity to PPV‐D EF569214. It is well adapted to herbaceous hosts and engineered to express the red fluorescence protein (RFP) coding sequence inserted between NIb and CP cistrons. At 6 days post‐inoculation (dpi) we observed fluorescence in the upper non‐inoculated leaves and collected two leaves from each plant, pooling them together for the analysis. Following tryptic digestion and a combination of liquid chromatography coupled with Orbitrap fragmentation, we identified a total of 26,627 peptides corresponding to 6257 proteins. Of these, 144 peptides were identified as belonging to PPVr (Table S1, sheet‐table 1). These accounted for the 10 proteins expressed in the main ORF and represented 29% coverage of the total viral proteome (Figure 1a). The identified peptides included partial cleavage products of P3‐6K1 and VPg‐NIaPro, alongside fully processed N‐termini of CI and CP, and C‐terminal fragments of 6K1, 6K2, VPg and NIb. No rORF‐encoded peptides were detected. To rule out a possible host‐specific effect, we performed a similar MS analysis using Prunus persica tissue inoculated with pLXB2‐SD‐PPV (González de Prádena et al. 2020), an infectious clone that expresses a chimerical virus, known as SD‐PPV, capable of infecting both herbaceous and Prunus species. In addition, this virus carries the green fluorescent protein (GFP) coding sequence inserted between NIb and CP cistron to facilitate infection tracking. A total of 7314 proteins and 39,605 peptides were identified, of which 205 peptides corresponded to SD‐PPV (Table S1, sheet‐table 2), representing 51% of the viral proteome. Partial cleavage products of P1‐HCPro, P3‐6K1, VPg‐NIapro and NIPro‐NIb were detected, as well as fully processed N‐termini of CI and CP and C‐termini of CI, NIapro and NIb. As in the previous analysis, rORF‐encoded peptides were not detected.

FIGURE 1.

Mass spectrometry analyses and rORF search and selection. (a) Schematic representation of the proteins encoded within the two main ORFs of plum pox virus (PPV) generated from the positive strand. Pale green regions and pale yellow regions indicate areas covered by peptides identified via mass spectrometry in the case of PPVr and SD‐PPV infected tissue, respectively. RFP and GFP cistrons and the corresponding peptides have been omitted for clarity. Selected rORFs, encoded in the negative strand, are shown below. rORFs studied in turnip mosaic virus (TuMV) are indicated and marked in red or purple if selected or not for this study. Newly selected rORFs are marked in blue. (b) rORF conservation analysis among potyviruses. Numbers are shown for rORFs with > 30% conservation; intraspecies conservation of PPV is indicated in parentheses; position in the reference genome of PPV‐D EF569214 is shown in square brackets for rORFs1‐6. Unselected rORFs are shown in grey. Selected rORFs follow the same colour code as in (a).

Assuming rORF‐derived peptides are expressed, several factors could account for their absence in the above analyses. These include (i) very low expression levels resulting in peptides being below the detection limit of the MS equipment; (ii) difficulties in detecting peptides following tryptic digestion, potentially due to their size or other technical limitations; and (iii) temporal differences between peptide expression and sample collection—specifically, peptides accumulate at a different time point from the sampling. To circumvent these potential problems in the identification of rORFs in PPV, we adopted a more direct approach: generating knockout mutants for specific rORFs to assess their impact on viral infectivity. If the introduced mutations disrupt the expression of a factor essential for viral fitness, we anticipated substantial differences during viral infection between wild type and mutant variants, or alternately, the emergence of reversion mutations or compensatory changes aimed at restoring initial fitness.

To reduce the number of rORFs to a manageable number, we selected and tested those most likely to be produced. As in the case of TuMV (Gong et al. 2023), we considered peptide conservation among potyviruses as a strong indicator for bona fide rORFs. We downloaded 147 complete genome sequences from different potyviral species available at the NCBI (Table S2, sheet‐table 1) and obtained their corresponding rORFs. A database was generated and used for BLAST analysis against the rORFs of PPV‐D EF569214. Sequences with an E‐value < 0.05 were considered conserved peptides among species. The size of the generated peptides and the level of conservation were used to map the rORFs of PPV. As shown in Figure 1b, rORFs were predominantly small (30 amino acids in average) and exhibited low degrees of conservation among potyviral species. We selected 10 rORFs that were present in more than 30% of the analysed potyviruses; to further refine this number, we considered the level of conservation among different PPV isolates. To achieve this, we downloaded 487 PPV isolates available on NCBI from all 10 strains (Table S2, sheet‐table 2) and performed a similar conservation analysis as for potyvirus species. Due to anticipated higher conservation rate, we discarded rORFs present in less than 95% of isolates, resulting in eight remaining rORFs. The conservation percentage for each is shown in Figure 1b. These included the four previously described in TuMV (rORF1‐4); rORF2, rORF3 and rORF4 were selected for experimental analysis. rORF1 was excluded due to its location in a region containing an important RNA secondary structure (Firth 2014; Mahajan et al. 1996). Any effect observed from a nucleotide change in this region would be difficult to attribute specifically, as it could be the result from the loss of the rORF product or from the disruption of the secondary structure. To complement and extend previous results, we included the two rORFs (rORF5 and rORF6) identified as having the greatest degree of interspecies conservation based on sequence similarity (Figure S1).

The five selected rORFs (rORF2–6) were distributed throughout the entire negative strand (Figure 1a), and corresponding mutations were then engineered for further experimental tests (rORFm2–6) (File S1). To account for potential host specificities, mutants were prepared in the pLXB2‐SD‐PPV background.

SD‐PPV and rORFm constructs were independently inoculated by agroinfiltration into N. benthamiana and P. persica plants at the same time. We used two independent clones per construct, inoculating three plants per clone for a total of six plants. At 10 days post‐inoculation (dpi), typical PPV symptoms were observed in all inoculated plants. At 13 dpi, GFP fluorescence was detected under UV light, revealing no discernible differences between SD‐PPV and rORFm‐inoculated plants. At 21 dpi, leaf samples were collected and photographs were taken under a fluorescent lamp. As before, no differences were observed among plants inoculated with SD‐PPV and those inoculated with mutant derivatives. The two plants that showed the highest fluorescence levels from each clone (four plants in total) were selected for western blot analysis. No significant differences in virus accumulation were detected among these samples (Figures 2 and S2). To evaluate the stability of the introduced mutations and assess whether any mutations had arisen to compensate for a potential loss of fitness, reverse transcription (RT)‐PCR analysis was performed. Using equal amounts of total RNA extracted from the same tissue of those plants analysed by western blot, cDNA was prepared and used as a template to amplify genome segments spanning the corresponding mutated rORFs. Sequencing of the PCR products revealed that the engineered mutations were stable and that no additional mutations appeared in the rORFm constructs, at least in the sequenced genome fragment (Figure S3).

FIGURE 2.

Inoculation experiment of wild type (WT) and rORFm constructs in plants. (a) Nicotiana benthamiana images were taken at 21 days post‐agroinfiltration (dpa) (left panel). One representative image was used per case. Western blot densitometry analysis is shown in the middle panel and reverse transcription‐quantitative PCR (RT‐qPCR) analysis of viral RNA titres is shown on the right. Lowercase letters indicate significant differences or lack thereof, determined by one‐way ANOVA with Tukey's post hoc test. (b) Prunus persica images were taken at 21 dpa (left panel). One representative image was used per case. Western blot densitometry analysis is shown in the middle panel and RT‐qPCR analysis of viral RNA titres is shown on the right. Uppercase letters indicate significant differences or lack thereof, determined by a nonparametric Kruskal–Wallis test because the data did not pass the normality criteria.

To facilitate viral evolution, infected N. benthamiana plants were used as inoculum sources in a passage experiment. For each construct, one leaf from each of those plants analysed by western blot was selected, pooled and homogenised; and each of the resulting extracts was then used to inoculate three new plants. These newly inoculated N. benthamiana plants developed clear symptomatology at 10 dpi, and samples were collected at 21 dpi for analysis. Total RNA was extracted from each plant, and equal amounts were pooled to prepare cDNA and perform PCR amplification of the previously analysed genome segment. No changes were observed in any of the constructs (Figure S4), not even in the rORFm2, which had completely abolished infection when tested in TuMV (Gong et al. 2023). In the case of P. persica trees, we implemented a vernalisation cycle using SD‐PPV‐ and rORF2m‐inoculated plants. After dormancy, the plants sprouted rapidly, promoting viral replication and increasing the likelihood of mutations (Martínez‐Gómez et al. 2000). Symptoms appeared in the newly emerging leaves of all plants at the same time. One leaf from each plant inoculated with rORF2m was collected and used for total RNA extraction and cDNA preparation. These cDNAs were then pooled and used as templates for PCR amplification. Amplicon sequencing revealed no detectable changes in the genome segment encompassing the rORF2 (Figure S4).

A useful method to assess fitness differences between mutant variants and their parental virus is through a competition experiment. While a small loss of fitness may be undetected in a single infection, it would become detrimental in a competitive scenario, where the more fit virus would dominate over the less fit one (Elena et al. 2008). Given the relevance of rORF2 shown in TuMV, and considering the results obtained in the single inoculation experiments described above, we selected rORF2m to compete with SD‐PPV, and prepared cartridges using a 1:1 and 1:2 ratio (parental:mutant) for biolistic inoculation of N. benthamiana and P. persica plants, four plants per mix. At 10 dpi, all Nicotiana plants and seven Prunus trees showed clear symptoms. At 28 dpi, individual samples were collected from each plant. Total RNA was extracted, followed by cDNA synthesis for subsequent PCR amplification and sequencing (Figures 3 and S4). Results did not show a consistent pattern: SD‐PPV predominated in some cases, whereas rORF2m was imposed in others, independently of the host, indicating a stochastic selection of the variants. This suggests that the rORF2m variant does not exhibit a specific loss of fitness.

FIGURE 3.

Analysis of the competition experiment between rORF2m and wild type plum pox virus (PPV). Electropherograms corresponding to reverse transcription‐PCR amplification and sequencing of the indicated samples are shown. Sequencing results from Nicotiana benthamiana and Prunus persica plants are displayed on the left and right, respectively; a cartridge amplification control is shown above; the mutated nucleotide in rORF2m is highlighted with an orange rectangle.

The level of conservation of the rORFs found in TuMV suggested that the peptides they encode might be common among potyviral species and could therefore play relevant roles in their respective infections (Gong et al. 2023). Our results with PPV lead us to take a more nuanced view. Mass spectrometry analysis of two independent experiments with experimental and natural hosts did not identify a single peptide for any of the 81 proteins encoded by the corresponding rORFs of the negative strand. Despite obtaining high coverage of those proteins encoded in the PPV main ORF, including partial processing products, we were unable to find any of the potential peptides derived from the rORFs. Subsequent experiments with the mutants corresponding to the most likely rORFs, conservation‐wise, reinforce the notion that in PPV there is no peptide relevant for the infection encoded in the negative strand. While the mass spectrometry results presented some limitations that could explain the absence of peptides, as previously stated, the second set of results based on silent point mutations overcame these limitations narrowing alternative explanations, at least in the case of the five rORFs tested. We cannot rule out the possibility that an untested rORF might be relevant for PPV infection. However, if that was the case, conservation among Potyviridae would not serve as a selective criterion. Additionally, we must consider the possibility that certain rORF products could be relevant at early stages of infection that were not analysed, play roles in untested functions such as transmission, or become significant under more challenging conditions. The production of rORFs from internal AUG or non‐AUG start codons is also a possibility that could be explored.

Given that identified rORFs are based on intraspecies conservation, it is interesting to question why these sequences are preserved among different potyviruses if they are not relevant or even expressed. One potential explanation is that the BLAST search parameters (E‐value < 0.05) may not be stringent enough to distinguish between truly conserved and non‐conserved proteins, which could lead to misidentification of potentially relevant rORFs. Analysing the level of identity of the different proteins produced by rORFs 2–6 in potyviral species with respect to PPV, we find they range between 12.5% and 81%. The lower limit is below the consensus threshold for homology (20%) and might support at least in part this possible explanation (Rost 1999). Another interesting point to consider is whether the rORFs are actually in frame with the main ORF. That is, nucleotides 1 and 2 of the 3‐letter codon in the rORF match nucleotides 2 and 1 of the codon for the main ORF. Considering that these nucleotides are very restrictive for specific amino acids, it is reasonable to assume that conserved amino acids in the main ORF would maintain a theoretical rORF if both frames follow this alignment. Remarkably, this 1,2–2,1 alignment is found in all rORFs with more than 40% interspecies conservation, including rORFs tested in this work. The exception to this is rORF6. In this case, nucleotides 1 and 2 of the 3‐letter codon in the main ORF correspond to nucleotides 2 and 3 of the codon in the rORF, so high conservation was not anticipated. Nonetheless, results with this rORF are clear, and there is no indication that a protein relevant for infection is produced in this case. Supporting this finding are the results showing that exchanging this part of the genome (HCPro cistron in the main ORF) by different sequences has no dramatic effect on viral infection, as long as the new sequence encodes for an RNA silencing suppressor (Carbonell et al. 2012; Maliogka et al. 2012; Valli et al. 2012). An undescribed RNA secondary structure in this region might explain the level of conservation found across different potyviruses (Rivas 2021).

Our findings indicate that PPV is unlikely to express functional proteins from its negative strand, at least from the most conserved rORFs. This supports the notion that potyviruses employ distinct strategies to exploit their genomes, highlighting the need for case‐by‐case analyses to uncover alternative infection mechanisms linked to noncanonical protein expression. In the case of PPV, this also represents a point of continuity when considering special requirements for the development of new varieties from a plant‐breeding perspective, while suggesting that a better understanding of species‐specific viral strategies may help guide future studies in this area.

Author Contributions

Beatriz García: investigation, Review and editing; José Vicente Del Álamo: investigation, Review and editing; Mario Rincón: investigation, Formal analysis, Review and editing; Irene Gonzalo: investigation, Review and editing; Xiaofei Cheng: conceptualization, Review and editing; Juan Antonio García: conceptualization, Review and editing; Carmen Simón‐Mateo: funding acquisition, Conceptualization, Review and editing; Adrián A. Valli: project administration, Funding acquisition, Conceptualization, Writing, Review and editing; Bernardo Rodamilans: formal analysis, Project administration, Conceptualization, Writing, Review and editing.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Sequence alignments of the different rORFs. Ten conserved sequences were used in each case for alignment using Jalview (Waterhouse et al. 2009). Viral species vary in each case depending on the rORF conservation. Red arrows indicate the amino acid mutated in the corresponding rORFm. NCBI codes can be found in Table S2. AnMV, anemone mosaic virus; AV‐1, asparagus virus 1; BaRMV, basella rugose mosaic virus; ChiRSV, chilli ringspot virus; ChiVMV, chilli veinal mottle virus; ENMV, endive necrotic mosaic virus; FreMV, freesia mosaic virus; HarMV, hardenbergia mosaic virus; HMV, hippeastrum mosaic virus; LMoV, lettuce mosaic virus; NDV, narcissus degeneration virus; NYSV, narcissus yellow stripe virus; PepMV, pepper mottle virus; PPV, plum pox virus; PRSV, papaya ringspot virus; PVA, potato virus A; PVMV, pepper veinal mottle virus; PVY, potato virus Y; SPFMV, Sweet potato feathery mottle virus; TeMV, telosma mosaic virus; TEV, tobacco etch virus; TuMV, turnip mosaic virus; TVMV, tobacco vein mottling virus; VVY, vervena virus Y; YBMV, yam bean mosaic virus.

Figure S2: Western blot analysis of plants at 21 dpi. Anti‐CP immunoblots of protein extracts from the indicated tissue; Constructs used to inoculate each plant are shown; numbers below indicate the plant number; NT, nontreated plant; MW, molecular weight; a Ponceau red‐stained blot (RbcL) showing the small subunit of the Rubisco is displayed below the membrane as loading control.

Figure S3: Analysis of the inoculated plants at 21 dpa. Electropherograms corresponding to RT‐PCR amplification and sequencing of the indicated samples are shown. Sequencing results from Nicotiana benthamiana and Prunus persica plants are displayed on the left and right, respectively; the mutated nucleotide in each case is highlighted with an orange rectangle and its number in the SD‐PPV sequence is shown in the left.

Figure S4: Analysis of the plants after passage and vernalization. Electropherograms corresponding to RT‐PCR amplification and sequencing of the indicated samples are shown. Sequencing results from Nicotiana benthamiana and Prunus persica plants are displayed on the left and right, respectively; the mutated nucleotide in each case is highlighted with an orange rectangle and its number in the SD‐PPV sequence is shown in the left.

Table S1: Viral peptides identified by mass spectrometry.

Table S2: Potyviral sequences used in the rORF conservation analysis.

Data S1: Materials and methods.

García, B. , Del Álamo J. V., Rincón M., et al. 2025. “Evaluating the Negative‐Strand Coding Potential in Plum Pox Virus.” Molecular Plant Pathology 26, no. 12: e70179. 10.1111/mpp.70179.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.