PELOTA at the crossroads of tomato fertility and viral resistance · New Phytologist · El Autómata
Martes, 17 de noviembre de 2026
Portada de las 23:15
Todos los periódicos y la radio, en unoEscrito por IA a partir de 26 medios · Cómo se hace
Biología y genética
New Phytologist · texto original
<fc>PELOTA</fc> at the crossroads of tomato fertility and viral resistance
PELOTA (PELO), a conserved ribosome surveillance factor underlying the tomato resistance allele ty-5, plays key roles in TYLCV susceptibility and reproductive development. Complete loss of PELO funct...
Tomato yellow leaf curl virus (TYLCV) is a major threat to tomato (Solanum lycopersicum) production, and durable resistance remains challenging to achieve.
We show that complete loss of function of PELOTA (PELO), a conserved ribosome surveillance factor underlying the resistance allele ty-5, confers near-immunity to TYLCV.
Unexpectedly, PELO is also required for reproductive development; pelo mutants display severe floral defects, impaired male and female fertility, and complete sterility, revealing a previously unknown role for PELO in tomato plant fertility.
Using CRISPR-mediated promoter editing, we generated novel PELO alleles with reduced expression that maintain fertility while conferring enhanced resistance to TYLCV, providing a new strategy.
Introduction
Tomato yellow leaf curl virus (TYLCV) is the causal agent of one of the most devastating diseases affecting tomato cultivation world-wide. It was first identified in the eastern Mediterranean basin and has since spread across tropical and subtropical regions, largely due to human-mediated factors (Lefeuvre et al., 2010). The virus is transmitted by whiteflies and causes severe symptoms, including apical leaf cupping, chlorosis, stunted growth, and flower abortion, resulting in considerable yield losses (Picó et al., 1996; Czosnek & Laterrot, 1997; Lapidot et al., 1997, 2008; Czosnek, 2007; Scholthof et al., 2011). TYLCV is a monopartite begomovirus belonging to the family Geminiviridae, with a genome consisting of a single circular c. 2.8 kb single-stranded DNA molecule. Its transmission depends on the sweet potato whitefly (Bemisia tabaci), and increases in whitefly populations are often associated with disease outbreaks (Gafni, 2003; Gronenborn, 2007; Hanley-Bowdoin et al., 2013). In fields containing TYLCV-infected plants, once whitefly populations reach high levels, controlling disease spread becomes extremely difficult. Therefore, the development of disease-resistant tomato cultivars is vital to mitigate viral impact. Wild tomato species such as Solanum cheesmaniae, S. arcanum, S. chilense, S. corneliomulleri, S. chmielewskii, S. galapagense, S. pennellii, S. habrochaites, S. pimpinellifolium, and S. peruvianum do not exhibit disease symptoms upon TYLCV infection (Ji et al., 2007; Vidavski et al., 2008; de la Pena et al., 2010; Pereira-Carvalho et al., 2010; Tomás et al., 2011). Therefore, they have been used as sources of resistance, which have been introgressed into commercial tomato cultivars. To date, five loci conferring varying levels of resistance to TYLCV have been identified, often acting via mechanisms such as inhibition of viral replication or restriction of virus movement: Ty-1/Ty-3 (Zamir et al., 1994; Ji et al., 2007; Verlaan et al., 2013), Ty-2 (Hanson et al., 2006), Ty-4 (Ji et al., 2009), ty-5 (Anbinder et al., 2009; Hutton et al., 2012), and Ty-6 (Scott et al., 2015; Gill et al., 2019). Ty-1 and Ty-3, alleles of the same gene introgressed from S. chilense, encode an RNA-dependent RNA polymerase (RDR) (Zamir et al., 1994; Ji et al., 2007; Verlaan et al., 2013). Ty-2, derived from S. habrochaites, encodes an NB-LRR protein (Yang et al., 2014; Yamaguchi et al., 2018). ty-5, originating probably from S. peruvianum, encodes PELOTA (PELO), a homolog of a messenger RNA surveillance factor (Anbinder et al., 2009; Lapidot et al., 2015). In addition, Ty-6, encoding DNA polymerase delta subunit 1, and Ty-4 were identified from S. chilense (Ji et al., 2009; Scott et al., 2015; Gill et al., 2019; Shen et al., 2025). Although such resistance genes have been introgressed into commercial cultivars, none alone provides high resistance against begomoviruses (Mejia et al., 2005; Vidavski et al., 2008; Anbinder et al., 2009; Hutton et al., 2015). Moreover, tomato varieties resistant to TYLCV usually show some level of symptoms and can still harbor the virus, generating strong selective pressure that enables TYLCV to overcome Ty-1/Ty-3-mediated resistance in the field (Belabess et al., 2016; Nakajima et al., 2026). In addition, breakdown of Ty-1/Ty-3-mediated resistance has been observed under co-infection with TYLCV and a DNA β satellite or TYLCV and Tomato chlorosis virus (ToCV) (Anfoka et al., 2014; Gelbart et al., 2020; Voorburg et al., 2020; Fortes et al., 2023). Similarly, an outbreak of Tomato curly stunt virus (ToCSV), which causes TYLCV-like symptoms, led to the breakdown of Ty-1/Ty-3 resistance in commercial tomato fields in South Africa (Pietersen et al., 2008). These observations indicate that effective and durable control of TYLCV will likely require the combination of multiple resistance genes. Pyramiding resistance genes is an effective strategy for improving resistance to begomoviruses (Vidavski et al., 2008; Anbinder et al., 2009; Hutton et al., 2015; Lapidot et al., 2015; Jaganathan et al., 2025; Shimoide et al., 2026). However, pyramiding multiple resistance genes can lead to trade-offs, as several studies have reported that introgressions from wild relatives are often associated with yield penalties, as well as reduced fruit size and quality (Lapidot et al., 1997; Rubio et al., 2012; García-Martínez et al., 2014, 2016; Cabrera et al., 2024).
The ty-5 locus, one of the strongest sources of TYLCV resistance, maps to a single valine-to-glycine (V16G) substitution in the PELO protein, which compromises its function and is consistent with the recessive nature of ty-5 (Vidavski et al., 2008; Anbinder et al., 2009; Hutton et al., 2012, 2015; Lapidot et al., 2015; Ren et al., 2022). Silencing of PELO using RNAi in susceptible plants confers resistance to TYLCV, while silencing PELO in resistant plants further increases resistance (Lapidot et al., 2015). Although considerable progress has been made in understanding PELO function in different species, the mechanism by which this recessive mutation confers virus resistance in tomato remains unknown. Given that PELO is an essential component of ribosome recycling during protein translation in plants, it is plausible that PELO is required for efficient viral protein synthesis. Consistent with this, PELO mutants also confer resistance to viruses in other crops (Li et al., 2025; Majumdar et al., 2026), including pepper yellow leaf curl Indonesia virus (PepYLCIV) and pepper yellow leaf curl Aceh virus (PepYLCAV) in Capsicum annuum, as well as bacterial blight and fijivirus southern rice black-streaked dwarf virus (SRBSDV) in rice (Ding et al., 2018; Zhang et al., 2018; Koeda et al., 2021; Sun et al., 2024). Moreover, knockdown of PELO in Nicotiana benthamiana and Arabidopsis thaliana confers resistance to geminiviruses (Ren et al., 2022; Onouchi et al., 2025). In contrast to its role as a susceptibility factor for geminiviruses, PELO contributes to resistance against potyviruses, and reduced PELO activity in Arabidopsis increases susceptibility to potyvirus infection (Ge et al., 2023, 2025). Thus, the role of PELO in plant-virus interactions appears to differ among viral families.
In this study, we generated pelo loss-of-function mutants that eliminate virus symptoms and viral accumulation. However, these mutations also cause defects in floral organs and a lack of fruit production, resulting in plant sterility. These findings indicate that PELO is required for both TYLCV infection and plant fertility, positioning it as a key factor in both processes. In an effort to overcome this trade-off between resistance and fertility, we performed CRISPR-Cas9-based promoter bashing (Rodriguez-Leal et al., 2017) and demonstrated that it is possible to mimic and potentially improve the native resistance allele, which was likely selected during evolution, to achieve a balance between resistance and fertility.
Materials and Methods
Genetic material, DNA constructs, and plant genetic transformation
Tomato (Solanum lycopersicum L.) M82, Rehovot 13 (R13), and ty-5 were described previously (Lapidot et al., 2015). The ty-5 line used throughout this study was backcrossed to M82 five times (BC5F6).
DNA constructs were generated using the Golden Gate and GoldenBraid cloning systems (Werner et al., 2012; Sarrion-Perdigones et al., 2013; Vazquez-Vilar et al., 2015).
For PELO promoter-reporter analysis, the PELO promoter (PELOpro, Solyc04g009810) was cloned into the entry vector pUPD2 (GoldenBraid) and subsequently used to assemble expression cassettes according to GoldenBraid guidelines (Sarrion-Perdigones et al., 2013; Vazquez-Vilar et al., 2015). A 2935 bp genomic fragment containing the PELO promoter region, including the 5′ untranslated region (UTR), was amplified by PCR from M82 genomic DNA using primers listed in Supporting Information Table S1 and cloned into pUPD2. The GUS-mCitrine reporter was PCR-amplified from the SlPIF8a:GUS-mCitrine construct (Kunta et al., 2025) using primers listed in Table S1 and cloned into pUPD2. The construct α2_PELOpro:GUS-mCitrine was generated by assembling PELOpro, GUS-mCitrine, and the nos terminator (nosT) (all in pUPD2) into pDGBα2 using BsaI. This assembly was subsequently combined with α1R_NOSpro:NPTII and cloned into the binary vector pDGBΩ1 using BsmBI to generate the final expression cassette Ω1_PELOpro:GUS-mCitrine_NPTII.
The CRISPR-Cas9 lines were generated as described (Brooks et al., 2014; Kunta et al., 2025). Two guide RNA (gRNA) target sites, each with at least four mismatches to any potential off-target sites, were selected within the PELO genomic sequence using CRISPR-P and the tomato genome SL2.50 (Liu et al., 2017) (Fig. 1a). To generate a binary vector containing a CRISPR cassette with 35Spro:Cas9 and the gRNAs, gRNA1 and gRNA2 were cloned into the Level 1 vectors pICH47751 (#48002; Addgene) and pICH47761 (#48003; Addgene), respectively, by PCR using a primer containing the target sequence and a universal primer (Table S1), with pICH86966::AtU6pro:gRNA_PDS (#46966; Addgene) as the template. The resulting gRNA constructs were subsequently combined into a Level 2 vector together with the Level 1 vectors pICH47732-NOSpro::NPTII (#51144; Addgene), pICH47742-35Spro:Cas9 (#49771; Addgene), and pICH41780 (#48019; Addgene), which were assembled into the binary Level 2 vector pAGM4723 (#48015; Addgene) using BpiI, to generate the final CRISPR-Cas9 binary construct, 35Spro:Cas9_g1_g2_NPTII.
CRISPR-edited pelo mutants are highly resistant to TYLCV. (a) Sequences of wild-type M82 (WT, top) and the two CRISPR-generated pelo mutant alleles used in this study (pelo-1CR and pelo-2CR, middle), along with the corresponding changes in the mutant protein sequences (bottom), are shown. The gRNA1 and gRNA2 binding sites, as well as the deletion (red dashed lines) or insertion (red letter), are indicated within the gene sequence. Both alleles harbor frameshift mutations within the first exon, proximal to the first exon-intron junction, resulting in the introduction of premature termination codons (red asterisks). Uppercase and lowercase letters represent exons and introns, respectively. Dotted lines represent gaps in the sequence, with the number of missing nucleotides or amino acids specified. Blue amino acids denote the WT sequence, whereas red amino acids indicate the predicted changes in the protein sequence resulting from the CRISPR-mediated edits. (b, c) Representative side (b) and top (c) views of tomato plants infected with TYLCV for wild-type M82 (WT), ty-5, pelo-1CR, and pelo-2CR mutants. Pictures were taken 35 d post-inoculation (DPI). Scale bars in (b, c) represent 5 and 2.5 cm, respectively. (d, e) Average viral DNA copy number determined by quantitative real-time PCR (qPCR) relative to the reference 18S ribosomal DNA (2−ΔCt, d) and disease severity index (DSI; e) are shown. Tomato plants of the indicated genotypes were inoculated with TYLCV using viruliferous whiteflies, and infectivity was assessed at 28 DPI. Data are presented as means ± SE from four to five biological replicates per genotype in (d) and at least eight biological replicates per genotype in (e). Gray dots indicate the individual data points. Due to the presence of many near-zero values in pelo mutants, resulting in a non-normal distribution, statistical significance in (d, e) was assessed using the Kruskal–Wallis test followed by two-sided pairwise Wilcoxon rank-sum tests with Benjamini–Hochberg correction. Different letters denote statistically significant differences between samples (P < 0.05).
For the PELO promoter-bashing lines, eight gRNA target sites were selected within the 3000 bp upstream region of the PELO start codon using CRISPR-P and the tomato genome SL2.50 (Liu et al., 2017), based on their predicted efficiency and off-target scores, while ensuring their distribution across the entire promoter region for comprehensive promoter bashing. The gRNAs were cloned into Level 1 plasmids using primers listed in Table S1. gRNA1 and gRNA8 were cloned into pICH47751 (#48002; Addgene), gRNA2 was cloned into pICH47761 (#48003; Addgene), gRNA3 was cloned into pICH47772 (#48004; Addgene), gRNA4 was cloned into pICH47781 (#48005; Addgene), gRNA5 was cloned into pICH47791 (#48006; Addgene), gRNA6 was cloned into pICH47732 (#48000; Addgene), and gRNA7 was cloned into pICH47742 (#48001; Addgene). All eight gRNAs were then assembled into two intermediate Level M vectors, pAGM8055 (#48039; Addgene) with pICH50927 (#48049; Addgene), and pAGM8093 (#48043; Addgene) with pICH50892 (#48046; Addgene), using the BpiI enzyme. The level 1 and Level M vectors pICH47732-NOSpro::NPTII (#51144; Addgene), pICH47742-35Spro:Cas9 (#49771; Addgene), pAGM8055 (#48039; Addgene), pAGM8093 (#48043; Addgene), and pICH41766 (#48018; Addgene) were then assembled into the binary Level 2 vector pAGM4723 (#48015; Addgene), using BpiI and BsaI, yielding the final binary plasmids containing all eight gRNAs, 35Spro:Cas9, and NOSpro::NPTII.
To generate the CRISPR Knockout-rescue (CKR) system expressing 35Spro:Cas9, two gRNAs targeting the PELO genomic region, and SlUbq10pro:PELOCDS, Golden Gate and GoldenBraid cloning systems were used. The PELO coding region (CDS, Solyc04g009810) was amplified by PCR from M82 cDNA using primers listed in Table S1 and cloned into pUPD2. The tomato Ubiquitin10 promoter, including the 5′ UTR (SlUbq10pro; Solyc07g064130), was amplified from pUPD_pro-SlUBIQ (Kumar et al., 2022) using primers listed in Table S1 and cloned into pUPD2. Both PELO CDS and SlUbq10pro were domesticated into the GoldenBraid system by removal of BsaI and BbsI restriction sites according to guidelines (Sarrion-Perdigones et al., 2013; Vazquez-Vilar et al., 2015). The α2_SlUbq10pro:PELOCDS construct was generated by assembling the pUPD2 plasmids SlUbq10pro, PELOCDS, and the nosT into pDGBα2 using BsaI. To combine the Golden Gate-based CRISPR-Cas9 constructs with the GoldenBraid-based SlUbq10pro:PELOCDS, the plasmids pICH47761 (#48003; Addgene) and pAGM4723 (#48015; Addgene) were modified using site-directed mutagenesis with primers listed in Table S1 and DpnI treatment before transforming E. coli. gRNA2 was subsequently recloned into Level 1 pICH47761_CKR (compatible with the CKR system), and the Level 1 pICH47751 (gRNA1), pICH47761_CKR (gRNA2), pICH47732-NOSpro::NPTII (#51144; Addgene), pICH47742-35Spro:Cas9 (#49771; Addgene), and α2_SlUbq10pro:PELOCDS were then assembled into the modified binary Level 2 vector pAGM4723_CKR (compatible with the CKR system) using BpiI and BsaI to obtain the final binary plasmid pAGM4723_35proCAS9_g1_g2_SlUbq10pro:PELOCDS_NPTII.
All constructs were introduced into Agrobacterium tumefaciens strain GV3101 and transformed into tomato M82 (sp) using the cotyledon transformation method (McCormick, 1997). The 35Spro:Cas9_g1_g2_NPTII construct was also transformed into R13 (SP). Segregation analysis of antibiotic resistance was used to isolate single-copy homozygous transgenic lines, and specific primers for the Cas9 sequence were used to identify pelo mutants, with or without the CRISPR cassette or the CRISPR SlUbq10pro:PELOCDS cassette (Table S1).
Imaging, microscopy, and GUS staining
GUS (β-glucuronidase) staining was performed as described (Ori et al., 2000). Plant tissue was vacuum-infiltrated for 30 min in a staining solution (25 mM phosphate buffer, pH 7, 0.25% Triton X-100, 1.25 mM potassium ferricyanide, 1.25 mM potassium ferrocyanide, 0.25 mM EDTA, 1 mg ml−1 5-bromo-4-chloro-3-indolyl β-D-glucuronide (X-Gluc)) and then incubated overnight at 37°C. The tissue was then cleared in 100% ethanol, gradually rehydrated to 50% ethanol through sequential washes, and finally transferred to 50% glycerol.
Scanning electron microscopy (SEM) was performed using a JEOL JCM-6000 instrument equipped with a GSD detector. Before imaging, floral organs were dehydrated or directly dissected, mounted on SEM stubs, and coated with palladium.
Confocal microscopy images were acquired using the Leica SP8 CLSM system at the Volcani Institute microscopy core facility. Transverse sections of anthers and ovary from pelo, PELOpro:GUS-mCitrine, and wild-type plants were stained with calcofluor white M2R (fluorescent brightener 28, FB; cat. no. 4359), and washed twice with 1× phosphate-buffered saline (PBS). Wild-type and pelo also stained with propidium iodide (PI; cat. no. J66584.AB), followed by two washes in 1× PBS. Samples were then mounted on 0.8% (w/v) agarose in small Petri dishes. Images were acquired using excitation/detection settings of 405 nm/425–520 nm for FB, 515 nm/522–580 nm for mCitrine, 568 nm/580–692 nm for PI, and 568 nm/601–696 nm for Chl autofluorescence. Emissions were collected using HyD detectors, and fluorescence lifetime imaging microscopy (FLIM) was routinely used to distinguish mCitrine fluorescence from autofluorescence.
Whitefly maintenance, plant inoculation, and disease severity scoring
Whitefly colonies (Bemisia tabaci, biotype B), reared on cotton plants (Gossypium hirsutum), were used for TYLCV (GenBank accession no. X15656) inoculation as described before (Lapidot et al., 1997, 2015; Lapidot, 2007). In short, young seedlings (2 wk post germination) were exposed to viruliferous whiteflies for 24 h, then sprayed with imidacloprid (Bayer, Leverkusen, Germany) and maintained in an insect-proof net house at 26–32°C. TYLCV-induced disease symptoms were evaluated for each plant separately at 28 d post-inoculation (DPI) according to a disease severity index (DSI) of 0 to 4, where 0 indicates no symptoms and 4 indicates severe infection (Friedmann et al., 1998; Lapidot & Friedmann, 2002).
Grafting experiment and plant inoculation using pelo mutants
Five-wk-old infected wild-type, pelo-1CR, and pelo-2CR plants at 21 DPI were used as virus sources for whitefly-mediated TYLCV transmission. Whiteflies that had fed on a single source plant for 24 h (acquisition) were transferred to three or four healthy wild-type test plants for 24 h (transmission). The inoculated test plants were then sprayed with imidacloprid (Bayer) and maintained in an insect-proof net house. One week after the whiteflies were removed, the same virus-source plants, together with additional infected source plants, were used for grafting experiments. The different graft combinations were generated using a V-shaped grafting method.
Tissue collection and RNA analysis
For quantitative RT-qPCR, RNA was extracted from the first leaves of 14-d-old tomato seedlings (two to three leaves per biological replicate) or from the apices of infected plants, using the Nucleospin RNA Plant Kit (Macherey-Nagel, Duren, Germany). For the analysis of plant genes, cDNA was synthesized from 400 ng of RNA using the Maxima first-strand cDNA synthesis kit (Thermo Fisher Scientific, Waltham, MA, USA). For the analysis of the viral gene, cDNA was synthesized using the Revert-Aid first-strand cDNA synthesis kit (Fermentas Life Sciences, Vilnius, Lithuania). RT-qPCR analysis was carried out using StepOnePlus Real-Time PCR Systems (ThermoFisher, Waltham, MA, USA), with qPCRBIO SyGreen Blue Mix Hi-ROX (PCR Biosystems, London, UK). Transcript levels were normalized to EXPRESS (Solyc07g025390) as an internal control for plant genes and to 18S ribosomal RNA for the viral gene. The primers used for the RT-qPCR analysis are listed in Table S1.
Quantification of TYLCV DNA in the infected plants
TYLCV DNA copy number in plants was determined using quantitative real-time PCR (qPCR). Total DNA was extracted from plant apices and diluted to 0.1 ng μl−1 before use. Individual plants, each analyzed with two technical replicates, were used for analysis, and 18S ribosomal DNA was used as a reference using primers listed in Table S1.
Cycle threshold (Ct) values were determined using a StepOnePlus real-time PCR system (Thermo Fisher Scientific) with qPCRBIO SyGreen Blue Mix Hi-ROX (PCR Biosystems). ΔCt was calculated by subtracting the Ct of the reference gene (18S) from the Ct of the virus. TYLCV DNA accumulation was plotted as 2−ΔCT.
Pollen germination and viability assays
For in vitro pollen germination assays, pre-anthesis flowers were marked 1 d before analysis. Anther cones were collected from three to four flowers per genotype, cut transversely, and immersed in pollen germination medium (2 mM boric acid, 2 mM calcium nitrate, 2 mM magnesium sulfate, and 1 mM potassium nitrate, supplemented with 10% w/v sucrose) as described before (Rutley et al., 2021). The anther cones were vortexed to release pollen into the solution. Pollen grains were then plated onto microscope slides coated with a thin layer of solidified germination medium containing 2% agarose and incubated at room temperature for 2 h in a closed, humid chamber. Germinated pollen grains were imaged using a microscope and quantified manually using ImageJ.
For pollen viability, Alexander staining was performed using five to six flowers per genotype. Another cone was dissected and placed onto a microscope slide containing a drop of Alexander stain. Slides were incubated in the dark until imaging. Pollen grains were visualized under a microscope and quantified using ImageJ.
Results
The pelo loss-of-function mutant is highly resistant to TYLCV
The ty-5 locus has been shown to play an important role in resistance against TYLCV. This resistance is present only in plants homozygous for the recessive ty-5 allele, which reduces viral load and symptom severity. However, because the virus can still propagate in ty-5 plants and mild symptoms are still observed, we hypothesized that a loss-of-function mutation in the tomato PELO gene, corresponding to the ty-5 locus, may confer enhanced resistance to TYLCV. Supporting this hypothesis, we have shown that RNAi-mediated silencing of PELO in wild-type plants resulted in stronger resistance than in ty-5 plants, as evidenced by reduced viral copy number and a complete lack of disease symptoms (Lapidot et al., 2015).
To generate loss-of-function pelo alleles that will enhance resistance relative to ty-5, we first attempted to produce transgene-free Cas9 mutants, lacking the inserted cassette encoding Cas9 and the gRNAs, in the R13 tomato cultivar. In R13, 23 T0 transgenic plants were generated. Among these Cas9-positive plants, most of them were sterile; only three (one homozygous and two hemizygous) produced a small number of parthenocarpic fruits (Table S2). Interestingly, no heterozygous Cas9-positive plants were found in the R13 background. Cuttings were obtained from six representative T0pelo mutant lines generated in the R13 background and inoculated with TYLCV using viruliferous whiteflies. Cuttings obtained from non-transgenic R13 plants served as controls. Because none of the pelo mutant lines developed symptoms, they were reinoculated 90 d after the first inoculation. Twenty-eight days after the second inoculation, all inoculated T0pelo mutant cuttings remained symptomless, tested negative for TYLCV DNA, and were positive for the Cas9 cassette. By contrast, the inoculated R13 control plants developed severe disease symptoms after the first inoculation, tested positive for TYLCV DNA, and were negative for the Cas9 cassette (Fig. S1). Since all T0 loss-of-function mutants were sterile, limiting propagation and making subsequent genetic analyses of virus resistance in this cultivar difficult, we decided to generate transgene-free Cas9 mutants in the M82 cultivar. In M82, we obtained a total of 19 T0 plants, of which 14 were heterozygous, and 5 were homozygous for the edited allele in the PELO gene coding sequence (CDS; Table S3). As in R13, all homozygous T0 mutant plants carrying early stop codons exhibited severe fertility defects, including the production of few or no parthenocarpic fruits. By contrast, the 14 heterozygous M82 plants were fertile.
To further investigate the role of PELO in TYLCV resistance, we selected two segregating mutant lines: pelo-1CR (one T deletion in exon 1 and a five base pair deletion in intron 1) and pelo-2CR (one T insertion in exon 1 and a two base pair deletion in intron 1) (Fig. 1a). Seedlings of pelo-1CR, pelo-2CR, wild-type, and ty-5 plants were inoculated with TYLCV using whiteflies, and disease severity was quantified using a DSI ranging from 0 to 4, where 4 indicates severe symptoms and 0 indicates no or minimal symptoms (Friedmann et al., 1998; Lapidot & Friedmann, 2002). In wild-type plants, disease symptoms peaked at 28 d post-infection (DPI) (DSI 3.3), with severe leaf curling and yellowing of leaf margins primarily in the shoot apex. By contrast, pelo-1CR and pelo-2CR mutants exhibited no symptoms (DSI 0.2), while ty-5 plants showed intermediate symptom severity (DSI 2.2; Fig. 1b–e). To determine whether reduced symptoms correlated with viral load, four to five plants were randomly selected from each group, and viral DNA accumulation was quantified at 28 DPI. While ty-5 plants exhibited significantly lower viral levels than wild-type plants, pelo mutants showed nearly undetectable viral levels (Fig. 1d). To better determine whether the pelo mutants that showed no disease symptoms and undetectable viral accumulation were indeed immune to TYLCV, or instead exhibited very strong resistance (near-immunity), with no symptoms and very low viral accumulation, we used inoculated pelo mutant plants as virus source in whitefly transmission and grafting experiments. For whitefly transmission, each mutant source plant was used to inoculate three healthy wild-type plants. In the grafting experiments, infected plants served as rootstocks for healthy wild-type scions. All 14 inoculated pelo mutant plants tested remained symptomless, although three (pelo-1CR #28, pelo-2CR #34 and #36) contained very low levels of both viral DNA and RNA (Fig. 2a–c). Nine of these plants were selected as virus source for whitefly transmission or grafting experiments. As expected, all wild-type plants inoculated by whiteflies that fed on infected wild-type sources acquired the virus and developed severe disease symptoms (Fig. 2d–f). By contrast, only two mutant source plants, pelo-2CR #34 and pelo-2CR #36, transmitted the virus by whiteflies, each to one of three wild-type recipients. Overall, 2 of the 24 wild-type plants inoculated from pelo mutant sources developed symptoms and accumulated viral DNA (Fig. 2e,f). In the grafting experiments, all wild-type scions grafted onto infected wild-type rootstocks became infected. Transmission from pelo mutant rootstocks occurred only when the source plants contained detectable, albeit very low, viral levels (Fig. 2g–i). Together, these results show that pelo mutations strongly restrict viral accumulation and confer strong resistance, but do not provide complete immunity, as even very low viral levels can be transmitted to healthy plants.
Loss of PELO function in tomato strongly restricts TYLCV accumulation. (a–c) Disease severity and viral accumulation in TYLCV-infected ty-5, wild-type (WT), pelo-1CR, and pelo-2CR plants, measured 21 d after inoculation. (a) Disease severity index (DSI), (b) relative viral DNA accumulation normalized to 18S ribosomal DNA (2−ΔCt), and (c) viral C1 gene expression normalized to 18S ribosomal RNA (2−ΔCt). In (b, c), black arrows indicate the source plants in which low viral accumulation was detected. Uninfected plants served as negative controls. (d) Schematic representation of the secondary-transmission experiment. Individual infected WT, pelo-1CR, and pelo-2CR plants were used as virus sources for whitefly acquisition feeding and subsequent TYLCV transmission to inoculate healthy WT plants. Colors identify the individual source plants and their corresponding recipient plants and are consistent across panels. (e, f) Disease severity and viral accumulation in WT recipient plants, determined 28 d following inoculation, from the indicated source genotypes. (e) DSI and (f) relative viral DNA accumulation normalized to 18S ribosomal DNA (2−ΔCt). In (a–c, e, f) boxes represent the interquartile range, horizontal lines indicate the median, whiskers show the minimum and maximum values, and dots represent individual plants. (g) Schematic representation of the grafting experiment. Infected WT, pelo-1CR, and pelo-2CR plants were used as rootstocks, and uninfected plants were used as scions. Colors identify the individual source plants used as rootstocks. (h, i) Disease severity and viral accumulation in the scions were assessed at 28 d after grafting. (h) Disease severity index (DSI), (i) relative viral DNA accumulation normalized to 18S ribosomal DNA (2−ΔCt). In (h, i), data are presented as means ± SE from 4 to 5 independent grafts per group. Colored dots represent individual grafts, with colors identifying the corresponding rootstock plants. In (e, f, h, i), red arrows indicate WT recipient plants that developed TYLCV symptoms and showed viral accumulation when pelo-1CR #28, pelo-2CR #34, or pelo-2CR #36 was used as the virus source. Due to the presence of many near-zero values in pelo mutants, resulting in a non-normal distribution, statistical significance in (a–c, e, f, h, i) was assessed using the Kruskal–Wallis test followed by two-sided pairwise Wilcoxon rank-sum tests with Benjamini–Hochberg correction. Different letters denote statistically significant differences between samples (P < 0.05). ns, not significant.
Since a previous study reported that pelo mutant plants do not exhibit any growth penalty under their experimental conditions (Pramanik et al., 2021), the next step was to confirm that both the resistance and sterility phenotypes observed in our pelo plants are caused by a loss-of-function mutation in PELO. We performed a rescue experiment in which we combined, in a single action, CRISPR-Cas9-mediated inactivation of the endogenous PELO gene with the expression of a PELO cDNA CDS driven by the constitutive SlUbq10 promoter. The mutagenic construct contained the Cas9 system under the control of the constitutive SlUbq10 promoter, along with two gRNAs designed to target the coding regions (exons) of the endogenous PELO, but not the introduced PELO cDNA (Fig. S2a). This specificity was achieved by designing the gRNAs to target the exon 1 intron 2 junctions, with the protospacer adjacent motif sequences located within the first intron (Fig. S2b).
Using this approach, which we named the CKR system, we obtained fertile T1 plants homozygous for the pelo mutation (Fig. S2c) that also carried the SlUbq10p:PELOCDS cassette, thereby confirming rescue of the sterility phenotype through overexpression of the PELO cDNA.
Next, we inoculated T2 plants homozygous for pelo and segregating for the SlUbq10p:PELO cassette with TYLCV and scored their DSI. Notably, pelo mutant plants carrying the SlUbq10p:PELO cassette were sensitive to TYLCV, similar to the wild-type control (Fig. S2d–f). By contrast, segregating plants that were homozygous for pelo but lacked the SlUbq10p:PELO cassette showed very strong resistance to the virus (Fig. S2d–g). These results confirm that, as for the sterility phenotype, resistance to TYLCV is also caused by loss of function of the PELO gene.
Collectively, these results indicate that PELO is required for TYLCV accumulation in the plant, as loss-of-function pelo mutants are highly resistant to infection with no disease symptoms or virus detected. At the same time, loss of PELO function markedly impairs plant fertility and development, consistent with its essential role in ribosome recycling and the dissociation of stalled ribosomes during the late stages of translation (Lee et al., 2007; Passos et al., 2009; Shoemaker et al., 2010; Becker et al., 2011; Majumdar et al., 2026; Sans-Coll & Merchante, 2026).
PELO plays a role in flower development and plant fertility
The heterozygous pelo mutant enabled us to investigate the role of PELO in plant growth and development throughout the plant life cycle. Because homozygous pelo mutants were sterile, we sowed T1 or T2 seeds derived from heterozygous pelo mutant plants. As no clear visible phenotype was detected among segregating seedlings, we performed PCR followed by sequencing to identify pelo mutant plants.
After identifying homozygous, heterozygous, and wild-type plants, we observed that some, but not all, pelo-1CR mutant seedlings were slightly smaller than wild-type seedlings. This phenotype was not observed in the pelo-2CR allele (Fig. 3a–c). In mature plants, both pelo-1CR and pelo-2CR mutants exhibited a bushy and compact growth habit, chlorotic leaves, malformed flowers, and the absence of fruit, compared with wild-type plants, which had already set fruit at the same growth stage (Fig. 3d–i). Although some pelo mutants eventually produced fruits, all fruits that developed were parthenocarpic (Fig. 3j–l).
Loss of PELO function results in sterile tomato plants. (a–f) Phenotypes of young (a–c) and mature (d–f) wild-type M82 (WT), pelo-1CR, and pelo-2CR mutant plants. At the mature stage, numerous fruits were present on the WT plant, whereas only a few fruits were occasionally observed on mutant plants. Bars in (a–c) and (d–f) represent 2.5 cm and 10 cm, respectively. (g–l) Representative images of leaves (g–i) and fruits (j–l) from WT, pelo-1CR, and pelo-2CR mutant plants. Mutant plants rarely produce fruits, and the fruits are seedless, as shown in (k, l). Bars in (g–i) and (j–l) represent 3 and 1 cm, respectively.
To understand the cause of sterility and parthenocarpic fruit formation in pelo mutant alleles, we closely examined the reproductive organs. While in wild-type the anthers form a tightly fused, symmetrical cone, this structure is disrupted in the pelo mutants. Both pelo-1CR and pelo-2CR exhibit loosely associated, partially unfused, and irregularly arranged anthers; notably, pelo-1CR forms a short, green anther cone, whereas pelo-2CR displays a wider, yellow anther cone (Fig. 4a–c). Transverse sections of mutant anthers revealed loose connections between adjacent anthers, resulting in inter-anther spaces, along with an increased number of cone-shaped anthers (Fig. 4a–f). Moreover, anther locules in the mutants exhibited a shrunken and disorganized morphology relative to the wild-type and contained abnormal pollen grains in both shape and size (Fig. 4d–i), as also observed by SEM (Fig. 4j–l). In addition, whereas in wild-type plants numerous pollen grains packed the anther locules, in the mutants they were localized at the periphery (Fig. 4d–f). Consistent with these developmental events, pollen viability and germination assays showed that mutant pollen was non-viable and failed to germinate (Fig. 4m–q).
Tomato pelo mutants display severe flaws in anther and pollen development, as well as pollen viability. (a–c) Confocal images of anther cross-sections stained with propidium iodide (red) and fluorescence brightener (blue) from wild-type M82 (WT), pelo-1CR, and pelo-2CR mutant flowers. Bar, 500 μm. Insets show the entire anther structure. Bar, 100 mm. (d–f) Higher-magnification images of the sections shown in (a–c). Bar, 200 μm. (g–i) Higher magnification of the pollen within the anther cross-sections shown in (a–c). White arrows indicate abnormal pollen. Bar, 50 μm. (j–l) Scanning electron microscopy (SEM) images of pollen grains from WT and mutant flowers. White arrows indicate abnormal pollen. Bar, 20 μm. (m–o) Microscopic analysis of pollen viability using Alexander's stain, showing reduced pollen viability in the pelo mutant compared with the WT. Scale bar: 100 μm. (p, q) Quantification of the percentage of germinated (p) and viable (q) pollen in WT and mutant lines. Mean values from six biological replicates per genotype ± SE are presented. Gray dots indicate the individual data points. Asterisks indicate statistically significant differences compared with WT (P < 0.001), as determined by the Wilcoxon rank-sum test.
The female reproductive organ (pistil) also showed abnormalities in both mutants. pelo-1CR exhibited a short, straight style with a flattened ovary, whereas pelo-2CR showed a tilted style with a morphologically normal ovary (Fig. 5a–c). Surface structure analysis of mutant stigmas using SEM revealed a flattened morphology lacking distinct lobes, which are typically observed in wild-type plants (Fig. 5d–i). Transverse sections of the ovaries revealed no observable differences in ovule morphology between mutant and wild-type plants (Fig. 5j–l).
Tomato pelo mutants display abnormal style and stigma. (a–c) Representative images of the pistil from wild-type M82 (WT, a), and abnormal pelo mutant (b, c) flowers. Bar, 0.5 cm. (d–i) Scanning electron microscopy (SEM) images of stigma from WT flowers (d, g) and abnormal pelo mutant flowers (e, f, h, i). Bar, 200 μm. (j–l) SEM images of ovary cross-sections from WT (j) and the abnormal pelo mutant flowers (k, l). Bar, 500 μm. (m) Percentage of successful crosses using WT pollen with the indicated emasculated flowers (listed to the left of each cross). Percentages were calculated from 20 crosses per genotype.
While the severe pollen defects and lack of pollen viability clearly demonstrated male sterility, the relatively mild phenotypes observed in the female reproductive organs prompted us to investigate whether the pelo mutant flowers also exhibit female sterility. To assess this, we pollinated mutant flowers with wild-type pollen. Pollination of pelo-2CR flowers resulted in fruit formation in only 30% of the crosses, whereas pelo-1CR flowers failed to produce any fruit. By contrast, pollination of wild-type flowers with wild-type pollen resulted in fruit formation in 85% of the crosses (Fig. 5m).
Together, these results suggest that PELO is required for proper reproductive organ development, and that defects in both male and female reproductive structures likely contribute to sterility in the mutants. These findings are consistent with previous studies in Drosophila and Saccharomyces cerevisiae, where PELO regulates germline development and sporulation, respectively, through control of meiotic and mitotic divisions (Eberhart & Wasserman, 1995; Davis & Engebrecht, 1998).
PELO is expressed in vegetative and reproductive organs in tomato
Based on the observation that the mutants exhibit abnormal flower development affecting the anthers, pollen, and stigma, we identified that in addition to its role in TYLCV resistance, PELO plays a key role in plant fertility. We therefore asked whether PELO is expressed in the vegetative and reproductive organs. To address this, we generated a reporter line, PELOpro:GUS-mCit, expressing β-glucuronidase (GUS) fused to mCitrine under the control of the native PELO promoter (2935 bp upstream of the ATG), and monitored its expression.
By following the expression pattern of the mCitrine reporter, we found that in wild-type plants, PELO is expressed in the pollen, as well as in the ovary, ovules, style, and stigma (Fig. 6a–f). All of these organs exhibit abnormalities in the pelo mutants (Figs 4, 5), further supporting a role for PELO in the development of reproductive organs. GUS staining revealed a tissue-specific expression pattern of PELO in young and mature buds, as well as in the petals and at the base of the sepals (Fig. 6g,h). In addition, PELO expression in leaves was primarily detected in young leaflets and was less pronounced in fully expanded mature leaflets (Fig. 6i,j). This correlates with the observation that whiteflies preferentially feed and infect young leaves rather than mature leaves. Interestingly, although PELO is expressed in leaves, loss of PELO has a relatively minor effect on leaf phenotype, resulting in chlorosis of mature leaves (Fig. 3g–i), compared to its pronounced impact on flower development.
PELO expression in tomato flowers and leaves. (a–f) Confocal images of tomato flower cross-sections stained with propidium iodide (red) and fluorescence brightener (blue), showing in the top panel the expression of the GUS-mCitrine reporter driven by the PELO promoter (PELOpro:mCit, green), and in the lower panel the wild-type M82 (WT) control, in the pollen (a, b), ovary (c), ovules (d), and pistil (e, f). Bars in (a, b), (c), (d), (e), and (f) represent 250 μm, 200 μm, 75 μm, 500 μm, and 100 μm, respectively. (g–j) Histochemical staining of β-glucuronidase (GUS) activity (blue). The top panel shows GUS-stained images of tomato flower buds, flowers, and leaves from plants carrying a PELOpro:GUS-mCitrine (PELOpro:GUS) reporter. The lower panel shows the WT control. (g), flower buds; (h), petals and sepals from a fully open flower; (i), leaflet from a young leaf; (j), leaflet from a mature leaf. Bars in (g–i) and (j) represent 1 mm and 0.5 cm, respectively.
PELO expression in leaves, the primary interface for plant-virus-whitefly interactions, suggests that PELO may support long-distance viral movement through the phloem. The relatively minor effect of the pelo mutation on leaf development, in contrast to its pronounced effect on flower development, suggests that while PELO is important for virus infection and floral development, it plays a limited role in leaf development. This may indicate redundancy with other factors or that ribosome recycling, in which PELO functions, is less critical in leaf cells.
Modification of PELO expression by promoter bashing mutagenesis
In the present study, pelo knockout mutants exhibit strong resistance to the virus and are sterile, a phenotype that has not been previously reported in the ty-5 allele or in RNAi lines (Lapidot et al., 2015). This suggests that ty-5 represents a missense or partial loss-of-function allele of PELO, rather than a complete loss-of-function. This led us to hypothesize that PELO function is dosage-dependent, and that fine-tuning its expression could enable the generation of plants in which PELO levels are maintained within a range that does not impair fertility while retaining strong TYLCV resistance. Therefore, we generated a series of PELO promoter-bashing alleles (Rodriguez-Leal et al., 2017), generating endogenous promoter variants with altered gene expression and associated phenotypic traits.
Using CRISPR–Cas9, we designed eight gRNAs targeting sequences distributed across the 2.8 kb PELO promoter region. The target site closest to the CDS was located c. 100 bp upstream of the ATG (Fig. 7a). We obtained 14 T0 promoter-edited plants carrying deletions, inversions, insertions, and single-nucleotide polymorphisms (Fig. S3). None of these lines exhibited developmental defects, and all produced fertile fruits.
Inducing edits in the PELO promoter using CRISPR-Cas9. (a) Sequencing of four selected PELO promoter-edited alleles, PELOCRpro-2, PELOCRpro-4, PELOCRpro-10, and PELOCRpro-31. Deletions (−), insertions (+), and inversions (gray dashed lines) are indicated by their positions and sizes. Blue arrowheads indicate the positions of gRNAs upstream of the start codon. (b) Levels of PELO expression in PELOCRpro-2 (#2), PELOCRpro-4 (#4), PELOCRpro-10 (#10), and PELOCRpro-31 (#31) were assayed by RT-qPCR relative to the reference gene EXPRESS and normalized to wild-type M82 (WT) expression. Data represent means ± SE from 4 to 5 biological replicates per genotype. (c) Disease severity index (DSI), ranging from 0 to 4, is shown. Infectivity was assessed 28 d post-inoculation (DPI). Mean values from at least 15 biological replicates per genotype ± SE are presented. (d) Viral DNA copy number determined by qPCR relative to the reference 18S ribosomal DNA (2−ΔCt). Tomato WT M82, PELOCRpro-4, and PELOCRpro-31 plants were inoculated with TYLCV using viruliferous whiteflies, and infectivity was assessed at 14–35 DPI. Mean values from 7 to 8 biological replicates per genotype and time point ± SE are shown. (e–g) Number of red fruits per plant (e), average single red fruit weight (f), and total fruit weight per plant (g) in WT M82 and the indicated genotypes. Mean values from at least four biological replicates (each replicate contained three plants) per genotype ± SE are presented. In (b, d–g), different letters denote statistical differences (P < 0.05) among samples as assessed by one-way ANOVA and Tukey HSD (honestly significant difference) test. ns, not significant. Gray dots indicate the individual data points.
All homozygote T2PELO promoter-edited lines (Fig. 7a, S3) were initially screened by visual assessment following TYLCV inoculation. Based on this screening, two lines that exhibited reduced disease symptoms (PELOCR-pro-4 and PELOCR-pro-31), together with two susceptible lines (PELOCR-pro-2 and PELOCR-pro-10), were selected for further characterization. Whereas the susceptible lines exhibited DSI values similar to those of the wild-type, PELOCR-pro-4 and PELOCR-pro-31 showed significantly reduced symptom severity and virus accumulation compared with wild-type plants (Fig. 7c,d). Both promoter-edited lines exhibited slightly, but not significantly, higher DSI values and viral accumulation than ty-5 plants, suggesting that the promoter edits may not fully recapitulate ty-5-mediated resistance (Fig. S4). Analysis of PELO expression revealed a significant reduction in transcript levels in PELOCR-pro-4 and PELOCR-pro-31 (Fig. 7b). Interestingly, the susceptible promoter alleles, PELOCR-pro-2 and PELOCR-pro-10, also exhibited reduced PELO expression, although to a lesser extent than PELOCR-pro-4 and PELOCR-pro-31 (Fig. 7b). These findings support the conclusion that PELO acts in a dosage-dependent manner. In addition, we observed that PELOCR-pro-4 and PELOCR-pro-31 are the only alleles with significant sequence changes in the region located between −124 and −267 bp upstream of the ATG, suggesting that this region, as well as sequences closer to the ATG, are important for PELO expression (Fig. 7a, S3).
Although all T0 promoter alleles were fertile, we next tested whether the resistance-associated promoter alleles impose any penalty on fruit number or size. To this end, we grew PELOCR-pro-4 and PELOCR-pro-31 plants in the net house and compared their fruit number and size with those of the resistant allele ty-5 and the wild-type M82 plants. Note that the ty-5 line used throughout this study was backcrossed to M82 five times (BC5F6) (Lapidot et al., 2015). We did not detect any significant differences in fruit number among the promoter bashing alleles, the ty-5 line, and the wild-type M82 control (Fig. 7e). Consistent with previous reports (Lapidot et al., 2015; Cabrera et al., 2024), ty-5 plants exhibited a significant reduction in single-fruit weight compared with M82 (Fig. 7f). Although total fruit weight was lower in ty-5 plants, this difference was not statistically significant (Fig. 7g). These findings suggest that ty-5 carries a fruit-size penalty, possibly resulting from linkage drag from its original genetic background or the presence of partially functional PELO protein. By contrast, neither single-fruit weight nor total fruit weight was reduced in PELOCR-pro-4 or PELOCR-pro-31 compared with M82 (Fig. 7f,g).
These results highlight the potential of CRISPR-based promoter bashing of PELO to generate novel alleles that may serve as alternatives to ty-5 in commercial programs while avoiding the significant reduction in single-fruit weight associated with ty-5. Nevertheless, because ty-5 plants accumulated less viral DNA than the promoter-edited lines generated in this study, additional promoter alleles should be developed to confer resistance levels comparable to, or greater than, those conferred by ty-5.
Discussion
The ty-5 locus has been known for many years to confer resistance to TYLCV (Anbinder et al., 2009; Hutton et al., 2012). This locus was mapped to a single-nucleotide polymorphism in the CDS of PELOTA (PELO), which encodes the tomato homolog of a messenger RNA surveillance factor (Lapidot et al., 2015). While a recent study reported that loss-of-function pelo mutant plants are resistant to TYLCV, this work was based on micropropagated shoots of biallelic CRISPR-Cas9 plants and did not include a comparison to ty-5-mediated resistance, leaving many questions unresolved (Pramanik et al., 2021). Here, we identify PELO as an important link between virus susceptibility and reproductive development in tomato. Complete loss of PELO function confers strong resistance to TYLCV, eliminating both disease symptoms and detectable viral accumulation. However, this resistance comes at the cost of severe failings in floral organ development and complete sterility. By contrast, partial reduction of PELO expression through CRISPR-based promoter editing generates alleles that maintain fertility while conferring intermediate resistance. Although the major effects of complete PELO loss were consistent in the M82 and R13 backgrounds, pleiotropic effects in tomato can vary considerably depending on the genetic background (Ramirez Gaona et al., 2023). Therefore, altered PELO activity may produce somewhat different phenotypic outcomes in other tomato backgrounds, and caution should be exercised when generalizing these findings.
Collectively, these findings show PELO as a key host factor whose precise expression and activity levels determine the balance between plant development and susceptibility to TYLCV, consistent with concepts proposed in previous studies (Onouchi et al., 2025; Majumdar et al., 2026).
PELO supports reproductive development
Recent work, including in plants, has established PELO as a key component of the Ribosome-Associated Quality Control (RaQC) pathways, including no-go decay and non-stop decay, which monitor ribosome stalling and ensure the removal of aberrant mRNAs and nascent peptides (Lee et al., 2007; Passos et al., 2009; Shoemaker et al., 2010; Becker et al., 2011; Szádeczky-Kardoss et al., 2018; Li et al., 2025; Majumdar et al., 2026; Sans-Coll & Merchante, 2026). These pathways not only maintain translational quality but also influence gene expression and developmental programs. Studies in yeast, animals, and plants have shown that PELO deficiency leads to defects in cell proliferation, differentiation, and gametogenesis. In Drosophila and mice, PELO is required for germline development and stem cell maintenance, while in yeast it regulates sporulation (Castrillon et al., 1993; Eberhart & Wasserman, 1995; Davis & Engebrecht, 1998; Adham et al., 2003; Raju et al., 2015). These conserved roles suggest that PELO-dependent ribosome recycling is particularly critical for maintaining cellular proteostasis in rapidly dividing or highly specialized cells. In plants, PELO is similarly essential for development, with Arabidopsis pelota1 and hbs1 mutants showing reduced germination, and rice (Oryza sativa) pelota mutants displaying spotted-leaf phenotypes, defective root systems, impaired pollen development, and reduced fertility (Ding et al., 2018; Zhang et al., 2018; Kong et al., 2021; Guo & Gregory, 2023). In addition, rice pelota mutants manifest enhanced disease resistance due to constitutive activation of defense pathways, supporting a role for PELO as a key regulator linking translational quality control with immune signaling and growth (Feng et al., 2013; Zhang et al., 2018).
Previous studies of partial loss-of-function and loss-of-function alleles in tomato and pepper did not report to have major developmental defects (Lapidot et al., 2015; Koeda et al., 2021; Pramanik et al., 2021). In the previously reported RNAi lines, PELO expression was reduced but not completely abolished, and the residual expression may have been sufficient to support normal reproductive development (Lapidot et al., 2015). Moreover, PELO expression and RNAi efficiency were not examined in floral tissues; therefore, the possibility that silencing was less effective or absent in reproductive organs cannot be excluded. By contrast, our data indicate that PELO is required for the proper development of both male and female reproductive organs. In addition, in wild-type plants, PELO expression was detected in pollen, ovules, the style, and the stigma, correlating with the abnormal anther cone formation, defective pollen, and loss of pollen viability. These data, consistent with the findings in Arabidopsis, rice, yeast, and animals (Castrillon et al., 1993; Eberhart & Wasserman, 1995; Davis & Engebrecht, 1998; Raju et al., 2015; Ding et al., 2018; Zhang et al., 2018; Kong et al., 2021; Guo & Gregory, 2023), show that in tomato, PELO deficiency might negatively affect cell proliferation, differentiation, and gametogenesis.
Variation in developmental severity among pelo mutants in tomato, rice, and Arabidopsis may result from differences in allele strength, residual PELO activity, genetic background, or functional redundancy with alternative ribosome-rescue pathways. The tomato genome contains an additional divergent PELOTA-like paralog, Solyc10g083770, whose predicted protein shares c. 35% amino acid sequence identity with PELO. Although the function of Solyc10g083770 has not been experimentally characterized, the complete sterility and strong resistance to TYLCV of the pelo loss-of-function mutants suggest that this paralog cannot fully compensate for PELO loss during reproductive development or replace its role in supporting TYLCV propagation. Conversely, the absence of pronounced vegetative defects suggests that Solyc10g083770, or other ribosome-rescue pathways, may provide partial or tissue-specific redundancy in vegetative organs. Greater functional redundancy in rice and Arabidopsis may preserve sufficient translational capacity to maintain at least partial fertility.
Floral organs, particularly the tapetum and developing gametophytes, have high metabolic and translational demands. Disruption of ribosome recycling in these tissues could therefore impair the synthesis of proteins needed for meiosis, cell differentiation, and pollen development and maturation. The observed malformations in anther structure, pollen morphology, and pollen viability provide evidence for this interpretation. Specifically, the disorganization of anther locules and the presence of abnormal pollen grains indicate impaired tapetum function, which is critical for supplying nutrients and regulatory signals during pollen development. We therefore propose the unifying model of PELO as a major supportive factor for reproductive development by ensuring effective translation during key developmental transitions, such as meiosis, gametophyte maturation, and fertilization.
PELO as a host factor for TYLCV infection
The strong resistance phenotype of pelo loss-of-function mutants indicates that PELO is required for TYLCV infection, propagation, or both. Notably, the strong expression of PELO in leaf vascular tissues is consistent with the phloem-associated biology of TYLCV. The virus is introduced by its whitefly vector, during phloem feeding, and subsequently infects phloem-associated cells and moves systemically through the vascular system. Accordingly, reduced PELO activity in vascular tissues may contribute to resistance by limiting TYLCV accumulation and systemic movement. However, our transmission experiments show that the very low viral levels detected in some pelo mutant plants can still lead to infection in healthy wild-type plants through whiteflly-transmission or grafting (Fig. 2). These findings argue against PELO being strictly required for viral movement or transmission, although they do not rule out a role in either process (Fig. 2).
Conversely, the loss or reduction of PELO activity in vascular tissues may restrict viral accumulation by limiting viral translation or enhancing antiviral defense responses. Although TYLCV mainly replicates in dividing tissues and this possibility requires further investigation, the spatial overlap between PELO expression and the tissues targeted by TYLCV provides a plausible link between its tissue-specific expression and the resistance observed in pelo mutants.
At the molecular level, one plausible explanation is that PELO facilitates the efficient synthesis of viral proteins through its central role in translation. Viral replication relies heavily on the host translational machinery, and ribosome stalling on viral RNAs may necessitate PELO-mediated recycling to sustain protein production. In the absence of PELO, stalled ribosomes are likely to accumulate on viral transcripts, leading to inefficient translation and consequently reduced viral replication. In addition, reduced levels of ribosomal components reported in pelo mutants (e.g. rice KO-OsPelo), suggest that impaired ribosome homeostasis may further limit viral protein synthesis, independently of its role in no-go decay (Sun et al., 2024).
An alternative, non-mutually exclusive explanation involves enhanced antiviral immunity. PELO has been proposed to function in non-stop decay-mediated suppression of RNA silencing amplification by promoting the clearance of aberrant RNA fragments generated during antiviral responses (Feng et al., 2013; Zhang et al., 2018). In pelo mutants, inefficient removal of these RNA substrates may prolong ribosome association with truncated transcripts, facilitating their entry into RNA-Dependent RNA Polymerase 6 (RDR6)-dependent amplification pathways. This results in increased production of secondary siRNAs and a hyperactivated antiviral response. Consistent with this, studies in rice have shown that OsPELOTA acts as a negative regulator of innate immunity, with pelota mutants exhibiting constitutive activation of defense pathways, including salicylic acid signaling, and enhanced disease resistance (Feng et al., 2013; Zhang et al., 2018). Together, these findings support a model in which PELO functions as a host susceptibility factor, required both for efficient viral translation and for restraining antiviral defense responses, a role that appears to be conserved across plant-pathogen systems.
Notably, recent evidence further indicates that viruses can actively counteract PELO-mediated RNA quality control (Ge et al., 2025). In potyviruses, the viral RNA-dependent RNA polymerase (NIb) acts as a SUMOylation decoy that competitively interacts with the host SUMO-conjugating enzyme SCE1, thereby reducing PELO SUMOylation and impairing its RNA quality control activity. This suppression of PELO-dependent RNA surveillance limits viral RNA degradation and promotes viral accumulation (Ge et al., 2023, 2025), highlight that PELO is not only a passive host factor required for viral infection, but also a direct target of viral counter-defense strategies, reinforcing its central role at the interface between host translational control, RNA surveillance, and antiviral immunity.
Linking virus infection to reproductive traits
TYLCV infection is known to cause pronounced developmental defects, including leaf abnormalities, flower abortion, reduced fruit set, and reduced yield. Our findings give a mechanistic explanation for these symptoms. If PELO is required both for viral replication and for normal reproductive development, viral infection may create a functional competition for PELO activity. Specifically, the virus may recruit PELO to sustain viral translation, thereby limiting its availability to support the endogenous developmental processes. Such competition could effectively phenocopy a partial pelo loss-of-function, leading to failings in reproductive development even in wild-type plants. This framework is consistent with the observed overlap between reproductive defects in infected wild-type plants and those seen in pelo mutants.
In addition, viral infection is likely to impose translational stress, increasing the demand for ribosome recycling and further constraining PELO availability. This effect would disproportionately impact tissues with high translational demand, such as developing floral organs, explaining their increased sensitivity compared to vegetative tissues. The relatively mild leaf phenotype observed in pelo mutants, in contrast to the more severe symptoms caused by TYLCV, may be explained by the presence of complementary or redundant surveillance pathways in leaves. Indeed, PELO is not the sole ribosome rescue factor; studies in yeast suggest that translation on aberrant or stalled mRNAs can, in some cases, be resolved by alternative mechanisms, such as eRF3-mediated termination, providing a parallel or early-acting pathway (Chiabudini et al., 2014; Guydosh & Green, 2017; Sans-Coll & Merchante, 2026).
Dosage-dependent effects and breeding implications
The recovery of fertile, virus-resistant plants through promoter editing demonstrates that PELO function is highly dosage-sensitive. Complete loss of function abolishes viral infection but disrupts development, whereas partial reduction maintains sufficient activity for fertility while limiting viral propagation. This highlights the importance of fine-tuning essential host factors rather than eliminating them. A similar conclusion was reached in a complementation study in Arabidopsis, which revealed that pepy-1, an allele of the pepper PELOTA homolog, is a partial loss-of-function allele that enhances virus resistance without severely compromising plant growth and development. The authors suggested that this balance may explain why the leaky allele was naturally selected in pepper and why it represents a useful source of resistance for breeding programs (Onouchi et al., 2025).
From a breeding perspective, these data provide a strategy to overcome the limitations of existing resistance alleles. Unlike ty-5, which still permits viral accumulation and may carry fitness penalties, engineered promoter alleles can be tailored to achieve an optimal balance between resistance and yield. Although the promoter-edited lines generated in this study did not exhibit any reduction in fruit size, they also did not confer stronger resistance than ty-5. Nevertheless, further refinement of promoter variants may enable stronger resistance while maintaining normal reproductive development and avoiding the fitness costs associated with complete PELO loss of function. This general strategy may be applicable to other crops. Nevertheless, orthologous promoters in Solanaceae species contain only short conserved regulatory regions embedded within otherwise divergent noncoding sequences (Hendelman et al., 2021). Moreover, the functional characterization of PELO in crop species remains limited. Consequently, extending this promoter-bashing strategy to additional species will require identification of the orthologous PELO gene, validation of its contribution to disease susceptibility, and species-specific characterization of its promoter before promoter optimization can be pursued. An additional consideration is that the role of PELO differs among viral families: although reduced PELO activity confers resistance to begomoviruses, it may increase susceptibility to potyviruses (Ge et al., 2023, 2025). Therefore, PELO promoter-edited alleles should be evaluated against other agriculturally important viruses before their incorporation into breeding programs.
Acknowledgements
The study was supported by grants from the United States-Israel Binational Agricultural Research and Development Fund (IS-5679-24D) to YB, VC, and ML, Israeli Ministry of Agriculture and Food Security (20-01-0337) to YB and ML, and the National Institute of Health (R35 GM144059) to VC We thank Hanita Zemach and Edward Belausov (Department of Plant Science, ARO, Israel) for their assistance with SEM and confocal microscopy, and we thank Dr Rina Kamentsky-Goldstein (Department of Plant Science, ARO, Israel) for critically reviewing the manuscript. The authors declare that GPT 5.6 was used to assist with English language polishing and readability improvement. The authors carefully reviewed and edited all text after using these tools and take full responsibility for the final content of the manuscript.
Competing interests
None declared.
Author contributions
ZS, TSR, VC, ML and YB designed the research and wrote the article. YB and ML supervised. ZS, TSR, YD, DG, SK, HW, BA, ML and YB performed the experiments and analyzed the data. ZS and TSR contributed equally to this work.
Data availability
The authors confirm that the data supporting the findings of this study are available within the article or its Tables S1–S3 and Figs S1–S4. Sequence data for genes used in this study can be found in the Sol Genomics Network under the following accession numbers: PELOTA (Solyc04g009810), EXPRESS (Solyc07g025390), and SlUbq10 (Solyc07g064130).
References
Disclaimer
The New Phytologist Foundation remains neutral with regard to jurisdictional claims in maps and in any institutional affiliations.