Discussion
The success of transgene RNA silencing transmission through grafting would be of practical importance in horticulture. Grafting wild-type scions onto transgenic silenced rootstocks could improve individual traits of well-established non-transgenic tree cultivars, particularly, for those recalcitrant to regeneration or transformation. It would be instrumental in the case of Prunus  species that are difficult or still impossible to transform, such as apricot (Petri et al. , 2015) or peach (Ricci et al. , 2020). Notably, it is expected that public concerns about using transgenic plants should be mitigated by the lack of transgenes spreading by outcrossing coupled with the consumption of non-transgenic edible parts of the plant.
Although encouraging, our first data (Table 1) studying the transmission of PPV resistance from transgenic plum rootstocks to wt apricot scions were limited in broadness due to the reduced sprouting of the wt apricot PPV-infected buds, possibly due to a high inoculum pressure. A similar technical limitation was also reported by Zhao and Song, (2014), which showed that grafting PNRSV infected buds onto sweet cherries results in the death of the buds.
To overcome the above limitation, we adopted the chip budding technique. This inoculum procedure permitted us to study the resistance behavior of one hundred and ten grafted wt apricots challenged with a PPV-D isolate.
The data clearly shows an increase with time in the number of PPV-free plants as evaluated by the RT-PCR analysis (Figure 1). In addition, the low percentage of PPV-infected apricot plants grafted onto resistant rootstocks accumulated significantly fewer amounts of the virus when compared to the susceptible rootstock (Figure 2). Notably, when non-transgenic scions that recovered from viral infection were re-challenged with PPV, the RNA silencing effectively eliminated the virus within the same growth cycle in about 50% of scions (Table 2). The ability to recover from virus infection is a peculiar characteristic of the RNA silencing-mediated resistance occurring both during natural viral infections (Covey et al. , 1997; Ratcliff et al. , 1997) and in herbaceous and woody transgenic plants (Dougherty et al. , 1994; García-Almodovar et al. , 2015; Ravelonandro et al. , 1993; Lindbo and Dougherty, 2005).
To the best of our knowledge, there are only three previous papers intending induced silencing from a rootstock to a scion in woody plants, with contradictory results. In apples, transgenic rootstock-mediated silencing in the scions was shown to occur for a gusA  transgene but not for an endogenous anthocyanidin synthase gene. Additionally, nor the transgene nor the endogen was silenced when the grafting experiment was conducted in the greenhouse (Flachowsky et al. , 2012). Authors hypothesized that lignification might influence cell-to-cell transport of siRNAs in living cells, thus explaining the lack of silencing effect.
In a recent paper (Sidorova et al. , 2021), two PPV-resistant transgenic plum cultivars transformed with a hairpin to silence the virus capsid gene were evaluated for their capacity to transfer the PPV resistance character to the wild-type grafts. They found that transgenic rootstocks remained virus free but could not protect the scion due to the lack of an efficient transfer of transgene-derived siRNAs from the rootstocks to the scions. However, scions accumulated specific endogen sRNAs characteristic of the rootstocks (Sidorova et al. , 2021). Conversely, Zhao and Song, (2014) showed that PNRSV-hpRNA-derived siRNAs were transmitted up to 1.2 m from the transgenic sweet cherry rootstocks to the non-transgenic scions conferring enhanced PNRSV resistance.
Variable results were also obtained in works dealing with grafting-mediated virus resistance in horticultural species. Baiet al. , (2016) showed that 66.7% to 83.3% of non-transgenic tomatoes were highly resistant to CMV. In tobacco, detached leaves from scions grafted on transgenic tobacco silenced for the endogenous NtTOM1  and NtTOM3  genes were shown to accumulate fewer tobamoviruses than the control plants (Ali et al. , 2013). Similarly, Nicotiana benthamiana  transgenic plants expressing a hairpin designed to silence PSTVd produced only attenuation of viral infection (Kasai et al. , 2013). The contradictory data can be attributed to differences between plant species, the transgenic construct used and/or the targeted sequence (exogenous infecting virus or endogenous gene transcripts).
Our data do not support the hypothesis of lignification as the primary cause of the lack of RNA silencing spreading from rootstocks to scions (Flachowsky et al. , 2012; Sidorova et al. , 2021). Apricot scions and plum rootstocks were well-lignified during the four years that the experiment lasted (from rooting and acclimatizing the rootstocks, grafting apricots, and infecting them by chip-budding to the final evaluation).
The siRNAs analyses identified the accumulation of the transgene-derived PPV UTR/P1 24 nt siRNAs in apricot scions grafted on the PPV resistant but not onto the susceptible rootstock (Figure 3). In contrast, faintly amount of 22 nt siRNAs were detected in the apricot scions on resistant rootstocks but more clearly seen on St5’-7 scions (Figure 3), suggesting that the 24 nt siRNAs can be: a) preferably transported over a long distance (Hamilton et al. , 2002; Molnar et al. , 2010); b) less prone to degradation or; c) less consumed by AGO in the traversed and recipient cells (Voinnet, 2022). When studying the molecular mechanisms associated with the resistance to sharka of C5 plum (’Honeysweet’), 24 nt siRNAs was related to systemic silencing (Kunduet al. , 2008). In particular, they were present only in resistant C5 plants but not in susceptible ones nor in C5 plants showing middle sharka symptoms. The evidence that the tolerant and resistant plum rootstocks could protect the apricot scions, and that 24 nt siRNAs were only found in these scions but never in those grafted onto susceptible St5’-7 line, agrees with results found in ‘Honeysweet’ plum.
Different works suggest that all siRNA classes (21, 22, and 24 nt long siRNAs) are mobile (Devers et al. , 2020), with the 22 nt siRNAs having a pivotal role in the siRNAs signal amplification and translational repression (Chen et al. , 2010; Cuperus et al. , 2010; Wu et al. , 2020). Trans-grafting movement of siRNAs is not a simple concentration dependent diffusion process, but probably requires a selective sRNA sorting mechanism and recent studies suggest that it might be dictated by sRNA biosynthetic pathways, sRNA sizes, sequence features such as 5’ nucleotide, or selective RNA-binding protein partners (Kong et al. , 2022). It will be interesting to evaluate the amounts and nature/diversity of 5′-nucleotide identities/sizes of siRNAs accumulating in the grafted apricot scions and transgenic rootstocks using a more sensible and specific technique.
Northern blot analysis identified, in addition to the siRNA in PPV-resistant plum, transgene-derived UTR/P1 siRNAs in all transgenic rootstocks independently on the level of PPV resistance, indicating that their accumulation is necessary, but not sufficient, to assure efficient PPV interference. These data agree with those obtained by López et al. , (2010) in Mexican lime transformed with sense, antisense, and intron-hairpin cDNAs from viral sequences and with data from tobacco (Alburquerque et al. , 2012) or plum (Alburquerque et al. , 2017) transformed with a chimerical transgene designed to silenceAgrobacterium oncogenes iaaM and ipt . In those works, all resistant lines accumulated transgene-derived siRNAs, but this was not necessarily associated with resistance to citrus tristeza virus (López et al. , 2010) or crown gall (Alburquerque et al. , 2012; Alburquerque et al. , 2017). Therefore, a lower amount of hpRNA seems to be better correlated with resistance. Although this could be due to lower expression or higher degradation, it seems logical to think that resistance is related to a more efficient degradation of the hpRNA.
Previous studies showed that transgenic C5 plants were resistant to PPV when exposed to natural viruliferous aphids while accumulating low-level PPV near the graft junction if graft-inoculated (Malinowski et al. , 2006). Based on the C5 plants data, we expected that the apricots grafted onto the PPV-resistant plum lines should also be resistant to PPV infection under natural field conditions. Importantly, since the PPV-derived h-UTR/P1 construct present in transgenic plum rootstocks was derived from a PPV isolate belonging to the M strain while the plants were challenged with a PPV-D isolate, it suggests that the resistance observed should be extended to, at the very least, the viral isolates of the two most important and widespread PPV strains.
As conclusion, the results demonstrate for the first time that PPV-resistant transgenic plums can effectively confer sharka resistance in grafted non-transgenic apricots scions. It is expected that using transgenic rootstocks can mitigate public concerns about transgene dispersions and eating transgenic food.
Additional studies on the long-distance movement of the RNA silencing signal are required to understand how broadly applicable this technique is to modulate the phenotype of wild-type grafted scions in woody plants. Uncovering the mechanism of sRNA selection for trans-grafting transport will potentially enhance success in designing artificial sRNAs to control plant disease.