Pre-symptomatic modified phytohormone profile is associated with lower phytoplasma titres in an Arabidopsis seor1ko line

The proteins AtSEOR1 and AtSEOR2 occur as conjugates in the form of filaments in sieve elements of Arabidopsis thaliana. A reduced phytoplasma titre found in infected defective-mutant Atseor1ko plants in previous work raised the speculation that non-conjugated SEOR2 is involved in the phytohormone-mediated suppression of Chrysanthemum Yellows (CY)-phytoplasma infection transmitted by Euscelidius variegatus (Ev). This early and long-lasting SEOR2 impact was revealed in Atseor1ko plants by the lack of detectable phytoplasmas at an early stage of infection (symptomless plants) and a lower phytoplasma titre at a later stage (fully symptomatic plants). The high insect survival rate on Atseor1ko line and the proof of phytoplasma infection at the end of the acquisition access period confirmed the high transmission efficiency of CY-phytoplasma by the vectors. Transmission electron microscopy analysis ruled out a direct role of SE filament proteins in physical phytoplasma containment. Time-correlated HPLC–MS/MS-based phytohormone analyses revealed increased jasmonate levels in midribs of Atseor1ko plants at an early stage of infection and appreciably enhanced levels of indole acetic acid and abscisic acid at the early and late stages. Effects of Ev-probing on phytohormone levels was not found. The results suggest that SEOR2 interferes with phytohormonal pathways in Arabidopsis midrib tissues in order to establish early defensive responses to phytoplasma infection.


Results
Phytoplasma transmission by Euscelidius variegatus. In this study, three treatment groups of the three Arabidopsis plant lines were investigated: (1) plants (no-Ev) not infested by E. variegatus, (2)

plants (H-Ev) infested by E. variegatus from a healthy colony that never fed on phytoplasma-infected plants, and (3) plants (CY-Ev) infested by CY-infected E. variegatus.
For each analysis plants were tested at two stages, i.e. 5 (T1) and 20 days (T2) after the end of the inoculation access period (IAP), which are referred to as the "early" and "late" stage of infection, respectively.
Three individuals of healthy E. variegatus or CY-infected E. variegatus were placed on wild-type, Atseorko1 and Atseor2ko Arabidopsis plants. The insects were manually removed after 7 days (i.e. at the end of IAP) and the number of living individuals was counted. The survival rates of healthy E. variegatus, which ranged between 59 and 79% (Fig. S1), showed no statistically significant differences between the respective Arabidopsis lines. Yet there was a tendency for the healthy individuals to survive slightly better on the Atseorko mutants (Fig. S1). This tendency became statistically significant (Fig. S1) for CY-infected-E. variegatus, which suffered from decreased fitness on wild-type plants as compared to the mutants (with a lower survival rate, ranging from 19.5 to 38.9%).
Healthy and CY-infected E. variegatus were then pooled (as reported below) and processed for phytoplasma detection. PCR analysis, with a 1,250 bp amplicon as a final product, confirmed the presence of phytoplasmas in all pools from 3 insects having fed on CY-infected chrysanthemum. The success of the infection in CY-Ev plants was further confirmed by symptom development in each Arabidopsis line under investigation (100% plants were positive to CY phytoplasma following inoculation with vectors having fed on CY-infected chrysanthemum, with a transmission rate (p) of 1 27 .
Five days after the IAP, no symptoms were visible in the lines exposed to the CY-infected E. variegatus ( Fig. 1A,C,E). Initial symptoms (leaf chlorosis and petiole elongation) emerged starting from the 14th day after IAP and characteristic CY symptoms became fully discernible 20 days after IAP (Fig. 1B,D,F). In fact, at this time point, all infected plant lines showed reduced growth and shorter, yellowish leaves, with a thick main vein. Chlorosis progressed from the youngest leaves towards the others. The appearance of the symptoms in wild-type and mutant lines was similar at this stage of infection (Fig. 1B,D,F).

Phytoplasma quantification in Arabidopsis lines. To quantify the phytoplasma inside the CY-Ev
Arabidopsis lines, qPCR was carried out using genomic DNA extracted from 12 plants for each infected Arabidopsis line.
At the early stage of infection (i.e. 5 days after IAP), none of the Atseor1ko plants tested positive for the presence of phytoplasma, while 58% of wild-type plants and 42% of Atseor2ko line did so ( Fig. 2A). At this timepoint, the phytoplasma titre was lower in comparison to that found at the late stage of infection, and did not significantly differ among the Arabidopsis lines (Fig. 2B). The highest cycle quantification (Cq) value to detect phytoplasma DNA in 100 mg of plant tissue was 32.83 for wild-type, corresponding to 7.76E+03 genome units (GUs), 34.30 for Atseor2ko, corresponding to 2.73E+03 GUs, while none of Atseor1ko plants resulted positive for phytoplasma presence. At the late stage of infection (i.e. 20 days after IAP), 100% of the plants treated with CY-Ev tested positive for phytoplasmas ( Fig. 2A), but the phytoplasma titre in the Atseor1ko line was significantly lower than in Atseor2ko or wild-type plants ( Fig. 2B and 9 ). The highest Cq value to detect phytoplasma DNA in 100 mg of plant tissue was 18.38 for wild-type, corresponding to 2.05E+08 GUs, 18.07 for Atseor2ko, corresponding to 2.70E+08 GUs, 20.17 for Atseor1ko corresponding to 6.62E+08 GUs.

Ultrastructural modifications in midrib phloem at the early and late stage of infection. Both
AtSEOR1 and AtSEOR2 8,28 are regarded as being necessary for the formation of SE protein filaments in Arabi-Scientific RepoRtS | (2020) 10:14770 | https://doi.org/10.1038/s41598-020-71660-0 www.nature.com/scientificreports/ dopsis. As expected 8 , the SEs in the no-Ev wild-type line contained protein filaments, scattered throughout the SE lumen (Fig. 3A,G) and accumulated in the proximity of the sieve plates (Fig. 3D,J). On the contrary, the two no-Ev mutants, which are unable to form the respective SEOR partner proteins, showed no filaments in either the lumen (Fig. 3B,C,H,I) or near the sieve plates (Fig. 3E,F,K,L). In order to examine leafhopper-or phytoplasma-induced ultrastructural modifications in SEs, 6 H-Ev and 6 CY-Ev plants of each line were sampled for TEM analysis and compared with the corresponding no-Ev plants at 5 and 20 days after IAP. Five days after IAP, the H-Ev plants (Fig. 4A-F) did not show ultrastructural changes as compared to the no-Ev plants (Fig. 3). The SEs in the wild-type Arabidopsis showed protein filaments both in SE lumen (Fig. 4A) and near the sieve plates (Fig. 4D). As in the no-Ev plants (Fig. 3), the H-Ev mutant plants did not show filaments in the lumen of SEs (Fig. 4B,C) or near the sieve plates (Fig. 4E,F).
At the early stage (5 days after IAP), the CY-Ev wild-type plants (Fig. 4G,J) did not differ from their controls (Fig. 4A,D), in that protein filaments had accumulated in SEs (Fig. 4G,J). At this stage, Atseor1ko and Atse-or2ko CY-Ev plants did not show any ultrastructural alterations as well, because SE filaments were not detected (Fig. 4H,I,K,L) just as in their controls (Fig. 4B,C,E,F). At this time-point, phytoplasmas were not found in TEM pictures in any of the CY-Ev Arabidopsis lines (Fig. 4G-L).
Twenty days after IAP, the H-Ev plants ( Fig. 5A-F) contained SE protein filaments only in the wild-type individuals (Fig. 5A,D) and not in the mutants (Fig. 5B,C,E,F) as in the no-Ev plants (Fig. 3). By contrast, SEs of all CY-Ev lines were characterized by the presence of filaments (Fig. 5G-L). Phytoplasmas were abundant throughout the entire SE, both in the SE lumen ( Fig. 5G-I) and in proximity of the sieve plates ( Fig. 5J-L).  The IAA levels were not affected by E. variegatus infestation at both time intervals in both tissues of H-Ev plants (Fig. 8A,B). The IAA concentrations in laminae decreased in the three CY-Ev lines 5 and 20 days after IAP (Fig. 8A), with a significant decrease in the two mutant lines at the early stage of infection. In midribs, the IAA concentration showed a statistically significant increase in wild-type CY-Ev samples, as compared to the  In comparison with the No-Ev or H-Ev samples, the ABA concentrations were significantly increased in CY-Ev in the midribs at both time-points in Atseor1ko plants (+ 177% at the early infection stage and + 530% at the late infection stage) and at the early stage of infection in Atseor2ko plants (+ 242%).
All in all, infestation by E. variegatus does not seem to seriously affect the hormonal balance (Figs. 6,7,8), whereas CY infection is associated with changes on JA, IAA and ABA levels, in particular in the midribs of Atseor1ko plants during the pre-symptomatic early stage (Tables 1, 2).

Discussion
the Atseor1ko line limits phytoplasma replication from an early stage of infection onwards. In search for a role for SE protein filaments in response to phytoplasma infection 9 , phytoplasma-infected Atseor1ko plants turned out to host a significantly lower phytoplasma titre in comparison to wild-type and Atseor2ko lines,   www.nature.com/scientificreports/ but a satisfactory explanation was not evident. AtSEOR1 and AtSEOR2 are thought to be necessary for filament formation through their heterodimeric interaction 8 . Therefore, it was speculated that in the Atseor1ko line, the AtSEOR2 protein in its free form, i.e. not linked to AtSEOR1, is involved in immune signalling through an interaction with defence-related plant proteins 10,11 . In line with this conjecture, AtSEOR2 was found to interact with AtRIN4, a PRR plasma membrane-anchored protein in a matrix-based yeast two-hybrid assay 10 . Expression of AtRIN4, and the associated AtRPM1 and AtRPS2 genes in healthy and phytoplasma-infected wild-type and Atse-or1ko lines revealed an upregulation in the mutant line as compared to the wild-type, which was suggestive of a role of AtSEOR2 in promoting defence mechanisms 11 . In this frame, interactions between AtSEOR2 and diverse transcription factors 29 as well as its intervention in the IAA and ABA signalling cascades were predicted 12,30 .
In the present work, the relation between AtSEOR2 and phytohormone synthesis and the impact on phytoplasma titres was investigated in wild-type, Atseor1ko and Atseor2ko Arabidopsis lines. As the effectiveness of the defence processes is related to the readiness of plants to counter pathogens 31 and very little is known about the early response to phytoplasma infection, analyses were performed both at an early and a late time point of infection, i.e., respectively, at 5 days and 20 days after IAP. Up to 5 days after the end of IAP, the plant lines did not show any phenotypic differences (Fig. 1). As reported before 32 , growth and development of infected plants at this stage were comparable to those of control plants exposed to non-infected insect vectors.
Real-time PCR analyses evidenced that at the early stage (Fig. 2), the average phytoplasma titres were low (Cq values ± SE wild-type: 29.11 ± 2.17; Atseor1ko: n.d.; Atseor2ko: 30.98 ± 3.71), which has been demonstrated previously 33 . Interestingly, none of the Atseor1ko individuals tested positive at all for the presence of phytoplasma at this stage (Cq value: n.d. and Fig. 2A). The very high transmission efficiency of E. variegatus under natural and experimental conditions [34][35][36][37] is confirmed in our experimental system (transmission rate 100%, p = 1 27 ). Moreover, the higher vector survival rates on mutant lines (Fig. S1) together with the assumption that the vector fitness reflects the feeding capacity, infers that the lower phytoplasma titre in the Atseor1ko line is logically due to plant properties and not to reduced insect-mediated transmission efficiency.
The fact that insects showed higher survival rate on infected mutant lines than in wild-type, could be due to the complementary capability of AtSEOR proteins (both expressed in wild-type Arabidopsis) to aggregate in www.nature.com/scientificreports/ presence of phytoplasmas [this work; 9 ], reducing the phloem mass-flow and impairing stylet sucking-activity.
Will et al. 38 demonstrated that SEOR-mediated plugging is induced by green peach aphid (Myzus persicae) feeding on Vicia faba and this mechanism impairs feeding in aphid-resistant varieties 39 . Phytoplasma titres at the early stage of infection showed trends similar to those 20 days after IAP (when the symptoms had become manifest): the phytoplasma titres in wild-type or Atseor2ko plants exceeded by far the titre in Atseor1ko plants ( Fig. 2 and 9 ). Therefore, phytoplasma multiplication was presumably impaired in the Atseor1ko line from the earliest stages of infection onwards.
Sieve-element protein filaments aggregate in sieve tubes of wild-type Arabidopsis line from the early stage of infection onward. We previously demonstrated that SE protein filaments play a role in the plant response to phytoplasma infection, even in the absence of genes that are considered necessary for their formation in healthy plants 8,9 . Nevertheless, the picture drawn at that time did not explain whether SE protein filaments per se are involved in some defence mechanisms, such as early pathogen containment. Ultrastructural analysis confirmed the presence, at the late stage of infection (i.e., 20 days after IAP, Fig. 5), of filamentous structures in each CY-Ev Arabidopsis line (wild-type, Atseor1ko and Atseor2ko) and revealed that, at the first stage of infection (5 days after IAP), SE protein filaments only aggregated in CY-Ev wild-type plants (Fig. 4). Their initial absence in the Arabidopsis line that is best equipped to suppress the pathogen invasion from the early stage of infection on (i.e. Atseor1ko), led us to exclude aggregation of SE protein filaments as a possible explanation for the better defence performance of the Atseor1ko line. Therefore, our investigations were further focused on the phytohormone levels, which are frequently related to defence mechanisms 15,16,[40][41][42][43] . Since leaf midribs are rich in phloem tissues, where phytoplasmas and plants physically and chemically interact 5 , plant responses were determined in leaf midribs and laminae separately.

Phytohormone levels in Arabidopsis lines are not affected by E. variegatus infestation at 5 and 20 days after the end of the IAP. Damage inflicted by insect feeding leads to the immediate activation
of phytohormone-related signaling 26 and in particular to the synthesis and accumulation of JA in tissues both proximal and distal to injury sites 44 .
As for the effects of leafhopper infestation, phytohormone levels were equal in the midribs or laminae of each line and at both time points (i.e. 5 and 20 days after the end of IAP) in H-Ev and no-Ev samples. Changes in phytohormonal balance following leafhopper infestation have been amply described 26 , but the results were quite variable, probably owing to variations in infestation times 45 . In general, defense phytohormones are immediately induced in plant tissues after recognition of invaders, but the amounts tend to level off during persistent insect infestation 44 . Moreover, the effects of cutting, necessary to isolate midribs from laminae, may overshadow insect-induced responses 44 .
The fact that the phytohormone levels are identical in H-Ev and no-Ev plants strongly supports the conclusion that the phytohormone modulation in CY-Ev plants is solely due to phytoplasma infection. This is in agreement with the observation that 6 days after Scaphoideus titanus infestation in grapevine, the expression level of genes involved in JA and ABA pathways were similar to those found in non-infested control leaves, while they had been upregulated 3 days after infestation 46 .
Jasmonates, but not salicylic acid seem to be involved in the response of Atseor1ko line to phytoplasma infection. There are several indications for phytohormone involvement in diverse phytoplasma-plant interactions 4 , but few studies have addressed plant responses at the early stage of infection 40,46 .
The most frequently studied phytohormones in relation with phytoplasma infections are SA and JA, which both confer signal transduction leading to plant resistance. JA-and SA-mediated signalling pathways are presumed to be antagonistic 41,42 . Traditionally, SA signalling is deemed to activate resistance against biotrophic and hemibiotrophic pathogens, while JA is mainly thought to induce resistance against necrotrophic pathogens and wounding 44,47 .
In CY-Ev wild-type and Atseor1ko lines SA levels did not change significantly in all tissues examined (Fig. 6). SA increased in the midribs of Atseor2ko line at the early stage of infection, followed by a drop at the late stage (Fig. 6). Enhanced amounts of SA resulted from different plant-phytoplasma interactions in whole leaves 18,40,48 , in midribs 22 and phloem sap 49 . On the other hand, cases, which showed reduced SA levels in response to phytoplasma infection were also described 17,50 . The reason why SA level was significantly higher in Atseor2ko line compared to the other lines is unknown. It has to be noted that SA levels depend on many factors, such as the developmental stage of the vegetative cycle 17 , varying environmental conditions 20 , the phytoplasma strain in question 51 , and distinct sets of virulence factors 49 .
With regard to the JA concentration, the levels of both jasmonates were virtually unaffected in laminae. In midribs of all Arabidopsis lines JA showed a non-significant increment, apart from the Atseor1ko line, which showed significantly higher jasmonate levels at an early stage of infection (Fig. 7B). An increase of JA levels during the early stages of infection process, followed by decrease at symptom appearance, was described in different plant-phytoplasma interactions 14,17,20,24,52 .
Sugio et al. 53 found that Arabidopsis plants infected with the 'Ca. P. asteris' strain AY-WB produced more JA in old asymptomatic leaves (as well as in uninfected plants) as compared to young symptomatic leaves. The authors also demonstrated that reduced JA levels affected plant development, leading to symptom appearance and increasing insect-vector fitness 53 . Decreased JA levels were also reported in Arabidopsis expressing phytoplasma virulence factors (i.e. TENGU 14 or SAP11 53  www.nature.com/scientificreports/ Janik et al. 17 hypothesized a relationship between an increased phytoplasma titre, the activity of phytoplasma effectors and the decrease of JA levels over the growth season. Interestingly, in perennial plants, increased JA synthesis was correlated to the phenomenon of recovery 40,48 , a resilience status characterized by the loss of disease symptoms in plants which previously showed them 54,55 . Hence, the higher content of JA in Atseor1ko line could be related to failing phytoplasma detection (and to the absence of activity by their effectors) at the first stage of infection.
A JA-SA antagonism 41,42 may be visible in Atseor1ko and Atseor2ko lines at the early stage of infection, because at the high JA levels in the one correspond the low SA content in the other and vice versa at the same time point (Figs. 6B, 7B).
IAA-and ABA-synthesis are activated in the midribs of the phytoplasma-infected Atseor1ko line. As for the IAA variations in phytoplasma-infected plants, expression analyses of IAA-related genes in whole leaves 56,57 , buds 58 , or leaf midribs 59 infer a downregulation in several infected hosts. On the other hand, IAA levels were reported to increase in phloem sap 49 , whole leaves 60 and leaf midribs 59 of phytoplasma-infected plants as compared to the respective control samples.
Here, phytoplasma infection induced a decrease in IAA levels in the laminae and an increase in the midribs in each Arabidopsis line (Fig. 8). This indicates a positive response of IAA synthesis, located in the midribs as reported previously 59 . The increase in IAA was statistically significant in infected midribs of Atseor1ko mutants at both time points, i.e. at 5 and 20 days after IAP (Fig. 8).
Previous studies evidenced an increase in ABA levels and the upregulation of ABA-related genes in all tissues of phytoplasma-infected plants 17,18,49,58,61 . In our study, a tissue-dependent ABA response was observed in wild-type plants; the ABA level decreased in laminae, while it increased in midribs. Variations in the expression of genes related to ABA biosynthesis have been associated with symptom expression 61 . ABA-promoted stomatal closure could induce pre-invasive defence by inhibiting the entry of pathogens through passive ports 62,63 . Furthermore, ABA signalling would initiate events such as callose accumulation and antagonize the signalling cascades of other phytohormones at an advanced state of infection 64,65 . It is worth noting that Atseor1ko was the only line, in which ABA is significantly enhanced in the midrib from the early infection stage on, and the higher level was maintained throughout the entire measurement period.
Potential modes of involvement of AtSEOR2. Thus far, the relationship between free AtSEOR2 and phytohormone synthesis is a mystery. Yet, there is a wide range of possibilities.
According to the COACH platform 66 , AtSEOR2 protein (but not AtSEOR1) may have Ca 2+ binding sites, which suggests a possible role for AtSEOR2 in the lowering of Ca 2+ levels as does the sequestration of Ca 2+ ions by SEOR-based forisomes in legumes 67 . The subsequent modification of Ca 2+ signatures may promote ABA and IAA synthesis [68][69][70][71][72] in Atseor1ko infected plants. This hypothesis seemingly makes a logical connection between AtSEOR2 and the low phytoplasma titre owing to enhanced IAA and ABA synthesis.
In addition or alternatively, AtSEOR2 may effect on the IAA and ABA signalling pathways. To the best of our knowledge, the number of reports about possible interaction of AtSEOR2 protein with components of IAAor ABA-signalling cascade is scarce. Yeast two-hybridization experiments demonstrated that AtSEOR2 is able to interact with the At4g04950 gene product, a monothiol glutaredoxin that is a key component involved in ROS accumulation and IAA signalling 73 . The role of AtSEOR2 in ABA-cascade signalling has also hardly been explored thus far. Some evidence has recently been presented for an interaction between AtSEOR2 and SUA (SUPPRESSOR OF ABI3-5), a main component of the ABA signalling pathway 74 , and its involvement in the increased sensitivity to ABA 74 . Moreover, Nakashima and co-workers 30 demonstrated that AtSEOR2 is one of the many genes whose expression changes in A. thaliana knock-out mutants of three SnRK2 kinases involved in ABA signalling. Interestingly, AtSEOR2 has been reported to interact with the At1G31280.1 gene product 12 , an ABA-regulated protein that controls plant response against virus infections 75 .
Furthermore, it is not excluded that AtSEOR2 directly interacts with receptor(s) in the SE-CC complex to elicit defence responses. SEOR proteins are characterized by a conserved C-terminal M1 motif, containing several conserved cysteine residues 28 characteristic of the so-called peptide ligands 76 , which are engaged in the regulation of developmental processes and defence mechanisms against pathogens 77 . Interestingly, in mulberry infected by yellow dwarf disease, a phytoplasma-responsive gene encoding a protein that shows structural similarity to peptide ligands, was identified. This gene is involved in signaling and metabolism of IAA, ABA and JA 78 .
In conclusion, AtSEOR2 indirectly manipulates plant response via increased phytohormone synthesis and phytohormone signalling and perhaps via interaction with membrane receptors. These responses emerge very early in the infection process, long before the appearance of infection symptoms.

Plant material and insect vectors.
The seeds of wild-type, Atseor1ko (SALK_081968C), and Atseor2ko www.nature.com/scientificreports/ the 35-day latency period (LP) after which they have become infectious. Twelve 45-day-old A. thaliana plants per line were then each exposed to three infectious insects (CY-infected E. variegatus, CY-Ev plants) for a 7-day phytoplasma inoculation-access period (IAP), after which the insects were manually removed. Twelve Arabidopsis plants per line, treated with three healthy leafhoppers (H-Ev plants) were used as a healthy control. Healthy leafhoppers have been collected from healthy colonies and were as old as the infected ones. For microscopy and phytohormone analyses, 12 plants per line not subjected to insect feeding (non-infested plants, no-Ev plants), were included as additional negative controls. For ultrastructural observations and phytohormone quantification, 6 H-Ev and 6 CY-Ev plants from each line (i.e. 6 independent biological replicates) were used, for the phytoplasma titre analyses12 H-Ev and 12 CY-Ev plants (i.e. 12 independent biological replicates) were used.
Evaluation of insect survival rate and phytoplasma detection. To ascertain successful phytoplasma transmission to the Arabidopsis plants, survival rates and the presence of phytoplasma were checked in leafhoppers that were removed from Arabidopsis plants at the end of the IAP. The survival rate was calculated using 12 biological replicates (i.e. 12 Arabidopsis plants) for each condition (i.e. various Arabidopsis lines and infection times). For phytoplasma detection, the 3 insects used on each Arabidopsis were pooled, to obtain 72 pools per condition. Total DNA was extracted as described 79 and the presence of phytoplasmas in each pool was assayed by conventional PCR using the primer pair R16F2/R2, as described by Pagliari and co-authors 9 .
Phytoplasma transmission by leafhoppers was evaluated on the basis of the number of Arabidopsis plants showing symptoms 20 days after the end of IAP. To estimate the proportion of infectious insects in our experiment, the maximum likelihood estimator of p, p = 1 − Q 1/k was used 27 , where Q is the observed fraction of noninfected plants and k is the number of insects per plant, assuming that the vectors acted independently 27 . The formula can be applied when transmission trials are carried out using groups of insects 35 .
Statistical analysis was performed using SigmaPlot 12.0 software (Systat Software, Inc., San Jose, CA, USA). The normal distribution of the data was checked with the Shapiro-Wilk normality test. A three-way ANOVA of the means (from 12 biological replicates and 3 technical replicates) followed by the Holm-Sidak test as the post hoc test for multiple comparisons demonstrated the significance for p < 0.05.
Phytoplasma quantification in Arabidopsis. Total DNA was extracted from 200 mg of whole-leaf tissue of H-Ev and CY-Ev plants according to Martini et al. 80 . The amount of CY phytoplasmas was quantified according to a real-time PCR protocol described in detail by Pagliari and co-authors 9 . Briefly, the ribosomal protein gene rplV (rpl22) was the target for amplification of CY phytoplasma DNA using the primer pair rp(I-B) F2/rp(I-B)R2 9 ) and a CFX96 real-time PCR detection system (Bio-Rad Laboratories, Richmond, CA, USA). A standard curve was established by tenfold serial dilutions of plasmid DNA containing the 1,260 bp ribosomal protein fragment from CY phytoplasma, amplified with the primer pair rpF1C/rp(I)R1A. Real-time PCR mixture and cycling conditions were as previously described 9 . The phytoplasma concentration was expressed as the number of CY phytoplasma genome units (GUs) per mg of leaf sample to normalize the data. Differences among the means were calculated using SigmaPlot 12.0 software (Systat Software). The normal distribution of the data was checked with the Shapiro-Wilk normality test. A two-way ANOVA of the means (obtained from 12 biological replicates and 3 technical replicates) followed by the Holm-Sidak test as post hoc test for multiple comparisons demonstrated the significance for p < 0.05.

Transmission electron microscopy.
To preserve the damage-sensitive sieve-element ultrastructure, a gentle preparation method was adopted following Pagliari et al. 9 . From each plant a 25 mm-long midrib portion was excised from rosette leaves. The midrib segments were submerged in MES buffer and then fixed with 3% paraformaldehyde and 4% glutaraldehyde solutions. Samples were rinsed, post-fixed overnight with 2% (w/v) OsO4, dehydrated in a graded ethanol series and then transferred to propylene oxide. From the central part of each midrib, 6-7 mm long segments were excised and embedded in Epon/Araldite epoxy resin (Electron Microscopy Sciences, Fort Washington, PA, USA).
Ultrathin sections (60-70 nm in thickness) were cut, stained with UAR-EMS (uranyl acetate replacement stain) (Electron Microscopy Sciences), and observed under a PHILIPS CM 10 (FEI, Eindhoven, The Netherlands) transmission electron microscope (TEM), operated at 80 kV, and equipped with a Megaview G3 CCD camera (EMSIS GmbH, Münster, Germany). Five non-serial cross-sections from each sample were analysed. Phytohormone analyses. We adopted a validated HPLC-MS/MS method 81 , optimized for A. thaliana and the low concentrations (nM to μM) of the phytohormones of interest. Phytohormone extraction was performed using 6 plants per experimental condition.
For each sample, roughly 250 mg of midribs and 250 mg of laminae were collected for phytohormone analysis, immersed immediately in liquid nitrogen and then stored at − 80 °C, as described by Pommerrenig et al. 82 .
The extracts were analysed using an HPLC-MS/MS method 81 on an Agilent 1,100 HPLC system (Agilent Technologies, Böblingen, Germany) connected to a LTQ Ion Trap mass spectrometer (Thermo Scientific, Bremen, Scientific RepoRtS | (2020) 10:14770 | https://doi.org/10.1038/s41598-020-71660-0 www.nature.com/scientificreports/ Germany). Chromatographic separation was carried out in a Luna phenyl-hexyl column (150 × 4.6 mm, 5 μm; Phenomenex, Aschaffenburg, Germany). Formic acid (0.05%, v/v) and MeOH with 0.05% (v/v) formic acid were used as mobile phases A and B, respectively. The elution profile was: 0-10 min, 42-55% B in A; 10-13 min, 55-100% B; 13-15 min 100% B; 15-15.1 min 100-42% B in A; 15.1-20 min 42% B in A. The mobile phase flow rate was 1.1 mL/min. The injection volume was 20 μL. Phytohormone quantifications were based on calibration curves, and the data obtained from each sample were subsequently analysed with XCalibur software (Thermo Fisher Scientific). Statistical differences between the means obtained from 6 individuals exposed to the different conditions (i.e. various Arabidopsis lines and infection times) were evaluated using R Studio 1.1.456 software (Northern Ave, Boston, MA, USA) using three-way ANOVA and the Tukey HSD test as post hoc test for pairwise multiple comparisons, with p < 0.05. The normal distribution of the data was checked with the Shapiro-Wilk normality test.