2.1 Plant materials and growth conditions

A. thaliana Columbia-0 ecotype (Col-0) plant seeds were plated on half-strength Murashige–Skoog (½ MS) media supplemented with 1% (w/v) sucrose and 0.8% (w/v) agar. The seedlings were subsequently transferred to soil and grown under a long-day (16-h light and 8-h dark) photoperiod in a growth room at 22°C. The growth conditions of N. benthamiana were the same as those described previously (Liu et al., 2010). p35S::PHYL1_JWB transgenic plants were generated using the full-length PHYL1_JWB coding sequences. The sequences of the PHYL1_JWB used for generating the transgenic plants are listed in File S1. The primers used for this study are listed in Table S1.

2.2 RNA extraction, reverse transcription and qPCR

The Arabidopsis plants to be used in gene expression assays were grown on ½ MS plates for 14 days under long-day conditions. Approximately 50 mg of plant tissue from three or four individual seedlings of the indicated genotype was collected as a single sample. Three biological replicates of each genotype were analysed. RNA extraction was performed using the RNAprep Pure Plant Kit (TIANGEN, DP441). The RNA was reverse-transcribed into complementary DNA (cDNA) with FastKing gDNA Dispelling RT SuperMix (TIANGEN, KR118). qPCR was performed with total cDNA using SuperReal PreMix Plus SYBR Green (TIANGEN, FP205). Three biological replicates of each treatment were performed. Threshold cycle values were calculated using iCycler software, and ZjACT or AtACT was used as an internal control. Relative transcript levels were calculated according to the 2^–ΔΔCT method (Bu et al., 2016).

2.3 Vector constructs

Entry clones encoding full-length AP1, AGL6, SEP1, SEP2, SEP3, SEP4, SOC1, and CAL from Arabidopsis were recombined into pCam1300-3FLAG. In addition, the entry clones encoding full-length PHYL1_JWB were recombined into pCam1300-GFP vectors. All the constructs and empty vector controls were transformed into competent Agrobacterium strain GV3101 using the freeze‒thaw method. The cultured cells were harvested, resuspended in 10 mM MgCl₂ plus 150 μM acetosyringone (Sigma‒Aldrich) and subsequently maintained at 25°C for at least 3 h without shaking. Agrobacterium suspensions were infiltrated into the leaves of N. benthamiana with a needleless syringe, and the proteins were then detected and coimmunoprecipitated.

2.4 Yeast two-hybrid (Y2H) analysis

We performed the initial Y2H screen of PHYL1_JWB against a fruit library of ‘Dongzao’. The full-length sequence of PHYL1_JWB with the secretory signal peptide was cloned and inserted into a pGBKT7 bait plasmid as a C-terminal fusion to the GAL4 domain. The prey library was constructed from the fruit cDNA library with pGADT7 as the prey plasmid. The candidate interactions were confirmed by Y2H assays (Joo et al., 2020; Teresinski et al., 2023). Yeast growth on medium lacking leucine and tryptophan (-L-W) indicated the presence of AD and BD constructs, and growth on medium lacking leucine, tryptophan and histidine (-L-W-H) and medium lacking leucine, tryptophan, histidine and alanine (-L-W-H-A) implied interactions between the AD and BD fusion proteins. The yeast plates were maintained in growth chambers at 28°C for 3 days before imaging.

2.5 Gene synthesis

The sequences of the different PHYL1_JWB mutants, AP1 (K → R), and AGL6 (K → R) were synthesised by AZENTA and subsequently constructed into the pGBKT7 and pCam1300-3Flag plasmids for Y2H and protein-level analyses. The sequence information is shown in File S1.

2.6 Western blotting analysis

Agrobacteria harbouring plasmids expressing FLAG-tagged AP1, AGL6, SEP1, SEP2, SEP3, SEP4, SOC1, and CAL were infiltrated into N. benthamiana leaves together with Agrobacteria harbouring plasmids expressing GFP-PHYL1_JWB in half of each leaf. At 42 h post-infiltration (hpi), approximately 0.1 g of N. benthamiana leaves expressing the indicated proteins were collected and ground into powder with liquid nitrogen. Then, twofold SDS buffer (100 mM Tris HCl pH 8.0, 0.2% SDS and 2% β-mercaptoethanol) was added to each protein sample, and the samples were boiled for 5 min before loading. For drug treatment, N. benthamiana leaves expressing the indicated proteins were injected at a final concentration of 20 mM MG132 (Sigma) at 8 h before the sampling period. A stock solution of 10 mM MG132 was prepared in dimethylsulfoxide (DMSO). An equivalent volume of DMSO was used as a mock control. The protein levels were detected by western blotting. Whole protein extracts from N. benthamiana leaves were separated on NuPAGE 10% Bis-Tris Gels (Invitrogen) and transferred to 0.45-µm polyvinylidene difluoride membranes (Thermo Scientific) using the Bio-Rad mini-PROTEAN Electrophoresis system. The membranes were blocked by incubation in 5% (w/v) milk powder in TSBT phosphate-buffered saline and 0.1% (v/v) Tween-20 for 1 h at room temperature. The membranes were incubated with primary antibodies (anti-FLAG (Sigma catalog no. F1804; 1:2000 dilution) and anti-GFP (Roche catalog no. 11814460001)) overnight at 4°C and then with horseradish peroxidase-conjugated secondary antibodies (Abcam) for 1 h at room temperature and washed with TBST. The blots were developed via the enhanced chemiluminescence method (EpiZyme).

2.7 Coimmunoprecipitation

For transient expression of the Flag-tagged AP1 or AGL6 proteins and GFP-tagged PHYL1_JWB in N. benthamiana leaves of approximately 4-week-old plants, the plants were infiltrated with an Agrobacterium suspension (OD₆₀₀ = 0.5). Two days later, approximately 2 g of tissue from the infiltrated area was collected and frozen in liquid nitrogen. The tissue was ground into powder using a mortar and pestle. All subsequent steps were conducted on ice or in a cold (4°C) room. Briefly, approximately two volumes of extraction buffer (10% glycerol, 25 mM Tris-HCl (pH 7.5), 1 mM EDTA, 150 mM NaCl, 0.15% NP-40, 1 mM NaF, 1 mM PMSF, 10 mM DTT, 2% PVPP, and 1× protease inhibitor cocktail from Roche) with MG132 were added to each sample to homogenise the powder. The resuspended samples were centrifuged at 14 000 rpm for 10 min, and the supernatant was subsequently transferred to 2-mL microcentrifuge tubes. The supernatant was centrifuged again to remove additional debris. Afterward, the proteins were transferred to a new tube containing anti-FLAG-conjugated beads (Sigma) and incubated for 2 h. The proteins were collected by centrifugation and washed four times with extraction buffer. Proteins bound to the beads were eluted by adding 1× SDS loading buffer (preheated to 95°C), and the elution was then incubated for 5 min at room temperature. The eluted proteins were analysed by western blotting using an anti-FLAG antibody (Sigma) or anti-GFP antibody (Roche).

2.8 Fluorescence microscopy observations

The sequence about base mutant of PHYL1_JWB were synthesised and constructed to pCam1300-GFP plasmids for the subcellular localisation. Agrobacterium with recombinant were injected into tobacco leaves, and the related experiments were performed according to the method described before (Zhao et al., 2024). N. benthamiana epidermal cells were observed 40 h after agroinfiltration using an Axio Imager M2 (ZEISS) with a ×20 immersion objective. GFP was excited by a 488 nm argon laser, and the emission was visualised at 509 nm. The images were scanned and captured by z-stack acquisition during line and frame switching.

2.9 Statistical analysis

The error bars in all of the figures represent the standard deviations. Statistical comparisons among different samples were performed by either one-way analysis of variance with Tukey's honestly significant difference post hoc test or Student's t test, as indicated in the figure legends.

3.1 PHYL1_JWB alters the flower architecture and reproduction of plants

Functional screens with candidate JWB phytoplasma effectors revealed that PHYL1_JWB perturbs plant developmental processes, as revealed by analysing the phenotypes of stable transgenic A. thaliana lines that constitutively express the PHYL1_JWB gene. These PHYL1_JWB-overexpressing plants exhibited a range of architectural differences compared with wild-type Col‒0 plants (Figure 1; Figure S1). At 4 weeks, the PHYL1_JWB-overexpressing plants presented cauliflower-like unclosed flower buds with thickened sepals full of solid thorns and greenish petals compared with the wild-type plants (Figure S2). Moreover, the PHYL1_JWB-overexpressing plants exhibited a delayed flowering time compared with the wild-type plants. In addition, the transgenic plants produced pistils that broke through the envelopes of sepals and petals (Figure S2d). Analysis of the flower bud structures showed that the sepals and petals of the transgenic plants were too short to parcel the stigma at the early stage (Figure S2a,b), indicating that PHYL1_JWB affects the formation of the flower bud architecture.

At 5 weeks, comparied to wide type (Figure 1a), the p35S::PHYL1_JWB transformed plants exhibited various aberrant floral phenotypes, characterised by thickened sepals and chlorophyll accumulation in petals. Additionally, inflorescence-like structures emerged from one or all of the stamens and pistils, leading to abnormal organ development in flowers (Figure 1b). Furthermore, premature seed pod swelling and reduced length were observed in PHYL1_JWB-overexpressing plants [Figure 1c(3)]. Subsequently, the seed coat colour transitioned to purple, accompanied by persistent petal and sepal retention [Figure 1c(4)]. The phenotype bears resemblance to that of diseased jujube, exhibiting branches that remain steadfast even in the grip of winter (Figure S2c). Taken together, these findings showed that the fertility of the PHYL1_JWB-overexpressing plants was strongly compromised (Figure 1c), especially the plants with the most severely diseased leaves/flowers (Figure S2e), which were completely sterile. This finding is similar to the phyllody observed in diseased jujube (Figure 1d), indicating that PHYL1_JWB has a severe effect on plant fertility. At approximately 7 weeks, the wild-type plants had aged and died, but the transformed plants continued to grow (Figure S2f), suggesting that PHYL1_JWB prolonged the lifespan of the host.

3.2 PHYL1_JWB can interact with A/E-type proteins and RAD23C/D

Based on the abnormal flower development induced by PHYL1_JWB, we wondered whether PHYL1_JWB interacts with proteins related to flower development, such as proteins related to the ABCDE model (Alejandra Mandel et al., 1992; Ditta et al., 2004; Goto & Meyerowitz, 1994; Jack et al., 1992; Jofuku et al., 1994; Pelaz et al., 2000; Rijpkema et al., 2010; Schauer et al., 2009; Smaczniak et al., 2012; Yanofsky et al., 1990). First, we evaluated the interactions between these proteins by using Y2H assay. Interactions between PHYL1_JWB and class-A/E proteins (AP1 and SEP1-4 in Arabidopsis and MADS30, 31, 47, and 48 in jujube), but not class-B/C/D proteins, were detected in both Arabidopsis and jujube (Figure 2). The AGL6 and CAL proteins from Arabidopsis also interacted with PHYL1_JWB (Figure 2). Moreover, we selected AP1 and AGL6 to confirm their interactions with PHYL1_JWB by Co-IP in vivo. As shown in Figure 2d,e, GFP-PHYL1_JWB also pulled down AP1 and AGL6 from N. benthamiana leaves, indicating that PHYL1_JWB interacts with the AP1 and AGL6 TFs. Moreover, the K domains of AP1 and SEP3 from both jujube and Arabidopsis were sufficient to mediate binding to PHYL1_JWB (Figure S3).

We also conducted an Y2H screen against a jujube cDNA library. Two proteins, Rad23C and Rad23D, were identified as potential interactors of PHYL1_JWB and further confirmed (Figure 3a). The A. thaliana 26S proteasome subunit RAD23 is located within the 19S regulatory particle of the 26S proteasome and serves as one of the main ubiquitin receptors that recruit ubiquitinated proteins for proteasomal degradation (Dantuma et al., 2009; Farmer et al., 2010; Saeki, 2017; Tsuchiya et al., 2017). RAD23C is composed of four main domains: UBL, UBA1, XPC-binding, and UBA2 (Figure 3b; Figure S4). The UBL domain is needed for proteasome association, and the N-terminal half of the domain with an ubiquitin-interacting motif (UBA2) is involved in recruiting ubiquitinated substrates. We found that PHYL1_JWB interacted with the UBA2 domain but not the UBL domain of Rad23C (Figure 3b).

3.3 PHYL1_JWB mediates the degradation of A/E-type proteins

Because PHYL1_JWB interacted with the 26S proteasome subunit RAD23C, we investigated whether PHYL1_JWB affects the protein levels of class A and E TFs in plant cells. We transiently co-expressed A/E genes and PHYL1_JWB (or GFP as a control) under the control of the constitutive 35S promoter in N. benthamiana leaves (Figure 4a). The SEP1-4 and CAL proteins were absent or less abundant in the presence of PHYL1_JWB compared with GFP (Figure 4b). Moreover, the AGL6 and AP1 protein levels were also reduced by PHYL1_JWB expression (Figure 5a,b). We also found that PHYL1_JWB mediated the degradation of ZjMADS31 (the homologous protein of AP1) and ZjMADS47 (the homologous protein of SEP3) (Figure S5). Through transient co-expression of AP1 with a luciferase tag and PHYL1_JWB (a control), we also found that PHYL1_JWB could reduce the Luc signal of AP1 protein after treatment with luciferin substrates in N. benthamiana leaves (Figure S6). These data showed that PHYL1_JWB mediates the destabilization of the AP1, class E, AGL6, and CAL TFs in plant cells.

Because class B/C/D genes are downstream genes of the class A/E during flower development (Liu et al. 2009; Lamb et al., 2002), we investigated whether PHYL1_JWB affects the expression of the class B and C/D genes in Arabidopsis and jujube by qRT-PCR. The results showed that the expression of class B (AtAP3) and class C/D (AtAG) genes was significantly lower in 35S::PHYL1_JWB plants than in wild-type plants (Figure S7a). Similarly, the transcriptional levels of ZjMADS39 (class B) and ZjMADS46 (class C/D) were greater in phyllody leaves than in healthy jujube flowers (Figure S7b), indicating that the degradation of class A and E proteins mediated by PHYL1_JWB may contribute to the suppression of class B and C/D gene expression.

3.4 PHYL1_JWB mediates the degradation of AP1 and AGL6 independent of the ubiquitination of substrates

Because Rad23C functions as one of the main ubiquitin receptors by recruiting poly-ubiquitinated proteins for proteasomal degradation (Farmer et al., 2010; Lin et al., 2011), we investigated whether the ubiquitination of lysine residues within PHYL1_JWB targets is necessary for their degradation. AtAP1 and AtAGL6 proteins in which all lysines were replaced by arginines were more abundant than wild-type proteins, as revealed by transient expression assays. Nonetheless, both the AP1 and AGL6 mutants were degraded in the presence of PHYL1_JWB (Figure 5a,b). Thus, lysine ubiquitination of the substrates may not be needed for their PHYL1_JWB-mediated degradation. These results indicate that the PHYL1_JWB-mediated target degradation occurs independently of substrate ubiquitination.

3.5 The PHYL1_JWB-mediated degradation of AP1 and AGL6 relies on the 26S proteasome

The 26S proteasome is the main pathway for protein degradation in eukaryotic cells (Ji & Kwon, 2017; Marshall et al., 2021). Therefore, we further investigated the effect of the 26S proteasome on the PHYL1_JWB-mediated degradation of the AP1 and AGL6 proteins. MG132, a potent 26S proteasome inhibitor, was co-injected into tobacco leaves with the expression vectors for PHYL1_JWB and AP1 or AGL6, and the protein stability was then examined. As shown in Figure 5c,d, the addition of MG132 inhibited the PHYL1_JWB-mediated instability of the AP1 and AGL6 proteins compared with the addition of DMSO (control). The above described results indicated that the AP1 and AGL6 proteins are destabilised by PHYL1_JWB through the host 26S proteasome.

3.6 The second α helix is needed for the interaction ability of the PHYL1_JWB protein

Prediction of the secondary and tertiary structure (https://npsa-prabi.ibcp.fr/cgi-bin/npsa_automat.pl?page=npsa_sopma.html) of the PHYL1_JWB protein revealed the presence of two α helixes, namely, α1 (amino acids from position 21 to 47) and α2 (amino acids from position 54 to 86), two random coils (amino acids from position 1 to 20 and from position 48 to 53), and the two helixes were found to be separated one of the random coils (amino acids from position 48 to 53) (Figure S8). To determine whether deletion of these fragments affected the interaction of PHYL1_JWB with AP1 and SEP1–4, the mutant with deletion of different fragments of PHYL1_JWB or mutant with amino acid substitutions at specific sites were produced to verify the interactions of these proteins by Y2H assay.

A schematic diagram of the PHYL1_JWB with fragments deletion mutants is shown in Figure 6a. Y2H assays were performed to compare the interactions of these PHYL1_JWB deletion mutants with AP1 or class E proteins from Arabidopsis and jujube. The results showed that colonies harbouring PHYL1_JWBΔ54–86 and AP1 or class E proteins (SEP1, SEP2, SEP3, and SEP4) derived from jujube and A. thaliana did not grow on -LWH or -LWAH media, indicating that deletion of the second α helix (Δ54–86) directly inhibited the binding ability of PHYL1_JWB (Figure 6b,c). In addition, yeast colonies containing the coexpressing mutants PHYL1_JWBΔ21-47 and AP1 did not grow on either -LWH or -LWAH media (Figure 6b), showing that the first α helix of PHYL1_JWB was also important for the interaction with AP1. In summary, two α-helix domains are pivotal for the ability of PHYL1_JWB to bind to AP1 in jujube and Arabidopsis.

3.7 Leu 75 is an essential functional site in PHYL1_JWB

To further confirm which site in PHYL1_JWB influences its interaction ability, we also obtained PHYL1_JWB mutants with amino acid substitutions in the two α helices, which are based on the amino acid substitution position of SAP54 (Aurin et al., 2020). The amino acid sequences are shown in Figure 7a, and a Y2H assay was conducted between these PHYL1_JWB mutants and MIKC-type proteins derived from jujube and Arabidopsis. The results showed that the interaction ability of PHYL1_JWB decreased when leucine at position 75 was replaced with proline (PHYL1_JWB-L75P). In the first α helix, simultaneous replacement of the amino acids at positions 27 and 37 caused PHYL1_JWB to lose its ability to interact with ZjSEP2 and ZjSEP3 (Figure 7b), indicating that these two amino acids play key roles in its binding ability. However, substitution of three sites in the first α helix did not affect the binding of PHYL1_JWB to class E proteins in Arabidopsis (Figure 7c; Figure S9).

Previous studies have found that the K domain of SEP3 is the key domain involved in its interaction with phyllody effectors (Liao et al., 2019); thus, the interaction between the K domain of SEP3 and these mutants was also verified by Y2H assay. As shown in Figure 7c and Figure S3, Leu 75 in the second α helix was found to be essential for interaction with the K domain of SEP3, and substitution at the first α-helix amino acid site did not affect the binding to the K domain but affected the interaction with full-length SEP3. This finding suggested that other domains in ZjSEP3 may also affect its binding to the first α helix. The PHYL1_JWB mutant in which Leu at position 26 in the first α helix was replaced by Pro (PHYL1-L26P) could not interact with the K domain of AP1 derived from jujube and Arabidopsis (Figure 7c; Figure S3). Lysine at position 37 is also necessary for the binding of PHYL1_JWB to the K domain of AP1 (Figure 7c; Figure S3). These results suggested that leucine at position 26 and lysine at position 37 are the binding sites in the first α helix and that leucine at position 75 is necessary for the function of the second α helix.

To determine whether the absence of each helix has an effect on the subcellular localisation of PHYL1_JWB, a subcellular localisation analysis of the PHYL1_JWB mutants with a GFP tag was performed in N. benthamiana. The results showed that deletion of either the first or second a helix did not affect the localisation of PHYL1_JWB in the nucleus or cytoplasm (Figure S10). And, the change in the amino acid at site 75 of PHYL1_JWB also did not affect its localisation (Figure S10). This finding suggested that conformational changes in the protein structure, rather than changes in the subcellular localisation, caused by deletion of helical fragments and changes in specific amino acids contribute to loss of the binding capacity of PHYL1_JWB.