Uninfected D. citri used were reared in a greenhouse at the National Navel Orange Engineering Research Center, Gannan Normal University, Ganzhou, China. CLas-infected D. citri were collected near Tandong orchard (latitude 25°47′5″ north and longitude 114°52′4″ east). Infected D. citri and uninfected D. citri were reared separately in cages (60 cm × 60 cm × 90 cm). For this experiment, one to two pots of healthy citrus or Murraya exotica were maintained in the rearing cages under the conditions of 27 ± 1°C and RH of 70 ± 5%, with a 14-h light/10-h dark cycle. Wild-type (WT) N. benthamiana plants were grown in a growth chamber maintained at 25 ± 2°C and an RH of 70 ± 5%, with a 16-h light/8-h dark cycle.

The full length of Vg_VWD (Gene ID: LOC103523873, XP_008487105.1, 557 aa) was obtained by quantitative real-time (qRT)-PCR using total RNA isolated from D. citri, cloned into the pCE2 TA/Blunt-Zero vectors (Vazyme, China), and then sequenced. The ClonExpress II One Step Cloning Kit (Vazyme, China) was used to insert Vg_VWD into the prokaryotic expression vector (pGEX-4T-1, GST-tagged protein), yeast two-hybrid (Y2H) vector PPR3-N (prey), vector pgR107 of potato virus X (PVX), and pull-down vector PMAL-C2X (PBM-tagged protein). The sequence of the CLas flagellin (flaA) was derived from the whole-genome sequence of the strain Psy62 (taxid: 537021, GenBank accession no. CP001677) (Duan et al., 2009). It was cloned into the Y2H vectors pBT3-N, pBT3-STE, pDHB1 (bait plasmid), pull-down vector (pGEX-4T-1, GST-tagged protein), and PVX vector pgR107.

The full flaA gene (Gene ID: CLIBASIA_02090, 1359 dp) was cloned in-frame into the vector (pBT3-N, pBT3-STE, and pDHB1) as the bait. The self-activation and functional validation of the bait plasmids referred to the methods of Than et al. (2016) and Liang et al. (2022). In brief, the NMY51 yeast cells containing the plasmid were transferred to DDO (SD/-His/−Leu with agar), TDO (SD/-His/−Leu/−Trp with agar), and QDO (SD/-His/−Leu/−Trp/−Ade with agar) for observation and recording of the growth. The certified bait plasmids were transformed into NMY51 yeast cells according to the manufacturer’s instructions to produce the yeast receptor cells containing the decoy plasmids. The D. citri membrane library plasmids were transferred into receptor cells, coated on QDO-medium, and incubated at 28°C for 3–5 days, following which the plasmids were extracted and sequenced from the grown single colonies. The obtained candidate proteins were verified by Y2H protein–protein, and the interaction intensity was verified by the β-galactosidase colorimetric reaction.

The full flaA (CLIBASIA_02090, 1,359 dp) was cloned into pGEX-4T-1 for fusion with the GST tag, and Vg_VWD was cloned into PMAL-C2X for fusion with the MBP tag. The recombinants were transformed into competent Escherichia coli Rosetta (DE3) cells for expression and purified with the GST Tag Protein Purification Kit (Beyotime Biotechnology, China) after IPTG (1 mM, 20°C for 8 h). For the experimental and control groups, 500 μg of GST and GST-flaA were added to fully bind the solutions with 50% glutathione-agarose resin at 4°C. The supernatant was removed by centrifugation, and 1 mL of PBST was added to wash off the unbound protein in the resin, following which 500 μg of the MBP/His-Vg_VWD protein was added and incubated overnight at 4°C. The products were rinsed three times with 1 mL of pre-cooled PBST, and RIPA-buffered cell lysate and loading buffer were added and mixed well, boiled for 5–10 min, and then the supernatant was collected by centrifugation. The supernatant was separated by SDS-PAGE and subjected to immunoblotting analysis.

The Co-IP assays were performed as previously described with a slight modification (Wang et al., 2016). In brief, His and Flag were tagged to the N terminus and C terminus of Vg_VWD, respectively, and HA was tagged to the N terminus of flaA. The construct fusion plasmids pFastBac1-Vg_VWD and pFastBac1-flaA were transfected into Sf9 cells to detect protein expression. The fusion plasmids were then co-transfected into Sf9 cells, and the expression of the co-transfected proteins was detected. The co-transfected cell lysate samples were collected for subsequent experiments. First, the co-transfected cell lysate was pretreated, the lysate was incubated with the protein A magnetic column to prevent non-specific binding between the lysate and the magnetic column, and the incubated flow-through was taken for Co-IP experiments. Second, the experimental and control groups were set up: the flag antibody or IgG antibody was hung onto the protein A magnetic column, then the supernatant was lysed after incubation with mixed magnetic beads (4°C, 3 h), followed by washing thoroughly with 1 mL of ice-cold PBS buffer for three times to elute the protein, which was collected for Western blot detection.

According to the morphological differentiation method for D. citri nymphs, 3–5 instar nymphs were prepared, with three biological replicates for each instar and 15 nymphs per replicate. For the collection of different gender adults, the 5th instar nymphs were selected and transferred to a single M. exotica seedling for feeding. Then, the adults were collected on days 3, 5, and 10 after emergence, respectively. Adults of different genders were distinguished under a microscope and were separately collected. There were three biological repeats and five male or female adults of D. citri per replicate.

The CLas-infected adults of D. citri were collected from orchards, where CLas-positive rates ranged between 80% and 95% (Supplementary Figure 2A). Six organs including the midgut, bacteriomes, testis, ovary, fat body, and hemolymph were dissected under the insect-dissecting microscope, and the hemolymph was collected according to Kruse et al.’s method (2018). Tissues collected from 30 D. citri adults (male:female = 1:1) were taken as one biological replicate, and three biological replicates were used in experiments. In addition, the salivary glands and Malpighian tubes were collected with three biological replicates each from 100 adults D. citri (male:female = 1:1). Dissected tissues were transferred to 1.5-mL centrifuge tubes (Trizol, 500 μL) using forceps. Subsequently, the total nucleic acids were extracted.

The DNA extraction of a single D. citri via method A was used to detect the CLas-infection rate of D. citri in the field and method B was used to detect the CLas titer in infected D. citri. Method A is performed according to the instructions for the Animal Tissue Direct PCR Kit as follows: (1) the D. citri were kept on ice for 3–5 min; (2) each D. citri was transferred via forceps to PCR tubes with premixed 12.5 μL buffer AL and 0.5 μL Foregene protease; (3) mashed with a small sterilized grinding rod homogenate, and the homogenate was treated under the conditions of 65°C for 20 min and 95°C for 5 min; and (4) centrifuged at 12,000 rpm for 5 min. The PCR detection system has been carried out according to Shen et al. (2022). Two μL of the supernatant obtained from method A was aspirated as the PCR template with primers OI1 and OI2 (Supplementary Table 1). Method B extracted DNA from D. citri according to the CTAB method, and the specific steps were performed according to Quintana et al. (2022). A two-step assessment of CLas titer in a single D. citri was addressed: (1) the presence of CLas was detected by ordinary PCR; and (2) the CLas titer in infected D. citri was quantified by qPCR.

To perform qPCR, the total DNA of a single CLas-infected D. citri was diluted to 100 ng/μL, and the copy number of CLas per 100 ng of total DNA was detected by absolute quantification. Probe qPCR experiments were performed with Premix (TakaRa, Dalian, China) using probe primers HLBr and HLB4G, and sequence information is provided in Supplementary Table 1. A mixture (20 μL) of Premix (10 μL), probe (0.3 μL), each primer (0.4 μL), and D. citri template DNA [1 μL (100 ng)] was reacted on a Light Cycler 96 SYBR Green I Master (Roche). The qPCR conditions were as follows: 3 min at 95°C; 40 thermal cycles (10 s at 95°C; 30 s at 60°C); and 30 s at 37°C. At least three technical replicates were performed for each sample. The equation for absolute quantification of CLas was y = −4.11x + 55.508 (R² = 0.9964), where y is the Ct value, x is the copy number, and the CLas titer is 10x per 100 ng of the D. citri DNA.

The D. citri total RNA was extracted using Trizol regent (Invitrogen, Carlsbad, CA, United States) according to the manufacturer protocol. The synthesis of cDNA was performed by using a PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Dalian, China) with 1 μg of total RNA reverse transcribed for each age of D. citri and 0.5 μg of total RNA reverse transcribed for each tissue. The cDNA was diluted 3–5 times and then subjected to RT-qPCR, and a reaction system (20 μL) was applied, including 10 μL of 2 × TB Green Premix (TaKaRa, Dalian, China), 7 μL of DNase/RNase-free water, 0.8 μL of each primer (10 μM), 0.4 μL of ROX reference, and 1 μL of diluted cDNA template. The PCR cycling consisted of an initial activation step at 95°C for 3 min, followed by 40 cycles of 95°C for 5 s and 60°C for 34 s, and a single collection at 72°C for 30 s. The DcGAPDH gene was used as an internal control. The relative expression value was calculated with three biological and technical replicates, and the calculation was accordant with the 2^−ΔΔCT quantification method (Livak and Schmittgen, 2001). The primers are listed in Supplementary Table 1.

The primers specific (Supplementary Table 1) for Vg_VWD and GFP were designed to synthesize dsRNA by using the T7 High Yield Transcription Kit (Thermo Scientific, Wilmington, DE, United States) according to the manufacturer’s instructions. The D. citri adults addressed for dsRNA injection were ca. 95% of CLas infection rates (Supplementary Figure 2B). The dsRNAs of Vg_VWD and GFP were injected as experimental and control groups, respectively, and the injected D. citri were transferred to M. exotica seedlings for rearing (Supplementary Figure 3). The total DNA of D. citri was extracted and quantified to 100 ng/μL at 6, 12, and 24 h after injection. The titer of CLas per 100 ng total DNA of treated D. citri was determined. The experiments for each group were designed with at least three biological replicates (each with five males or females) and three technical replicates.

The full-length sequences of flaA (1,359 dp) and Vg_VWD (1,674 dp) were inserted into the binary vector potato virus X (PVX) digested by ClaI/SalI (Liu et al., 2019). PVX-GFP was used as a negative control and PVX-BAX and PVX-NIF1 as positive controls. The plasmids PVX-GFP, PVX-BAX, PVX-NIF1, PVX-flaA, and PVX-Vg_VWD were transformed into A. tumefaciens GV3101 for culturing, which then were centrifuged and resuspended with the buffer [10 mM 2-(N-morpholino) ethanesulfonic acid (MES), 10 mM MgCl₂, and 100 μM acetosyringone] to OD₆₀₀ = 0.6. After maintaining the suspension at room temperature and in the dark for 2 h, it was infiltrated into four to six leaves of N. benthamiana using sterile syringes. The symptoms were observed and photographed later.

For the callose deposition assay, the sampled N. benthamiana leaves were incubated in a mixed solution of acetic acid:glycerol:ethanol (1v:1v:3v) until the green color faded away, and then rinsed with 150 mM K₂HPO₄ for 30 min. Finally, samples were stained with aniline blue solution (150 mM K₂HPO₄, 0.05% w/v aniline blue) for 2 h, and the callose deposition was observed via fluorescence microscope using a DAPI filter (excitation filter 390 nm; dichroic mirror 420 nm; emission filter 460 nm) (Schenk and Schikora, 2015).

A phylogenetic tree of D. citri for Vg_VWD and 11 other insect Vg-VWDs was constructed using the neighbor-joining (NJ) method with 1,000 bootstrap replicates in MEGA 7. Sequence information is provided in Supplementary Table 3.

All statistical analyses were performed using SPSS 24.0 software. A one-way ANOVA was used, followed by least-significant difference (LSD) multiple comparisons tests. Pairwise comparisons were performed by an independent samples t-test. Graphs were illustrated using GraphPad Prism 8.2.1 software. Data were expressed as means ± standard deviation.

To find D. citri proteins targeting CLas flagellum proteins, flaA was used as bait. The pDHB1-flaA plasmid passed the self-activation assay and functional validation (Supplementary Figure 1), and a Y2H screen was performed on the D. citri membrane library. As a result, 16 candidate proteins with potential interactions with flaA, of which 11 were annotated and five were unannotated (Supplementary Table 3). Following structural domain analysis and functional prediction, a vitellogenin-1-like protein XP_008487105.1 (named Vg_VWD thereafter) containing a conserved domain-VWD with 189 amino acids was selected for further analysis (Figure 1A). The phylogenetic analysis showed that the Vg-VWD sequences of D. citri and other representative Hemiptera insects clustered together. In addition, the Vg_VWD sequences of D. citri and the potato psyllid (Bactericera cockerelli) sequence were closest in the evolutionary relationship (Figure 1A).

The verification of the flaA and Vg_VWD cotransformants showed that the yeast grew well and formed a certain gradient on DDO and QDO (Figure 1B). The point-to-point Y2H method and β-galactosidase assays showed that flaA-Vg_VWD interacted in the yeast system (Figures 1C,D). Subsequent GST pull-down and Co-IP assays further confirmed the interaction between flaA and Vg_VWD (Figures 1E,F). Taken together, these results indicate that flaA interacts with Vg_VWD in vivo and in vitro.

Quantitative RT-PCR was used to measure the transcript levels of the Vg_VWD gene in D. citri. In uninfected D. citri, the expression level of Vg_VWD gradually decreased in 3–5 instar nymphs, and Vg-VWD expression levels increased in females and decreased in males at 3–7 days after emergence (Figure 2A). The transcript abundance of Vg_VWD in CLas-infected D. citri was much higher than that in uninfected D. citri (p < 0.05), especially in females. This induction was 74,227-fold (CLas-infected/uninfected) in females, but only 38.3-fold (CLas-infected/uninfected) in males. In addition, the transcript abundance of Vg_VWD in CLas-infected females was 36,289-fold higher than that in CLas-infected males. These results indicate that Vg_VWD is expressed at significantly higher levels in females than males after infection (p < 0.05) (Figures 2B,C).

The proteomics analysis showed that Vg was upregulated in the hemolymph of the CLas-infected D. citri (Kruse et al., 2018). Here, we further analyzed the expression patterns of the 16srRNA and Vg_VWD genes of D. citri in the midgut, salivary glands, malpighian tubules, bacteriomes, testes, ovaries, fat body, and hemolymph of the CLas-infected D. citri using RT-qPCR. The gene expression levels of 16S rRNA in the midgut and salivary gland were significantly higher than those in the other tissues (p < 0.05), while the transcription of Vg_VWD was highest in the fat body, followed by the salivary glands and bacteriomes (Figure 2D).

As insect hemolymph and salivary glands play important roles in the transmission of pathogens (Jiang et al., 2019; Chen et al., 2021), we next compared the transcription of Vg_VWD in the salivary glands and hemolymph of infected and uninfected D. citri. The results showed that the expression levels of Vg_VWD in the salivary gland and hemolymph were 338.39-fold and 61.67-fold higher in infected D. citri than those in uninfected D. citri, respectively (Figure 2E). Collectively, these results demonstrate that CLas acquisition induces Vg_VWD transcription in D. citri, with significant differences between males and females (p < 0.05).

Next, according to the characteristics of Vg_VWD expression (Figure 2A), we constructed Vg_VWD-silenced D. citri using dsVg_VWD-mediated RNAi silencing to determine whether Vg_VWD is involved in CLas proliferation. After dsVg_VWD injection in CLas-infected D. citri, the expression of Vg_VWD was reduced by 70% compared with the same dsGFP injection dose for 24 h (Figure 3A). The CLas titer was detected in the CLas-infected male and female of D. citri at 6, 12, and 24 h after the injection. The results showed that the CLas titer of female D. citri was significantly different from the control group at 12 and 24 h (Figure 3B), with an increase of 1.76 and 1.58 times (CLas copies in 100 ng of female D. citri DNA), respectively. There was a significant difference in the CLas titer in male D. citri at 24 h compared with the control group, which was 1.98-fold higher (CLas copies in 100 ng of male D. citri DNA), whereas no significant differences were observed at 6 and 12 h (Figure 3C).

Since Vg_VWD was highly expressed in the salivary glands of the CLas-infected D. citri and directly interacts with CLas flagellin, we speculated that Vg_VWD may attach to the CLas flagellin protein during psyllid feeding along with the secretion of psyllid saliva into the citrus phloem, which is likely to be a stress factor. The mouse BAX and Phytophthora infestans INF1 are well-known inducers of cell death and are widely used to identify the PCD inhibitors of pathogens (Kamoun et al., 1998; Lacomme and Cruz, 1999). Vg_VWD was transiently expressed in N. benthamiana to evaluate its inhibitory effect on PCD, and the results showed that it could inhibit the BAX- and INF1-induced hypersensitivity response (Figures 4B,C). Callose deposition is one important indicator of plant immunity reaction, and the CLas flagellin has PAMP-triggered immunity and causes N. benthamiana leaf necrosis and callose deposition (Zou et al., 2012). Our results showed that Vg_VWD attenuates flaA-induced callose deposition in N. benthamiana (Figure 4D). Therefore, these results suggest that Vg_VWD can suppress N. benthamiana immunity and inhibit the phenomenon of callose deposition induced by flaA.