Clone of the parthenogenetic grain aphid S. miscanthi was initially established from single aphid collected from a wheat field in Langfang Experimental Station, located in Langfang city, Hebei Province, northern China, and maintained under laboratory conditions (L: D = 16 h: 8 h, 20 ± 1°C) on wheat plants (cv Mingxian 169) for more than 3 years (25-30 generations per year). Winter wheat (Triticum aestivum) cultivars Mingxian169 were grown in a climate chamber at 20°C ± 1°C under 16-h light/8-h dark photoperiod until reaching the two-leaf stage (12 day-old). Tobacco plants N. benthamiana were grown in an environmentally controlled greenhouse at 23°C ± 1°C under 16-h light/8-h dark photoperiod. N. benthamiana plants were used in transient transformation experiments after 4–5 weeks.

The multiple alignment of amino acid sequences was performed using Clustal Omega (RRID : SCR_001591) (https://www.ebi.ac.uk/Tools/msa/clustalo/). A phylogenetic tree was constructed by the neighbor-joining method via MEGA7.0 (RRID : SCR_000667). Bootstrap values were calculated as a percentage from over 1000 replications. The signal peptides were predicted using the SignalP 4.1 server (RRID : SCR_015644) (http://www.cbs.dtu.dk/services/SignalP/) and iPSORT (http://ipsort.hgc.jp/). The protein molecular weight was predicted by the Compute pI/Mw tool (http://web.expasy.org/com-pute_pi/) on ExPASy (RRID : SCR_007944). The prediction of transmembrane regions, Pfam domains and nuclear localization signals in protein sequences was performed using TMHMM 2.0 (RRID : SCR_014935) (http://www.cbs.dtu.dk/services/TMHMM-2.0/), Pfam (RRID : SCR_004726) (http://pfam.xfam.org/) and NLStradamus (Nguyen Ba et al., 2009) (http://www.Moseslab.csb.utoronto.ca/NLStradamus/), respectively. The prediction and localization of protein motifs were conducted in MEME (RRID : SCR_001783) (http://meme-suite.org/tools/meme).

The ORF sequences of two effectors without an N-terminal signal peptide were cloned into pBWA(V)HS-GLosgfp vector (BioRun, Wuhan, China) with TagGFP fused at the C-terminus using Eco31I restriction enzymes (Thermo Fisher Scientific, MA, USA), and all constructs were then transformed into Agrobacterium tumefaciens strain EHA105 (Biomed, Beijing, China). After culturing overnight at 28°C, the cells were harvested and resuspended in infiltration buffer (10 mmol/L 2-(N-morpholino) ethanesulfonic acid, 20 mmol/L acetosyringone, and 10mmol/L MgCl₂, all reagents from Sigma-Aldrich, MO, USA) to reach an optimal density at an OD₆₀₀ of 0.3. The suspensions were kept at room temperature for 4-6 h and then infiltrated into N. benthamiana (4-week-old) leaves with a 1 mL syringe without a needle. The infiltrated leaves were collected and observed using a Zeiss LSM 880 laser confocal microscope (Zeiss, Jena, Germany) post four days infiltration. GFP and chlorophyll autofluorescence were excited at 561 and 488nm, respectively, and collected at 586–647 and 680–700nm, respectively. Three biological replicates were performed.

Predicted signal peptides of Sm10 and SmC002 were functionally validated using the yeast secretion system as reported previously (Oh et al., 2009). Briefly, the DNA fragments of Sm10, SmC002, Avr1b (positive control) and the Mg47 (negative control) signal peptide were introduced into pSUC2T7M13ORI (pSUC2, provided by Professor Zhongyue Wang from Chinese Academy of Agricultural Sciences) by EcoRI and XhoI (Thermo Fisher Scientific, MA, USA) restriction sites (all primers used in this study are listed in Supporting information Table S1 ) and then transformed into the yeast strain YTK12 (Coolaber, Beijing, China). Transformants were grown on yeast minimal medium with sucrose in place of glucose at 30°C. To assay for invertase secretion, colonies were replica-plated on YPRAA plates containing raffinose and lacking glucose. The activity of invertase was further determined by monitoring the reduction of 2,3,5-triphenyltetrazolium chloride (TTC) (Sigma-Aldrich, MO, USA) dye to the insoluble red colored triphenylformazan (Oh et al., 2009). The yeast cell pellet was collected, washed, and resuspended in distilled sterile water, and an aliquot was incubated at 35°C for 35 min with 0.1% of the colorless dye TTC. Colorimetric changes were recorded after 5 min of incubation at room temperature. Three independent biological replicates were performed.

Candidate effector gene sequences were amplified using the Phusion high fidelity DNA polymerase (Thermo Scientific, MA, USA) according to the manufacturer’s instructions. All primers used in this study were designed using Primer Premier 5.0 (PREMIER Biosoft, Canada) and are listed in Supporting Information Table S1 . The pEDV6: candidate effector recombinant plasmids were constructed using Gateway technology (Invitrogen, CA, USA) as described previously (Cheng et al., 2017). A stable T3SS delivery system of P. fluorescens has been developed and applied to deliver pathogen effectors into plant cells (Sohn et al., 2007; Thomas et al., 2009; Yin and Hulbert, 2011). The pEDV6: Sm10 and pEDV6: SmC002 vectors were constructed, and T3SS of P. fluorescens was then used to deliver candidate effectors in wheat leaves. The pEDV6: Sm10, pEDV6: SmC002 and pEDV6: AvrRpt2 (positive control) (Cheng et al., 2017) constructs and pEDV6 empty vector (negative control) were transformed into the P. fluorescens EtHAn strain by electroporation. The pEDV6 empty vector, pEDV6: AvrRpt2 and EtHAn strain were provided by Academician Dr. Zhensheng Kang from Northwest A & F University. Recombinant strains of EtHAn were grown in KB liquid medium for 48 h and resuspended in infiltration medium (10 mM MgCl₂) after being washed two times using 10 mM MgCl₂ (Sigma-Aldrich, MO, USA). EtHAn suspensions were infiltrated at an OD₆₀₀ of 1.5 into the second leaves of wheat seedlings at the two-leaf stage (12-day-old) using a syringe without a needle. The infiltrated plants were grown and maintained in a cultivation room at 25°C for 2 days with a 16h: 8h light: dark cycle.

Wheat leaves infiltrated with recombinant EtHAn carrying the empty vector or the bacterial avirulence protein AvrRpt2 were used as negative or positive controls, respectively. After 2 days of infiltration, the accumulation of hydrogen peroxide in wheat leaves was examined by staining with 1 mg/ml diaminobezidin (DAB) (Sigma-Aldrich, MO, USA) solution (10mmol/L Na₂HPO₄, pH3.8, Sigma-Aldrich, MO, USA) as described previously (Wang et al., 2007). The infiltration area of the wheat leaves was cut off, immediately immersed in DAB solution, and dyed for 24 hours in the dark. To visualize necrotic cell death, after washing with distilled water, leaf segments were fixed and destained in ethanol/acetic acid (1:1 v/v). The stained leaves were imaged using an Olympus SZX-16 (Olympus Corporation, Japan). Three independent biological replicates were performed.

For the visualization of callose, the leaves infiltrated with P. fluorescens EtHAn at 2 days were fixated and destained in 1: 3 acetic acid/ethanol (v/v) solution overnight. Callose deposition in the infiltrated leaves was detected using aniline blue diammonium salt solution (75 mmol/L K₂HPO₄, pH9.5, Sigma-Aldrich, MO, USA) according to the histochemical methods described by Hood and Shew (1996). Callose deposits were observed and photographed with an Echo Revolve Hybrid Microscope (Echo, USA) using a DAPI filter. Thirty sites (1.0 mm²/site) were selected randomly from infiltrated areas of each treated leaf, and the number of callose deposits was counted from each site using Echo Revolve Hybrid Microscope. Three independent biological replicates were conducted.

Total RNA from aphid S. miscanthi and wheat leaves was extracted with the TRIzol reagent (Invitrogen) according to the manufacturer’s protocols. After 6 hours starvation, five adults S. miscanthi were fed on healthy wheat for 0h, 6h, 12h, 24h, 48h, and then collected to characterize the transcript levels of Sm10 and SmC002 in S. miscanthi at different feeding stages. To detect the expression of glucan synthase genes and defense genes in wheat, leaf tissues from the infiltration sites of each plant were collected using sterilized scissors at 2 d and 4 d post inoculation. All of the samples were transferred to liquid nitrogen immediately and stored at −80°C until use. First-strand cDNA was synthesized from 1 μg RNA using TransScript One-Step gDNA Removal and cDNA Synthesis SuperMix (TransGen Biotech, Beijing, China). The Sm10 and SmC002 specific primers for qPCR were quoted from Zhang et al. (2017b). β-actin of aphid was used as internal reference gene (Yang et al., 2014). The marker gene of the phenylpropanoid pathway phenylalanine ammonia-lyase (PAL) and the induced salicylic acid (SA) marker protein pathogenesis-related protein 1 (PR-1), and the genes tested for the jasmonic acid (JA)-responsive pathway included lipoxygenase (LOX) and Ω-3 fatty acid desaturase (FAD) were selected as target genes for the detection of wheat defense responses (Zhang et al., 2017a). Gene-specific primers for callose synthase genes were synthesized according to a previous study (Fu et al., 2014). The housekeeping gene β-actin of wheat was used as a reference (Liu et al., 2011; Zhang et al., 2017a). All primers are listed in Supplementary Table S1 . RT-qPCR was performed using an ABI 7500 Real-Time PCR System (Applied Biosystems, CA, USA). cDNA was diluted 10-fold and used as the templates to examine the relative expression of genes in a 20 μL reaction system containing 2 μL of cDNA, 0.5 μL each of 10 μmol/L forward primer and reverse primer, 10 μL of 2× SYBR premix Ex Taq (Tli RNaseH Plus, Takara), and 0.4 μL of 50× ROX Reference Dye II (Tli RNaseH Plus, Takara) under the following conditions: 30 s at 95°C, followed by 40 cycles of 30 s at 95°C and 30 s at 60°C. The relative expression level of each gene was determined by using the 2^−ΔΔCT method (Livak and Schmittgen, 2001). There were three biological replicates for each treatment and time point.

Quantifications of phytohormone levels were conducted according to a method described previously (Wu et al., 2007). In brief, at 2 d and 4 d post infiltration, approximately 500 mg infiltrated leaves from 10 individual plants with the same treatment were harvested and pooled to generate one biological replicate. Samples were homogenized in 1 mL acetonitrile (1% famic acid). Four nanogram of D4-SA, D6-JA, 1 ng JA-Ile-D6 to were added to the aqueous extraction solvent as internal standards. All samples were then vortexed for 2 min and centrifuged at 14,000 rpm for 10 min at 4°C. Eight hundred microlitres of supernatants was collected and evaporated to dryness using a vacuum concentrator to dry. Residues were resuspended in 200 µL of acetonitrile: H₂O (1:1, v/v) and centrifuged at 14,000 g for 10 min. The supernatants were then collected and analyzed with a high-performance liquid chromatography-tandem mass spectrometry system (Agilent 1290 Infinity LC, QTRAP 5500, AB SCIES). Wheat leaves incubated with the pEDV6 empty vector were used as control groups. There were five biological replicates for each treatment and time point.

After 2 days of infiltration, five alate adult S. miscanthi were transferred to infiltrated leaves within transparent plastic clip cages (2.7cm×2.7cm×2.7cm) to produce a population of age-synchronized nymphs. On the following day, all adults and other nymphs were removed, in order to avoid overcrowding and effects on plant defense responses, leaving only eight new-born nymphs on each treated leaf. The number of newly produced nymphs was recorded daily for a period of 14 days, and the nymphs were removed after each count to avoid overcrowding again. The number of surviving aphids was also recorded on the end day. After each 4-day period, the wheat plants were replaced by freshly infiltrated plants to guarantee continuous expression of effectors in the leaves. The pEDV6 empty vector was used as a control. A total of fifteen independent biological replicates were performed for each treatment.

All the data were analyzed using SPSS Statistics 20.0 software (RRID : SCR_002865) (http://www-01.ibm.com/software/uk/analytics/spss/), and the differences between or among groups were examined through independent sample t-test or one-way analysis of variance (ANOVA) LSD test. P values less than 0.05 were considered statistically significant.

The two putative effector orthologs obtained from the salivary gland transcriptome of S. miscanthi (Zhang et al., 2017b) were designated Sm10 and SmC002. The open reading frames of these two candidate effectors were cloned and sequenced. We identified the homologous proteins of S. miscanthi using BLAST searches against the GenBank nonredundant (nr) protein database and performed phylogenetic analyses ( Figure 1 ). Sm10 was orthologous to Me10 (M. euphorbiae), NP_001155747.1 (A. pisum), and Mp58 (M. persicae), sharing 76.3%–92.4% sequence identity. The similarity of SmC002 with ApC002 was 83.13% (which was the highest), followed by that with MpC002 (72.32%).

All candidate effectors and homologous proteins had putative signal peptides at the N-terminal region (Me10 was a partial sequence), which is an important feature of the secretion of effectors (Kale, 2012). In the homologous proteins, the pfam domain, transmembrane helices, and nuclear localization signal were not found in the amino acid sequence, as assessed using bioinformatics analysis ( Table 1 ). Amino acid sequence alignments revealed varying degrees of sequence divergence across the selected aphid effectors and proteins ( Figure 2 ). Sm10 and its orthologs had a highly conserved domain within the 57–87 amino acid sequence region ( Figure 2A ). Compared with the amino acid sequence of MpC002, those of SmC002 and ApC002 lacked the NDNQGEE repeat in the N-terminal region ( Figure 2B ). Motifs with a high degree of conservation were detected and their positions in sequences were identified; these sites had a considerable impact on protein function and structure ( Supplementary Figure 1 ).

To determine the subcellular localization of candidate effectors in the cellular compartment, effector-GFP fusion proteins (C-terminal GFP tag) were transiently expressed in the leaves of N. benthamiana. As shown in Figure 3 , the control expressing GFP exhibited fluorescence throughout the whole cell (cytoplasm and nucleus), as revealed via confocal microscopy. Compared with the control, the fluorescence of SmC002-GFP fusion proteins was only observed in the plasma membranes of the epidermal cells of N. benthamiana. In contrast, the fluorescence of Sm10-GFP fusion proteins was detected in both the cytoplasm and nucleus in the leaves of N. benthamiana.

The transcript profiles of Sm10 and SmC002 in various feeding stages of S. miscanthi were characterized using qRT-PCR. The relative expression of Sm10 and SmC002 was differentially induced during the feeding of aphids on wheat and was found to peak at infection stages. The relative expression of Sm10 peaked when aphids fed for 6 hours post-infection (hpi; 4.20-fold, P < 0.001), whereas that of SmC002 peaked when aphids fed for 12 hpi (3.42-fold, P < 0.001). However, Sm10 and SmC002 exhibited no significant difference at 48 and 24 hpi, respectively ( Figure 4 ). These results indicate that Sm10 and SmC002 are induced at the early feeding stage of aphids on wheat and may contribute to the interaction between aphids and plants.

The predicted signal peptide sequences of the Sm10 and SmC002 candidate effectors were cloned and fused into the vector pSUC2. As shown in Supplementary Figure 2 , the negative control (Mg87) could not grow on YPRAA medium. In contrast, both Avr1b (a signal peptide of the Avr1b effector from Phytophthora sojae, which has been reported as a secretory leader and was used as the positive control) and two candidate effector-fused constructs enabled the invertase-mutated yeast strain to grow on CMD-W (YTK12 had the pSCU2 vector and showed growth without invertase secretion) and YPRAA (which allows growth only when invertase is secreted) media ( Supplementary Figure 2A ). Invertase secretion was further confirmed using an enzymatic activity test based on the reduction of TTC to the insoluble red-colored triphenylformazan. Both the candidate effector and Avr1b constructs enabled the conversion of TTC to the insoluble dark-red-colored triphenylformazan. However, the TTC-treated cultures remained almost colorless in the negative controls (the YTK12 and YTK12 yeast strains carrying Mg87; Supplementary Figure 2B ).

These results confirmed the secretory functions of the putative N-terminal signal peptides of the candidate effectors Sm10 and SmC002, which were secreted from S. miscanthi. Moreover, Sm10 (LOC100167427) and SmC002 (LOC100167863) were previously detected in the saliva proteomes of S. miscanthi (Zhang et al., 2021), which further suggests that Sm10 and SmC002 can be secreted into plant cells by aphids.

A bacterial type III secretion system (T3SS) including the expression vector pEDV6 and the P. fluorescens EtHAn strain was used to deliver the effectors of S. miscanthi into wheat and investigate their functions. The AvrRpt2-inoculated wheat plants (positive control) displayed obvious chlorotic reaction phenotypes at 4 days, and DAB staining results revealed high levels of H₂O₂ accumulation in the infiltration areas, indicating that the T3SS expression/delivery system was effective on wheat. In contrast, inoculation of the EtHAn strain carrying the empty vector pEDV6 into wheat did not result in any phenotypic changes or H₂O₂ accumulation. Moreover, infiltration of the EtHAn strain with the pEDV6:Sm10 and pEDV6:SmC002 constructs also did not induce any obvious chlorosis symptoms or H₂O₂ accumulation in the infiltrated regions ( Figure 5 ).

Callose deposition is regarded as a response associated with the later stages of PTI. Aniline blue staining showed that callose deposition was present on wheat plants inoculated with EtHAn and carrying the empty pEDV6 vector (control group), indicating that infection with nonpathogenic EtHAn triggered PTI in wheat ( Figure 6A ). Compared with the control group, a lower amount of callose deposition was observed in wheat leaves inoculated with Sm10 and SmC002 ( Figures 6A, B ), and the number of callose deposits in wheat leaves significantly reduced after the delivery of Sm10 and SmC002 (P < 0.001) compared with AvrRpt2 (positive control) and pEDV6. These results indicated that Sm10 and SmC002 inhibited callose deposition and PTI.

Glucan synthases are key enzymes involved in the biosynthesis of callose. To verify the effects of the effectors on wheat callose deposition, the expression levels of five glucan synthase genes were examined using RT-qPCR (Fu et al., 2014). Our results showed that the expression levels of GSL2 (5.13- to 5.61-fold, P = 0.001) and GSL12 (1.72-fold; P < 0.05) in Sm10-delivered wheat leaves and of GSL23 (1.56-fold; P < 0.05) in SmC002-delivered wheat leaves were significantly upregulated compared with those detected in the controls. However, the transcript levels of GSL8 (0.41- and 0.47-fold, P < 0.001), GSL19 (0.38- and 0.47-fold, P < 0.001), and GSL23 (0.50- and 0.54-fold, P < 0.05) in wheat leaves were significantly decreased at 4 days post-infiltration of EtAnH carrying Sm10 and SmC002 ( Figures 7A–E ). These results further suggested that Sm10 and SmC002 inhibited leaf callose accumulation by affecting the expression of glucan synthase genes.

To further investigate the roles of Sm10 and SmC002 in the regulation of plant defenses, the relative expression of key genes associated with SA and JA and the levels of plant defense hormones in treated wheat leaves were examined. The results showed that the transcript levels of the SA biosynthesis enzyme PAL (2.5- to 7.5-fold) were significantly upregulated after treatment with Sm10 and SmC002 at 2 days (P < 0.001) and 4 days (P = 0.003) and that the expression levels of PR1 (4.6- to 5.2-fold) were significantly increased at 2 days compared with the control (P < 0.001; Figures 8A, B ). Furthermore, delivery of Sm10 and SmC002 into wheat significantly increased the levels of SA at 4 days compared with the control ( Figures 9A, B ), indicating that these two effectors can induce the SA signaling pathway in wheat.

The expression levels of LOX (2.02- to 3.30-fold, P = 0.002) were significantly upregulated at 2 days after SmC002 or Sm10 infiltration. However, the expression levels of FAD (0.21- to 0.32-fold, P < 0.001) and LOX (0.62- to 0.78-fold, P = 0.004), which are involved in the JA signaling pathway, were significantly reduced at 4 days post-infiltration of the two candidate effectors ( Figures 8C, D ). JA and JA-isoleucine were not detected in our study.

To investigate the roles of these two putative effectors in the modulation of host susceptibility, the performance of aphids was determined when the aphids were allowed to feed on wheat leaves incubated with EtHAn carrying the effector. As shown in Figure 10A , the survival of S. miscanthi feeding on SmC002-expressing leaf areas (95.83% ± 4.45%) was significantly higher than that of S. miscanthi feeding on empty vector (control)-infiltrated areas (89.17% ± 4.62%). In addition, the number of nymphs produced by adults and the average weight of adults were both significantly higher in aphids feeding on Sm10- and SmC002-treated leaves (P = 0.001; P = 0.006; Figures 10B, C ). These results suggested that Sm10 and SmC002 enhanced aphid performance and host susceptibility.