A novel salivary effector, BtE3, is essential for whitefly performance on host plants

Abstract

The whitefly Bemisia tabaci is a piercing-sucking herbivore that reduces the yields of crops both by feeding on plants and transmitting plant viruses. Like most plant feeders, B. tabaci has evolved ways to avoid plant defence responses. For example, B. tabaci is known to secrete salivary effectors to suppress host defences. However, the nature of B. tabaci effectors is not completely understood. In this study, we used B. tabaci genomic and salivary gland transcriptomic data and an overexpression system to identify a previously unknown B. tabaci salivary effector, BtE3. BtE3 is specifically expressed in the head (containing primary salivary glands) and is secreted into hosts during B. tabaci feeding. In planta overexpression of BtE3 blocked Burkholderia glumae-induced hypersensitive response (HR) in both Nicotiana benthamiana and Solanum lycopersicum. Silencing of BtE3 by plant-mediated RNAi prevented B. tabaci from continuously ingesting phloem sap, and reduced B. tabaci survival and fecundity. Moreover, overexpression of BtE3 in planta up-regulated the salicylic acid- (SA-) signalling pathway, but suppressed the downstream jasmonic acid- (JA-) mediated defences. Taken together, these results indicate that BtE3 is a B. tabaci-specific novel effector involved in B. tabaci-plant interactions. These findings increase our understanding of B. tabaci effectors and suggest novel strategies for B. tabaci pest management.

Introduction

Interactions between plants and insect herbivores are known to be mediated by diverse molecular processes (Howe and Jander, 2008). Once damage-associated and/or herbivore-associated molecular patterns (DAMPs and/or HAMPs) are recognized by plants, the affected plants activate early signalling components such as Ca²⁺, reactive oxygen species, and mitogen-activated protein kinases (MAPKs). Phytohormones, plant secondary metabolites, and plant transcription factors form signalling networks that trigger specific plant defences (Erb and Reymond, 2019). Other deterrents include toxins, defence proteins, physical barriers, and tolerance traits (Erb and Reymond, 2019). At the same time, herbivores use feeding strategies and other weapons like salivary effectors to circumvent plant defences (Walling, 2008).

Salivary effectors secreted from insect herbivores are indispensable for the establishment of compatible insect-plant interactions. Broadly defined, effectors include all pathogen/herbivore proteins and small molecules that alter the structure or function of host cells (Hogenhout et al., 2009). These alterations may trigger defence responses, or may promote infection by pathogens or infestation by herbivores (Hogenhout and Bos, 2011). Effectors that nematodes and other non-arthropod herbivores secrete in order to alter plant tissues, or otherwise interfere with plant defences, have been well documented (Haegeman et al., 2013). As more and more effector proteins secreted by herbivorous arthropods have been identified, it has been increasingly suspected that many herbivorous insects rely on the secretion of effectors to overcome the defences of host plants (Villarroel et al., 2016; Huang et al., 2021; Naalden et al., 2021).

A number of salivary effectors from herbivores have been identified. The first effector identified in chewing insects was glucose oxidase (GOX), which Helicoverpa zea secretes to suppress the accumulation of nicotine in tobacco (Musser et al., 2002). The effector HARP1 from Helicoverpa armigera was recently reported to directly interact with host JASMONATE-ZIM-domain (JAZ) repressors and to thereby block jasmonic acid (JA) signal transduction in the host plant (Chen et al., 2019). Among piercing-sucking herbivores such as aphids, planthoppers, and spider mites, several effectors that manipulate host defences have been identified (Mutti et al., 2008; Bos et al., 2010; Elzinga et al., 2014; Naessens et al., 2015; Jonckheere et al., 2016; Kaloshian and Walling, 2016; Villarroel et al., 2016; Ji et al., 2017; Cui et al., 2019; Su et al., 2019; Xu et al., 2019). The first well studied effector from an aphid was C002, which helps Acyrthosiphon pisum contact host sieve elements and maintain ingestion of sap (Mutti et al., 2008). A functional genomics pipeline was also applied to screen for effectors released by the aphid Myzus persicae, and three effectors (Mp10, Mp42, and MpC002) were identified and found to be involved in the manipulation of host cell processes to the benefit of aphid performance (Bos et al., 2010). Similarly, the effector proteins Tu28, Tu84, and Tu84 promote performance of the mite Tetranychus urticae by suppressing host salicylic acid (SA) defences (Villarroel et al., 2016).

As a piercing-sucking herbivore, the whitefly Bemisia tabaci (Gennadius) (Hemiptera: Aleyrodidae) is adept at overcoming host defences (Walling, 2008; Naalden et al., 2021). Its specialized mouthparts (stylets) enable B. tabaci to penetrate into the phloem without causing substantial mechanical damage, so as to reduce detection and defence induction by the plant host (Walling, 2008). In addition, B. tabaci have been reported to secrete effectors that manipulate plant defences by altering the interplay between JA and SA signalling pathways; such pathways are significant components of plant defence against aphids (Su et al., 2019; Wang et al., 2019; Xu et al., 2019; Du et al., 2022). Although effectors are known to help B. tabaci counter plant defences, effector biology remains incompletely understood.

In the current study, we conducted a large-scale search for possible B. tabaci effectors by using the previously reported B. tabaci salivary gland transcriptome (Su et al., 2012), B. tabaci genome data (Xie et al., 2017), and a modified effector-screening pipeline that was previously used to identify multiple herbivore effectors (Bos et al., 2010; Villarroel et al., 2016). To identify effectors that reduce plant defences, we used the effector detector vector (pEDV) system, which enables expression and delivery of candidate effectors encoded by the introduced exogenous genes from B. tabaci. In addition, the type III secretion system (T3SS) of the rice pathogen Burkholderia glumae was used as an efficient tool for translocating B. tabaci candidate effectors into plant cells (Sharma et al., 2013; Shangguan et al., 2018). B. glumae causes rot and panicle blight in rice and bacterial wilt in many field crops, as well as a strong hypersensitive response (HR) in Nicotiana benthamiana (Ham et al., 2011; Sharma et al., 2013). The B. glumae and pEDV combined (BG-pEDV) system provides a practical tool to identify herbivore virulence effectors. To verify that candidate effectors inhibited plant defences, we also used the commonly employed AG-pGR107 transient expression system (a combination of Agrobacterium tumefaciens and the overexpression vector pGR107), which was previously used to screen for effectors in aphids and mites (Bos et al., 2010; Elzinga et al., 2014; Thorpe et al., 2016).

Using bioinformatic prediction and transient expression, we identified a novel B. tabaci effector, BtE3, that inhibits the HR caused by B. glumae in N. benthamiana. Bioassays with RNAi transgenic tobacco lines targeting BtE3, and BtE3-overexpression lines indicated that BtE3 plays an important role in B. tabaci feeding and oviposition on host plants. We also found that BtE3 acts as an effector that up-regulates the SA signalling pathway and suppresses downstream JA-regulated defences of host plants, thereby facilitating herbivory of host plants by B. tabaci.

Materials and methods

Insects and plants

The colony of B. tabaci Mediterranean (MED; previously known as the Q biotype) used in this study was maintained on cotton plants in climate chambers at 26 ± 2 °C with a photoperiod of 16 h:8 h (light: dark) and an artificial light intensity of ~80 µmol m^–2 s^–1, and 70 ± 10% relative humidity. All insect-host plant experiments were conducted at this temperature and with this light regime. The genetic purity of the B. tabaci MED cultures was monitored every 3–5 generations based on amplified fragment-length polymorphism (AFLP) analysis and on the sequence of the mitochondrial cytochrome oxidase I (mtCO1 gene), which have been widely used to differentiate the genetic groups of B. tabaci (Zhang et al., 2005).

Seeds of cotton (Gossypium hirsutum ‘Zhongmian 49’), tobacco (Nicotiana tabacum ‘K326’ and N. benthamiana), and tomato (Solanum lycopersicum ‘Zhongza 9’) were from our laboratory. Seeds of NahG tobacco (N. tabacum var. Samsun NN, a transgenic plant that does not accumulate SA) were kindly provided by Prof. Feng Liu (Shandong Agricultural University, China). All plants were cultivated singly in pots in climate chambers at 25 ± 2 °C, with a photoperiod of 16 h light: 8 h darkness with a light intensity of ~80 µmol m^–2 s^–1 and 70 ± 10% relative humidity. Cotton plants were used to maintain B. tabaci strains; N. benthamiana and S. lycopersicum were used for plant HR assays; and N. tabacum was used for construction of transgenic plants and subsequent bioassays.

In silico prediction of B. tabaci candidate effectors

To select candidate B. tabaci effectors, we used a B. tabaci salivary gland transcriptome dataset from B. tabaci MED raised on cotton (Su et al., 2012), and a B. tabaci genome dataset (Xie et al., 2017). We performed BLASTn searches to obtain the coding sequences of those transcriptome unigenes, and then translated all of the coding sequences into amino acid sequences. Deduced amino acid sequences were submitted to the SignalP 5.0 server (https://services.healthtech.dtu.dk/service.php?SignalP-5.0) and the TMHMM 2.0 server (https://services.healthtech.dtu.dk/service.php?TMHMM-2.0) for prediction of secretory and transmembrane proteins. The non-redundant (Nr) dataset of NCBI was used to run BLASTp for annotation of those predicted secretory proteins. Transcriptomes of different developmental stages of B. tabaci (Tian et al., 2017) were used to screen those proteins exhibiting a relatively high expression level in the adult. For genes that were barely expressed in the egg, which could not be used to calculate the increased expression in the adult, we directly compared their FPKM values in the adult.

Transient expression analysis

The suppression of the HR elicited by B. glumae (a gift of Prof. Bingyan Xie, Chinese Academy of Agricultural Sciences) on N. benthamiana and S. lycopersicum leaves was performed as described previously (Sharma et al., 2013). The coding sequences of candidate effectors lacking the signal peptide and GFP were amplified by PCR using the primers listed in Supplementary Table S1, and the sequences were then cloned into the pEDV vector (kindly provided by Prof. Bingyan Xie, Chinese Academy of Agricultural Sciences) to generate pEDV::effector and pEDV::GFP constructs, respectively. All of the constructs were transformed into B. glumae by electroporation, and the transformants were then cultured in LB medium with 25 μg ml^–1 gentamycin. Positive clones were verified by PCR using specific primers and were cultured in liquid LB medium with gentamycin. Strains carrying pEDV::effector and pEDV::GFP were suspended in 0.9% NaCl at 600 nm (OD₆₀₀=0.4). B. glumae strains carrying pEDV::GFP or pEDV::effector were infiltrated separately into leaves of N. benthamiana and S. lycopersicum using a 1 ml needleless syringe. Plants were maintained at 25 °C, and leaf symptoms were observed at 4 d post-inoculation. Three leaves (the third to fifth leaf counted from bottom to top) of an individual plant were used for each biological replicate, and at least three independent replicates were performed.

For Agrobacterium-mediated transient expression analysis, the coding sequences corresponding to the mature protein of candidate effectors were amplified and ligated into vector pGR107 (a potato virus X vector) with a cauliflower mosaic virus 35S promoter (kindly provided by Prof. Yuanchao Wang, Nanjing Agriculture University, China) to generate recombinant vectors. The recombinant vectors were validated by sequencing and were transformed into A. tumefaciens strain GV3101 by electroporation. The cultivated cells of recombinant A. tumefaciens strains were collected by centrifugation and resuspended in infiltration buffer (10 mM MES, 10 mM MgCl₂, 150 μM acetosyringone, pH 5.6) to a final OD₆₀₀ of 0.6, as described previously (Bos et al., 2010). A. tumefaciens strains GV3101 carrying pGR107::GFP and pGR107::Bax (Bax is a mouse pro-apoptotic protein which elicits plant HR)were used as negative and positive controls, respectively. Three hours after resuspension, leaves of tobacco at the four-to-five true-leaf stage were infiltrated with A. tumefaciens, and alterations of systemic PVX symptoms were recorded 2 d after infiltration.

Generation of transgenic tobacco plants and RNAi experiments

The hairpin RNA expression vector (pCAMBIA2300-silencer::BtE3, containing a cauliflower mosaic virus 35S promoter) was introduced into tobacco (N. tabacum ‘K326’) for generation of RNAi transgenic lines. For construction of pCAMBIA2300-silencer::BtE3, a 314 bp target fragment amplified from B. tabaci with sense-BtE3 primers (Supplementary Table S1) was inserted into BamHⅠ-EcoRⅠ-cut pCAMBIA2300 (kindly provided by Prof. Xia Cui, Chinese Academy of Agricultural Sciences, China), generating pRNAi-sense-BtE3. Antisense-BtE3 primers (Supplementary Table S1) were then used to clone the antisense fragment of BtE3 from B. tabaci, after which the purified fragment was ligated to SacⅠ-KpnⅠ-cut pRNAi-sense-BtE3, generating pCAMBIA2300-silencer::BtE3 (Supplementary Fig. S1). A. tumefaciens LBA4404-based transformation was used to transfer the recombinant pCAMBIA2300-silencer::BtE3 plasmid into tobacco, as described previously (Ren et al., 2015).

For BtE3-overexpression transgenic tobacco lines, the BtE3 fragment without the signal peptide sequence was cloned from B. tabaci using OE-BtE3 primers (Supplementary Table S1), and the purified fragment was then introduced into BamHⅠ-EcoRⅠ-cut pCAMBIA2300, generating pCAMBIA2300-35S::BtE3 (Supplementary Fig. S1). The recombinant pCAMBIA2300-35S::BtE3 plasmid was also transferred into A. tumefaciens LBA4404 in the same manner as the hairpin RNA expression vector.

Primary tobacco transformants (T₀) were selected with kanamycin (100 mg l^–1). The regenerated transformants were transplanted in a greenhouse with 80% humidity, 280 μmol m^–2 s^–1 light intensity, and a 16 h/8 h light/dark cycle (Supplementary Fig. S2). RNAi transgenic lines were confirmed by PCR using specific primers and the gDNA as template. Overexpression transgenic lines were confirmed by PCR using cDNA from leaves as templates and primers specific to BtE3 gene sequence.

To silence BtE3, groups of 40 newly emerged B. tabaci were allowed to feed on individual RNAi transgenic and control tobacco plants (producing dsBtE3 or dsGFP, as described above). After 2 d of feeding, B. tabaci were collected, and the level of BtE3 expression in these treated B. tabaci was determined by RT–qPCR to assess the efficiency of RNAi.

Bemisia tabaci bioassays

To measure the effect of BtE3 silencing on B. tabaci survival and fecundity, newly emerged adults were allowed to feed on RNAi transgenic and control tobacco plants that produced dsBtE3 or dsGFP, as described earlier, for 2 d; this generated BtE3-silenced individuals or control individuals, respectively. Adult B. tabaci (BtE3-silenced or control individuals) were placed in clip cages (five female and five male adults per cage) attached to the third true leaf (counted from the bottom to the top) of individual BtE3- and GFP-overexpressed tobacco plants; clip cages were 3 cm in diameter with two cages per plant. In the artificial diet experiment, groups of 10 adult B. tabaci (BtE3-silenced or control individuals) were introduced into individual feeding chambers (20 mm diameter and 50 mm long) containing 5% yeast extract and 30% sucrose, as described previously (Su et al., 2018). The number of surviving adults in each clip cage or feeding chamber was recorded daily for 7 d. Each clip cage or feeding chamber was treated as a biological replicate, and 12 replicates were used for each treatment. For fecundity, 7 d after the release of adult B. tabaci, the eggs laid on the leaf within clip cages were counted with the aid of a stereomicroscope. In total, 16–18 replicate clip cages were assessed for each treatment before average fecundity per female was calculated.

Analysis of B. tabaci feeding behaviour

The feeding behaviour of individual B. tabaci on cotton was recorded using a Giga-8 direct-current electrical penetration graphing (DC-EPG) system (Wageningen University, The Netherlands) as described previously (Liu et al., 2013). Newly emerged females were allowed to feed on RNAi transgenic and control tobacco plants (as described earlier) for 2 d, and the treated females were then transferred to cotton plants on which their feeding behaviour was continuously monitored for 8 h. The recording EPG waveforms were then analysed with PROBE V. 3.4 software (Wageningen University). Relationships between EPG waveforms and feeding behaviour were described by Liu et al. (2013). Fresh plants and insects were used for each replicate, with 17 replicate females per treatment.

Protein expression, purification, and antibody preparation

The nucleotide sequence encoding BtE3 lacking the N-terminal secretion signal was amplified by PCR using the primers listed in Supplementary Table S1, and then inserted into the expression vector pATX-SUMO (Supplementary Fig. S3). The recombinant plasmid was transformed into the Escherichia coli BL21 (DE3) strain for expression after sequence verification. Induced expression was conducted after adding 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) at 37 °C for 4 h. The products from the recombinant were purified by using Ni-NTA columns (Qiagen, Valencia, CA, USA) and concentrated with a YM-10 Microcon centrifugal filter device (Millipore, Billerica, MA, USA) to remove imidazole. Purified recombinant protein of BtE3 was used as the antigen to produce the rabbit polyclonal antibodies, and the polyclonal antibodies were made and purified by Atagenix™ (Atagenix, Nanjing, China).

Western blot analysis

To determine whether BtE3 is a secretory protein that can be delivered into host plants via salivation, proteins extracted from B. tabaci whole bodies, from B. tabaci honeydew, and from B. tabaci-infested and uninfested cotton leaves were subjected to western blot analysis, as described by Su et al. (2019). About 200 adult B. tabaci were collected and homogenized in 1 ml of phosphate-buffered saline (PBS). The extract was centrifuged at 12 000×g for 10 min at 4 °C, and the supernatant was collected. Cotton plants were individually confined in a ventilated cage in which 5000 B. tabaci adults were released, and Petri dishes were placed under cotton leaves for honeydew collection (Supplementary Fig. S4). After 48 h, all of the released B. tabaci and their eggs laid on leaves were removed carefully. Plants without B. tabaci were used as controls. The honeydew deposited in Petri dishes was subsequently collected with a pipette by adding 1 ml of PBS; the mixture was homogenized and then sterilized by passage through 0.2 μm filters (Millipore, Billerica, MA, USA). The top three leaves from three cotton plants (~1 g per plant) were pooled and homogenized in 4 ml of PBS. The extract was centrifuged at 12 000×g for 10 min at 4 °C, and the supernatants were collected as samples. After protein extraction, proteins were quantified using the Bradford method with bovine serum albumin (BSA) as a standard, and then flash frozen and kept in aliquots at –80°C until further use (Bradford, 1976). After SDS loading buffer was added, the samples were subjected to SDS–PAGE on a 12% gradient gel, and proteins were transferred onto a polyvinylidene fluoride (PVDF) membrane. The blot was incubated with the purified rabbit polyclonal anti-BtE3 antibody (1:500 dilution), visualized with a goat anti-rabbit IgG conjugated horseradish peroxidase antibody (CWBIO, Beijing, China) at a 1:5000 dilution, and detected with the ECL Plus Detection System (Bio-Rad, Hercules, CA, USA).

The B. tabaci CYP6CM1 antiserum and protein blot methods were previously described by Yang et al. (2020). This antiserum was used to assess the possible contamination of cotton leaf proteins with residual B. tabaci material.

Reverse transcription real-time quantitative PCR (RT–qPCR)

Total RNA from B. tabaci tissues (head, thorax, and abdomen) and at developmental stages (eggs, first- and second-instar nymphs, third-instar nymphs, fourth-instar nymphs, and newly emerged adults) was collected and extracted using Trizol reagent (Invitrogen, Carlsbad, CA, USA). Total RNA from tobacco leaves was extracted using a RNeasy plus Mini kit (Qiagen, Hilden, Germany). The RNA quantities were determined using a Nanodrop 2000 (Thermo Scientific, Wilmington, DE, USA), and purity was determined on 1% agarose gels. First-strand cDNA synthesis and RT–qPCR were performed using the SYBR Green PCR Master Mix (Tiangen, Beijing, China) on the ABI QuantStudio3 real-time PCR system, as described previously (Su et al., 2016, 2018). Primers used for all of the tested genes are indicated in Supplementary Table S1. Expression levels of target genes were normalized with internal reference genes encoding RPL-29 (ribosomal protein L29) and EF-1α from B. tabaci, and RPL25 and UBC2 (ubiquitin-conjugating enzyme 2) from N. tabacum. Each gene was analysed in triplicate for each of four biologically independent treatments.

Jasmonic acid, JA-isoleucine, and salicylic acid analysis

Leaf samples were harvested and ground in liquid nitrogen, and the phytohormones were extracted with ethyl acetate spiked with labelled internal standards containing D4-SA, D6-JA, and D6-JA-Ile (CDN Isotopes, Pointe-Claire, Canada), and analysed with a triple–quadrupole LC-MS/MS system (Thermo Scientific, Waltham, MA, USA) following the method as described in Su et al. (2019). Each treatment was replicated three times.

Statistical analyses

Before analysis, data were checked for normality and homogeneity of variance using the Kolmogorov-Smirnov test and Levene’s test, respectively. Data were subjected to either Student’s t-test or one-way analysis of variance (ANOVA) followed by a comparison of the means according to a Tukey’s honestly significant difference (HSD) test at P<0.05. All statistical analyses were performed with GraphPad Prism 8.0 (GraphPad Software Inc., San Diego, California, USA).

Results

In silico screening for candidate effectors in B. tabaci

A broadly used strategy (Bos et al., 2010) to search for effectors that may potentially suppress plant defences was applied to generate a list of candidate effectors in B. tabaci (Fig. 1). We used the previously reported B. tabaci salivary gland transcriptome (Su et al., 2012) combined with our B. tabaci genome data (Xie et al., 2017) to obtain potential full-length salivary proteins. There were 4048 coding sequences from our genome data that matched the 13 615 unigenes of the salivary gland transcriptome. These coding sequences (CDS) were then translated into amino acid sequences for prediction of signal peptide using SignalP v4.0 (Petersen et al., 2011). Finally, 194 sequences predicted to contain a signal peptide, but no transmembrane domain, were retained as candidate effectors. Subsequently, two additional steps were added to the filtering pipeline for the specific screening of candidates. Previous studies indicated that the majority of pathogen effectors are highly species-specific (Göhre and Robatzek, 2008; Thomma et al., 2011), and based on a B. tabaci salivary proteome Huang et al. (2021), showed that 20% of the B. tabaci salivary proteins were B. tabaci-specific. Therefore, to identify novel B. tabaci effectors, 71 proteins with unknown functions and no homologs in other insect species were prioritized for this study. Because effectors should be especially important during B. tabaci feeding, we assumed that candidates should display a higher expression level in B. tabaci at the feeding stages (nymph or adult) than at the non-feeding stage (embryo). Our previous B. tabaci transcriptome data at different developmental stages (Tian et al., 2017) were utilized for screening, and we arbitrarily selected 10 candidates (a workable number) whose expression was ≥20-fold greater during the peak of adult feeding than in the egg. In addition, four genes barely expressed in the egg, but with high expression levels in the adult, were added to the candidate pool, bringing the total to 14 candidates; these were BtE1 – BtE14 (Fig. 1; Supplementary Fig. S5).

Transient expression of the candidate effector BtE3 suppresses the hypersensitive response in N. benthamiana

To determine whether the 14 candidates function in plant immune suppression, we used the rice pathogen, B. glumae, to deliver B. tabaci candidate effectors into N. benthamiana plants via a pEDV-based type-Ⅲ system (T3SS). The bacteria alone can induce an immune HR response (rapid and localized cell death) in the non-host N. benthamiana, but transient expression of candidate effectors may suppress the immune response. Three days after infiltration, the HR symptoms were recorded. Our results showed that, when compared with the green fluorescent protein (GFP) negative control, two of the 14 candidates (BtE3 and BtE8) notably suppressed the B. glumae-induced HR in N. benthamiana (Fig. 2). To confirm that the two candidates function in immune suppression, we conducted the same transient expression experiments with the two candidates (BtE3 and BtE8) in tomato plants, and observed similar results (Fig. 2).

We also performed Agrobacterium-mediated transient expression of these two candidates in N. benthamiana. Interestingly, though both candidates were unable to induce local cell death in the infiltration areas, only BtE3 could moderately suppress the HR induced by the immune-elicitor Bax (Fig. 3). We therefore focused on the function of BtE3 in B. tabaci -plant interactions.

Characterization of the candidate effector BtE3

The transcript of BtE3 includes 1109 bp, and sequence analysis indicated that BtE3 encodes a 268 amino acid protein that has an extracellular signal peptide of 22 amino acids at its N-terminus and that lacks a transmembrane domain, suggesting that BtE3 might be a secretory protein. The molecular mass of the predicted mature protein is 28.3 kDa, with a pI of 9.25. BtE3 also has a high (5%) cysteine content, which may promote the formation of disulphide bonds and thereby stabilize the protein’s higher order structure (Fig. 4A). Spatial and temporal expression analysis showed that BtE3 is highly expressed in the head (which contains the primary salivary glands) and has a significant increase in expression in the adult, i.e. an active feeding stage (Fig. 4B, C).

BtE3 is secreted into the host plant during B. tabaci feeding

To determine whether B. tabaci secretes BtE3 during feeding, we extracted proteins from cotton leaves that had been kept B. tabaci-free or B. tabaci-infested for 48 h. We also extracted proteins from B. tabaci honeydew to exclude the possibility that BtE3 is present in the B. tabaci excrement. Western blot was performed using polyclonal anti-BtE3 rabbit antibodies. A band of ~30 kDa, corresponding to the molecular weight of BtE3, was detected in the protein extracts from B. tabaci heads (lane 1) and was also detected in B. tabaci-infested leaves (lane 4; Fig. 4D). In contrast, a 30 kDa band corresponding to BtE3 was not detected in B. tabaci honeydew or in leaves without B. tabaci infestation (lanes 2 and 3). To assure that the BtE3 detected in leaf protein extracts was reflective of BtE3 secretion into leaf tissues, and not due to a contamination of B. tabaci tissue, we used an antiserum to the B. tabaci CYP6CM1 to detect residual whitefly proteins (Supplementary Fig. S6). These results demonstrated that B. tabaci was able to transfer BtE3 into the host plant via salivation rather than via excretion.

Silencing of BtE3 impairs the ability of B. tabaci to exploit host plants

To investigate the function of BtE3 during B. tabaci feeding, we generated 35S:BtE3-RNAi plants that systematically synthesized double-stranded RNA targeting BtE3 transcripts. In B. tabaci adult females, transcript levels of BtE3 were significantly lower after feeding on 35S:BtE3-RNAi, than after feeding on control plants (Fig. 4E). Adult females of B. tabaci were also allowed to feed on 35S:BtE3-RNAi (which generated BtE3-silenced females) or control plants (which generated control females) for 2 d, after which the feeding behaviour of individual females on healthy cotton plants was monitored with electrical penetration graph (EPG) equipment. Four main phases were analysed: the non-probing phase, pathway phase, salivation phase, and phloem ingestion phase. The duration of the non-probing phase did not significantly differ (P=0.408) between BtE3-silenced and control females (Fig. 5; Supplementary Table S2). However, the duration of the pathway phase was longer, and the phloem phase was shorter for BtE3-silenced females than for control females (Fig. 5; Supplementary Table S2). Interestingly, BtE3-silenced individuals reached the phloem phase less frequently than control females (Supplementary Table S2). We also measured the survival rate and fecundity of BtE3-silenced B. tabaci that fed on BtE3-overexpression (35S:BtE3) and GFP-overexpression (35S:GFP) transgenic tobacco plants. The survival (P<0.0001) and fecundity (P=0.0061) were lower for BtE3-silenced individuals than for control individuals on 35S:GFP plants (Fig. 6A, D). However, survival (P=0.893) and fecundity (P=0.852) did not significantly differ between BtE3-silenced and control individuals on 35S:BtE3 plants (Fig. 6B, E). Moreover, knock down of BtE3 did not result in an obvious difference in survival rate when B. tabaci fed on artificial diets (Fig. 6C), indicating that the secreted BtE3 allows B. tabaci to continuously ingest phloem sap and increase its oviposition and survival.

BtE3 suppresses induced plant defences

To further investigate the mechanism by which BtE3 enhances B. tabaci performance, we measured the expression levels of SA- and JA-related marker genes in 35S:BtE3 and 35S:GFP transgenic tobacco plants. Compared with the expression levels in 35S:GFP plants, three SA-related marker genes, PAL (phenylalanine ammonia lyase), NPR1(non-expressor of pathogenesis-related genes 1), and PR1a (pathogenesis-related protein 1), were significantly up-regulated (P<0.0001) in 35S:BtE3 tobacco (Supplementary Fig. S7A). In contrast, two JA-related marker genes, PDF1.2 (P<0.0001) and COI1(P<0.0001), were markedly down-regulated in 35S:BtE3 tobacco (Supplementary Fig. S7B). Besides, transcript levels of both SA- and JA-related marker genes were similar in 35S:BtE3 plants when they were infested by BtE3-silenced B. tabaci (B. tabaci previously fed on RNAi plants) and control B. tabaci (Fig. 7A, B). In contrast, the expression of JA-related marker genes was significantly higher (P=0.0184 for LOX; P=0.032 for COI1) in 35S:GFP plants treated with BtE3-silenced B. tabaci than with control B. tabaci, but the expression of SA-related genes was significantly lower (P=0.0136 for PAL; P=0.0404 for NPR1; P=0.0166 for PR1a; Fig. 7C, D). These results suggested that BtE3 could activate the SA signalling pathway and repress the JA signalling pathway.

To determine if B. tabaci utilizes BtE3 to alter the concentrations of SA, JA and JA-Ile in tobacco after B. tabaci infestation, these phytohormones were quantified. SA content of tobacco plants treated with BtE3-silenced B. tabaci was significantly repressed (P=0.0008), while the JA and JA-Ile content was slightly increased, but not significant (Fig. 8). Together, these results indicate that BtE3 from B. tabaci activates the SA signalling pathway, represses downstream JA-mediated defences, and thereby breaks the balance of SA- and JA-signalling crosstalk in host plants.

Suppression of JA defences by BtE3 is dependent on the SA signalling pathway

To determine whether BtE3 activation of the SA signalling pathway promotes B. tabaci performance on host plants, we infested the SA-suppressing NahG transgenic tobacco line with BtE3-silenced B. tabaci and control B. tabaci. Survival and fecundity were significantly lower (P<0.0001) for BtE3-silenced B. tabaci than for control B. tabaci on wild-type tobacco plants (Fig. 9A, C). In contrast, survival and fecundity of these insects was similar, indicating that knock down of BtE3 did not affect the survival or the fecundity of B. tabaci on NahG transgenic tobacco (Fig. 9B, D). In addition, BtE3-silenced B. tabaci produced fewer progeny than control B. tabaci after 7 d of feeding on wild-type tobacco plants, but not after 7 d of feeding on NahG transgenic tobacco (Supplementary Fig. S8). These results demonstrated that induction of the SA signalling pathway by BtE3 helps explain how BtE3 improved the performance of B. tabaci on host plants.

Discussion

When feeding on hosts, B. tabaci inject a variety of molecules that can serve as effectors to manipulate host cell processes (Naalden et al., 2021). In the current study, with the help of a bioinformatic pipeline and transient expression systems, we identified a B. tabaci salivary effector, BtE3. Our results showed that BtE3 could significantly reduce pathogen-triggered plant HR symptoms, and acted as an elicitor of SA which enhanced the performance of B. tabaci on its hosts.

The HR is a frequent consequence of avirulence effectors or of the recognition of conserved molecular patterns by plant immune systems (Pitsili et al., 2020). Researchers often use in planta transient expression of effector candidates from herbivores or pathogens to identify effectors that repress plant defence (Bos et al., 2010; Sharma et al., 2013; Lee et al., 2018). The B. glumae type Ⅲ secretion system (BG-T3SS)-based effector delivery system is useful for the identification and characterization of novel pathogen cytoplasmic effectors (Sharma et al., 2013; Shi et al., 2018). As B. glumae alone induces HR ­symptoms, the system does not require the introduction of other immune elicitors, which simplifies the screening for effectors that suppress HR. By using the BG-T3SS-based system, we successfully identified two candidates, BtE3 and BtE8, that suppressed B. glumae -induced HR in both N. benthamiana and S. lycopersicum (Fig. 2), implying a virulence factor activity of these two candidates. Besides, transient expression of BtE3 and BtE8 in N. benthamiana did not lead to a HR symptom (Fig. 3), suggesting that both candidates function as plant defence suppressors, rather than elicitors.

However, BtE3 has a broader spectrum of defence suppression than BtE8, as BtE3 also suppresses Bax (an immune elicitor)-induced HR (Fig. 3). As HR involves a series of events that include calcium influxes, oxidative bursts originating in different cellular compartments, hormone signalling, mitogen-activated protein kinases, and transcriptional reprogramming (Adachi and Tsuda, 2019), B. glumae-triggered and Bax-induced HR may be associated with different signalling components. This might also help explain why BtE8 suppressed the HR induced by B. glumae, while it failed for Bax. The M. persicae effector Mp10 also specifically interfered with the plant immune pathway that requires the recognition of flg22, but failed to interfere with the plant immune pathway that requires the recognition of chitin (Bos et al., 2010).

Piercing-sucking mouthparts (i.e. stylet bundles) enable B. tabaci to feed on phloem sap without causing massive ­mechanical damage to host tissues; this reduces detection and defence induction by the plant host (Naalden et al., 2021). However, phloem-localized resistance traits can often deter phloem-feeders (Mutti et al., 2008; Jiang et al., 2019; Kloth et al., 2021). Correspondingly, B. tabaci and other piercing and sucking herbivores release a series of effectors that may reduce phloem resistance (Mutti et al., 2008; Ji et al., 2017; Su et al., 2019; Xu et al., 2019). The prolonged pathway phase and shortened sap-feeding stage of BtE3-silenced B. tabaci (Fig. 5; Supplementary Table S2) indicated that they may encounter difficulties in stylet penetration, salivation, and continuous sap ingestion (Tjallingii, 1978; Liu et al., 2013), perhaps because of phloem resistance. Besides, callose synthesis and deposition appears to be a common response to B. tabaci feeding, although the role of callose in resistance to B. tabaci needs to be further demonstrated (Kempema et al., 2007; Su et al., 2019; Walker, 2022). We therefore speculate that BtE3-enhanced B. tabaci feeding might link to suppression of phloem resistance, although the mechanisms remain to be studied.

To better colonize host plants, herbivores use multiple effectors to increase compatibility with their host plants (Mutti et al., 2008; Bos et al., 2010; Rodriguez and Bos, 2013; Chaudhary et al., 2014; Naessens et al., 2015; Wang et al., 2015; Ji et al., 2017; Su et al., 2019; Xu et al., 2019; Guo et al., 2020; Du et al., 2022). The ability of BtE3 to improve the performance of B. tabaci on their hosts (e.g. increased fecundity and adult survival) can be associated with the ability of BtE3 to accumulate SA, activate SA signalling, and decrease expression of JA-responsive genes (Figs 6–8). As sessile organisms, plants are confronted with multiple invaders, so they arm themselves with hormone signalling pathways to rapidly adapt and utilize limited resources for development in a cost-efficient manner (Pieterse et al., 2012). Herbivores have evolved ingenious strategies to rewire the plant’s hormone signalling circuitry to repress or circumvent host defences. For instance, B. tabaci induce SA signalling pathways and constrain the up-regulation of JA signalling pathways for their own benefits in some host plants (Kempema et al., 2007; Zarate et al., 2007; Zhang et al., 2013; Su et al., 2015). Although the mechanisms underlying the hijacking of plant hormone pathways by B. tabaci remain elusive, effectors are reported to engage in the manipulation of JA and SA signalling pathways (Su et al., 2019; Xu et al., 2019; Du et al., 2022), and BtE3 provides compelling evidence of such modification. BtE3-induced accumulation of SA did not significantly influence JA and JA-Ile levels, indicating that it might be associated with phytohormone concentrations. SA can both initiate the expression of JA-responsive genes, and neutralize the effect of JA against herbivores, depending on subtle changes in relative titres of SA (Wei et al., 2014; Liu et al., 2016). Besides, the accumulation of SA induced by BtE3 may account for the attenuated pathogen-triggered HR. Exogenous application of SA on Arabidopsis Col-0 plants blocks HR, and the Arabidopsis mutants deficient in SA accumulation or perception showed enhanced cell death after pathogen infection, which indicates that SA is involved in negative feedback regulation of HR (Devadasand Raina, 2002; Radojičić et al., 2018; Zhou and Zhang, 2020).

Meanwhile, the use of the tobacco SA-deficient mutant NahG confirmed that induction of the SA pathway by BtE3 is sufficient to increase B. tabaci performance (Fig. 9). However, mechanisms on how SA-JA interplay steers the final outcome of plant defences by B. tabaci requires further investigation (Zarate et al., 2007; Pieterse et al., 2012; Xu et al., 2019). It is possible that B. tabaci uses BtE3 as an elicitor of SA, similar to its use of Bt56, to promote its performance on plants (Xu et al., 2019). However, B. tabaci performance is also enhanced by BtFer1, which directly represses JA in the host, and by BtArmet, which inhibits the SA pathway in the host (Su et al., 2019; Du et al., 2022). These B. tabaci effectors achieve the same end, but by different routes, which together contribute to the formidable capacity of B. tabaci to counter plant basal resistance mechanisms. Therefore, B. tabaci probably uses a repertoire of unknown effectors that target different immune pathways, which increases the ability of B. tabaci to use a broad range of hosts. In our study, we identified genes with high expression at the adult stage as candidate effectors (Supplementary Fig. S3). There is also an opportunity to understand the effector profile of developing nymphs, which feed at the same site for over 17 d. They may secrete a similar or distinct set of effectors from feeding adults. With the rising number of effectors discovered, integrative mechanisms for the modulation of phytohormones by B. tabaci can be disclosed. Furthermore, the identification of B. tabaci -specific effectors could provide promising targets for RNAi-based pest control strategies.

Supplementary data

The following supplementary data are available at JXB online.

Fig. S1. Vector construction for tobacco transformation.

Fig. S2. Transgenic tobacco plants cultured in soil.

Fig. S3. Expression and purification of BtE3 and specificity of anti-BtE3 antibodies.

Fig. S4. Diagram of method used to collect honeydew of B. tabaci feeding on cotton.

Fig. S5. Expression profiles of genes without an annotation in the Nr database throughout B. tabaci MED life cycle.

Fig. S6. Bemisia tabaci antiserum against CYP6CM1 was used as a control.

Fig. S7. Expression levels of genes involved in phytohormone pathways in transgenic tobacco plants without B. tabaci infestation.

Fig. S8. Fecundity of BtE3-silenced B. tabaci feeding on NahG mutant tobacco plants was restored.

Table S1. List of primer sequences used in this study.

Table S2. Feeding behaviour parameters from B. tabaci after feeding on RNAi transgenic plants expressing dsRNA of BtE3 (dsBtE3) or GFP (dsGFP).

Author contributions

YJZ and QS conceived and designed the experiments; ZKP, JR, and LXT conducted the experiments; YZ and YTY helped perform the experiments and participated in data analysis; ZKP and QS wrote the manuscript; SLW, WX, QJW and ZYL helped revise the manuscript. All authors gave final approval for publication.

Conflict of interest

The authors declare no conflict of interest.

Funding

This work was supported by the National Key R & D Program of China (2022YFD1400800, 2022YFD1401200), the National Natural Science Foundation of China (32172388), China Agriculture Research System (CARS-24-C-02), the Beijing Key Laboratory for Pest Control and Sustainable Cultivation of Vegetables, and the Science and Technology Innovation Program of the Chinese Academy of Agricultural Sciences (CAAS-ASTIP-IVFCAAS), the Outstanding Youth Science and Technology Innovation Team Project of Colleges and University in Hubei Province (T2022009).

Data availability

The data supporting the findings of this study are available from the corresponding author (YJZ) upon request.

References

Author notes