Rickettsia transmission from whitefly to plants benefits herbivore insects but is detrimental to fungal and viral pathogens

ABSTRACT Bacterial endosymbionts play important roles in the life histories of herbivorous insects by impacting their development, survival, reproduction, and stress tolerance. How endosymbionts may affect the interactions between plants and insect herbivores is still largely unclear. Here, we show that endosymbiotic Rickettsia belli can provide mutual benefits also outside of their hosts when the sap-sucking whitefly Bemisia tabaci transmits them to plants. This transmission facilitates the spread of Rickettsia but is shown to also enhance the performance of the whitefly and co-infesting caterpillars. In contrast, Rickettsia infection enhanced plant resistance to several pathogens. Inside the plants, Rickettsia triggers the expression of salicylic acid-related genes and the two pathogen-resistance genes TGA 2.1 and VRP, whereas they repressed genes of the jasmonic acid pathway. Performance experiments using wild type and mutant tomato plants confirmed that Rickettsia enhances the plants’ suitability for insect herbivores but makes them more resistant to fungal and viral pathogens. Our results imply that endosymbiotic Rickettsia of phloem-feeding insects affects plant defenses in a manner that facilitates their spread and transmission. This novel insight into how insects can exploit endosymbionts to manipulate plant defenses also opens possibilities to interfere with their ability to do so as a crop protection strategy. IMPORTANCE Most insects are associated with symbiotic bacteria in nature. These symbionts play important roles in the life histories of herbivorous insects by impacting their development, survival, reproduction as well as stress tolerance. Rickettsia is one important symbiont to the agricultural pest whitefly Bemisia tabaci. Here, for the first time, we revealed that the persistence of Rickettsia symbionts in tomato leaves significantly changed the defense pattern of tomato plants. These changes benefit both sap-feeding and leaf-chewing herbivore insects, such as increasing the fecundity of whitefly adults, enhancing the growth and development of the noctuid Spodoptera litura, but reducing the pathogenicity of Verticillium fungi and TYLCV virus to tomato plants distinctively. Our study unraveled a new horizon for the multiple interaction theories among plant-insect-bacterial symbionts.

In the case of Rickettsia, plants can serve as reservoirs for horizontal transmission of endosymbiotic Rickettsia among whiteflies.Caspi-Fluger et al. (21) and Li et al. (23) demonstrated that horizontal transmission of Rickettsia occurred between two different cryptic species of whitefly Bemisia tabaci (Gennadius), Middle East-Asia Minor 1 (MEAM1), and Mediterranean (MED), via cotton plants.These two species are biologi cally differentiated and reproductively isolated (26), which excludes the possibility that transmission occurs during mating.
The aforementioned studies have focused on the transmission routes and the direct effects of endosymbiont acquisition by herbivorous insects.However, they did not consider the possibility that, endosymbionts that are transmitted to plants may influence plant traits, such as growth and defense, and thereby indirectly influence the perform ance of the insects (19).We here specifically test if the exogenous introduction of a bacterial endosymbiont may trigger host plant responses that, in turn, may affect the performance of insects and pathogens on the same host plant.
We previously demonstrated that Rickettsia belli can infect all life stages of B. tabaci, including eggs, nymphs, and male and female adults (27).Importantly, Himler et al. (28) found that, compared to uninfected whiteflies, Rickettsia-infected whiteflies produce more offspring, have higher survival to adulthood, develop faster, and produce a higher proportion of daughters.However, the mechanism that mediates these benefits to the whitefly host has remained unknown.In the current study, we used a combination of molecular tools, biochemical analyses, and performance experiments to investigate how the horizontal transmission of Rickettsia endosymbionts from phloem-feeding whitefly adults into tomato plants affect insect and pathogen performance on these plants (Fig. S1).For the first time, our study has revealed the consequent effects of endosymbiont persistence in host plants.The transmission of Rickettsia into tomato plants makes the plants more suitable for their normal herbivorous host, the whitefly B. tabaci, as well as a chewing herbivore, the noctuid S. litura (Fabricius).In contrast, the Rickettsia-induced changes made the plants more resistant to a pathogenic fungus (Verticillium dahliae Klebahn), and two begomoviruses (Tomato yellow leaf curl virus, TYLCV; Papaya leaf curl China virus, PaLCuCNV) vectored by whiteflies.

The distribution and persistence of Rickettsia in tomato plants
The distribution and persistence of Rickettsia in tomato plants was detected by PCR, fluorescence in situ hybridization (FISH), and transmission electron microscopy (TEM).The titer of Rickettsia in tomato plants at different time periods was detected by quantitative real-time PCR (qRT-PCR).The PCR analyses confirmed that the endosym biont, Rickettsia belli, can be transmitted from B. tabaci whiteflies into the phloem of tomato leaves.We also detected Rickettsia gene PCR products in the leaves adjacent to whitefly-infested leaves (Fig. 1A).The BLAST results of three genes specific for Rickettsia, namely, gltA, Pgt, and 16S rRNA, confirmed that the Rickettsia endosymbionts from B. tabaci and tomato leaf samples were identical to each other.They fully matched the gltA, Pgt, and 16S rRNA gene sequences of Rickettsia in GenBank (KX645660-KX645662).
The visualization using FISH confirmed that, after entering the tomato plant, Rickettsia was located exclusively inside the phloem vessels of plant leaves.From there, Rickettsia moved from the whitefly-infested leaf through the phloem to the younger and older neighboring leaves (Fig. 1B).There was no FISH signal found in the control plant leaves (Fig. S2).
The TEM images revealed the presence of Rickettsia in the B. tabaci adult abdomen (Fig. 1C).The Rickettsia cells are rod-shaped with a cell wall structure and are approxi mately 2.0 µm long and 0.5 µm wide.In tomato leaves, we found Rickettsia in the phloem sieve vessels.The Rickettsia cells in the tomato leaves were morphologically similar to those in the abdomen of B. tabaci adults though the size varied among individual cells (Fig. 1D).
Once Rickettsia was transmitted to the tomato plant by the whitefly, they spread to undamaged leaves.The relative quantities of Rickettsia, as determined with qPCR, in the leaves neighboring the infested leaf increased during the first 21 days.After that, the quantities started to decrease (Fig. 2).The relative quantity of Rickettsia on the 21st day was approximately 300 times higher than that on the 7th day, and the relative quantities on the 14th, 28th, and 35th days were about 48, 175, and 132 times higher than that of the 7th day, respectively.
In the gene ontology (GO) analysis, we found 17 GO terms, 3 in biological process (GOBP), 7 in cellular component (GOCC), and 7 in molecular function (GOMF), were significantly enriched in both Ctrl vs R− and Ctrl vs R+ (Fig. S5).However, these 17 GO terms were not significantly enriched in R+ vs R−.Between Ctrl vs R+ and R+ vs R−, 6 GOBP and 7 GOMF were significantly enriched in both comparisons.The Kyoto Encyclopedia of Genes and Genomes (KEGG) signaling pathway analyses revealed that plant-pathogen interaction (ko04626) and plant hormone signal transduction (ko04075) were significantly enriched (Fig. S6).
To validate the RNA-Seq results, 16 DEGs with different expression patterns were selected for qPCR validation (Fig. S7).Similar results were obtained from both qPCR and RNA-Seq analyses: feeding by R+ whiteflies increased expression levels of SA-responsive WRKY70, PR1, VRP, and TGA 2.1 compared with feeding by R− whiteflies (Fig. 3A).In contrast, expression levels of JA-responsive PI-II, AOC, and LOX genes were lower, while expression level of JAZ1 gene was higher, in tomato plants fed upon by R+ whiteflies compared to the plants fed upon by R− whiteflies (Fig. 3A).
To further examine the effect of Rickettsia on tomato defenses, we measured the levels of endogenous salicylic acid (SA), jasmonic acid (JA), and JA-Ile in plants infested by R+ or R− whiteflies and undamaged plants.After 7 days infestation, the level of SA in plants infested by R+ whiteflies was significantly higher than that in plants infested by R− whiteflies (P = 0.02; Fig. 3B).In contrast, the levels of JA and JA-Ile in plants infested by R+ whiteflies were significantly lower than those in plants infested by R− whiteflies (JA: P = 0.01; JA-Ile: P = 0.02; Fig. 3C and D).To examine whether the increased SA levels affect the suppression of JA-regulated defense in infested R+ plants, we measured the expression of SA-and JA-regulated genes in Moneymaker (wild type) and SA-deficient NahG tomato plants infested by R− or R+ whiteflies.In Moneymaker, compared to feeding by R− whiteflies, feeding by R+ whiteflies increased the expression of SA-regulated PR-1, TGA2.1, and VRP but decreased the expression of JA-regulated PI-II and LOX (Fig. 4).In contrast, the suppression of SAregulated genes (PI-II and LOX) and the induction of JA-regulated genes did not occur in NahG plants infested by R+ whiteflies (Fig. 4).These results indicate that negative crosstalk between the JA and SA pathways was involved in JA defense suppression by Rickettsia infection.The expression patterns of VRP were similar on both plant types, implying that its expression is independent of SA.

Rickettsia infection promotes the susceptibility of tomato plants to herbi vores
To investigate whether Rickettsia infection affects the susceptibility of tomato plants to subsequent infestations by herbivores, we evaluated the performance of whitefly and S. litura larvae on plants pre-infested by R− or R+ whiteflies.Female whiteflies laid more eggs on plants pre-infested by R+ whiteflies than on plants pre-infested by R− whiteflies (Fig. 5A).In addition, the female proportion of progeny reared on plants pre-infested by R+ whiteflies was about 50% higher than on plants pre-infested by R− whiteflies (Fig. 5B).Yet, Rickettsia infection did not significantly affect the survival rate and developmental time of whitefly (Fig. 5C and D).
Feeding S. litura larvae on plants that were pre-infested with R+ whiteflies resulted in larvae with higher body weights and shortened pupation times as compared to feeding them on control plants (Fig. 5E and F).In contrast, the body weight and pupation time of S. litura reared on plants pre-infested with R− whiteflies showed no such effect (Fig. 5E  and F).These results indicate that Rickettsia infection played a crucial role in suppressing the defense of tomato against herbivores.

Rickettsia infection induced resistance of tomato plants against pathogens
To investigate if the Rickettsia infection affected the resistance of tomato plants against pathogens, we measured the performance of the fungal pathogen V. dahlia and the expression levels of the viruses TYLCV and PaLCuCNV, which are all known to be vectored by whiteflies.At the 14 and 28 days after V. dahliae infection, the wilt symptoms caused by V. dahliae on plants pre-infested with R+ whiteflies were significantly lower than those on plants pre-infested with R− whiteflies (Fig. 6A and 7).The feeding by R+ whiteflies also suppressed expression levels of TYLCV and PaLCuCNV at 72 h as compared to those by R− whiteflies but not at 48 h after virus infection (Fig. 6B and C).These results imply that the infection with Rickettsia positively increased the resistance of tomato plants against pathogens.

DISCUSSION
Manipulation of the plant immune system by insects has largely focused on effectors that are produced by the plant antagonists themselves (29)(30)(31)(32)(33), or microorganisms that they carry with them (34).Here, we revealed the mechanism how the transmission of an endosymbiont by an insect to a host plant enhances the suitability of the plant for the insect, which was also previously reported by Chung et al. (34) and Chen et al. (35).Various studies have revealed that host plants can vector the horizontal transmission of endosymbionts between different populations, or even different species, of phloem feeding insects (20,21,23,24,36).On the other hand, herbivore-associated microorgan isms can be transmitted to a plant and spread throughout a plant and persist there while the herbivore insect continuously feeds (23,37).Here, we reveal that Rickettsia is, indeed, vectored by whiteflies and infects tomato plants (Fig. 1).It spreads and stays active in the phloem for at least 5 weeks (Fig. 2).Gene-expression and phytohormone analyses further showed that the infection of Rickettsia vectored by whiteflies enhances SA-regulated defenses, but suppresses JA-regulated defenses (Fig. 3), which are defenses that strongly affect whitefly performance (38)(39)(40).Our study demonstrates that endosymbionts can infect and modify plant defenses in a manner that benefits their insect host.
Various studies have shown that microbial symbionts can provide herbivore insects with essential nutrients including B vitamins and also enhance their defenses against predators (15,41), but the role of insect symbionts as effectors in modifying plantinsect interactions has received scant attention.In apple, Wolbachia-infected leaf miners (Phyllonorycter blancardella) elicit a green-island phenotype, which preserves photosyn thetically active tissue in senescent leaves and indirectly enhances leaf miner perform ance (42).In maize, it has been proposed that Wolbachia may suppress plant defenses against the western corn rootworm (Diabrotica virgifera virgifera) (43), but this has been questioned (44).It remains uncertain that insects secrete symbionts or symbiont-derived compounds into plants during feeding for the purpose of plant manipulation.Our study reveals that Rickettsia associated with whiteflies are secreted into tomato leaves and stay active and reproduce for more than 1 month.Moreover, the plant defense suppression by Rickettsia was found to enhance the performance of phloem-feeding B. tabaci and chewing S. litura caterpillars (Fig. 5).The long-term colonization of Rickettsia in a tomato plant in our study has shown that it can change the plant defenses.However, previous studies revealed that some aphid symbiotic bacteria, which are transmitted into the plant by aphids, can manipulate the plant defenses via attenuating the plant's volatile emissions although these bacteria do not multiple themselves within the plant (45).This kind of manipulation may occur through the symbiont effectors delivered with aphid saliva, or through the deposition of symbionts on the plant surface via honeydew that are then injected into the plant while aphids are piercing (46,47).In our current study, the same situation may also exist; therefore, further studies need to be undertaken to isolate the manipulators of plant defenses including the symbiotic bacteria in plant phloem and in the honeydew of phloem-sucking insects, as well as in the saliva of piercing vector insects.All these findings imply that insect symbionts play a more intricate role in the interaction between plants and insects than previously thought (15).
Our validation experiments using SA-deficient NahG mutants and wild-type tomato plants revealed the importance of crosstalk between the SA and JA signaling pathways.Rickettsia-mediated defense suppression was not observed in SA-deficient NahG plants, indicating that suppression of JA-regulated defenses is linked to the upregulation of the SA signaling pathway (48).Indeed, it has been well documented that pathogens may manipulate the antagonistic effects between JA and SA and, thereby, suppress host plant defenses so enhancing the performance of insect vectors (49,50).It is important to note that also in NahG plants, feeding by R+ whiteflies increased the expression level of the pathogen-resistance gene VRP compared to feeding by R− whiteflies (Fig. 4).This indicates that Rickettsia presence enhances the expression of VRP via a mechanism that is not dependent on the SA or JA pathway.
The behavioral and performance assays showed that R+ B. tabaci females laid more eggs than R− B. tabaci females (Fig. 5A).There may be three non-exclusive reasons for this.First, carrying Rickettsia might increase whitefly fecundity as we also found on cotton plants (Shi et al., unpublished data).Second, by suppressing the JA-mediated plant defenses, the Rickettsia infection may increase plant quality and, thus, indirectly the fitness of B. tabaci.Third, the females may detect the superior quality of the Rickettsiainfected plants and prefer to lay more eggs on them than on uninfected plants.The females did not only lay more eggs on R+ plants, but these eggs also resulted in more female progeny, which may further increase their fitness (Fig. 5B).
It is important to note that feeding by R− whiteflies triggered the expression of JA-regulated defense genes (LOX, AOC, and PI-II; Fig. 3A).This means that, without a Rickettsia infection resulting from whitefly feeding, host plants can perceive the infesting insects and respond with appropriate defense responses.A similar phenom enon was observed in the interaction between the Colorado potato beetle (Leptinotarsa decemlineata) and tomato plants, whereby larvae without symbionts activated stronger JA-regulated defenses than those with symbionts, which resulted in reduced larval growth (36).In the current study, the activation of JA-regulated genes induced by R− whiteflies seems not to significantly affect the survival and development of whitefly adults (Fig. 5C and D) as previously found (51), but compared to infections by R+ whiteflies, it reduces fecundity (Fig. 5A) and affects the insect's sex ration (Fig. 5B).
Our study revealed that whitefly-infestation significantly reduced the plants' susceptibility to Verticillium wilt (Fig. 7) and to the whitefly-vectored viruses TYLCV and PaLCuCNV (Fig. 6).We deduce that by itself whitefly induction of the SA-pathway enhances the resistance to these important plant pathogens and that Rickettsia further synergizes the plant responses (Fig. 3 and 4), making the plants even more resistant (Fig. 6).As far as we are aware, this is the first study to show that feeding by whitefly induces resistance to pathogen infestation and that Rickettsia further synergizes this effect.
In conclusion, our study answers two important and unresolved questions in whitefly-plant interactions.Walling (52) hypothesized that the mechanism by which the whitefly B. tabaci down-regulates JA defenses via SA cross-talk in host plants involves a salivary component synthesized by the whitefly or one of its endosymbionts.In another insightful review on the horizontal transmission of endosymbionts via the host plants, Chrostek et al. (19) asked the question "how do insects symbionts influence plants?" after their horizontal transmission.Here, we show that the horizontal transmission of Rickettsia from whitefly to plants enhances the plants' suitability to insect herbivores and simultaneously makes them more resistant to pathogens.In addition, a field survey from 2011 to 2022 showed that the B. tabaci population that harbors Rickettsia was gradually increasing and becoming the dominant species in Guangdong, China (Qiu et al., unpublished data).These data further indicate that the bacterial symbionts may be consequently changing the structure of the herbivore via the plant defenses at a community level.As far as we know, our study represents the first report of an adaptive transmission by insects of an endosymbiont to plants in order to manipulate the plants' immunity in a manner that benefits the insect as well as the symbiont.If and how the endosymbiont may also provide certain benefits to plants is worthy of further investiga tion.

Host plants
Seeds of the tomato plant Lycopersicon esculentum Miller (var.Xinjinfeng no. 1, Changhe Seed Co. Ltd., Guangzhou), a salicylic acid-deficient NahG mutant, and its wild type (variety Moneymaker) were sown in 15 cm diameter plastic pots containing a soil-sand mixture (10% sand, 5% clay, and 85% peat).The seedlings were cultured at ambient temperature and photoperiod in a glasshouse at South China Agricultural University (SCAU) (with a mean of 27.5°C and 12.2 h light per day during September-October, 2017).The plants were watered as required and were used for experiments at the six to eight expanded leaf stage.All the leaf surfaces were disinfected with 75% alcohol and allowed to air dry to remove other phyllospheric microorganisms before undertaking experiments.

Chewing herbivore Spodoptera litura
Spodoptera litura larvae (Rickettsia negative [see the supplemental material]) were originally collected on broccoli plants at the training farm of SCAU, Guangzhou in the fall of 2015.Thereafter, it was mass reared on artificial diet (53) for at least eight generations.Both the whitefly and the noctuid insects were reared under the same conditions as described in the above section.

Pathogenic fungus Verticillium dahliae
V. dahliae (strain v991) was initially grown on potato dextrose agar (PDA) media and stored at 4°C before use (see the supplemental material).

Viruses
Clones of tomato yellow leaf curl virus (TYLCV; GenBank accession no.AM282874) and Papaya leaf curl China virus (PaLCuCNV; GenBank accession no.AM691554) were used for virus inoculation and transmission.

Horizontal transmission of Rickettsia from whitefly to tomato plants
To transmit Rickettsia to the tomato plants, 30 pairs of 24-48 h old R+ B. tabaci adults were collected from the R+ subcolony and released into a nylon bag (10 × 15 cm, 70 mesh/cm 2 ) that covered one tomato composite leaf.After 1 week of whitefly feeding, the distribution and persistence of Rickettsia in the tomato plant was examined as described below.

The distribution and persistence of Rickettsia in tomato plants
The persistence and localization of Rickettsia in the tomato plants was assessed using PCR, fluorescence in situ hybridization (FISH) and transmission electron microscopy (TEM) techniques.

PCR detection
One upper or lower tomato leaf neighboring the leaf infested with R+ whitefly (about 2-3 cm distance between each of the two leaves on the stem) was cut and homogenized in lysis buffer for DNA extraction (TIANamp Genomic DNA Kit, Tiangen Biotech Co. Ltd, Beijing, China).The specific primers used for Rickettsia detection were citrate synthase (gltA), phosphoglycerol transferase (Pgt), and 16S rRNA genes, which were amplified according to Caspi-Fluger et al. (21) and Gottlieb et al. (54) (Table S1).Portiera aleyrodida rum DNA was used as a positive control, and ddH 2 O was used as a negative control.All the gltA, Pgt, and 16S rRNA genes of Rickettsia amplified from plant leaves were sequenced and then searched using BLAST in GenBank to confirm their strains.Twelve tomato plants were used in the PCR detection.

Fluorescence in situ hybridization detection
A strip of tomato leaf (10 mm × 5 mm, 0.01 g) was cut longitudinally along both sides of the midrib from the uninfested leaf neighboring the leaf infested with R+ whitefly.This leaf strip was used to visually inspect for Rickettsia presence and localization using FISH.The leaves were placed in Carnoy's fixative, and FISH was performed with the symbiont-specific 16S rRNA probe for Rickettsia (Rb1-Cy5: 5′-TCCACGTCGCCGTCTTGC-3′) (54).Stained tomato leaves were mounted and viewed under a Nikon eclipse Ti-U inverted microscope.Healthy tomato leaves infested with R− whitefly and leaves infested with R+ whitefly but without symbiont-specific 16S rRNA probe hybridization were used as negative controls.Eighteen tomato plants were used in the FISH detection.

Transmission electron microscope detection
The Rickettsia localization in tomato leaves was detected with a transmission electron microscope (TEM) according to the method of Li et al. (24).Briefly, samples of Rickettsia infested tomato leaves (1.0 mm × 0.5 mm) were fixed in 4% glutaraldehyde in cacody late buffer (pH 7.4) at 4°C for 24 h and then overnight in 1% osmium tetroxide.The fixed leaf samples were dehydrated through an alcohol series and embedded in Spurr's resin.Ultrathin sections were collected on copper grids with a single slot, stained with 1% uranyl acetate and lead citrate, and finally examined under transmission electron microscopy (JEOL, Tokyo, Japan).Nine tomato plants were used in the TEM detection.

Quantitative real-time PCR detection
The persistence of Rickettsia in tomato leaves was quantified with quantitative real-time PCR (qPCR) after 7, 14, 21, 28, and 35 days of R+ whitefly feeding (n = 3 per time point).The leaf sampling and DNA extraction were as described above.The primers used for Rickettsia qPCR detection were gltA, and two housekeeping genes of the tomato plant, RuBisCo and β-actin, were selected as internal controls for data normalization and quantification (38).Statistically significant differences of the relative quantity of Rickettsia were determined by analyzing the data with ANOVA followed by Duncan's test at alpha = 0.05 (SPSS vs 18.0, SPSS Inc. Chicago, USA).

Plant pre-treatments with whitefly
In order to determine the response of tomato plants to whitefly feeding and Rickettsia persistence, tomato plants were grown as described above for 4 weeks after which they were subjected to the following three treatments: (1) plants were pre-infested with R+ whiteflies for 7 days (Rickettsia positive [R+) (2); plants were pre-infested with R− whiteflies for 7 days (Rickettsia negative [R−]) (3); uninfested plants (as control plants in the experiment, hereafter abbreviated to "Ctrl").Approximately 60 pairs of R+ or R− whitefly adults (24-48 h old) were collected and divided into two nylon bags that covered the upper and lower surfaces of one tomato leaf, whereas control plants received empty sleeves.
The R+ and R− B. tabaci adults were used to induce the plant defense with a 7 day-infestation in all experiments.This was based on our previous finding (55) that, when using an exotic biological or chemical factor to induce plant defenses, the endogenous JA and SA usually displayed distinct changes from the 4th day onwards and peaked on the 11th day after whitefly feeding.Thus, we chose the plants pre-infested with whiteflies for 7 days to investigate the JA-and SA-affected performance of the herbivores and pathogen in our current study.

Transcriptomics analysis of tomato plants feeding with different whiteflies
The differentially expressed genes (DEGs) in different treatments of tomato plants were analyzed.The middle tomato leaves neighboring the whitefly-infested ones were harvested for RNA extraction after 7 days of whitefly feeding.Six biological replicates (leaves) in each treatment were randomly split into two groups of three plants each for RNA sequencing and qPCR analysis, respectively.

RNA sequencing and bioinformatics analysis
The total RNA of a tomato plant was extracted using TRIzol Reagent (Invitrogen, Guangzhou).The mRNA was purified using the NucleoTrap mRNA kit (Macherey-Nagel, Düren, Germany).Then, the cDNA samples were constructed using the Standard cDNA Synthesis Kit (Takara, Japan) according to Marioni et al. (56).All cDNA were sent to Beijing Genomics Institute (Shenzhen, China) for sequencing with the BGISEQ-500RS platform.SOAPnuke1.5.6 was used for reads trimming to remove adaptors and low-quality bases.HISAT (version 2.0.4) and Bowtie2 (version 2.2.5) were used to index reference genome and reads mapping.Tomato genome annotation ITAG3.2 was used as the reference genome.RSEM (version 1.2.8) was used to generate the count matrix.Reads count was normalized with FPKM, and NOISeq was used to identify differentially expressed genes (57).We selected the genes with a log 2 fold change ≥1.5 and deviation probability value ≥0.8 of each plant sample for further analyses.Gene set enrichment analysis was performed with AgriGO (version 2.0) (58).All three Gene Ontology (GO) (i.e., Molecular Function, Biological Process, and Cellular Component) and KEGG Plant pathways (59) were used.An adjusted P value < 0.05 cutoff was used for selecting significantly enriched GO terms and KEGG pathways.

Validation of differential expression genes by qPCR
To validate the RNA seq result, 16 DEGs were randomly selected for further qPCR validations (Fig. S2).To identify potential genes that were involved in the response of the tomato plant caused by exogenous Rickettsia, we validated the expression changes of 6 DEGs associated with immune response-related GO terms and KEGG pathways.These genes are WRKY70 and PR1-(P4) (pathogenesis-related protein 1), indicative for activation of the salicylic acid (SA) pathway, and AOC (allene oxide cyclase), LOX (linoleate 13S-lipoxygenase 2-1), PI-II (proteinase inhibitor II), and JAZ1 (jasmonate ZIM-domain protein 1), indicative of a response in the jasmonic acid (JA) pathway.In addition, we selected two pathogen-resistance genes, VRP (Verticillium wilt disease resistance 2) and TGA2.1, for further quantification by qPCR.
The primers and protocol for qPCR quantification are specified in Table S2.The PCRs were performed on Bio-Rad CFX96 Real-Time System using SYBR Green (Bio-Rad, USA).The relative expression levels were calculated with the 2 −ΔΔCT method and further log 2 transformed, and the RuBisCo gene (ribulose-bisphosphate carboxylase) of the tomato plant was used as a reference gene.

Analysis of JA, JA-Ile and SA
The phytohormones JA, JA-Ile, and SA were analyzed as described by Engelberth et al. (60) with modification.In brief, plant material (250-300 mg) was frozen and ground in liquid nitrogen.For quantification purposes (9, 10), dihydro-JA (15 ng; Sigma-Aldrich, St Louis, MO, USA) and D6-SA (20 ng; CDN Isotopes, Pointe-Claire, Quebec, Canada) were added as internal standards with 2 mL of 80% methanol.JA, JA-Ile, SA, and the internal standards were partitioned to an aqueous phase by centrifugation and vaporization.Subsequently, they were extracted from the aqueous phase with an equal volume of ethyl acetate and then dried.The dried extract was resuspended in 0.1 M acetic acid and loaded onto a C18 column (Waters Company, Milford, MA, USA).The C18 column was sequentially eluted with a series of solvent mixtures [acetic acid/methanol (vol/vol) at 83/17, 60/40, and 40/60].The effluents of the last 4 mL in 40% methanol and the first 3 mL in 60% methanol were collected.After evaporation of the solvent and esterification of the residue using excess ethereal diazomethane, samples were analyzed using a gas chromatograph coupled to a mass selective detector (6890N/5973 MSD, Agilent Technologies, Inc., Palo Alto, CA, USA), which was operated in electron impact ionization mode.The compounds in the samples were separated on an HP-5-MS column (30 mm × 0.25 mm × 0.25 mm; 19,091 S-433, J&W Scientific, Agilent Technologies).JA, JA-Ile, and SA were quantified by correlating the peak area (extracted ion) of the compound with the peak area of the respective deuterated internal standard.Instead of measuring D6-MeSA, we measured D4-MeSA due to the loss of two deuterium ions during sample preparation.

Validation of related expression genes in SA-deficient NahG plants
Based on the results of the DEGs and qPCR analyses, the role of SA and cross-talk between the JA and SA signal transduction pathways were further studied.To do so, the same treatments (R+, R−, and Ctrl, n = 6 per treatment) were applied to SA-deficient NahG mutant tomato plants and their wild-type, cv.Moneymaker.The expression of WRKY70, PR1, AOC, LOX, PI-II, VRP, and TGA2.1 was quantified by qPCR as described above.Three technical qPCR replicates were analyzed for each biological replicate, and the statistically significant differences between the relative expression quantities of DEGs were analyzed using ANOVA followed by Duncan's test.

Effects of Rickettsia persistence on the performance of Bemisia tabaci
Tomato plants in this experiment were pre-treated as above.Approximately 30 pairs of R+ or R− B. tabaci adults were allowed to feed for 7 days before being removed.Hereafter, one pair of B. tabaci adults (24-48 h old) was introduced into another nylon bag (20 × 30 cm) that covered one healthy, fully expanded tomato leaf of a pre-treated plant.The fecundity of the female (numbers of eggs/female) was examined until the female died.In a second experiment, 30 pairs of B. tabaci adults (24-48 h old) were introduced into another nylon bag covering one healthy, fully expanded leaf of a pre-treated plant.The adult whiteflies were removed after 24 h.Then, the development of whitefly progeny including the survival (egg-adult), developmental time of whitefly nymphs as well as the sex ratio of F1 generation adults were assessed.
There were three biological repeats for each treatment, and each repeat included 10 plants.Statistically significant differences in fecundity, survival, development time of B. tabaci among the differently treated tomato plants were analyzed using ANOVA followed by Duncan's test.

Effects of Rickettsia persistence on the performance of Spodoptera litura
Tomato plants in this experiment were pre-treated as above.After 7 days of R+ or R− whitefly feeding, one freshly molted 2nd instar S. litura larva was introduced onto an uninfested leaf of R+, R-, and Ctrl plants within a nylon bag using a soft brush.On the 22nd day, the body mass of the S. litura larva was measured with a digital balance (0.1 mg precision, CP214, Ohaus Co. Ltd., China).In addition, the period of S. litura from larvae to pupae was recorded.The experiments were repeated three times, and each time 10 individuals of S. litura larvae were investigated for every treatment.Statistically significant differences in the developmental time and body weight of S. litura fed on the differently treated plants were analyzed using ANOVA followed by Duncan's test.

FIG 1
FIG 1 Location of Rickettsia in Bemisia tabaci and in tomato leaves.(A) PCR detection of Rickettsia in different tomato leaves.M: DNA marker; P: 16S rRNA of Portiera, positive control; N: ddH2O, negative control.Lanes 1-3: DNA samples extracted from tomato leaf infested by R+ whiteflies, and the leaves situated above and below of an infested leaf, respectively.(B) Rickettsia visualization in tomato leaves with fluorescence in situ hybridization.Tomato leaves were taken from plants infested by R+ whiteflies.Left: sample of a leaf above the infested leaf; Right: sample of a leaf below the infested leaf.Rickettsia-specific 16S rRNA was used as probe.(C) Detection of Rickettsia in the abdomen of an adult R+ whiteflies by transmission electron microscope; (D) Localization of Rickettsia in phloem sieve tube cells of a tomato leaf infested by R+ whiteflies.ST, phloem sieve tube; CW, cell wall of the plant phloem; V, vacuole; M, mitochondrion; R, Rickettsia endosymbiont.

FIG 2
FIG 2 The persistence of Rickettsia in tomato leaves adjacent to a whitefly infested leaf.For Rickettsia detection, glt A qRT-PCR primers were used.The expression levels of two tomato housekeeping genes, RuBisCo and β-actin, were used for normalization.Different letters over the bars indicate statistically significant differences, error bars are standard deviations (n = 15; Duncan's multiple range test, alpha = 0.05 level).

FIG 4
FIG 4 Transcript levels of SA-and JA-regulated genes in wild-type Moneymaker and SA-deficient NahG plants infested with R− whiteflies or R+ whiteflies.Values are untransformed means ± SE (n = 3).The tomato household RuBisCo gene was used for normalization of the expression levels.Different letters over the bars indicate statistically significant differences (Duncan's multiple range test, alpha = 0.05 level).Ctrl, undamaged plants; R−, plants infested with R− whiteflies; R+, plants infested with R+ whiteflies.

FIG 5
FIG 5 Infection of Rickettsia increases susceptibility to insect herbivores.(A) Average fecundity, (B) sex ratio of the progeny, (C) survival rate from egg to adult, (D) developmental period from egg to adult of Bemisia tabaci infested on different tomato plants, (E) larval mass at day 22, and (F) pupation time of S. litura fed on different tomato plants.Values are means ± SE (n = 10).Different letters over bars indicate significant differences between treatments (Duncan's multiple range test, alpha = 0.05 level).Ctrl, undamaged plants; R−, plants pre-infested by R− whiteflies; R+, plants pre-infested by R+ whiteflies.

FIG 6
FIG 6 Infection of Rickettsia induces resistance against pathogens.(A) Disease severity assessment of V. dahliae, (B) expression of TYLCV, and (C) expression of PaLCuCNV on different tomato plants.The tomato household RuBisCo gene was used for normalization of the expression levels.Values are means ± SE (n = 3).Different letters over bars indicate significant differences between treatments (Duncan's multiple range test, alpha = 0.05 level).Ctrl, undamaged plants; R−, plants pre-infested by R− whiteflies; R+, plants pre-infested by R+ whiteflies.

FIG 7
FIG 7 The disease symptoms of V. dahliae infesting different tomato plants.(a-c) V. dahliae symptoms of Ctrl, R− and R+ plants at day 14, (d-f ) V. dahliae symptoms of Ctrl, R− and R+ plants at day 28.Treatment R+: tomato pretreated with R+ B. tabaci for 7 days, R−: tomato pretreated with R− B. tabaci for 7 days, Ctrl: uninfested tomato leaves as controls.