Proteomic analysis of watery saliva secreted by white-backed planthopper, Sogatella furcifera

The white-backed planthopper, Sogatella furcifera, is a phloem sap feeder that secretes watery and gelling saliva during feeding. In this study, we identified the major proteins in watery saliva of S. furcifera by shotgun LC-MS/MS analysis combined with transcriptomic analysis. A total of 161 proteins were identified, which were divided into 8 function categories, including enzymes, transporter, calcium ion binding protein, salivary sheath protein, cytoskeleton protein, DNA-, RNA-, and protein-binding or regulating proteins, other non-enzyme proteins and unknown proteins. Gene expression pattern of 11 secretory proteins were analyzed by real time quantitative-PCR. We detected the mucin-like protein, which had a unique expression level in salivary gland, most likely to be a candidate effector involved in regulation of plant defense. This study identified the watery saliva component of S. furcifera and it provided a list of proteins which may play a role in interaction between S. furcifera and rice.


Introduction
Saliva is an important biochemical interface determining the compatibility between sap-sucking insects and their host plants [1,2]. The saliva includes bioactive compounds that have a range of functions from degrading plant cell and digesting nutrients to eliciting or inhibiting plant defense [3]. Hemipterans secrete two types of saliva: gelling saliva and watery saliva. Gelling saliva is used to form salivary sheath in order to facilitate stylet penetration, while watery saliva involves in the regulation of plant defense [4].
White-backed planthopper, Sogatella furcifera, a critical phloem sap feeder, causes considerable damage by sucking plant sap and transmitting plant viral diseases [5]. It also secretes gelling and watery saliva during feeding process like other hemipterans. Previous studies have detected the saliva composition in many species of aphids [1,[6][7][8][9], Nephotettix cincticeps [10], Lygus Hesperus [11] and Nilaparvata lugens [4,12]. Huang et al. [13] compared the secreted saliva composition of three planthopper species and revealed the ubiquitous and specific saliva a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 rpm. Then centrifuge at 14 000 g for 15 min. Sample was added 100 μL UA buffer and centrifuged at 14 000 g for 15 min. Perform this procedure twice. Then sample was digested with 1 μg trypsin (Promega, Madison, USA) in 40 μL 100 mM NH 4 HCO 3 (Sigma, St. Louis, MO, USA) buffer. After vortexing for 1 min at 600 rpm, the mixture was reacted for 16-18 h at 37˚C. Sample was centrifuged at 14 000 g for 15 min and transferred to a new collection tube. 40 μL 25 mM NH 4 HCO 3 was added to the tube and then centrifuged at 14 000 g for 15 min to collect filtrate. C18 Cartridge (Sigma) was used to desalinate the peptides. 15 μL 0.1% formic acid (FA, Sigma) was added after lyophilizing the peptides. Peptides were qualified at OD 280 .
Peptide separations were analyzed using a Q Exactive mass spectrometer by dynamically choosing the 20 most abundant ions from one full mass scan (300-1800 m/z) for high-energy collisional dissociation (HCD) fragmentation. Normalized collision energy was 27 eV and dynamic exclusion was 60 s. Under fill ratio was defined as 0.1%. First level of mass spectrum resolution was 70 000 at m/z 200 and second level was 17 500 at m/z 200. Automatic gain control (AGC) target was 3e6.

Bioinformatic analysis
Gene Ontology (GO) annotations of identified proteins were assigned according to information available in the Swiss-Prot/TrEMBL (http://www.uniprot.org/) and Gene Ontology database (http://geneontology.org/). MS-identified proteins were categorized by molecular function, biological process and cellular component.

Prediction of secretory protein and amplification of candidate sequence
Probable secretory proteins were predicted by signal peptide and transmembrane domain. In eukaryotes, there are two cases for secretory proteins: (1) presence of a signal peptide and absence of transmembrane domain; (2) presence of a signal peptide and a transmembrane domain, and meanwhile, the transmembrane domain was in range of the signal peptide [9]. Proteins qualified the conditions were searched in the transcriptomic database of S. furcifera salivary gland and verified using blastx to search for similar sequences in NCBI non-redundant protein database with default parameters. The open reading frame (ORF) of each candidate gene was determined using the ExPASY Translate Tool (http://web.expasy.org/translate/). Gene-specific primers (S1 Table) designed by Primer Premier 5.0 were used to clone the complete ORF or partial sequences of each salivary protein gene. The total RNA from S. furcifera adult was extracted by using TRIzol Reagent (Invitrogen, MA, USA) according to manufacturer's instructions. Template cDNA was synthesized using the Fast Quant RT kit (TIANGEN, Beijing, China). The PCR products were cloned into the pEASY-T1 vector (TransGen Biotech, Beijing, China) and the insert was sequenced with standard M13 primers.

Real-time quantitative PCR (qPCR) for gene expression analysis
To investigate the tissue-, developmental stage-and sex-specific expression patterns of S. furcifera secretory protein genes, real-time qRCR was conducted using an ABI 7500 Real-Time PCR System (Applied Biosystems, Carlsbad, CA, USA). The tissues including salivary gland, head (without salivary gland), gut, testis, ovary and remaining body were dissected from S. furcifera adults under an anatomical lens (Leica Microsystems GmbH, Wetzlar, Germany) with microforceps (Shanghai Medical Instruments Ltd., Corp.) in chilled 1×Phosphate Buffered Saline (1×PBS) solution (pH 7.2, Life Technologies Corporation, NY, USA). Different developmental stages of S. furcifera including egg, 1 st -2 nd instar nymphs, 3 rd -4 th instar nymphs, 5 th instar nymphs, newly emerged female and male adults, and female and male adults after molting for 5 days were collected. For tissues, total RNA was isolated using the PureLink RNA Mini Kit (Life Technologies, Carlsbad, CA, USA). Total RNA of the whole insect was extracted by using TRIzol Reagent (Invitrogen).
Gene-specific primers (S2 Table) were designed using primer3 web (version 4.0.0) (http:// primer3.ut.ee/) and Beacon Designer 7.90, and the cDNA was prepared according to the instruction. The S. furcifera housekeeping gene Ribosomal protein L9 (GeneBank accession number KP735523) and Ribosomal protein L10 (GeneBank accession number KP735524) were used as internal control [16]. The specificity and efficiency of each primer was validated by analyzing standard curves with a fivefold cDNA dilution series.
Each qPCR reaction was conducted in a 20 μl mixture containing 10 μl of Bester 1 Sybr-Green qPCR mastermix (DBI 1 Bioscience, Germany), 0.4 μl of each primer (10 μM), 0.04 μl of 50 × ROX Reference Dye, 4 μl of sample cDNA and 5.16 μl of sterilized H 2 O. The firststrand cDNA and a no-reverse-transcription control were used as templates for three biological replicates under the following reaction program: an initial denaturation step at 95˚C for 5 min, followed by 40 cycles of 95˚C for 10 s and 60˚C for 31 s, melt curves stages at 95˚C for 15 s, 60˚C for 1 min, and 95˚C for 15 s. Relative quantification of salivary protein genes in different tissues and developmental stages was performed with the 2 -ΔΔCt method [17]. Data analysis was performed using the SPSS Statistics 20.0 software (IBM SPSS Statistics Inc., Chicago, IL, USA). A one-way nested analysis of variance (ANOVA) and Duncan's multiple range test (p < 0.05) were used to calculate the relative expression of each target gene. The values were presented as the mean ± SE when applicable.
Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis were used to identify the potential functions of S. furcifera water salivary proteins. For GO annotation, saliva components were classified at the second level under three root GO domains: biological process, molecular function and cellular component (Fig 1). The most two categories were metabolic process (including 70 proteins) and cellular process (69) in biological process, binding (75) and catalytic activity (71) in molecular function, cell (55) and cell part (54) in cellular component. KEGG pathways involved in the salivary proteins were divided into 5 branches: metabolism, genetic information processing, environmental information processing, cellular processes and organismal systems (Fig 2). Most proteins were distributed in metabolism (34) and genetic information processing (33) at the first level, while the majority proteins were related to energy metabolism (14), transcription (14) and transport and catabolism (15) at the second level.

Predicting, cloning and sequencing of the secretory proteins
Among the 161 proteins, 21 proteins have a putative secretory peptide and no transmembrane domain or the only transmembrane domain was in range of the signal peptide, suggesting secretory proteins. Other proteins without signal peptide indicate unknown secretory mechanism. The putative secretory proteins were searched in the transcriptome of S. furcifera salivary glands, 11 proteins were found to have the complete ORF, suggesting reliable for sequencing and cloning. Other proteins are not found in the transcriptome or only have a short fragment, so we excluded these proteins. All of the 11 salivary secretory proteins were manually searched by blastx program and then named according to the highest protein similarities with the high amio acid identities range from 54% to 100% in National Center for Biotechnology Information (NCBI) ( Table 2). Complete ORFs of 9 proteins were verified by cloning and sequencing, while other two proteins were confirmed with partial ORF. The data were deposited on NCBI (accession number from MF189025 to MF189034) except mucin-like protein, which was the same as the Accession number KX670544.

Tissue, development and sex-specific expression analysis
The expression of 11 secretory salivary protein genes in different tissues (salivary gland, head, gut, testis, ovary and remaining body) and different developmental stages and sexes (egg, 1 st -2 nd instar nymphs, 3 rd -4 th instar nymphs, 5 th instar nymphs, newly emerged female and male adults, and female and male adults after molting for 5 days) were determined by using qPCR. Figs 3 and 4 showed the results. Oxidoreductases are common enzymes in insect saliva. These enzymes can detoxify phenolic compounds in plant-defense reactions by changing the redox balance [3]. Oxidoreductase peptidylglycine α-hydroxylating monooxygenase was highly expressed in head, followed by ovary. This enzyme was extensively expressed in all developmental stages and no significant difference between female and male.
Hydrolases are related to plant-cell degradation by facilitating stylets penetration and movement in plant cell [3]. Carboxylesterase had the highest transcript level in ovary, followed by salivary gland, and then testis. Other tissues were almost not detected gene expression. This gene was highly expressed in female adult after molting for 5 days, suggesting it may play an important role in reproduction.
Proteases (peptidases) generally function as digestive enzymes and are important in detoxification of plant defense compounds [7]. We investigated two proteases in this study. Serine protease 6 had a unique expression in gut, indicating its digestive function. However, neprilysin-11-like isoform X4 was highly expressed in testis, followed by head and remaining body. The developmental stages and sex specific expression pattern between these two enzymes were also different. The expression level of serine protease 6 was increasing along with the development of insect in general except a decline in female adult. Neprilysin-11-like isoform X4 had a significantly high transcript level in newly emerged male adult. Lyases degrade substance without hydrolysis and oxidation and lyases in insect saliva were not widely known [3]. Carbonic anhydrase was expressed in all tissues but highly expressed in   gut. Adult expression level of this gene was higher than nymph and male expression was higher than female adult. Adult molting for 5 days was expressed higher than newly emerged adult. We tested an isomerase, protein disulfide-isomerase, which is involved in protein folding and synthesis. This gene was highest expressed in testis and then salivary gland and gut. Protein disulfide-isomerase was expressed in all developmental stages and had the highest expression level in newly emerged male adult. However, the expression decreased significantly in male adult after molting for 5 days, indicating it may function in the male mating with female adult.
ATP synthases are related to energy metabolism. Vacuolar ATP synthase subunit E had an extensive expression in all tissues and developmental phases, but it was expressed higher in gut and head than other tissues. Expression level of this gene in male adult was higher than female.
Lipophorin precursor and vitellogenin are lipid transport proteins that function in modulating immune responses in insects [18,19]. They share a similar expression pattern in tissues. They were both expressed highly in remaining body and head. As to developmental stages and sexes, they preform differently. Vitellogenin expressed significantly high in female adult molting for 5 days, indicating an important role in female reproduction. Lipophorin precursor expressed in all developmental stages and the expression level was increasing together with the insect development. Calcium ion binding proteins are recorded in saliva of many aphids. This kind of protein facilitates insects to feed on phloem sap of plants by preventing sieve element occlusion [6]. Calexcitin-1 isoform X1 was mainly expressed in remaining body and testis but hardly expressed in salivary gland. It will be interesting to find how calexcitin enters saliva of S. furcifera. Calexcitin-1 isoform X1 had a higher expression level in newly emerged male and female adult and male expression was higher than female.
Mucin-like protein was reported to play an integral role in the formation of salivary sheath [20]. The transcript of this protein was detected at significantly high level in salivary gland but was not detectable in head, gut, testis, ovary and remaining body. It was expressed in all stages and sexes while egg period had the lowest expression level and male adult had the highest one.

Discussion
We investigated the watery saliva component of S. furcifera in an attempt to understand the interaction between the insects and their host plants. This study has identified 161 proteins which have been classified into 8 categories according to the function. Recently, Huang et al. detected 177 proteins in the secreted saliva of S. furcifera. After comparison, we found that 10 salivary proteins in our study were the same with Huang's research, including salivap-5, plancitoxin-1, golgin subfamily A member 4-like, lipophorin precursor, stubble-2, RNA polymerase-associated protein CTR9-like protein, nucleoprotein TPR, AT-rich interactive domaincontaining protein 4B, carboxylesterase and vitellogenin. Some similar cytoskeleton proteins like actin, myosin, kinesin-like protein and a transcription factor like elongation factor were also found in our research. However, the majority of identified proteins were different from them, which may be caused by insect geographical population, saliva collection method and LC/MS-MS analysis method. For example, Chaudhary et al. [8] reported saliva of M. euphorbiae obtained in different diet (0.4% resorcinol, 15% sucrose plus amino acids or water) had different compositions. Our investigated salivary proteins will be useful in increasing the existing knowledge of watery saliva of S. furcifera and enriching the saliva composition database of S. furcifera. Some salivary proteins of N. lugens and Laodelphax striatellus that have been found in the study of Huang et al. [13] was also detected in our research, such as carbonic anhydrase, EF-hand domain containing protein, protein disulfide isomerase and mucin-like protein. It is worth mentioning that Huang et al. [20] found the mucin-like protein in N. lugens was necessary in feeding, especially reared on the resistant rice variety. This protein was confirmed to be involved in the formation of salivary sheath and suppression of resistant plant defense [20].
Since secretory proteins are most likely to be candidate effectors that can secrete into plant tissues and modulate plant defense [21,22,23], we selected 11 secretory proteins from the proteomic and transcriptomic analysis for intensive study.
Peptidylglycine α-hydroxylating monooxygenase (PHM), a copper binding oxidoreductase, was highly expressed in the head of S. furcifera and intensively expressed in all developmental stages. Zabriskie et al. [24] isolated a PHM from heads of honeybees (Apis mellifera) and found the enzyme catalyze the amidation of C-terminal peptides, which play important roles in insect reproduction, development and defense. The research about PHM was rarely conducted in invertebrates. The expression pattern about this gene was firstly reported in sap-sucking insects. It may play the same role in formatting of neuropeptides in head as in mammals, and it works in all developmental stages in S. furcifera.
Carboxylesterases (CarEs) occupy crucial roles in detoxification of xenobiotics, degradation of pheromone, neurogenesis and developmental regulations [25]. We found the CarE was highly expressed in ovary, followed by the salivary gland, suggesting its function in regulation of reproduction and detoxification of plant allelochemicals. The significantly high expression level in female after molting for 5 days also verified its important role in reproduction. CarE was previously detected in the saliva of N. lugens [4,12], while the expression pattern was a little different. CarE was expressed in all tissues, with the highest level in salivary gland and the lowest in ovary [12]. This could be because different families of carboxylesterases were detected in saliva of S. furcifera and N. lugens.
Peptidases (proteases) are the protein-hydrolysing enzymes which have been reported in many hemipterans [3]. We detected two secreted peptidases in saliva of S. furcifera. Serine protease 6, acting as digesting and detoxifying enzyme, had a specific expression in gut of S. furcifera and an extensive expression in all developmental stages except egg. Serine protease was detected in saliva of Schizaphis graminum biotype GB-E, GB-G and GB-H other than GB-C (biotype GB-H, GB-G, GB-E, GB-C ranging in virulence from high to low) [6], suggesting its function in detoxification. Serine protease 6 found in saliva of S. furcifera may play a role in preoral digestion. Neprilysins belong to zinc metalloendopeptidase and play an important role in turning off peptide signaling events at the cell surface [26]. Research on neprilysin and neprilysin-like were conducted on Locusta migratoria [27], Drosophila melanogaster [28] and Bactrocera dorsalis [29], but report about this protein in insect saliva was not found. Neprilysin-11-like isoform X4, detected in S. furcifera saliva, was highly expressed in testis and newly emerged male adults, suggesting its role in testis development and spermatogenesis. This result was in accord with expression pattern of neprilysin in B. dorsalis and D. melanogaster. Neprilysin of B. dorsalis was specifically expressed in testis [29] and different types of neprilysin in D. melanogaster were reported to be expressed in malpighian tubules and testis, suggesting roles for the peptidase in excretory function and in spermatogenesis [28]. However, the role of this protein in insect saliva requires further research.
Carbonic anhydrase (CA) and vacuolar ATP synthase were responsible for intracellular pH homeostasis [30]. Slaymaker et al. [31] claimed that tobacco chloroplast carbonic anhydrase binds salicylic acid (SA), indicating function in plant defense response. CA was previously detected in saliva of S. graminum [6], Macrosiphum euphorbiae [8] and N. cincticeps [10]. It was also found in midgut of larval Aedes aegypti [32] and malpighian tubules of D. melanogaster [33]. CA in S. furcifera had a highest expression level in gut and relatively high level in testis and salivary gland, while it was specifically expressed in salivary gland in N. lugens [12]. This may be because CA in different species owns different tissue distribution and function. Vacuolar ATP synthase exists in almost all eukaryotic cells functioning in membrane trafficking, cytosolic alkalinization and extracellular acidification [34]. It was highly expressed in gut and head in S. furcifera. Hu et al. [35] found that vacuolar ATPase plays an important role in male reproductive physiological processes in B. dorsalis. CA and vacuolar ATP synthase expression in male adult both were higher than female, indicating more important function in male. These two proteins in saliva of S. furcifera could involve in the regulation of pH homeostasis.
Protein disulfide isomerase (PDI) is essential for protein folding [36]. PDI was expressed relatively high in testis and salivary gland in S. furcifera, suggesting its role in regulation of seminal fluid and salivary protein folding. It has also been detected in saliva of Acyrthosiphon pisum [9] and N. lugens [12] previously. Carolan [9] indicated that PDIs were related to an increased yield of salivary proteins. PDI in saliva of S. furcifera may function in synthesis of salivary proteins.
Lipophorin precursor and vitellogenin (Vg), function as lipid transporter, were found in saliva of S. furcifera. Lipophorin was reported to act in lipid-based plant defense [6] and previously discovered in water saliva of S. graminum, A. pisum, M. euphorbiae and N. lugens and L. striatellus [1,4,6,8,13]. Vg is known as a nutritional source for embryonic development [37]. It has been reported in gelling saliva of N. lugens [12]. These two proteins were both expressed highly in remaining body and head. This may be because they were synthesized in fat body, which is abundant in remaining body and head. Lipophorin precursor had a valid expression in all developmental stages while Vg had a significantly high transcript level in female molting for 5 days, indicating its crucial role in reproduction. However, it is still unknown how Vg secretes into the saliva of S. furcifera and what its role in saliva.
Calcium ion binding proteins exist in saliva of many sap-sucking insects. Innate plant defense mechanism enables occlusion of sieve-tube element when suffering damage by insects [3]. Calexcitin (CE), a calcium ion binding protein, appears as a Ca 2+ -activated signaling molecular [38,39] and involves in suppressing plant defense. CE had a high transcript level in remaining body and testis, and was highly expressed in newly emerged adults, suggesting a role in regulation of mating and reproduction. CE had the lowest transcript level in salivary gland, leaving us the confusion that how it works in inhibiting of plant defense.
Mucin-like protein was related to the formation of salivary sheath [4], existing in saliva of N. lugens [4,12,13], L. striatellus [13] and N. cincticeps [10]. Huang et al. [12,20] investigated mucin-like protein had a unique expression in salivary gland of N. lugens and its role in N. lugens virulence and adaptation to host resistance. Shangguan et al. [40] revealed that mucinlike protein of N. lugens can induce plant cell death, the expression of defense-related genes and callose deposition. Expression pattern of mucin-like protein was generally the same in S. furcifera, indicating its function in feeding and interacting with host plant. Male adult had a higher expression than female, suggesting it plays a more important role in male.
In conclusion, we investigated the water saliva component in S. furcifera and tested the expression pattern of 11 secretory proteins. Many proteins were expressed relatively high in salivary gland of S. fuicifera, suggesting its role in saliva. While some proteins had a low expression level in salivary gland but were found in saliva, leaving us to find their potential function in saliva. Mucin-like protein was specifically expressed in salivary gland, which is most likely to be an effector functioning in plant defense. Intensive studies are needed to understand the function of this protein.
Supporting information S1