Proteomic Analyses of Whitefly-Begomovirus Interactions Reveal the Inhibitory Role of Tumorous Imaginal Discs in Viral Retention

In nature, plant viruses are mostly transmitted by hemipteran insects, such as aphids, leafhoppers, and whiteflies. However, the molecular mechanisms underlying the interactions between virus and insect vector are poorly known. Here, we investigate the proteomic interactions between tomato yellow leaf curl virus (TYLCV, genus Begomovirus, family Geminiviridae), a plant virus, and its vector whitefly (Bemisia tabaci) species complex. First, using a yeast two-hybrid system, we identified 15 candidate whitefly proteins interacting with the coat protein of TYLCV. GO and KEGG pathway analysis implicated that these 15 whitefly proteins are of different biological functions/processes mainly including metabolic process, cell motility, signal transduction, and response to stimulus. We then found that the whitefly protein tumorous imaginal discs (Tid), one of the 15 whitefly proteins identified, had a stable interaction with TYLCV CP in vitro, and the DnaJ_C domain of Tid301−499aa may be the viral binding site. During viral retention, the expression of whitefly protein Tid was observed to increase at the protein level, and feeding whiteflies with dsRNA or antibody against Tid resulted in a higher quantity of TYLCV in the whitefly body, suggesting the role of Tid in antiviral infection. Our data indicate that the induction of Tid following viral acquisition is likely a whitefly immune response to TYLCV infection.


INTRODUCTION
Many plant viruses, such as species of the Luteoviridae, Geminiviridae, and Nanoviridae families, are transmitted by hemipteran insects in a persistent, circulative manner (1). During the long-term virus-vector interactions, insect vectors have developed two inevitable physical barriers to virus movement: midgut and salivary glands (1,2). Initially, the vector ingests virions from virus-infected plants; then, virions enter the insect midgut lumen and subsequently cross through the midgut epithelial cells to be released into the hemolymph. Afterwards, virions move along with the hemolymph and reach the salivary glands from which they are injected into plants together with whitefly saliva secretion (1,2). During this circulative journey, viruses need to interact with the insect vector in a coordinated manner for successful transmission to occur; at the same time, viral infection may activate immune reactions from its vector (3,4).
Begomoviruses (genus Begomovirus, family Geminiviridae) are a group of single-stranded circular DNA viruses, which are transmitted by whiteflies of the Bemisia tabaci species complex in a circulative manner (5,6). Some begomoviruses are serious viral disease agents of many crops worldwide. For example, tomato yellow leaf curl virus (TYLCV) is transmitted by a notorious invading species of whitefly, provisionally named as Middle East-Asia Minor 1 (MEAM1), of the B. tabaci species complex and has caused enormous damage to the production of tomato and some other crops in many countries/regions in the last three decades (7)(8)(9). Similar to other begomoviruses and other circulatively transmitted viruses, ingested TYLCV moves along the path of stylet-midgut-hemolymph-salivary glands in whitefly vectors. During the movement, TYLCV depends on clathrin-mediated endocytosis to enter the midgut cells and then accumulates in intracellular vesicle-like structures (10)(11)(12). At the same time, a viral infection activates the whitefly autophagy pathway, which plays an important role in the antiviral response (13,14).
Up to now, the coat protein (CP) is the only structural protein of begomoviruses known to be involved in viral movement in the vector (15). CP gene replacement results in dramatic changes in characteristics of viral acquisition and transmission by whitefly vector (16)(17)(18)(19). However, so far, only a few whitefly proteins have been reported to interact with the viral CP. The heat shock protein 70 (HSP70) and vesicle-associated membrane protein-associated protein B (VAPB) show inhibitory roles in virus transmission (20,21), and a peptidoglycan recognition protein BtPGRP acts in whitefly immunity (22). In contrast, GroEL produced by secondary endosymbionts Hamiltonella or Arsenophonus may protect the virus from degradation in vector hemolymph (23,24), and the midgut protein, cyclophilin B and collagen protein may assist in viral transmission (25)(26)(27). Vitellogenin may enable transovarial transmission of virus to the next generation of whitefly (28). The putative roles of BtHSP16, thioredoxin-like protein (TLP) and protein BtR242 produced by Rickettsia in the viral transmission are yet unclear (29)(30)(31). Despite this progress, the functions of some of the abovementioned proteins require further validation, and many more vector components remain to be discovered to achieve an adequate understanding of begomovirus-whitefly interactions.
In this study, first, using the yeast two hybrid (Y2H) system, we identified 15 candidate whitefly proteins interacting with TYLCV CP, including the evolutionarily highly conserved protein tumorous imaginal discs (Tid). As the mammalian homolog of whitefly Tid has been implicated for its role in a variety of signaling pathways and autophagy (32,33), we then conducted a series of molecular experiments and bioassays to examine in vitro interaction between whitefly Tid and TYLCV CP. Following viral infection, increase of whitefly Tid at the protein level exerted constraints on viral retention. Our data provide novel insights into begomoviruswhitefly interactions, indicating the negative impact of Tid on viral retention.

Y2H Assay System
The Y2H assay based on the matchmaker gold yeast two-hybrid system (Cat. No. 630489; Clontech) was used to explore the interactions between whitefly proteins and TYLCV CP. The cDNA library of whitefly was constructed in the prey plasmid of SfiI-digested pGADT7. The full-length of TYLCV CP gene was cloned into the bait plasmid of pGBKT7 after Nde I and EcoR I restriction. Primers used for cloning are listed in Supplementary Table 1. We used the following procedure for the Y2H assay: (1) transform the recombinant plasmid pGBKT7-TYLCV CP into the Y2H Gold yeast strain; (2) select the yeast strain on synthetic defined minimal medium lacking tryptophan (S.D./-Trp); (3) extract the yeast protein by yeast total protein extraction kit (Cat. No.C500013; Sangon Biotech) and confirm the expression of TYLCV CP in yeast in a Western blot by anti-TYLCV CP antibody (provided by Professor Jian-Xiang Wu); (4) conduct the auto-activation detection; (5) transform the cDNA library of whitefly into the Y2HGold yeast strain containing the bait plasmid pGBKT7-TYLCV CP; (6) observe the growth of yeast strain on the double dropout medium (DDO: S.D./-Leu/-Trp) and triple dropout medium (TDO: S.D./-His/-Leu/-Trp) with 40 µg/ml X-alpha-Gal and 125 ng/ml aureobasidin A (AbA) (TDO/X/A), select the positive clones on TDO/X/A; (7) restreak these positive clones on quadruple dropout medium (QDO: S.D./Ade/-His/-Leu/-Trp) with 40 µg/ml X-alpha-Gal and 125 ng/ml AbA (QDO/X/A) to eliminate the false positives; (8) recover the prey plasmids from the positive clones and transform them into Escherichia coli strain DH5α, sequence, and identify their interactions with TYLCV CP again. The different fragments screened from the whitefly cDNA library were used in a BLAST search of the NCBI database (http://blast.st-va. ncbi.nlm.nih.gov/Blast.cgi), and the sequences of these fragments screening in the Y2H assay were deposited in GenBank.

Bioinformatic Analysis
Whitefly proteins identified from the Y2H assay system were categorized according to their gene ontology (GO) annotation using the Blast2GO software and then performed using the OmicShare tools, a free online platform for data analysis (http://www.omicshare.com/tools). The metabolic pathway analysis of these proteins was conducted according to the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway annotation (https://www.kegg.jp/blastkoala/). Network diagrams were created using the database search tool for the retrieval of interacting genes/proteins (STRING 9.1; http://stringdb.org). All of these analyses were conducted by the full length of amino acid sequences.

Real-Time PCR
Quantitative (q) PCR was performed on CFX connect real-time PCR system (Bio-Rad, USA) using the SYBR Premix Ex Taq II (Cat. No. RR820A, Takara). β-Actin was used as an internal reference, and relative abundance of TYLCV or transcript was calculated by 2 − Ct . Primers used for real-time PCR are listed in Supplementary Table 1. dsRNA Synthesis DNA templates with a T7 promoter at both ends of selected genes were used to synthesize dsRNA following the manufacturer's instruction of the T7 high-yield transcription kit (Cat. No.TR101-02; Vazyme). Then, dsRNA was purified using phenol: chloroform extraction, isopropanol precipitation, and resuspended in nuclease-free water. The size and quality of the dsRNA were confirmed by 1% agarose gel electrophoresis, and its quantity was measured using Nanodrop (Thermo Scientific, USA). DsGFP was used as a control. Primers used for DNA template synthesis are listed in Supplementary Table 1.

Membrane Feeding on dsRNA or Antibody
Whitefly adults within 7 days post-emergence were collected from cotton plants. A group of 250 adults were released into a glass tube 1.5 cm in diameter and 10 cm in length. According to Pan et al. (10), for dsRNA silencing, whiteflies were fed on 15% sucrose solution containing 200 ng/µl dsRNA for 48 h, and 15% sucrose solution with the same amount dsGFP was used as control. For antibody feeding, Tid polyclonal antibody (PcAb) was mixed with 15% sucrose solution with a dilution rate of 1:50 for 24 h, and 15% sucrose solution with the same dilution of rabbit pre-immune serum was set as control.

Viral Acquisition
For viral acquisition after dsTid or dsUBR7 feeding (dsGFP was used as control), whiteflies were caged with leaves from the same branch of TYLCV-infected tomato plants for 6, 12, or 24 h, respectively, and then transferred to feed on cotton for 48 h for viral retention. Female adults were collected in groups of 10 each and homogenized in 100 µL lysis buffer for relative viral quantity analysis (10). Three biological replicates were conducted for relative viral quantity analysis by real-time PCR. For the subsequent experiments of membrane feeding of dsRNA or antibody against Tid, whiteflies were caged with leaves of two symmetrical leaves of the same height on the same branch of TYLCV-infected tomato plants for 12 h and then transferred to feed on cotton for 48 h for viral retention. Three to five biological replicates were conducted for relative viral quantity analysis by real-time PCR.

Full-Length Amplification, Protein Expression, and Antibody Production
The ORF of Tid-FL (GenBank: MT505750) was amplified from the whitefly cDNA using PrimerSTAR max DNA polymerase (Cat. No. R045A; Takara) and then cloned into pET28a plasmid for fusion with His tag. His-Tid-FL was expressed in inclusion bodies of E. coli strain Rosetta, and following renaturation and purification of inclusion body protein, His-Tid-FL was used to immunize rabbits to obtain a Tid-specific PcAb by HuaBio  Thermo Scientific) to determine and unify the concentration of protein samples. Western blot analysis was conducted by anti-Tid antibody, using anti-β-actin antibody (Cat. No. E021020-01; Earthox) as a control. The translational-level analysis was repeated three times, and ImageJ was used to quantify the relative protein level, Following dsTid membrane feeding, 12 h viral acquisition, and 48 h viral retention, the expressions of Tid at transcriptional and translational levels were analyzed as described above.

Statistical Analysis
Comparison of the relative abundance of virus in whitefly and expression levels of genes were performed using an independent t-test with P < 0.05 as the threshold of significant difference ( * P < 0.05, * * P < 0.01, and * * * P < 0.001). All the statistical analyses were performed using SPSS 20.0 (SPSS Inc., USA).

Analysis of the Interactions Between Whitefly Proteins and TYLCV CP
As shown in Supplementary Figure 2A, the Y2H system was used to examine the interactions between whitefly proteins and TYLCV CP. The titer of the primary whitefly cDNA library was ∼5.0 × 10 6 cfu with an average insert size of 1 kb, meeting the requirements of a standard cDNA library (Supplementary Figure 2B). The fusion expression of TYLCV CP with GAL4 DNA-BD in the yeast (≈46 kDa) was verified using Western blot analysis (Supplementary Figure 2C). The auto-activation detection showed that the bait plasmid pGBKT7-TYLCV CP could be used in this Y2H system (Supplementary Figure 2D). After Y2H screening, 26 positive clones were isolated, and 15 unique whitefly proteins were identified ( Table 1). To identify the one-to-one interaction between bait and prey protein, the interactions between these 15 screened whitefly proteins and TYLCV CP were validated using the Y2H assay (Figure 1A), combined with reported interactions between MEAM1 whitefly and TYLCV, and a protein interaction network was generated, including the predicted interactions among whitefly proteins ( Figure 1B).

In silico Analysis of the Whitefly Proteins Screened by Y2H Assay
According to GO and KEGG analyses (Figures 2, 3), the 15 interactors from the Y2H assay (Table 1) were classified into different groups, mainly including metabolic process, cell motility, signal transduction, and response to stimulus. The GO analysis suggests that the 15 proteins may be responsible for 17 different biological processes, mainly involved in cellular and metabolic processes with different distributions inside and Frontiers in Immunology | www.frontiersin.org outside of cells; most of them shared the binding activity, and about half of them possess catalytic activity (Figure 2). The KEGG pathway analysis suggests that the 15 proteins can be classified into 7 groups (Figure 3). For example, gelsolin-like isoform X2 belongs to the pathway of cell motility; protein phosphatase 1B as a member of the MAPK signaling pathway belongs to the group of signal transduction. The whitefly autophagy pathway and ubiquitin-proteasome system have been shown to play a role in antiviral response (13,14,35). Among the 15 whitefly proteins, there is a ubiquitin-protein ligase (UBR7) and a protein Tid related to macro-autophagy (33). Both of these two proteins belong to the biological process of response to stimulus (GO: 0050896).

Effects of dsRNA Interference of Tid and UBR7 on Viral Retention
To examine the role of proteins Tid and UBR7 on virus retention, whiteflies that had received dsRNA interference treatment were transferred to feed on a TYLCV-infected tomato for 6, 12, or 24 h, respectively, and then transferred to feed on cotton for  48 h. At the end of each of the three time points, after dsTid interference, the relative viral quantity in whiteflies significantly increased compared to the control (Figures 4A-C) although, for UBR7 dsRNA interference, following a viral acquisition access period for 12 h, the defense ability of the whitefly against TYLCV retention significantly decreased (Figure 4B). When the intervals of the viral acquisition access period lasted 6 or 24 h, the defense ability of the whitefly against TYLCV retention had non-significant decrease (Figures 4A,C). Based on these results, we selected Tid for the following experiments.

Structural and Phylogenetic Analysis of the Protein Tid
After sequencing the Tid prey plasmid screened from Y2H, we obtained an 855 bp long (285 aa) Tid-S sequence (GenBank: MT505751), having a 60% coverage (164-448aa) of Tid-FL. Tid-FL (GenBank: MT505750, ≈52 kDa) has a DnaJ domain (N-terminal, 76-138aa), a DnaJ_C domain (C-terminal, 301-419aa), and four repeats of a CXXCXGX(G) motif (239-299aa). Tid-S contains only the CXXCXGX(G) motifs and DnaJ_C domain (Figure 5A). Tid-FL has no transmembrane domain or signal peptide, and its 3-D structure model is shown in Figure 5B. Phylogenetic analysis of B. tabaci Tid and 16 other insect Tid proteins showed that B. tabaci Tid forms a monophyletic lineage with species of Hymenoptera and appears closely related to the genus Drosophila ( Figure 5C). This DnaJ domain-containing protein is evolutionarily highly conserved; Tid in mammals and that of Drosophila show 54.9% identity in amino acid sequences (36), and the Tid of whitefly and that of Drosophila melanogaster show 54.0% identity in amino acid sequences.

In vitro Evidence Supports the Interaction Between Tid and TYLCV CP
TYLCV CP fused with GST and Tid-S tagged with MBP were used to verify their interaction through GST pull-down analysis ( Figure 6A). Using the fusion protein GST-TYLCV CP as a bait protein and native whitefly proteins extracted by cytoplasmic extraction buffer (Cat. No.SC-003; Invent) as prey proteins, whitefly endogenous Tid could co-elute with GST-fused TYLCV CP but not with GST ( Figure 6B). Further, we tested the interaction between TYLCV CP and different regions of Tid-FL mentioned above: Tid 76−138aa (DnaJ domain), Tid 239−299aa (four repeats of a CXXCXGX(G) motif), and Tid 301−419aa (DnaJ_C domain). The results showed that Tid 76−138aa and Tid 239−299aa show no binding activity with TYLCV CP (Figure 6C); the binding site of TYLCV CP may be located in the C terminal of Tid-FL ( Figure 6D).

The Increase of Tid at Protein Level During Viral Retention
Following viral infection, the expression of Tid at both transcriptional and translational levels was tested. Data demonstrates that there was no significant change of the expression of Tid at transcriptional level ( Figure 7A). However, Western blot analysis showed TYLCV infection could significantly increase the expression of Tid at protein level ( Figure 7B).

Effects of Tid Interference on TYLCV Retention
Following dsRNA feeding, the adults were transferred to feed on TYLCV-infected tomato plants for 12 h for virus acquisition and then were transferred to feed on cotton for 48 h for observation on virus retention. Data showed that the expression of Tid in whiteflies was effectively knocked down via dsRNA interference (Figures 8A,B), and knockdown of Tid expression resulted in significant increases of relative virus quantity in whiteflies ( Figure 8C). In addition, blocking Tid function by anti-Tid antibody likewise resulted in significantly higher relative virus quantity in whiteflies during virus retention ( Figure 8D).

DISCUSSION
Investigation of the interactions between begomoviruses and whitefly proteins can provide new knowledge of the virus transportation journey in vector. In this study, 15 candidate whitefly proteins of various categories were detected that may interact with TYLCV CP. In further tests of Tid and UBR7, two of the 15 candidate proteins detected showed that both proteins posed an adverse effect on viral retention, and Tid had a stronger effect than UBR7. A stable interaction between whitefly Tid and TYLCV CP was then observed, and the C-terminal of Tid was observed to be the likely binding site. Viral infection could increase the expression of whitefly Tid at the protein level; feeding whiteflies with dsRNA or antibody against Tid resulted in a significantly higher quantity of TYLCV in the body of whiteflies following viral acquisition. Altogether, these data reveal one novel whitefly protein that may function in antiviral response. The insect innate immune system incurs physical, cellular, and humoral responses to invaders (37), and it is common for insect vectors to take advantage of their immunity to fight against viral infection. Wang et al. (22) demonstrate that whitefly protein BtPGRP with antibacterial activity acts in multiple immuneresponse functions. Wang et al. (38) show that insect vectors could operate the c-Jun N-terminal kinase (JNK) signaling pathway for controlling viral transmission, causing a significant reduction in virus accumulation and transmission. The studies of Luan et al. (13) and Wang et al. (14) indicate that autophagy is involved in whitefly repression of begomovirus infection and triggers complex interactions between virus and insect vector. A previous study reported a mammalian homolog of whitefly Tid, which acted as a key regulator in mediating autophagy independently of HSP70 (33). Data available to date indicate that both Tid and HSP 70 play a role in repressing virus infection [(20); this study]; however, the relationships among whitefly autophagy, Tid, HSP70, and TYLCV CP remain unclear. Molecular mechanisms underlying the activation of autophagy pathway by TYLCV-infection in whiteflies warrant further investigation. Our findings provide clues for future studies on these issues.
Additionally, the roles of other candidate proteins detected in this study are also worth exploring. Gelsolin is a key regulator of actin filament assembly and disassembly (39), and actin has been shown to interact with several viral proteins and plays important roles in viral transmission. For example, interactions between non-structural protein Pns10 of rice dwarf virus and the cytoplasmic actin of leafhoppers is correlated with insect vector specificity (40); the non-structural protein P7-1 of reovirus southern rice black-streaked dwarf virus generates tubules and this tubules associate with the actin cytoskeleton in insect vector (Sogatella furcifera) cells (41,42). In addition, MAPK signaling pathway is known to be activated by a diverse group of viruses and has important roles in viral replication (43), such as supporting assembly and maturation of West The expression of Tid following viral acquisition at protein level was analyzed by Western blot analysis; 100 whitefly adults were collected as one sample for protein extraction and BCA protein assay was used to determine and unify the concentration of protein samples. Three biological replicates were set, and the results were quantified by ImageJ, t = −3.077, P = 0.0370. Independent t-test was used here and the differences between treatments were considered significant when *P < 0.05.
Nile virus and dengue virus (44,45), regulating multiple steps of influenza A virus replication (46) and so on. In view of the potential role of protein phosphatase in regulating the life cycle of Simian Virus 40 (47), a study of the relationship of protein phosphatase 1B (a member of the MAPK signaling pathway) with TYLCV infection may be worthwhile. These investigations may lead to a comprehensive recognition of whitefly binding partners of viral CP and better understanding of the complex interactions between begomoviruses and their whitefly vectors.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
JZ and TG designed this study and conducted most experiments as well as data analysis. JZ drafted and revised the manuscript. TL did the qPCR analysis and participated in statistical analysis. J-CZ made bioinformatic analysis. FW participated in Y2H screening. X-WW participated in manuscript preparation. S-SL provided supervision for the study and participated in manuscript preparation and revision. All authors read and approved the final version of the manuscript.

FUNDING
This study was financially supported by the National Natural Science Foundation of China (Project Nos. 31930092 and 31925033).

ACKNOWLEDGMENTS
We thank Professor Jianxiang Wu, Institute of Biotechnology, Zhejiang University, for providing monoclonal antibodies of TYLCV CP. We also thank our lab engineer Gen-Hong Yan for his effort in plant growing and maintenance and Yuenan Zhou and Wenqiang Xia of our institute for their valuable discussions during this study.