The α-tubulin of Laodelphax striatellus mediates the passage of rice stripe virus (RSV) and enhances horizontal transmission

Rice stripe virus (RSV, genus Tenuivirus, family Phenuiviridae) is the causal agent of rice stripe disease transmitted by the small brown planthopper (SBPH, Laodelphax striatellus) in a persistent propagative manner. The midgut and salivary glands of SBPH are the first and last barriers to the viral circulation and transmission processes, respectively; however, the precise mechanisms used by RSV to cross these organs and transmit to rice plants have not been fully elucidated. We obtained the full-length cDNA sequence of L. striatellus α-tubulin 2 (LsTUB) and found that RSV infection increased the level of LsTUB in vivo. Furthermore, LsTUB was shown to co-localize with RSV nonstructural protein 3 (NS3) in vivo and bound NS3 at positions 74–76 and 80–82 in vitro. Transient gene silencing of LsTUB expression caused a significant reduction in detectable RSV loads and viral NS3 expression levels, but had no effect on NS3 silencing suppressor activity and viral replication in insect cells. However, suppression of LsTUB attenuated viral spread in the bodies of SBPHs and decreased RSV transmission rates to rice plants. Electrical penetration graphs (EPG) showed that LsTUB knockdown by RNAi did not impact SBPH feeding; therefore, the reduction in RSV transmission rates was likely caused by a decrease in viral loads inside the planthopper. These findings suggest that LsTUB mediates the passage of RSV through midgut and salivary glands and leads to successful horizontal transmission.


Introduction
The survival of plant viruses is largely dependent on the efficient transmission to plant hosts by virus-specific vectors [1,2]. Over 70% of all known plant viruses are transmitted by insects, and approximately 55% are vectored by Hemipteran insects (e.g. leafhoppers, planthoppers, aphids and whiteflies) [3]. These insects have distinctive piercing-sucking mouthparts with needle-like stylet bundles that are comprised of two maxillary and two mandibular stylets, making insect more suitable for virus transmission [3][4][5]. Four categories of insect vectorplant virus transmission relationships have been described as follows: nonpersistent; semipersistent; persistent circulative and persistent propagative [2,6]. Plant viruses with persistent relationships enter vectors via the alimentary canal during feeding. After traversing several key tissues, these viruses eventually enter the salivary glands, which consist of principal and accessory glands. Furthermore, some even enter ovaries, which is another type of terminal insect tissue. In the principal salivary glands, viral particles are released into insect saliva when the vector is probing the host plant. In the ovaries, viruses are vertically transmitted to offspring [7]. Barriers to the persistent transmission of plant viruses in insect vectors include the following: (i) midgut infection barriers; (ii) dissemination barriers, including midgut escape and salivary gland infection barriers; (iii) salivary gland escape barriers; and (iv) transovarial transmission barriers [8,9]. A deeper understanding of the mechanistic basis of virus transmission through these four barriers will facilitate the development of novel methods to control the systemic spread of plant viruses.
Previous studies demonstrated that the persistent transmission of viruses in different insect tissues requires specialized interactions between components of the virus and vector [10]. For example, in the aphid Myzus persicae, the coat protein read-through domain (CP-RTD) of Beet western yellows virus bound Rack-1 and membrane-bound glyceraldehyde-3-phosphate dehydrogenase to facilitate transcytosis of luteoviruses in the aphid midgut and accessory salivary glands [11]. The coat proteins of Tomato leaf curl New Delhi virus and Cotton leaf curl Rajasthan virus were shown to interact with a Bemisia tabaci midgut protein to facilitate trafficking of viral particles from the midgut into the insect hemolymph [12]. Furthermore, the Rice ragged stunt virus nonstructural protein Pns10 interacted with the Nilaparvata lugens oligomycin-sensitivity conferring protein to enhance viral titer in salivary gland cells [13]. Such interactions in different insect vectors are highly complex and diverse, and their effects on the horizontal transmission of viruses remain unclear.
RSV particles initially establish infection in the midgut epithelium, then disseminate to the midgut visceral muscles, and ultimately enter the SBPH salivary glands before the virus can be transmitted to rice plants along with saliva [24]. Recently, molecular interactions between RSV and various SBPH tissues have received increased attention [25]. In the midgut, a direct interaction between NS4 and RSV CP promoted viral spread in viruliferous SBPH [24]. Furthermore, the interaction between the SBPH sugar transporter 6 (LsST6) and RSV CP was shown to be essential for RSV transfer across the midgut infection barrier [26], whereas the interaction between RSV CP and SBPH vitellogenin (LsVg) facilitated vertical transmission of the virus [27]. Further work revealed that LsVg expression was tissue-specific, and that LsVg produced in hemocytes was responsible for vertical transmission of RSV [28]. We previously reported that RSV was horizontally transmitted to rice plants via salivation during the feeding of insect vectors [29]. A series of salivary gland-specific transcriptome and proteome analyses revealed numerous genes involved in RSV transmission [30,31]; however, only cuticular protein (CPR1) and a G protein pathway suppressor 2 (GPS2) impacted RSV transmission and replication in salivary glands [31,32]. Obviously, the mechanism that RSV uses to overcome the salivary gland barrier and then undergo horizontal transmission to the plant warrants further investigation.
In the present study, proteomic analysis revealed that α-tubulin (TUB) was highly expressed in viruliferous SBPH compared to non-infected SBPH, suggesting that L. striatellus TUB (LsTUB) may have a role in mediating RSV transmission. Our data show that LsTUB facilitated the passage of RSV through the midgut and salivary gland barriers and enhanced viral transmission from SBPH to rice plants. Yeast two-hybrid and pull-down assays provided evidence that the interaction of LsTUB and the RSV NS3 likely constitutes a critical step in RSV transmission.

cDNA cloning and sequence analysis of LsTUB
Proteomic analysis of SBPH salivary glands identified 33 differentially expressed proteins in viruliferous and non-infected SBPH, including LsTUB (Table 1). Because tubulin heterodimers are known to function in viral assembly and transport [33][34][35], a full-length cDNA sequence of LsTUB (1,658 bp, GenBank accession no. KF934411) was cloned from female SBPH adults, using the conserved sequence of Tubulin α-2 from NCBI (GenBank accession no. AY550922.1) as an in silico probe. LsTUB contains a 1,353-bp open reading frame encoding a putative protein of 450 amino acids, a 93-bp 5' untranslated region (UTR), and a 212-bp 3' UTR. The translated cDNA of LsTUB yields a protein with a mass of approximately 50.0 kDA, and theoretical isoelectric point (pI) of 5.01. SMART analysis showed that LsTUB contains two conserved domains, including a GTPase domain (amino acids 49-246) with a GDPbinding site (amino acids 142-147) and a C-terminal domain (amino acids 248-393) (Fig 1A). Alignment of the LsTUB predicted protein sequence with other TUB proteins indicated a high level of identity with Hemipteran TUB proteins, including NlTUB in N. lugens (GenBank accession no. ANJ04673.1, 100% identity) and LlTUB in Lygus lineolaris (GenBank accession no. AHG54247.1, 99% identity) (Fig 1B).

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission

RSV infection in SBPH increases LsTUB expression
To further evaluate differential expression of LsTUB in viruliferous vs. non-infected SBPH, qRT-PCR and Western blot analyses were conducted to quantify its mRNA and protein expression levels, respectively. The mRNA expression levels of LsTUB were significantly upregulated in viruliferous SBPH (Fig 2A). The trend in protein expression was consistent with changes in gene expression as determined by immunoblotting with anti-LsTUB antisera ( Fig  2B, S1 Fig).

LsTUB interacts with RSV NS3 in vitro
The interaction between LsTUB and RSV NS3. LsTUB was used as bait and a RSV cDNA library as prey in a yeast-two-hybrid (Y2H) assay designed to identify RSV proteins that potentially interact with LsTUB. After screening, RSV NS3 was found to interact with LsTUB in the Y2H assay ( Fig 4A, S2 Fig). The expression of LsTUB and RSV NS3 in yeast was confirmed by Western blot analysis ( Fig 4B). To further examine the interaction between LsTUB and RSV NS3, a pull-down assay was performed with glutathione S-transferase-tagged LsTUB (GST-TUB). When the extracts from viruliferous SBPH were incubated with GST-TUB, NS3 co-immunoprecipitated with GST-TUB as shown by Western blot analysis ( Fig 4C).
The interaction between LsTUB and other viral proteins. Considering that other RSV proteins may also bind LsTUB, we evaluated whether LsTUB could interact with four RSV proteins (e.g. NSvc2, CP, SP and NSvc4) using Y2H analysis. Yeast strains containing fulllength LsTUB as bait and each of the four proteins as prey failed to grow on synthetic dextrose dropout medium (S3 Fig). This result suggests that the interaction between LsTUB and NS3 is specific.

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission NS3 mut SFL 74-76 AAA , RSV NS3 mut IWI 80-82 AAA ), the interaction between NS3 and LsTUB was abolished ( Fig 4E). These results reveal that the amino acid residues at positions 74-76 and 80-82 of NS3 are required for the interaction between LsTUB and RSV NS3. The interaction of LsTUB and RSV NS3 has no effect on silencing suppressor activity. To determine whether the interaction of LsTUB and RSV NS3 would affect the latter's silencing suppressor activity, we carried out a GFP (green fluoresce protein) silencing assay by transient ectopic expression of NS3 and its substitution mutants in Nicotiana benthamiana leaves. The results from the GFP silencing assay for the two mutants screened above showed that the substitutions of NS3 amino acids at positions 74-76 and 80-82 led to loss of their silencing suppressor ability as previously reported [36] (Fig 5A). However, leaves co-transfected with pBin438-NS3 and pBin438-LsTUB (see Methods) also exhibited high levels of GFP expression, and expression levels were similar to those in leaves transfected with pBin438-NS3 or the positive control (pBin438-P19) ( Fig 5B). The expression of NS3 and LsTUB in N. benthamiana was verified by Western blot analysis ( Fig 5C). These results clearly demonstrate that the presence of LsTUB has no effect on the silencing suppressor activity of RSV NS3 that would affect viral infection and accumulation.

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission

Silencing of LsTUB via RNAi reduces NS3 protein levels and RSV loads in vivo
To further explore the potential role of LsTUB in NS3-mediated transmission of RSV, 3 rd instar viruliferous SBPH nymphs were supplied with 0.5 mg/ml dsRNAs derived from GFP (dsGFP) or LsTUB (dsTUB) via membrane feeding. After seven days of feeding, qRT-PCR analysis showed that LsTUB mRNA in dsTUB-treated SBPHs was significantly reduced by more than 75% compared to the controls (untreated and dsGFP-treated SBPHs) (S4 Fig).
These results indicate that RNAi-mediated knockdown of LsTUB is highly effective. The midgut and salivary glands of viruliferous SBPHs were also examined by immunoblotting and confocal microscopy. Immunoblotting showed that application of dsTUB led to a decrease in LsTUB in both midgut and salivary glands, accompanied by a decrease in RSV-NS3 (Fig 6A & 6B, S5 Fig). Immunofluorescence and qRT-PCR indicated that dsTUB

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission treatment caused a substantial reduction in RSV loads in the midgut and salivary glands. Furthermore, LsTUB and NS3 were co-localized in tissues treated with dsGFP (see arrows), but not in tissues treated with dsTUB ( Fig 6C-6F"'). Taken together, these results indicate that the interaction of LsTUB and NS3 is essential for RSV accumulation in the insect vector.

Repression and overexpression of LsTUB do not affect RSV replication in SBPH cells
We examined whether the LsTUB-dependent decrease in RSV loads would affect viral replication in the insect. LsST6-expressing Spodoptera frugiperda 9 (Sf9) cells [26] were transfected with LsTUB and changes in RSV infection and replication were evaluated. At 1, 6, 12 and 24 h following RSV infection, RSV loads and NS3 protein levels were analyzed by immunofluorescence, Western blot and qRT-PCR in Sf9 cells in which LsTUB was silenced or overexpressed (Fig 7). Immunofluorescence analyses showed that NS3 co-localized with LsTUB in the Sf9 cell membrane, and stronger fluorescence signals of NS3 were detected in the cytoplasm of Sf9 cells at 24 h ( Fig 7A). However, differential expression of LsTUB did not affect NS3 signal strength in Sf9 cells (Fig 7A). Western blot analysis showed that the transfection with LsTUB

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission resulted in a significant increase of LsTUB mRNA abundance and protein levels, and treatment with dsLsTUB significantly reduced LsTUB expression compared to the controls ( Fig  7B, S6 Fig). The RNA copy number equivalents of RSV were obviously 4 to 5-fold higher at 24 h than at 12 h post-RSV infection ( Fig 7C). However, when the LsTUB was overexpressed or knocked down in Sf9 cells, the copy number equivalents of RSV RNA were not significantly reduced at 1, 6, 12 and 24 h post-infection compared to GFP RNA and dsGFP control treatments ( Fig 7C). Taken together, these results suggest that overexpression or suppression of LsTUB does not significantly affect RSV replication in virus-infected insect cells.

RNAi mediated silencing of LsTUB attenuates viral spread in vivo
After eliminating the above possibilities, we proposed that LsTUB might play a critical role in the passage of RSV through midgut and salivary glands. We investigated the distribution of Viruliferous SBPH were fed on artificial diets supplemented with dsGFP or dsTUB. SBPH midgut (A) and salivary glands (B) were analyzed for viral loads by qRT-PCR (red columns indicate dsTUB; blue columns indicate dsGFP). Levels of LsTUB, NS3 and GAPDH were analyzed by immunoblotting, and GAPDH was used as control. Each treatment was replicated three times for qRT-PCR, and values represent means ± SE, A student's t-test was used to analyze significance; �� represents P<0.01. Midguts (Fig 5C-5D"') and salivary glands (Fig 5E-5F"') of viruliferous SBPH were immunolabeled with anti-RSV NS3 (Alexa Fluor 488, green) and anti-LsTUB (Alexa Fluor 555, red) and then examined by confocal microscopy. Abbreviations: mg, midgut; sg, salivary glands; and vm, visceral muscle. Bar = 50 μm. https://doi.org/10.1371/journal.ppat.1008710.g006

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission LsTUB and RSV NS3 in different organs from 400 dsGFP/dsLsTUB-treated non-infected SBPHs at 1, 3, 5, and 7 days after acquiring the virus from the infected plants. Once RSV was ingested by SBPH, the virus established its primary infection in the midgut epithelial cells and then spread to other insect tissues (Fig 8). At 3 days post-initial access to diseased plants (padp), viral infection was initially observed in a limited number of midgut epithelial cells in 14-15% of SBPHs treated with dsGFP or dsLsTUB ( Table 2

Silencing of LsTUB via RNAi results in decreased RSV transmission efficiency
The ability of dsLsTUB-treated viruliferous SBPH to transmit RSV was evaluated. Following a two-day acquisition period on RSV-infected rice plants, each viruliferous SBPH treated with dsLsTUB or dsGFP was transferred to healthy rice seedlings, allowed to feed for two days, and then evaluated for virus infection rates by qRT-PCR (Table 3). At 15 days post-infection, 24.1% of rice plants fed on by dsLsTUB-treated SBPH contained RSV, compared to 61.3% of plants fed on by dsGFP-treated SBPH. The significance of transmission efficiency was evaluated by χ 2 analysis, and P-values indicated that the infection rate of dsLsTUB-treated SBPH was significantly lower than the dsGFP-treated viruliferous SBPH (Table 3).
Furthermore, the effect of dsLsTUB on the transmission efficiency of SBPH acquiring RSV from parents was evaluated by qRT-PCR. At 15 days post-infection, about 12.0% of rice plants fed on by dsLsTUB-treated SBPH contained RSV, compared to over 40.0% of plants fed on by dsGFP-treated or untreated (control) SBPH (Table 3), indicating that RNAi-mediated knockdown of LsTUB inhibits horizontal transmission of the virus. These results indicate that LsTUB plays a function in RSV transmission from SBPH to rice plants.

Silencing of LsTUB in viruliferous SBPH has no effect on its feeding behavior
To investigate whether the decrease in RSV transmission efficiency was caused by RNAi-mediated changes in SBPH feeding behavior, the electrical penetration graph (EPG) technique was used to monitor SBPH feeding [37]. EPG signals were classified into seven different waveforms including NP, N1, N2-a, N2-b, N3, N4, and N5, which represent the following phases: non-RSV NS3 antibody labeled with Alexa Fluor 488 (green) and then examined by confocal microscopy. Bars = 50 μm. (B) The protein expression level of LsTUB and RSV NS3 were analyzed by immunoblotting. GAPDH was used as control. (C) RSV copy number equivalents were tested by qRT-PCR in Sf9 cells. Total RNA was extracted from Sf9 cells at 1, 6, 12 and 24 h after adding virus particles to analyze the RSV loads. Values represent means ± SE; significance was evaluated by t-test analysis. https://doi.org/10.1371/journal.ppat.1008710.g007

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission penetration, penetration, stylet movement with salivary secretion, sustained salivary secretion, extracellular movement of the stylet around the phloem, phloem feeding, and xylem feeding, respectively [29]. Representative EPG waveforms were not significantly different between dsGFP-treated and dsTUB-treated SBPH (Fig 9), indicating that LsTUB knockdown does not alter the feeding behavior of SBPH. As mentioned above, plants fed on by dsLsTUB-treated

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission viruliferous SBPH showed a lower infection rate (Table 3). Thus, the decrease in RSV transmission efficiency in dsTUB-treated viruliferous SBPH can be attributed to reduced RSV accumulation and transmission in the insect.

Discussion
The intracellular tubulin family is highly-conserved [38] and consists of α, β and γ subfamilies; the α-and β-tubulin subunits are highly heterogeneous and numerous isotypes exist with different expression patterns [39]. The α-and β-tubulin heterodimers interact laterally and longitudinally to form microtubules in the cytoskeleton [40,41], and function in intracellular transport and cell division in eukaryotic organisms [42]. The tubulins and formed microtubules are also utilized by Chikungunya virus and Suid alphaherpesvirus 1 for viral internalization [43,44]; whereas Human immunodeficiency virus-1, dengue 2 virus, Vaccinia virus and John Cunningham virus use them for trafficking [45,46,47,48]. Furthermore, tubulins and microtubules are used by Human parainfluenza virus type 3 for viral replication [49], and Japanese encephalitis virus and West Nile virus use them for viral assembly [50,51]. SBPH was shown to contain two or more genes encoding both α-and β-tubulin, and additional fulllength sequence of the β-3 tubulin gene has also been identified [52]. However, no prior reports exist documenting a role for SBPH tubulin in virus transmission. In the present study, we cloned the gene encoding LsTUB (Fig 1) and showed that it was highly expressed in RSVinfected salivary glands of SBPH (Fig 2). Subsequent experiments revealed that suppression of LsTUB expression resulted in attenuated spread of RSV from midgut to salivary gland and reduced RSV transmission to rice plants (Fig 8, Table 2). These findings document a novel function for LsTUB in enabling RSV to overcome the midgut and salivary gland barriers of SBPH, leading to the dissemination of the virus to other organs in the insect vector.

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission In insect vectors, the interaction between host tubulin and viral proteins has been associated with viral transport. Recently, α-tubulin in the leafhopper Nephotettix cincticeps was shown to interact with nucleorhabdovirus matrix (M) protein from rice yellow stunt virus (RYSV), which impacted viral transport through neural tissue [53]. Our results showed that knockdown or overexpression of LsTUB did not alter RSV replication in insect cells (Fig 7). However, when LsTUB-NS3 binding was disrupted, the spread of RSV from the initiallyinfected midgut cells to salivary glands was inhibited (Fig 8). Thus, we suggest that LsTUB-NS3 binding facilitates RSV transport in SBPH cells, which is similar to the interaction of α-tubulin and the M protein from RYSV [53]. Meanwhile, the distribution analysis revealed the co-localization of LsTUB with NS3 or CP in SBPH cells (Figs 3 & 4). Although there is no direct interaction between NS3 and CP, the two proteins aggregated and formed inclusion bodies in Sf9 cells [54]. Thus, we speculate that LsTUB and NS3 associate with CP via an unknown mechanism. Taken together, our results suggest that LsTUB-generated microtubules bind to NS3 and recruit CP to facilitate virion transport through SBPH cells, thereby overcoming organ barriers; however, the actual mechanism in vivo warrants further exploration.

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission RSV requires NS3 silencing suppressor activity to counter host defenses, which facilitates systemic infection of both rice and SBPH [55,56]. A recent report demonstrated that RSV NS3 can hijack the 26S proteasome by interacting directly with the SBPH RPN3 protein [55], suggesting. RSV NS3 may have additional functions that have not yet been elucidated. Our observations using Y2H and GST pull-down assays showed that NS3 interacts with LsTUB, and the binding site of NS3 to LsTUB was required for NS3 self-interaction and silencing suppressor activity (Fig 4). Furthermore, our results confirm that, in addition to silencing suppressor activity, the binding site of NS3 to LsTUB also plays a critical role in RSV transmission (Fig 6). Thus, RSV NS3 interacts with LsTUB to enhance dissemination of the virus in the insect vector.
Multiple reports have demonstrated that circulative-propagative viruses establish infections in insect midgut [32,57], which is the initial barrier. These viruses then disseminate into other organs until they reach the final barrier, e.g. the salivary glands [24,58]. Based on previous reports, two SBPH components are known to interact with the RSV CP to help the virus overcome the midgut barrier. One interaction consists of LsST6-CP binding, which mediates viral entry into midgut epithelial cells [26]. Another is GPS2-CP binding, an interaction that activates the SBPH JNK (c-Jun N-terminal kinase) signaling pathway in the midgut, which is beneficial to viral replication [32]. With respect to the salivary gland barrier, only CPR1-or GPS2-CP binding were reported to facilitate viral infection in the salivary glands [31,32]. The spread of RSV in SBPH is obviously complex and requires multiple components. In this study, the interaction between LsTUB and NS3 facilitated RSV infection of midgut and salivary glands, which results in successful virus transmission to rice plants. Our findings thus complement and improve overall knowledge of the mechanistic basis of viral transmission in the SBPH vector; however, whether the LsTUB-NS3 interaction functions differently in the midgut vs. salivary glands requires further study.
Plant virus transmission is closely associated with the feeding behavior of insect vectors; therefore, monitoring the feeding process of dsRNA-treated SBPH can reveal the impact of dsRNA on feeding behavior and subsequent transmission to rice plants. EPG recordings have been used to investigate stylet penetration behavior in hemipteran insects [37,59]. In the present study, the RSV infection rate in dsTUB-treated SBPH was significantly lower than the control group; however, the feeding behaviors of dsTUB-treated and control SBPH were not significantly different. Thus, the decrease in RSV transmission rate was not the result of altered feeding behavior, but was instead attributed to inefficient crossing of midgut and salivary gland barriers due to dsTUB treatment.
In summary, our results indicate that LsTUB helps RSV overcome the midgut and salivary gland barriers and enhances horizontal transmission of the virus. This conclusion is supported by immunofluorescent monitoring of LsTUB and RSV in midgut and salivary glands and by Y2H and pull-down assays with LsTUB and NS3 in vitro. Silencing of LsTUB expression by RNAi also reduced NS3 levels and consequently reduced viral dissemination into midgut and salivary glands, which ultimately reduced re-inoculation into the plant. These insights provide a better understanding of the interaction between plant viruses and vectors and may ultimately reveal new avenues for therapeutic intervention.

Insects
Non-infected (RSV-free) and viruliferous strains of L. striatellus were originally collected from Jiangsu Province, China, and were maintained in the laboratory for eight years. Both noninfected and viruliferous SBPHs were reared independently on 2-3 cm seedlings of rice cv.

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission Wuyujing 3 in glass beakers containing soil at a depth of 3-5 cm. Plants were maintained in a growth incubator at 25 ± 1˚C, with 80 ± 5% RH and a 12-h light-dark photoperiod. Under these environmental conditions, SBPH development took 7 days for the egg stage, 15-18 days for the nymph stage, and 10-12 days for the adult stage. During the 30-37 day developmental period, both non-infected and viruliferous strains were transferred to fresh seedlings every 10-12 days for sufficient nutrition. To ensure that insects were viruliferous, individual female insects were allowed to feed independently. Parents and resulting offspring were collected in two-week intervals and analyzed via Dot-ELISA using monoclonal RSV CP-specific antibodies (anti-CP, 1:500) [60]. Highly viruliferous colonies were retained and used in subsequent studies.
To obtain viruliferous SBPHs, non-infected second instar nymphs were exposed for two days to RSV-infected rice plants with visible symptoms. Thereafter, RSV-positive individuals were identified via Dot-ELISA.

Virus
More than 50 viruliferous SBPHs were placed into a glass beaker containing rice seedlings (2-3 cm high), and the RSV-infected rice plants were then cultivated until symptoms appeared. The RSV-infected leaves were collected and stored at −80˚C in the laboratory until RSV purification.

Cloning of LsTUB
Total RNA was isolated from 15-20 SBPHs using TRIzol reagent according to the manufacturer's protocol (Invitrogen). The quality and concentration of total RNA were determined by spectrophotometry (NanoDrop, Thermo Scientific). The extracted RNA (500 ng) was subsequently used for reverse transcription in a 10 μl reaction with the PrimeScript RT reagent kit and gDNA Eraser as recommended by the manufacturer (Takara, Dalian, China). Based on the α-tubulin mRNA sequence downloaded from NCBI (GenBank accession no. AY550922.1), 5' and 3' RACE were conducted to obtain the full-length transcript of α-tubulin using 5'-and 3'-RACE kits (Takara). The predicted LsTUB protein sequence was subjected to Blast analysis using DNAman software (LynnonBiosoft, USA), and domains of the predicted protein were deduced using SMART (http://smart.embl-heidelberg.de/) [61].

Tissue collection
SBPH adults and nymphs were anesthetized with CO 2 for 1-3 min, and then placed into a culture dish (35 mm diameter). First, forelegs were removed from the coxa-trochanter joint using forceps (WPI, Sarasota, FL, USA), and hemolymph droplets were collected to avoid contamination by the fat body [62]. The remaining SBPHs were then dissected in prechilled and sterile PBS buffer (pH 7.4, 10 mM) to obtain midguts and salivary glands. All tissues were washed three times in sterile PBS buffer to remove other tissues, proteins, and surface contaminants.

Real-time qRT-PCR
To measure LsTUB expression levels and RSV copy number equivalents in SBPH, total RNA was isolated from 50 whole bodies, 50 midgut or 100 salivary glands of SBPH adults (female / male ratio = 1:1) using the TRIzol Total RNA Isolation Kit (Takara, Dalian, China). Total RNA concentrations were quantified, and first-strand cDNA was synthesized as described above. The primers (S2 Table) used for detecting RSV titers were designed based on RSV CPspecific nucleotide sequences. Similarly, LsTUB and LsActin (control) primers (S2 Table) were

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission designed based on LsTUB and LsActin sequences, respectively. qRT-PCR was conducted using a CFX96 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) and SYBR Premix Ex Taq (Takara, Dalian, China) as follows: denaturation for 3 min at 95˚C, followed by 40 cycles at 95˚C for 10 s, and 60˚C for 30 s. Relative expression levels for triplicate samples were calculated using the ΔΔCt method, and expression levels of target genes were normalized to the SBPH Actin gene. Three technical repeats were performed for each of the three biological replicates.

RNA interference (RNAi)
The coding sequences of LsTUB and GFP were cloned into pMD19-T vectors (Takara, Japan). The primers for dsGFP and dsTUB amplification are listed in S2 Table. Using the cDNA templates obtained above, dsRNAs were synthesized using the T7 RiboMAX Express RNAi System kit as recommended by the manufacturer (Promega, USA). A membrane feeding approach was used to introduce dsRNAs into SBPHs as described previously [57,63,64]. Briefly, second instar nymphs of viruliferous SBPH were maintained on a mixed diet containing 0.5 mg/ml dsRNAs for four days via membrane feeding and then transferred to healthy rice seedlings. To investigate the RNAi effect of dsLsTUB on viral spread in SBPH, a microinjection method was used as previously reported [26,28]. Third instar non-infected nymphs were microinjected with 23 nl dsLsTUB (2.5 μg/μl) or dsGFP (2.5 μg/μl) using an UMP3-2 UltraMicroPump (UMP3) with SYS-Micro4 Controller (WPI, FL, USA), followed by a two-day acquisition

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission period on RSV-infected rice plants. All individuals were transferred and maintained on healthy rice seedlings until evaluation immunofluorescence, qRT-PCR or Western bolt analysis. The effects of dsRNA on LsTUB expression was investigated by qRT-PCR.

Immunofluorescence microscopy
More than 30 SBPH adults were maintained on rice plants for seven days after RNAi treatment and dissected to obtain midgut and salivary glands. The dissected samples were fixed with 4% paraformaldehyde for 1 h. Samples were then blocked using 10% fetal bovine serum at ambient temperature for 2 h. Samples were incubated for 16 h at 4˚C with preimmune serum and anti-LsTUB or anti-RSV CP antibody at 1:500 dilution before incubation with the following secondary antibodies: Alexa Fluor 555-or Alexa Fluor 488-labeled secondary goat anti-rabbit IgG. Samples were then washed three times in PBS, and stained with 100 nM DAPI and CM-Dil (Sigma-Aldrich, USA) for 2 min at room temperature. Fluorescence was observed with a Leica TCS SP8 STED confocal microscope (Leica, Germany).

Yeast two-hybrid assay
Yeast two-hybrid assays were conducted using protocols supplied with the Yeastmaker Yeast Transformation System 2 (Takara-Clontech, USA). Briefly, the cDNA library of RSV was cloned as prey in plasmid vector pGADT7 using the Easy Clone cDNA library construction kit (Dualsystems Biotech), and a full-length LsTUB was cloned as bait in pGBKT7. Positive clones were selected on SD quadruple-dropout (QDO) medium (SD/-Ade/-His/-Leu/-Trp), and interacting prey constructs were recovered and sequenced. To distinguish positive from falsepositive interactions, we co-transformed empty pGADT7 and pGBKT7 into yeast strain Y2HGold. ß-galactosidase activity was detected with the HTX Kit (Dualsystems Biotech).

GST pull-down assay
LsTUB cDNA fragments were amplified and cloned into pGEX-3X as glutathione-S-transferase (GST) translational fusions. Recombinant proteins were produced in Escherichia coli strain BL21 and purified. For pull-down assays, viruliferous SBPH extract (1 mg), immobilized glutathione-Sepharose beads (200 μl) and GST-LsTUB protein (500 μg) were added to 1 ml of pulldown buffer (50 mM Tris, 150 mM NaCl, 0.1% Triton X-100, 1 mM PMSF, 1% protease inhibitor cocktail [pH 8.0]), and then incubated at 4˚C for 16 h. Similarly, insect extracts were incubated with GST protein as a negative control. Beads were washed four times with pull-down buffer, and retained proteins were released by adding 2× loading buffer and incubating for 5 min at 95˚C. Proteins were then separated by SDS-PAGE and detected using anti-GST (Cusabio, China) and anti-NS3 antibodies.

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission dsGFP, NS3 or the NS3 mutants using 1 ml syringes. GFP silencing in leaf tissue was assessed under UV light at 3-4 days post inoculation, and leaves were photographed with a Nikon digital camera (model D800).

LsST6-transfected Sf9 cell line
The S. frugiperda Sf9 insect cell line was provided by Prof. Fei Ma (Nanjing Normal University) and incubated at 27˚C in Sf-900 III SFM medium (Gibco, USA) containing 5% newborn calf serum. LsST6 was amplified based on the SBPH sequence deposited in NCBI (GenBank accession no. MG589412). The PCR product containing LsSt6 was cloned as a BamHI/XbaI fragment in pFastBacHTB (Invitrogen, USA) and introduced into Sf9 cells to generate the LsST6-Sf9 cell line as described previously [26].

Transfection and RSV infection of Sf9 cells expressing LsST6
A baculovirus expression system (Bac-to-Bac, Invitrogen, USA) was used to overexpress GFP and LsTUB in LsST6-Sf9 cells. The LsST6-Sf9 cells were transfected with 0.5 mg/ml dsGFP or dsLsTUB and Cellfectin II reagent (Invitrogen, USA) for 24 h as described previously [26]. At 48 h post-transfection, RSV particles (1.5 μg/μl) were added to the Sf9 cells as described [26], and RSV-infected Sf9 cells were collected at 1, 6, 12 and 24 h post-infection and washed five times in double-distilled water. The Sf9 cell lines were fixed and prepared for confocal microscopy. Total RNA was also extracted from these cells using TRIzol Total RNA Isolation Kit (Takara, Dalian, China) for qRT-PCR, and total protein was extracted for Western blot analysis. Each well of the plate was considered to be a single replicate, and each treatment contained three replicates.

RSV acquired from infected plants spread in SBPHs
The SBPHs can acquire and transmit RSV at almost all stages of development [15]. The latent period of RSV in SBPHs ranges from 3 to 10 days [26,66]. No significant sexual dimorphism in regard to virus acquisition efficiency was observed. However, in general, females were more efficient vectors than males in terms of RSV transmission [15,67]. More than 400 3 rd instar nymphs of SBPH were microinjected with dsGFP or dsLsTUB, followed by a two-day acquisition period on RSV-infected rice as described above. At 1, 3, 5, and 7 day post-initial access to diseased plants (padp), 100 SBPHs were collected and dissected to obtain midguts, hemocytes, and salivary glands for RSV detection using RSV NS3 antibody and the number of infected organs was recorded.

Electrical penetration graph (EPG) recording and data analysis
More than 50 nymphs from RSV-infected SBPH strains were reared for seven days on artificial liquid diets [68] supplemented with one of the following: dsGFP, dsLsTUB, or no dsRNA (control). After a 30-min starvation, the mesonotum of L. striatellus was affixed with a gold wire (20 μm diameter, 2-3 cm long) using a soluble conductive adhesive. Each L. striatellus individual was then connected to an eight-channel EPG recorder (Model: CR-8 DC-EPG I). Nymphs were placed on the culms of rice seedlings (three-leaf stage) in Ferrari insect cages; activity was recorded for 8 h in a greenhouse maintained at 25-26˚C, with 60±5% RH. Nymphs were removed at the end of the 8 h period and analyzed for RSV by Dot-ELISA. If an insect tested virus-free, the data were considered invalid. Each treatment contained 20-30 replicates, and all recorded signals were analyzed.

PLOS PATHOGENS
Insect α-tubulin mediates the passage of RSV and enhances horizontal transmission

RSV transmission efficiency
The transmission efficiency of SBPH acquiring RSV from infected plants. More than 100 SBPH individuals acquired RSV from infected plants were microinjected with dsGFP or dsLsTUB as described above. SBPH individuals were kept independently on non-infected rice seedlings for 2 days. Each individual was then removed and analyzed by dot-ELISA to confirm viral infection. Non-infected (non-viruliferous) individuals and corresponding seedlings were eliminated from the experiment. The remaining rice seedlings colonized by RSV-viruliferous insects (25-30 plants/treatment) were incubated another 10-15 days. RSV infection of rice seedlings was confirmed by qRT-PCR using RSV CP primers (S2 Table).
The transmission efficiency of SBPH acquiring RSV from parents. Following the EPG analysis above, rice seedlings fed on by RSV-infected SBPH were also incubated an additional 10-15 days to ensure virus propagation. RSV infection of rice seedlings used in EPG analysis was determined by qRT-PCR as described above.
Evaluation of RSV transmission rates. The number of infected rice seedlings was recorded and calculated to determine transmission rates; rice seedlings colonized by dsGFPtreated or untreated SBPH were considered controls. Each seedling was considered to be one replicate, and each treatment had 25-30 replicates. Transmission rates were calculated as follows: transmission rate (%) = (number of infected seedlings/total number of seedlings) × 100. A χ 2 test was performed with SPSS statistical software [69] to detect differences between treatments.