Host Plants Indirectly Influence Plant Virus Transmission by Altering Gut Cysteine Protease Activity of Aphid Vectors*

The green peach aphid, Myzus persicae, is a vector of the Potato leafroll virus (PLRV, Luteoviridae), transmitted exclusively by aphids in a circulative manner. PLRV transmission efficiency was significantly reduced when a clonal lineage of M. persicae was reared on turnip as compared with the weed physalis, and this was a transient effect caused by a host-switch response. A trend of higher PLRV titer in physalis-reared aphids as compared with turnip-reared aphids was observed at 24 h and 72 h after virus acquisition. The major difference in the proteomes of these aphids was the up-regulation of predicted lysosomal enzymes, in particular the cysteine protease cathepsin B (cathB), in aphids reared on turnip. The aphid midgut is the site of PLRV acquisition, and cathB and PLRV localization were starkly different in midguts of the aphids reared on the two host plants. In viruliferous aphids that were reared on turnip, there was near complete colocalization of cathB and PLRV at the cell membranes, which was not observed in physalis-reared aphids. Chemical inhibition of cathB restored the ability of aphids reared on turnip to transmit PLRV in a dose-dependent manner, showing that the increased activity of cathB and other cysteine proteases at the cell membrane indirectly decreased virus transmission by aphids. Understanding how the host plant influences virus transmission by aphids is critical for growers to manage the spread of virus among field crops.

Aphids are small insects that feed exclusively on the phloem sap of plants, and cause significant damage to agronomic crops. However, their major economic importance is that they are the most numerous vectors of plant viruses, such as the poleroviruses in the Luteoviridae, which we will refer to collectively as luteovirids in this manuscript. Luteovirids are single stranded, positive sense, nonenveloped RNA viruses that infect a range of economically important crops and weedy hosts. Luteovirids, including Potato leafroll virus (PLRV) 1 , cause severe yield losses in agronomic crops around the world and are transmitted exclusively by aphids in a circulative manner. Circulative transmission requires a series of spatially and temporally regulated, largely unknown protein interactions with the virus structural capsid proteins (1,2). There is no cure for viral infection in plants, therefore, the only options are to prevent or avoid infection (3). Host resistance is the ideal method to prevent infection, but despite intensive efforts to identify or breed for resistance, few commercialized luteovirid-resistant cultivars have been released. Controlling aphid vectors using pesticides is costly, and to be effective, information about vector phenology is necessary. Disrupting an aphid's ability to transmit a virus into or within a crop represents a different approach and a promising means by which to control virus spread (3,4).
Aphids acquire and transmit luteovirids as intact virions, not viral RNA, and there is no evidence to show that luteovirids replicate in their aphid vectors (3,5). Luteovirids are nonspecifically ingested from the phloem sap together with sap proteins (6) while the aphid is feeding on an infected plant. To be transmitted to a new plant, luteovirids must overcome physical barriers within the insect, the gut and the accessory salivary glands, a process that is mediated by virus-vector species-specific protein interactions. The virus must first be internalized by gut cells (6 -12). Detailed microscopic investigations revealed that the virus moves via endosomes in the gut, with different virus species displaying different affinities to various regions of the gut (i.e. midgut or hindgut). PLRV is acquired into midgut epithelial cells (13). Virions bind to the luminal (apical) plasma membrane, stimulating the formation of coated pits and enter the gut epithelial cells via a receptormediated endocytosis mechanism (14). Aphid membrane alanyl aminopeptidase N (APN) has been identified as a cell surface receptor for Pea enation mosaic virus (PEMV, genus Enamovirus) in the pea aphid, Acyrthosiphon pisum (2,15). Once inside the aphid cell, the virus particles remain in membrane-bound vesicles, during transport through the cytoplasm, and this is universally true for every species of luteovirid studied by microscopy, to date. Unlike when the virus is in plant cells (16), in aphid cells, virions are never observed free in the cytoplasm. The observation that virus-containing tubular vesicles connect to aphid cellular organelles supports hypothesis that the virus is transported intracellularly through the gut endomembrane system. Membrane-bound vesicles containing virions in gut cells of Myzus persicae and other aphid species have been observed to connect to lysosomes and lysosomal-like organelles (10,13). Following transport through the endosome, PLRV and other luteovirids can be observed between the plasmalemma and the basal lamina of the gut epithelia where they are then released into the open circulatory system of the aphid and quickly diffuse (10,13). Once the virus reaches the accessory salivary glands, the virus is endocytosed (17,18), transported through the cells within vesicles, and released into the salivary duct where it can be inoculated into plants together with the saliva as the insects feed. Aphids that acquire luteovirids from an infected plant remain viruliferous for their entire life (5).
Luteovirids promote their own plant-to-plant spread by influencing plant host selection and feeding behavior of the insect vector (19,20) as well as affecting the production of winged, migratory individuals (21,22). Aphids are more attracted to plants infected with circulative viruses that they transmit than to healthy plants or to plants infected with viruses that have other modes of transmission. Aphids tend to remain on plants infected with circulative viruses longer than plants infected with cuticle-associated viruses. These findings suggest that different transmission modes shape the extent to which viruses influence their vectors (22)(23)(24)(25)(26)(27)(28). Positive or neutral effects on vector performance have been extensively reported for persistently transmitted viruses that are dependent on their insect vectors for transmission (20, 29 -36). On the other hand, negative and sometimes neutral effects on insects have been reported mainly for plants infected with viruses and other pathogens that are not transmitted or not exclusively transmitted by the insect species studied (24,(37)(38)(39)(40)(41). Collectively, these studies show that viruses have been favored by natural selection to alter vector behavior via controlling vector interactions with their host plants.
Here, we observed that M. persicae changed its vectoring ability in a host-dependent manner, and we investigated how the aphid's host plant impacts virus transmission at the molecular level. A clonal lineage of M. persicae reared on a PLRV host plant (physalis) and PLRV nonhost plant (turnip) showed significant variation in PLRV transmission efficiency. Using organismal, biochemical, molecular, proteomic, and imaging approaches, we show that high levels of cysteine proteases at the cell membrane in aphids reared on turnip are indirectly responsible for the host-dependent change in the virus transmission phenotype in M. persicae.

EXPERIMENTAL PROCEDURES
Insects-Parthenogenic reproducing colonies of the same clonal lineage (genetically identical individuals) of M. persicae Sultz (the green peach aphid) were maintained on caged physalis (Physalis floridana) or turnip (Brassica rapa) at 20°C with an 18-hour photoperiod for a minimum of four months prior to the experiments and proteomics analyses.
PLRV Transmission Assays-To test the host switch effect on PLRV transmission, aphids from both colonies, turnip (T-Myzus) and physalis (P-Myzus), were transferred to PLRV-infected hairy nightshade (Solanum sarrachoides, HNS) detached leaves for an acquisition access period (AAP) of 24 h, which is enough time for virus acquisition by aphids ( Fig. 1) (42,43). After collecting the data for this initial experiment, another experiment was performed with an AAP of 48 h to test whether a longer AAP would increase the virus transmission rate by T-Myzus. HNS seedlings were Agrobacterium-infiltrated with the PLRV infectious clone as described (44) and used as the source of virus. After that, 10 aphids were transferred to each healthy potato cv. Red Maria seedling (n ϭ 10 plants) for an inoculation access period (IAP) of 48 h, as described (42,43). Potato seedlings were treated with imidacloprid to eliminate aphids after the IAP. After three weeks, systemic infection of PLRV was detected in the recipient plants by double-antibody sandwich enzyme-linked immunosorbent assay (DAS-ELISA) using ␣-PLRV antibodies (Agdia, Elkhart, IN).
To test whether the impact of turnip feeding on PLRV transmission was transient, T-Myzus were transferred to physalis plants for three days and then transferred to PLRV-infected HNS for a 48 h AAP. Similarly, P-Myzus were transferred to turnip plants for three days and then fed to PLRV-infected HNS detached leaves for the same 48 h AAP. Five aphids were transferred to healthy potato seedlings for the transmission assay for a 48 h IAP, in 10 replicates per treatment. Three weeks later, the systemic PLRV infection in potato plants was detected by DAS-ELISA using ␣-PLRV antibodies. The proportion of plants infected with PLRV was compared with noninfected plants using the Chi square test.
Aphid Reproduction on Different Host Plants-We measured the effect of each host plant on aphid reproduction and weight. Fourth instar nymphs of P-Myzus were transferred to either a turnip or a physalis plant, and after 24 h, nymphs that had molted to adults were transferred to a fresh turnip or physalis plant in four biological replicates to measure the host effect on aphid reproduction. Fifteen days later, the progeny was counted. The individual weight of adults was obtained by averaging the weight of 15-30 adults per replicate. Progeny counts and adult weights were analyzed by one-way ANOVA, and the means were compared using the Student's t test.
PLRV Acquisition by Aphids-P-and T-Myzus were fed on PLRVinfected HNS plants for a 24 h AAP. Aphids were then transferred to an artificial diet (45) sandwiched between thinly-stretched parafilm membranes for gut clearing. Aphid cohorts were collected at 24 h or 72 h after the start of the AAP for PLRV quantification by digital droplet PCR (ddPCR), using the QX100 droplet digital PCR system (Bio-Rad, Hercules, CA). The ddPCR reaction for PLRV consisted of 10 l of 2X ddPCR Evagreen SuperMix (Bio-Rad), 1 l of combined primers at 10 M each (PLRV CP qPCR F1 5Ј-CTAACAGAGTTCAGC-CAGTGG-3Ј; PLRV CP qPCR R1 5Ј-TGTCCTTTGTAAACACGAAT-GTC-3Ј), 7 l of dH2O and 2 l of DNA diluted at 1:800 in a final volume of a 20 l reaction. The entire 20 l reaction was then loaded into a disposable DG8 droplet generator cartridge secured in the cartridge holder (Bio-Rad). A total of 70 l of droplet generator oil for Evagreen (Bio-Rad) was also loaded into the disposable DG8 droplet generator cartridge. The cartridge holder was placed into the QX100 droplet generator (Bio-Rad) where droplets were generated. Droplets were then transferred to a 96-well plate (Eppendorf, Hamburg, Germany) and the plate was sealed with an easy pierce foil seal (Bio-Rad). PCR amplification was carried out on the Applied Biosystems 2720 Thermocycler. The thermocycling conditions started at 95°C for 5 min, followed by 40 cycles of 95°C for 30 s and 60°C for 1 min, 1 cycle at 4°C for 5 min, 1 cycle at 90°C for 5 min and ending at 12°C. Following amplification, the plate was inserted into the droplet reader cassette and loaded into the droplet reader (Bio-Rad). The droplets were automatically read at a rate of 8 wells per 15 min. The ddPCR droplet data were analyzed using the QuantaSoft analysis software Version 1.7.4 2014 (Bio-Rad), which presents the target results as copies per l of PCR mixture. The default settings to analyze the data were as follows: experiment set to ABS; SuperMix: QX200 ddPCR EvaGreen Super; Channel 1: FAM; and Channel 2: Vic/Hex. The digital droplet reader determined the number of copies per l using the Poisson Distribution calculator. The number of copies obtained in two independent experiments were averaged and analyzed by One-Way ANOVA.
2-D DIGE-We used gel-based separation and quantification of intact proteins, 2D-DIGE, to measure the relative quantification of protein expression between P-Myzus and T-Myzus. Three biological replicates of T-and P-Myzus (all life stages) were weighed and frozen at Ϫ80°C in 50 ml BD-Falcon (Franklin Lakes, NJ) for protein extraction. Care was taken to remove all plant and soil debris from the aphids before freezing, so as not to contaminate the aphid protein samples. Proteins were extracted using a TCA-acetone protein precipitation protocol optimized for 2-D gel electrophoresis of aphid proteins (46). Protein samples were labeled with Cy3 or Cy5 according to the manufacturer's instructions (GE Healthcare; Piscataway, NJ). A Cy2 internal standard containing an equal amount of proteins from all the biological replicates was used for relative quantification by DIGE technology. Cy-dye labeled samples were grouped randomly during 2-D gel electrophoresis so that each gel contained a Cy3 and a Cy5 labeled sample, together with the Cy2-labeled, pooled internal standard. A dye swap was performed to control for any labeling bias. Analytical gels containing Cy dye-labeled samples were used for quantitative analysis and preparative gels containing nonlabeled samples were used for spot picking. A total of 50 g of each Cy dyelabeled sample or 500 g of nonlabeled protein were loaded onto immobilized pH gradient (IPG) strips (pH 4 to 7, 18 cm; GE Healthcare) during an overnight passive rehydration of the strips according to the manufacturer's specifications for the analytical and preparative gels, respectively. The first dimension was run on the IPGphor II (GE Healthcare) at 20°C with the following settings: step 1: step and hold for 500 V, 1 h; step 2: gradient 1,000 V, 2 h; step 3: gradient 8000 V, 3 h, and step 4: step and hold 8000V until 34,000 V for a total focusing time of 10 h. Next, the IPG strips were reduced for 15 min with 64.8 mM of dithiothreitol in SDS equilibration buffer (50 mM Tris-HCl [pH 8.8], 6 M urea, 30% glycerol, 2% SDS, 0.002% bromphenol blue) and then alkylated for 15 min with 135.2 mM of iodoacetamide in SDS equilibration buffer. The second dimension was carried out using 8 -15% gradient tris-glycine gels (Jule, Inc, Milford, CT). Gels were cast 1 mm thick by 25.5 cm wide by 20.2 cm tall with an acrylamide: bis ratio of 38:1. The Ettan DALT Six system (GE Healthcare) was used to run the second dimension at 25°C with the following settings: step 1, 10 mA/gel, 1 h; step 2, 40 mA/gel, 6 h or until the bromphenol blue front ran to the bottom of the gels. The preparative gels were fixed in a solution of 15% methanol, and 7.5% acetic acid for one h, stained overnight in Colloidal Coomassie Blue (Invitrogen, Carlsbad, CA) and destained in water for 12 h prior to spot picking.
Gel Analysis-Gels were scanned on the Typhoon Variable Mode Imager (GE Healthcare) at 100 dpi according to the manufacturer's specifications for Cy dyes and Colloidal Coomassie Blue (Invitrogen) stained gels were visualized with the 632.8 nm helium-neon laser with no emission filter. DIGE gel images were analyzed using Progenesis Samespots v. 3.1 (Nonlinear Dynamics; Newcastle Upon Tyne, United Kingdom). Fifty manual alignment seeds were added per gel (ϳ12 per quadrant) and the gels were then auto-aligned and grouped according to host plant for analysis. Spots were selected as being differentially expressed if they showed greater than a 2.0-fold change in spot density and an ANOVA p value of Ͻ0.05. Mass Spectrometry-Gel plugs were picked from the preparative gels and proteins were prepared for mass spectrometry as described previously (46). Peptides were analyzed using a Q Exactive (Thermo Fisher Scientific, Waltham, MA) mass spectrometer. For analysis on the Q Exactive, 5 l of the in gel digest was loaded onto a 5 cm C-18 PepMap trapping column with an EasyNanoLC 1000 (Thermo Fisher Scientific) system. The peptides were loaded at a maximum pressure of 280 bar and washed with 3 l of 0.1% formic acid (FA) at the same pressure settings. The peptides were separated by a 20-cm hand packed and pulled C-18 PepMap column where the voltage for ionization was applied at a liquid junction prior to the column. A 30-min gradient was applied, beginning at 4% acetonitrile (ACN), 0.1% FA and ramping linearly to 35% ACN, 0.1% FA. The remaining time was used for a high organic and re-equilibration of the trapping and analytical columns. Nanospray ionization was achieved with a Nano-Flex ion source (Thermo Fisher Scientific) operating at 1.9 kV and an ion transfer tube temperature of 275°C. MS/MS analysis was performed using software version 2.2 SP1. Profile MS1 spectra were obtained from an m/z of 400 to 1600 at 140 k resolution with an automatic gain control (AGC) target of 3.0e 6 charges and a maximum fill time of 50 ms. The 20 most intense ions from each MS1 scan with charge states of 2-7 were selected for fragmentation with preference given to those with an isotopic distribution matching that of average using the peptide match option "preferred." Selected ions were isolated in a 1.6 Da window centered on the most intense isotope with an AGC target of 1e 5 and a maximum fill time of 50 ms. Fragmentation was normalized to a collision energy of 27 for each selected ion. Targeted ions were then placed on the dynamic exclusion list for 15 s to allow multiple fragmentations of each ion. All raw files were converted to Mascot generic format (mgf) using Proteome Discoverer v. 1.3. For searching the 2-D DIGE data, a custom database was built containing M. persicae proteins sequences available from Aphidbase and available Buchnera sequences containing a total of 2736 proteins. MGF files were searched using Mascot v. 2.5 with oxidation of methionine and deamidation of glutamine and asparagine as variable modifications using trypsin as the enzyme and three missed cleavages allowed. No fixed modifications were used for the gel spot searches. Mass error tolerances were set at 20ppm for the precursor ion and 0.02 Da for the product ions. An expect score of 0.05 or less was used as a threshold cut-off for peptide identification. As is typical for gel spot analysis, no false discovery rate is provided as the data set for each gel spot is too small for FDR estimation (46). Differential expression of cathepsin B (cathB) in aphids was verified using relative peptide quantification by selected reaction monitoring mass spectrometry (SRM). Nano-flow liquid chromatography was performed using an Easy nLC 100 (Thermo Fisher Scientific) in a vented-tee configuration. . Emitter tips (New Objective, Woburn, MA) were trimmed to 4 cm. Two l of the 1 g/ml digested aphid protein extracts were loaded onto the trapping column and eluted with a flow-rate of 300 nL/mn. The gradient ramped from 5% B (95:5 acetonitrile/formic acid) to 37% B across 110 min, and then increased to 80% B and held constant for 5 min. Electrospray ionization (ESI) was initiated using a CorConnex plug and play nanoLC-ESI interface applying 1.4 kV via a liquid junction distally from the ESI tip. The capillary voltage and temperature were 42 V and 275°C, respectively. Selected Reaction Monitoring (SRM) analysis was performed using a TSQ Vantage (Thermo Fisher Scientific) operating in SRM mode. For SRM-mass spectrometry, the doubly charged precursor ions were monitored in Q1 with a resolution of 0.7 full width at half-maximum (FWHM) and singly charged y3 to n-1 ions for each peptide were monitored in Q3 at 0.7 FWHM. Each transition was monitored for 20 ms (dwell time) enabling a maximum duty cycle of 2.0 s. A digest of bovine serum albumin was analyzed every fifth run for signal intensity, retention time reproducibility, and peak width and shape to verify chromatography and instrument performance.
Targeted protein sequences for cathB were imported into Skyline (47) and converted into trypsin fragments. Methods were refined as described (6) with the following exception: all CID fragment y-ions (y3-yn-1) were monitored in all replicates. Three biological and three analytical replicates were analyzed and a Student's t test was used to compare total peak areas.
The same P-and T-Myzus samples analyzed using 2-D DIGE were also subjected to a 1-D separation and analysis using LC-MS/MS. For each sample, 1 g of total tryptic peptides were loaded onto a 5 cm C-18 PepMap trapping column with an EasyNanoLC 1000 (Thermo Fisher Scientific) system. The peptides were loaded with at a maximum pressure of 280 bar and were washed with 3 l of 0.1% formic acid at the same pressure settings. Following sample cleanup, the peptides were separated by a 20-cm hand-packed and pulled C-18 PepMap column where the voltage for ionization was applied at a liquid junction prior to the column. The gradient was a total of 230 min, beginning at 4% acetonitrile, 0.1% formic acid and ramping linearly to 35% acetonitirile, 0.1% formic acid in 185 min. The remaining time was used for a high organic and re-equilibration of the trapping and analytical columns. Nanospray ionization was achieved with a NanoFlex ion source (Thermo Fisher Scientific) operating at 1.9 kV and an ion transfer tube temperature of 275C.
MS/MS analysis was performed on a Q Exactive (Thermo Fisher Scientific) using software version 2.2 SP1. Profile MS1 spectra were obtained from an m/z of 400 to 1600 at 140k resolution with an AGC target of 3.0e 6 charges and a maximum fill time of 50ms. The 20 most intense ions from each MS1 scan with charge states of 2-7 were selected for fragmentation with preference given to those with an isotopic distribution matching that of average using the peptide match option "preferred." Selected ions were isolated in a 1.6 Da window centered on the most intense isotope with an AGC target of 1e 5 and a maximum fill time of 50 ms. Fragmentation was normalized to a collision energy of 27 for each selected ion. Targeted ions were then placed on the dynamic exclusion list for 15 s to allow multiple fragmentations of each ion. Protein identification was done as described above for gel spots except that a different database was used. A custom database was constructed for the label-free analysis and included a total of 35,482 sequences, including the proteins from the former database as well as the predicted protein sequences from the pea aphid genome available on NCBI and common contaminant proteins. Additional changes for this analysis included one missed tryptic cleavage was allowed and methylthio was selected as a fixed modification (digests were treated with methyl methanethiosulfonate as described in (48). Spectral counting was performed using Scaffold (Proteome Software, Portland, OR) and differentially expressed proteins were selected using a p value Ͻ0.05 and the Fisher's Exact test. Data are available via ProteomeXchange with identifier PXD004893.
CathB Quantification in Aphids at the Transcript Level-The levels of CathB transcripts were compared between P-Myzus and T-Myzus by qRT-PCR. Ten adult aphids were collected from both colonies and flash frozen for RNA isolation. Quantitative reverse transcribed PCR (qRT-PCR) was used to measure cathB expression at the transcript level in T-and P-Myzus. Pools of ten whole aphids were milled to a fine powder in a cryogrinder (Retsch, Haan, Germany) and total RNA was extracted by using the RNeasy mini kit (Qiagen), followed by cDNA synthesis using 1 g of total RNA and the SuperScript III First Strand Synthesis kit (Invitrogen) with oligo dT primers. qRT-PCR reactions were performed using 2 ng of cDNA and 10 l of SYBR Green PCR Master Mix (Applied Biosystems). Gene specific primers were used to amplify CathB (5Ј-ACAAGCGACTACATGGAAGG-3Ј and 5Ј-CCCAACACGATCCACAATTTC-3Ј) and B-actin (5Ј-TCGTCT-TGGATTCTGGTGATG-3Ј and 5Ј-GCAAGATCGAGACGAAGGATAG-3Ј) M. persicae genes. Ct values of Cathepsin were normalized to the ct values of the reference gene B-actin. Three biological replicates and three analytical replicates were performed for each gene.
Structural Modeling-Structural modeling was performed using the Phyre2 Protein Fold Recognition Server in "normal" model (49). Structural models were visualized using Molsoft.
Coimmunolocalization of CathB and PLRV in M. persicae Gut-Immunostaining assays were used to localize CathB in the aphid digestive system and to test the hypothesis that this protein colocalizes with PLRV in the aphid gut. The assays were performed with guts dissected from P-and T-Myzus fed on a PLRV-infected HNS detached leaves for an AAP of 48 h, after which aphid guts were dissected and prepared as described (50). Briefly, guts dissected in 1ϫ PBS were fixed in 4% paraformaldehyde and then permeated by adding 0.1% Triton X-100. After washing three times with 1x PBST (PBSϩ0.5% Tween 20), tissues were blocked for 1h in blocking buffer (1ϫ PBST containing 1% bovine serum albumin) and incubated overnight at 4°C with ␣-CathB human monoclonal antibody raised in rabbit (ABCAM ab125067). Midguts were washed three times with PBST and incubated for 1 h at room temperature with donkey antirabbit IgG secondary antibody conjugated to Cy3 (Jackson Immunoresearch 711-165-152, West Grove, PA). For the colocalization, tissues were blocked again for 1 h and then incubated with ␣-PLRV coat protein antibody at 4°C overnight. Tissues were then washed three times with PBST and incubated for 1 h at room temperature with donkey ␣-rabbit IgG secondary antibody conjugated to Cy2 (Jackson Immunoresearch 711-225-152). Gut tissues were washed again for three times with PBST, mounted in 1 x PBS buffer containing DAPI, covered and sealed for analysis under a confocal microscope. For each treatment, three guts were analyzed. Two different controls were performed: (1) healthy controls: aphids fed only on healthy plants and prepared with CathB and PLRV antibodies; (2) guts of aphids fed on PLRV-infected HNS leaves were prepared using only secondary antibodies (no primary antibodies).
Immunocapture RT-PCR-The potential interaction between the PLRV coat protein and M. persicae CathB was investigated by Immunocapture PCR as described (51) with the addition of the reverse transcription step. Briefly, the wells of a microtitre plate were coated in a coating buffer (0.05 M sodium carbonate pH 9.5) for 4 h at room temperature with 4 g of the following antibodies: ␣-CathB human monoclonal antibody raised in rabbit (ABCAM ab125067, Cambridge, United Kingdom), ␣-CathB polyclonal antibody raised in rabbit (ABCAM ab92955), and as a positive control, ␣-PLRV coat protein (Agdia). Negative controls were performed with ␣-IgG antibody and with no antibody. P-and T-Myzus aphids previously fed on PLRVinfected HNS detached leaves for 48 h were homogenized in one volume of extraction buffer (50 mM Tris-HCl (pH 8.0), 5 mM EDTA, 2% polyvinylpyrolidone and 0.05% Tween-20) and centrifuged at 13,000 ϫ g for 15 min at 4°C. The plate was washed with 1 ϫ PBS three times to remove unbound antibody, and 200 l aliquots of aphid extract were added into the wells in three replicates. The plate was incubated overnight at 4°C to allow maximum capturing of the particles. The overnight-incubated plate was washed with 1ϫ PBS to remove all unbound material. The bound particles were released by adding 30 l/well of extraction buffer and the suspension was stored at 4°C until further use. A total of 5 l of the extracted particles was used to synthesize cDNA by using Improm II reverse transcriptase, in a 15 l reverse transcription reaction. PCR amplification of the viral cDNA from the virions bound to the antibodies used in the capture was performed with 3 l of cDNA using specific primers to amplify a 660 bp fragment of the PLRV coat protein (44).

CathB Activity Assays and Inhibition of Cysteine Proteases
Using E-64 -The activity of cathB in P-Myzus and in T-Myzus was compared in a fluorescence assay, by using the cathB activity assay kit from ABCAM. A pool of adult aphids with the same weight (10 mg) was collected from both colonies, in three biological replicates. Aphids were ground in 200 l of the lysis buffer provided in the kit in a cryogrinder with metal beads and incubated in ice for 10 min, followed by 5 min of centrifugation at 13,000 rpm and 4°C. Fifty l of the aphid lysate was mixed with 50 l of the CB reaction buffer provided in the kit in each well of a 96-well plate. Two l of the substrate Ac-RR-AFC was added to each sample/well and mixed by pippeting. The cysteine protease inhibitor E-64 (10 M) was added to the negative controls. Samples were incubated at 37°C for 90 min and then read in a spectrofluorometer with 400 nm excitation and 505 nm emission filter. Relative Fluorescence Units (RFU) were compared between the two treatments.
In the second experiment, concentrated samples were prepared by grinding 35 mg of whole aphids in 200 l of lysis buffer, as described above. Aphid lysate was then diluted to 10, 100, and 1000-fold to perform a sample dilution curve and then mixed with the CB reaction buffer provided in the kit (1:1) and the substrate Ac-RR-AFC, following incubation at 37°C and absorbance reading at a spectrofluorometer, as described above. A substrate dilution curve was also performed by diluting the substrate to 10, 100, and 1000 fold and keeping the sample concentration constant.
Cathepsin Inhibition and PLRV Transmission Assays-Adult aphids were placed in dishes covered by a Parafilm® sandwich membrane containing an aphid diet prepared with balanced amounts of amino acids (45) for 48 h. The adults were then removed, and first instar nymphs were transferred to a fresh diet containing the cysteine protease inhibitor E-64 (AG Scientific) at 0, 10, 30, or 50 M, for a 48h AAP. Nymphs were transferred to PLRV-infected HNS detached leaves for a 24 h AAP, after which ten nymphs per plant were then transferred to healthy potato cv. Red Maria seedlings for an IAP of 48 h, with 20 replicates per treatment. Plants were treated with imidacloprid to kill the insects. After 3 weeks, PLRV systemic infection was confirmed by DAS-ELISA. These experiments were repeated three times. The proportion of plants infected with PLRV was compared with noninfected plants using the Chi square statistic test.
ICP Analysis of Nutrients in Aphids-Inductively Coupled Plasma Atomic Emission Spectrometer (ICP-AES) was used to quantify the amounts of macro and micronutrients in whole aphids (P-versus T-Myzus), using three biological replicates per treatment. Aphid tissue was dried at 60°C for 4 days and then macerated using a mortar and pestle. Samples were weighed and then incubated overnight on 2.5 ml of 60:40 HNO3 and HClO4 mixture into a Pyrex glass tube to degrade organic matter. The mixture was then heated to 120°C for 2 h and 0.25 ml of 40 ppm Yttrium added as an internal standard to compensate any drift during run in the ICP-AES. The temperature of the heating block was then raised to 190°C for 10 min and turned off. Samples were then cooled down and diluted to 20 ml, vortexed and transferred into auto sampler tubes to run in the ICP-AES Thermo iCAP 6500. The amount of each nutrient was converted from ppm to mg/g of sample based on the original weight of each sample. The normalized amount of each nutrient was compared between treatments by one-way ANOVA, and the means were compared using the Student's t test.

Turnip-reared Aphids are Less Efficient in Transmitting PLRV as Compared with Physalis-reared Aphids-When
aphids were given an AAP of 24 h, a significant difference in the proportion of plants infected with PLRV was observed ( Fig. 2A). P-Myzus infected 70% of the plants, whereas T-Myzus infected only 20% of the plants ( Fig. 2A, p ϭ 0.035). When aphids were given a longer AAP of 48 h, the transmission efficiency increased for both aphid lineages, but it was still significantly higher for P-Myzus (p ϭ 0.025). P-Myzus infected 80% of the plants with PLRV, whereas T-Myzus infected 33% of the plants with PLRV ( Fig. 2A).
The Host Switch Effect is Transient-To test whether the host switch effect was transient, we transferred aphids from the physalis colony to turnip plants (PT-Myzus) and aphids from the turnip colony to physalis plants (TP-Myzus) for a three-day feeding period prior to allowing a 48 h AAP on PLRV-infected HNS. The transmission efficiency was significantly different between the two groups of aphids (p ϭ 0.0235). The host switch caused the transmission phenotypes to be reversed. Aphids that were reared on physalis and then transferred to turnip for 3 days showed a dramatic decrease in their transmission efficiency, infecting 27.3% of the plants (Fig.  2B). Surprisingly, aphids from the turnip colony that were transferred to physalis for 3 days recovered their ability to efficiently transmit PLRV, infecting 72.7% of the plants (Fig. 2B). (Fig. 3). A 10-fold increase in the progeny number of T-Myzus as compared with P-Myzus ruled out the possibility that the difference in transmission was because of an inability to feed on or colonize turnip (Fig. 3A). Additionally, T-Myzus adults were double the mass of P-Myzus adults (Fig. 3B).

The Difference in Transmission is Not Because of An Inability to Feed or Colonize Turnip-T-Myzus reproduced better compared with P-Myzus
P-Myzus Do Not Acquire More Virus Than T-Myzus-Virus acquisition by aphids is a specific process that requires crossing physical barriers in the aphid gut and salivary gland cells (52)(53)(54), a process which is mediated by receptor molecules in the cell membranes (2). Even in compatible aphidvirus species pairs, aphids do not acquire all viruses they ingest. Some virus particles might not cross the barriers presented by the aphid cells, others may be degraded. We tested whether the difference in PLRV transmission could be because of a difference in the aphids' ability to acquire the virus when they are reared on different host plants by comparing PLRV genomic RNA levels in T-and P-Myzus after gut clearing on artificial diets. Viral genomic RNA is unstable in the

Molecular & Cellular Proteomics 16 Supplement 4 S235
aphid gut (44,55) and so the gut clearing step removes ingested virus that is not acquired and enables us to obtain an accurate measure of the virus that is acquired by the aphid. At both time points of 24 h and 72 h after virus acquisition there is a trend of higher number of PLRV copies in P-Myzus than in T-Myzus, although not significant (Fig. 4).

2-D Difference Gel Electrophoresis Reveals that cathB is the Major Proteome Change Between T-and P-Myzus-
The proteomes of T-and P-Myzus, were analyzed by quantitative 2-D DIfference Gel Electrophoresis (DIGE). A total of 1324 spots were visualized using DIGE on the Cy2 reference gel and matched across all gels in the experiment (Fig. 5). Because Pand T-Myzus have the same genetic background, it was expected that their proteome profile would have only a few differences in protein expression. Accordingly, of the 1324 total spots, only 15 spots were differentially expressed (Fig. 5) by at least a 2.5-fold change (p value Ͻ 0.005) using a oneway ANOVA, all of them up-regulated in T-Myzus and all located in one region of the 2-D gels (Fig. 5, spots in red). A Cy3-Cy5 dye swap on biological replicates (Fig. 5, inset) showed the pattern of protein spot up-regulation in the T-Myzus samples was highly reproducible and not an artifact of dye-labeling bias. The spot pattern is indicative of a protein with multiple isoforms varying by charge, with isoelectric points ranging from pH 4.2-5.5 (Fig. 5) and molecular mass from 45 to 35 kDa (Fig. 5). The proteins in these spots were identified as isoforms of the aphid cysteine protease, cathB, with at least eight matching peptides, using nLC-MS/MS analysis (Table I). Peptides from lower abundant proteins (more than 20-fold lower) were identified in these spots, but this is common for gel spot analysis by LC-MS/MS. Other spots appearing as green or red on the image DIGE provided in Fig.  5 did not represent statistically reproducible differences between T-and P-Myzus in the 2-D DIGE experiment. These spots may represent proteins with high biological variability requiring more replicates to show statistically different fold-changes, or analytical variation from 2-D gel, DIGE and/or protein extraction artifacts.
CathB in M. persicae-The draft genome of M. persicae is available from Aphidbase (http://www.aphidbase.com/ node_94263/Myzus-DB/Downloads). Three cathB genes were identified from the available sequencing data (supplementary Fig. S1). CathB transcripts were 11-fold higher in T-Myzus compared with P-Myzus (supplementary Fig. S2), indicating that the difference in expression of this protein between T-and P-Myzus results from differences in transcript abundance.
Two genes mined from the sequencing data were full length, contigs 1 and 2, with isoelectric points (pIs) of 5.3 and

FIG. 5. 2-D Difference In Gel Electrophoresis shows minimal proteome level differences between Myzus persicae reared on turnip (T-Myzus) and physalis (P-Myzus).
In this image, red spots are specific to T-Myzus, green spots specific to P-Myzus. Yellow spots are proteins expressed equally in both aphid colonies. Red spot train was identified as cathepsin B using mass spectrometry. Inset shows cathB spots in a dye-swap experiment. 5.2 and molecular mass at 37 kDa, which corresponded to the protein spots identified on the 2-D gels. CathB contig 3 was truncated and predicted to be 14 kDa in size with an isoelectric point at 4.2. All cathB peptides isolated from the 2-D gels were either shared among all three contigs or between cathB contigs 1 and 3. To test whether cathB contig 2 was also up-regulated in T-Myzus, we used selected reaction monitoring mass spectrometry (SRM) targeting peptides that were shared and specific to cathepsin B contigs 1 and 2 ( Fig. 6A and  6B, respectively). Peak areas from all peptides measured using SRM were higher in abundance in T-Myzus (Fig. 6), confirming that cathB protein expressed from both contigs 1 and 2 were up-regulated in T-Myzus. The structural topology of the three catalytic residues of M. persicae cathepsin B were conserved with human cathepsin B (supplementary Fig. S3), further verifying our annotation of these contigs as M. persicae cathB.

Other Enzymes with Predicted Lysosomal Function Are Upregulated in T-Myzus-To expand on the 2-D DIGE analysis,
we performed quantitative label-free proteomics comparing tryptic digests of P and T-Myzus. This analysis revealed that other enzymes with predicted lysosomal functions were upregulated in T-Myzus, compared with P-Myzus, such as ␤-glucuronidase, peroxidasin, legumain-like, and aminopeptidase-N (Fig. 7). CathB and cathB-16 were also identified as up-regulated in T-Myzus in the label-free experiment (Fig. 7). CathB-16 was the major proteome level change between Pand T-Myzus in this experiment, with a 29-fold higher expression in T-Myzus, suggesting that aphid feeding on turnip plants induces a higher expression of other enzymes in the aphid gut, including additional cysteine proteases.
CathB Colocalized with PLRV In Gut Brush Border Membranes of T-Myzus But Not P-Myzus-In guts dissected from P-Myzus, the majority of PLRV and cathB were in distinct subcellular compartments, with cathB localized to a particular organelle along the midgut epithelial cytoskeleton and the virus was diffused throughout the cell (Fig. 8A-8C). In contrast, nearly complete colocalization of cathB and PLRV was observed in T-Myzus guts on the cell membranes (Fig. 8D-8F). Although the assay is not quantitative, we observed that the amount of cathB in P-Myzus was lower than in T-Myzus, consistent with the proteome data, whereas the abundance of PLRV was clearly greater inside of P-Myzus gut epithelial cells, consistent with the quantitative PCR data. Although cathB and PLRV colocalize to the cell membranes in T-Myzus, no direct interaction was detected using immunocapture-RT-PCR with the cathB antibodies and PLRV-specific primers (not shown). The localization of the virus is markedly different in P and T-Myzus and colocalization with aphid cathepsin B only occurs in T-Myzus, predominately on the brush border membranes of the midgut lumen. Controls to demonstrate the specificity of each antibody with nonviruliferous aphids, antibody controls and controls for the emission and excitation of each cyanine dye in the double labeling experiments are shown in supplementary Fig. S4 -S6, respectively.
CathB Activity In Aphids-The increased level cathB in T-Myzus, suggests that there may be a higher activity level in these aphids as compared with P-Myzus. To test this, we performed a cathB activity assay. The activity of cathepsin B was significantly higher in T-Myzus than in P-Myzus (Fig. 9). The cysteine protease inhibitor E-64 significantly reduced the activity of cathepsin B in both P-and T-Myzus (Fig. 9), enabling us to develop a functional assay using E-64 to test the effect of inhibiting cathB on virus transmission.
Cathepsin Inhibition in T-Myzus Recovered the Efficient Vectoring Phenotype-Feeding aphids on a cysteine protease inhibitor, E-64, resulted in a contrasting phenotype (Fig. 10)  Two peptides abbreviated as DQG and DYY, are shared between the two cathepsin contigs. The four remaining peptides are unique to each cathB contig. The full tryptic peptide sequences corresponding to the three amino acid abbreviations can be found in supplementary Table S2.

FIG. 8. Coimmunolocalization of CathB and PLRV in P-Myzus (A-C) and T-Myzus (D-F) alimentary canals.
C is a single focal plane and at higher magnification of the inset in B, and F is a higher magnification of the inset in E. Blue in all panels is DAPI staining of the nuclei. Red is immunostaining of PLRV using specific primary antibody and secondary antibody conjugated to Cy2. Green is immunostaining of cathB using specific primary antibody and secondary antibody conjugated to Cy3. Colocalization of PLRV and cathB appears in yellow. ag: anterior midgut (stomach); pg: posterior midgut; hg: hindgut; e: embryo. Note that localization of PLRV and PLRV-cathB colocalization appears in some or all portions of the posterior midgut only.  (Fig. 10A). In contrast, PLRV transmission by P-Myzus significantly decreased following treatment with E-64 (Fig. 10B, p ϭ 0.0004), although the level did not differ among the concentrations of E-64 (Fig. 10B). No aphid mortality was observed in the inhibitor treatments compared with the controls (data not shown). The contrasting phenotypes observed in these E-64 inhibitor studies suggests that when cathB is highly expressed and on the cell periphery, it negatively regulates virus acquisition and when cathB is contained in the lysosomes, its activity is required but not sufficient for proper virus acquisition into the gut.
Varying Enzyme and Substrate Concentration Showed Evidence for a Cathepsin Inhibitor in P-Myzus-The initial rate of a catalyzed reaction is directly proportional to the enzyme concentration over a wide range (56). The presence of inhibitors, that are, molecules combining with the enzyme or the substrate, can easily be detected by the failure of proportionality in experiments measuring enzymatic activity as a function of substrate or enzyme concentration. To test for the presence of an aphid endogenous cathB inhibitor in P or T-Myzus, we performed fluorescence cathB activity assays in T-and P-Myzus by varying enzyme and substrate concentrations (Fig. 11). Holding all other conditions constant, the initial rate of reaction increased faster with a rise of substrate concentration in T-compared with P-Myzus (Fig. 11A), suggesting the presence of an inhibitor in P-Myzus. Similarly, controlling for weight and varying the aphid lysate concentration showed that the P-Myzus sample lost proportionality (Fig.  11B), also providing evidence for a cathB inhibitor in the P-Myzus sample. It is not known whether the inhibitor is derived from the plant or the aphid.
There Are Higher Levels of Ca 2ϩ in T-Myzus-Based on the fact that T-Myzus are larger and more fecund than P-Myzus conspecifics, the proteome data that showed a higher level of lysosomal enzymes in T-Myzus, the colocalization of PLRV with cathB in the membranes of T-Myzus midgut epithelial cells, and our ability to increase the virus transmission efficiency of T-Myzus with the cysteine protease inhibitor E-64, we hypothesized that there is an increase in Ca 2ϩ -mediated lysosomal functions in T-Myzus and this directly or indirectly reduces acquisition and transmission of PLRV by T-Myzus. We measured the amounts of Ca and other nutrients in P-and T-Myzus. Accordingly, the amount of Ca is higher in T-Myzus than in P-Myzus (Fig. 12). DISCUSSION The ability of M. persicae to transmit PLRV when reared on turnip was impaired, although T-Myzus presented a fitness advantage, producing more progeny and larger adults than P-Myzus. A significant growth reduction was reported by Rahbé and collaborators (57) when they fed M. persicae on transgenic oilseed rape plants (Brassica napus) expressing the cysteine protease inhibitor oryzacystatin, which also suggests that cathepsin activity in the aphid is required for proper growth and low levels of cathepsin impairs aphid growth. These observations are consistent with our studies showing that T-Myzus reproduced better and were larger than P-Myzus, which correlated with the higher levels of cathB in T-Myzus. Collectively, the data show that the reduction in virus transmission efficiency in T-Myzus is not because of an impairment of aphid growth when reared on turnip plants.
Two proteomics studies have examined the impact of the host on the aphid vector (58,59). To our knowledge, our study is the first to look at the impact of a host switch on vectoring ability, i.e. the efficiency at which an insect vector transmits a virus, at the molecular level and revealed cathB expression may regulate virus transmission by aphids. Two experimental approaches were available for testing the role of cathB in virus transmission, RNA interference (RNAi), which has been shown to be functional in aphids (60 -64), or the use of a chemical cathB inhibitor, E-64, a fungal-derived secondary metabolite that has specificity for thiol-proteases, including cathB but not cathepsin A or D (65). Because the cath gene family is highly expanded in the aphid (66), both approaches have a drawback of some degree of lack of specificity. CathB silencing constructs will have off target effects and E-64 will inhibit other thiol proteases in the aphid gut. In our hands, silencing of cathB in the aphid using RNA interference was not amenable for conducting virus transmission experiments using different host plants because of the length of time required for transmission experiments (data not shown). Because the inhibition of cathB is irreversible by E-64, it was our method of choice, and we probed the function of cathB in transmission using E-64. In addition, we sought to use a functional assay that interfered with enzyme function, which E-64 provided, not just reduce the level of the enzyme by lowering transcript levels. Although RNAi is certainly a powerful approach for functional genomics and given the paucity of tools for functional genomics in aphids and other hemipterans, the vast array of fungal secondary metabolites with specificity for different proteins makes this approach a highly attractive alternative to RNAi for functional genomics in these insects. The proteome-guided, cell biological analyses in our study show that feeding on turnip influences the expression of several predicted lysosomal enzymes in the aphid. It is possible and likely that other proteins, including other cysteine proteases, play a regulatory role, directly or indirectly, in virus acquisition. The functions of these additional proteins in virus transmission should be further investigated.
The cathB gene family is massively expanded in aphids, which is thought to be an adaptation for feeding on phloem sap (66). This raises an interesting question about the potential role of cathepsins in regulating the ability of an aphid to transmit viruses. Cathepsins are papain cysteine proteases under physiologic conditions (67) and are widely expressed in the aphid gut (68,69), including in the anterior aphid midgut (70). They function as both endopeptidases and as peptiddyldipeptases in the lysosome (67), cleaving substrate proteins as they enter into the organelle. It is notable that "putative cathepsin B-S" was the fifth most abundant transcript in the gut of the soybean aphid and "cathepsin B-16A" the tenth most abundant (71), indicating that cathB clearly has important functions in the aphid gut. In our study, cathB-16 was also identified as up-regulated in T-Myzus by the label-free proteomics. Other cysteine proteases might have been inhibited by E-64 in the inhibition assays, as E-64 is a cysteine protease inhibitor, not specific to cathB. Therefore, a role for the gut luminal cysteine proteases such as cathL cannot be ruled out.
CathB and PLRV do not directly interact in immunocapture experiments and argues for an indirect association between presence of Cathepsin B on the gut surface, it's up-regulation, and reduced virus transmission. Aphids have multiple cathepsin genes that are tightly regulated in a tissue specific manner (66,68,69,71). CathL and cathB are the major cathepsins in the aphid gut and cathB is one of the four most important detoxification enzymes (70,(72)(73)(74). CathB is directly involved in the hydrolysis of toxic proteins in the diet (79). Brassicaceae plants (including turnip) contain protease inhibitors that are overcome by up-regulation of cathepsins in the gut (76). What is observed in our experiments is up-regulation of and increase in activity of gut localized cathB in response to the host plant and an associated decrease in the transmission of PLRV.
A number of different hypotheses are possible to explain these observations. The first of these is that the effect is indirect resulting from the presence of turnip proteins. Turnips contain lectins that bind to glycosylated proteins on the surface of the gut epithelium in herbivores (77). Such binding may sterically interfere with association of PLRV with its receptor. This scenario is unlikely because our data show that by inhibiting cathepsin activity using E-64, virus transmission efficiency can be restored. A second, and more plausible alternative is that the increase of cathepsin B in T-Myzus is degrading turnip phloem proteins that facilitate virus entry. Previous work has shown evidence for plant proteins associating with purified luteovirids (6) and that phloem proteins may assist in virus transmission (78). E-64 may block cysteine protease mediated degradation of turnip proteins that facilitate virus entry in T-Myzus, rendering them poorer vectors than their P-Myzus conspecifics.
Other scenarios are also possible. Inhibition of cathepsins in the gut of P-Myzus could result in up-regulation of other proteolytic enzymes, which could have a similarly detrimental effect on the virus, and these other proteolytic enzymes may have slightly reduced the PLRV transmission efficiency. Reduced cysteine proteases in the gut have been reported to have detrimental effects on aphid gut cells (68). Damage to the gut cells could hamper virus acquisition into the aphid vector; however, no such effects were observed during our experiments. Our data suggest a biphasic effect of cathB in PLRV transmission by M. persicae, where some cathepsin activity is required inside a particular organelle in the midgut (possibly lysosomes), but very high levels of the enzyme on the brush border membranes inhibits virus transmission, as we observed in T-Myzus. Other nonmutually exclusive ideas include that different protein complexes are formed with cathB with distinct functions and the expression of cathB in the gut cell membranes of T-Myzus also relocalizes a receptor critical for virus acquisition; the pH of the gut lumens are different and this change protease activity which changes vector competency; or that the proteolytic specificities of the enzyme are not the same in T-and P-Myzus because of their distinct subcellular localizations. Finally, the larger amount of calcium in T-Myzus as compared with P-Myzus may cause an increase in lysosomal activities in the aphid midgut, a hypothesis that is also consistent with the label-free proteomics data showing an increase in the expression of other enzymes with predicted lysosomal function in T-Myzus. As lysosomes are the hub of cellular homeostasis (79) , an increase in lysosomal exocytosis or other lysosomal functions may impair intracellular trafficking, egress, or uptake of the virus or one of its receptors.
The generalist aphid M. persicae feeds on a large array of host plants, from more than 40 plant families (80), which requires a broad adaptive capacity and phenotypic plasticity. This aphid is one of the most important vectors of PLRV (23,81,82). It has been shown in the past that the host plant on which this aphid is reared on influences its vector competence (83)(84)(85). For example, a difference in virus transmission was observed for Zucchini yellow mosaic virus (ZYMV), a virus that is nonpersistently transmitted by M. persicae, when aphids were reared on two plants that are both hosts of M. persicae but nonhosts of ZYMV, the Brassica mustard (Brassica juncea) and the Malvaceous okra (Abelmoschus esculentus) (85). Overall, aphids reared on mustard plants had higher transmission rates of ZYMV to host recipient plants (Cucurbita pepo) than okra-reared aphids. Interestingly, the host switch had a complex effect, giving intermediate transmission rates when mustard-reared aphids were given a preacquisition time of 24 h on okra plants, showing that the host switch effect was transient and in a continuum. In a previous study with PLRV, 75% of the nymphs born on physalis transmitted PLRV to recipient plants, whereas only 49% of the nymphs born on rape (Brassica rape) transmitted the virus (84). Similarly to our study, these authors used a host plant of PLRV, physalis, and a nonhost of PLRV, the Brassica rape. A similar effect was observed in another study, when M. persicae reared on rape were less efficient in transmitting Beet Yellows Virus (BYV) than aphids reared on beet, a host of the virus (83). Again, in this case, rape is a nonhost of BYV. Therefore, it seems that rearing the aphids on a host plant of the virus increases the efficiency of virus transmission by this generalist aphid, which may have implications for virus epidemiology. Our data add to this body of knowledge and for the first time provide important insights into how the host plant is impacting vectoring ability at the molecular level with the support of proteomics, biochemical and imaging data. ¶