Article pubs.acs.org/jpr
Model System-Guided Protein Interaction Mapping for Virus Isolated from Phloem Tissue Stacy L. DeBlasio,†,‡ Richard S. Johnson,§ Michael J. MacCoss,§ Stewart M. Gray,†,∥ and Michelle Cilia*,†,‡,∥ †
Agricultural Research Service, USDA, Ithaca, New York 14853, United States Boyce Thompson Institute for Plant Research, Ithaca, New York 14853, United States § Department of Genome Sciences, University of Washington, Seattle Washington 98109, United States ∥ Department of Plant Pathology and Plant-Microbe Biology, Cornell University, Ithaca, New York 14853, United States ‡
S Supporting Information *
ABSTRACT: Phloem localization of plant viruses is advantageous for acquisition by sapsucking vectors but hampers host−virus protein interaction studies. In this study, Potato leafroll virus (PLRV)−host protein complexes were isolated from systemically infected potato, a natural host of the virus. Comparing two different co-immunoprecipitation (co-IP) support matrices coupled to mass spectrometry (MS), we identified 44 potato proteins and one viral protein (P1) specifically associated with virus isolated from infected phloem. An additional 142 proteins interact in complex with virus at varying degrees of confidence. Greater than 80% of these proteins were previously found to form high confidence interactions with PLRV isolated from the model host Nicotiana benthamiana. Bioinformatics revealed that these proteins are enriched for functions related to plasmodesmata, organelle membrane transport, translation, and mRNA processing. Our results show that model system proteomics experiments are extremely valuable for understanding protein interactions regulating infection in recalcitrant pathogens such as phloem-limited viruses. KEYWORDS: Luteoviridae, polerovirus, Potato leafroll virus, host−pathogen interactions, phloem-limited, co-immunoprecipitation, high-resolution mass spectrometry, model system, Solanum tuberosum, SAINT
■
INTRODUCTION Potato leaf roll virus (PLRV) is an agriculturally important phloem-limited pathogen in the family Luteoviridae, which causes significant yield loss in potato (Solanum tuberosum), especially in third world countries where economically challenged farmers do not have access to commercially available control agents. Members of this family of +ssRNA viruses (collectively referred to as luteovirids) are interesting in that they have evolved a way to actively restrict themselves to the phloem of their host plant, a tropism that facilitates acquisition by aphids during feeding.1−4 Unlike stylet-borne plant viruses such as Potato virus Y, which is mechanically transmitted by probing insects, luteovirids are transmitted in a persistent, circulative manner. Virions must migrate across different tissue barriers in their insect vector in order to be transmitted to plants.5 In general, aphids must feed on infected phloem for at least 6 h to become competent for virus transmission.6 Consistent with the host−vector manipulation hypothesis,7 luteovirids have been selected to alter the physiology of their host plant to attract vectors as well as to maintain some level of host health that prevents death of the phloem to sustain prolonged feeding by aphid vectors. Indeed, aphids are more attracted to luteoviridinfected plants than noninfected hosts, but once virus is acquired, viruliferous aphids are less responsive to these same virusinduced changes and prefer to seek out healthy hosts.8−10 © XXXX American Chemical Society
Once deposited in host phloem cells via the saliva of a feeding aphid, luteovirids replicate most efficiently in companion cells and then move systemically throughout the plant as assembled virions.11−13 The capsid structure for this family consists of a nonenveloped, icosohedral shaped virion composed of two structural proteins, the coat protein (CP) making up the majority of the virion and a minor component known as the readthrough protein (RTP). The RTP is generated from the readthrough of a leaky stop codon at the end of the CP open-reading frame to create a readthrough domain (RTD) that protrudes from the surface of the virion. It is the RTD C-terminal domain, which restricts virions to the phloem. Truncation of this region in PLRV permits the virus to invade mesophyll tissues, where it becomes unavailable for aphid acquisition and alters symptom development, in a host specific manner.3,4 The RTD also provides specificity for insect interactions.1−4,14,15 Although much is known about how each of the proteins encoded by luteovirids regulates different aspects of the viral infection cycle in planta,2−4,16−19 very little is known about the host proteins these viruses need to interact with in order to replicate and traffic through the plant. Previously, we reported the identification of over 1000 plant and three viral proteins Received: August 4, 2016 Published: October 20, 2016 A
DOI: 10.1021/acs.jproteome.6b00715 J. Proteome Res. XXXX, XXX, XXX−XXX
Article
Journal of Proteome Research in complex with PLRV particles or structural proteins using co-IP-MS.15,20 In mammalian systems, this approach has been successfully used to capture and characterize the dynamic progression of protein−protein interactions formed and destroyed over the course of a viral infection cycle.21−25 The technique is particularly useful in identifying interaction networks between proteins that exist as multimeric complexes within cells like assembled virions compared to protocols such as yeast-two hybrid that are limited to studying one on one interactions. In our previous experiments, a PLRV infectious clone26 was overexpressed in Nicotiana benthamiana leaf cells (mesophyll and vasculature) and plant−virus protein complexes were captured using Protein A Dynabeads coated with an in-house anti-PLRV polyclonal antibody (α-PLRV). Our workflow was optimized for rapid isolation of these complexes using conditions shown to preserve endogenous virus−host interactions in mammalian cells.15,20 Bioinformatics analysis showed that plant proteins in the PLRV−N. benthamiana interaction network regulate key biochemical processes including ion transport and trafficking.20 Some of those interactions were found to be lost or weakened in the absence of the RTP,15 suggesting a role in processes regulated by this viral domain such as tissue tropism and symptom development.2,4 Although N. benthamiana is a model host of PLRV,27 this species is not ideal to study some of the hostspecific effects on PLRV movement and tissue tropism that have been reported.3,28,29 For example, truncation of the RTP leads to phloem-unloading in a natural, weed host of PLRV, but not in N. benthamiana.3 Therefore, in this study, we extended our co-IPMS analysis to identify proteins interacting with PLRV in the context of a systemic infection in a natural host, S. tuberosum cv. Russet Burbank (potato).
■
eluates and input fractions (diluted 1:1 in sample buffer) were first separated on a 10% Mini-PROTEAN TGX precast gel (Bio-Rad), followed by wet transfer to 0.45 micron NitroBind pure nitrocellulose (GE Water and Process Technologies) using the Mini-PROTEAN tetra cell system (Bio-Rad).20 Levels of PLRV CP and RTP were simultaneously detected using an alkaline phosphatase conjugated, α-PLRV primary antibody (Agdia) following the protocol outlined in DeBlasio et al. (2015).20 To visualize proteins co-immunoprecipitating with virion, 10 μL of co-IP eluates were separated on a gradient (4−20%) MiniPROTEAN TGX precast gel (Bio-Rad) and stained with SYPRO Ruby stain (Invitrogen) following the manufacturer’s instructions and imaged on a Typhoon variable mode imager (GE Healthcare). Co-Immunoprecipitation and Sample Preparation for Mass Spectrometry
An in-house PLRV polyclonal antibody raised against cesium chloride purified virions was first cross-absorbed by sodium sulfate precipitation32 using cryolysed, noninfected S. tuberosum cv. Russet Burbank leaf tissue and then purified on-column using a Pierce Protein A IgG Purification Kit (ThermoFisher Scientific). Purified antibody was covalently coupled to M-270 epoxy magnetic Dynabeads (Invitrogen) following the protocol described in Cristea and Chait (2011)33 at a concentration of 10 μg/mg of beads. PLRV−potato protein complexes were extracted from 1 g of cryolysed, systemically infected or healthy S. tuberosum leaf tissue in 5 mL of lysis buffer (50 mM HEPES, 0.4% Triton X-100, 110 mM potassium acetate, 2 mM MgCl2 plus protease inhibitors) on ice as described in DeBlasio et al. (2015).20 The resulting plant homogenate was diluted 1:2 in fresh lysis buffer and 10 mL of diluted homogenate incubated with 5 mg of α-PLRV conjugated M-270 epoxy or α-PLRV (cross-absorbed antibody) coated Protein A Dynabeads20 (Invitrogen) for 1 h at 4 °C. Beads were washed four times with 1 mL of lysis buffer and twice with 1 mL of 1× PBS. Proteins were subjected to on-bead reduction, alkylation, and trypsin digestion following DeBlasio et al. (2015).20
EXPERIMENTAL SECTION
Plant and Virus Material
S. tuberosum cv. Russet Burbank plants were infected with a wild-type PLRV infectious clone26 by graft inoculation30 using scions from S. sarrachoides that were systemically infected via Agrobacterium-mediated transformation.18,19 Infected daughter tubers were sprouted and grown at 22 °C for 3−4 months under 12 h days. Leaf tissue was harvested from these plants and cryogenically lysed using a Mixer Mill 400 (Retsch) for six sets of 3 min grinds at a vibrational frequency of 30 Hz to achieve near total cell disruption.20 The resulting leaf powder was never thawed and was stored at −80 °C until used.
Mass Spectrometry
Peptides were analyzed using an Orbitrap Fusion Tribrid mass spectrometer (Thermo Scientific). At least 1 μg of each digest was loaded from the autosampler onto a 150 μm Kasil fritted trap packed with Reprosil-Pur C18-AQ (3 μm bead diameter) to a bed length of 2 cm at a flow rate of 2 μL/min. After loading and desalting using a total volume of 10 μL of 0.1% formic acid plus 2% acetonitrile, the trap was brought online with a pulled fusedsilica capillary tip (75 μm i.d.) packed with the same Reprosil C18-AQ that was mounted in an in-house constructed microspray source and placed in line with a Waters Nanoacquity binary UPLC pump plus autosampler. Peptides were eluted off the column using a gradient of 2−35% acetonitrile in 0.1% formic acid over 120 min, followed by 35−60% acetonitrile over 10 min at a flow rate of 250 nL/min. The mass spectrometer was operated using electrospray ionization (2 kV) with the heated transfer tube at 275 °C using data dependent acquisition (DDA) in “Top Speed” mode, whereby one orbitrap mass spectrum (m/z 400−1600 with quadrupole isolation) was acquired with multiple linear ion trap tandem mass spectra every three seconds or less. The resolution for MS in the orbitrap was 120,000 at m/z 200, and for MS/MS the linear ion trap provided unit resolution. The automatic gain control target for MS in the orbitrap was 2e5, whereas for MS/MS it was 1e4, and the maximum fill times were 20 and 35 ms, respectively. The MS/MS spectra were acquired
Immunodetection of PLRV
Cryogenically lysed leaf powder from agro-infiltrated N. benthamiana20 and systemically infected S. tuberosum (described above) were solubilized in 1× phosphate-buffered saline (PBS, pH 7.4) at a concentration of 200 mg of tissue per mL of buffer. Each homogenate was serially diluted (1:2 or 1:5) in fresh buffer and relative levels of PLRV measured by double-antibody sandwich ELISA (DAS-ELISA) following procedures outlined in Barker and Solomon31 and Lee et al.19 using commercially available PLRV capture and alkaline-phosphatase conjugated detection antibodies (Agdia). Average Δ absorbance (405 minus 490 nm) was measured on an Epoch spectrophotometer (BioTek) 1 h after addition of substrate. All values were normalized to the absorbance readings for 1× PBS. For Western blot and SDS−PAGE analysis, proteins from co-IPs were eluted from magnetic beads in 50 μL of 2× Laemmli sample buffer (Bio-Rad) supplemented with 5% (v/v) 2-mercapethanol at 70 °C for 10 min. Forty microliters of co-IP B
DOI: 10.1021/acs.jproteome.6b00715 J. Proteome Res. XXXX, XXX, XXX−XXX
Article
Journal of Proteome Research using quadrupole isolation with an isolation width of 1.6 m/z and HCD collision energy (NCE) of 30%. The precursor ion threshold intensity was set to 1e4 in order to trigger an MS/MS acquisition. MS/MS acquisitions were allowed for precursor charge states of 2−7. Dynamic exclusion (including all isotope peaks) was set for 30 s using monoisotopic precursor selection with a mass error of 15 ppm. The fragment ions were analyzed in the linear trap using the “rapid” scan rate. Duplicate runs were acquired under identical conditions except that the dynamic exclusion time was set to 5 s.
consensus localization obtained using the subcellular localization database for Arabidopsis proteins, SUBA3,45 for the corresponding Arabidopsis thaliana orthologues or their identification in the A. thaliana plasmodesmal proteome identified by FernandezCalvino et al. (2011).46
Protein Identification and Label-Free Quantification
A systemic infection with the same WT PLRV cDNA clone used in our N. benthamiana study20 was established in S. tuberosum cv. Russet Burbank (potato) via graft inoculation. Leaf tissue from plants grown from infected tubers (secondary infection) was used in subsequent experiments and exhibited slight interveinal chlorosis, a common PLRV disease symptom47 at the time of harvest. The concentration of PLRV was significantly lower in this tissue compared to the agro-infiltrated N. benthamiana leaf tissue that was used in our previous analysis20 as revealed by DAS-ELISA (Figure 1A). PLRV levels in potato (n = 3) peaked at an average Δ absorbance A405−490 (ΔABS) of 0.42 SE ± 0.013 in undiluted samples compared to >3.5 ΔABS for undiluted N. benthamiana homogenate (n = 3). At the concentration (1:5 dilution) used for co-IP of PLRV in the N. benthamiana study,20 levels of PLRV in systemically infected potato resulted in a ΔABS value of 0.29 SE ± 0.005 compared to 3.58 SE ± 0.01 for N. benthamiana. Thus, PLRV accumulation in Agrobacteriuminfiltrated N. benthamiana tissue is >625-fold higher than in systemically infected potato tissue under our growth conditions. Immunoprecipitation of PLRV from undiluted, systemically infected S. tuberosum leaf homogenate following the same workflow used in the N. benthamiana interactome study20 also revealed a lower enrichment of the PLRV capsid proteins (CP and RTP) by both Western blot (Figure 1B) and SDS−PAGE (Figure 1C), compared to using Agrobacterium-infiltrated N. benthamiana.20 The level of CP in the undiluted potato input fraction was barely detectable on Western blots while both capsid proteins could still be detected in diluted N. benthamiana homogenate (Figure 1B). Enrichment of CP/RTP, albeit to a lesser extent in co-IPs from systemically infected potato, was detected in Western analysis of eluates from both species after a 1 h incubation with α-PLRV coated Protein A Dynabeads (Figure 1B). The truncated and oligomeric forms of the RTP could also be detected in co-IPs from potato (Figure 1B), similar to what we observed using N. benthamiana.15 Thus, multiple forms of the RTP are expressed in the context of a natural, systemic infection and are not artifacts of overexpression in N. benthamiana. Our results support mutational studies, which show that truncated forms of the RTP play a role in regulating different aspects of the viral infection cycle either in cis or in trans including phloem-retention and systemic movement.1,3 When analyzed by SDS−PAGE, the full-length forms of CP and the RTP were also detected in 20% the volume of α-PLRV co-IP eluates from agro-infiltrated N. benthamiana tissue as well as differences in the composition of proteins co-immunoprecipitating with virus compared to negative control immunoprecipitations using mockinfiltrated (N. benthamiana) or healthy (S. tuberosum) tissue.20 However, the only detectable difference observed by SDS−PAGE in the same volume of co-IP eluate from systemically infected potato homogenate compared to negative controls was the presence of CP (Figure 1C). Along with our DAS-ELISA results, these data show that replication of PLRV in N. benthamiana mesophyll cells initially driven by the 35s cauliflower mosaic virus
■
RESULTS AND DISCUSSION
Co-Immunoprecipitation Can Successfully Isolate PLRV from Systemically Infected Phloem Tissue
Mass spectrometry data were converted into Mascot generic format (mgf) files using tools in Proteowizard34 and were searched using Mascot35 with the following modifications considered: methylthio on cysteines (fixed), oxidation of methionine (variable), and deamidation of glutamine/asparagine (variable). One missed tryptic cleavage was allowed. Mass measurement accuracy was set at 30 ppm for precursor ions and 0.8 Da fragment ions. The database used for protein identification was created using amino acid sequences corresponding to all coding gene sequences from the recently released assembled genome of the doubled monoploid S. tuberosum group Phureja DM1-3,36 amino acid sequences from species of Luteoviridae, and common mammalian contaminant protein sequences obtained from NCBI. Search results were imported into Scaffold-Q+ version 4.4.6 (Proteome Software) for labelfree quantification by spectral counting using the proteinclustering feature20,37−39 with the following identification filter thresholds: two-peptide minimum, a peptide identification threshold of ≥95%, and a protein identification threshold of ≥80%. The false discovery rate for this analysis was 0.3, Student t test) was observed in the levels of IgG between co-IPs from systemically infected tissue compared to healthy tissue (Table S1), indicating that the levels of beads in each of the experiments (M-270 and Protein A) were equal between bait and control co-IPs. Therefore, plant proteins found enriched in viral co-IPs compared to the controls can be attributed to direct or via network interaction with PLRV and not to variability in the amount of beads between samples.
Figure 1. Levels of PLRV are significantly lower in systemically infected S. tuberosum (S.t.) compared to agro-infiltrated N. benthamiana (N.b.) leaf tissue as measured by (A) DAS-ELISA using α-PLRV (Agdia), error bars = ±standard error of n = 3, *p < 0.01, Student t test, and by (B) Western blot analysis showing enrichment of the PLRV capsid proteins (CP and RTP) in co-IPs from 5 mL of diluted (1:5) N. benthamiana or undiluted S. tuberosum homogenate using Protein A (Prot A) Dynabeads coated with an in-house PLRV antibody raised against purified virus. The input levels of CP and RTP in each infected homogenate (INPUT) as well as negative controls (mock-infiltrated and healthy) are also shown. N-mers = multimers and trunc RTP = truncated RTP. (C) SDS−PAGE analysis of 20% of the co-IP eluates analyzed in panel B. Gel was stained with SYPRO Ruby and imaged using a Typhoon variable mode imager (GE Healthcare).
The Majority of High-Confidence PLRV-Interacting Proteins Were Identified Using Both Protein A and M-270 Beads for Co-Immunoprecipitation
After excluding common mammalian contaminants, a total of 664 and 419 prey proteins were detected in M-270 and Protein A Dynabead co-IP reactions, respectively. To assess the interaction specificity of each of these prey proteins with PLRV, we used SAINT,40−42 a proteomics tool that statistically models total spectral counts between negative controls (healthy) and bait (PLRV) co-IP samples in a semisupervised manner to calculate
promoter of the infectious clone results in a much higher accumulation of virus and as a result leads to a higher co-IP enrichment compared to systemically infected S. tuberosum, where virus endogenously replicates only in the cytosol of phloem companion and parenchyma cells.13,48 D
DOI: 10.1021/acs.jproteome.6b00715 J. Proteome Res. XXXX, XXX, XXX−XXX
Article
Journal of Proteome Research
RNA helicase,54 dynamin ADL4,55 and translocon at the outer envelope membrane of chloroplasts (TOC) 75-III.56 Proteins in category II and III were identified as enriched in PLRV co-IPs in either the M-270 or Protein A data set, respectively (Figure 3, Table 1, Table S1). Proteins in categories II and III tend to have lower spectral counts as compared to category I proteins. It is possible that the number of spectra sampled for these ions are limited by resampling of peptides from the most abundant protein species in the co-IPs.57 Other, nonmutually exclusive explanations include that these proteins may interact transiently with PLRV or are at lower abundance overall in their protein complexes. A total of 72 potato proteins were classified in category II, identified using the M-270 beads but not the Protein A bead matrix data set. Eight of these had high confidence interaction scores (SP ≥ 0.8, Table 1). These proteins include pre-mRNA-splicing factor (SR1), vesicle-fusing ATPase (NSF), calreticulin-3-like (CRT3), and a serine carboxypeptidase (SCPL27). Forty-nine host proteins were identified as category III interactions, identified as enriched in PLRV co-IPs using Protein A beads but not detected in the M-270 data set (Table S1). Thirteen of these proteins were identified as having SP ≥ 0.8 (Table 1). Surprisingly, 10 were ribosomal proteins. The three nonribosomal proteins identified included flotillin1, a lipid-raft associated protein,58 TOC159, and dynamin DRP4C.
Figure 2. Co-IP bead matrix has no effect on enrichment of PLRV from systemically infected tissue. (A) Western blot analysis of SDS eluates from co-IPs from systemically infected S. tuberosum leaf tissue using Protein A (ProtA) or M-270 epoxy (M-270) magnetic Dynabeads coated with our in-house α-PLRV antibody. Levels of PLRV coat protein (CP) as well as the full length, truncated, and oligomeric (N-mers) forms of the readthrough protein (RTP) in SDS eluates are shown. Negative controls with healthy tissue are not shown but were similar to healthy control co-IPs shown in Figure 1B. Blots were probed with an alkaline phosphatase conjugated, α-PLRV primary antibody (Agdia). (B) Bar graph shows the average total spectral counts (SPC) for PLRV CP/RTD (readthrough domain), rabbit IgG, and Protein A detected by nLC-MS/MS and Scaffold analysis of on-bead digests of replicate α-PLRV co-IPs mentioned above. Error bars represent the standard error for n = 3. Student t test comparing total SPCs measured in co-IPs using M-270 epoxy to Protein A Dynabeads shows significant differences in the average total SPCs for all three proteins. *p < 0.01, Student t test.
Protein Interaction Data from a Virus-Infected Model Host Is a Blueprint To Interpret Interaction Data from a Natural Host
Interestingly, orthologues of 88% of all category I S. tuberosum proteins were found to be highly enriched (8−210-fold, SP ≥ 0.8) in our PLRV−N. benthamiana interactome despite some having SP scores
