key: cord-0907012-sgswo5s4
authors: Sun, Yanting; Hu, Boli; Fan, Chengfei; Jia, Lu; Zhang, Yina; Du, Aifang; Zheng, Xiaojuan; Zhou, Jiyong
title: iTRAQ‐based quantitative subcellular proteomic analysis of Avibirnavirus‐infected cells
date: 2015-06-17
journal: Electrophoresis
DOI: 10.1002/elps.201500014
sha: 08864f104f76d29320882bcbb20b6393247ecaa2
doc_id: 907012
cord_uid: sgswo5s4

Infectious bursal disease virus (IBDV) enters the host cells via endocytic pathway to achieve viral replication in the cytoplasm. Here, we performed LC‐MS/MS coupled with isobaric tags for relative and absolute quantification labeling of differentially abundant proteins of IBDV‐infected cells using a subcellular fractionation strategy. We show that the viral infection regulates the abundance and/or subcellular localization of 3211 proteins during early infection. In total, 23 cellular proteins in the cytoplasmic proteome and 34 in the nuclear proteome were significantly altered after virus infection. These differentially abundant proteins are involved in such biological processes as immune response, signal transduction, RNA processing, macromolecular biosynthesis, energy metabolism, virus binding, and cellular apoptosis. Moreover, transcriptional profiles of the 25 genes corresponding to the identified proteins were analyzed by quantitative real‐time RT‐PCR. Ingenuity Pathway Analysis clustered the differentially abundant proteins primarily into the mTOR pathway, PI3K/Akt pathway, and interferon‐β signaling cascades. Confocal microscopy showed colocalization of the viral protein VP3 with host proteins heterogeneous nuclear ribonucleoprotein H1, nuclear factor 45, apoptosis inhibitor 5, nuclear protein localization protein 4 and DEAD‐box RNA helicase 42 during the virus infection. Together, these identified subcellular constituents provide important information for understanding host–IBDV interactions and underlying mechanisms of IBDV infection and pathogenesis.

Avibirnavirus infectious bursal disease virus (IBDV), belonging to the Birnaviridae family, is a pathogenic agent of a highly contagious disease that damages immature B lymphocytes in the bursa of Fabricius in poultry and causes severe immunodeficiency and mortality [1] . IBDV has a bisegmented double-stranded RNA genome (segments A and B) [2] . The large ORF of the segment A encodes a polyprotein that can undergo cotranslational self-proteolytic processing and release three polypeptides, the precursor capsid protein (pVP2), the protease VP4 and the multifunctional polypeptide VP3. The small ORF in segment A encodes a nonstructural protein VP5 with a molecular weight of 17 kDa. Segment B contains a single ORF encoding the RdRp VP1 polypeptide. In 2007, a cDNA microarray was used to analyze the differentially abundant transcripts of IBDV-infected chicken embryo fibroblasts (CEFs), which provides an overview of the mRNA profiles of cells during virus infection [3] . However, mRNA is not always the same with the protein level. Consequently, potential cellular factors involved in viral infection are more likely obtained from the proteomic profile of the host cellular response. Now, large-scale screening is well recognized to be very effective for identifying host proteins of interest, many of which are unexpected and may lead to new discoveries regarding the host-virus interactions after additional functional validation. In the past decade, researchers have been paying increasingly close attention to the host response to virus infection at the translational level. 2DE coupled with MALDI-TOF MS has been widely applied to detecting the dynamics of host-pathogen interactions at various stages of virus infection [4] [5] [6] [7] [8] [9] [10] [11] [12] . However, proteins with low abundance and those that are very large or very small have been proven to be difficult to resolve using 2DE gels [13] .

Using isobaric tags for relative and absolute quantification (iTRAQ) combined with LC and ESI-MS/MS analysis is emerging as a powerful methodology in the search for disease-specific targets [14] . In the past few years, iTRAQ has been used to explore and discover previously unidentified factors in the cell proteomic response to infection with porcine circovirus type 2 [15] , porcine reproductive and respiratory syndrome virus [16] , human immunodeficiency virus 1 [17] , human papilloma virus [18] , simian immunodeficiency virus [19] and influenza virus [20] . However, iTRAQ was rarely used to determine the subcellular proteomic response to IBDV infection. In this study, we utilized an iTRAQ-based quantitative subcellular proteomics approach followed by bioinformatic analysis and Western blot coupled with quantitative real-time RT-PCR and colocalization assays to observe the differentially abundant proteins profiles of IBDV-infected cells at 24 h after inoculation. We find that viral infection regulates the abundance and/or subcellular localization of more than 1000 host proteins in either the nuclear or the cytoplasmic fraction at the early phase of infection. These data provide clues to further understanding the replication and pathogenesis of IBDV and virus-host interactions.

IBDV NB strain with CPE characteristics, including the infected cells became round, increasing light refrangibility, netting, and nucleus broken, was propagated in CEF as previously described [21, 22] . DF-1 cells (ATCC CRL-12203) were continuous cell line of CEF and routinely cultured at 39°C with 5% CO 2 in DMEM (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (GibcoBRL Life Technologies, Grand Island, NY). DF-1 cell monolayers were inoculated with the NB strain (10 7.2 TCID 50 /0.1 mL) at a multiplicity of infection (MOI) of 1. Mock-infected DF-1 cells were used as the negative control. The IBDV-and mockinfected cells were harvested at 24 h postinfection (hpi) through scraping followed by centrifugation at 5000 × g for 10 min. Each sample of cells was washed thrice with ice-cold PBS followed by subcellular fractionation.

IBDV-and mock-infected DF-1 cells were harvested at 24 hpi for the flow cytometric determination of IBDV infection [23] . Cells were washed twice with 1× PBS followed by fixation in 4% paraformaldehyde at room temperature for 30 min and subsequently permeabilized with 0.1% Triton X-100 in PBS for 10 min at room temperature. After washing twice with PBS, cells were incubated with the FITC-labeled anti-VP4 mAb at 4°C for 1 h in the dark. Cells were washed twice with 1× PBS, resuspended in 1 mL PBS, and analyzed by the flow cytometry (Beckman Coulter, Miami) using Flowjo7.6 TM software. The percentage of IBDV-positive cells was calculated from the FITC fluorescence histogram using a region that was defined based on analysis of the noninfected control.

According to the manufacturer's instructions, cytosolic and nuclear protein fractions were sequentially extracted by the ProteoExtract R Subcellular Proteome Extraction Kit (Merck, Darmstadt, Germany). Briefly, cells were harvested by centrifugation at 600 × g for 10 min at 4°C and washed twice with cold wash buffer. Thereafter, the cells were resuspended in 1 mL of cytosol extraction buffer containing protease inhibitor cocktail (included in the fractionation kit), incubated for 10 min at 4°C with gentle agitation and centrifuged at 4°C at 800 × g for 10 min. The resultant supernatant was harvested and considered as the cytosolic fraction. The membranes and membrane organelles were dissolved with extraction buffer II. Subsequently, the pellets were resuspended in 500 L nuclear extraction buffer containing a protease inhibitor mixture and Benzonase R Nuclease, incubated at 4°C for 10 min and centrifuged at 7000 × g for 10 min. The supernatant containing predominantly nuclear proteins was collected. Finally, protein concentrations were measured using a BCA Protein Assay Kit (Novagen, Darmstadt, Germany).

In the light of the previously FASP (filter aided sample preparation) procedure, proteins were digested [24] . Each sample about 32 g was labeled using iTRAQ reagent 8-plex Multiplex Kit in accordance with the manufacturer's instructions (Applied Biosystems). The cytoplasmic fractions (F1) of IBDV-infected samples including three biological repeats were labeled as (Sample1)-113, (Sample2)-114 and (Sample3)-115, respectively, while mock-infected samples were labeled as (Sample4)-116, (Sample5)-117 and (Sample6)-118. The nuclear fraction (F2) samples were labeled similarly.

Peptides labeled with iTRAQ were fractionated by strong cation exchange (SCX) chromatography using the AKTA Purifier system (GE Healthcare). After reconstituted and acidified with 2 mL buffer A (10 mM KH 2 PO 4 in 25% of ACN, pH 2.7), the dried peptide mixture was loaded onto a Poly-SULFOETHYL 4.6 × 100 mm column (5 m, 200Å, PolyLC, MD, USA). Then, the peptides were eluted at flow velocity of 1 mL/min with a gradient of 0-10% buffer B (500 mM KCl, 10 mM KH 2 PO 4 in 25% of ACN, pH 2.7) for 2 min, 10-20% buffer B for 25 min, 20-45% buffer B for 5 min, and 50-100% buffer B for 5 min. The fractions were collected every 1 min and the elution process was monitored by absorbance at 214 nm. The collected fractions were desalted on C18 Cartridges (Empore TM SPE Cartridges C18 [standard density], bed id 7 mm, volume 3 mL, Sigma) and then combined into six pools. Each fraction was concentrated by vacuum centrifugation following reconstituted in 40 L of 0.1% (v/v) TFA. Afterwards, the samples were performed using Q Exactive mass spectrometer that was coupled with the Easy nLC apparatus (Thermo Fisher Scientific). The samples were added into C18-RP column (10 cm long, 75 m inner diameter, 3 m resin) in buffer A (0.1% formic acid in water), and then separated with buffer B (80% ACN in 0.1% formic acid) by IntelliFlow technology. The MS data were collected in accordance with following methods: after each full scan collected ten fragment patterns (MS2 scan).

MS/MS data were searched and quantified using Maxquant 20 ppm, MS/MS tolerance = 0.1 Da, enzyme = trypsin, missed cleavage = 2; fixed modification: carbamidomethyl (C), iTRAQ8plex (K), iTRAQ8plex (N-term), variable modification: oxidation (M), FDR ࣘ 0.01. The protein identification that was inferred from the unique peptide identification in three independent biological replicates was considered. In addition, a p-value less than 0.05 was considered to be statistically significant. The fold changes ࣙ1.2 or ࣘ0.83 of proteins were designated significant.

We have designed specific primers (Supporting Information Table 1 ) to amplify target genes corresponding to the MS/MSidentified proteins. Using cDNA from IBDV-infected DF-1 cells, gene products were analyzed using the 7500 real-time PCR system (Applied Biosystems). The reaction contained 200 ng of cDNA template, 200 nM of each primer, and 1× SYBR premix Ex Taq (Perfect Real Time, TakaRa, Dalian, China), supplemented with water to a total volume of 20 L. After an initial denaturation at 95°C for 30 s, the samples were amplified by 40 cycles of 95°C for 5 s and 60°C for 34 s. We obtained the melting curves, and then performed quantitative analysis of the data by the 7500 System SDS software version 1.3.1 (Applied Biosystems) with relative quantification (cycle threshold [⌬ ⌬ Ct]) method. Simultaneously, mock-infected samples in parallel were used as negative controls.

In order to evaluate the differentially abundant proteins identified by iTRAQ-labeled LC-MS/MS system, IFN-induced protein with tetratricopeptide repeats 5 (IFIT5), signal transducer and activator of transcription 1 (STAT1), nuclear factor 45 (NF45), heterogeneous nuclear ribonucleoprotein H1 (hn-RNPH1), and apoptosis inhibitor 5 (API5) were selected for Western blot analysis. Equal amounts of proteins from each sample were separated on 8% SDS-PAGE gels and transferred to NC membranes (Amersham Biosciences, Berks, UK) by using a semidry transfer cell (Bio-Rad Trans-blot SD). The membranes were blocked with 5% nonfat dried milk and then stained with a rabbit anti-STAT1 mAb (Epitomics, Cambridge, UK), mouse polyclonal antibody (pAb) to IFIT5 prepared in our laboratory (unpublished data), goat anti-NF45 pAb, goat anti-hnRNPH1 pAb, rabbit anti-API5 pAb, or rabbit anti-API5 pAb (Santa Cruz Biotechnology, Santa Cruz, CA). After washing three times with 0.05% Tween-20 in PBS (PBST), the membranes were incubated with horseradish peroxidase (HRP) conjugated goat anti-mouse IgG, HRPconjugated goat anti-rabbit IgG, or HRP-conjugated donkey anti-goat IgG (Kirkegarard & Perry Laboratories, Gaithersburg, MD) at 37°C for 1 h. Finally, the detection was performed using chemiluminescence luminal reagents (Su-perSignalWest Pico Trial Kit, Pierce, Rockford, IL).

DF-1 cells inoculated with IBDV at MOI = 1 were cultured for 24 h. At the end of this infection period, the cells were fixed with cold acetone/methanol (1/1) for 20 min at −20°C, and then dried in air. Fixed cells were incubated with a pAb adding to the NF45, hnRNPH1, API5, nuclear protein localization protein 4 (NPL4), and DDX42 at 37°C for 90 min. After washing three times with PBST, the cells were incubated with FITC-conjugated goat anti-rabbit IgG (Sigma), FITCconjugated rabbit anti-goat IgG, and Alexa Fluor 566 goat anti-mouse IgG (Life Technologies) for 1 h at 37°C. Next, cell nuclei were stained blue using 4 ,6-diamidino-2-phenylindole (DAPI, Sigma). Finally, the cells were washed three times with PBS and visualized using a Zeiss LSM780 laser confocal microscope.

In order to gain further insight into the most relevant interaction networks and biological functions of the differentially abundant proteins, data of all proteins were analyzed using Ingenuity Pathway Analysis (IPA, Ingenuity Systems, www.ingenuity.com). Network construction was carried out as described previously [25] . In other words, we uploaded differentially abundant proteins (showing fold change of ࣙ1.2 or ࣘ0.83) identified into the IPA application, and then constructed the interaction network. Possible connections between those mapped genes were evaluated on the basis of knowledge curated in the IPA database. Each protein represented a node, and the interaction relationships between the nodes were linked by line. The linkages have been supported by at least one or more reference. Different shapes and the color of nodes represented differentially abundance and the functional classes of the proteins, respectively. Furthermore, IPA-curated canonical pathways were used to explore possible metabolic and cell signaling pathways.

For analysis of the subcellular proteome, DF-1 cells were first infected with IBDV for 12, 18, 21, and 24 h. To select the optimal time point for iTRAQ analysis, virus infection profiles were identified by IFA using a mAb to VP3 protein of IBDV as the primary antibody. IFA results revealed that IBDV-specific immunofluorescence was observed at 12, 18, 21, and 24 hpi (with the monolayer appearing nearly 90% positive at the last time point), but not in mock-infected DF-1 cells (Fig. 1A) . Additionally, we confirmed the rate of infection by flow cytometry (Fig. 1B) , which showed that the percentage of IBDV-positive cells was 99.4 ± 0.2% at 24 hpi. These results indicated that 24 hpi would be the best sample collection time point for iTRAQ analysis. Therefore, cytoplasmic and nuclear fractions were extracted from the cells at 24 hpi. The enrichment of proteins in these two fractions was confirmed by Western blot using GAPDH as a cytoplasmic protein marker and histone 3 as a nuclear protein marker (Fig. 1C) . The result showed that the protein markers GAPDH and histone 3 appeared in cytoplasm and nucleus, respectively, indicating the procedure used for cytoplasmic and nuclear fractionation was suitable for iTRAQ analysis. 

Protein identification and quantification from cytoplasmic fraction and nuclear fraction were performed, respectively, using 6-plex iTRAQ labeling combined with LC-MS/MS analysis. Two fractions iTRAQ sample sets were analyzed, including three independent biological replicates of cytoplasmic and nuclear fractions (Supporting Information Fig. 1 ). Based on the LC-MS/MS data, 1152 and 2117 proteins were identified from the cytoplasmic and nuclear cell fractions, respectively. Of the identified proteins, we reliably quantified 3211 proteins, of which 23 and 34 proteins were differentially abundant proteins in the cytoplasmic and nuclear fractions (fold changes ࣙ1.2 or ࣘ0.83) as a result of infection [26] , respectively (Table 1 , Fig. 2A ). However, of 57 identified proteins, 19 proteins were identified by one unique peptide from three biological replicates, the MS/MS spectrum for each of these proteins are included in Supporting Information Fig. 2 . In addition, these differentially abundant proteins were clustered hierarchically with Cluster 3.0 software (Fig. 2B ). Based on this analysis, these proteins were found to belong to two groups where the amount of protein increased and decreased at 24 hpi in the cytoplasmic and nuclear fractions, respectively.

In order to understand our proteomic profiles, the identified and quantified proteins were analyzed further by different bioinformatic tools. The numbers of proteins for different subcellular location were quantified using the gene ontology (GO) annotations (Fig. 2C) . It can be seen that cytoplasmic localization of proteins annotated was the largest proportion in the cytoplasmic fraction. Similarly, nuclear proteins were most abundant in the nuclear fraction. Additionally, some proteins showed cytoplasmic annotation using our dataset also appeared in the nuclear fraction. Thus, some proteins apparently had shifted between the nucleus and cytoplasm after IBDV infection. Simultaneously, to better understand the implications of host responses to IBDV infection, we analyzed the differentially abundant proteins with the "biological process" GO annotation. Shown in Fig. 3 , about half of the proteins are involved in regulation of cellular, metabolic, and biological processes. Based on the statistical analysis of cytoplasmic proteins, the cell proliferation, developmental process, multicellular organismal process, and locomotion were identified in the upregulated proteins (Fig. 3A, in red) , indicating that these processes were more active inside the cytoplasm after IBDV infection. Interestingly, these processes were found in the downregulated nuclear fraction (Fig. 3B, in red) , but not in the upregulated nuclear fraction. The above data showed a dynamic shift of differentially abundant proteins between the cytoplasm and nucleus.

We assessed the mRNA abundance of 25 identified proteins by quantitative real-time RT-PCR (Fig. 4) . These proteins had differentially abundant in MS/MS. The GAPDH mRNA transcript was used as a control housekeeping gene. In general, most of the change trends of 25 mRNA were similar to their proteins in iTRAQ analysis (Table 1) . Among these genes, the abundance of the innate immune associated genes IFIT5 [27] and STAT1 were evidently increased, while the genes eukaryotic translation initiation factor 4E binding protein 1 (EIF4EBP1), DEAD-box RNA helicase 1 (DDX1), and DEAD-box RNA helicase 42 (DDX42) were decreased. Transcriptional levels of chemokinetic and inflammatory associated genes c-c motif chemokine 4 homolog (CCL4) and FK506 Binding Protein (FKBP) were increased. Similarly, metabolic process and signaling associated genes hn-RNPH1, interleukin enhancer-binding factor 3 (ILF3), NPL4, and NF45 were also increased, but the transcription of antioxidant enzyme gene peroxiredoxin 4 (PRDX4) was decreased in abundance. Additionally, although the mRNA level of apoptosis inhibitor associated gene API5 was increased, that of the voltage-dependent anion channel 3 (VDAC3) gene was decreased. A variety of factors can affect gene expression. In this study, changes in mRNA of most genes were consistent with its corresponding protein abundance, which indicated these proteins may be regulated at the mRNA level.

To provide additional evidence to validate the differentially abundant proteins identified by the iTRAQ-labeled LC-MS/MS system, five proteins in both the cytoplasmic and nuclear fractions (IFIT5, hnRNPH1, NF45, API5, and STAT1) were detected by Western blot analysis (Fig. 5 ). GAPDH and histone 3 were used as the marker and loading controls of cytoplasmic and nuclear proteins, respectively. The significant increases of IFIT5 were observed in the cytoplasmic fractions of IBDV-infected cells. The abundance of the nuclear STAT1 protein was identified as increase in the nuclear fractions by MS analysis, and the Western blot results supported this observation. The API5 protein was shown to be nonsignificantly increased in abundance in the nucleus of IBDV-infected cells compared with mock-infected cells, which was consistent with the MS results. However, the intracytoplasmic API5 protein was significantly increased in abundance, which was not found in the MS analysis. Significant increases of NF45 and hnRNPH1 were observed in the cytoplasmic fraction of IBDV-infected cells. Interestingly, NF45 mostly appeared in the nucleus, but then shifted to the cytoplasm after IBDV infection. Generally, except for API5, the change trends in protein abundance of other four proteins were similar to the change patterns of their corresponding proteins in iTRAQ analysis.

Confocal microscopy can be used to monitor not only the subcellular localization of proteins, but also changes in protein abundance [28] . Thus, several quantitative proteomics studies have employed confocal microscopy to validate their MS results [25, 29] . In our study, colocalization analysis was performed by the double staining indirect immunofluorescence assay (IFA) analysis using the pAbs to NF45, hnRNPH1, DDX42, NPL4, and API5, and the mAb to VP3 protein of IBDV. In mock-infected cells (Supporting Information Fig. 3) , hnRNPH1, API5, and NPL4 were mainly expressed in the cytoplasm, while NF45 and DDX42 appeared in the nucleus. Distribution patterns of Figure 5 . Western blot analysis of selected differentially abundant proteins in cytoplasmic and nuclear fractions of IBDVinfected cells. The differentially abundant proteins NF45 was increased in the cytoplasmic fraction and translocated from nucleus to the cytoplasm upon virus infection. Equal amounts of protein fractions were loaded, and the abundance of IFIT5, hnRNPH1, NF45, API5, STAT1, and viral protein VP3 were determined using specific pAbs/mAbs. Mock-infected cells were used as a negative control. GAPDH and histone 3 were used as the marker and loading control for the cytoplasm and nucleus, respectively.

the above-mentioned cytoplasmic proteins in IBDV-infected cells were the same as those in mock-infected cells, but the fluorescence intensities of these proteins were increased in infected cells compared with mock-infected cells. In IBDV-infected cells at 24 h after inoculation, the VP3 viral protein signal overlapped with that of NF45, hnRNPH1, DDX42, NPL4, and API5 (Fig. 6) . Interestingly, NF45 and DDX42, which are normally expressed in the nucleus, were colocalized with the viral protein VP3 in the cytoplasm of IBDV-infected cells, indicating that NF45 and DDX42 were translocated from the nucleus to cytoplasm during infection. Moreover, the abundance of NF45 clearly increased and accumulated in the cytoplasm of IBDV-infected cells.

We constructed the pathways of differentially abundant proteins by canonical pathway analysis. For the pathway analysis, 11 canonical pathways were affected significantly during IBDV infection (p ࣘ 0.05). As shown in Fig. 7A , telomere extension by telomerase (p = 3.51 × 10 −4 , ratio 0.111) was the top priority in the cytoplasmic proteome, other favored the role of p14/p19ARF in tumor suppression (p = 1.43 × 10 −3 , ratio 0.057) and inosine-5 -phosphate biosynthesis II (p = 5.6 × 10 −3 , ratio 0.062). Meanwhile, the ratio in telomere extension by telomerase pathway showed the most significant among the ten canonical pathways. Additionally, the nuclear proteome (Fig. 7B ) tended to heavily favor eIF2 signaling as the primary pathway (p = 1.09 × 10 −4 , ratio 0.025), which is highly conserved and vitally important for antiviral responses [30] . Other significant associations included gluconeogenesis I, intrinsic prothrombin activation pathway, and mitochondrial dysfunction (Fig. 7B) .

In order to determine how IBDV interacts with the host proteins and how they affect cellular functions, we constructed virus-host interaction networks by IPA. Network maps of both cytoplasmic proteins and nuclear proteins were created ( Fig. 8A and B) . For the network analysis, cytoplasmic proteins involved in RNA posttranscriptional modification, molecular transport, and RNA trafficking were grouped together (Fig. 8A) . Most of the cytoplasmic proteins were linked to the mTOR signaling, ERK/MAPK signaling, and protein ubiquitination pathways. Additionally, in the nuclear fraction, proteins involved in DNA replication, repair, cellmediated immune response, and cellular development were grouped together (Fig. 8B) . The bioinformatic analysis suggests that the PI3K/Akt signaling pathway and IFN-␤ signaling cascades are vitally involved in IBDV infection. Certain proteins increased in abundance in infected cells are known to be products of IFNs, such as IFIT5 [27, 31] and STAT1 [32, 33] , which implies a specific antiviral response in the host cells. Due to limitations of the database or unknown functions of the host protein, some of the differentially abundant proteins were not included in this network.

Reliable quantitative methods are essential in biomarker discovery and to interpret proteomic data. In recent years, subcellular prefractionation has been employed as a popular technique to reduce sample complexity, extend protein coverage, and more critically, to provide a spatial description of proteins in the cell. Understanding virus-host interactions is critical for elucidating molecular mechanisms associated with functional alterations of IBDV-infected cells. Until now, no study had examined the subcellular proteome of IBDV-infected cells, although proteomic analyses of in vitro IBDV-infected CEFs [6] and in vivo IBDV-infected bursa of Fabricius [9] from our research group were previously reported. In the current work, we used a subcellular iTRAQ proteomic approach to characterize differentially abundant proteins in the cytoplasmic/nuclear fractions of IBDVinfected DF-1 cells. Comparative analysis of the results by MS revealed that in the cytoplasmic proteome of virus-infected cells, altered cellular proteins included 17 increased proteins Figure  6 . Immunofluorescence analysis of subcellular distribution of cellular NF45, hnRNPH1, DDX42, NPL4, and API5 with IBDV VP3 protein by confocal microscopy. IBDV-infected DF-1 cells reacted with anti-VP3 mAb and anti-hnRNPH1, anti-NF45, anti-API5, anti-NPL4, or anti-DDX42 pAbs as a primary antibody, followed by the FITC-conjugated (green) goat/rabbit IgG and Alexa Fluor 546 conjugated (red) mouse IgG. The fixed cell nuclei were stained with DAPI (blue).

in abundance and six decreased proteins in abundance. In the nuclear proteome, 34 altered cellular proteins were identified, including 14 increased proteins in abundance and 20 decreased proteins in abundance. Additionally, IFA showed the interplay between the viral protein VP3 and cellular protein NF45. Clearly, further studies are necessary to understand the roles of the altered cellular proteins in IBDV infection. Possible biological functions and pathways associated with these differentially abundant proteins were explored by the IPA package, and the identified cellular proteins were found to be mainly involved in immune system process, biological regulation, molecular transport, signaling, and metabolic process.

In present study, except for that two altered proteins were found identical in IBDV-infected CEF cells by 2DE and iTRAQ, we also found that most of the differentially increased proteins in abundance were different in comparison with our previous research reports using 2DE [6, 9] . We identified more cytosolic and nuclear proteins with iTRAQ-based quantitative subcellular proteomics. Proteins not identified in previous 2DE proteomics included IFIT5, NF45, STAT1, and API5 involved in immune response and cellular apoptosis. Apart from this, RNA processing and energy metabolism related proteins hnRNP H1, DDX42, and DDX1 also were supplemented from our data. Generally, the iTRAQ method provided the finer details of the subcellular proteome of IBDV-infected cells, and more information for different compartments of host cells. To our knowledge, this study is the first to report a subcellular proteomic analysis of the response to IBDV infection.

IFIT genes encode a family of proteins that are induced after IFN treatment, viral infection, or PAMP recognition [34] . IFIT proteins are poised to confer inhibitory effects after infection, and they have been found to be increased in abundance in target cells infected with various viruses, such as H3N2 swine influenza virus [11] , Japanese encephalitis virus (JEV) [35] , and porcine reproductive and respiratory syndrome virus [16] . Recently, progress was made in identifying how IFIT proteins inhibit through distinct mechanisms of action, and the replication of multiple families of viruses [36] , such as hepatitis C virus [37] , influenza A virus, and vesicular stomatitis virus [38] . One of the IFIT family genes, IFIT5 (ISG58), is involved in anti-viral responses through enhancing innate immune signaling pathways [27] . In the present study, IFIT5 was revealed to be increased in abundance during IBDV infection (increased ß2.001-fold in the cytoplasmic fraction). Therefore, it is conceivable that IFIT5 may be an important modulator in the antivial innate immune response during IBDV infection. STAT1 is a downstream target of the IFN pathway, which is central to the regulation of immune responses in viral invasion. Type I IFN binds to a common IFN-␣/␤ receptor that initiates a signaling cascade that results in the phosphorylation and nuclear translocation of STAT1 and the induction of the expression of hundreds of IFN-stimulated genes [33, 39] . In the present study, STAT1 was notably upregulated during IBDV infection. However, the upregulated mechanism of STAT1 needs further investigation in IBDV-infected DF-1 cells.

Our analysis found increased abundance of the RNA processing and translation-associated gene hnRNPH1 in the cytoplasmic fraction of IBDV-infected cells. As RNA binding proteins, hnRNPs are associated with pre-mRNAs in the nucleus and appear to influence pre-mRNA processing and other aspects of mRNA metabolism and transport [40] [41] [42] . Previous reports have shown that poliovirus infection cause redistribution of the cellular hnRNPs A1 and K from the nucleus to the cytoplasm [43] , and hnRNP U is colocalized with vesicular stomatitis virus [44] . HnRNP H1 has also been implicated in assisting hepatitis C virus replication [41] . Moreover, confocal microscopy analysis showed colocalization of hnRNP H1 protein with the viral protein VP3 in the cytoplasm of IBDV-infected cells (Fig. 6B ). VP3 has multiple functions that contribute to efficient virus replication [45] . Therefore, we hypothesized that the increased abundance of hnRNP H1 may play an important role in the replication of IBDV. Among the differentially abundant proteins, we also found decrease in abundance of cellular proteins associated with RNA processing and biosynthesis, including 60S ribosomal protein l5 (RPL5, for RNA binding) and malate dehydrogenase 2 (MDH2, operates the metabolic coordination between cytosol and mitochondria). Notably the proteomic data further revealed that the key glycolytic enzymes, including aldolase C (ALDOC), were extensively decreased in abundance. Based on these proteomic analyses, we speculate that IBDV replication may extensively inhibit the host cellular metabolic pathways involved in glycolysis and energy metabolism as well as mRNA processing.

We also found differentially abundant proteins associated with DEAD-box RNA helicases DDX1 and DDX42 (Fig. 6) , which are involved in many aspects of RNA metabolism. Additionally, DDX1 was shown to be an RNAactivated ATPase, wherein Rev-bound RNA was equally effective at stimulating ATPase activity as protein-free RNA. DDX1 interacted with HIV-1 Rev and involved in virus replication in Hela cells [46] . Moreover, DDX1 interacted with NF-B subunit RelA (p65) and took part in regulating NF-B-mediated regulatory pathway [47] . Futhermore, the study found that DDX1 interacted with nsp14 from both coronavirus infectious bronchitis virus (IBV) and severe acute respiratory syndrome coronavirus (SARS-CoV). Intranuclear DDX1 was translocated to cytoplasm in IBV-infected Vero cells. In other words, DDX1 was conducive to enhance coronavirus replication [48] . ATP-dependent RNA helicase DDX42 interacted with JEV NS4A in vitro and in vivo. Importantly, the expression of DDX42 inhibited the confrontation of the response to IFN-␤ by JEV [49] . Together, these results provided insight into the understanding of IBV pathogenesis.

NF45, also known as ILF2, has been indicated to act as an important regulatory factor involved in transcription, microRNA processing, nuclear export of RNAs, and posttranscriptional and translational regulation of gene expression [50] [51] [52] . NF45/NF90 heterodimeric protein complex have involved multitudinous functions, one of the functions was transferred from the nuclear to the cytoplasm in many different cell types [53] . In the infected DF-1 cells, we found predominantly the NF45 protein in the cytoplasmic fraction (Fig. 6 ). In addition, we also found that NF45 colocalized with VP3 and translocated from the nucleus to cytoplasm during IBDV infection (Fig. 6) . This observation was similar to a previous observation of IBDV infection at 12 hpi [54] and consistent with the conclusion that the cytoplasmic accumulation of NF45 as a consequence of the viral infection occurs independently from the expression of IFN-regulated genes.

Some proteins associated with cell death and apoptosis, such as API5, was identified as being altered by IBDV infection. API5 is an antiapoptotic protein that is increased in abundance in various cancer cells, the roles of API5 may be to mediate protein-protein interactions [55] . Morris et al. [56] reported that API5 play an important role in dE2F1-induced apoptosis. Proviral integration of moloney virus-2 can activate API5 through phosphorylation mechanism in order to inhibit the apoptosis of liver cells, and NF-B is the key regulator in this antiapoptotic pathway [57] . The antioxidant stress protein PRDX4 was decreased in abundance, suggesting that IBDVinfected inhibits host cell oxidative stress. Virus-induced oxidative stress is associated with the activation of phagocytosis and the release of reactive oxygen that positively modulate the immune activation and eradicate viral infection and immuneinduced cellular injury [58] . Moreover, Tavender et al. [59] found that PRDX4, an endoplasmic reticulum localized peroxiredoxin and known to play an important role in disulfide formation, can be associated with the endoplasmic reticulum stress response.

In addition, among the differentially abundant proteins, we identified nucleolar protein nucleophosmin 1 (NPM1) in the cytoplasmic fraction. NPM1 has been linked to a variety of important cellular processes, both in and out of the nucleolus, including ribosome processing, molecular chaperoning, maintenance of genomic integrity, centrosome duplication, and transcriptional regulation [60, 61] . It has been reported to function as a chaperone in the viral chromatin assembly process in infected cells [62] . This shuttling of proteins between the nucleus and cytoplasm is now recognized as a key mechanism for ensuring proper cell cycle progression [63] . In previous report, others have identified NPM as a novel p53independent target of the ARF tumor suppressor protein [64] . Our investigation demonstrated an increased level of NPM1 in IBDV-infected DF-1 cells, which can be hypothesized as an attempt by the host to arrest viral replication inside the cells.

As demonstrated in this study, we first use iTRAQ subcellular proteomic analysis to understand DF-1 cell responses of IBDV infection. Using accurate methods, a total of 57 cellular proteins that were significantly altered postinfection were identified. These data provide clues to elucidate the molecular mechanisms of the interaction between IBDV and target cells.

Birnaviridae

The MS proteomics data have been deposited to the Pro-teomeXchange Consortium [65]