Tandem mass tag-based quantitative proteomic analysis of lycorine treatment in highly pathogenic avian influenza H5N1 virus infection

Virus, cells, and lycorine

GD178 (N0. AY737296-737300) and MDCK cells were obtained from the Key Laboratory of Veterinary Vaccine Innovation of the Ministry of Agriculture, P. R. China. The MDCK cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Invitrogen) containing 10% (v/v) fetal bovine serum (FBS, Gibco). A plaque assay was used to determine viral activity in MDCK cells (He et al., 2013). Three separate biological cell cultures were made for the tandem mass tag (TMT) proteomic experiments. Lycorine (Fig. 1A) was obtained as previously described (Wang et al., 2009). All experiments involving live H5N1 influenza virus were carried out in biosafety level-3 facilities.

TMT assay design and protein preparation

Three groups were designed: (1) M, mock group (control); (2) V, virus-infected group. They were then inoculated at 3 multiplicity of infection (MOI) for 1 h (V), removed, and inoculated with DMEM containing 10% FBS until 12 h post-infection (h.p.i.); and (3) L, lycorine-treated post-viral infected group. After 1 h adsorption of 3 MOI GD178, lycorine was added at 0.52 µM for 12 h.p.i. For each sample, the proteins were diluted in 300 µL of SDT buffer (containing 4% SDS, 100 mM Tris-HCl, and 100 mM DTT, pH 8.0) following a standardized protocol (Wisniewski et al., 2009), homogenized, heated at 100 °C for 3 min, and then centrifuged at 14,000 × g for 30 min at room temperature. Each sample (one µL) was taken and quantified using the bicinchoninic acid method. The remaining lysate was frozen at −80 °C until use.

Trypsin digestion

Protein digestion was conducted using the FASP procedure (Wisniewski et al., 2009). In brief, 300 mg of proteins were loaded onto an ultrafiltration filter (30 kDa cutoff; Sartorius, Goettingen, Germany) containing 200 µL of UA buffer (8 M urea, 150 mM Tris-HCl, pH 8.0), centrifuged at 14,000× g for 30 min, and then washed with 200 µL of UA buffer. Then, 100 µL of 50 mM iodoacetamide in UA buffer was subsequently added to the filter to block reduced cysteine residues. The samples were incubated for 30 min at room temperature in the dark and then centrifuged at 14,000× g for 30 min. The filters were washed three times with 100 µL of UA buffer and then centrifuged at 14,000× g for 30 min after each washing step. Next, 100 µL of dissolution buffer (Applied Biosystems, Foster City, CA, USA) was added to each filter and then centrifuged at 14,000× g for 30 min, which was repeated three times. The protein suspensions were then digested with 40 µL of trypsin (Promega, Madison, WI, USA) buffer (6 µg trypsin in 40 µL of dissolution buffer) at 37 °C for 18 h. Finally, the filter unit was transferred to a new tube, and 40 µL of dissolution buffer was added followed by centrifugation at 14,000× g for 30 min. The resulting peptides were collected as a filtrate, and the peptide concentration was analyzed at OD280.

Off-Line High-pH Reversed-Phase Fractionation

Nine serial samples from each subject and the common reference sample were included in one TMT10 labeling experiment set as shown in Fig. 1C. This labeling strategy avoids the potential for missing value issues resulting from data-dependent MS/MS acquisitions (Zimmer et al., 2006). TMT-labeled peptides were subjected to High-pH Reversed-Phase Fractionation in 1100 Series HPLC Value System (Agilent) equipped with a Gemini-NX (Phenomenex, 00F-4453-E0) column (4.6 × 150 mm, 3 µm, 110 Å) (Wang et al., 2011; Batth, Francavilla & Olsen, 2014). The peptides were eluted at a flow rate of 0.8 mL/min. Buffer A consisted of 10 mM ammonium acetate, pH 10.0, and Buffer B consisted of 10 mM ammonium acetate, 90% v/v acetonitrile (ACN), pH 10.0. Both buffers were filter sterilized. The gradient to perform the separation consisted of several steps: (1) 100% Buffer A for 40 min, 0%–5% Buffer B for 3 min; (2) 5%–35% Buffer B for 30 min; (3) 35%–70% Buffer B for 10 min; (4) 70%–75% Buffer B for 10 min; (5) 75%–100% Buffer B for 7 min; (6) 100% Buffer B for 15 min; and (7) 100% Buffer A for 15 min. The elution process was monitored by measuring absorbance at 214 nm, and fractions were collected every 1.25 min. The collected fractions (approximately 40) were finally combined into 10 pools. Each fraction was concentrated via vacuum centrifugation and reconstituted in 10 µL of 0.1% v/v formic acid. All samples were stored at −80 °C until LC-MS/MS analysis.

Quantitative LC- MS/MS

The labeled samples were analyzed using the Easy-nLC nanoflow HPLC system connected to the Thermo Scientific™ Orbitrap Fusion™ Tribrid™ mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA). A total of 1 µg of each sample was loaded onto the Thermo Scientific EASY column (two columns) using an autosampler at a flow rate of 200 nL/min. Sequential separation of peptides on the Thermo Scientific EASY trap column (100 µm × 2 cm, 5 µm, 100 Å, C18) and analytical column (75 µm × 25 cm, 5 µm, 100 Å, C18) was accomplished using a segmented 60 min gradient from 5% to 28% Solvent B (0.1% formic acid in 100% ACN) for 40 min followed by 28%–90% Solvent B for 2 min and then 90% Solvent B for 18 min. The column was re-equilibrated to its initial highly aqueous solvent composition before each analysis. The mass spectrometer was operated in a positive ion mode, and MS spectra were acquired over a range of 375–1,500 m/z. The resolving powers of the MS and MS/MS scans at 200 m/z for the fusion were set as 120,000 and 50,000, respectively. The data-dependent mode was top speed, cycle time was 3 s, and ions were fragmented through high energy collisional dissociation. The maximum ion injection times were set at 50 ms for the survey scan and 105 ms for the MS/MS scans, and the automatic gain control target value was set to 4e5 for MS and to 1e5 for MS/MS. The dynamic exclusion duration was 40s.

Database search

All raw files were analyzed using the Proteome Discoverer 2.1 software (Thermo Fisher Scientific). A search for fragmentation spectra was performed using the MASCOT search engine (Matrix Science, London, UK; Version 2.2) embedded in Proteome Discoverer against the Uniprot Canidae protein sequence database (released in November 2016, 31,596 sequences). Several search parameters were used: (1) monoisotopic mass; (2) trypsin as the cleavage enzyme; (3) two missed cleavages; (4) TMT labeling; (5) cysteine carbamidomethylation as fixed modifications; (6) peptide charges of 2+, 3+, and 4+; and (7) methionine oxidation. The mass tolerance was set to 20 ppm for precursor ions and to 0.1 Da for the fragment ions. Peptide spectral matches were filtered to a 1% false discovery rate. The relative quantitative analysis of the sample proteins was based on TMT reporter ion ratios from all unique peptides representing each protein. This analysis was performed using Proteome Discoverer (version 2.1). The relative peak intensities of the TMT reporter ions released in each of the MS/MS spectra were used. The final ratios obtained from the relative protein quantifications were normalized based on the median average protein quantification ratio. Only unique peptides obtained with a confidence percentage of >95% were included in the ratio ≥1.20, or ≤0.8333-fold cutoff value was used to identify upregulated and downregulated proteins with p < 0.05. The mass spectrometry proteomic data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD lycorine (accession number PXD012936). Protein-protein interaction (PPI) analysis of DEP proteins STRING (Search Tool for the Retrieval of Interacting Genes/Proteins) version 11.0 was employed in this study for the potential PPI analysis of the DEPs proteins (http://string-db.org/) (Szklarczyk et al., 2015; Shannon et al., 2003). The parameter for confidence score was set to 0.4, and the yield PPI results were visualized by Cytoscape software (Shannon et al., 2003). The light blue nodes indicated upregulated-proteins and the light green nodes indicated downregulated-proteins. The edge thickness was proportional to the combined score of the proteins.

Bioinformatics and enrichment analysis

Bioinformatics and enrichment analysis was performed as previously described (Hui Bin Huang, 2017).

Western blot

Cells were homogenized on ice for 20 min in 1 × RIPA buffer with cocktail protease inhibitors (Roche) and then centrifuged at 15,000× g for 20 min at 4 °C. The supernatant was collected and then stored at −80 °C. For western blot analysis, equal amounts of proteins were separated using SDS-PAGE and subsequently transferred onto a polyvinyl fluoride membrane (Bio-Rad). The membrane was blocked with 5% BSA and then incubated with monoclonal antibody Nup93 (ab168805) and monoclonal antibody anti- β-actin (Cell Signaling Technologies) at 4 °C overnight. Membranes were washed in phosphate-buffered saline with Tween20 and then incubated for 2 h at RT with goat anti-rabbit horseradish peroxidase-conjugated secondary antibody (1:2,000; Cell Signaling Technologies).

Statistical analyses

Statistical analyses were conducted as previously described (Hui Bin Huang, Zhang & Li, 2017).

Conditions for lycorine treatment, viral infection, and group design

The concentration and exposure time of lycorine were determined to avoid any cytotoxic effects on the host cells and to ensure that the proteome of the host cells reflected the true response to the presence of lycorine. A prior study found through CCK-8 assay that 0.52 µM lycorine is not toxic to cells at 12 h post-infection (h.p.i.) (He et al., 2013). In addition, the time of lycorine addition was optimized. Results showed that the earlier lycorine was added, the better effect on viral titers. Therefore, a single replication cycle of influenza virus (12 h.p.i.) within lycorine treatment at 0.52 µM was chosen because it reflects a single viral replication and is a non-toxic concentration of lycorine at the time addition assay. Dose–effect relationship experiments showed that inhibition of lycorine of AIV becomes invalid within high-dose virus (Jun He et al., 2012; He et al., 2013). Hence, we ensured that the response of the whole cell proteome is caused by early infection and that the activity of lycorine is effective. Thus, 3 MOI was selected as the minimum infectious dose for a single cycle of viral infection for studying the single viral replication cycle (Peschel et al., 2013). Therefore, 3 MOI of AIV infections, lycorine concentration of 0.52 µM, and 12 h of incubation time were selected as the optimal dose and exposure time for further experiments.

In this study, we applied a TMT method, which has been widely applied for quantitative proteomic cell biology studies and many model organisms, to explore potent anti-viral agents and different signaling pathways (Lombardi et al., 2015; Burton et al., 2016). A schematic of the experimental workflow and Venn diagram of V/M, L/V, and L/M is shown in Fig. 1C. Each experiment group consisted of three biological replicates (n = 3). In total, 5,786 proteins were identified and relatively quantified at 0.01% FDR in the proteome using MaxQuant (Table S1). Among the statistically significant proteins detected by ANOVA (p < 0.05), protein abundances that changed <1.2-fold, >0.8333-fold, and p > 0.05 were discarded. The proteome profiles of mock control and infected cells were compared (V/M) to determine the effects of AIV infection on MDCK cells. Among those protein groups, 1101 proteins were identified, of which 340 DEPs were determined in the V/M group. The proteome profiles of the infected cells and lycorine-treated infected cells were compared to determine the effects of lycorine treatment on the infected cells. A total of 1059 proteins were identified in the V/L group, of which 258 were DEPs. These results indicated wide differences of protein expression in the HPAIV infection and lycorine treatment. Inoculating with HPAIV and lycorine resulted in 691 identified DEPs (76 + 86 + 42 + 54 + 189 + 159 + 85) among the three groups of MDCK cells in response to HPAIV infection and lycorine treatment, as shown in the Venn diagram (Fig. 1B).

H5N1 AIV-induced cellular protein content changes in MDCK cells

The proteome profiles of the GD178-treated MDCK cells and mock control were compared (V/M) to determine the effects of GD178 influenza virus on MDCK cells. The V/M group contained 123 upregulated (>1.2-fold) and 217 downregulated (to <0.8333-fold) proteins in the V/M group (Table S2). Some of the proteins in the same network were direct interaction or through intermediate partner at the PPI level, as shown in Fig. S1. Gene ontology analyses of the proteins upregulated and downregulated by GD178 were performed to map the genes involved in different cellular processes, including biological and functional events (Fig. 2A). A large number of biological process GO terms were identified. The DEPs were strongly represented by “catabolism, metabolism, and physiological process”; the infected cells were also assigned to numerous molecular functions, of which “binding, catalytic and molecular transducer activities” were the main ones, and cellular components, of which “extracellular region part, organelle and membrane” were dominant. GO enrichment analysis was conducted on cell periphery (90 proteins), plasma membrane (88 proteins), intrinsic component of membrane (59 proteins) and molecular transducer activity (39 proteins) to further study the impact of DEPs in cell physiological processes and discover their internal relations (Fig. 2B). Several identified proteins have known interactions with one another and function in the same biochemical pathways. Therefore, a KEGG pathway-based enrichment analysis was applied to identify the main pathways that were potentially affected by the DEPs in the V/M group (Fig. 2C). The enriched pathways showed that the proteins were involved in pathway in cancer (path: ko05200), extracellular matrix (ECM)–receptor interaction (path: ko04512), focal adhesion (path: ko04510), mitogen-activated protein kinase signaling pathway (path: ko04010), and PI3K-Akt signaling pathway (path: ko04151). These data indicated that there are many changes in the protein profile in response to influenza virus infection at 12 h.p.i.

Differentially expressed proteins in lycorine-treated and virus-infected cells

The proteome profiles of the infected cells and the infected cells within lycorine treatment were compared (V/L) to determine the effects of lycorine treatment on the infected cells. In total, 258 DEPs were found (Table S3). The PPI network was performed to determine whether these DEPs interact with each other and form protein complexes (Fig. S2). Two-hundred six proteins with at least 1.2-fold upregulation between the lycorine-treated and virus control group were found (V/L), and 52 proteins were significantly downregulated. The GO terms of DEPs were strongly represented by “cellular process, single-organism process, and biological regulation” in the biological process and “binding, catalytic and molecular transducer activities” in molecular functions; the lycorine-treated cells were also assigned to numerous cellular components, of which “membrane-enclosed lumen, organelle and membrane” were dominant (Fig. 3A). GO enrichment analysis showed that the three main parts were nuclear division (15 proteins), chromosome segregation (14 proteins), and condensed chromosome (eight proteins) (Fig. 3B). The functional classification of DEPs was conducted by KEGG enrichment analysis, and each protein was assigned to at least one of the following pathways: human T-lymphotropic virus-1 infection pathway (path: ko05166), cell adhesion molecules (CAMs) (path: ko04514), epidermal growth factor receptor tyrosine kinase inhibitor resistance (path: ko01521), Janus kinase-STAT signaling pathway (path: ko04630), and pancreatic cancer (path: ko05212) (Fig. 3C). These results showed differences between V/M and V/L, and further analyses of these genes that were co-changed in the two groups may shed light on the antiviral mechanism of lycorine.

Characterization of proteins in the same trend in the V/M and V/L groups

The proteins co-regulated in V/M and V/L were selected for further analysis. These 71 DEPs were investigated and annotated using GO terms and subjected to GO functional analysis (Table S4). Fifty-four proteins were upregulated in the V group but downregulated in the L group; this finding indicated that some of the proteins may participate in viral infections and may be blocked by lycorine to avoid progeny virus budding. Seventeen proteins were downregulated in the V group but induced in the L group, implying that these proteins were downregulated after influenza virus infection but increased after lycorine treatment. A total of 71 proteins showed upregulated or downregulated difference in abundance, and a heat map of these proteins was obtained (Figs. 4A and 4C). All co-upregulated and co-downregulated proteins were further analyzed by GO enrichment analysis, and each protein belonged to at least one term. Among the upregulated proteins, the three main terms in the BP category were cellular process (40 proteins), metabolic process (29 proteins), and organic substance metabolic process (29 proteins). The main terms in the MF category were transcription factor activity (42 proteins) and nucleic acid binding transcription factor activity (33 proteins). The main terms in the CC category were membrane part (49 proteins) and cell junction (41 proteins). In the downregulated proteins, the main terms in the BP category were localization (15 proteins), immune system process (14 proteins), and reproduction (12 proteins). The main terms in the MF category were structural molecule activity (15 proteins) and binding (11 proteins). The main terms in the CC category were extracellular region (16 proteins) and membrane part (16 proteins) (Figs. 4C and 4D). The analysis suggested that these biological functions may be affected by the treatment of AIV-infected MDCK cells with lycorine.

In addition, pathway analysis showed that these proteins were involved in viral replication-related pathways. For instance, oligoadenylate cyclase (Oasl1), serine/threonine protein kinase (STK10), and other proteins (8/71) were involved in adenosine and guanine triphosphatase (ATP and GTP) activity-related pathways; other pathways such as Nup93, mitochondrial translational initiation factor (MTIF3), and other proteins (9/71) participated in host cell responses to lycorine-related transcription or translation processes, such as cellular responses to stimulus and oxoacid-related metabolic processes. This may or may not be related to the antiviral properties of lycorine.

Validation of protein profiles and Western blot of DEPs

According to our prior study, vRNP is retained in the nucleus under conditions of lycorine treatment to stop the cycle of influenza virus, which leads us to search for the proteins involved in the vRNP export pathway (He et al., 2013). For several pathways relevant to proteins targeting nuclear export were enriched in the “biological process” category. The proteins with the same trend in the V/M and V/L groups were our candidate proteins. Fortunately, a gene directly related to nuclear transport, Nup93, was upregulated by 1.29-fold, in contrast to the low expression level (0.8) found in lycorine treatment over controls (Table S2). From the GO terms, Nup93 is related with export of material from the nucleus, such as mRNA, tRNA, viral material, and viral transcription (Table S4). Given its possible role in viral inhibition (Muhlbauer et al., 2015), western blotting assay with specific antibodies was performed to determine if the expression regulation for these proteins is at the protein level. As shown in Fig. 5, the trends in Nup93 level changes in the infected cells and lycorine treatment groups were similar to the changed patterns as was observed. Moreover, the protein levels of Nup93 in the uninfected cells but treated with the same concentrations of lycorine were not changed as such as the cells treated with lycorine after influenza virus infection. These results indicate that Nup93 expression was induced after influenza virus infection but was dramatically decreased after lycorine treatment at 0.52 µM. This concentration did not exhibit toxicity to the uninfected cells, which was in accordance with the TMT results. As a result, AIV infection may induce Nup93 to complete the viral cycle, and the process of protein targeting into Nup93 after lycorine treatment may partly be explained by the blockage of vRNPs in the host cellular nucleus.