Identification of Differentially Expressed Proteins in Porcine Alveolar Macrophages Infected with Virulent/Attenuated Strains of Porcine Reproductive and Respiratory Syndrome Virus

The highly pathogenic porcine reproductive and respiratory syndrome virus (HP-PRRSV) is still a serious threat to the swine industry. However, the pathogenic mechanism of HP-PRRSV remains unclear. We infected host porcine alveolar macrophages (PAMs) with the virulent HuN4 strain and the attenuated HuN4-F112 strain and then utilized fluorescent two-dimensional difference gel electrophoresis (2D-DIGE) to screen for intracellular proteins that were differentially expressed in host cells infected with the two strains. There were 153 proteins with significant different expression (P<0.01) observed, 42 of which were subjected to mass spectrometry, and 24 proteins were identified. PAM cells infected with the virulent strain showed upregulated expression of pyruvate kinase M2 (PKM2), heat shock protein beta-1 (HSPB1), and proteasome subunit alpha type 6 (PSMA6), which were downregulated in cells infected with the attenuated strain. The upregulation of PKM2 provides sufficient energy for viral replication, and the upregulation of HSPB1 inhibits host cell apoptosis and therefore facilitates mass replication of the virulent strain, while the upregulation of PSMA6 facilitates the evasion of immune surveillance by the virus. Studying on those molecules mentioned above may be able to help us to understand some unrevealed details of HP-PRRSV infection, and then help us to decrease its threat to the swine industry in the future.


Introduction
Porcine reproductive and respiratory syndrome (PRRS) is an infectious disease characterized by reproductive failure in female pigs and respiratory tract illness in young pigs [1,2]. The highly pathogenic porcine reproductive and respiratory syndrome (HP-PRRS) that recently emerged in China is associated with additional symptoms, such as a high fever in infected pigs, high morbidity, and high mortality [3][4][5]. The pathogen causing the syndrome is the porcine reproductive and respiratory syndrome virus (PRRSV), which is an enveloped, non-segmented, singlestranded positive-sense RNA virus belonging to the family Arteriviridae, order Nidovirales [6,7]. Based on antigenic and genetic differences, PRRSV isolates are divided into two distinct genotypes: the European type and the North American type [8]. The pandemic PRRSV strains in China primarily correspond to the North American type. PRRSV exhibits strict cell tropism, and porcine alveolar macrophages (PAMs) are the main target cells that become infected in vivo. PRRSV can replicate in infected PAMs [9][10][11] and cause changes in the morphology and function of PAM cells [12,13].
In recent years, many new discoveries have been made concerning host cell changes at the molecular level that are caused by PRRSV infection. Technologies such as gene chips have been utilized to investigate changes in the host transcriptome following PRRSV infection, and it was found that PRRSV infection of host cells could cause changes in a large number of host genes at the transcriptional level [14][15][16][17]. Because there are certain differences between changes at the mRNA level and the actual protein expression level, several proteomic technologies have also been applied in the investigation of PRRSV infections. Zhang et al [18] utilized the JXwn06 strain of the virus (HP-PRRSV) to infect PAM cells in vitro and identified 23 differentially expressed proteins using two-dimensional electrophoresis. Ding et al. [19] utilized the JL/07/SW strain (HP-PRRSV) to infect PAM cells in vitro and detected 29 differentially expressed proteins. Xiao et al. [20] identified 45 differentially expressed proteins using the virulent Chinese strain and the attenuated North American strain to infect pigs in vivo and then compared the proteomic changes induced by the virulent/attenuated strain in lung tissue. However, the specific proteins involved in viral replication in these infections remain unknown. In addition, although the virulent HP-PRRSV strain that is pandemic in China and the classic North American attenuated strain both belong to the North American type of the virus, there are significant genomic sequence differences between the two strains. In particular, the highly pathogenic virulent strain carries deletions of 1 and 29 amino acids (aa) in the Nsp2 gene [4,5], which are unrelated to the highly virulence [21]. The evolutionary history of the genetic backgrounds of the two strains remains unclear, which makes it difficult to explore the differences between PRRSV virulent and attenuated strains.
In contrast, the virulent HuN4 strain of HP-PRRSV isolated in our laboratory can cause disease and death in infected pigs [3,5], and the continuous passage of this strain generated the attenuated vaccine strain HuN4-F112, which can provide protective immunity in pigs. This pair of virulent/attenuated strains has a relatively clear genetic background [22,23], thus providing better experimental material for our investigation of the pathogenic differences between HP-PRRSV virulent/attenuated strains at the cellular level using two-dimensional difference gel electrophoresis (2D-DIGE) technology. This study employed this pair of virulent/ attenuated strains to infect PAM cells to identify differentially expressed proteins. We annotated the Gene Ontology (GO) of the differentially expressed proteins, built a protein-protein interaction network through protein interaction network analysis, and, finally, validated the differentially expressed proteins via real-time PCR.

Ethics statement
The animal studies were approved by the Animal Care and Use Committee of Shanghai Veterinary Research Institute, CAAS.

Virus and experimental pigs
Both the virulent HP-PRRSV strain HuN4 and the attenuated vaccine strain HuN4-F112 were isolated, passaged, and preserved in our laboratory [3,5,23]. Newborn piglets were purchased from the Experimental Animal Center, School of Agriculture and biology, Shanghai Jiao Tong University. The ethics committee of the college approved the study.

Preparation of PAMs
Healthy newborn piglets were selected for isolation of PAMs based on a previously reported method [11]. PAM cells harvested from wash fluid were cultured in RPMI-1640 medium (Gibco Invitrogen, Shanghai, China) containing 10% fetal bovine serum (FBS) (Gibco), 100 units/mL penicillin, and 100 mg/mL streptomycin (Gibco) in cell incubators at 37uC with 5% CO 2 .

Virus inoculation and sample preparation
When the PAM cell density exceeded 95% confluence, the cells were washed three times with phosphate buffered saline (PBS) and 2610 6 PAM cells were then inoculated with 10 3 TCID 50 (50% tissue culture infective dose) of either the HuN4 strain or the HuN4-F112 strain. PAM cells without virus were used as the control. After 1 h of infection, the culture medium was discarded and replaced with RPMI-1640 medium containing 2% FBS. The cell culture was continued, and periodic observations of the cytopathic effects were performed. The culture medium was discarded 48 h after inoculation. After gently washing the culture plate three times with pre-cooled sterilized PBS (pH 7.2-7.4) at 4uC, the PBS was discarded and the culture plate was placed on ice. Lysis buffer (pH 8.5) containing 7 M urea, 2 M thiourea, 65 mM Tris, 2% dithiotreitol (DTT), 4% 3-[(3-cholamidopropyl) dimethylammonio] propanesulfonate (CHAPS), 0.2% IPG buffer (GE Healthcare, Munich, Germany), and 0.1% v/v protease inhibitor mixture was added into each well. All the cells were scraped into a 1.5 mL centrifuge tube and homogenized with a Dounce's homogenizer. The cells were disrupted by sonication (80W, five times for a 10 s interval each time at 15 s intervals). All procedures were performed on ice. The mixture was centrifuged at 15,000 g for 45 min at 4uC. Protein content in the supernatants was determined by Bradford assays (Bio-Rad, USA). Samples of aliquots were stored at 280uC until use for proteomic analysis.

Protein labeling
Three protein samples were labeled using the CyDye DIGE Fluor Minimal Labeling Kit (GE Healthcare, Munich, Germany). The dyes were designed to ensure that a given protein originating from different samples would have the same relative mobility regardless of the dye used to tag them. Fifty micrograms of an internal standard (IS) containing an equal amount of three protein samples collected at 48 h post-inoculation (hpi) were labeled with 400 pmol Cy2. And fifty micrograms protein sample of group A, B, and C were labeled with 400 pmol Cy3 or Cy5 using the CyDye DIGE Fluor Minimal Labeling Kit (GE Healthcare, Munich, Germany) according to the manufacturer's instructions. Details of the experimental design using the three-dye approach is illustrated in Table 1. Briefly, each CyDye minimal dye was redissolved in fresh N, N-dimethylformamide (DMF). To label proteins, 400 pmol of fluorescent dye was used to label 50 mg of protein at a pH of 8.5. The labeling reaction was carried out on ice for 40 min in the dark. Termination of the reaction was also carried out in the dark via reaction with 1 mL of 10 mM lysine for 10 min.

2D-DIGE
For each gel, Cy2-, Cy3-, and Cy5-labeled proteins (50 mg each) were combined and an equal volume of rehydration buffer (8 M urea, 4% w/v CHAPS, 130 mM DTT, and 2% v/v Pharmalyte TM pH 3-10) was added. The pooled protein samples were subjected to isoelectric focusing carried out on nonlinear IPG strips, length 13 cm, pH 3-10 (GE Healthcare), rehydrated at 30 V for 12 h at room temperature. Isoelectric focusing was conducted at 500 V for 1 h, followed by 1000 V for 1 h, then 8000 V for 3 h and held at 8000 V to reach a total of 40000 Vh at 20uC and a maximum current setting of 50 mA per strip using Ettan IPG-phor apparatus (GE Healthcare). After IEF, individual strips were incubated in equilibration buffer (50 mM Tris-HCl, 6 M urea, 30% glycerol, 2% SDS) supplemented with 1% DTT.

Gel image acquisition and analysis
The gel images for analysis were obtained by using the Typhoon 9700 Imager (Amersham Bioscience) and were processed in DeCyder 6.5 differential analysis software (Amersham Bioscience). The spots on the gels were codetected automatically as 2D DIGE image pairs, which intrinsically links a sample to its in-gel standard. Matching between gels was performed utilizing the ingel standard from each image pair. The experimental setup and relationship between samples were assigned in DeCyder software. Each individual Cy3 or Cy5 gel image was assigned an experimental condition, and all Cy2 images were classified as standards. The gel with the highest spot count was considered the master gel. Statistical analysis was carried out for every matchedspot set, comparing the average and standard deviation of protein   Table 3. Equal amounts of total protein from infected and uninfected whole cell lysates were resolved by 2D-DIGE. doi:10.1371/journal.pone.0085767.g001 abundance for each spot between HuN4 and HuN4-F112 or between HuN4-F112 and PAM using Student's t-test, and the comparison among the three groups was carried out using ANOVA. The procedure was performed using the DeCyder DIA (Difference In-gel Analysis) and the DeCyder Biological Variation Analysisb (BVA) software module. Protein spots with significant differences in abundance (more than 1.2-fold, P,0.01) were selected in the stained preparative gels for spot selection.

Mass spectrometry (MS) identification
Spots of interest from the preparative gels were manually excised. The gel samples were placed in a tube and were washed twice with 500 mL and 250 mL ddH 2 O for 15 min. For trypsin digestion, the gel samples were washed twice with 50 mM w/v NH 4 HCO 3 and covered with 0.7 mL Porcine Trypsin solution (Promega, Madison, WI, USA) in 50 mM w/v NH 4 HCO 3 . After incubation overnight at 37uC, the supernatant was transferred to a second tube and 40 mL 50 mM w/v NH 4 HCO 3 was added. Gel samples were washed with 40 mL of 50 mM w/v NH 4 HCO 3 , the supernatant was collected, and both collected supernatants were combined. Then, the collected solution was washed with 70% v/v ACN and dried in a Speed Vac (Vacuum Concentrator, Bachhofer). The peptide mixtures were desalted using ZipTip C-18 RP tips (Millipore, Billerica, MA, USA) that were moistened with 100% ACN and equilibrated with 0.1% TFA. Peptide samples, which were redissolved in 10 mL 0.5% TFA, were eluted with 50% ACN/0.1% TFA and were then dried in a Speed Vac (vacuum concentrator).
The purified peptides were spotted on a MALDI plate and covered with 0.7 mL of 2 mg/mL 3,5-Dimethoxy-4-hydroxycinnamic acid matrix (Sigma) with 10 mM NH 4 H 2 PO 4 in 60% ACN. All samples were then analyzed by MALDI-TOF/TOF MS with a 4800 Proteomics Analyzer (Applied Biosystems, Foster City, CA). Monoisotopic peak masses were acquired in a mass range of 800-4,000 Da, with a signal-to-noise ratio (S/N) 200. Five of the most intense ion signals, excluding common trypsin autolysis peaks and matrix ion signals, were selected as precursors for MS/MS acquisition. The peptide mass fingerprint (PMF) combined MS/ MS data were submitted to MASCOT version 2.1 (Matrix Science) for identification according to the SwissProt Sus scrofa database. The searching parameters were set as follows: Sus scrofa, trypsin cleavage (one missed cleavage permitted), fixed modifications, methionine oxidation as a variable modification, peptide mass tolerance set at 100 ppm, and fragment tolerance set at 0.8 Da. The required criteria for the successful identification of a protein were as follows: an ion score confidence interval (C.I.%) for PMF and MS/MS data $95%, and a peptide count (hit) $4. At least two peptides with distinct sequences were identified in the MS/MS analysis.

Gene annotation and functional classification
Gene Ontology (GO) annotation and functional classification of the identified proteins was performed with Blast2GO [24,25], the version is V2.6.2 and the current public database is b2g_aug12 (www.blast2go.com). A nonredundant database was used as reference for Blastp searches with an expectation value minimum of 16e 23 and a high scoring segment pair cut-off of 33. Annotations were made with default parameters. Briefly, the pre-eValue-Hit-Filter was 16e 26 , the Annotation cut-off was 55, and the GO Weight was 5. The percentages of assigned GOs in level 2 of biological process, molecular function and cellular component were calculated.

Construction of protein-protein interaction networks
We performed a protein-protein interaction network analysis on the identified differentially expressed proteins. We utilized String software (http://string-db.org/) to search the String database and the protein-protein interaction network database and subsequently built a protein-protein interaction network for all of the differentially expressed proteins.

Validation of differentially expressed proteins via real-time PCR
The virulent HuN4 strain and the attenuated HuN4-F112 strain were used to infect PAM cells separately. Infected cells were harvested 24 hours and 48 hours after infection, followed by extraction of total cellular RNA in accordance with the instructions of the RNeasy Plus Mini Kit (Qiagen). The RNA concentration and purity were determined based on the A260 and A280 values, and the RNA concentration was adjusted to 250 ng/ mL. Real-time PCR was performed using a kit from Fermentas Inc. according to the manufacturer's instructions. Total RNA was reverse transcribed into cDNA. PCR was performed in a reaction volume of 25 mL containing 12.5 mL of Maxima SYBR Green ROX qPCR Master Mix (26), 0.5 mL of each of the primers (10 mM) ( Table 2), and 1 mL of cDNA. The reaction conditions were as follows: 50uC for 2 minutes, 95uC for 10 minutes, and 40 cycles of 95uC for 15 secondsR65uC for 60 seconds (for fluorescence detection). Additionally, a melting curve analysis was performed at 65uC to 95uC at the end of the reaction. All samples were used in three independently repeated experiments, and the average value obtained from the three experiments was taken as the quantitative result. Quantitative analysis was performed using the quantitation-comparative C T (DDC T ) mode in ABI 7500 software, version 2.0.1 (USA). The uninfected group was used for calibration (relative expression = 1), and b-actin was used as the internal reference gene. Finally, a single factor analysis of variance (ANOVA) was used to analyze the data statistically, with P,0.05 indicating statistically significant differences and P,0.01 indicating highly statistically significant differences.

2D-DIGE screening of differentially expressed protein spots
The reproducibility of the system was evaluated by running an identical sample three times. Figure 1A shows a representative 2D-DIGE image. Approximately, a total of 1492 Protein spots were resolved in a single master gel image of this reproducibility experiment. Image analysis revealed a total of 153 differentially expressed protein spots in the PAMs infected with the virulent HuN4 strain or the attenuated HuN4-F112 strain (with an average ratio .1.2 or ,21.2, P,0.01) ( Table 3). 73, 70, and 98 protein

MALDI-TOF/TOF MS identification of differentially expressed proteins
MALDI-TOF/TOF MS was used to analyze the samples and the protein identity and partial sequence information was obtained by GPS Explorer and MASCOT software. We picked and analyzed 42 spots of the total 153 spots, 27 of which were successfully identified as 24 proteins (Table 4). Spots 1017 and 1028 were identified as triosephosphate isomerase 1. Spots 1043, 1039 and 993 were identified as heat shock protein beta-1(heat shock protein 27 kDa). Spots 616 and 1034 were identified as keratin 18.

GO annotation and functional classification
Gene Ontology annotation was performed for the 24 differentially expressed proteins. First, the Blast2GO was used to download annotated porcine protein data from the NCBI nonredundant database, and GO functional classification was performed for every annotated known protein. In the present study, 24 single sequences were successfully matched with one or more GO phrases, and these 256 matched GO phrases were functionally classified into three unrelated GO functions (GO ontologies): biological processes, molecular functions, and cellular components (Figure 2). Among these categories, biological processes accounted for 145 sequences (56.7%), molecular functions for 38 sequences (14.8%), and cellular components for 73 sequences (28.5%). The biological processes class involved the following four major categories: cell processes, metabolism, biological regulation, and cell structural components. This class was also associated with the functions immune regulation and signal transduction, including six sequences (4%) involved in immune system processes, eight sequences (6%) involved in signaling pathways, and eleven sequences (8%) involved in response to stimulus. In the molecular functions class, 18 sequences (47%) presented the function of binding with other molecules, and 12 sequences (31%) had catalytic functions.

Real-time PCR analysis of differentially expressed genes
Real-time PCR assay was done to confirm some differences found by 2D-DIGE and MS in the current study. The result, PKM2 was upregulated at 24 hpi by infection with both the HuN4 strain and the HuN4-F112 strain. However, it is interesting that at 48 hpi PKM2 was still upregulated in cells with HuN4 infection but downregulated in cells with HuN4-F112, and the difference is significant (P,0.01). The expression of HSPB1 was upregulated upon infection with the HuN4 strain and downregulated upon infection with the HuN4-F112 strain, and the differences between cells infected with HuN4 and HuN4-F112 strains were significant at 48 hpi (P,0.01). Similar to HSPB1, PSMA6 was also upregulated with HuN4 infection and downregulated when infected with HuN4-F112, and the difference at 48 hpi between them was significant (P,0.05), too. So, all the differences in Realtime PCR assay confirmed those we found by 2D-DIGE and MS (Figure 4).

Discussion
There were significant differences between the virulent HP-PRRSV strain, HuN4, and its attenuated form, HuN4-F112, detected in our in vivo pathogenicity tests [5,22]. Focusing on the causes of the differences in pathogenicity between the two strains, we previously analyzed the strains at the level of viral genes and revealed that there were only 41 amino acid differences between the two strains [23]. Two infectious molecular clones of virulent HuN4 and HuN-F112 strains were created, and several chimeric strains by mixing those two were also made, to study viral pathogenicity. However, we have not yet identified any amino acid differences being responsible for the different pathogenicity between HuN4 and HuN-F112 [26], and we are still working on it. On the other hand, we also attempted to identify relevant factors underlying the different responses to infection with HuN4 and virulent HuN4-F112 strains within host PAM cells, so in the present study, we did 2D-DIGE followed by MS to screen differential protein expression after infecting of PAM cells with HuN4 and HuN-F112 strains. Knowing protein level changing caused by viral infection may be able to allow us to better understand the mechanism of viral pathogenicity, too.
PAM cells infected with the virulent HuN4 strain and the attenuated HuN4-F112 strain were collected for utilizing fluorescent two-dimensional difference gel electrophoresis (2D-DIGE) to screen for intracellular proteins that were differentially expressed in host cells infected with the two strains. There were 153 proteins with significant different expression (P,0.01) observed, 42 of which were subjected to mass spectrometry. Among those 42 samples, in this study, we identified 24 differentially expressed proteins that might be involved to varying degrees and could play different roles in the processes leading to the differences between the two strains. The validation of a number of differentially expressed proteins showed that the expression of PKM2, HSPB1, and PSMA6 was significantly different during the infection course between the virulent HuN4 strain and the attenuated HuN4-F112 strain.
Different viruses have different methods for regulating cell metabolism by regulating certain cellular protein level or concentration of particular lipids, like Hepatitis C virus [27], hepatitis B virus [28], human cytomegalovirus (HCMV) and herpes simplex virus 1 (HSV-1) [29]. Viruses may first regulate enzymes involved in cellular metabolic pathways to utilize the cellular metabolism of energy and substances for viral synthesis [29][30][31]. Here, we found that PKM2 was upregulated when PAMs were infected with the HuN4 strain and downregulated when PAMs were infected with the HuN4-F112 strain. PKM2 is an enzyme that participates in the metabolism of carbohydrates and is mainly involved in glycolysis. It has also been observed that when an oncogenic virus causes tumor, glycolysis is enhanced in the tumor cells, and that it is glycolysis, not aerobic metabolism, provides the energy supply (the mitochondrial oxidative phosphorylation reaction is inhibited). Although less ATP is produced per monosaccharide molecule during glycolysis than during aerobic metabolism, increased glycolysis and reduced oxidative phosphorylation could increase the production of ATP without generating reactive oxygen species, which was known as Warburg effect [32,33]. The Warburg effect is required for the infectivity of many viruses, such as the Kaposi's sarcoma herpesvirus [33], feline leukemia virus [34], Rous sarcoma virus [35], cytomegalovirus [36], and white spot syndrome virus [37]. The Warburg effect facilitates viral replication and long-term viral existence in cells [32]. Here, potential role of PKM2 in PRRSV infection needs to be addressed by further study.
Our study showed that HSPB1 in PAM cells was up-and downregulated by the virulent HuN4 strain and the attenuated HuN4-F112 strain, respectively, which confirmed related previous studies [18,20]. In addition, infections involving many other viruses, such as the avian influenza H9N2 [38], the Epstein-Barr virus [39], the African swine fever virus [40], IBDV [41] and CSFV [42] also induce upregulation of HSP27 expression. HSPB1, also known as heat shock protein 27 (HSP27), belongs to the small heat shock protein (HSP) family and is a ubiquitous protein induced by stress. To release large quantities of virus particles during the virus release phase, the virus delays cellular apoptosis through different strategies in the early stages of the viral infection of cells to provide sufficient time for viral replication. Hsp27 regulates apoptosis through an ability to interact with key components of the apoptotic Figure 4. Confirmation of the transcriptional regulation of DEPs by real-time RT-PCR. Transcript alteration of three selected genes in PAM cells from the PRRSV-infected group compared with the mock-infected group. Total RNA extracted from PAM cells was measured by real-time RT-PCR analysis; relative expression levels were calculated according to the 2 DD CT method, using b-actin as an internal reference gene and the mock-infected group as a calibrator (relative expression = 1). Error bars represent the standard deviation. Please refer to Table 2 for the identification of gene symbols that represent different genes. doi:10.1371/journal.pone.0085767.g004 signaling pathway, in particular, those involved in caspase activation and apoptosis [43]. HSPB1 is also involved in the cellular protective responses to various stresses, such as heat shock, toxins, and oxidative stress, as well as stress-induced HSP regulation, the mitogen-activated protein kinase (MAPK) signaling pathway, and the regulation of translation initiation, molecular chaperones, actin organization, and cell motility, and, any change of which could affect PRRSV infection, but we still need to do more study to find out.
In the protein-protein network diagram, PSMA6, PSMB3, PSMD9 (proteasome (prosome, macropain) 26S subunit, non-ATPase, 9), and PSMB7 (proteasome (prosome, macropain) subunit, beta type 7) all belong to the ubiquitin-proteasome pathway (UPP), and all of these proteins showed significant changes between the PAM cells infected with the virulent strain and the cells infected with the attenuated strain. PSMA6 was upregulated following infection with the HuN4 strain in the present study and was downregulated by infection with the HuN4-F112 strain. The UPP is the major intracellular degradation system that performs degradation of foreign proteins and is the system that a virus must avoid to evade immune surveillance and achieve virus particle maturation, release, replication, and activation from the latent state [44,45]. It has been reported that many viruses employ unique approaches for co-opting the UPP pathway for the replication of their viral particles [44]. Activation of the UPP pathway is necessary for hepatitis E virus replication [46] and is also required by other viruses, such as the hepatitis B virus [47,48], influenza A virus [49], vaccinia virus [50], herpes simplex virus [51], and rotavirus [52]. Zhang et al. [18] studied the infection of PAMs by HP-PRRSV using proteomic methods and found that HP-PRRSV upregulated the expression of UPP-related proteins, suggesting that these proteins might enhance HP-PRRSV replication in PAMs. Our results also showed that the virulent HuN4 strain upregulated the expression of UPP-related proteins, while the attenuated HuN4-F112 strain downregulated the expression of UPP-related proteins, suggesting that the virulent PRRSV strain might utilize the UPP pathway to promote viral replication. However, determination of whether PRRSV employs a similar approach to evade host immune surveillance and whether the virulent and attenuated strains differ in their ability to regulate and utilize the UPP pathway, thus leading to their different replication capacities in cells, still requires further experimental validation.
In summary, we utilized proteomic approaches to identify differential expression of cellular proteins between PAM cells infected with either the virulent HuN4 strain or the attenuated HuN4-F112 vaccine strain. Functional analysis of the proteins showed that the differentially expressed proteins were correlated with the degree of infectivity of the virulent/attenuated strains in the target PAM cells and suggested that these differentially expressed proteins are involved in the pathogenic mechanisms of PRRSV strains with different pathogenicities. Our findings provide new information that can be used in future studies addressing the differences in pathogenicity between PRRSV virulent and attenuated strains. Table S1 The differential expressed protein spots between HuN4 and PAM(with an average ratio .1.2 or ,21.2, P,0.01).

Supporting Information
(DOC)