Proteomic analysis reveals down-regulation of surfactant protein B in murine type II pneumocytes infected with influenza A virus

Abstract

Infection of type II alveolar epithelial (ATII) cells by influenza A viruses (IAV) correlates with severe respiratory disease in humans and mice. To understand pathogenic mechanisms during IAV infection of ATII cells, murine ATII cells were cultured to maintain a differentiated phenotype, infected with IAV-PR8, which causes severe lung pathology in mice, and proteomics analyses were performed using liquid chromatography-mass spectrometry. PR8 infection increased levels of proteins involved in interferon signaling, antigen presentation, and cytoskeleton regulation. Proteins involved in mitochondrial membrane permeability, energy metabolism, and chromatin formation had reduced levels in PR8-infected cells. Phenotypic markers of ATII cells in vivo were identified, confirming the differentiation status of the cultures. Surfactant protein B had decreased levels in PR8-infected cells, which was confirmed by immunoblotting and immunofluorescence assays. Analysis of ATII cell protein profiles will elucidate cellular processes in IAV pathogenesis, which may provide insight into potential therapies to modulate disease severity.

Introduction

Influenza A virus (IAV) is a common cause of human respiratory tract infections with epidemics causing significant morbidity and mortality worldwide. While most infections are mild, spread of IAV to the lungs can result in severe pneumonia during annual epidemics or due to infection by highly pathogenic strains, such as avian H5N1 or H7N9 viruses (Guarner and Falcon-Escobedo, 2009; Mauad et al., 2010; Shieh et al., 2010; Uiprasertkul et al., 2007). Like other viruses, IAV depends on the host cellular machinery for replication, which is orchestrated by complex interactions between host and viral proteins. Further, the host cell is able to mount antiviral responses to counter-act the effects of the virus. An understanding of the complex interactions between IAV and host cells is important for the development of novel strategies to inhibit the viral replication cycle or enhance effective host antiviral responses.

The host factors that contribute to pathogenesis during IAV infection are frequently cell type-specific. Epithelial cells of the upper and lower respiratory tracts are the primary sites of virus replication in IAV infections. However, infection of the lower respiratory tract is the most severe as it may result in damage of the alveolar epithelium, thus compromising the gas exchange in the lungs (Blazejewska et al., 2011; Guarner and Falcon-Escobedo, 2009; Shieh et al., 2010). The alveolar type II (ATII) epithelial cells located in the lower respiratory tract are involved in regeneration to repair injured alveolar epithelium and produce pulmonary surfactant phospholipids and proteins, which are critical to prevent alveolar collapse (Rooney et al., 1994). ATII cells are infected by IAV in fatal human cases of avian H5N1 and 2009 pandemic H1N1 virus infections (Mauad et al., 2010; Shieh et al., 2010; Weinheimer et al., 2012). Influenza A/PR/8/34 H1N1 virus (PR8), a mouse adapted influenza strain, provokes destruction of ATII cells in the alveoli of mice (Loosli et al., 1975), making it a suitable model to study IAV pathogenesis in the lung. Primary murine ATII cells thus provide a powerful in vitro model to determine the effects of viral infection on these specialized pulmonary cell types, which corresponds to pathology in the mouse model. We have shown that differentiated cultures of primary murine ATII cells are susceptible to PR8 infection in vitro, resulting in secretion of various pro-and anti-inflammatory cytokines and chemokines (Kebaabetswe et al., 2013). This approach, however, did not offer a comprehensive analysis of viral-host protein interactions that could be important to pathogenesis during infection of ATII cells. Hence, examining protein expression in IAV-infected ATII cells could offer a global picture of how these cells interact with IAV at the cellular level.

Proteomics studies of influenza virus infections in macaques (Brown et al., 2010), mice (Kumar et al., 2014), continuous cell lines (Coombs et al., 2010; Kummer et al., 2014; Vester et al., 2009), and primary human cells (Kroeker et al., 2012; Lietzen et al., 2011; Liu et al., 2012) have provided insight into how IAV strains alter the host proteome at cellular and organismal levels. However, host responses to IAV infection are also, in part, specific to the cell type. Analysis of primary murine ATII cells allows us to characterize the host response to infection in cells that are relevant to severe disease and correlate these findings to pathological findings in the mouse model. In this study, our aim was to utilize the in vitro model of highly differentiated ATII cells and examine cellular protein profiles following infection with PR8. We used liquid chromatography-mass spectrometry (LC-MS) and the accurate mass and time (AMT) tag approach (Zimmer et al., 2006) to quantitate cellular proteins in PR8-infected and mock-inoculated ATII cells. The proteomics data were confirmed for a select set of proteins by immunoblotting and/or immunofluorescence assay. These data provide information on the differentiated phenotype of primary murine ATII cells and the effects of PR8 infection on the cellular proteome, which suggest potential mechanisms of pathogenesis during IAV infection of this critical cell type.

Results and Discussion

Infection of primary, differentiated ATII cells by PR8

Primary ATII cells were isolated from mice and cultured to maintain an ATII phenotype in vitro, as previously described (Kebaabetswe et al., 2013). ATII cultures were inoculated with PR8 or mock-inoculated and infection was analyzed by immunofluorescence assay and immunoblotting 24 h later. Viral hemagglutinin (HA) was detected by immunofluorescence in PR8-, but not mock-inoculated ATII cells (Fig. 1A). Confocal image analysis demonstrated that HA-specific fluorescence was detected within the cytoplasm of infected cells (Fig. 1B). Further, the different forms of HA: the uncleaved precursor, HA0 and cleaved forms, HA1 and HA2 were detected by immunoblot analysis (Fig. 1C). The presence of the cleaved HA protein in cell lysates suggests that differentiated ATII cells are capable of post-translational processing of HA, which is required for viral infectivity and correlates with virulence of IAV strains (Blazejewska et al., 2011; Steinhauer, 1999). These data are in agreement with our previous study that demonstrated productive infection of murine ATII cells by PR8 (Kebaabetswe et al., 2013). Infection of primary murine ATII cells provides a robust system for analysis of PR8-regulated changes to the cellular proteome in physiologically relevant target cells. Protein profiles were analyzed 24 h after inoculation with PR8 because this allowed time for the virus to spread through-out the culture without inducing overt cytopathic effects (Fig. 1A). In a previous study, PR8 replication in primary murine ATII cells peaked at 24 h (Kebaabetswe et al., 2013). This time point is consistent with other proteomic studies using the same strain of influenza A virus in different cell types, including primary cultures (Coombs et al., 2010; Kroeker et al., 2012; Liu et al., 2012).

Fig. 1. Influenza viral infection of ATII cells.

Primary murine ATII cells were mock-treated or inoculated with PR8 for 24 h. Immunofluorescence assay of viral hemagglutinin protein (red) and nuclei stained with DAPI (blue) was imaged using (A) epifluorescence microscopy at 40X and (B) confocal microscopy at 60X. Orthogonal views along the indicated cross lines are shown to the right and below the projected Z-stack. Scale bar indicates 10 µm. (C) Immunoblot analysis of viral HA protein isoforms in whole cell lysates. β-actin was used as a loading control. Molecular weights in kDa are provided to the left. The images shown are representative of three replicate experiments.

Influenza viral proteins identified by proteomics analyses

Four biological replicates each of mock-inoculated and PR8-infected ATII cultures were analyzed by LC-MS-based proteomics using the AMT tag approach. Five IAV proteins were detected (Table 1): HA, neuraminidase (NA), nucleoprotein (NP), matrix 1 (M1), and non-structural protein 1 (NS1). These viral proteins were present in all four replicates of the PR8-infected ATII cells. Peptides mapping to NP and HA were also detected in two or fewer of the mock samples. Kummer et al. demonstrated that these five proteins are the most abundant viral proteins in PR8-infected MDCK cells (Kummer et al., 2014). HA, NA, NP, and M1 are present in infectious virions and are produced in great abundance during viral assembly (Hutchinson et al., 2014; Kummer et al., 2014). Therefore, detecting these proteins late during the replication cycle (24 h post-infection) is not surprising and suggests that influenza virus is actively replicating in these cells. NS1 has multiple roles during IAV infection, including inhibition of cellular gene expression and the type I interferon response, and is also expressed at high levels in infected cells (Fernandez-Sesma et al., 2006; Garcia-Sastre et al., 1998). Proteins that form the viral RNA-dependent RNA polymerase complex, PB1, PB2, and PA, the nuclear export protein, NEP, and M2 protein are expressed in relatively low abundance in infected cells (Kummer et al., 2014) and were not detected by our analysis. Overall, we detected a similar profile of viral proteins as others have reported in proteomics analyses of influenza virus- infected cells and animals (Brown et al., 2010; Lietzen et al., 2011; Liu et al., 2012).

Table 1.

Influenza A virus proteins identified by proteomics analysis

Uniprot ID	Viral Protein	Number of Peptides Identified	Fold Change1 (P-value)
P03466	Nucleoprotein (NP)	61	18.2 (0.049)
P03452	Hemagglutinin (HA)	39	15.3 (0.075)
P03468	Neuraminidase (NA)	8	*(0.001)
P03485	Matrix protein 1 (M1)	8	*(0.014)
P03496	Non-structural protein 1 (NS1)	4	*(0.001)

¹Fold change expressed as ratio of signal intensity in PR8-infected/mock

P-value determined by Student’s t-test, or

^*G-test when not detected in mock

PR8-induced changes to the cellular proteome

Proteomics analyses were used to compare the ATII protein profiles in PR8-infected and mock-inoculated cells using four biological replicates for each treatment. Comparative statistical analyses of control and infected samples were performed using a two-sample t-test to assess differences in protein average abundance, and a G-test to assess associations among factors due to the presence or absence of a protein. This analysis identified 151 differentially regulated proteins: 46 up-regulated and 105 down-regulated by PR8 infection (Fig. 2). The complete data set of peptides, proteins, and statistical analyses are provided as Supplemental Dataset. Proteins that mapped to a single peptide were excluded from further analyses. Proteins that were up-regulated by PR8 infection are associated with host defense, including interferon-induced cellular response, MHC class I antigen processing and presentation, and pathogen detection, and cytoskeletal proteins (Tables 2 and 3). Proteins that were decreased by infection included those involved in cellular metabolism, response to stress or cell damage, and regulation of cell death (Table 4). The significance of these functional groups is discussed below.

Fig. 2. Murine proteins with statistically significant expression in mock vs. PR8-infected ATII cells.

Proteomics analysis was performed on four replicates each of mock-treated and PR8-infected murine ATII cell cultures. Comparative statistical analyses of mock with PR8-infected samples were performed using a 2-sample t-test to assess differences in protein average abundance, and a G-test to assess associations among factors due to the presence/absence of response.

Table 2.

Type I interferon-induced proteins increased by PR8 infection of ATII cells

Uniprot ID	Gene	Protein Name	Reference
Q8R2Q8	BST-2	Bone marrow stromal antigen 2/tetherin	Der et al., 1998
Q9CQW9	IFITM3	Interferon induced transmembrane protein 3	Dolken et al., 2008
Q99MB1	TLR3	Toll-like receptor 3	Dolken et al., 2008
Q9R233	Tapasin	TAP binding protein	Dolken et al., 2008
P01897	H-2L	MHC Class I alpha chain (H2L)	Der et al., 1998
P01898	H-2Q10	MHC Class I alpha chain (Q10)	Der et al., 1998
P14426	H-2D(K)	MHC Class I alpha chain (D–K)	Der et al., 1998
P01887	B2M	β-2 microglobulin	Der et al., 1998
Q07797	Lgals3bp	Galectin-3 binding protein	Dolken et al., 2008
P05533	Ly6A	Lymphocyte antigen 6 complex, locus A	Dolken et al., 2008
P62631	EF1a2	Eukaryotic translation elongation factor 1 alpha 2	Der et al., 1998

List of proteins with two or more peptides identified and significantly increased abundance in PR8-infected samples (see Table 3) that were identified by Interferome v. 2.1 to be induced by type I IFN.

Table 3.

Proteins significantly up-regulated in PR8-infected ATII cells

Uniprot ID	Protein Name	Number of Peptides Identified	Fold Change1 (P-value)
Host Defense
Q8R2Q8	Bone marrow stromal antigen 2	4	25.4 (<0.001)
P05533	Lymphocyte antigen 6A-2/6E-1	4	4.23 (0.018)
Q9R233	TAP binding protein (Tapasin)	4	*(0.001)
Q07797	Galectin-3 binding protein	3	3.51(<0.001)
P01897	MHC Class I alpha chain (H2L)	3	*(0.014)
Q9CQW9	Interferon induced transmembrane protein 3	2	*(0.001)
P01898	MHC Class I alpha chain (Q10)	2	*(0.014)
P14426	MHC Class I alpha chain (D–K)	2	*(0.014)
P01887	β-2 microglobulin	2	*(0.001)
Q99MB1	Toll-like receptor 3 (TLR3)	2	*(0.014)
Cytoskeleton
Q6IFZ6	Cytokeratin 1B	4	2.28 (0.049)
P26040	Ezrin	4	1.40 (0.045)
Q9Z2K1	Cytokeratin 16	2	2.89 (0.013)
P08730/Q61414	Cytokeratin 13 or 15	2	2.23 (0.006)
Protein Synthesis
P62631	Eukaryotic translation elongation factor 1 alpha 2	2	1.33 (0.004)
Cellular Metabolism
Q05769	Cyclooxygenase-2	3	2.68 (0.006)
Q64467	Glyceraldehyde-3-phosphate dehydrogenase	2	1.53 (0.044)

¹Fold change expressed as ratio of signal intensity in PR8-infected/mock

P-value determined by Student’s t-test, or

^*G-test when not detected in mock

Table 4.

Proteins significantly down-regulated in PR8-infected ATII cells

Uniprot ID	Protein Name	Number of Peptides Identified	Fold Change1 (P-value)
Cell Death and Stress Response
Q60930	Voltage-dependent anion channel 2	21	0.80 (0.042)
Q60932	Voltage-dependent anion channel 1	6	0.71 (0.002)
Q01853	Transitional endoplasmic reticulum ATPase	5	0.51 (0.044)
Q69ZN7	Myoferlin	5	0.63 (0.008)
Q6P069	Sorcin	2	*(0.014)
Q8R1V4	ER stress response protein 25	2	0.64 (0.049)
Proteolysis
P49722	Proteasome, subunit, alpha type 2	17	0.76 (0.032)
O55234	Proteasome, subunit, beta type 5	13	0.66 (0.017)
P57716	Nicastrin	3	0.57 (0.002)
P17047	Lysosomal-associated membrane protein 2	2	0.25 (0.014)
Chromatin Formation and DNA Damage Response
P62806	Histone H4	55	0.68 (0.026)
Q9D2U9	Histone H2ba, H2bb	34	0.87 (0.025)
Q9QZQ8	Core histone H2A.1	5	0.62 (0.001)
O54962	Barrier to autointegration factor 1	2	0.72 (0.008)
Mitochondrial
P43024	Cytochrome c oxidase, subunit VI	6	0.5 (0.023)
Q8BW75	Monoamine oxidase B	6	0.64 (0.016)
Q9CZ13	Ubiquinol-cytochrome c reductase core protein 1	5	0.69 (0.005)
Q9CPQ8	ATP synthase, H+ transporting, mitochondrial F0 complex	4	0.55 (0.02)
Q99JR1	Sideroflexin 1	4	0.61 (0.037)
Q99JB2	Stomatin (Epb7.2)-like 2	4	0.72 (0.002)
Q9R112	Sulfide quinone reductase-like (yeast)	3	0.52 (0.01)
Q9Z2Z6	Mitochondrial carnitine/acylcarnitine translocase, 20	3	0.52 (0.018)
Q9DCS9	NADH dehydrogenase 1 beta subcomplex, 10	3	0.6 (0.028)
O55126	Protein NipSnap homolog 2	3	0.66 (0.037)
Q8BGH2	Sorting and assembly machinery component 50	3	0.68 (0.017)
Q921G7	Electron transferring flavoprotein, dehydrogenase	2	*(0.014)
Q9DCJ5	NADH dehydrogenase 1 alpha subcomplex, 8	2	0.52 (0.013)
Q8R127	Saccharopine dehydrogenase (putative)	2	0.56 (0.002)
Q9D855	Ubiquinol-cytochrome c reductase binding protein	2	0.71 (0.005)
Q9D0M3	Cytochrome c-1	2	0.73 (0.002)
Cell Signaling and Adhesion
Q8R2Y2	Melanoma cell adhesion molecule MUC18	6	0.5 (0.006)
Q61490	Activated leukocyte cell adhesion molecule	6	0.64 (0.002)
P11688	Integrin alpha 5 (fibronectin receptor alpha)	4	0.60 (0.002)
Q07113	Insulin-like growth factor 2 receptor	3	*(0.014)
Lipids and Polysaccharides Metabolism
Q8C166	Copine I	11	0.67 (0.029)
P50405	Surfactant-associated protein B	8	0.37 (<0.001)
Q64435	UDP glucuronosyltransferase 1–6	6	0.78 (0.05)
Q80UM7	Mannosyl-oligosaccharide glucosidase	2	*(0.014)
Other Metabolism
P22437	Prostaglandin-endoperoxide synthase 1	2	0.44 (0.007)
Miscellaneous Transmembrane
Q9DBS1	Transmembrane protein 43	11	0.77 (0.037)
Q8VCM8	Nicalin	2	0.47 (0.002)

¹Fold change expressed as ratio of signal intensity in PR8-infected/mock

P-value determined by Student’s t-test, or

^*G-test when not detected in PR8

Cellular proteins increased by PR8 infection

Virus-infected cells produce type I interferons (IFNs), pleiotropic cytokines that contribute to both inhibition of viral replication and activation of downstream immune responses. Many viruses counter-act the type I IFN response to promote successful infection. This tug-of-war between type I IFN production and antagonism is central to the outcome of infection in favor of either the host or virus. To identify proteins that are induced by type I IFNs in murine model systems, the Interferome database (http://interferome.its.monash.edu.au/interferome/) was queried with the subset of proteins whose levels were significantly increased by PR8 infection and mapped to two or more peptides (Rusinova et al., 2013). This analysis identified several IFN-induced proteins that were up-regulated by PR8 infection, including antiviral effector proteins (BST-2, IFITM3), pathogen recognition receptors (TLR3), and multiple IFN-induced components of the MHC class I antigen processing and presentation pathway (Table 2) (Der et al., 1998; Dolken et al., 2008). Our proteomics analysis did not include secreted proteins; therefore, we did not directly identify type I IFNs. IFNs stimulate cells bearing the IFN α/β receptor, which can be the cells that produced the IFNs in addition to neighboring cells. Primary cultures of murine ATII cells are not homogenously susceptible to PR8 infection, as can be seen by heterogeneous staining of viral antigens in our cultures (Fig. 1A) (Kebaabetswe et al., 2013). Therefore, it is possible that the IFN-stimulated proteins we are detecting come from either infected or uninfected cells within the culture. Furthermore, it is possible that the type I IFN response is protecting some cells in the culture from infection. Additional studies to determine the role of this response in limiting infection of the alveolar epithelium are needed.

The NS1 protein of influenza virus is a potent interferon antagonist that targets multiple steps, including upstream viral detection and signaling, IFN production, and downstream inhibition of antiviral effector proteins. There are differences in the effectiveness of IFN antagonism by different influenza virus strains and among host cell types. The NS1 protein of PR8 suppresses expression of type I IFNs and IFN-induced genes in infected human dendritic cells and lung epithelial cells, and is required for virulence in mice with functional IFN responses (Fernandez-Sesma et al., 2006; Garcia-Sastre et al., 1998; Kochs et al., 2009). However, PR8 NS1 has been found to inhibit the IFN response by human dendritic cells by a lesser degree than human isolates of IAV (Haye et al., 2009). Furthermore, the multiple functions of NS1 are regulated by post-translational modification, including phosphorylation, SUMOylation, and ISGylation (Hsiang et al., 2012; Santos et al., 2013; Zhao et al., 2010). It is unknown how these modifications of NS1 might affect IFN antagonism by PR8 specifically in ATII cells. Finally, in addition to NS1, the PB2 and PB1-F2 proteins of IAV have been demonstrated to inhibit RIG-I signaling, which is critical for production of type I IFN in response to IAV infection (Graef et al., 2010; Opitz et al., 2007; Varga et al., 2011). Our results suggest that PR8 does not completely impair IFN signaling in differentiated murine ATII cells. However, we cannot rule out the possibility of IFN-independent expression of these proteins, as some can also be induced by other signaling pathways (Bego et al., 2012).

One of the proteins that was most highly increased by PR8 infection was IFN-induced, bone marrow stromal antigen 2 (BST-2/tetherin/CD317; Table 3). BST-2 is a lipid raft-associated membrane protein that inhibits the release of a diversity of enveloped viruses during the budding process. Although IAV uses lipid rafts for budding (Takeda et al, 2003), there have been contradictory findings regarding the role of BST-2 in restricting the release of IAV from infected cells. Multiple studies report that BST-2 inhibits the release of influenza virus-like particles but not infectious viruses (Bruce et al., 2012; Watanabe et al., 2011; Winkler et al., 2012; Yondola et al., 2011). In contrast, other studies found that BST-2 reduces the amount of wild-type viral particles released from infected cells (Mangeat et al., 2012; Narkpuk et al., 2014). Both of these studies used IAV generated through reverse genetics, which may account for the contrasting findings with the studies that saw no effect of BST-2 on release of infectious virus. In addition to a direct role in viral inhibition, BST-2 also functions as a viral detection and signaling molecule, leading to activation of NFκB (Galao et al., 2012). While our results demonstrate that BST-2 is part of the host cellular response to infection, further studies are needed to determine whether BST-2 inhibits release of PR8 and/or activates proinflammatory gene expression during infection of ATII cells.

An additional antiviral effector protein that is induced by IFN, IFITM3, was detected in PR8-infected ATII cells and not in mock-inoculated cells (Tables 2 and 3). In contrast to BST-2, IFITM3 inhibits an early step in viral infection, preventing effective viral entry prior to replication. IFITM3 inhibits membrane fusion of IAV, thereby trapping viruses in the endocytic pathway, leading to lysosomal degradation of virions (Desai et al., 2014). There is strong evidence for a role for IFITM3 in protecting both mice and humans against severe influenza disease. Mice lacking IFITM3 have increased viral replication, morbidity, and mortality upon IAV infection (Bailey et al., 2012; Everitt et al., 2011). A minor human IFITM3 allele that has reduced ability to inhibit IAV infection in vitro is associated with severe disease upon infection with seasonal or 2009 pandemic H1N1 viral strains (Everitt et al., 2011; Zhang et al., 2013). IFITM3 appears to have a protective role against severe disease upon IAV infection (Bailey et al., 2012; Everitt et al., 2011). As severe disease corresponds with infection of pulmonary ATII cells, expression of IFITM3 in these cells may be an important determinant of influenza disease severity.

Toll-like receptors (TLRs) detect pathogen-associated molecular patterns, triggering a signaling cascade that leads to expression of inflammatory cytokines and type I IFNs. TLR3 specifically recognizes dsRNA produced during viral infections, and type I IFNs up-regulate expression of TLR3. Our proteomics analyses detected TLR3 in PR8-infected ATII cells, but not in mock-inoculated cells (Tables 2 and 3). IAV infection of respiratory epithelial cells triggers TLR3 signaling, resulting in cytokine expression (Guillot et al., 2005). The TLR3-mediated response to IAV infection can restrict viral production, but also contribute to inflammatory pathology in the lungs. Because of this, the outcome of IAV infection in TLR3−/− mice is highly dependent upon the strain and dose of virus. Mice deficient in TLR3 have increased survival than wild-type mice when infected with a highly pathogenic H5N1 strain or mouse-adapted H3N2 strain (Le Goffic et al., 2006; Leung et al., 2014). In contrast, TLR3−/− mice have no survival advantage when infected with a 2009 pandemic H1N1 virus (Leung et al., 2014). A common polymorphism in the TLR3 gene has been associated with severe disease upon infection with 2009 pandemic H1N1 virus (Esposito et al., 2012). Thus, TLR3 may contribute to either pathogenesis or viral suppression during IAV infection of the alveoli.

Type I IFNs also activate the MHC class I antigen processing and presentation pathway, which is critical for recognition of virus-infected cells by cytotoxic T cells. IFN-β produced by RSV-infected epithelial cells stimulates expression of various components of the MHC class I processing and presentation pathway (Jamaluddin et al., 2001). Several components of MHC I proteins such as β-2-microglobulin and various alpha chains were up-regulated by PR8 infection of ATII cells (Tables 2 and 3). Under the influence of type I IFNs, enhanced MHC I antigen processing and presentation by the ATII cells may render these cells to be targeted by cytotoxic CD8⁺ T cells during PR8 infection in the lungs.

Proteins that are associated with the cellular cytoskeleton and intracellular trafficking had increased abundance in PR8-infected ATII cells (Table 3). These include components of intermediate filaments (cytokeratins) and the actin network (ezrin). Other proteomics studies have reported an increased abundance of cytokeratins in various mammalian cell types infected by influenza A viruses (Liu et al., 2008; Vester et al., 2009). Arcangeletti et al. demonstrated co-localization of influenza NP with cytokeratin and a disruption of intermediate filaments late in the viral replication cycle (Arcangeletti et al., 1997). Furthermore, treatment of cells with an inhibitor of intermediate filaments resulted in reduced viral production, which corresponded with reduced cellular and viral protein synthesis (Arcangeletti et al., 1997). Thus, the cellular cytoskeleton, specifically intermediate filaments, is important for influenza viral replication. Potential roles of the cytoskeleton in promoting virus production include facilitating viral endocytosis, regulating gene expression, and trafficking of viral components to their correct location within the cell. Intermediate filaments are also important in maintaining polarization of epithelial cells. As influenza virions bud from the apical membrane of polarized cells, maintaining the intermediate filament network in ATII cells may be important for directing virus release from the correct side of the cell.

Cellular proteins decreased by PR8 infection

The majority of proteins that were down-regulated by PR8 in our study were associated with either the inner or outer mitochondrial membranes (Table 4). Proteomics analysis identified reduced levels of polypeptides that are components of the three main enzyme complexes in the oxidative electron transport chain: NADH dehydrogenase complex I, cytochrome b-c₁ complex, and the cytochrome oxidase complex, in PR8-infected samples. Several other mitochondrial proteins that regulate apoptosis were also detected in lower abundance in PR8-infected samples. The PB1-F2 protein of PR8 enhances mitochondrial permeability through interactions with cellular proteins in the inner and outer mitochondrial membranes (Zamarin et al., 2005). Proteins belonging to these two mitochondrial compartments were down-regulated in our study (Table 4). Reduced expression of these proteins may alter the permeability and membrane potential of mitochondria, thereby interfering with energy transduction and promoting cellular apoptosis. IAV has mechanisms to delay apoptosis of the host cell, which may be necessary to allow for sufficient time for viral replication (Zhirnov et al., 2002). IAV also benefits from cellular apoptosis. Caspase 3 activity enhances production of IAV by promoting the transport of viral ribonucleoprotein complexes out of the nucleus (Wurzer et al., 2003). In addition to decreased levels of proteins that regulate cell death at mitochondrial membranes, PR8-infected cells had reduced levels of other proteins involved in response to DNA damage, other cell stress, or protein degradation (Table 4).

Proteins involved in chromatin formation were down-regulated by PR8 infection including core histone proteins H2A.1, H2B, and H4 (Table 4). Influenza viruses undergo transcription and replication in close association with cellular chromatin and have mechanisms that alter chromatin structure. The NP and M1 proteins of IAV bind to core histone proteins (Garcia-Robles et al., 2005; Zhirnov and Klenk, 1997). NP binding may be involved in anchoring the viral RNA to cellular chromatin and M1 may be important for release of the viral ribonucleoprotein complex from the nucleus. Modification of cellular gene expression through chromatin remodeling may have important roles during viral infection. As influenza virus steals caps from cellular mRNAs, enhancement of cellular transcription concurrent with viral transcription would support viral gene expression. In contrast, inhibition of cellular gene expression would benefit the virus by inhibiting the cellular antiviral response. For example, the NS1 protein of an H3N2 strain of IAV suppresses cellular transcription of antiviral genes through a histone-like sequence (Marazzi et al., 2012). Other proteomics analyses have identified histone proteins as being down-regulated (Coombs et al., 2010; Liu et al., 2012) or up-regulated (Coombs et al., 2010; Kroeker et al., 2012) by IAV infection. The differences in these findings may be due to different viral strains, host cell species and types, the specific histone proteins and subcellular fractions analyzed by the study, and the timing within the viral replication cycle.

Validation of proteomics data

To validate the mass spectrometry based quantification, four proteins with increased abundance in PR8-infected ATII cells (BST-2, ezrin, TLR3, and IFITM3) and one with decreased abundance (histone H2B) were selected for immunoblot and densitometry analysis (Fig. 3). The proteomics data showed that BST-2 and ezrin had significantly increased levels in PR8-infected compared to mock-inoculated ATII cells (Table 3). In agreement with the proteomics analysis, the immunoblot shows that expression of BST-2 was enhanced in PR8-infected cells, showing multiple glycosylated forms of BST-2 (Fig. 3A), which have been stipulated to correlate with the antiviral activity of BST-2 (Mangeat et al., 2012). Ezrin also had a stronger signal in PR8-infected compared to mock-inoculated cells by immunoblot analysis (Fig. 3A). TLR3 and IFITM3 were only detected in PR8-infected samples by mass spectrometry (Table 3), whereas both proteins were detected in mock samples at lower abundance than PR8-infected ATII cells by immunoblot analysis (Fig. 3A). Multiple histone proteins had decreased abundance in PR8-infected, compared to mock-inoculated, ATII cells as determined by LC-MS (Table 4). Histone 2B was analyzed further by immunoblot analysis, which confirmed reduced levels in PR8-infected cells (Fig. 3B).

Fig. 3. Validation of proteomics data by immunoblotting.

At 24 h post-infection of murine ATII cells with PR8 or mock treatment, equal volumes of whole cell lysates were analyzed by immunoblot analysis of proteins that had (A) increased or (B) decreased abundance as determined by proteomics analysis. Densitometry was used to quantify band intensities, which are expressed as a ratio of the intensity of the specific protein band to the corresponding band for β-actin, after background correction. The data are representative of three replicate experiments. BST-2 = bone marrow stromal antigen 2, H2B = histone 2B, IFITM3 = interferon-induced transmembrane protein 3, TLR3 = tolllike receptor 3.

ATII-specific proteins identified by LC-MS

We previously showed that murine ATII cells cultured on collagen/matrigel in the presence of keratinocyte growth factor maintain expression of an ATII phenotype-specific protein, LBP180 (ABCA3) (Kebaabetswe et al., 2013). Through proteomics analyses, we identified ATII cell-specific proteins in both PR8-infected and mock-inoculated samples, further confirming that these cultures maintained an ATII phenotype in vitro. Single peptides that mapped to ATII-specific proteins LBP180 and lipocalin were detected in our samples (Supplemental Dataset). As shown in Table 5, we also identified multiple ATII-specific proteins by the detection of two or more peptides, including the four pulmonary surfactant proteins (SP)-A, B, C and D. These proteins have been identified by numerous studies to be specifically expressed by ATII, and not ATI cells in the alveoli (Gonzalez et al., 2005; Kalina et al., 1992; Mulugeta et al., 2002; Voorhout et al., 1992; Walker et al., 1986; Weaver et al., 1988). PR8 infection inhibited expression of SP-B, but did not alter the expression of SP-A, C, D, or other ATII-specific proteins (Table 5).

Table 5.

ATII phenotypic proteins identified by proteomics analysis

		Positive
		Number of	Samples1		Fold Change2
Uniprot ID	Protein Name	Peptides	Mock	PR 8	(P- value	Reference
Q02496	Mucin-	3	4	4	1.009 (0.974)	Gonzalez 2005
Q8VCM7	Fibrino gen gamma chain	5	4	4	0.955 (0.865)	Gonzalez 2005
P35242	Surfact ant protein A	3	4	4	0.891 (0.614)	Walker 1986
P50405	Surfact ant protein B	8	4	4	*0.37 (<0.001)	Weaver 1988
P21841	Surfact ant protein C	2	4	4	0.905 (0.771)	Kalina 1992
P50404	Surfact ant protein D	5	4	3	0.893 (0.774)	Voorhout 1992
P14094	Na/K-transporting ATPase, beta-1	10	4	4	0.865 (0.328)	Gonzalez 2005
Q91VS7	Microsomal glutathi one S-transferase 1	10	4	4	0.494 (0.068)	Gonzalez 2005

¹Number of replicate samples (of four for each treatment) in which peptides were detected.

²Fold change expressed as ratio of PR8-infected/mock, P-value determined by Student’s t-test

^*Significantly regulated by PR8 infection.

ATII cells are capable of trans-differentiation into cells with an ATI phenotype in the alveoli and during in vitro culture (Adamson and Bowden, 1974; Dobbs et al., 1988). Therefore, we also searched the proteomics data for proteins that are specific to ATI cells. Two proteins commonly used to distinguish ATI from ATII cells, T1α and aquaporin 5, were not detected in our samples (Nielsen et al., 1997; Williams et al., 1996). However, other proteins that have been identified in ATI, but not ATII, cells by immunohistochemistry in rat or human lung tissue: tissue inhibitor of metalloproteinase (TIMP3) and agrin (Dahlin et al., 2004; Gonzalez et al., 2005), were detected in our ATII cell cultures (Supplemental Table). Three additional ATI markers, SPARC, caveolin-1, and carboxypeptidase M, were identified by single peptides in our analysis (Supplemental Dataset). Thus, our ATII cultures express some proteins that are specifically expressed by ATI cells. This is not surprising because the ATII cells were cultured for six days prior to proteomics analysis. Although culture on matrigel/collagen with keratinocyte growth factor supports maintenance of the ATII cell phenotype, some cells may undergo trans-differentiation to an ATI cell phenotype (Kebaabetswe et al., 2013). None of the ATI-specific proteins that were identified had significantly altered levels in PR8-infected samples (Supplemental Table).

Analysis of SP-B expression by PR8-infected ATII cells

SP-B is required for the production of functional pulmonary surfactant, which is essential for preventing alveolar collapse (Hawgood, 2004). Our proteomic analysis indicated significantly reduced expression of SP-B in PR8-infected ATII cells (Fold change 0.37; P-value <0.001) (Table 5). Due to the critical role of SP-B in the biosynthesis of pulmonary surfactant, we further validated the LC-MS data by immunofluorescence and immunoblot analysis of PR8-infected ATII cells in vitro. Mock-inoculated cells had distinct, punctate immunofluorescence, indicating the SP-B in cytoplasmic lamellar bodies (Fig. 4A, C and D). In contrast, there were very few cells with SP-B-specific fluorescence in PR8-infected cultures (Fig. 4B). Immunoblotting demonstrated the absence of the 43 kDa pro-SP-B band in PR8-infected ATII cell lysates (Fig. 4E), indicating that the sparse antigen detected by immunofluorescence assay was below the limit of detection of the immunoblot. In agreement with the immunofluorescence assay, peptides that mapped to SP-B were also detected in PR8-infected samples by mass-spectrometry (Table 5). Thus, both immunofluorescence and immunoblotting assays confirmed the reduced expression of SP-B in PR8-infected ATII cultures demonstrated by proteomic analysis.

Fig. 4. PR8 infection inhibits expression of SP-B in ATII cells.

ATII cells were mock-treated or infected with PR8. AT 24 h post-infection, cells were analyzed for SP-B expression. (A–D) Immunofluorescence assay with SP-B labeled red and nuclei stained with DAPI (blue). Panels A and B show epifluorescence microscopy of 40X fields in (A) mock and (B) PR8-infected ATII cells. The dashed rectangle in (A) is magnified in panel C to show the punctate pattern of fluorescence in SP-B expressing cells. (D) Imaging of SP-B signal in mock-inoculated ATII cells by confocal microscopy. Scale bar indicates 5 µm. The projected Z-stack and its corresponding orthogonal views are indicated by the cross lines. (E) Immunoblot analysis, with band sizes in kDa shown on the left. Densitometry was used to quantify band intensities, which are expressed as a ratio of the intensity of the SP-B band to the corresponding band for β-actin, after background correction. The data are representative of two replicate experiments.

The mature form of SP-B is a small hydrophobic peptide that organizes the phospholipids of surfactant, resulting in a structure that effectively reduces the surface tension at the air-liquid interface of the alveoli. Targeted disruption of the SP-B gene in mice and hereditary SP-B deficiency in humans are associated with alterations in surfactant production and lethal respiratory failure at birth (Clark et al., 1995; deMello et al., 1994; Nogee et al., 1994). In contrast, deletion of each of the other three surfactant protein genes in mice alters surfactant homeostasis, but does not result in neonatal death (Botas et al., 1998; Glasser et al., 2001; Korfhagen et al., 1996). Furthermore, studies have shown that polymorphisms in SP-B are associated with severe disease upon infection by influenza A virus (To et al., 2014), respiratory syncytial virus (Puthothu et al., 2007), or non-infectious causes of acute respiratory distress (Currier et al., 2008; Gong et al., 2004). Down-regulation of SP-B during a pulmonary influenza virus infection may disrupt surfactant homeostasis and contribute to respiratory distress.

The decreased SP-B levels in PR8-infected samples may be a direct result of viral infection or an indirect effect of proinflammatory cytokines produced by the infected cells. Expression of SP-B is regulated by various growth factors, hormones, and inflammatory mediators, including ceramide, nitric oxide, and cytokines (Salinas et al., 2003; Sparkman et al., 2006; Whitsett et al., 1992a; Wispe et al., 1990). Reduced expression of SP-B has been observed in multiple systems associated with pulmonary inflammation, including infections and endotoxin-induced acute lung injury (Beers et al., 1999; Ingenito et al., 2001; Kerr and Paton, 1999). Tumor necrosis factor (TNF-α) is an inflammatory cytokine that decreases expression of SP-B in cultured respiratory epithelial cells and in murine lungs (Bein et al., 2013; Pryhuber et al., 1996; Whitsett et al., 1992b; Wispe et al., 1990). IFN-γ and IL-13 have also been shown to down-regulate production of the mature SP-B protein by human and rat alveolar ATII cells, respectively (Ito and Mason, 2010). While secreted proteins, including cytokines, were not evaluated by the current study, we previously demonstrated that PR8 infection of murine alveolar epithelial cells induces dramatic upregulation of TNF mRNA expression (Kebaabetswe et al., 2013). In addition to TNF, expression of other proinflammatory cytokines is induced by PR8 infection of murine alveolar epithelial cells (Kebaabetswe et al., 2013). The potential role of these cytokines in down-regulation of SP-B production during PR8 infection is the focus of ongoing studies.

Materials and Methods

Virus preparation

Human influenza A/PR/8/34/H1N1 virus (PR8) was obtained from BEI Resources. Madin-Darby canine kidney (MDCK) cells were purchased from American Type Culture Collection. PR8 was propagated and titrated by plaque assay in MDCK cells in media containing 0.1% BSA and TPCK-trypsin (Sigma-Aldrich, St. Louis, MO; 1 ug/ml).

Alveolar type II cell culture and virus infections

Animal protocols were approved by the University of Idaho Animal Care and Use Committee according to the National Research Council Guide for the Care and Use of Laboratory Animals. ATII cells were isolated from 6–8 week old C57BL/6 mice as previously described (Kebaabetswe et al., 2013). To maintain an ATII cell phenotype, ATII cells from 6 mice were pooled and cultured on millicell inserts (EMD Millipore, Billerica, MA) coated with 70% rat tail collagen and 30% BD Matrigel (BD Biosciences, San Jose, CA), in DMEM/10% FBS supplemented with keratinocyte growth factor (KGF; 10 ng/ml; ProSpec, Rehovot, Israel) for 5 days. Quadruplicate cultures of cells were inoculated with PR8 (MOI 0.1-1) in the presence of TPCK-trypsin or mock treated with media for 24 h.

Mass spectrometry sample preparation, data acquisition, and analysis

Four samples each of mock-inoculated and PR8-infected ATII cells were isolated from collagen/matrigel by dissolving the matrix in collagenase and dispase. The cells were washed in 150 mM ammonium bicarbonate buffer followed by lysis in 8 M urea, which also served to inactivate PR8. The samples were then sent to the Pacific Northwest National Laboratory (Richland, WA) on dry ice for proteomics analyses in duplicate using LC-MS and the AMT tag approach (Zimmer et al., 2006) (see supplemental methods for full experimental details). The peak area values (i.e., abundances) for the final peptide identifications were processed in a series of steps using MatLab® R20111 that included quality control, normalization, protein quantification, and comparative statistical analyses. Peptides were normalized using a statistical procedure for the analysis of proteomic normalization strategies (SPANS) that identifies the peptide selection method and data scaling factor that introduces the least amount of bias into the dataset (Webb-Robertson et al., 2011). The peptide abundance values were normalized across the technical replicates with a rank invariant peptide subset using a mean centering of the data. Normalized log₁₀ abundance values were averaged across all technical and biological replicates (n = 8). Peptides were evaluated with a 2-sample t-test and a G-test to identify quantitative and qualitative significance patterns, respectively, in the peptide data (Webb-Robertson et al., 2010). Peptide level significance patterns were used for protein roll-up to select appropriate peptides for protein quantification. Proteins were quantified using a standard R-Rollup method (Matzke et al., 2013; Polpitiya et al., 2008) using the most abundant reference peptide, after filtering the peptides that were redundant, had low data content, or were outside the dominant significance pattern. Protein roll-up identified 1,156 mouse proteins and 6 viral proteins. Comparative statistical analyses of control and infected samples were performed using a 2-sample t-test to assess differences in protein average abundance, and a G-test to assess associations among factors due to the presence/absence of response. Identification of IFN-regulated proteins was performed using the Interferome v.2.0 database (http://interferome.its.monash.edu.au/interferome/) (Rusinova et al., 2013).

Immunoblotting and densitometry analysis

Mock and PR8 infected ATII cells were lysed in NP40 lysis buffer (1% NP40, 0.1% SDS, 0.5% sodium-deoxycholate, protease inhibitors). Equal volumes of samples were electrophoresed on 4–20% SDS page gels (Thermo Scientific, Rockford, IL) and transferred to polyvinylidene difluoride (Thermo Scientific) or nitrocellulose (Schleicher and Schuell, Keene, NH) membranes. After blocking with either 5% BSA or milk, membranes were incubated with antibodies against BST-2 (Santa Cruz Biotechnology, Dallas, TX), ezrin (Abcam, Cambridge, MA), histone H2B (Abcam), SP-B (Santa Cruz Biotechnology), IFITM3 (Santa Cruz Biotechnology), or TLR3 (eBioscience, San Diego, CA). With the exception of TLR3, all blots were incubated with either anti-rabbit (Thermo Scientific) or anti-goat (Southern Biotech, Birmingham, AL) horseradish peroxidase (HRP)-conjugated secondary antibodies. Anti-TLR3 was detected with biotinylated anti-rat IgG (eBioscience) followed by avidin-HRP (eBioscience). A β-actin antibody conjugated to HRP was used as a loading control (Abcam). HRP was detected on film with Supersignal West Pico chemiluminescence substrate (Thermo Scientific). Protein bands were quantified by densitometric scanning of exposed film and analysis with ImageJ (National Institutes of Health, Bethesda, Maryland). For each protein, the specific densitometry was determined by subtracting the background signal using areas adjacent (above and below) to the specific protein band and dividing by the corresponding background-corrected β-actin value.

Indirect immunofluorescence assay

After 24 h, PR8 and mock-infected cells were fixed with 4% formaldehyde and permeabilized with 0.2% Triton X-100. Viral antigen was detected with goat antiserum NR-3148, which recognizes the hemagglutinin protein of PR8 (BEI Resources). SP-B was detected using a goat polyclonal SP-B antibody (Santa Cruz Biotechnology). Anti-goat-555 secondary antibody (Invitrogen, Carlsbad, CA) was used to detect both the PR8 infection and SP-B. Cells were stained with 4',6-diamidino-2-phenylindole (DAPI) to visualize the nuclei and were photographed on a Nikon Eclipse Fluorescence Microscope with Hamatsu digital camera and Metamorph software (Molecular Devices, Sunnyvale, CA). Confocal imaging was performed on an Olympus Fluoview 1000 upright microscope with a 60X Plan Apo oil objective lens and FluoView ASW software. Stacked images were captured using step sizes ranging from 0.32 µm to 0.42 µm. Z-stack analysis was performed in Imaris version 7.2.1 (Bitplane, Zurich, Switzerland).

Research Highlights.

• Proteomic analysis was performed on primary murine ATII cells infected by PR8.

• Proteins expressed by ATII cells in vivo are expressed by primary ATII cells.

• PR8 increases expression of antiviral and class I MHC proteins in ATII cells.

• PR8 decreases expression of surfactant protein B in ATII cells.

• Reduced SP-B levels may contribute to severe lung damage during influenza.