Cellular networks involved in the influenza virus life cycle

Abstract

Influenza viruses cause epidemics and pandemics. Like all other viruses, influenza viruses rely on the host cellular machinery to support their life cycle. Accordingly, identification of host functions that participate in viral replication steps is of great interest to understand the mechanisms of the virus life cycle as well as to find new targets for the development of antiviral compounds. Multiple laboratories have used various approaches to explore host factors involvement in the influenza virus replication cycle. One of the most powerful approaches is an RNAi-based genome-wide screen, which has thrown new light on the search for host factors involved in virus replication. In this review, we examine the cellular genes identified to date as important for influenza virus replication in genome-wide screens, assess pathways that were repeatedly identified in these studies, and discuss how these pathways might be involved in the individual steps of influenza virus replication, ultimately leading to a comprehensive understanding of the virus life cycle.

Introduction

Viruses, which consist of nucleic acid encased in a protein shell, are parasites of their host organisms. Despite their simple structures, viruses have sophisticated mechanisms to replicate themselves in their hosts. A ‘host cellular factory’ has thousands of machines, which viruses use during each step of their life cycle. Viruses generally initiate their life cycle by attaching to host cell surface receptors, entering the cells and uncoating their viral nucleic acid. They then replicate their viral genome. After new copies of viral proteins and genes are synthesized, these components assemble into progeny virions, which then exit the cell. For each step, viruses use many host cellular functions. Identifying these host functions has long been of great interest in virology to further our understanding of the precise mechanisms of the viral life cycle.

Influenza viruses cause annual epidemics and recurring pandemics with potentially severe consequences for public health and the global economy. The first influenza pandemic of the 21^st century was caused by the 2009 H1N1 virus and has so far resulted in 42–86 million cases of infection worldwide (http://www.cdc.gov/h1n1flu/estimates_2009_h1n1.htm). In addition, highly pathogenic avian H5N1 influenza A viruses have spread throughout Asia, Europe, and Africa, overcoming host species barriers to infect humans, with fatal outcome in many cases (Webster and Govorkova, 2006; Yen and Webster, 2009). Antiviral drugs, such as oseltamivir, zanamivir (Hayden, 2001), and amantadine/rimantadine (Davies et al., 1964), are available for prophylaxis and treatment of influenza virus infection; however, most human H3N2 and H1N1, including pandemic 2009 H1N1, influenza viruses are resistant to amantadine/rimantadine (Bright et al., 2005; Bright et al., 2006; Dawood et al., 2009). Moreover, the frequency with which H1N1 influenza viruses are becoming oseltamivir-resistant is a cause for concern and highlights the urgent need for new antiviral drugs (Nicoll et al., 2008) (http://www.who.int/csr/disease/influenza/h1n1_table/en/index.html).

Influenza A viruses are enveloped negative-strand RNA viruses with eight RNA segments encoding at least 10 viral proteins (reviewed in (Palese, 2007)). Two additional viral proteins, PB1-F2 and PB1 N40, have been identified, although not all strains encode these proteins (Chen et al., 2001; Wise et al., 2009). The virus particles are enclosed by a lipid envelope, which is derived from the host cellular membrane. Three viral proteins, the surface glycoproteins hemagglutinin (HA) and neuraminidase (NA), and the M2 ion channel protein, are embedded in the lipid bilayer of the viral envelope. The HA protein binds to sialic acid-containing receptors on the host cell surface and mediates fusion of the viral envelope with the endosomal membrane after receptor-mediated endocytosis (Palese, 2007; Skehel and Wiley, 2000). By contrast, the NA protein plays a crucial role late in infection by removing sialic acid from sialyloligosaccharides, thereby releasing newly assembled virions from the cell surface and preventing the self-aggregation of virus particles. Underneath the lipid bilayer lies the matrix protein (M1), a major structural protein. Within the virus shell are eight viral ribonucleoprotein (vRNP) complexes, each composed of viral RNA (vRNA) associated with the nucleoprotein NP and the three components of the viral RNA polymerase complex (PB2, PB1, and PA). The NS1 protein, which counteracts the cellular interferon response, is synthesized from an unspliced mRNA, while a spliced mRNA yields the NS2, or NEP, protein, which mediates nuclear export of vRNP complexes. In addition, the recently identified PB1-F2 protein, which is encoded by the PB1 segment, is thought to play a role in viral pathogenicity (Chen et al., 2001; Conenello et al., 2007; McAuley et al., 2007; Zamarin et al., 2006).

While the functions of the viral proteins have been studied extensively over the last decade, relatively little is known about the cellular factors involved in influenza virus life cycle. Here, we review recent genome-wide screens that aim to close this critical gap in influenza virus research.

Dawn of a new era: The application of genome-wide screens to identify host factors involved in influenza virus replication

A number of interactions between viral components and specific host cell gene products have now been identified. Although most host molecules remain elusive, emerging data indicate that their identification and characterization will provide new insights into the mechanisms by which viruses complete their life cycle. Moreover, such knowledge would illuminate potentially useful targets for therapeutic intervention. Indeed, antiviral drugs targeting host cell factors involved in viral replication have been tested for the treatment of human immunodeficiency virus type I (HIV-1), with promising results (Coley et al., 2009). However, this goal would generally take several decades to achieve with conventional genetic screening methods and mammalian cell cultures.

Genome-wide RNAi screens

Many laboratories have expended great effort in the search for new host factors involved in the virus replication cycle by using various strategies. One of the most powerful approaches is systematic, genome-wide RNA interference (RNAi) analysis. RNAi is a regulatory mechanism that uses double-stranded RNA (dsRNA) molecules to direct homology-dependent suppression of gene activity (reviewed in (Mello and Conte, 2004)). This powerful cellular process has been extensively applied for studies of gene functions and therapeutic approaches for disease treatment. Now, this technology offers a new exciting tool for the exploration of novel host genes involved in virus replication (Brass et al., 2008; Brass et al., 2009; Cherry et al., 2005; Goff, 2008; Hao et al., 2008; Konig et al., 2010; Konig et al., 2008; Krishnan et al., 2008; Li et al., 2009; Zhou et al., 2008).

The first genome-wide screen for the identification of host factors involved in influenza virus replication was reported by Hao et al. (2008) (Table 1). Since RNAi-based screening was not well-established in mammalian cells at that time, the authors used Drosophila RNAi technology. Studies in Drosophila have made numerous fundamental contributions to our understanding of mammalian cell biology because of the high degree of genetic conservation between Drosophila and vertebrates. Hao et al. (2008) first established an influenza virus infection system in Drosophila cells by modifying the influenza virus genetically to infect Drosophila cells and express a reporter gene product in infected cells. This system supported influenza virus replication from post-entry to the protein expression step in the life cycle (Table 1). The authors tested an RNAi library against 13,071 genes (90% of the Drosophila genome) and identified 110 genes whose depletion in Drosophila cells significantly affected reporter gene expression from the influenza virus-like RNA (see Supplementary Table S3 for the list of genes). Of those 110 candidates, they validated roles for the host proteins ATP6V0D1 (an ATPase), COX6A1 (a cytochrome C oxidase subunit), and NXF1 (a nuclear RNA export factor) in the replication of H5N1 and H1N1 influenza A viruses in mammalian cells. These studies showed the feasibility and power of genome-wide RNAi screens to identify previously unrecognized host proteins required for influenza virus replication.

Table 1.

Summary of genome-wide screens for the identification of host factors involved in influenza virus replication.

Reference	Methods (approach)	Cells used for RNAi screen	Virus strain	Detection methods	Steps in the viral life cycle covered by the screen	Number of hits/number of genes tested
Brass et al. (2009)	RNAi screen	Osteosarcoma cells (U2OS)	A/Puerto Rico/8/34 (PR8; H1N1)	HA expression on cell surface	Entry Uncoating Nuclear import Transcription Translation HA trafficking to cell surface	133/17,877
Hao et al. (2008)	RNAi screen	Drosophila cells	Recombinant A/WSN/33 (WSN; H1N1) virus possessing vesicular stomatitis virus glycoprotein G and NA-Luciferase (Luc) genes	Luciferase activity	Uncoating Nuclear import Transcription Translation	110/13,071
Konig et al. (2010)	RNAi screen	Human lung cells (A549)	Recombinant WSN possessing HA-Luc reporter gene	Luciferase activity	Entry Uncoating Nuclear import Transcription Translation	295/19,628
Karlas et al. (2010)	RNAi screen	Human lung cells (A549)	WSN (H1N1) H1N1 2009	Virus yield determined by NP staining and luciferase activity	Entry Uncoating Nuclear import Transcription Translation Nuclear export Protein Transportation/processing Assembly Budding	287/22,843
Shapira et al. (2009)	Y2H Microarray RNAi screen	HBEC	PR8 (H1N1) A/Udorn/307/72 (H3N2)	Virus yield determined by luciferase activity	Entry Uncoating Nuclear import Transcription Translation Nuclear export Protein transportation/processing Assembly Budding	616/1,745 (RNAi screen)
Sui et al. (2009)	RHGP	MDCK cells	Udorn (H3N2)	Cloning of cells resistant to influenza virus infection	Entry Uncoating Nuclear import Transcription Translation Nuclear export Protein transportation/processing Assembly Budding	110 clones, sequenced

The development of mammalian RNAi-based screening now enables the comprehensive analysis of mammalian host cell functions in influenza virus replication. Recently, Brass et al. (2009), Konig et al. (2010), and Karlas et al. (2010) reported genome-wide RNAi screens in mammalian cells that identified host proteins important for influenza virus replication (Table 1). They tested siRNAs targeting more than 17,000 human genes in human osteosarcoma (Brass et al., 2009) and the human lung cell line A549 (Karlas et al., 2010; Konig et al., 2010), respectively. Because the readout of the assay was the measurement of a viral protein or a reporter protein encoded by the viral genome, the systems used by Brass et al. (2009) and Konig et al. (2010) allowed the identification of host gene products important for the early to mid stages of the influenza virus life cycle, including virus entry, uncoating, vRNP nuclear import, genome transcription, and viral protein translation. These systems identified 133 (Brass et al., 2009) and 295 (Konig et al., 2010) human genes whose depletion affected the efficiency of HA cell surface expression or reporter gene expression, respectively (Table 1; see Supplementary Table S3 for the list of genes). In contrast, Karlas et al. (2010) studied the entire viral replication cycle, from viral entry to budding, by determining virus infectivity titers in culture supernatants from siRNA-treated, virus-infected A549 cells. This approach identified 287 human genes important for influenza virus replication (Karlas et al., 2010) (see Supplementary Table S3 for the list of genes).

Shapira et al. (2009) employed a different tactic. They combined the results of yeast two-hybrid analyses, genome-wide transcriptional gene-expression profiling, and an RNAi screen (Shapira et al., 2009). First, they built a physical interaction map for viral proteins and human cellular molecules by applying computational analysis to their comprehensive yeast two-hybrid assay. This physical map indicated that the viral proteins of influenza interacted with ‘a significantly greater number of human proteins than expected from the human interaction network, even when compared to other viruses’ (Shapira et al., 2009), suggesting influenza viruses have to maximize the diversity of functions for each protein (Shapira et al., 2009). They then examined which cellular gene products were differentially expressed in primary human bronchial epithelial cells exposed to influenza virus or viral RNA. Collectively, these approaches identified 1,745 potential host factors with roles in the influenza virus life cycle (Shapira et al., 2009). Further validation with using siRNAs targeting to 1,745 genes narrowed this list to 616 human genes whose products affected influenza virus replication.

Other approaches

Proteomics approaches and yeast two-hybrid analyses have also been used to identify host molecules that interact with influenza virus components (reviewed in (Nagata et al., 2008)). Several host factors involved in the steps of viral genome replication and transcription have been identified (Deng et al., 2006; Engelhardt et al., 2005; Huarte et al., 2001; Jorba et al., 2008; Kawaguchi and Nagata, 2007; Momose et al., 2001; Momose et al., 2002; Resa-Infante et al., 2008), as well as NS1 protein-interacting host factors associated with immune responses (reviewed in (Hale et al., 2008)). Using mass spectrometry, Mayer et al. (2007) recently identified 41 host cellular molecules that interact with influenza vRNPs, including nucleoplasmin which may play a role in viral RNA synthesis.

A library encompassing about 4,800 yeast single-gene deletion strains, which covers 80% of all yeast genes, has been used to explore host gene involvement in virus replication, especially that of several positive-strand RNA viruses (Kushner et al., 2003; Panavas et al., 2005). Naito et al. (2007) developed a system for influenza virus genome replication and transcription in Saccharomyces cerevisiae, and conducted a screen with a sub-library of 354 deletion strains, each of which lacked a gene encoding a putative nucleic acid-binding/-related functional protein. This approach resulted in the identification of several host factors that affected influenza RNA synthesis (Naito et al., 2007).

Sui et al. (2009) employed a new technique, called ‘Random Homozygous Gene Perturbation’ (RHGP), to identify host factors involved in influenza virus replication. This system uses a lentiviral-based genetic element containing a promoter (e.g., the cytomegalovirus immediate early promoter); the genetic element is designed to integrate at a single site in one allele of the genome in either the sense or antisense orientation. Integration of this element in the antisense orientation disrupts gene expression in one allele. Moreover, because this element contains a promoter, antisense RNA is expressed and can abrogate expression of the same gene in the other allele. If the element is integrated in the sense orientation in one allele, either overexpression of the gene, or expression of a dominant-negative form occurs, leading to an altered phenotype. Unlike RNAi screens, the RHGP system can theoretically knockdown or overexpress any gene, without any prior knowledge or annotation of that gene, resulting in alteration of the phenotype. Sui et al. (2009) generated an RHGP library in mammalian cells, followed by infection with a dose of influenza virus that results in cell death. They then isolated the influenza virus-resistant cell clones and identified 110 human genes that contributed to the host cell resistance to viral infection, demonstrating the usefulness of this RGHP system for genome-wide screening.

Pair-wise comparison among six independent genome-wide screens for human genes involved in influenza virus replication

In the six independent genome-wide screens described above (Brass et al., 2009; Hao et al., 2008; Karlas et al., 2010; Konig et al., 2010; Shapira et al., 2009; Sui et al., 2009), a total of 1,449 human genes were identified as potential host factors involved in influenza virus replication (including 110 Drosophila genes that have human orthologs; Supplementary Table S1A, B), indicating that about 5.8% of human protein-coding genes may contribute to the life cycle of influenza virus. However, there is the possibility that these high-throughput screens may contain false-positive. To narrow down the genes important for influenza viral replication, we compared the candidate genes identified in these screens by pair-wise comparison and found 128 human genes that affected influenza virus replication in at least two screens (Supplementary Table S2). The number of genes common between pairs of screens is relatively small (from 0 to 32 genes in the 15 pair-wise comparisons). This low incidence of overlap has also been observed in three RNAi screens designed to identify host factors for HIV-1 replication (Brass et al., 2008; Konig et al., 2008; Zhou et al., 2008), and it could be due to differences in screening systems (e.g., RNAi screen vs. RHGP library screen), or other conditions such as cell type, virus strains, and/or detection methods. Indeed, the number of common genes was highest in the comparison between the Konig and Karlas screens (32 human genes) (Supplementary Table S2), presumably because both screens were based on RNAi and used the same human lung cells. In addition, the analysis of raw data, particularly the ‘cut-off’ used to determine candidates, likely differed among the studies. Nonetheless, a number of factors were identified in several screens (see below for more details), suggesting their importance in the influenza virus life cycle.

Network analysis of influenza-host protein interactions

For the set of 128 human genes that were identified in two or more screens and are likely involved in influenza virus replication, we next determined major gene categories. Our analysis with the PANTHER Classification system (Mi et al., 2005) revealed several enriched gene categories associated with molecular functions, such as nucleic acid-binding proteins, kinases, transcription factors, ribosomal proteins, hydrogen transporters, and proteins involved in mRNA splicing. Moreover, we found enrichment for cellular proteins involved in biological processes such as protein metabolism and modification, signal transduction, protein phosphorylation, nucleoside, nucleotide and nucleic acid metabolism, and transport. Analysis by Reactome, which is a curated knowledgebase of biological pathways (Matthews et al., 2009), revealed several other over-represented events, including eukaryotic translation initiation, regulation of gene expression, processing of capped intron-containing pre-mRNAs, and Golgi to ER retrograde transport. Further, we analyzed interactions of the set of these 128 genes with using GeneGo (GeneGo Inc, MI, USA) (Supplementary Table S4A). We then integrated this data set with information on the viral and cellular interaction partners of viral proteins as reported by Konig et al. (2010) and Shapira et al. (2009) (Supplementary Table S4B), resulting in a network of host-influenza virus interactions (Figure 2). In addition, this network included several sub-networks associated with specific biological processes, such as translation initiation, processing of pre-mRNA, and proton-transporting ATPase functions.

Figure 2. A network of host-influenza virus interactions.

Using GeneGo (GeneGo Inc, MI, USA), we analyzed the interactions of the 128 human genes identified to be involved in influenza virus replication in at least two genome-wide screens (Supplementary Table S4A). We then integrated this information with data on host protein-virus interactions and interactions among viral proteins, as reported by Konig et al. (2010) and Shapira et al. (2009) (Supplementary Table S4B). Interaction networks among host and viral proteins (i.e., PA, PB1, PB2, NP, HA, NA, M1, M2, NS1, NEP/NS2, and PB1-F2)(Fig 2A), virions (Fig 2B), or vRNP (Fig 2C) were visualized by using Cytoscape (http://cytoscape.org/). Yellow and lilac nodes indicate influenza viral components and host proteins, respectively. Red circles indicate clusters associated with particular biological processes in the interaction networks, as identified by MCODE (“Molecular Complex Detection”) analysis (Bader and Hogue, 2003): 1) translation initiation, 2) pre-mRNA processing, and 3) proton-transporting V-type ATPase.

Insights into host factors involved in the influenza virus life cycle

Based on the analyses described above, influenza virus-host interactions can be mapped to individual steps of the influenza virus life cycle (Fig. 1), as outlined below.

Figure 1. Influenza virus life cycle and host factors.

Upon infection, influenza virus is internalized by receptor-mediated endocytosis. After membrane fusion between the virus envelope and the endosome, the viral ribonucleoprotein (vRNP) complex is released into the cytoplasm and transported into the nucleus by the active import machinery of the host cell nuclear pore complex. Replication and transcription of the viral genome take place in the nucleus and many host factors are believed to be involved in this process. Newly synthesized viral RNA associates with NP and forms the viral ribonucleoprotein (vRNP) complex together with viral polymerase proteins, which is then transported from the nucleus to the cytoplasm. HA and NA are processed posttranslationally during their transport from the ER to the Golgi apparatus. During assembly and budding, virion components are transported to the assembly site, the lipid raft microdomains at the apical plasma membrane of polarized epithelial cells. Progeny viruses then bud from the cells. The light orange rectangles indicate individual steps of the influenza virus life cycle. The light blue rectangles indicate host cellular processes that may be involved in the virus life cycle. Red circles indicate host factors identified in the screens discussed here, while yellow circles indicate host factors identified in other previous studies.

1) Endocytosis, fusion, and uncoating

Once virus particles attach to a cell surface receptor, they are mainly internalized by clathrin-dependent endocytosis, although influenza virus also uses a non-clathrin-dependent endocytic pathway (Marsh and Helenius, 2006; Matlin et al., 1981; Rust et al., 2004; Sieczkarski and Whittaker, 2002). The importance of both early and late endosomes in influenza virus entry has been demonstrated with dominant-negative forms of the GTPases Rab5 and Rab7/protein kinase C β II (PKCβII), which regulate the functions of early and late endosomes, respectively (Sieczkarski et al., 2003; Sieczkarski and Whittaker, 2003). Rab10 (RAB10), which controls the movement of endosomes (Conenello et al., 2007; Glodowski et al., 2007), was identified in two screens (Supplementary Table S3). Moreover, four of the seven subunits of the coatomer 1 (COPI) vesicular transport complex (i.e., ARCN1, COPA, COPB2, and COPG), which is required for the formation of intermediate transport vesicles between the early and late endosomes and for retrograde Golgi-to-ER transport (Aniento et al., 1996; Cai et al., 2007; Whitney et al., 1995), were identified in five independent screens (Supplementary Table S3), indicating the biological significance of this host cell machinery in the early steps of the influenza virus life cycle.

Uncoating of viral particles occurs upon fusion between the viral and endosomal membranes, is mediated by HA, and leads to the release of vRNPs into the cytoplasm (Stegmann et al., 1990). The M2 ion channel protein functions in the endosomes to lower the pH of virion interior, resulting in the disassociation of the viral matrix M1 protein from the vRNP complex, which can then be imported to the nucleus (Helenius, 1992). Proton-transporting V-type ATPase functions in both the acidification and fusion of the cellular compartments (Stevens and Forgac, 1997). Four of the six screens identified at least two subunits of this V-type ATPase complex as host factors required for influenza virus infection (Supplementary Table S3). Although this complex was thought to be generally involved in the low pH-dependent membrane fusion between endosomes and virions, knockdown of ATP6V0D1 caused significant inhibition of influenza virus production but not VSV production, which also requires a low pH for fusion (Hao et al., 2008). This finding implies a specific role for the complex in influenza virus replication.

2) Transport of vRNP into the host nucleus

Replication/transcription of influenza virus genomic RNA occurs in the nucleus (reviewed in (Engelhardt and Fodor, 2006; Palese, 2007)). Therefore, the RNP complexes released from the incoming virus particles must be transported into the nucleus through the nuclear pores. The nuclear import of RNP complexes is dependent on the active import machinery of the host cell nuclear pore complex (NPC) because of the size of these complexes (10–15 nm wide and 30–120 nm long). In a classic nuclear import pathway, importin α recognizes a cargo protein containing nuclear localization signals (NLS), binds to importin β, and the cargo/importin α/β complex is transported into the nucleus through the NPC. For influenza virus, RanBP5 (or importin-5), importin α1/α2, and importin α1/importin α5 have been shown to bind to PB1-PA dimers (Deng et al., 2006), NP (O’Neill et al., 1995; Wang et al., 1997), and PB2 (Gabriel et al., 2008; Resa-Infante et al., 2008; Tarendeau et al., 2007; Tarendeau et al., 2008), respectively. Our list of 128 genes includes the importin subunit β1 (KPNB1) and nuclear pore complex proteins Nup98 and Nup153 (NUP98 and NUP153). Likely, these factors are involved in the nuclear import of influenza virus RNPs, as well as in the nuclear import of newly synthesized influenza viral proteins.

3) Genome replication/transcription and translation

Once the vRNP complexes are imported into the nucleus, the vRNA of influenza virus is transcribed into mRNA and replicated via complementary RNA (cRNA) (reviewed in (Engelhardt and Fodor, 2006; Palese, 2007)). The viral polymerase subunits (PB1, PB2 and PA) and NP catalyze both genome replication and transcription. Several host cellular molecules play a part in these steps. Host factors shown to stimulate viral RNA synthesis include BAT1 (or UAP56), heat shock protein (Hsp) 90, the minichromosome maintenance (MCM) complex, Tat-SF1, and DNA-dependent RNA polymerase II (PolII) (Chan et al., 2006; Engelhardt et al., 2005; Kawaguchi and Nagata, 2007; Momose et al., 2001; Momose et al., 1996; Momose et al., 2002). BAT1, which is a splicing factor for mRNA and also functions in the nuclear export of cellular mRNA, interacts with NP to facilitate the formation of the NP-RNA complex. Similarly, Tat-SF1 facilitates NP-vRNP complex formation (Naito et al., 2007). Viral RNA synthesis is then initiated by the viral RNA polymerase (Momose et al., 2001; Momose et al., 1996). Hsp90 stimulates viral RNA synthesis through its interaction with PB2 (Momose et al., 1996; Momose et al., 2002). MCM, which functions as a DNA replication fork helicase (Forsburg, 2004), plays a role in influenza viral genome replication by promoting the association between nascent cRNA and the viral polymerase complexes during the transition from initiation to elongation, thus stabilizing replication elongation complexes and allowing the synthesis of full-length cRNA (Kawaguchi and Nagata, 2007). As described above, the yeast library screen identified Tat-SF1 as a factor involved in vRNA-NP complex formation and viral RNA synthesis (Naito et al., 2007). In its active state, PolII, which catalyzes the synthesis of mRNA precursors and most snRNA and microRNAs, is required for viral mRNA synthesis during the viral transcription step (Chan et al., 2006; Engelhardt et al., 2005). Included in the 128 genes identified in our pair-wise comparison (Supplementary Table S2) are genes associated with pre-mRNA splicing (i.e., PTBP1, NHP2L1, SNRP70, SF3B1, SF3A1, P14, and PRPF8), suggesting the importance of the host mRNA splicing machinery for influenza virus gene expression. Interestingly, no splicing factors were found in the screen carried out in Drosophila cells (Hao et al., 2008), which is consistent with the finding that influenza virus mRNAs were not spliced in this system (Hao et al, 2008).

Four screens identified NXF1 as an important host factor in influenza virus replication (Supplementary Table S3). NXF1 exports spliced, cellular mRNAs but rarely exports unspliced cellular mRNAs (Reed and Cheng, 2005). Nonetheless, several viruses use the NXF1-pathway for the nuclear export of unspliced viral RNAs (Cullen, 2003). For influenza virus, NP interacts with BAT1/UAP56(Momose et al., 2001), which forms a bridge between the NXF1/NXT1/Ref complex and the spliced mRNA. In addition, NS1 binds NXF1 to suppress host mRNA export (Satterly et al., 2007), perhaps to retain host mRNAs in the nucleus for ‘cap-snatching’. In addition, the requirement of NXF1 for nuclear export of viral mRNAs has recently been shown (Read and Digard, 2010). Thus, influenza virus may exploit the NXF1-mediated RNA export pathway for multiple purposes.

Influenza virus also uses the cellular machinery for mRNA translation. Influenza virus infection shuts-off host protein synthesis, thereby enhancing translation of viral mRNAs (Chen and Krug, 2000; Chen et al., 1999; Fortes et al., 1994; Nemeroff et al., 1998; Qiu and Krug, 1994). GRSF-1 (the cellular RNA-recognition motif containing the RNA-binding protein G-rich sequence factor 1) and P58IPK (the cellular inhibitor of PKR, an interferon-induced kinase that targetsthe eukaryotic translation initiation factor eIF2) appear to be involved in the selective synthesis of influenza viral proteins (Goodman et al., 2007; Kash et al., 2002; Park et al., 1999). Analysis with Reactome (Matthews et al., 2009) of the set of 128 genes we identified above revealed several host genes involved in the formation of the translation initiation complex and the small ribosomal subunit (i.e., FAU, EIF3S5, RPS3A, RPS4X, RPS10, RPS14, RPS5, RPS16, EIF4A2, RPS20, and EIF3S8). Possibly, some of these host factors promote virus-specific translation by inhibiting host cellular mRNA translation and/or stimulating viral mRNA translation.

4) vRNP transport from the nucleus to the cytoplasm

Newly synthesized vRNPs are exported from the nucleus to the cytoplasm (reviewed in (Boulo et al., 2007)). The viral matrix protein, M1, and nuclear export protein, NEP/NS2, have critical roles in this step. After M1 binds to the RNP complex, NEP/NS2 attaches to the M1/RNP complex through its C-terminal domain (Akarsu et al., 2003). NEP/NS2 can also bind to a host cell nuclear export protein, CRM1 (Boulo et al., 2007; Elton et al., 2001; Neumann et al., 2000). Additionally, heat shock cognate (Hsc) 70 protein and the MAP kinase cascade may have roles in the nuclear export of vRNP complexes (Pleschka et al., 2001; Watanabe et al., 2006).

Recently, phosphatidylinositol-3-kinase (PI3K) and its downstream effector protein (the kinase Akt) were identified as host factors involved in influenza virus-induced signaling (Ehrhardt and Ludwig, 2009; Ehrhardt et al., 2006; Hale et al., 2006; Shin et al., 2007; Zhirnov and Klenk, 2007). Consistent with this finding, three genes involved in the PI3K/AKT signaling pathway (i.e., AKT1, MDM2, IKBKE) were among the 128 genes identified by the screens discussed here. Roles for this pathway in antiviral activity and prevention of premature apoptosis have been reported (Ehrhardt et al., 2006; Ehrhardt et al., 2007; Sarkar et al., 2004). Interestingly, a PI3K inhibitor obstructs the nuclear export of vRNPs, as well as viral RNA and protein synthesis (Shin et al., 2007).

5) Assembly and budding

In the late stage of the life cycle, the virion components, including vRNPs, the matrix protein M1, and the viral envelope proteins (HA, NA and M2) are transported to the assembly site, the apical plasma membrane in polarized epithelial cells (Nayak et al., 2004). Relatively few host factors have been identified to be involved in this stage. Of the six genome-wide screens discussed in this review, only two studies (Karlas et al., 2010; Shapira et al., 2009) were designed to evaluate the assembly and budding steps. Depletion of subunits of the COPI complex reduced HA expression on the cell surface relative to the total HA level in the screen by Brass et al. (2009)(Fig. 1), suggesting that the COPI complex plays a role in the transport of influenza viral glycoproteins (i.e., HA and NA) to the cell surface, in addition to its function in virus entry described above. The HA and NA proteins contain signals for association with lipid rafts, which are non-ionic detergent-resistant lipid microdomains within the plasma membrane. Influenza viruses preferentially bud from these lipid raft microdomains (Barman et al., 2001; Leser and Lamb, 2005; Nayak et al., 2004; Takeda et al., 2003; Zhang et al., 2000). Although the precise mechanism of M1-vRNP complex transport to the cell surface is still unclear, the viral M2 protein is thought to play a role in the incorporation of M1-vRNPs into virus particles via an interaction between M1 and the M2 cytoplasmic tail (Iwatsuki-Horimoto et al., 2006; McCown and Pekosz, 2005, 2006; Wu and Pekosz, 2008). NP and M1 interact with host cytoskeletal components, such as actin, presumably facilitating vRNP transport to the budding site (Avalos et al., 1997). In the final step of the life cycle, the viral membrane components are concentrated in the microdomains of the lipid rafts, which may provide the stage for virus particle release from the cell surface (Takeda et al., 2003). Although few host factors have been found that participate in the budding step, the actin cytoskeleton and Rab11, which belongs to the Rab family of small GTPases, have been shown to contribute to the budding of filamentous influenza virus particles (Bruce et al., 2010; Roberts and Compans, 1998; Simpson-Holley et al., 2002).

Perspectives and Concluding Remarks

In 2003, the human genome project completed its mission of identifying the approximately 20,000–25,000 genes in human DNA. To determine host gene involvement in virus replication, genome-wide screens, such as RNAi screens, offer an effective approach. For influenza virus, among six independent genome-wide screens, 1,452 human genes, including 110 human orthologs identified by a screen in Drosophila cells, have been identified as host factors involved in influenza virus replication; of these, 128 genes were found in at least two screens. Bioinformatics analysis of these 128 human genes revealed gene clusters related to host cellular functions, such as endocytosis, translation initiation, and nuclear transport, all of which can be connected to the influenza viral life cycle. Fig. 1 illustrates influenza virus-host interactions mapped to individual steps of the influenza virus life cycle based on the known functions of the identified host proteins. However, it should be borne in mind that these proteins may be involved in other steps of the viral life cycle with, as yet, undetermined functions.

Systematic, genome-wide RNAi screens are powerful tools for identifying host genes important for viral replication. However, they do have limitations; for example, the RNAi library does not yet covered all of the human genes, the knockdown efficiency of target gene expression could vary among siRNAs, and genes whose siRNA knockdown leads to cytotoxicity would be eliminated from a ‘hit-list’ even though they may be important for viral replication, resulting in a number of false-positives and false-negatives. Therefore, the identification of host gene candidates involved in viral replication by any primary screen really represents no more than a starting point for exploration of such host factors. More detailed functional analyses of human genes identified in genome-wide screens will allow us to find novel cellular pathways and/or host gene sets important for influenza virus replication, leading to a greater understanding of the precise mechanisms of influenza virus replication.