Genome Scale Transcriptomics of Baculovirus-Insect Interactions

Baculovirus-insect cell technologies are applied in the production of complex proteins, veterinary and human vaccines, gene delivery vectors‚ and biopesticides. Better understanding of how baculoviruses and insect cells interact would facilitate baculovirus-based production. While complete genomic sequences are available for over 58 baculovirus species, little insect genomic information is known. The release of the Bombyx mori and Plutella xylostella genomes, the accumulation of EST sequences for several Lepidopteran species, and especially the availability of two genome-scale analysis tools, namely oligonucleotide microarrays and next generation sequencing (NGS), have facilitated expression studies to generate a rich picture of insect gene responses to baculovirus infections. This review presents current knowledge on the interaction dynamics of the baculovirus-insect system‚ which is relatively well studied in relation to nucleocapsid transportation, apoptosis, and heat shock responses, but is still poorly understood regarding responses involved in pro-survival pathways, DNA damage pathways, protein degradation, translation, signaling pathways, RNAi pathways, and importantly metabolic pathways for energy, nucleotide and amino acid production. We discuss how the two genome-scale transcriptomic tools can be applied for studying such pathways and suggest that proteomics and metabolomics can produce complementary findings to transcriptomic studies.


Introduction
Baculoviruses are enveloped viruses that contain a circular double stranded DNA genome (80-180 kb) [1]. Baculoviruses are specifically infectious to invertebrates, most commonly insect species belonging to the Lepidoptera order, and there is no evidence of cross-order transmission [2]. Baculoviruses can also enter a range of mammalian cell types including human, primates, rodent, rabbit, porcine, bovine, and other animals such as fish and avian species, but neither replicate nor transpose baculovirus DNA into host chromosomes in these cell types [3,4]. Hence, the viruses are safe to vertebrates and have been approved by the US Food and Drug Administration and European Medicine Agency for manufacturing of veterinary and human vaccines [5].
Over 40 years of research on molecular biology of baculoviruses has produced a rich knowledge about baculovirus gene expression profiles and gene functions during infection, including: virus entry, movement into the cell nucleus, DNA unpackaging, early transcription, DNA replication, late transcription, translation, budded virus (BV) assembly, BV budding, occlusion derived virus (ODV) assembly, occlusion, and release of occlusion bodies (for an overview of baculovirus molecular biology studies, refer to Rohrmann [6]). However, little is known about host responses towards baculovirus infections.
Baculoviruses exert vigorous effects on host cells. Within a short infection time frame, the viruses rapidly propagate and change host cells by arresting host cell cycle progression, creating a "viral pseudo S phase", remodeling the cellular cytoskeleton, forming virogenic stroma in the nucleus as a specialized compartment for replication, undertaking transcription and packaging, and eventually breaking-down the host nuclei. In the Helicoverpa armigera nucleopolyhedrovirus (HearNPV)-H. zea system, within 48 hours post infection (h.p.i), baculoviruses can produce over 250,000 genomes per cell, which is 20 times more DNA than that produced in an uninfected host cell, and the total viral mRNA from 134 virus genes is higher than the total cellular mRNA produced from over 15,000 genes during normal cell growth [7]. Similarly, in the Autographa californica multicapsid nucleopolyhedrovirus (AcMNPV) infected Spodoptera frugiperda cell line (Sf9) and AcMNPV infected Trichoplusia ni cell line (Tnms42), as high as one million viral genomes per haploid host insect genome was reported [8].
Although infected cells are controlled by baculoviruses, and a global reduction of host proteins and mRNAs occurs, host cells also respond by up-regulating or repressing certain genes and cellular pathways. A number of genes escape from the global expression shutoff and several host cellular pathways such as DNA damage responses, heat shock responses, pro-survival responses, apoptosis, energy metabolism, iron-ion transport, ubiquitin-proteasomal degradation and miRNA pathways have been shown to be up-regulated following infections by recent studies [9][10][11][12][13].
Transcriptomics is a powerful tool to study baculovirus-insect systems at the mRNA level, which is arguably the most important level of regulation during a baculovirus infection. Previous studies by Iwanaga et al. [14] using BmNPV infected cells suggested that infection events in the host cell were strongly regulated at the mRNA expression level rather than at the translational level. Gatehouse et al. [15] proposed that protein and RNA degradation are the key processes for controlling the global shutoff pattern induced by a baculovirus infection. In addition, polysome profiles, which were mRNA-ribosome complexes extracted by sucrose gradient centrifugation of cytosolic fractions of infected cells, were similar between Sf9 cells with and without an AcMNPV infection, indicating that active translation activities take place in infected cells even at late infection stages (16 h.p.i) [16]. Usually‚ it was found that changes in protein levels were due to prior changes in mRNA levels. Huynh et al. [17] showed that for high cell density infections in cultures, a reduction in the expression of the β-galactosidase protein in Sf9 cells infected by a recombinant AcMNPV virus was due to early infection events, most likely leading to a decrease in mRNA levels following a reduction in virion genome numbers.
To study host responses, the lack of insect genome sequences is a challenge. To overcome this challenge, rapid developments of RNA sequencing (RNA-seq) technologies for generating genome-scale transcript sequences in non-model organisms can be utilized [18,19]. This review shows that most of our understanding about insect host responses to baculovirus infections so far has been derived from microarray and RNA-seq studies in the last five years. These studies, however, investigated different host-baculovirus systems, under different infection conditions, and hence identified different sets of regulated genes. This review identifies possible conservative changes and knowledge gaps relating to host-baculovirus interactions, and proposes the further use of transcriptomics to investigate genome-scale interactions, with more focus on host responses. Finally‚ the review briefly discusses challenges and opportunities for genetic engineering of host genes, determined by transcriptomic studies, to moderate the infection process in order to increase product yields using insect cell technology, which is important for commercialization of the technology. For expression of some recombinant proteins, efficacy remains the key challenge in relation to commercial production but for many protein products and certainly for wild type baculoviruses to be used as biopesticides, yield improvements are essential for insect cell technology to be competitive in relation to alternative manufacturing systems.

Transcriptomic Studies
Transcriptomic studies of baculovirus gene expression have expanded current understanding extensively. One of the first microarray studies for baculovirus genes was by Yamagishi et al. [20], in which PCR amplified fragments of virus genes were printed on 192 spots of a microarray chip and used as probes for hybridization. The study found four virus genes, namely p10, p35, lef-3 and lef-6, were involved in differential responses of host cells to AcMNPV infection, which would lead to different productivities by two insect cell lines Sf9 and High-Five. Also, by applying DNA microarrays with cDNA probes from PCR, Jiang et al. [21] characterized sequential expression patterns of all 155 AcMNPV genes and found 12 virus genes that depended on the virus pe38 gene for expression. Recently, the use of NGS for analysis of baculovirus gene expression patterns produced unprecedented genome-scale analysis of baculovirus infection processes, including mapping of all transcription start sites and polyadenylation sites, splicing variants, and putative interactions of virus proteins expressed at certain infection phases with host proteins [8,11]. A recent study using NGS for differential gene expression analysis and bioinformatic prediction of protein-protein interaction network suggested 22 viral proteins that could interact with 2,326 host proteins, totaling almost 8,907 interactions [11]. Another RNA-seq study comprehensively characterized 218 transcription start sites and 120 polyadenylation sites of all AcMNPV virus genes [8]. The authors also detected the possible encapsidation of viral mRNAs for late genes, suggesting a novel mechanism that the virus potentially uses to establish early infection, which is currently believed to depend entirely on the host RNA polymerase. Their study also suggested that 12 virus genes carry splicing variants, rather than only one gene as previously thought. The study also found an unexpected scale of antisense transcription for 50 virus genes. Such genome-scale transcriptomic studies offer great capacity to answer many unanswered questions on baculovirus gene regulation. Table 1 outlines the current knowledge on functions of a number of baculovirus genes that interact with host genes, which have been identified, mainly by molecular experiments, to interact with insect hosts.
For transcriptomic studies in an insect host, genome-scale sequences of coding genes are valuable resources for studying insect responses to baculovirus infections. With international efforts, only two Lepidopteran genomes are available to date, namely the B. mori genome [22] and the recently sequenced Diamondback moth (Plutella xylostella) genome [23]. To overcome challenges due to the lack of genomic sequences, several research groups have applied the traditional Sanger sequencing method to different collections of expressed sequence tag (EST) libraries such as those from the fall armyworm S. frugiperda [24,25], or from H. virescens [26], but these efforts could generate sequences for only a portion of the genome, usually fewer than 5,000 coding genes. Recently‚ more powerful RNA-sequencing technology has revolutionized the exome sequencing of a variety of non-model organisms, including insect species such as Manduca sexta [27], H. virescens [28], Galleria mellonella [29], L. dispar [30], and H. zea [18]. These sequence resources have been contributing to the expression analysis of baculovirus-insect interactions.
Expression analysis using microarrays have been applied to different baculovirus infection setups to compare different virus-host interactions, including: permissive vs. non-permissive hosts [31], infection of fat body vs. haemocyte [31], infection of high-yield vs. low-yield cell lines [20], uninfected vs. infected cells [15,18,32,33] or to compare infections at different time-points [7,10]. RNA-sequencing experiments are also useful for the analysis of RNA differential expression of all insect genes and of non-coding RNA [8,11,19]. Reviewing findings by these studies may show common genes and pathways being regulated by baculovirus infections ( Table 2).
The following subsections review in more detail the virus-insect interactions presented in Table 1, Table 2, and summarized in Figure 1 according to major pathways and cellular processes.

Metabolism
In relation to energy, no gene for an enzyme involved in energy metabolism has been found in any sequenced baculoviruses. However, a number of processes during virus infection require a high energy supply. For example, DNA packaging into nucleocapsids requires Adenosine triphosphate (ATP) for motor protein activities [95], elongation during translation uses Guanosine triphosphate (GTP) energy for binding of tRNA onto ribosomes, and active substrate transport, such as that provided by ABC transporters, the uptake of amino acids into cells, the export of viral RNA and the import of viral proteins cross the nuclear membrane, plus the directional transport of nucleocapsids all consume extensive energy [44]. Therefore, baculoviruses rely on host enzymes for energy metabolism. Transcriptomic studies show that ABC transporters and Citrate synthetase increased in Sf9 cells infected by AcMNPV [32,84]. These studies also identified an increase in the level of Glyceryldehyde 3-phosphate dehydrogenase (GAPDH) in HzAM1 cells upon infection by HearNPV [7], and in other hosts infected by DNA viruses‚ including Vaccinia virus infected human monocytes [96] and Rock bream Iridovirus infected marine teleost cells [97]. Similarly, up-regulation of Aldehyde dehydrogenase was found in both S. exigua cells infected by AcMNPV [19] and AcMNPV infected Sf9 cells [12]. Importantly, how the virus manipulates host energy generation and metabolic pathways to foster virus replication is still largely unknown.
For synthesis of virus building blocks such as nucleotides and amino acids, there are known virus genes for nucleotide metabolism, encoding a Ribonucleotide reductase (RR) large subunit and a small subunit, and a dUTPase protein, which prevents incorporation of mutagenic dUTP into DNA, suggesting that nucleotide metabolism may be especially important for baculovirus infections [57,58]. Via transcriptomics, Breitenbach et al. [33] found an increase in fatty acid metabolism in infected insects, including genes encoding lipid reductases and lipid desaturases in H. virescens larvae infected by HzSNPV. However, an extensive range of responses by metabolic genes have not been identified in baculovirus infected cells, especially those related to nucleotide and amino acid metabolism, which are important for production of new viruses in infected cells.

Translation
For translation, baculoviruses are entirely dependent on the host translational machinery including ribosomes, tRNAs, amino acid metabolism and transport, chaperones for protein folding, endoplasmic reticulum (ER) for glycosylation/phosphorylation/transport, and other translation factors. Several mechanisms allowing baculovirus mRNAs to compete with insect host mRNAs for translation have been found. These include the presence of AT-rich regions and unstructured 5' untranslated (UTR) regions [8,98], which confer a higher binding affinity for host translation factors by virus mRNAs compared to that of host mRNAs; the relative independence of virus late mRNAs on the 5'-cap-binding eukaryotic initiation factor (eIF4E) for the ribosomal recruitment process, in contrast to most insect mRNAs, which are dependent on cap-binding proteins for being translated [98,99]; and the ability of several virus genes, namely protein kinase 2 (pk-2) and host range factor 1 (hrf-1), to inhibit translational shutoff [64,66]. Emerging evidence also suggests that baculoviruses up-regulate translation initiation factors to accelerate the translation process [7,11].
Up-regulation of host genes for protein folding, translation initiation, and translational arrest have been reported. Like many other viruses, baculoviruses utilize host chaperones to facilitate the rapid synthesis of a large amount of viral proteins [87]. Heat shock proteins (HSPs) and heat shock protein cognates (HSCs) were commonly up-regulated in AcMNPV-S. frugiperda, HearNPV-H. zea, HzSNPV-H. virescens, B. mori-BmNPV, and S. exigua-AcMNPV infections [10,18,19,32]. Similarly, the up-regulation of other translation factors was commonly found, including: eukaryotic initiation factors (eIF3-6, eIF1A, eIF3-2b) and an elongation factor (EF1d) in B. mori-BmNPV infections [11]; eIF in S. exigua-AcMNPV infections [19]; and IF-2 and IF-3 in H. zea-HearNPV infections [7]. In addition, baculovirus infections trigger phosphorylation of the Eukaryotic initiation factor (eIF2α), which potentially causes translation arrest, but the host protein synthesis machinery is still functional at the late infection stage, as shown by detection of polysomal translation activities [16]. Protein degradation is one of the major responses of the host cell to a baculovirus infection. Proteasome 26S subunit and the ubiquitin-proteosomal degradation pathway in general were up-regulated in silkworm cells infected by BmNPV [11,32]. There is also evidence for baculovirus infections causing Endoplasmic reticulum (ER) stress [33]. More research is needed to better understand the roles of the ER, ribosomes, and translation factors needed for facilitating viral protein production.

Transport
For transport processes in infected cells, nucleocapsid movement is relatively well understood, while there is less knowledge on transport of viral proteins for virus DNA packaging, nucleocapsid assembly, and occlusion processes. Baculoviruses have devised effective transport mechanisms that are dependent on host Actin and cause a cytoskeleton remodeling phenomenon, in which both microfilaments and microtubules are used [44][45][46]49,100]. Nucleocapsid transport was detected within the cytoplasm as early as 5-30 min post infection, and inside the nucleus within one hour post infection [45]. A number of virus genes have been found to be involved in the nucleocapsid transport process, including: VP80, P78/83, EXONO, p10, VP39, ODV-c42, Ac66 and Arif-1 (Actin rearrangement inducing factor) [46]. For assembly, traditional protein techniques using yeast-two-hybrid screening for studying ODV protein-protein interactions have identified five triple-interactions, eight double-and nine self-interactions among virus ODV proteins [101]. Interactions of virus proteins with nuclear actin filaments and microtubule systems are essential for nucleocapsid assembly [44,45]. For transport of ODV proteins from the ER to the inner nuclear membrane, the host Importin-α-16, the viral FP25K (interacts with all known ODV proteins) and the ODV-E26 protein (interacts with some ODV proteins) play important roles [51]. The nuclear importins and exportins are also important for exporting viral mRNA to the cytoplasm, but their specific roles for transporting viral mRNA have not been investigated. BV envelope proteins are N-glycosylated and are transported from the ER to the cellular membrane, where nucleocapsids bud out. DNA packaging involves a packaging motor complex, possibly contains a virus Ac66 protein which interacts with host myosin and ATPase, and likely involves host actin and microtubules [6]. The occlusion process uses micro vesicles derived from the inner nuclear membrane [102]. Besides temporal differences in relation to BV versus ODV formation, it is unknown how virus genomes are directed to BV assembly rather than to the ODV occlusion process. The finding that a number of host proteins are incorporated into BV structures indicates that a number of unknown host proteins may be important for virus assembly, or for BV functions when entering new cells [11]. Thus, although the movement and remodeling of the host cytoskeleton system has been well studied, the roles of host genes in viral mRNA and viral protein transport and in virion assembly are still poorly understood.

Replication
For virus replication, baculoviruses have evolved a near complete replication machinery, which concentrates in a separate compartment within the nucleus, called the virogenic stroma [103]. Known replication enzymes that baculoviruses possess include: DNA polymerase (capable of both leading and lagging strand synthesis as well as proof reading by the 3'-5' exonuclease activities), helicase (p143), late expression factor LEF-1 (late expression factor, primase), LEF-2 (primase accessory factor), LEF-3 (single stranded DNA binding protein), p143 (helicase), origin binding protein IE1 (immediate early), and origin of replication with homologous regions (hrs) and non-homologous regions (non-hrs) (reviewed by Rohrmann [6] and Vanarsdall et al. [59]). Compared to virus transcription, replication may require more host factors, such as host DNA topoimerases (to untwist the double stranded DNA, thereby facilitating helicase activities) and DNA ligases (to link Okazaki fragments at the lagging strand). However, the host DNA polymerase is unable to substitute the viral DNA polymerase to complete virus DNA replication [103]. On the other hand, a baculovirus DNA polymerase can be interchanged with that from another baculovirus or even with a distantly related ascovirus DNA polymerase [104], suggesting that there is flexibility for baculovirus DNA polymerase activities, and hence it is possible that host DNA replication factors can facilitate viral replication. It remains unknown which host replication factors are used for virus replication, an efficient process that eventually leads to over 20 times more virus DNA than total host DNA in an infected cell [7].
For host DNA replication, cellular DNA replication is arrested, but viruses can utilize the host replication machinery to facilitate viral replication. Under baculovirus infection conditions, while cell cycle progression is abandoned at the G1/S or G2/M checkpoint (by 4 h.p.i, or 10 h.p.i in AcMNPV and 24 h.p.i for HearNPV systems), virus DNA replication increased rapidly [105]. Baculovirus replication, by a rolling cycle and recombination process, causes a cellular DNA damage response (DDR) following phosphorylation of several Phosphoinositide 3-kinases (Pi3K), which phosphorylate targeted substrates in the nucleus, for example the Histone 2A variant, and eventually causes cell cycle arrest or apoptosis [9]. The host DDR is a conserved response found by many host cells following infection by DNA viruses [106]. The up-regulation of Histone 2A variants was also found in B. mori cells infected by BmNPV [11]. Interestingly, even though the DDR is considered as a defence response by the cells, baculoviruses require this process for promoting virus replication and inhibition of DDR reduced virus production by 100,000 times [9]. Cellular replication factors, such as DNA polymerase subunits, could also be utilized by the virus for viral replication [7]. Knowing how the virus makes full use of the cellular replication machinery is important for maximizing virus production.

Transcription
For virus late transcription, baculoviruses possess a complete and effective transcription system to express viral genes under stringent temporal patterns. The viruses have evolved a unique RNA polymerase complex that can transcribe genes and process mRNAs at both ends by capping and polyadenylation. The complex contains 4 subunits‚ which are: LEF-4, an RNA capping enzyme; LEF-8, similar to the large subunit β in Bacteria; LEF-9, similar to the large subunit β' in Eukaryotes; and p147 similar to the alpha subunit in Bacteria. Virally encoded transcription enhancers include: a very late factor VLF-1‚ which binds to a burst sequence to make very late promoters accessible to RNA polymerase; a LEF-5 gene as a transcription initiation factor; and different DNA sequence motives within early and late promoters. Virus genes for mRNA processing include: the LEF-4 and an RNA cap 2'O-methyltransferase that are responsible for capping by RNA triphosphatase and Guanine transferase activities; an ADP-ribose pyrophosphatase for decapping; and a gene for polyadenylation‚ which is the viral RNA polymerase II with a distinct mechanism different to the host's polyadenylation process [107]. The viruses also encode a putative mRNA exporter (LEF-6). Nevertheless, the host factors are essential for immediate early transcription, when the virus transcription machinery has not been established. In addition, it is likely that the viruses also utilize host factors to facilitate its late transcription processes, such as a polyhedrin promoter binding protein (PPBP), which binds to promoters of both p10 and polyhedrin genes to enhance their transcription [89]. Xue et al. [11] predicted that all five virus proteins, including protein kinases (PK-1 and PK-2), CG-30, IE-2 and PE38, which display the most extensive interactions with over 1,000 insect genes, are related to transcription regulation, indicating that a high number of host proteins may contribute to virus transcription. Up-regulation of host transcription factors to facilitate the viral infection process deserves to be explored further.

Immune responses
For apoptosis and cell cycle arrest‚ it is well known that baculoviruses have evolved anti-apoptosis genes (virus inhibitor of appoptosis IAPs, p25, and p49) to inhibit cell death responses and they carry a cell cycle arrest gene, a Cyclin homolog, to block G2/M progression and create a "pseudo S" phase conducive for virus replication. On the other hand, there is little research on virus infection and host signaling pathways to identify cellular modulators and mediators involved in apoptosis and translation arrest [108]. There is evidence that the virus DNA replication process via a rolling cycle and recombination mechanism trigger a host DNA damage response, via the P3i-Akt pathway, which leads to an apoptotic response. The virus anti-apoptosis counter-defense is by direct binding of virus anti-apoptotic proteins to host caspases or by triggering proteasomal degradation of caspases. It has been found for AcMPNV and BmNPV that one of the mechanisms for the viruses to trigger cell cycle arrest at G2/M phase is by a virus Cyclin, a homolog of the host Cylin-B, which stably binds to the host cell division cycle (cdc-2) protein, leading to prolonged cdc-2 activities and preventing cell cycle progression to anaphase [41,109].
Over-expression of antimicrobial peptides and apoptosis are common host defense responses found in insect-baculovirus systems. Gloverin, encoding an antiviral peptide‚ increased in BmNPV infected B. mori and in S. exigua larvae infected by AcMPNV [19,31]. Another antiviral peptide, Cecropin, is also commonly found up-regulated, such as in H. virescens infected by Helicoverpa zea single nucleopolyhedrovirus (HzSNPV) [33] and in the H. zea-HearNPV system [7]. Similarly, induction of apoptosis by baculovirus infections has been found in many expression studies using baculovirus-insect systems [10,11,18].
Recently‚ microRNA (miRNA) and small interfering RNA (siRNA) responses have been identified as important control mechanisms for insect-virus interactions. Rapid discovery of miRNAs is aided by the development of NGS technologies and computational analysis followed by validation of selected putative miRNAs. For baculoviruses, only four viral miRNAs have been detected by using NGS to sequence small RNAs extracted from BmNPV-infected B. mori cells [77]. These four miRNAs are evolutionarily conserved among many baculoviruses. On the other hand, many insect miRNAs, from 90-62 miRNAs, have been discovered. Using Illumina NGS, over 90 host insect miRNAs were found in S. frugiperda [13]. By sequencing 6,720 clones of small RNAs in B. mori, Yu et al. [91] found 35 miRNAs, while the computational prediction identified 262 genes for miRNAs. Comparing outputs by the two approaches, NGS generated more miRNAs than traditional sequencing. For siRNA, NGS also discovered siRNA responses in the H. armigera insect [92]. The insect cells encode a Dicer 2, which cuts double stranded viral dsRNA to generate viral small interfering siRNAs that target viral mRNA products during late infection periods [92]. In addition, some virus miRNAs may regulate host miRNA populations by inhibiting nuclear export of host miRNAs [76].
In relation to signaling pathways‚ the BmNPV virus induced the B. mori mitogen protein kinase (MAPK) pathways, including extracellular signal-regulated kinase (ERK), and c-Jun NH 2 -terminal kinase activity (JNK), possibly via the virus PK-1 protein [83]. AcMNPV infection increased Akt phosphorylation from 2-6 h.p.i and activated the Pi3-Akt pathway in infected cells [81]. The MAPK and Akt pathways are pro-survival pathways, often being activated before cells commit to apoptosis. Sagisaka et al. [32] found up-regulation of a Toll protein in B. mori cells infected by BmNPV (BmToll10-3, a surface protein receptor), but a change in the Toll pathway was not identified. Similarly, there have been no findings of changes to the Janus kinase (JAK)-signal transducer and activator of transcription (STAT) (JAK-STAT) pathway or the immune deficiency (IMD) pathway in baculovirus infected insect cells to date. Therefore how baculoviruses control, and insect cells respond, via signaling pathways‚ especially those involved in innate immune responses‚ remain to be explored.

Manipulation of Host Cells or Larvae to Study Baculovirus-Insect Interactions
Common methods applied for studying pathways in insects use chemical inhibitors to moderate a pathway, for example: using 5-fluoro-29deoxyuridine (FdUrd) for cell cycle arrest at G1/S [105], using U0126 and PD98059 for moderation of Extracellular signal regulated kinases (ERK) [69], using SP600125 for moderation of the JNK pathway [69], and using PS-341 for proteasome modifications [11]. Similarly; to trigger a pathway, Actinocyn D is used for inducing apoptosis [110]. Molecular biology methods for testing roles of a pathway or a gene to baculovirus infection in insects most commonly use gene knockdown by RNA-interfering (RNAi) technology, for example using RNAi to repress the MAPK pathway (JNK and ERK) [69] or apoptotic caspases [111]. Until recently, RNAi silencing technology has been used successfully for studies involving 15 Lepidopteran species [112]. A technique for targeted gene knockout, using Zinc Finger Nuclease technology, was also successfully applied for studies with B. mori [113]. For over-expressing a gene or introducing a heterologous gene into insect cells, transient transfections have been widely applied with a range of transfection reagents, including inexpensive and effective reagents such as polyethylenimine (PEI) [114]. Tools are also available for stable transfection, most commonly using antibiotics for selecting cells containing heterologous DNA randomly inserted into insect cell chromosomes [115]. More details about methods for genetic engineering of the baculovirus-insect system are reviewed elsewhere [116,117].

Conclusions and Future Perspectives
Overall, while baculoviruses have evolved to control insect hosts‚ there is increasing evidence of host cell responses against baculovirus infections. Figure 1 summarizes many of the bidirectional interactions between cell and baculovirus genes/proteins that have been discussed in this review. However, host responses are relatively poorly understood compared to current knowledge on virus genes. Metabolic pathways including energy and nucleotide/amino acid/lipid metabolism are expected to be up-regulated‚ but little is known about which genes, if any, are affected. Likewise‚ how the viruses control various pathways such as signaling pathways are still largely unanswered questions. Similarly‚ although heat shock responses are now known to be a conserved mechanism beneficial to baculoviruses and have been found in all studied baculovirus-insect cell systems, the trigger of the response by these host genes to baculovirus infections remains unknown. Virus genes/processes are in red, and insect genes/processes are in green. Major pathways are indicated by ellipses. The left half of the diagram represents pathways that are relatively well understood. The right half shows pathways that deserve more research, especially those indicated by filled ellipses. The ellipse with dashed boundary represents the nuclei. The nucl./aa./lp. abbreviation stands for nucleotide/amino acid/lipid metabolism processes.
"Omics" tools can greatly improve the knowledge landscape of baculovirus-insect interactions by investigating cellular responses at a genome scale. It can be seen from Figure 1 that the right half of the diagram represents cellular pathways rather than virus-related processes, indicating a lack of research on cellular processes affected by virus infections. This lack of understanding of cellular responses is mainly due to the unavailability of genomic sequences as well as the longer time required to engineer an insect genome than the time required to edit a virus genome. To date, RNA-seq technology is revolutionizing research in organisms without genomic sequences by generating transcript sequences and allowing the analysis of expression levels of these transcripts [118,119]. Transcriptomic approaches enable more comprehensive studies, which are at a whole genome scale (involving over 15,000 genes), generally far more genes compared to the number of proteins or metabolites being measured by proteomic or metabolomic approaches.
A number of studies comparing microarray and RNA-sequencing (RNA-seq) have shown that the two tools produce highly correlated results (r > 0.91) and they complement each other [119,120]. RNA-seq has advantages over microarrays in the discovery of novel transcripts, analysis of isoforms, and detection of antisense, splice junctions, siRNA, and miRNA. However, for non-model organisms that do not have reference genomes, accurately assembling transcripts and mapping of reads onto transcripts are fundamentally important for determining differential expression results using RNA-seq [119]. In addition, it is speculated that RNA-seq might produce bias for highly expressed virus genes at late infection stages when the polyhedrin and p10 genes are hyper-expressed, and detection of low abundant genes could be compromised [11]. Combining results from microarray and RNA-seq analysis can increase detection power and accuracy [120].
Compared to transcriptomics, proteomics and metabolomics are in an early stage of development, but rapid progress is being made with these tools. Proteomics with mass spectroscopy based approaches and activity-based functional validation procedures can provide insights into host-virus interactions [121]. A proteomic study of infections in permissive, semi-permissive and non-permissive cell lines, H. virescens and H. zea, by AcMNPV and HzSNPV identified 21 proteins, among which the up-regulation of two Calreticulin proteins, two DNA supercoiling factors, and two heat shock proteins by 5-fold were found [122]. Courtiade et al. [123] carried out a proteomic study using H. armigera cells following induction of apoptosis and identified changes in 13 proteins compared to control cells. Chen et al. [124] found nine proteins being changed in susceptible and resistant B. mori midgut tissues following virus infections. The current metabolomics technology only allows measurements of a part of the metabolome [125]. Among a few metabolic studies in insects so far, measurements of amino acids, sugars and several redox agents have been undertaken. Tran et al. [126] measured amino acids and adenosine phosphates Bernal et al. [127] measured glucose, glutamine, glutamate, lactate, ammonia, maltose, sucrose, glucose, NAD + /NADH and performed assays for 10 enzyme activities. At this stage, the rapid development of quantitative proteomics allows identification and measurements of thousands of proteins [128]. Targeted metabolomics is also powerful in unraveling changes in host metabolism [129], an area that remains largely unexplored for baculovirus-insect systems.
Engineering insect cells for functional analysis of target genes in insect-virus interactions can use high throughput screening technologies, involving the use of RNAi systems [130]. Although no high throughput gene knocking technologies have been established to date in insect-baculovirus systems, temporary expression by transient transfection of expression vectors or creating stably transfected polyclonal cells can be good options to validate genes that require up-regulation or over-expression to improve production outcomes using insect cell technology. These technologies together with the power of "omics" tools, especially the two well established transcriptomic tools, can unravel many host pathways that are yet to be fully explored.