MIAMI UNIVERSITY the Graduate School CERTIFICATE FOR

Home , Ascoviridae, Deltabaculovirus

MIAMI UNIVERSITY The Graduate School

CERTIFICATE FOR APPROVING THE DISSERTATION

We hereby approve the Dissertation

Jianli Xue

Candidate for the Degree: Doctor of Philosophy

______Xiao-Wen Cheng, Ph.D., Director

______Mitchell F. Balish, Ph.D., Reader

______Eileen K. Bridge, Ph.D., Reader

______Gary R. Janssen, Ph.D., Reader

______Michael Novak, Ph.D., Graduate School Representative

ABSTRACT

COMPARISON OF ASCOVIRUS AND BACULOVIRUS GENOMES AND THEIR REPLICATION AND GENE EXPRESSION STRATEGIES

by Jianli Xue

A new member of the newly discovered insect virus family Ascoviridae, Spodoptera frugiperda ascovirus-1d (SfAV-1d) was identified from South Carolina, USA. The genome size of SfAV-1d is estimated to be about 100 kb which makes SfAV-1d the smallest ascovirus genome so far. SfAV-1d is closely related to the previously reported SfAV-1a with 99% DNA sequence identity to SfAV-1a. A deletion of 14 kb was found in the SfAV-1d genome that corresponds to the inverted repeat region in SfAV-1a. Cloning and sequencing revealed that the deleted region is highly variable in the SfAV-1d genome with different lengths deleted in individual isolates. SfAV-1d has a narrower host-range than SfAV-1a and it can only develop full infections in S. frugiperda but not in S. exigua in which SfAV-1a is highly infectious. In order to understand better ascoviruses, ascoviruses gene transcription and expression strategies were studied and compared with baculoviruses. The DNA polymerase gene of SfAV-1d was demonstrated to be an early gene while the major capsid protein gene was confirmed to be a late gene. Three RNA polymerase homologues were found in the SfAV-1d genome. In vitro transcription assays showed that an early gene of SfAV-1d was transcribed by the nuclear extract from Sf21 insect cells while a late gene of SfAV-1d was not, suggesting that host RNA polymerase transcribes early genes of SfAV-1d while late gene transcription needs viral factors. Therefore, SfAV-1d follows the same transcription patterns as baculoviruses: early genes are transcribed by the host RNA polymerase while late genes transcription needs viral factors. We further characterized specifically a baculovirus late gene gp37 expression to understand the gene expression strategies. The 3’ untranslated region (UTR) of gp37 gene from the well-studied baculovirus Autographa californica Multicapsid NPV (AcMNPV) was studied. It has multiple polyadenylation sites and can reduce polyhedron production at the polyhedrin locus without changing the total amount of protein expressed.

COMPARISON OF ASCOVIRUS AND BACULOVIRUS GENOMES AND THEIR REPLICATION AND GENE EXPRESSION STRATEGIES

A Dissertation

Submitted to the Faculty of

Miami University

in partial fulfillment of the requirement for the degree of

Doctor of Philosophy

Department of Microbiology

Jianli Xue

Miami University

Oxford, Ohio

2011

Dissertation Director: Xiao-Wen Cheng, Ph.D.

Jianli Xue

2011

Table of Contents

General Introduction………………………………………………………….………...1

Chapter One…………………………………………………….……………………….7

Abstract………………………………………………………………………….…....8

Introduction……………………………………………………….………….…….....9

Materials and Methods………………………………………………….…………...12

Insects, insect cells and viruses…………………………….. ………….………..12

Viral DNA purification……………………………….…….…………………...12

Restriction fragment length polymorphism (RFLP) and DNA hybridization analysis……………………………………………………………………..….…12

Pulse-field gel electrophoresis…………………………………….……………..13

Generation of SfAV-1d libraries for genome sequencing………...……………..13

PCR analysis, DNA cloning and sequencing……………………………….……14

Cell cultures and virus infection…...…………………………………………….14

Bioassay …………………………………………………………………………15

Results……………………………………………………………………………….16

Restriction fragment length polymorphism and Southern hybridization analysis of SfAV-1d and SfAV-1a…………………………………………….……………..16

Genomic DNA sequence deletions of SfAV-1d…………………………...... 19

Cloning and sequencing of the deletion region of SfAV-1d……….…………….19

Comparison of cell infection between SfAV-1d and SfAV-1a………………….26

iii

SfAV-1d has a narrower host-range than SfAV-1a………………….………….30

Discussion...……………………………………………………………….………....32

Acknowledgements……………………………………………………….……….....35

Chapter Two………………………………………………………………….………....36

Abstract…………………………………………………………………….………...37

Introduction………………………………….…………………………….…………38

Materials and Methods……………………….………………………….…...……...40

Results…………………………………………..……………………….….………..43

28S rRNA shows lower Ct in qRT-PCR…….…………………………...….…..43

28S rRNA has the least variation during viral infection…..…………………….43

Inclusion of the 28S rRNA-R primer into oligo-dT-primed cDNA synthesis decreases Ct and variation of 28S rRNA…………………..………….…………50

Inclusion of the 28S rRNA-R primer into oligo-dT-primed cDNA synthesis does not interfere with the detection of other gene transcription.……..…….………...50

Discussion……………………………………………………...………….…..……..57

Acknowledgements……………………………………………………….…..……...59

Chapter Three…………………………………………………………………..……....60

Introduction………………………………….………………………………...….….61

Materials and Methods……………………….………………………………………64

DNA sequence analysis……………………………………………..………..….64

Phylogenetic analysis……………….…………………………….……...... ….64

Plasmids, cloning and sequencing…………………………………….…..….….64

Viral replication inhibition analysis……….…………………………………...... 64

In vitro transcription assay……………………………………………………….65

RT-PCR analysis………………………………………………………………....66

Results…………………………………………….………………………………….68

Partial genome annotation…………………….…………………....….…………68

Phylogenetic analysis of ascovirus………………………………………………68

Identification of early and late genes…………………………………………….68

Dependence of viral genes transcription on the host RNA polymerase…………72

Discussion……………………………………………………………………………77

Chapter Four……………………………………………………………………………80

Introduction…………………………………………………………………………..81

Materials and Methods……………………………………………………………….84

Cell lines and viruses…………………………….………………………………84

Comparison of 3’ downstream sequences of AcMNPV gp37, polyhedrin gene and CfDEFNPV spindlin gene……..…………………………………….………84

Construction of viruses…………………….………………..…………………...84

Quantitative analysis of polyhedron production and size measurement of polyhedra……………………………………………….……….………………..86

Protein yield assay……………………………………………………………….86

Results………………………………………………………………………………..88

3’ UTR sequence analysis…………………………………….……….…….…..88

Construction of two sets of viruses………………………….……………….….88

Polyhedron production and their sizes comparison between AcpolUTR and Acgp37UTR...……………………………….…….…....………………………..88

Comparison of total polyhedrin protein production…...…………….……..…....97

Discussion …………………………………………………………………………..102

General Conclusions…….…………………………………………………………….105

References…………………………………………………….…………..…………...110

List of Tables

Table 1-1. Summary of the variants of the SfAV-1d deletion region……………..…….27

Table 1-2. Susceptibility of noctuid larvae to SfAV-1a and SfAV-1d (% mortality)…...31

Table 2-1. A list of primers used in this study…………………………………………..42

Table 2-2. Comparison of average Ct of the possible reference genes…………………..46

Table 2-3. Comparison of SD [±Ct] by Bestkeeper analysis……………………….…...49

Table 2-4. Comparison of ΔΔCt of the possible reference genes………………….…….51

Table 3-1. Oligonucleotide primer sequences used in real-time PCR and in vitro transcription…………………………………………………………………………...... 67

Table 3-2. Partial genome annotation of the SfAV-1d genome………….……….……..69

vii

List of Figures

Figure 1-1. RFLP and Southern hybridization analysis of the new ascovirus isolate

SfAV-1d………………………………………………………………………………...17

Figure 1-2. Assembly of the SfAV-1d genome from the shotgun library sequences with

SfAV-1a as the reference using Sequencher……………….……………………...……20

Figure 1-3. Confirmation of the major 14kb deletion of SfAV-1d comparing to

SfAV-1a….……………………………………………………………….…………….22

Figure 1-4. Characterization of the variable sequence in the deletion region of

SfAV-1d….……………………………………….…………………………………….24

Figure 1-5. Infection of the insect cell line, IOZCAS-Spex-II by SfAV-1a or SfA-1d..28

Figure 2-1. Melting curve analysis of housekeeping gene amplification using gene specific primers by qRT-PCR…………………………………………………………..44

Figure 2-2. Comparison of cycle threshold (Ct) fluctuation of housekeeping genes in

Sf21 cells infected with viruses at different time points post infection………………...47

Figure 2-3. Comparison of cycle threshold (Ct) in qRT-PCR using mixture of oligo-dT/28S-R and oligo-dT in cDNA synthesis for 28S gene amplification…………52

Figure 2-4. Comparison of Cts of PPI gene amplification in qRT-PCR between oligo-dT/28S-R and oligo-dT synthesized cDNA for interference detection..…………55

Figure 3-1. Evolutional relationship between baculovirus and ascovirus………...……70

viii

Figure 3-2. Determination of early or late gene in the SfAV-1d genome……………..73

Figure 3-3. In vitro transcription analysis of different baculovirus and ascovirus promoters with insect cells nuclear extract……………………………………………..75

Figure 4-1. 3’ downstream sequence analysis of AcMNPV gp37 (A), polyhedrin (B) genes and the spindling gene from CfDEFNPV (C)……………...…………………….89

Figure 4-2. A schematic presentation of four constructed viruses……...………...…....92

Figure 4-3. The insect cell lines were infected by AcpolUTR or Acgp37UTR…...…...94

Figure 4-4. Bio-Rad protein assay of polyhedrin amount in Sf21 cells infected by

AcpolUTR or Acgp37UTR……………………………………………………………..98

Figure 4-5. GFP production comparison between AcgfppolUTR infected Sf21 cells and

Acgfpgp37UTR infected Sf21 cells at 4 d p.i…………………………………………100

General Introduction

Insect viruses are important in controlling insect populations in agriculture and forestry. Baculoviruses are the most studied insect viruses, whereas ascoviruses belonging to the Ascoviridae family are recently discovered and characterized (Federici et al., 2005). They were identified in the late 1970s in the Southeast region of the United States. Ascoviruses infect lepidopteran insects, which include moths and butterflies at the larval and pupal stages, causing a chronic but fatal disease (Federici et al., 2000). Among those insects infected by ascoviruses are many economically important insect species that impact agricultural production, such as Helicoverpa zea, Heliothis virescens, Spodoptera frugiperda and Spodoptera exigua (Carner and Hudson, 1983; Hamm et al., 1985; Hamm et al., 1986; Cheng et al., 2005). These insect pests damage a wide range of crops, such as vegetables, fruits, and corn, and cause significant economic loss every year.

Four species of ascovirus are recognized by the International Committee on Taxonomy of Viruses (ICTV): Spodoptera frugiperda ascovirus 1 (SfAV-1), Trichoplusia ni ascovirus 2 (TnAV-2), Heliothis virescens ascovirus 3 (HvAV-3), and Diadromus pulchellus ascovirus 4 (DpAV-4) (Federici et al., 2005). Ascoviruses have large enveloped virions, which are typically bacilliform (rod shaped) or allantoid (egg shaped) (Federici, 1983; Bideshi et al., 2005). Ascovirus infection in cells triggers an apoptotic pathway which produces vesicles containing virions (Federici, 1983). The ascovirus encodes its own viral caspase which by itself can induce apoptosis (Bideshi et al., 2005). The developing apoptotic vesicles in the infected cells are used to incorporate virions. These vesicles are released into the hemolymph of the infected insects which makes the hemolymph milky white, a unique symptom of ascovirus infection of insects (Federici, 1983).

In general, per os (by feeding) infectivity of insects by ascovirus is poor. It was also reported that per os infection of insects by ascoviruses varies among different species and the infection rates can be from 0 to 59% (Govindarajan and Federici, 1990). Therefore, transmission of ascovirus is primarily through the ovipositor of the endoparasitic female wasps during their oviposition (Federici et al., 2005). These

endoparasitic wasps are the most common known vectors of ascovirus. When the wasps lay eggs in the ascovirus-infected larvae, thereby contaminating their ovipositor, and then lay eggs in the healthy larvae, they infect the healthy larvae with ascoviruses through the ovipositor (Hamm et al., 1985). Thus, the circulation of the vesicles packed with virions in the hemolymph may have evolved to help the transmission of ascoviruses in the wild insect population by wasps (Hamm et al., 1986; Hamm et al., 1998; Lopez et al., 2002; Tillman et al., 2004). The host-range of ascoviruses varies among different species. TnAV-2 and HvAV-3 have a broad host-range and they infect different noctuid species and some species belonging to other families of Lepidoptera, whereas SfAV-1 has a narrow host-range and can only infect and produce mortality to the species of Spodoptera (Federici et al., 2009).

Ascoviruses have a large circular double-stranded DNA (dsDNA) genome. Four ascovirus genomes were sequenced and are available to the public: SfAV-1a, TnAV-2c, HvAV-3e, and DpAV-4a (Cheng et al., 1999; Bideshi et al., 2006; Wang et al., 2006; Asgari, 2007; Bigot et al., 2009). The size range of ascoviruses varies from 119 to186 kb. So far ascoviruses have been found in the United States, Australia, France, and Indonesia (Federici et al., 1991; Bigot et al., 1997b; Cheng, et al., 2000; Asgari et al., 2007). Since ascoviruses infect insects in the largest family Noctuidae of the order Lepidoptera and the vectors of ascoviruses are found throughout the world, ascoviruses are probably distributed worldwide (Federici et al., 1991). All the sequenced ascovirus genomes encode RNA polymerase subunit homologues. DpAV-4a has five RNA polymerase subunit homologues while all the other ascoviruses have three RNA polymerase subunit homologues. This is in contrast to baculoviruses, which have four RNA polymerase subunits with specific functions in late gene transcription. Whether these ascovirus RNA polymerase subunit homologues have any function similar to that of baculovirus RNA polymerase in viral transcription is still unknown.

Baculoviruses in the family Baculoviridae were first identified as insect-specific viruses. It is a well-studied family of occluded viruses that have their virions embedded or occluded in the polyhedra. Baculoviruses are divided into four genera, alpha-, beta-,

gama- and deltabaculovirus. The alpha- and betabaculoviruses are classically named nucleopolyhedrovirus (NPV) and granulovirus (GV), respectively. The gamabaculoviruses infect dipteran insects (mosquito), whereas deltabaculoviruses infect hymenopteran insects (wasps) (Jehle et al., 2006). NPV is the most studied baculovirus due to its importance in biological control of insect pests of agriculture and forestry. Furthermore, the availability of the cell culture systems for baculovirus genetic studies results in a vast amount of information on baculovirus gene transcription mechanisms and DNA replication strategies (Miller, 1997). In the late phase of baculovirus infection of insect cells, NPVs produce large polyhedron-shaped structures called polyhedra that contain many virions. Polyhedra protect the virions from UV inactivation and desiccation in the natural environment and the virions in the polyhedra can survive outside of the cells for years (Miller, 1997).

Baculoviruses infect more than 500 species of invertebrates, including many insects but also shrimp (Miller, 1997). The genome size range of baculoviruses is 80-180 kb (van Oers and Vlak, 2007). There are multiple homologous repeats (hrs) at various locations throughout baculovirus genomes, and the repeats from different baculoviruses have high homology (Cochran and Faulkner, 1983). However, the number and distribution of hrs are variable across species (van Oers and Vlak, 2007). Analysis of defective genomes lacking major segments of the wild-type viral genomes but containing viral repeats, as well as transient replication assays from plasmids containing the hr sequences, suggest that hrs serve as the origins of replication (Pearson et al., 1992; Pearson et al., 1993; Pearson and Rohrmann, 1995; Xie et al., 1995). There are multiple origins in the baculovirus genome, and it has been suggested that the replication ability of the different hrs in different cell lines is variable, which may be partly responsible for the host-range difference among different baculoviruses (Pearson et al., 1992; Leisy and Rohrmann, 1993; Pearson et al., 1993; Lee and Krell, 1994; Ahrens et al., 1995; Leisy et al., 1995; Pearson and Rohrmann, 1995; Xie et al., 1995).

Transcription of baculovirus occurs in a cascade with three major phases: early, late, and very late. Viral proteins synthesized in the previous phase allow the expression

of genes in the subsequent phase. Early genes are transcribed by host and/or viral factors. Baculovirus early genes include immediate early gene 1 (IE1) (Guarino and Summers, 1986). As the viral replication progresses, host and early genes are shut off and late and very late genes are transcribed. Late genes include virion structural proteins such as P74 (Wang et al., 2009). Very late genes include the polyhedrin gene that is expressed in large amounts to form polyhedra to occlude the virions (Miller, 1997). All the early genes are transcribed by the host RNA polymerase whereas all the late genes are transcribed by viral RNA polymerase. The virus-encoded RNA polymerase includes four subunits: LEF- 4, LEF-8, LEF-9, and P47 (Guarino et al., 1998; Berretta and Passarelli, 2006).

One of the late genes that is believed to be transcribed by the viral RNA polymerase is gp37. The function of the gp37 gene is not clear but it is known that it is not required for replication (Cheng et al., 2001). Autographa californica Multicapsid NPV (AcMNPV) gp37 is a homologue of the Choristoneura fumiferana defective nucleopolyhedrovirus (CfDEFNPV) spindle gene, a highly expressed gene whose product forms spindle-like crystals in the cytoplasm of the infected cells (Doerfler, 1986; Li et al., 2000). The function of the spindle gene is still not understood but it was reported that it is a homologue of fusolin protein from entomopoxvirus which can enhance baculovirus infection (Xu and Hukuhara, 1992; Li et al., 2000). There are three different lengths of the gp37 transcripts and the percentage of each size of the transcript changes as the infection progresses (Wu and Miller, 1989). The ends of three kinds of transcripts match the classic polyadenylation signal (AAUAAA) sites downstream of the stop codon of gp37. All three different length transcripts appear starting from 6 h post infection. The amount of the longest transcript increases with infection time whereas the amount of the other two shorter transcripts decreases with time. The switch from shorter transcripts to the longest transcript after 6 h p.i. leads to different 3’ untranslated region (UTR) of gp37 transcripts (Wu and Miller, 1989). Alternative polyadenylation signals are found in 3’ UTRs, and usage of different polyadenylation signals can lead to different expression levels of the gene (Lutz, 2008). The gp37 gene has multiple polyadenylation signals in its UTR region. Whether the 3’ UTRs of gp37 has any regulatory effect on its expression and whether the usage of different polyadenylation signals leads to the regulation of gp37 4

are still unknown. Nevertheless, use of alternative polyadenylation is a common phenomenon for baculovirus mRNAs whose shortest forms are more common in the earlier stages and whose longest forms are more common in the later stages of transcription (Westwood et al., 1993). By using alternative polyadenylation, viruses can efficiently regulate gene expression at different stages.

Phylogenetic analysis based on biological characteristics and DNA polymerase protein sequence suggested that ascovirus is closely related to baculoviruses and may represent the transition in evolution between nucleus-replicating viruses and cytoplasmic viruses (Cheng et al., 2007). Baculovirus and ascovirus share many similarities in the genome structure and virion morphology. In addition, the ascoviruses, SfAV-1a and TnAv-2c share 20 and 16 homologous predicted proteins with baculovirus, respectively (Bideshi et al., 2006; Wang et al., 2006). The close relationship between ascovirus and baculovirus may help ascovirus research by using the well-studied baculovirus as a reference.

All the baculovirus NPVs replicate exclusively in the nuclei of infected cells. NPV replication in the nucleus is associated with the virogenic stroma (VS), which is a virus-induced electron-dense structure observed by electron microscopy (Miller, 1997). It was suggested that ascovirus replication occurs in the nucleus because similar VS-like structures were observed in the nucleus during ascovirus infection in insect cells (Federici, 1983; Cheng et al., 2000). As the infection continues, the nucleus of the infected cell is disrupted, resulting in virion formation in the cytoplasm (Bideshi et al., 2005). However, whether replication of ascovirus actually occurs inside the nucleus has not been experimentally tested. Since ascovirus and baculovirus are closely related viruses, it is likely that their DNA replication and gene transcription strategies are conserved. Since all the sequenced ascovirus genomes have homologous RNA polymerase genes, it is likely that they function on their own or with other viral or cellular factors to transcribe ascovirus genes. But whether ascovirus can access the host transcription machinery in the nucleus is still unknown and needs to be addressed.

Similar to baculovirus, there are large interspersed repeats of 1-3 kb in the ascovirus genomes (Bigot et al., 2000). Some of the repeats contain open reading frames (ORFs) and are identified as baculovirus repeated ORF (bro) homologs, with 23 found in HvAV-3e, 3 in TnAV-2c, 7 in SfAV-1a, and 12 in DpAV-4a (Bideshi et al., 2006; Wang et al., 2006; Asgari et al., 2007; Bigot et al., 2009). In HvAV-3e, one repeat region encodes a conserved transposase (Asgari, 2007). Other than the ORF-containing repeats, other non-coding hrs are also found in the ascovirus genomes. Two hrs are identified in the TnAV-2c genome and 5 hrs are identified in HvAV-3e, whereas 4.5 inverted repeats are identified in the SfAV-1a genome (Bideshi et al., 2006; Wang et al., 2006; Asgari et al., 2007; Bigot et al., 2009). However, the functions of these homologous sequences are not determined. The research described in this dissertation compares the role of the hrs in allowing ascovirus replication in different cell lines, the dependence of ascovirus gene transcription on host cell transcription enzymes, and baculovirus late gene 3’ end processing, in hopes of providing information for further genetic studies of insect viruses.

Chapter One

Submit to Journal of General Virology Corresponding author: Xiao-Wen Cheng Department of Microbiology 32 Pearson Hall Miami University Oxford, Ohio 45056 USA Phone: (513) 529-5429 FAX: (513) 529-2431 E-mail: [email protected]

Comparative analysis of a major inverted repeat variable region of the genomes and host range between Spodoptera frugiperda ascovirus 1d (SfAV-1d) and SfAV-1a

Jian-Li Xue and Xiao-Wen Cheng

Department of Microbiology, 32 Pearson Hall, Miami University, Oxford, Ohio 45056, USA

Abstract The recently discovered ascoviruses have a worldwide distribution. Here we report a new member of the family Ascoviridae, Spodoptera frugiperda ascovirus-1d (SfAV-1d) with a major variable region in the genome. Restriction fragment length polymorphism, Southern hybridization and genome sequencing analyses confirmed that SfAV-1d and the earlier reported SfAV-1a are closely related but not identical. The genome size of SfAV- 1d is about 100 kb which is about 57 kb smaller than SfAV-1a. The SfAV-1d genome lacks the 14 kb inverted repeat (IR) region of SfAV-1a. Cloning and sequencing this region of SfAV-1d revealed that this region is highly variable with different lengths of DNA sequence deletions. In all the variants, the whole 14kb IR is missing with 88.2% of the variants missing part or the whole adjacent SfAV-1d ORF71, 94.1% missing part or the whole adjacent ORF72 and 64.6% missing part of or the whole ORF73. Both SfAV- 1d and SfAV-1a infect Sf21 cells derived from S. frugiperda producing unique cytopathic effects like other ascoviruses. SfAV-1a can infect and kill almost all the IOZCAS-Spex-II cells derived from S. exigua while in SfAV-1d infections about 3% of the IOZCAS-Spex- II cells are resistant to killing and continued to grow. SfAV-1a is highly virulent to S. exigua, Tricoplusia ni, S. frugiperda and Pseudoplusia includens larvae but not to Anticarsia gemmatalis. SfAV-1d has a narrower host-range compared to SfAV-1a, only produced mortality in S. frugiperda. Collectively, this variable region of the SfAV-1d genome might be involved in host-range in cells and insects.

Introduction Ascoviruses are members of the family Ascoviridae. They are recently discovered insect specific viruses that cause a chronic and ultimately fatal disease in lepidopteran larvae, mostly noctuids (Federici et al., 2005). Their later discovery compared to other insect viruses such as baculovirus is likely due to the fact that no obvious symptom is seen when a larva is infected with an ascovirus, although the hemolymph of infected larvae becomes uniquely milky white due to the accumulation of high concentration of vesicles packed with virions (Federici and Govindarajan, 1990). Unlike the well-studied baculovirus that can cause per os infection to susceptible larvae, ascovirus is poorly infectious per os to susceptible larvae. Instead it is transmitted mechanically to the insect hosts by parasitoid wasps during oviposition in the natural environment (Hamm et al., 1985). Per os infection of baculovirus in the natural environment is largely enhanced by the protective polyhedra that occlude the enveloped virions (Vlak and Rohrmann, 1985). Since ascovirus does not encode a polyhedrin-like protein to protect the enveloped virions, transmission of ascovirus in insects by parasitoid wasps in natural environment is likely to be an evolutionary adaptation. The enveloped virion of ascovirus has a circular double-stranded DNA (dsDNA) genome with a size range from 119 kb to 186 kb (Federici and Govindarajan, 1990; Cheng et al., 1999; Bideshi et al., 2006; Wang et al., 2006; Asgari et al., 2007; Asgari et al., 2007; Bigot et al., 2009). A unique cellular pathology of ascovirus is its ability to produce apoptotic bodies called vesicles that contain the ascovirus virions. During infection, the nucleus of the infected cell enlarges followed by cleavage of host cells, and eventually the nuclear envelope ruptures (Federici et al., 1991). After nuclear lysis, the apoptotic bodies start to form, and instead of being degenerated, they are rescued by the virus and gradually form large virion-containing vesicles. It is reported that a virus- encoded executioner caspase plays a direct role in inducing apoptosis (Bideshi et al., 2005). The formation of vesicles during ascovirus infection in cells seems to be a general pathway in all the ascoviruses reported (Federici et al., 2009). Of all the ascoviruses reported, four ascovirus species have been accepted by the International Committee on Taxonomy of Viruses (ICTV): Spodoptera frugiperda

ascovirus (SfAV-1), Trichoplusia ni ascovirus (TnAV-2), Heliothis virescens ascovirus (HvAV-3), Diadromus pulchellus ascovirus (DpAV-4) (Federici et al., 2005). Four ascovirus genomes have been sequenced and reported, TnAV-2c, SfAV-1a, HvAV-3e, and DpAV-4a (Bideshi et al., 2006; Wang et al., 2006; Asgari et al., 2007; Bigot et al., 2009). All of these ascoviruses infect noctuids except DpAV-4a which infects the leak moth, Acrolepiopsis assectella (family Hyponomentoidae) (Bigot et al., 1997a). Host- ranges of ascoviruses are different with SfAV-1a having a narrower host range than TnAV-2c and HvAV-3e (Cheng et al., 2005; Bideshi et al., 2006; Bigot et al., 2009). The ascovirus genome is methylated and contains large interspersed repeats of 1-3 kb (Bigot et al., 2000). Some of these repeats are identified as baculovirus repeated ORF (bro) with 23 found in HvAV-3e, 3 in TnAV-2c, 7 in SfAV-1a, and 12 in DpAV-4a (Bideshi et al., 2006; Wang et al., 2006; Asgari et al., 2007; Bigot et al., 2009). Other than the bro genes, there are other repeated DNA regions distributed in the ascovirus genomes. In HvAv-3e, the repeat region encodes a conserved transposase (Asgari et al., 2007). Two homologous direct repeats (DRs) are identified in the TnAV-2c genome and 5 drs are identified in HvAV-3e whereas 4.5 homologous inverted repeats (IRs) are identified in the SfAV-1a genome (Bideshi et al., 2006; Wang et al., 2006; Asgari et al., 2007; Bigot et al., 2009). The IRs of SfAV-1 are non-coding and they differ in size. For example, they are smaller in two SfAV-1 variants, SfAV-1b and SfAV-1c. It was suggested that these IRs may be the region where the changes of genome configuration between linear and circular might occur during genome replication and only some of these large repeats may be essential for viral replication (Bideshi et al., 2006). A new ascovirus isolate, SfAV-1d, was first isolated from a S. frugiperda larva collected in a soybean field in South Carolina, USA in 2002. In this study, we show the high homology and the major difference between the SfAV-1a and SfAV-1d genomes. We further estimate the total size of the SfAV-1d genome to be about 100 kb, which makes it the smallest ascovirus genome so far. A major deletion of 14 kb in the IRs of SfAV-1a is found in SfAV-1d. The deletion region of SfAV-1d genome is highly variable. Since SfAV-1d has the smallest genome in ascoviruses, the study of the SfAV-1d

genome provides a better understanding of the requirement of essential genes for ascovirus replication and replication strategy in insect cells.

Materials and Methods Insects, insect cells and viruses. S. frugiperda, T. ni and S. exigua, Anticarsia gemmatalis and Pseudoplusia includens used in this project were lab maintained cultures on a pinto bean-based artificial diet in an insect rearing room with controlled temperature (27°C) and relative humidity (75%) and photoperiod (L:D 16:8 h). Insect cell lines for viral infection comparison included Sf21 cells derived from S. frugiperda and IOZCAS- Spex-II derived from the fat body of S. exigua. IOZCAS-Spex-II (Zhang et al., 2006) was a gift from Dr. Qi-Lian Qin (Institute of Zoology, Chinese Academy of Sciences). Sf21 cells were lab maintained cells. All cells were cultured in TNM-FH media supplemented with 10% fetal bovine serum at 27°C. SfAV-1d was isolated from a single S. frugiperda larva from cotton field in South Carolina and propagated in S. frugiperda larvae by piercing at the base of a proleg with a size 000 insect pin contaminated with SfAV-1d. The infected larvae were placed on an artificial diet and at day 7 p. i., opaque white hemolymph was collected by piercing the base of a proleg of the diseased larvae. The hemolymph was stored at -20°C (Govindarajan and Federici, 1990). The other ascoviruses used in this paper were SfAV-1a from S. frugiperda (a gift from Prof. B. A. Federici, University of California, Riverside, CA, USA), TnAV-2d and HvAV-3f isolated from H. virescens (Cheng et al., 2005). SfAV-1a was in the form of concentrated vesicles. TnAV-2d is actually a variant of TnAV-2a (Federici and Govindarajan, 1990). Viral DNA purification. Virus-containing vesicles from infected larvae were disrupted by ultrasonication and the virions were purified through sucrose density gradients (Federici and Govindarajan, 1990). Intact viral DNA was purified on CsCl-ethidium bromide gradients by ultracentrifugation (Cheng et al., 1999). Restriction fragment length polymorphism (RFLP) and DNA hybridization analysis. Genomic DNAs (500 ng) of TnAV-2d, HvAV-3f, SfAV-1a and SfAV-1d were digested with HindIII and BamHI (New England BioLabs) and separated on a 0.7% agarose gel. The lambda HindIII DNA marker (NEB) and the 1 kb DNA ladder marker were used as DNA size markers in the following hybridization analyses. Fragments were blotted onto a

Hybond N• nylon membrane (Amersham) (Southern, 1975; Chomczynski, 1992). Genomic DNA probe (400 ng) of SfAV-1a was digested with Sau3AI and the digested products (about 200 bp in length) were labeled with biotin (NEB Phototope Kit). Hybridization was carried out in 6 x SSC, 0.5% SDS and 5 x Denhardt's solution at 65°C for 21 h. The membrane was washed twice in 2 x SSC for 5 min at room temperature (23°C) and twice in 2 x SSC ± 0.1% SDS for 30 min at 65°C . The membrane was incubated with streptavidin and then biotinylated alkaline phosphatase. Chromogenic substrate (BCIP/NBT) was used for color development of hybridization products (Invitrogen). Pulse-field gel electrophoresis. Viral DNA (450 ng) was digested with HindIII, EcoRI and PstI separately under the conditions recommended by the manufacturers (NEB) and analyzed on 1% agarose gel for 17 hr at 5V/cm, 0.5-1.5 sec interval at 15ºC with DNA PFGE marker (NEB). The gel was stained with ethidium bromide and photographed with an Alphaimage gel documentation system. The accompanying software (version 5.1) was used to estimate the molecular mass of DNA fragments. Generation of SfAV-1d libraries for genome sequencing. Two SfAV-1d genomic libraries were generated for genome sequencing, one shotgun library and another HindIII fragment library. For the shotgun library construction, SfAV-1d DNA (2 µg) was partially digested with Sau3AI and separated by agarose gel electrophoresis (0.7%). The partially digested DNA in the range of 1 to 2 kb size was gel purified with the glassmilk method (Vogelstein and Gillespie, 1979) and ligated to the BamHI site of a cloning vector pBluscript SK+ using T4 DNA ligase (NEB). The vector had been previously dephosphorylated with calf intestinal phosphatase (CIP). The ligation was used to transform competent DH10B Escherichia coli cells by electroporation. The transformed cells were plated on Ampicillin (Amp)/Xgal LB plates and incubated at 37°C for 19 h. White colonies (960) were propagated in LB broth with Amp (100 μg/ml) in 96 deep- well plates and incubated for 19 h. The bacteria were pelleted by centrifugation and plasmid DNA was extracted by the alkaline method (QIAGEN). Plasmid DNA sequencing was performed using the universal M13 forward and reverse primers with the ABI BigDye version 3.1 in 96 well PCR plates following recommended procedure for

PCR amplification (Applied Biosystems by Life Technologies). The amplified products were analyzed with an ABI 3700. A total of 960 sequencing reactions were performed to produce an average of 700 bp readable DNA sequences. The obtained sequences were aligned against the genome sequence of SfAV-1a (Bideshi et al., 2006) to detect DNA deletions in SfAV-1d using computer program Sequencher (version 4.8). The HindIII fragment library was generated by digestion of 1 µg of SfAV-1d DNA with 10 units of HindIII (NEB) to completion. The cloning plasmid vector pCUGI- 1 that has two plasmids, pGEM-4Z and pIndigobac536 (Luo et al., 2001) was digested with HindIII and dephosphorylated with CIP. The digested SfAV-1d DNA and pCUGI-1 DNA were ligated with T4 ligase. Competent DH10B cells were transformed with the ligation and plated on LB-X-gal agar plates with either Amp (50 μg/ml) or Chloramphenicol (Chl) (12.5 μg/ml) antibiotics. DNA inserts smaller than 10 kb were screened on Amp plates and DNA inserts larger than 10 kb were screened on Chl plates. Positive clones carrying the SfAV-1d HindIII fragments were sequenced from both ends using M13 forward and reverse primers. PCR analysis, DNA cloning and sequencing. To confirm the deletion region of SfAV- 1d, a primer pair (DelF, 5’-TCACCCATGTGAGGATCG -3’ and DelR, 5’- TCATCGCCGTCACAACAC -3’) was designed according to the sequencing information in this region (Bideshi et al., 2006). Amplification of this region with the DelF and DelR primers was performed with LongAmp DNA polymerase (NEB) using SfAV-1a genomic DNA (20 ng) as the template or with regular Taq DNA polymerase (NEB) using SfAV- 1d (20 ng) as the template. The amplified products from SfAV-1a or SfAV-1d were analyzed on a 0.7% agarose gel. The amplified products in the range of 0.6 - 1.5 kb from SfAV-1d were gel purified by the glassmilk method (Vogelstein and Gillespie, 1979) and cloned to pGEM-T Easy vector (Promega). White colonies (17) were selected and sequenced using the M13 forward or reverse primers. Cell cultures and virus infection. Cell lines IOZCAS-Spex-II and Sf21 were absorbed separately with SfAV-1a and SfAV-1d at a multiplicity of infection (MOI) of 10

TCID50/cell for 1 h (O'Reilly et al., 1992) and observed and imaged at 24, 48, 72, 168, and 192 h p. i. . The rate of survived IOZCAS-Spex-II was calculated as the number of

cells which were able to divide after the infection (one cluster of dividing cells was counted as one) over the number of total cells seeded before the infection. Bioassay. Lab-reared S. exigua, T. ni, S. frugiperda, A. gemmatalis and P. includens were used in this study. Third instar larvae of each species were inoculated with SfAV-1a or SfAV-1d by puncturing them at the base of a proleg with a minute pin (size 000) that had been dipped in vesicle-containing hemolymph (Federici and Govindarajan, 1990). After inoculation, larvae were transferred to a pinto bean-based artificial diet and reared individually at 27 C and 16:8 h light dark conditions. For each ascovirus and host species, there were three replications with 30 larvae per replicate. Control larvae were inoculated with 0.1 M phosphate buffer, pH 7.0. Larval mortality was recorded. Mortality was reported as relative mortality using control as a reference (Cheng et al., 2000).

Results Restriction fragment length polymorphism and Southern hybridization analysis of SfAV-1d and SfAV-1a. SfAV-1d is closely related to SfAV-1a but its genome size is smaller than SfAV-1a. This conclusion was drawn from restriction fragment length polymorphism (RFLP) and Southern hybridization comparison between the two ascovirus isolates. In conventional agarose gel electrophoresis, the restriction fragment patterns of individual restriction endonuclease (REN) HindIII and BamHI digestion of genomic DNA of the two isolates showed that SfAV-1d is similar to SfAV-1a with 7 HindIII and 9 BamHI co-migrating fragments, but 3 HindIII and 7 BamHI fragments showed different mobility in agarose gel electrophoresis suggesting that SfAV-1d is distinct from SfAV-1a (Fig. 1-1A). Sequencing of the SfAV-1d HindIII fragment library end sequences showed that 4 of the 7 co-migrating HindIII REN fragments with SfAV-1a displayed 99% identity with only a few single nucleotide polymorphisms (data not shown). The largest fragments of HindIII and BamHI digestion of SfAV-1a were absent in the HindIII and BamHI digestion of SfAV-1d (Fig. 1-1A). To further confirm the similarities and differences between the two isolates, pulse field gel electrophoresis (PFGE) was conducted to increase the separation resolution of large REN fragments. The three large HindIII fragments (> 24.5 kb) in SfAV-1a were not present in the SfAV- 1d. One (>24.5 kb) and two (>29.9 kb) large fragments of BamHI and PstI fragments of SfAV-1a, respectively, were not present in SfAV-1d (Fig. 1-1C). After all the restriction fragments of each REN digestion of SfAV-1d were estimated and added up, the genome size of SfAV-1d was estimated to be about 100 kb, which is about 50 kb smaller than that of SfAV-1a (Bideshi et al., 2006). To further confirm that SfAV-1d is more closely related to SfAV-1a than to TnAV-2d and HvAV-3f, Southern hybridization was performed under high stringency hybridization conditions using SfAV-1a genomic DNA as the probe. Strong hybridization occurred between SfAV-1a and SfAV-1d, confirming the high homology between SfAV-1a and SfAV-1d. No or weak hybridization occurred between SfAV-1a and TnAV-2d or HvAV-3f (Fig. 1-1B). Collectively, these data suggest that the SfAV-1d is closely related to the reported SfAV-1a.

Figure 1-1. RFLP and Southern hybridization analysis of the new ascovirus isolate SfAV- 1d. (A) Gel electrophoresis of the restriction fragments (BamHI and HindIII) of SfAV-1a (a), SfAV-1d (d), TnAV-2d and HvAV-3f. (B) Southern hybridization with SfAV-1a genomic DNA labeled with biotin as probe. (C) Comparison of the restriction profiles of genomic DNA of SfAV-1a (a) and SfAV-1d (d) digested with HindIII, BamHI and PstI by pulse-field gel electrophoresis. Co-migrating fragments of SfAV-1a and SfAV-1d are indicated by slanted lines (white for fragments with identical DNA sequences between SfAV-1a and SfAV-1d and black for fragments with different DNA sequences).

Genomic DNA sequence deletions of SfAV-1d. From the estimation of the SfAV-1d genome size based on the restriction fragments, SfAV-1d is smaller than the published SfAV-1a genome. In order to confirm the deletion and study the essential genes for ascovirus replication in different cells and insects, a shotgun library of SfAV-1d genomic DNA was constructed and sequenced (unpublished data). When these shotgun library sequences were used to align against the published SfAV-1a genome sequence using the Sequencher program at 80% homology, several deletions including a major deletion of 14 kb were found (Fig. 1-2). The largest deletion is from the nucleotide (nt) position 79,979 to nt 93,882 of the SfAV-1a genome (Bideshi et al., 2006), which is the region with 4 non-coding IRs of SfAV-1a. Primers (DelF and DelR) were designed upstream and downstream of the deletion region (Fig. 1-2). The forward primer (DelF) anneals to the intergenic region between ORF70 and ORF71 (nt 79,918-79,935) of SfAV-1a; the reverse primer (DelR) anneals to ORF73 (nt 94,646-94,663) of SfAV-1a (Bideshi et al., 2006). The amplified products from SfAV-1d showed a major expected size of about 600 bp with some smears of DNA products in the range of 600 to 1,500 bp, whereas 6 discrete DNA fragments in the sizes of about 14, 10.5, 9, 5, 1.7 and 0.6 kb are amplified from SfAV-1a. Weak signal PCR products and smearing DNA were also observed between these discrete PCR products of SfAV-1a (Fig.1-3). Based on published genome sequence of SfAV-1a, the 14 kb PCR product is the expected size and these other DNA PCR products of SfAV-1a might also be from this variable region of the SfAV-1a genome as found in SfAV-1d (Fig. 1-3). These results also suggest that the wild type SfAV-1a and SfAV-1d are each a mixture of viruses containing different genomes with different lengths of DNA sequences at this locus. Cloning and sequencing of the deletion region of SfAV-1d. Since the amplified PCR products from SfAV-1d genomic DNA using the primers (DelF and DelR) were concentrated at 600 bp with DNA smears in the ranges of 600-1,500 bp (Fig. 1-3), PCR products in a large area covering the size distribution of 600-1,500 bp were gel purified and cloned. During colony screening, different sizes of inserts were identified in the ranges of 600-1,500 bp (Fig. 1-4A). Seventeen clones with different sizes of inserts were selected and sequenced. The sequence results showed that the deletion region of

Figure 1-2. Assembly of the SfAV-1d genome from the shortgun library sequences with SfAV-1a as the reference using Sequencher. All 960 shotgun library sequences were aligned to genome sequence of SfAV-1a using Sequencher (version 4.8) with the homology setting at 80%. The box shows the deletion region on the assembled SfAV-1d genome. The two arrowheads show the position of DelF and DelR primers.

Del F

Del R

Figure 1-3. Confirmation of the major 14 kb deletion of SfAV-1d comparing to SfAV-1a. DelF and DelR primers were used to amplify from viral genomic DNA of SfAV-1a and SfAV-1d separately. The amplified products from SfAV-1d are concentrated at 600 bp with DNA smears in the ranges of 600 to 1,500 bp while the amplified products from SfAV-1a show 6 discrete DNA fragments in the sizes of about 14, 10.5, 9, 5, 1.7 and 0.6 kb.

Figure 1-4. Characterization of the variable sequences in the deletion region of SfAV-1d. (a) PCR verification of the variable sizes of 17 clones of the deletion region using DelF and DelR. (b) Schematic representation of the variable deletion region of SfAV-1d. Linear diagram of the repeat region and the adjacent ORFs on SfAV-1a genome is shown as the reference. Solid lines indicate the sequences that are maintained in the SfAV-1d genome and the dotted lines indicate the missing sequences in SfAV-1d genome compared to that of the SfAV-1a.

80976 93888 79385 80029 94424 SfAV-1a 2.9kb repeat 2.9kb repeat 2.9kb repeat 2.9kb repeat ORF70 ORF71 ORF72 Primer F Primer R 79918 80433 SfAV-1d I 79918 80404 II 94365 94663

III 79918 80294 94542 94663

IV 79918 79995 94130 94663

V 79918 80318 94514 94663

VI 79918 80486

VII 79918 80837

79918 80202 93901 94663 VIII

SfAV-1d is a highly variable region (Fig. 1-4B). The shortest sequence covers nt 79,918- 80,294 and 94,542-94,663 of SfAV-1a, missing the whole IRs, ORF72 and part of ORF71; the longest sequence covers from 79,918-80,202 and 93,901-94,663, missing the whole IRs as well as part of ORF71 and ORF72. All 17 variants analyzed miss the whole IRs (Table 1-1). Among all these variants, 70.5% lack the whole ORF72 and 17.7% lack the whole ORF71, which suggests that ORF71 and ORF72 may not be essential for SfAV-1a or SfAV-1d replication. Since the wild type SfAV-1 could be a mixture of viruses containing different genomes with different DNA sequences in this variable region, it is also possible that the viruses which miss ORF71 and ORF72 are defective. Comparison of cell infection between SfAV-1d and SfAV-1a. Two cell lines, IOZCAS-Spex-II derived from S. exigua and Sf21 derived from S. frugiperda, were found permissive for SfAV-1d and SfAV-1a infection. In IOZCAS-Spex-II cell infection, SfAV-1d showed similar cytopathic effects to SfAV-1a, causing apoptosis in the infected cells leading to partition of the cells into virion containing vesicles and eventually to vesicle release. Unlike SfAV-1a that killed almost all the IOZCAS-Spex-II cells in 48 h p. i. , SfAV-1d was not able to kill a subpopulation of the IOZCAS-Spex-II cells. At 48 h p. i. , when almost all the IOZCAS-Spex-II cells were killed by SfAV-1a, some IOZCAS- Spex-II cells stayed uninfected showing distinct fibroblast-like shape typical of the uninfected or mock infected cells. These uninfected cells, about 3% of the total seeded cells for infection, continued dividing and eventually formed clumps of IOZCAS-Spex-II cells due to cell division at 192 h p. i. (Fig. 1-5). SfAV-1d developed similar infection efficiency in the Sf12 cells as SfAV-1a (data not shown).

Table 1-1. Summary of the clones of the SfAV-1d deletion region corresponding to the inverted repeat region of SfAV-1a

Name of the variant Position in SfAV-1a Percentage in Missing sequence total 17 clones SfAV-1d I 79918-80433 17.6% whole IRs and ORF72; part of ORF71 and ORF73 SfAV-1d II 79918-80404 + 11.8% whole IRs; part of 94365-94663 ORF72 SfAV-1d III 79918-80294+ 5.9% whole IRs; part of 94542-94663 ORF71 SfAV-1d IV 79918-79995+ 5.9% whole IRs and 94130-94663 ORF71; part of ORF72 SfAV-1d V 79918-80318+ 5.9% whole IRs and 94514-94663 ORF72; part of ORF71 SfAV-1d VI 79918-80486 29.4% whole IRs and ORF72; part of ORF71 and ORF73 SfAV-1d VII 79918-80837 17.6% whole IRs and ORF72; part of ORF71 and ORF73 SfAV-1d VIII 79918-80202+ 11.8% whole IRs and 93901-94663 ORF71; part of ORF72

Figure 1-5. Infection of the insect cell line IOZCAS-Spex-II by SfAV-1a or SfA-1d. IOZCAS-Spex-II cells were infected with SfAV-1a or SfAV-1d at an MOI of 10 viruses/cell and infection was monitored and image-recorded at 24, 48, 72, and 168 and 192 h p. i. .

Mock SfAV-1a SfAV-1d

24 h

48 h

72 h

168 h

192 h

SfAV-1d has a narrower host-range than SfAV-1a. SfAV-1 has a narrow host-range among all the ascoviruses that infect noctuids. However, the host-range diversity within SfAV-1 is still unknown. Five different insect species were used to compare mortality reported as relative mortality using control as a reference (Cheng et al., 2000) between SfAV-1d and SfAV-1a in this study. Based on larval mortality, S. exigua, T. ni and Pseudoplusia includens are susceptible to SfAV-1a, but not to SfAV-1d. All the larvae treated with SfAV-1d proceeded to pupa and adult moths. S. frugiperda larvae are susceptible to SfAV-1a and SfAV-1d and both viruses showed similar virulence. Anticarsia gemmatalis was resistant to both SfAV-1a and SfAV-1d (Table 1-2).

Table 1-2. Susceptibility of noctuid larvae to SfAV-1a and SfAV-1d (% mortality).

Viruses Host

S. exigua T. ni S. frugiperda A. gemmatalis P. includens

SfAV-1a 88.9 95.8 98.9 3.3 84.4

SfAV-1d 0 * 0 * 100 0 0*

* indicates the numbers in the same column are significantly different at p=0.05.

Discussion Since the discovery of ascovirus in the late 1970’s in Southeast US, ascovirus has been reported from other parts of the world such as Indonesia, Australia and France (Carner and Hudson, 1983; Bigot et al., 1997b; Cheng et al., 2000; Asgari, 2006). This suggests a worldwide distribution of ascoviruses. Interest in ascovirus research is rooted in the unique cytopathic effects of ascovirus on permissive cells resulting in apoptotic vesicles filled with virions. Further genetic study showed the virally encoded caspase of SfAV-1a is responsible for the apoptosis (Bideshi et al., 2005). One area, however, that has not been well addressed is identification of the genes required for ascovirus replication and infection in cells and its DNA replication strategy in cells. Here, to the best of our knowledge, we for the first time present data showing highly variable DNA deletions in a region of the SfAV-1d genome which corresponds to the 14 kb inverted repeat region in the SfAV-1a genome. Four ascovirus genomes have been sequenced and published. They are TnAV-2c, SfAV-1a, HvAV-3e and DpAV-4a and the genome sizes of these ascoviruses vary from 119 to 186 kb (Bideshi et al., 2006; Wang et al., 2006; Asgari et al., 2007; Bigot et al., 2009). SfAV-1a was isolated from S. frugiperda in Georgia, USA in 1982 (Hamm et al., 1986) and SfAV-1d was isolated from a single S. frugiperda larva in South Carolina. Geographically, SfAV-1a and SfAV-1d were isolated from the same region and theoretically they are closely related. Both conventional agarose gel electrophoresis and PFGE analysis showed that SfAV-1d and SfAV-1a have similar restriction profiles and Southern hybridization confirmed the relatedness (Fig. 1-1). The differences that contribute to the separation of the SfAV-1d from SfAV-1a are within the deletions of repetitive sequences in SfAV-1a since the genome sequence data obtained from SfAV-1d showed 99% identity with SfAV-1a at the DNA level (unpublished data). These DNA deletions in the SfAV-1d genome result in the formation of a smaller or smallest ascovirus genome so far reported. Comparing to SfAV-1a, the large fragments in SfAV- 1a are missing in SfAV-1d. Genome sequencing of SfAV-1d showed the 14 kb inverted repeat deletion in SfAV-1d corresponds to the largest 49 kb Hind III fragment of SfAV- 1a (Fig. 1-1 and unpublished data).

SfAV-1a was reported to have a genome size of 157 kb based on genome sequencing (Bideshi et al., 2006). This prediction of the SfAV-1a is likely based on the largest SfAV-1a genome with the longest 14 kb inverted repeat from. The PCR analysis result with DelF and DelR primer set using SfAV-1a as the template (Fig. 1-3) suggests that the IR region in SfAV-1a is also variable. Early estimation of SfAV-1a genome by REN showed a 140 kb with the largest HindIII fragment to be 43.9 kb (Fig. 6 in Federici et al., 1990). However the largest HindIII fragment from the genome sequencing of SfAV-1a is 49.2 kb (Bideshi et al., 2006). This further supports the notion that the 14 kb inverted repeats of SfAV-1a is a variable region as shown in Fig. 1-2. In addition to the deletion of the 14 kb inverted repeats region of the SfAV-1a in the SfAV-1d genome, the adjacent regions of the 14 kb inverted repeats region are also missing in SfAV-1d, which may suggest that this region is neither essential for SfAV-1d nor for SfAV-1a replication in permissive cells. The largest clone of the deletion region misses only the IRs and part of the adjacent ORF71 and ORF72 of SfAV-1a in SfAV-1d. The smallest clone misses the IRs and the whole ORF71 and ORF72 of SfAV-1a. ORF71 has no homology to known genes while ORF72 has high homology to the bro gene of baculovirus (Bideshi et al., 2006). It was reported that ascovirus is closely related to baculovirus (Cheng et al., 2007). In baculovirus genomes, homologous regions (hr) which are a set of closely related repeated sequences were found interspersed throughout the genome (Cochran and Faulkner, 1983; Ayres et al., 1994; Xie et al., 1995; Garcia-Maruniak et al., 1996; Tillman et al., 2004). Transient replication from a plasmid containing the hr sequence suggests that these hrs in the genome serve as replication origins and the replication ability of different hrs from different baculoviruses in heterologous systems can be variable, which may partially explain the host specificity of baculovirus replication (Pearson et al., 1992; Leisy and Rohrmann, 1993; Pearson et al., 1993; Lee and Krell, 1994; Ahrens et al., 1995; Leisy et al., 1995; Pearson and Rohrmann, 1995; Xie et al., 1995). Based on the study of hrs in baculovirus and the close relationship between ascovirus and baculovirus, we speculate that the inability to infect some of the S. exigua cells by SfAV-1d might be due to the missing inverted repeats from SfAV-1a that leads

to poor DNA replication in some of these cells. However, whether the missing IRs actually cause the host-range difference between SfAV-1a and SfAV-1d still needs to be investigated. Bideshi et al. reported that the IRs show different lengths and numbers of the repeats among SfAV-1a, SfAV-1b and SfAV-1c and suggested that the IRs are not essential for the viral replication of SfAV-1, which is supported by our data (Bideshi et al., 2006). The highly variable region reported in this study may serve as the origin of replication in a rolling circle replication manner, similar to other viruses such as baculovirus, herpes simplex virus and bacterial lambda phage, inside the permissive insect cells (Boehmer & Lehman, 1997; Oppenheimer & Volkman, 1997; Tanner et al., 2009). One possible explanation for the formation of this variable region is that during the cleavage of the concatenated linear ascovirus genomes by endonucleases either encoded by the virus or the cell, exonucleases from either the virus or the cell progressively degrade the ends of the linear genome of the ascovirus before the genome is circularized by ligation.

Acknowledgements The authors would like to thank Drs. Yayi Kusumah and Gerald R. Carner for providing the newly isolated SfAV-1d and help in insect infection. Dr. Brian A. Federici is greatly appreciated for donating SfAV-1a for the comparison between the two isolates. Dr. Qi-Lian Qin is thanked for providing the IOZCAS-Spex-II cell line. Undergraduates Andrew Kilianski, Samuel Murakami, Michael Howell are credited for viral DNA cloning and sequencing. The authors would like to thank Racheal A. Desmone for proof reading of this manuscript.

Chapter Two

Strategy of the use of 28S rRNA as a housekeeping gene in real-time quantitative

PCR analysis of gene transcription in insect cells infected by viruses

Jian-Li Xue, Tamer Z. Salem, Colin M. Turney, Xiao-Wen Cheng

Journal of Virological Methods (2010) 163: 210-215.

Abstract

Quantitative real-time reverse transcription-PCR (qRT-PCR) has been used widely to measure gene transcription regulation in cells. qRT-PCR must include one or more internal housekeeping genes to normalize data collection. A strategy to use the host cell 28S rRNA as a housekeeping gene in qRT-PCR analysis of gene transcription of insect cells infected by baculovirus and ascovirus was developed. It has been found that the 28S rRNA reverse primer can be incorporated in the oligo-dT-primed cDNA synthesis reaction. In such a way, amplification of 28S cDNA showed lower and less variable cycle thresholds in cells infected by viruses than by using only oligo-dT and other published housekeeping genes such as the TATA box binding protein (TBP) gene, the peptidyl prolyl isomerase A (PPI) gene and the ribosomal protein 13 (L13) gene. Incorporation of the 28S reverse primer in oligo-dT-primed cDNA synthesis also does not interfere with the detection of other polymerase II transcribed genes.

Introduction

Quantitative real-time reverse transcription-PCR (qRT-PCR) has been used widely in gene transcription analysis and clinical diagnosis of pathogens since its discovery in 1996 due to its simplicity of use, reproducibility, sensitivity and low amounts of RNA template input (Heid et al., 1996; Mackay et al., 2002; Radonic et al., 2004; Lu et al., 2006). To ensure reproducibility and accuracy in small reaction volumes, it is crucial to normalize the template inputs in qRT-PCR by selection and use of one or more appropriate internal control genes or housekeeping genes. The glyceraldehyde 3- phosphate dehydrogenase (GAPDH) gene and the actin (Act) gene have been commonly used as housekeeping genes for qRT-PCR to normalize reactions (Radonic et al., 2005). However, reports suggested that GAPDH and Act were not appropriate housekeeping genes to be used because they showed large variations in 16 different tested tissues; instead the RNA polymerase II (RNAPII) gene was the most stably expressed gene (Zhong and Simons, 1999; Selvey et al., 2001; Glare et al., 2002; Radonic et al., 2004).

In virus infected cells, it is common for the viruses to hijack and manipulate the cellular machinery for the sake of the virus replication (Hardwick, 2000; Ojala et al., 2000). Therefore, these reported housekeeping genes may not be suitable for qRT-PCR analysis of gene transcription regulation using total RNA from virus infected cells. A few candidate housekeeping genes for the virus infected cell samples were analyzed and it was reported that the TATA box binding protein (TBP) gene and the peptidyl prolyl isomeraseA(PPI) gene were more suitable housekeeping genes than GAPDH, Act and RNAPII (Radonic et al., 2005). Other structure-related genes, such as the ribosomal protein L13 (L13) gene, the tubulin (Tub) gene were also analyzed as potential housekeeping genes in qRT-PCR, but those genes turned out to be unstable at the mRNA levels in virus infected cells (Radonic et al., 2005). It is, therefore, suggested that these reported housekeeping genes may not be suitable for qRT-PCR analysis of gene transcription regulation using total RNA from virus infected cells. Eukaryotic 18S and 28S rRNA are used in Northern blot analysis as reference genes to normalize sample loading (Zhong and Simons, 1999; Hussain and Asgari, 2008; Salem et al., 2008). Their

use, however, as internal housekeeping control has been controversial in literature; some suggested that rRNA should be always used as an internal control (Schmittgen and Zakrajsek, 2000; Goidin et al., 2001; Bustin, 2002), whereas other rejected it (Pfaffl et al., 2004; Nicot et al., 2005). One of the reasons why 18S and 28S rRNA were not recommended as housekeeping genes for qRT-PCR might be due to the fact that the 18S and 28S rRNA are cleaved products from a large rRNA precursor. Unlike mRNA transcribed by RNA polymerase II that contains a poly(A) tail, 18S and 28S rRNA contain no poly(A) tail so that 18S and 28S rRNA cannot be annealed by the oligo-dT primer in cDNA synthesis before the qRT-PCR measurement (Radonic et al., 2004). Therefore, experiments using rRNA as housekeeping genes in qRT-PCR had to use random primers in the cDNA synthesis step (Thellin et al., 1999; Schmittgen and Zakrajsek, 2000; Blanquicett et al., 2002; Bond et al., 2002). In this study, this no- poly(A)-tail problem for the 28S rRNA was solved by using a 28S rRNAgene specific primer to synthesize cDNA. Reproducibility of 28S rRNA in qRT-PCR was compared with other studied housekeeping genes in virus infected cell samples. These genes include TBP, PPI and a ribosomal protein gene L14 (Radonic et al., 2005). The results of the studies suggest that the 28S rRNA is the most suitable housekeeping gene among the candidate housekeeping genes examined in virus infected samples.

Materials and Methods

Two insect-specific viruses with large DNA genomes, baculovirus that replicates in the nucleus and ascovirus that forms virus in the cytoplasm, were used for the tests. The baculovirus strain AcGFP (Autographa californica M nucleopolyhedrovirus or AcMNPV) (Cheng et al., 2001) and the ascovirus strain, Spodoptera frugiperda ascovirus 1a or SfAV-1a (Bideshi et al., 2006) were propagated according to standard procedures (O'Reilly et al., 1992). S. frugiperda insect cells (Sf21) were absorbed separately with AcGFP and SfAV-1a at a multiplicity of infection (MOI) of 10 p.f.u./cell for 1 h (O'Reilly et al., 1992) and harvested at 6, 12, 24, 36 and 48 h post infection (p.i.). Mock infected Sf21 cells were harvested at 48 h p.i. . Total RNA from 3×106 cells was extracted with the guanidinium thiocyanate–phenol–chloroform method (Chomczynski and Sacchi, 1987).

A two-step qRT-PCR method was used to evaluate the potential of the 28S rRNA gene as a housekeeping gene in qRT-PCR. The first step is the cDNA synthesis and the second step is quantitative PCR (qPCR). Total RNA (1 g) from each harvesting time p.i. was first treated with RQ1 RNase-Free DNase (Promega, Madison, USA) to remove contaminating DNA following recommended conditions of the enzyme supplier, and the effectiveness of this procedure was validated by PCR analysis as demonstrated (Cheng et al., 2001). The DNA-free RNA was used as a template with a combination of the Sf21 28SrRNAgene specific reverse primer (28S-R) and the oligo-dT primer (250mM final concentration each) using a DyNAmo cDNA Synthesis Kit (New England Biolabs, Beverley, USA) for first strand cDNA synthesis following the recommended conditions of the kit supplier (see Table 2-1 for primer sequences). The S. frugiperda 28S rRNA partial sequence was obtained from a cDNA clone when the Caenorhabditis elegans H4 gene primers were used to amplify the H4 gene of Sf21 cells (Salem et al., 2008). To determine if there was any interference of 28S-R primer to the detection of other genes using the oligo-dT primer in cDNA synthesis, another cDNA synthesis reaction with only the oligo-dT primer was performed. After the cDNA synthesis was completed, the cDNA was diluted four-fold with sterile water for qPCR.

qPCR was performed with a BIO-RAD iCycler iQTM system using a final reaction volume of 20 μl. Each reaction consisted of 1× iQTM SYBR green Supermix (BIO-RAD, Hercules, USA), 0.2 μM each of the gene specific forward primer and gene specific reverse primer as well as 2.5 μl template cDNA according to the instruction of the kit supplier. The qRT-PCR primer pairs were from genes of TBP, PPI, L14 and 28S rRNA (see Table 2-1). Amplifications were performed in triplicate. Amplifications started with an initial 1.5 min template denaturation step at 94°C, followed by 35 cycles of denaturation at 94°C for 5 s, primer annealing at 55°C for 20 s and primer extension at 72°C for 20 s. A DNA melting curve analysis was performed with an initial temperature at 55°C for 10 s followed by 80 cycles of a 0.5°C increase in each cycle for 10 s each.

Data acquisition and cycle threshold (Ct) calculation were performed using the associated Optical System software 3.1 (BIO-RAD, Hercules, USA). Data analysis was performed by t-test for the means using the Excel program (Microsoft). The ∆∆Ct was calculated as described (Radonic et al., 2005). First, ∆Ct for each time course was calculated as the difference between the Ct values of virus infected and mock infected samples. Then the maximum difference among all the time courses was calculated as ∆∆Ct. Conventional PCR using the same conditions as used in qPCR was performed to confirm the data collected from qRT-PCR. The PCR products were analyzed by agarose gel electrophoresis.

Table 2-1. A list of primers used in this study.

Primer names Primer sequences

Oligo-dT 3’ RACE 5’-GAGCACAGAATTAATACGACTCAC TAT AG GT12VN-3’ adapter primer*

TBP-F 5’- CAGCCACAAACACCACAA-3’

TBP-R 5’- CCAGCCAAGCTACCACTT-3’

L14-F 5’- CTGGCTTTCTGTCCACTT-3’

L14-R 5’- CCGTCTGAACTACCCATTC-3’

PPI-F 5’- CACCTGCAATGGCGTATG-3’

PPI-R 5’- GAGTTTGGTGGGATGACTG-3’

28S-F 5’- CGACGTTGCTTTTTGATCCT-3’

28S-R 5’- GCAACGACAAGCCATCAGTA-3’

* The primer is from Ambion Inc.

Results

28S rRNA shows lower Ct in qRT-PCR. One characteristic of a reliable housekeeping gene in qRT-PCR is low Ct. This was examined and compared with other reported housekeeping genes. In both AcGFP and SfAV-1a infected Sf21 cells over 5 sampling time periods (6, 12, 24, 36 and 48 h p.i.), the Ct values of the 28S gene were about two- fold significantly lower than that of TBP, L14 and PPI (Table 2-2). It is important that the primer pair for each gene (Table 2-1) was specific to the sequences they intended to amplify to produce a discrete product (amplicon) that leads to the Ct calculation in Table 2-2. A melting curve analysis of post amplification revealed that each gene primer pair produced a single peak (Fig. 2-1). This suggested that the differences of Ct values obtained from qRT-PCR on total RNA isolated from Sf21 cells infected by both insect viruses represented the differences of RNA copy numbers among 28S, TBP, L14 and PPI (Table 2-2).

28S rRNA has the least variation during viral infection. A typical housekeeping gene should also have the least RNA level variation when a cell is subjected to different experimental conditions including viral infection. We found that the 28S RNA levels had the least variation or fluctuation in Sf21 cells infected by either vAcGFP or SfAV-1a during the first 48 h p.i. among all the candidate genes examined (Fig. 2-2A and B). The standard deviation (SD) [±Ct] of the 28S gene was smaller than or close to one in Sf21 cells infected by AcGFP or SfAV-1a, respectively, whereas the other genes examined had greater SD [±Ct] values (>1) than that of the 28S gene (Table 2-3). A lower Ct SD value suggests a better reference gene while a higher SD indicates that the gene transcription levels are affected by virus replication. According to the calculated SD of the Ct values, the 28S rRNA gene showed a smaller SD value than TBP, L14 or PPI which suggested that 28S rRNA had the highest stability in Sf21 cells during both baculovirus and ascovirus infection (sum = 1.79, 4.88, 5.44 and 3.32 for 28S, TBP, L14 and PPI, respectively; Table 2-3).

To further confirm the highest stability of the 28S rRNA gene among all cellular genes examined during viral infection, another calculation mode that is based on the 43

Figure 2-1. Melting curve analysis of housekeeping gene amplification using gene specific primers by qRT-PCR. (A) S. frugiperda 28S rRNA. (B) S. frugiperda ribosomal protein L14. (C) S. frugiperda peptidyl prolyl isomerase A (PPI). (D) S. frugiperda TATA box binding protein (TBP).

Table 2-2. Comparison of average Ct of the possible reference genes.

TBP L14 PPI 28S

AcGFP 27.28 20.93 24.66 10.84*

SfAV1a 26.12 22.9 18.98 9.71*

* indicates significant difference to other tested genes (p<0.05)

Figure 2-2. Comparison of cycle threshold (Ct) fluctuation of housekeeping genes in Sf21 cells infected with viruses at different time points post infection. (A) Fluctuation of Cts of housekeeping genes (TBP, L14, PPI and 28S) in Sf21 cells infected by AcGFP. (B)

Fluctuation of CTs of housekeeping genes (TBP, L14, PPI and 28S) in Sf21 cells infected by SfAV-1a. Vertical lines on the bars represent the standard error of the mean in qRT- PCR (n = 3).

Table 2-3. Comparison of SD [±Ct] by Bestkeeper analysis.

TBP L14 PPI 28S

AcGFP 2.37 1.46 2.06 0.78

SfAV1a 2.51 3.98 3.31 1.01 sumv 4.88 5.44 5.37 1.79

calculation of ∆∆Ct values was used. Again, 28S rRNA showed the lowest ∆∆Ct value in Sf21 cells infected with either AcGFP or SfAV-1a (1.7 and 3.04, respectively; Table 2-

4). The combination of ∆∆Ct values of AcGFP and SfAV-1a also showed smaller than that of TBP, L14 and PPI (Table 2-4). This confirmed that the 28S rRNA was the most stable reference gene of Sf21 cells infected by AcGFP and SfAV-1a among all candidate genes examined.

Inclusion of the 28S rRNA-R primer into oligo-dT-primed cDNA synthesis decreases Ct and variation of 28S rRNA. Even though 28S rRNA is a cleaved product of a large rRNA precursor transcribed by cellular RNA polymerase I, qRT-PCR detected amplified products of 28S rRNA from cDNA synthesized with the oligo-dT primer. However, the Ct values were significantly higher than that from cDNA synthesized with the gene specific 28S-R primer in both virus infections (p < 0.01; Fig. 2-3A and B). Conventional PCR showed more amplified DNA products from cDNA synthesized using the 28S-R primer than using the oligo-dT primer, which confirmed the data collected from qRT-PCR (Fig. 2-3C and D). Not only was the average Ct value smaller in qPCR using cDNA synthesized with oligo-dT/28S-R primers than that from the oligo-dT primer, but also the variations (standard error of the mean) of the Ct during the two viral infection of Sf21 were smaller than that from oligo-dT-primed cDNA synthesis (Fig. 2- 3A and B). Comparing the average Cts of using a mixture of oligo-dT/28S-R primers with using the oligo-dT primer only showed that the Ct value increased from 10.4 to 16.92 (AcGFP) and from 9.71 to 23.8 (SfAV-1a) (p < 0.01; Fig. 2-3A and B).

Inclusion of the 28S rRNA-R primer into oligo-dT-primed cDNA synthesis does not interfere with the detection of other gene transcription. One major concern of this strategy by including the 28S-R primer into the oligo-dT-primed cDNA synthesis reaction was its potential interference with the detection of other target gene transcripts. Our results suggested that this was not the case. In Sf21 cells infected with AcGFP and SfAV-1a, no significant differences of Ct values of PPI were detected between qPCR using cDNA template synthesized from a mixture of the oligo-dT/28S-R primers and qPCR using cDNA template synthesized only with the oligo-dT (p = 0.7156 and 0.9470;

Table 2-4. Comparison of ΔΔCt of the possible reference genes.

TBP L14 PPI 28S

AcGFP 5.53 3.4 5.0 1.7

SfAV1a 7.53 9.4 6.97 3.04 sumv 13.06 12.8 11.97 4.74

Figure 2-3. Comparison of cycle threshold (Ct) in qRT-PCR using mixture of oligo- dT/28S-R and oligo-dT in cDNA synthesis for 28S gene amplification. (A) Comparison of average and standard deviation of 28S rRNA Cts in Sf21 cells infected with AcGFP at 48 h p.i. (n = 3). (B) Comparison of average and standard deviation of 28S rRNA Cts in Sf21 cells infected with SfAV-1a at 48 h p.i. . Vertical lines on the bars represent the standard error of the mean in qRT-PCR (n = 3). (C) Comparison of PCR product yields of 28S rRNA in Sf21 cells infected with AcGFP at 48 h p.i. by conventional PCR. (D) Comparison of PCR product yields of 28S rRNA in Sf21 cells infected with SfAV-1a at 48 h p.i. by conventional PCR.

respectively; Fig. 2-4A and B). This result was confirmed further by conventional PCR that no detectable difference could be visualized between the amplicons of PPI using cDNA synthesized with the oligo-dT primer with or without the 28S-R primer (Fig. 2-4C and D). Similar tests were also conducted on L14 and TBP transcription changes in Sf21 cells infected by AcGFP and SfAV-1a and the results were similar to these obtained from the PPI tests (data not shown). All these confirmed that inclusion of the 28S-R primer into the oligo-dT-primed cDNA synthesis reaction did not affect the detection of transcripts synthesized by RNA polymerase II in Sf21 cells.

Figure 2-4. Comparison of Cts of PPI gene amplification in qRT-PCR between oligo- dT/28S-R and oligo-dT synthesized cDNA for interference detection. (A) Comparison of average Cts of PPI in Sf21 cells infected with AcGFP at 48 h p.i. (n = 3). (B) Comparison of average Cts of PPI in Sf21 cells infected with SfAV-1a at 48 h p.i. (n = 3). (C) Comparison of PCR product yields of PPI in Sf21 cells infected with AcGFP at 48 h p.i. by conventional PCR. (D) Comparison of PCR product yields of PPI in Sf21 cells infected with SfAV-1a at 48 h p.i. by conventional PCR. Vertical lines on the bars represent the standard error of the mean in qRT-PCR (n = 3).

Discussion

For a gene to be considered as a housekeeping gene in qRT-PCR analysis of gene transcription regulations in cells, the candidate housekeeping genes must satisfy two criteria: (1) presence in relatively large abundance and (2) maintaining high stability in cells subjected to different experimental conditions including virus infection. The more transcripts the gene produces, the lower the Ct it will produce in qRT-PCR. In eukaryotic cells, the cellular structural genes such as actin and tubulin are highly active and their transcripts are very abundant in the cells. When the cells are infected by viruses, the transcription levels of these structural genes are affected depending on the cell–virus system (Radonic et al., 2005).

The rRNA is a major component of the ribosome and each ribosome in the eukaryotic cell has a copy of each of the rRNAs. Ribosomes are highly abundant in the cytoplasm of mammalian cells and the rRNA constitutes about 85% of the total RNA isolated from most of the eukaryotic cells (Bhatia et al., 1994; Sambrook and Russell, 2001). Since virus infected eukaryotic cells need host ribosome to synthesize their own viral proteins, it is reasonable to hypothesize that virus infection in general should not affect the ribosome levels dramatically. Therefore, the rRNA should be a good candidate for use as a housekeeping gene in qRT-PCR. Early studies used random primers to synthesize cDNA from 18S rRNA for housekeeping gene studies in qRT-PCR and suggested rRNA was stably maintained in cells under different growth conditions (Thellin et al., 1999; Schmittgen and Zakrajsek, 2000; Blanquicett et al., 2002; Bond et al., 2002). The use of oligo-dT primers has several advantages over using random primers in cDNA synthesis. For example, (1) oligo-dT primers are cheaper than the random primers, (2) cDNA synthesis using oligo-dT primer starts at the 3’-end of mRNAs whereas cDNA synthesis using random primers can initiate anywhere in the RNA leading to partial cDNA of transcripts that may affect accuracy of detection (Avison, 2007). Another potential problem associated with the use of random primers for cDNA synthesis is the large percentage of rRNA in any isolated total RNA samples that may use a biased large portion of the random primers that leads to reduction in

availability of random primers for other RNAs during cDNA synthesis. Therefore, a different strategy such as the one reported in this study to use rRNA as a house keeping gene in qRT-PCR can be employed.

Conventional qRT-PCR analysis on gene transcription regulation involves a first- step cDNA synthesis that uses the oligo-dT primer to anneal to the poly(A) tail of the transcripts (Sambrook and Russell, 2001). It was also reported that non-abundant polyadenylated transcripts of 28S rRNAs were detected in human cells and many of the poly(A) tails were composed of heteropolymeric poly(A)-rich sequences containing the other nucleotides in addition to adenosine (Slomovic et al., 2006). This explained the amplification of the 28S rRNAgene in the qRT-PCR and conventional PCR using the cDNA synthesized from oligo-dT primer in this study (Fig. 2-3). However, the Ct values were significantly higher and more variable during the course of viral infection when only the oligo-dT primer was used in the cDNA synthesis step than when the 28S-R primer in conjunction with the oligo-dT primer was used (Fig. 3A and B). It is, therefore, essential for the 28S-R primer to be added to the oligo-dT-primed cDNA synthesis reaction as a housekeeping gene since incorporation of the 28S-R primer did not interfere with the amplification of other genes (Fig. 2-4). However, this strategy of including the 28S-R primer into the oligo-dT-primed cDNA synthesis reaction is not applicable for cDNA synthesis with purified mRNA as template and is only applicable when total cellular RNA is used as template for cDNA synthesis.

The stability of the mammalian 28S rRNA gene in cells infected by viruses was not examined by qRT-PCR, but mammalian rRNA has long been used as reference in Northern blot analysis of gene transcription. It is, therefore, reasonable to assume that human rRNA gene can also be used as a housekeeping gene in qRT-PCR analysis of gene transcription of cells infected by human viruses following the same strategy described above.

Acknowledgements

We would like to thank Dr. C.M. Senthil Kumar for proof reading of this manuscript. This work is partially supported by a USDA cooperative agreement (58- 3148-7-164) awarded to X-WC.

Chapter Three

In vitro transcription from different SfAV-1d promoters with nuclear extract from uninfected insect Sf21 cells

Introduction

DNA-dependent RNA polymerase (DdRP or RNAP) is an essential enzyme in the transcription process of prokaryotic and eukaryotic organisms as well as cytoplasmic DNA viruses, such as poxviruses, iridoviruses, asfarviruses, and ascoviruses (Sonntag and Darai, 1995; Cheng et al., 2000). Cytoplasmic DNA viruses have their own RNA polymerase since they have no direct access to the cellular transcription machinery that is located in the nucleus of a cell. However, among the cytoplasmic DNA viruses that encode their own viral RNA polymerase, their independence on host RNA polymerase is quite different depending on their degree of access to the nucleus during their life cycles. Poxviruses are the most independent of the host cell gene expression machinary among the cytoplasmic DNA viruses, their whole life cycle being completed in the cytoplasm (Granados, 1973; Moss, 2001). They have all the proteins required for transcription of the early genes packaged within the large virus particle (virion) to support early gene transcription without the requirement of host factors, although the late genes of poxviruses require some host factors (Broyles and Kremer, 2004). The iridoviruses replicate first in the nucleus then continue in the cytoplasm and they depend on the host RNA polymerase II for their early gene transcription (Goorha et al., 1981; Williams et al., 1998). It is rare for nucleus-replicating viruses to have their own viral RNA polymerase, baculovirus being the only one known (Federici et al., 2000). The expression cascade of baculovirus genes are divided into four categories, immediate-early, early, late, and very late (Westwood et al., 1993). It uses host RNA polymerase to transcribe immediate-early and early genes including viral RNA polymerase subunits, LEF-8, LEF- 9, P47 and LEF-4. Then the virus encoded four-subunit RNA polymerase is expressed and assembled to transcribe late and very late genes (Passarelli, 2007).

The newly defined virus family Ascoviridae has a double-stranded DNA (dsDNA) genome and the location where their DNA replication occurs in the cell has not been experimentally identified (Federici et al., 2000; Cheng et al., 2007). However, it was reported that replication of ascoviruses may start in the nucleus based on the fact that a virogenic stroma (VS) is often observed in the nuclei of infected cells (Federici, 1983; Cheng et al., 2000; Cheng et al., 2007). Infection of a permissive cell by an ascovirus 61

often results in the rupture of the nucleus, after which virion morphogenesis occurs in the cytoplasm (Federici et al., 2009). In the late phase of ascovirus infection, ascoviruses partition the host cells by inducing apoptosis and eventually the virions become encompassed within vesicles followed by cell disintegration (Federici, 1983). Whether ascoviruses have access to the host replication, transcription machinery in the nucleus has not been studied.

The evolutionary origin of cytoplasmic DNA viruses and how ascoviruses are related to them has not been well studied. It was predicted that cytoplasmic DNA viruses might have evolved from nuclear replicating baculoviruses (Cheng et al., 2007). Ascoviruses are more closely related to baculoviruses than other cytoplasmic DNA viruses (Cheng et al., 2007). Four ascovirus genome sequences are available and all the genomes have host RNA polymerase subunit homologues (Bideshi et al., 2005; Wang et al., 2006; Asgari, 2007; Bigot et al., 2009): SfAV-1a has three RNA polymerase subunit homologues and two transcription factors; TnAV-2c has three RNA polymerase subunit homologues and two transcription factors; HvAV-3 has three RNA polymerase subunit homologues; DpAV-4a has five RNA polymerase subunit homologues and three transcription factors. Since ascoviruses are closely related to baculoviruses and also have RNA polymerase subunit homologues, it is possible that they follow the same transcription strategy as baculoviruses. A new member of the Ascoviridae, SfAV-1d, has been isolated. Preliminary studies showed that SfAV-1d has a smaller genome than the earlier reported SfAV-1a (Chapter 1). Since SfAV-1d has the smallest known ascovirus genome, it will be helpful to use it to study the essential genes for the viral life cycle including the transcription strategy for expressing virus genes.

In vitro transcription is a routinely used system to understand gene transcription dependence on certain RNA polymerases (Hoopes and Rohrmann, 1991; Hamm et al., 1998). It was reported that poxvirus early promoters can be recognized by virion protein extract in in vitro transcription studies and baculovirus early promoters can be recognized by host cell nuclear extract (Hoopes and Rohrmann, 1991; Broyles and Kremer, 2004; Passarelli, 2007).

In order to determine whether ascoviruses gene transcription follows the same strategy as baculoviruses, partial genome annotation has been performed for SfAV-1d to identify open reading frames that would produce products that could possibly be involved in transcription. Three RNA polymerase subunit homologues were found in the SfAV-1d genome. Early and late genes were defined using the aphidicolin treatment to inhibit the viral DNA replication. Nuclear extract was obtained from uninfected insect cells to perform in vitro transcription with early or late promoters from the SfAV-1d genome. It was found that the DNA polymerase gene is an early gene and the major capsid protein (mcp) gene is a late gene. The promoters of these genes were tested for transcription in nuclear extracts prepared from uninfected host cells. The early promoter can be recognized by host cell nuclear extract while the late promoter cannot be recognized by host cell nuclear extract.

Materials and Methods

DNA sequence analysis. Open reading frames (ORFs) were searched with the known large contigs obtained in Chapter 1 by using the Glimmer program (Version 1.02, TIGR), and the ORFs were selected according to the following criteria: (1) larger than 180 bp; (2) not contained within larger ORFs and (3) no major overlap with the adjacent ORFs.

Phylogenetic analysis. A phylogenetic tree was inferred from biological characters of the four cytoplasmic viruses and nuclear-replicating baculovirus by the discrete character method, CLIQUE, in PHYLIP version 3.5 (J. Felsenstein, Department of Genetics, University of Washington). A neighbor-joining phylogenetic tree was inferred from predicted amino acid sequences of DNA polymerse genes from different viruses by PAUP version 4 (D. L. Swofford, Sunderland, MA) (Cheng et al., 2007).

Plasmids, cloning and sequencing. Primers DNApolPF and DNApolPR (Table 3-1) were used to amplify the promoter region of the DNA polymerase gene (1 kb upstream of the start codon of the DNA polymerase gene) using SfAV-1d viral DNA as the template. Primers MCPPF and MCPPR (Table 3-1) were used to amplify the promoter region of the MCP gene (1 kb upstream of the start codon of the MCP gene) using SfAV-1d viral DNA as the template. PCR products from the amplifications of DNA polymerase gene and MCP gene were gel purified and subcloned into pGEM-T easy. Clones (DNApolP-T and MCPP-T) were screened and confirmed by sequencing. BamHI and HindIII were used to retrieve the promoter region of the two genes and the fragments were gel purified and ligated to BamHI- and HindIII-linearized pBlueGFP (Cheng et al., 2001). Clones (DNApolP-GFP and MCPP-GFP) were screened and confirmed by sequencing. Clone AcE#9 is from an EcoRI genomic library of AcMNPV which includes the promoter and ORF of the lef-8 gene.

Viral replication inhibition analysis. To determine the time course of gene transcription of the DNA polymerase gene and the mcp gene (early or late), aphidicolin (APH) was used to treat the cells infected by SfAV-1d. Sf21 cells (1×106) were seeded in each well of a 6-well plate and then infected with SfAV-1d at an MOI of 10 TCID50/cell. The inoculum was removed and replaced with fresh cell growth media containing APH at 5 64

mg/ml in DMSO to inhibit viral DNA replication (Miller et al., 1981; D'Costa et al., 2001). The treated cells were incubated at 27°C. The controls were cells infected by SfAV-1d without APH but with DMSO. Cells were harvested for RNA extraction at 24 h p.i. (Chomczynski and Sacchi, 1987). Real-time PCR was used to detect the inhibitory effect of APH on the transcription of the DNA polymerase gene and the mcp gene. Total RNA (1 mg) from every sample was ﬁrst treated with DNase to remove contaminant DNA following recommended procedures (Promega). The DNA-free RNA was used as the template to anneal to an oligo-dT 3’ RACE adapter primer (Salem et al., 2008) and a 28S rRNA (S. frugiperda) reverse primer (Xue et al., 2010) at 25°C for 10 min, and then followed by a primer extension reaction with M-MuLV reverse transcriptase (NEB) at 42°C for 60 min to synthesize cDNA. Equal amounts of cDNA were used as templates for Real-time PCR analysis. The primer pairs used for the DNA polymerase gene and the mcp gene transcripts were DNApolRTF and DNApolRTR as well as MCPRTF and MCPRTR (Table 3-1). 28S rRNA from S. frugiperda was detected by a pair of 28S rRNA primers for ampliﬁcation to serve as an internal control (Xue et al., 2010). Each gene detection was performed in triplicate for each cDNA sample. The effects of APH on transcription of the DNA polymerase gene and the mcp gene, expressed as differences normalized to 28S rRNA, were statistically analyzed by the t-test (Xue and Cheng, 2010).

In vitro transcription assay. Nuclear extract preparation was performed using Sf21 cells at a density of 3×106 cells/ ml following the protocol described by Hoopes and Rohrmann (Hoopes and Rohrmann, 1991). The nuclear extract was stored in liquid nitrogen. In vitro transcription assays with the nuclear extract from Sf21 cells were carried out in a 25-μl volume, including 2 μg linearized DNA template, 20 mM HEPES (pH 8.4 at 25°C), 6 mM MgCl2, and 5 μl of nuclear extract. The linearized DNA templates include: IEIGFP linearized by BglII; AcE#9 linearized by XbaI; pBluGFP linearized by XhoI; DNApolP- GFP linearized by XhoI and MCPP-GFP linearized by XhoI. The reaction with the DNA template, nuclear extract and the reagents were pre-incubated at 30°C for 20 min and then nucleotides (600 μM each of ATP, CTP, GTP, UTP) were added to the reaction to initiate transcription. The reaction was incubated at 30°C for 40 min and then stopped by addition of 25 μl stop buffer (0.5% SDS/10 mM EDTA/100 mM NaOAc, pH 5.2). Equal 65

volume of 1:1 phenol/chloroform was used to extract the in vitro transcription products. The phenol phase was back-extracted by adding 55 μl stop buffer and the combined aqueous phases were precipitated with 2.5X volume of 100% ethanol and 1 μl precipitation carrier (Molecular Research Center Inc.). After 25 min incubation at -80°C, RNA was pelleted using a microcentrifuge at a maximum speed (13,200 rpm) for 15 min at 4°C and air dried at room temperature for 25 min, then suspended in DEPC-treated

H2O. Concentrations of the suspended RNA were estimated by a ND-1000 Spectrophotometer (NanoDrop).

RT-PCR analysis. RNA (1 μg) was treated by DNase (Promega) and amplified by the one step reverse transcription PCR using the Access RT-PCR system (Promega). The transcripts of baculovirus IE1, LEF8 and polyhedrin genes were detected by primer pairs of IE1F and IE1R, LEF8F and LEF8R, POLF and POLR in the RT-PCR (Table 3-1). The transcripts of the SfAV-1d DNA polymerase and mcp genes were detected by primer pairs of DNApolF and DNApolR, MCPF and MCPR (Table 3-1). The RT-PCR program was 45°C for 45 min; 94°C for 2 min; 40 cycles of 94°C 30 sec; 60°C 1 min; 68°C 2 min; 68°C for 7 min; store at 4°C. The RT-PCR products were analyzed by agarose gel electrophoresis.

Table 3-1. Oligonucleotide primer sequences used in real-time PCR and in vitro transcription.

Primer name Primer sequence DNApolRTF 5’- CGACCAGTTCGTCACATC -3’ DNApolRTR 5’- TCGTCTTCGTGAGCATCT -3’ MCPRTF 5’- GGTCTCTAGGATCGGTATTAAGTG -3’ MCPRTR 5’- CGTGGTATGGTTCCGTGTA -3’ SF28SF 5’- CGACGTTGCTTTTTGATCCT -3’ SF28SR 5’- GCAACGACAAGCCATCAGTA -3’ GFP112F 5’- GCAAGCTGACCCTGAAGTTCATC -3’ GFP506R 5’- CGGATCTTGAATTCACCTTGATG -3’ AcLEF8F2 5’- AGACCTGACCGAGTGGAT-3’ AcLEF8R 5’- GTTGCACATCAACAGGCT-3’ DNApolPF 5’- AAGCTTGTACAAGAATTTCGTCCTC -3’ DNApolPR 5’- GGATCCTAACTATTCCAGCGTTC -3’ MCPPF 5’- AAGCTTCCGTTATAGAGGA -3’ MCPPR 5’- GGATCCATCTAACGATT -3’

Results

Partial genome annotation. Three RNA polymerase homologues were found in the genome of SfAV-1d from the sequences of the major contigs assembled in Chapter 1 (Table 3-2). ORFP1 encodes a putative 98.4-kDa protein which is a homologue of DNA- directed RNA polymerase β’ subunit. ORFP2 encodes a putative 128.1-kDa protein which is a homologue of DNA-directed RNA polymerase β subunit. ORFP3 encodes a putative 52.2-kDa protein which is a homologue of DNA-directed RNA polymerase α subunit.

Phylogenetic analysis of ascovirus. Since three RNA polymerase homologues were found in SfAV-1d and all the ascovirus genomes have three RNA polymerase homologues, it was important to test whether these three RNA polymerase homologues are actually functional proteins in the virus life cycle. Cytoplasmic viruses usually encode their own RNA polymerase and baculoviruses are the only nuclear-replicating virus which encodes its own RNA polymerase. Since at the late infection stage the nuclear envelope is disrupted by ascoviruses, no virus has been found in the nucleus. Whether ascoviruses ever gets into the nucleus during its life cycle is still unknown.

In order to study the possible function of the three RNA polymerase homologues in SfAV-1d genome, phylogenetic analysis was performed to study the relationship between ascoviruses and other nuclear-replicating or cytoplasmic viruses. Phylogenetic analysis based on the predicted amino acid sequences of the DNA polymerase genes (Fig. 3-1B) and biological characters, such as virus structure, genome configuration, viral budding etc. (Fig. 3-1A) showed that nuclear-replicating baculoviruses are more closely related to ascoviruses than other cytoplasmic viruses, which makes it more likely that ascovirus and baculovirus might share similar transcription strategies.

Identification of early and late genes. In baculoviruses, early genes including viral RNA polymerase subunit genes are transcribed and expressed by the host cell transcription machinery and the viral RNA polymerase subunits assemble to form a functional enzyme to transcribe late genes. In order to test whether host RNA polymerase

Table 3-2. Partial genome annotation of the SfAV-1d genome.

ORF Fragment of No. of Molecular Homologues SfAV-1d amino acids mass (kDa) P1 Frag B 885 98.4 DNA-directed RNA polymerase β’ subunit P2 Frag A 1157 128.1 DNA-directed RNA polymerase β subunit P3 Frag A 462 52.2 DNA-directed RNA polymerase α subunit

Figure 3-1. Phylogenetic relationship between baculovirus and ascovirus. (A) Phylogeny of ascovirus based on biological characters. Phylogenetic tree was inferred from biological characters of the four cytoplasmic viruses and nuclear-replicating baculovirus by PHYLIP version 3.5 (J. Felsenstein, Department of Genetics, University of Washington) with baculovirus served as an out-group. (B) Phylogeny of ascovirus based on genetic data. NJ phylogenetic tree was inferred from amino acid sequences of DNA polymerase genes from different viruses by PAUP version 4 (D. L. Swofford, Sunderland, MA).

is involved in ascovirus gene transcription and whether a viral RNA polymerase is used in the ascovirus life cycle, early and late genes have to be identified in the ascovirus genome. Since early and late genes are defined based on whether a gene is transcribed before or after viral DNA replication, APH was used to inhibit viral DNA replication in order to define early and late genes in the ascovirus genome. The host 28S rRNA was used as an internal control gene in the real-time PCR analysis of APH treated or non- treated samples. The transcriptional level of the DNA polymerase gene was not affected by the APH treatment suggesting that the SfAV-1d DNA polymerase gene is an early gene. However, the transcription level of the mcp gene was reduced 195-fold in APH- treated cells infected with SfAV-1d compared to no APH treatment. This confirmed that the mcp gene is a late gene (Salem et al., 2008; Fig. 3-2).

Dependence of viral genes transcription on the host RNA polymerase. In order to study the dependence of viral gene transcription on the host cell transcription machinery, an in vitro transcription assay with host cell nuclear extract was used in this study. The early gene promoter and the late gene promoter from the ascovirus genome were used in the in vitro transcription assay using the early gene promoters and the late gene promoter of AcMNPV as the reference. The baculovirus early promoters of ie-1 and lef-8 were recognized by host cell nuclear extract, whereas a faint signal was detected in the RT- PCR analysis from the late AcMNPV polh gene promoter (Fig. 3-3). RT-PCR detected transcripts in the in vitro transcription reaction when SfAV-1d DNA polymerase promoter was used but no transcript was detected from the ascovirus late mcp gene promoter. Taken together, this suggests that the early genes from ascovirus are transcribed by the host RNA polymerase, but the late genes are transcribed by either a viral RNA polymerase or by the host RNA polymerase with some viral factors involved.

Figure 3-2. Identification of early and late genes in the SfAV-1d genome. Sf21 cells 6 (1×10 ) were infected with SfAV-1d at an MOI of 10 TCID50/cell. The inoculum was removed and replaced with fresh cell growth media containing APH at 5 mg/ml in DMSO to inhibit viral DNA replication. The controls were cells infected by either SfAV- 1d without APH but with DMSO. Samples were harvested for RNA extraction at 24 h p.i.. Total RNA (1 mg) from every sample was ﬁrst treated with DNase and then used as the template to anneal to an oligo-dT 3’ RACE adapter primer and a 28S rRNA (S. frugiperda) reverse primer to synthesize cDNA. Equal amounts of cDNA were used as templates for real-time PCR analysis. Detection of each gene was performed in triplicate for each cDNA sample. The effects of APH on transcription of the DNA polymerase gene and the mcp gene, expressed as differences normalized to 28S rRNA (calculation was performed as described in Xue and Cheng, 2010) were statistically analyzed by the t- test. * indicates significant difference to other tested genes (p<0.05).

Figure 3-3. In vitro transcription analysis of different baculovirus and ascovirus promoters with insect Sf21cell nuclear extract. In vitro transcription products (1 μg) were treated with DNase and amplified by the one-step reverse transcription PCR using the Access RT-PCR system with or without the reverse transcriptase. The transcripts of baculovirus IE1 (early 1), LEF8 (early 2) and polyhedrin (late) genes as well as ascovirus DNA polymerase gene (early) and MCP gene (late) were detected with specific primer pairs. The RT-PCR products were analyzed by agarose gel electrophoresis.

Discussion

Transcription is a fundamental step in gene expression in cells regardless of whether they are prokaryotic or eukaryotic. In viruses that are obligate intracellular parasites, they depend on cellular machinery for certain steps in the virus infection cycle. In cells, RNA polymerase plays a major role in transcription (Sonntag and Darai, 1995). Eukaryotes have three types of RNA polymerases, RNA polymerase I, II, and III. Eukaryotic RNA polymerase II is mainly responsible for transcribing precursors of mRNA and most snRNAs as well as microRNAs and it is composed of 8-12 subunits whereas the core bacterial RNA polymerase has 5 subunits (Sonntag and Darai, 1995). Dependence of viruses on the host machinery for gene expression varies from virus to virus. For example, cytoplasmic viruses encode their own viral RNA polymerase to transcribe virus genes since they don’t have access to host RNA polymerase (Sonntag and Darai, 1995; Cheng et al., 2007). Baculoviruses which replicate in the nucleus, also encode their own four-subunit RNA polymerase for late gene expression that is independent of the host RNA polymerase (Passarelli, 2007). We found that the cytoplasmic ascovirus can use the nuclear host RNA polymerase for early gene transcription, similar to baculovirus.

Transcription dependence of cytoplasmic ascoviruses on host cellular RNA polymerase has not been studied before. Furthermore, TnAV-2c, SfV-1a, and HvAV-3e have three host RNA polymerase subunit homologues while DpAV-4a has five homologues. SfAV-1d also has three RNA polymerase subunit homologues (Table 3-2). Whether the three-subunit RNA polymerase of ascovirus can form a functional polymerase enzyme has not been tested. Part of the difficulty in testing the functionality of the three subunits RNA polymerase has been poor expression of the three subunits in either the baculovirus expression vector system or the bacterial expression system for in vitro biochemical studies (unpublished data). Since no RNA polymerase has been reported to have three subunits, it will be interesting to find out whether these three homologues have a function in viral gene transcription similar to that of the four-subunit RNA polymerase enzyme of baculovirus (Passarelli, 2007).

Phylogenetic studies showed that ascoviruses are more closely related to the nuclear-replicating baculoviruses than all the other cytoplasmic viruses (Cheng et al., 2007). It was speculated that ascovirus has an evolutionary position between nuclear- replicating viruses and cytoplasmic viruses which are more independent of the host cell transcription machinery (Cheng et al., 2007). Since ascoviruses and baculoviruses are closely related, it was important to investigate whether ascoviruses are dependent on host RNA polymerase like baculoviruses. Early and late gene groups were identified, with the DNA polymerase gene from the SfAV-1d genome representing an early gene and the mcp gene representing a late gene (Fig. 3-2).

In vitro transcription was used to test the dependence of ascovirus gene transcription on the host RNA polymerase. Baculovirus genes were used as references in the assay. The baculovirus promoters of early genes like ie-1, lef-8 were both recognized by cellular RNA polymerase in the Sf21 nuclear extract which is consistent with results in previous publications (Fig. 3-3) (Sonntag and Darai, 1995; Passarelli, 2007). But in the in vitro transcription assay with the baculovirus late polh promoter, a faint signal was detected in the RT-PCR analysis of cDNA product from the in vitro transcription of gene under the control of the late polh promoter. It was reported that only early genes from baculovirus genomes are transcribed by the host RNA polymerase. The small amount of the detected transcripts initiated from the polh promoter with Sf21 cellular nuclear extract could be due to the basal level of transcription with high structural conservation of the RNA polymerases. The early promoter from the ascovirus genome was recognized by the host nuclear extract but the late promoter was not recognized by the host, which suggests that ascoviruses can use the host RNA polymerase for early gene transcription but the late genes of ascoviruses might be transcribed by the virus-encoded RNA polymerase or the host RNA polymerase along with additional viral factors.

In this study, nuclear extract from uninfected cells was used. To further study the ascovirus gene transcription patterns, nuclear extract from ascovirus-infected cells could be used to test whether the late gene promoter of ascovirus can be recognized by infected cell nuclear extract. Since the nuclei of ascovirus-infected cells break down during the

infection, it will be helpful to determine the time at which the nuclear breakdown occurs during ascovirus infection before making the infected cell nuclear extract.

Chapter Four

3’ UTR of AcMNPV gp37 reduces polyhedron production at the polyhedrin locus of AcMNPV

Introduction

Baculoviruses are large double-stranded DNA viruses that have been used both as insect pest control agents and protein expression vectors (Doerfler, 1986; Miller, 1988). The type species of the family Baculoviridae, Autographa californica multicapsid nucleopolyhedrovirus (AcMNPV), has been used as a model system for baculovirus genetic studies since a number of insect cell lines are highly permissive for AcMNPV infection (Doerfler, 1986). The genome of AcMNPV was the first sequenced genome among baculoviruses, revealing that AcMNPV has a highly compressed genome of 134 kb (Ayres et al., 1994).

During a productive infection cycle of AcMNPV in permissive insect cells, AcMNPV enters the cells probably through receptor binding and endocytosis. Once inside the cytoplasm, the endosomes that contain the virions travel to the nucleus via actin filaments (Charlton and Volkman, 1993). Membrane fusion with the endosome releases the virions just outside the nucleus, and the nucleocapsids then enter the nucleus through the nuclear pore complex. In the nucleus, the nucleocapsids uncoat to release the viral DNA genome for viral gene transcription. The viral genes are expressed in four phases: immediate-early, delayed early, late, and very late. During the very late phase of AcMNPV infection, the polyhedrin gene is abundantly expressed to form large polyhedrin protein complexes called polyhedra, which occlude or embed virions (Westwood et al., 1993). Production of polyhedra by baculoviruses to protect the occluded virions from UV inactivation is believed to be an evolutionary adaption of the baculovirus to spread in the insect population in the natural environment (Miller, 1997).

The gp37 gene of AcMNPV, which is a homologue of Choristoneura fumiferana defective nucleopolyhedrovirus (CfDEFNPV) spindlin gene, is a late gene that is not highly expressed. The spindlin gene is highly expressed by CfDEFNPV in permissive cells. Unique spindle-shaped inclusions are observed in the cytosol of CfDEFNPV- infected cells, which are also present in other NPV infections, like Choristoneura murinana NPV, Orgyia pseudotsugata MNPV, Cadra cautella NPV, and Archips cerasivoranus NPV (Adams and Wilcox, 1968; Li et al., 2000). The proteins encoded by

the gp37 homologues from AcMNPV and CfDEFNPV share 59.09% amino acid identity (Li et al., 2000). However, no spindle-shaped inclusion is observed in AcMNPV- infected cells, which suggests that either gp37 (spindlin) is not expressed abundantly by AcMNPV, or does not crystallize to form discernable spindles. The function of AcMNPV gp37 is unclear but it was reported not to be essential for viral replication (Cheng et al., 2001). Transcriptional studies of AcMNPV gp37 showed that multiple species of transcripts were produced with different lengths of the 3’ untranslated region (3’ UTR) during AcMNPV infection in permissive Sf21 cells (Wu and Miller, 1989).

The 3’ UTRs of mRNAs can have multiple functions, e. g. directing formation of the poly (A) tail, regulating mRNA stability and having binding sites for microRNA (Wahle and Kuhn, 1997; Bartel, 2009). The proper formation of the 3’ end is important for mRNA gene expression. Cleavage and polyadenylation signals, which are usually located at the end of 3’ UTRs, are required for the maturation of most mRNA transcripts (Proudfoot, 1991; Colgan and Manley, 1997). It was reported that alternative polyadenylation, the use of more than one polyadenylation signal, often can eliminate large parts of the 3’ UTR and protect the mRNA transcripts from the stronger regulatory potential of longer 3’ UTRs (Mayr and Bartel, 2009). Alternative polyadenylation events could result in RNA transcripts with very subtle changes, such as having different 3’ UTR lengths, but could also lead to RNA transcripts with dramatic difference, such as producing different proteins with different domains. In addition to changing the sequences of RNA transcripts and proteins, alternative polyadenylation events affect the localization, stability, and transport of transcripts (Lutz, 2008). Since viral genomes are usually very compact and used efficiently, alternative polyadenylation has been found in many viruses, such as papillomavirus type 16 (HPV 16) and avian retroviruses (Graham, 2008; McNally, 2008; Schwartz, 2008).

Three different lengths of transcripts of AcMNPV gp37 were detected post infection (Wu and Miller, 1989). The ends of the transcripts contain three instances of three classic polyadenylation signal (AAUAAA) downstream of the stop codon of gp37. All three different lengths transcripts appeared starting from 6 h post infection. The

amount of the longest transcript increased with time post-infection while the amounts of the two shorter transcripts decreased. The switch from shorter transcripts to the longest transcript after 6 h p.i. leads to different 3’ UTRs of gp37 (Wu and Miller, 1989). The abundance of the shortest forms in the earlier stages of transcription seems to be a common phenomenon for baculovirus mRNAs with alternative polyadenylation sites (Westwood et al., 1993). The spindle gene of CfDEFNPV has only one polyadenylation signal downstream of the stop codon. Despite the high homology between AcMNPV GP37 and CfDEFNPV spindlin, their expression levels are quite different (Li et al., 2000; unpublished data). Considering important role of alternative polyadenylation in gene expression regulation and the difference between the 3’ downstream sequences of gp37 and the spindlin gene, we hypothesize that the 3’ UTR of gp37 is responsible for the reduced gene expression as compared with that of the spindlin gene of CfDEFNPV.

In order to test the role of the 3’ UTR of gp37, two sets of viruses were constructed. One set uses the polyhedrin gene as the reporter gene followed by the polyhedrin 3’ UTR or the gp37 3’ UTR; the other set uses the green fluorescent protein (GFP) gene as the reporter gene followed by the polyhedrin 3’ UTR or the gp37 3’ UTR.

Materials and Methods

Cell lines and viruses. The insect cell lines Sf21 and Hi5 were maintained at 27°C in Grace’s medium supplemented with 0.33% yeastolate, 0.33% lactalbumin hydrolysate and 10% fetal bovine serum. AcMNPV DNA was used as the template in PCR for cloning polyhedrin ORF and 3’ UTR as well as gp37 ORF and 3’ UTR.

Comparison of 3’ downstream sequences of AcMNPV gp37, polyhedrin gene and CfDEFNPV spindlin gene. Sequences downstream from the stop codon of the three genes were retrieved from GenBank. The DNAstar program was used to search for AATAAA in the 3’ downstream sequences. For the AcMNPV gp37 and the polyhedrin genes, since the transcripts have been mapped, the search was performed within the 3’ untranslated regions (Wu and Miller, 1989; Westwood et al., 1993). For the CfDEFNPV spindlin gene, since its transcript has not been mapped, the search was performed within 1000 nt downstream of the stop codon. The GT-rich regions downstream of the AATAAA signal sequences were visually examined and compared.

Construction of viruses. Specific primers were designed to amplify the fragments including the polyhedrin ORF and 3’ downstream sequences (DS) using AcMNPV DNA as the template and verified by agarose gel electrophoresis. The amplified PCR product was cut out from the gel and purified by the glassmilk method (Vogelstein and Gillespie, 1979), then cloned into pGEM-T Easy vector (Promega) and confirmed by DNA sequencing. The insert was subcloned into pFastBac1 vector (Invitrogen) to make pFastBacpolUTR. The specific primers including REN sites were designed to amplify the polyhedrin ORF and gp37 DS separately using AcMNPV DNA as the template and verified by agarose gel electrophoresis. The amplified PCR products were gel extracted using the glassmilk method and then cloned to pGEM-T Easy vector separately to make pGEMT-polh and pGEMT-gp37UTR. Both inserts were verified by DNA sequencing. The fragments of the polyhedrin ORF and gp37 3’DS were retrieved separately using EcoRI and XbaI as well as XbaI and XhoI from the plasmids of pGEMT-poly and pGEMT-gp37UTR. The two fragments were ligated to pFastBac1 in a three-piece DNA

ligation reaction with T4 DNA ligase to produce pFastBacgp37UTR. pFastBacpolUTR and pFastBacgp37UTR were confirmed by DNA sequencing and REN analysis.

Specific primers including REN sites were designed to amplify the GFP ORF using pBlueGFP (Cheng et al., 2001) as the template and to amplify the polyhedrin 3’ DS using AcMNPV DNA as the template and verified by agarose gel electrophoresis. The amplified PCR products were gel-extracted by the glassmilk method (Vogelstein and Gillespie, 1979), then cloned to the pGEM-T Easy vector to construct pGEMT-gfp, which has the GFP ORF and pGEMT –polhUTR, which has the polyhedrin 3’ UTR. The GFP ORF and polyhedrin 3’ UTR were retrieved from the plasmids pGEMT-gfp and pGEMT-polhUTR; then these two fragments were ligated to pFastBac1 in a three-piece DNA ligation reaction to construct pFastBacGFPpolhUTR. The GFP ORF and gp37 3’UTR fragments were retrieved from the plasmids pGEMT-gfp and pGEMT-gp37UTR; then the two fragments were ligated to pFastBac1 in a three-piece DNA ligation reaction to produce plasmid pFastBacGFPgp37UTR.

The plasmids pFastBacpolhUTR and pFastBacgp37UTR as well as pFastBacgfp- polhUTR and pFastBacgfp-gp37UTR were used to transform competent DH10Bac E. coli cells (Invitrogen) and then plated on triple antibiotic plates (50 µg/ml kanamycin, 7 µg/ml gentamicin, 10 µg/ml tetracycline) with 100 µg/ml X-gal and 40 µg /ml IPTG. White colonies from each transformation were picked and confirmed by PCR with specific primers. DNAs were extracted from the bacteria which contain the right insertion in the bacmid and then used to transfect Sf21 cells to produce the viruses AcBac- polhUTR, which has the polyhedrin ORF and polyhUTR in the bacmid genome under the control of the polyhedrin promoter; AcBac-gp37UTR, which has the polyhedrin ORF and gp37UTR in the bacmid under the polyhedrin promoter; AcBac-gfppolyhUTR, which has the GFP ORF and polyhUTR in the bacmid under the polyhedrin promoter; and AcBac- gfpgp37UTR, which has the GFP ORF and gp37UTR in the bacmid under the polyhedrin promoter. The newly constructed viruses were confirmed by polyhedron formation or by showing green fluorescence post-infection, and then propagated and titered (O'Reilly et al., 1992).

Quantitative analysis of polyhedron production and size measurement of polyhedra. Sf21 cells or Hi5 cells were infected by AcpolUTR or Acgp37UTR at an MOI of 5 p.f.u./cell in 25 cm2 TC flasks with 3x106 cells/flask (O'Reilly et al., 1992). The infected cells were incubated at 27°C for 96 h. Images were recorded from the infected cells using a SPOT Insight digital camera attached to a Nikon Eclipse TE2000-U inverted microscope. After image capture, media was removed and 1 ml of 0.1% SDS was added to each of the TC flasks; then the TC flasks were rocked at 27C for 1 h to lyse the cells for polyhedra releasing. The cell lysate of each TC flask was transferred to a 1.5-ml centrifuge tube. An aliquot (10 µl) was taken from each sample for polyhedron yield estimation with a hemocytometer.

To compare the size difference of polyhedra, 50 polyhedra were randomly selected and measured from Sf21 cells infected with either AcpolUTR or Acgp37UTR. Similar measurements were also performed on Hi5 cells infected with the two viruses. Diameters were compared between polyhedra from Sf21 cells infected with AcpolUTR and Acgp37UTR and Hi5 cells infected with AcpolUTR and AcpolUTR.

Protein yield assay. Polyhedrin protein yields were estimated by using a Bio-Rad protein assay kit. A standard curve was first constructed by using BSA from 0.2 mg/ml to

1.4 mg/ml in Na2CO3 (0.1 M, pH 10.5). The 5 × dye reagent was diluted to 1 ×. The diluted dye reagent was added to each BSA dilution and the OD595 was measured with a spectrophotometer. Three readings were taken for each sample. Fifty μl of purified polyhedra from AcpolUTR or Acgp37UTR infected Sf21 cells were mixed with 50 µl

Na2CO3. The diluted dye reagent was added to each sample and the absorbance of OD595 was recorded. The standard curve was used to estimate the protein yields from polyhedra produced in Sf21 cells infected with either AcpolUTR or Acgp37UTR. Polyhedrin yields from polyhedra of AcpolUTR and Acgp37UTR were statistically analyzed.

GFP expression yields were estimated using a spectrofluorometer. Sf21 cells were infected by AcgfppolUTR or Acgfpgp37UTR at an MOI of 5 p.f.u./cell in 25 cm2 TC flasks with 3×106 cells/flask. The infected cells were incubated at 27C for 96 h. After incubation, the infected cells were dislodged from the TC flasks by jet flashing with 86

media in the flasks using a Pasteur pipette and the resuspended cells were centrifuged at 500x g for 5 min. The supernatant was discarded and the residue media were blotted from the cell pellets. The cells were lysed with 500 μl 0.1% SDS in each sample. The lysate containing the EGFP was measured for fluorescence by a Shimadzu RF-5301PC spectrofluorometer (Ex 488 nm, Em 507 nm). The reading was taken in triplicate. The emission values were used to quantify the GFP expression between the two constructed viruses for statistical analysis using T-test in Excel (Microsoft).

Results 3’ UTR sequence analysis. Analysis of the 3’ downstream sequences (DS) of AcMNPV gp37 and polyhedrin gene, and the spindlin gene from CfDEFNPV revealed different polyadenylation signal sequences. The 3’ DS of AcMNPV gp37 has three polyadenylation sites (Fig. 4-1A) which include the polyadenylation signal AATAAA and the downstream GT-rich sequences. The 3’ DS of the homologue of gp37, the spindlin gene of CfDEFNPV has only one polyadenylation site (Fig. 4-1C). The 3’ DS of AcMNPV polyhedrin gene has two polyadenylation signals (AATAAA) but only the second one has the downstream GT-rich sequences (Fig. 4-1B).

Construction of two sets of viruses. In order to understand the regulatory effect of the alternative polyadenylation signals in the 3’ DS of AcMNPV gp37, the 3’ DS of gp37 was fused downstream of two reporter genes, the polh gene and the gfp gene. The 3’ DS of the polh gene was used as a strong polyadenylation signal with high protein expression to compare with the gp37 3’ DS. The reporter genes and the 3’ DSs were under the control of a strong polh late gene promoter. The AcpolUTR had the polyhedrin ORF followed by 3’ DS of polyhedrin whereas the Acgp37UTR had polyhedrin ORF followed by the 3’ DS of gp37 at the polyhedrin locus. Since these two viruses were constructed using the AcMNPV bacmid, the endogenous polyhedrin gene is not present. The AcgfppolUTR and Acgfpgp37UTR viruses had the gfp gene followed by the 3’ DS of polyhedrin or the 3’ DS of gp37 at the polyhedrin locus, respectively (Fig. 4-2).

Polyhedron production and size comparisons between AcpolUTR and Acgp37UTR. The 3’ DS of gp37 increased polyhedron sizes but reduced polyhedron yields. In the polyhedrin virus set, the polyhedrin gene was used as the reporter gene. Polyhedra are formed in the nuclei of infected cells by baculovirus at the late stage of infection, which give a good tracking of the dynamics of polyhedrin expression. The Sf21cells infected with Acgp37UTR showed much larger but fewer polyhedra in the nuclei of infected cells compared to AcpolUTR (Fig. 4-3A). When the sizes of polyhedra formed by AcpolUTR and Acgp37UTR in Sf21 cells were measured and compared, the polyhedron size of

Figure 4-1. 3’ downstream sequence analysis of AcMNPV gp37 (A), polyhedrin (B) genes and the spindling gene from CfDEFNPV (C). All the downstream sequences start with the stop codon and all the AATAAA sequences are bolded and the GT rich regions downstream of the AATAAA are underlined.

A. AcMNPV gp37 3’ downstream sequence: taataaaacaaacaaaattttaattacatattatatttagcaagaagtaataacagatc tctacaatcaatacgaacaatttcaaacatgtcttcaaattcgtagtctgctatgctaa aattgttaattaaaaaatattcttccatattttcacacctgtcataattatgtaataaa aatagtttaaagccatcttgatctcgtttggatatttcgtcatcgtcaaaaagatattg caaagcaaactgtacttctttggcgtaaggatttaacagtccatttttgcagactacgc tgtctacgagcgtattgagactttccgcagttaaagtgttgcaactgattgccggtttc agacgagtttcacaaaaccgcttgatttcatcggctacttgattttcaaaatccatttg agacattaaatatttgtagtattttgcttctccgtgcagcccattcaagtgaagataga tcataaatttatcgtacatgttgctctaaaattaataaaacactgattgcatgcaatca aatgcgctctatttatattatcattcaatgaacatgtcgagcaaatgttctggaaacag aaactcagagggcaaagcgcaaatattcatgattgtacgttcgtcgtttttgtaacaaa catagttatttttaataaacacgctctgtaacacaaccgtgttatgatatatgtcggcg ccacacacttttaccttattgtgagaagatatgtaactatttaaactggtggtggctcg tgaagacacagtttcaatatctgtatttaaagtctg

B. AcMNPV the polyhedrin gene 3’ downstream sequence: taaaacacgatacattgttattagtacatttattaagcgctagattctgtgcgttgttg atttacagacaattgttgtacgtattttaataattcattaaatttataatctttagggt ggtatgttagagcgaaaatcaaatgattttcagcgtctttatatctgaatttaaatatt aaatcctcaatagatttgtaaaataggtttcgattagtttcaaacaagggttgtttttc cgaaccgatggctggactatctaatggattttcgctcaacgccacaaaacttgccaaat cttgtagcagcaatctagctttgtcgatattcgtttgtgttttgttttgtaataaaggt tcgacgtcgttcaaaatattatgcgcttttgtatttctttcatcactgtcgttagtgta caattgactcgacgtaaacacgttaaataaagcttggacatatttaacatcgggcgtgt tagctttattaggccgattatcgtcgtcgtcccaa

C. CfDEFNPV the spindlin gene 3’ downstream sequence: tgagcgcgtcattgtgttgtttgcggtcgttgaccaacaacgtatgtttaaataaataa aatgactcaagtaaaaattggtcaatttaaattcggcgaagacacgttcacgctgcggtac gtgttggagcgc

Figure 4-2. A schematic presentation of four constructed viruses: AcpolUTR and Acgp37UTR which have polyhedrin ORF under polyhedrin promoter with polyhedrin UTR or gp37 UTR at the polyhedrin locus as well as AcgfppolUTR and Acgfpgp37UTR which have GFP ORF under polyhedrin promoter with polyhedrin UTR or gp37 UTR at the polyhedrin locus.

Polyhedrin ORF Polyhedrin UTR

GFP ORF GP37 UTR

AcpolUTR

Acgp37UTR

AcgfppolUTR

Acgfpgp37UTR

Figure 4-3. Insect cell lines were infected by AcpolUTR or Acgp37UTR.

A. Sf21 cells were infected by AcpolUTR (a) or Acgp37UTR (b) at an MOI of 5 p.f.u./cell and observed at 4 d p. i. . White arrowheads point to cells with polyhedra in the nucleus. Bar markers represent 10 μm.

B. Comparison of the total polyhedra production between samples from Sf21 cells infected by AcpolUTR or Acgp37UTR at an MOI of 5 p.f.u./cell at 4 d p. i. . At the time of total polyhedra harvest, medium was removed and 0.1% SDS was added to each flask and rocked at 27C for 1 h to lyse the cells for polyhedra release. An aliquot (10 µl) of the cell lysate was taken from each sample for polyhedron yield estimation with a hemocytometer. * indicates significant difference to the other tested sample (p<0.05).

C. Comparison at 4 d p. i. of polyhedra size between samples from Sf21 cells infected by AcpolUTR or Acgp37UTR at an MOI of 5 p.f.u./cell. 50 polyhedra from each sample were measured. * indicates significant difference to the other tested sample (p<0.05).

D. Hi5 cells were infected by AcpolUTR (a) and Acgp37UTR (b) at an MOI of 5 p.f.u./cell and observed at 4 d p. i. . White arrowheads point to cells with polyhedra in the nucleus. Bar markers represent 10 μm.

E. Comparison of the total polyhedra production between samples from Hi5 cells infected by AcpolUTR or Acgp37UTR at an MOI of 5 p.f.u./cell at 4 d p. i.. At the time of harvest of total polyhedra, media were removed and 0.1% SDS was added to each flask and rocked at 27C for 1 h to lyse the cells for polyhedra release. An aliquot (10 µl) of the cell lysate was taken from each sample for polyhedron yield estimation with a hemocytometer. * indicates significant difference to the other tested sample (p<0.05).

F. Comparison of the polyhedra size between samples from Hi5 cells infected by AcpolUTR or Acgp37UTR at an MOI of 5 p.f.u./cell at 4 d p. i. . 50 polyhedra from each sample were measured. * indicates significant difference to the other tested sample (p<0.05).

A. a b

B. *

AcpolUTR Acgp37UTR

D. a b

AcpolUTR Acgp37UTR F.

AcpolUTR Acgp37UTR

Acgp37UTR is 1.63 times larger in diameter than that of AcpolUTR (Fig. 4-3C). When the total polyhedron yields of Acgp37UTR and AcpolUTR were compared, the total polyhedra produced from Acgp37UTR infected Sf21 cells showed a 2-fold reduction comparing to AcpolUTR (Fig. 4-3B). Immunoblotting assay using an anti-VP39 (virion capsid protein) antibody was used to determine whether these large polyhedra contain virions (Wang et al., 2009).The immunoblot showed that the large polyhedra produced from Acgp37UTR still have virion occlusion similar to the polyhedra from the wild type AcMNPV infection (data not shown). To further confirm the reduction of polyhedra yields and increase of polyhedron size from Acgp37UTR infected cells, Hi5 cells were infected by AcpolUTR or Acgp37UTR for polyhedron production comparison in terms of polyhedron size and yield. The same phenomena were observed in Hi5 cells infection as Sf21 cells infection (Fig. 4-3D-F). The infection of Hi5 cells by Acgp37UTR produced 1.44-fold larger polyhedra in diameter than AcpolUTR (Fig. 4-3F) and the reduction of total polyhedron production is about 6-fold (Fig. 4-3E). Therefore, infections of both Sf21 cells and Hi5 cells by the two viruses confirmed that the gp37 3’ DS reduces polyhedron yield but increases polyhedron size compared to the polyhedrin 3’ DS.

Comparison of total polyhedrin protein production. Although polyhedron yields were reduced when the gp37 DS replaced the polyhedrin 3’ DS, the total protein yield was not affected. This conclusion is drawn from two independent assays. The Bio-Rad protein assay was used to compare the total protein expression levels of polyhedrin from Sf21 cells infected with Acgp37UTR and AcpolUTR. It was found that the total polyhedrin amounts from Sf21 cells infected with Acgp37UTR and Sf21 cells infected with AcpolUTR were at the same level (Fig. 4-4). This result is supported by comparing GFP expression levels in Sf21 cells infected by AcgfppolUTR and Acgfpgp37UTR. GFP expression by the two viral constructs in Sf21 cells showed no difference. When Sf21 cells infected with either AcgfppolUTR or Acgfpgp37UTR were used for GPF expression comparison at day 4 p. i., similar GFP expression between the two viruses was obtained (Fig. 4-5). This result further confirms that the total protein production levels either by polyhedrin or GFP were not affected by the 3’ DS of gp37.

Figure 4-4. Protein yield assay of polyhedrin in Sf21 cells infected by AcpolUTR or Acgp37UTR. A standard curve was constructed to estimate polyhedrin yields from different viral constructs using 0.2 mg/ml to 1.4 mg/ml BSA in Na2CO3. The diluted dye reagent was added to each sample and the absorbance value was recorded at OD595. The standard curve was used to estimate the protein concentration for each sample.

AcpolUTR Acgp37UTR

Figure 4-5. GFP production comparison between AcgfppolUTR infected Sf21 cells and Acgfpgp37UTR infected Sf21 cells at 4 d p. i. . GFP expression yields were estimated using a fluorescence spectrometer. At the time of measurement, the infected cells were dislodged from the flasks and the resuspended cells were centrifuged at 500 g for 5 minutes. The cell pellets were lysed with addition of 500 μl 0.1% SDS to each sample. The lysate containing the GFP was measured for fluorescence by a Shimadzu RF-5301PC spectrofluorometer (Ex 488 nm, Em 507 nm). The reading was taken in triplicates.

100

AcgfppolUTR Acgfpgp37UTR

101

Discussion

mRNA processing plays a major regulatory role in eukaryotic cells. Alternative polyadenylation has been recently identified as an important contributor to gene regulation (Lutz, 2008). Alternative polyadenylation is known to regulate gene expression in various ways (Lutz, 2008). Other than producing different proteins, alternative polyadenylation can be used to produce the same coding region of an mRNA but with different 3’ UTRs (Mayr and Bartel, 2009). It has only been recently appreciated that nearly all genes have additional polyadenylation signals in their 3’ UTRs, and more than half of the human genes are alternatively polyadenylated (Tian et al., 2005). Many alternative polyadenylation signals are evolutionarily conserved (Tian et al., 2005). The use of alternative polyadenylation signals can protect the transcript from the stronger regulatory potential of longer 3’ UTRs. For example, the shorter 3’ UTR can lose microRNA (miRNA) complementary sites (Mayr and Bartel, 2009). It was reported that alternative polyadenylation can activate oncogenes in cancer cells by shortening the 3’ UTRs of oncogenes; these shorter isoforms of oncogenes have higher stability and produce more proteins in part through the loss of miRNA-mediated repression (Mayr and Bartel, 2009). The alternative polyadenylated isoforms of the same gene can also have different translation efficiency (Stark et al., 2005).

In this study, it was demonstrated that 3’ DS of the spindle gene from CfDEFNPV has one putative polyadenylation site whereas the 3’ DS of gp37 has multiple polyadenylation sites (Fig. 4-1). It was also reported that the AcMNPV gp37 transcripts have different lengths and the different 3’ end of the transcripts match well with the three polyadenylation signals in the 3’ UTR of gp37. The percentage of the largest transcript decreases during later times post-infection while the percentage of the smaller transcripts increase (Wu and Miller, 1989).

In order to test the regulatory function of the gp37gene 3’ DS, the 3’ DS of polyhedrin from AcMNPV, which is similar to 3’ DS of the spindlin gene and has a high expression level, was used for comparison. When the polyhedrin gene from AcMNPV was used as the marker gene, the virus Acgp37UTR which has gp37 3’ DS makes fewer, 102

larger polyhedra in the nuclei of both infected Sf21 and Hi5 cells than does AcpolUTR, which has the polyhedrin 3’ DS. In Sf21 cells the reduction is 2-fold, and in Hi5 cells the reduction is 6-fold, both of which are significant. Surprisingly the total polyhedrin protein levels from AcpolUTR- and Acgp37UTR-infected Sf21 cells are the same. This may reflect the size difference between polyhedra in AcpolUTR and Acgp37UTR infected cells. When GFP was used as the reporter gene instead of polyhedrin in order to confirm that there is no difference at the final protein amount, the fluorescence of the cells infected with virus were very similar, which confirmed the observation that there is no difference in the final protein amount. The different size and number of polyhedra from the two virus infections indicates that there is a regulatory effect of the gp37 3’ UTR since the coding regions of the two viruses are identical, the only difference being in the 3’ DS of the reporter gene, polyhedrin. But the regulation is not at the level of total protein production.

Polyhedra are crystals and the formation of the crystals has two components, nucleation and growth. High protein concentration will lead to the supersaturating of the solution and drive the nucleation. The growth is favored by then reducing the protein concentration since the reduced protein concentration can prevent further nucleation from competing with the growth of established nuclei (Schmit and Dill, 2010). A possible explanation for the fewer, larger polyhedra from Acgp37UTR-infected cells is that the 3’ UTR of gp37 changes the protein synthesis kinetics. During the early stage of the infection, AcpolUTR may have higher polyhedrin expression than Acgp37UTR, making it possible for AcpolUTR to produce more polyhedral crystal nuclei. At later stages of the infection, Acgp37UTR catches up in polyhedrin expression but can only build on the established polyhedra crystal nuclei. It will be interesting to test whether the different lengths of gp37 transcripts have different RNA stability and translation efficiency.

How the alternative polyadenylation occurs in gp37 transcripts and regulates gene expression is unknown. It was reported that the efficiency of the usage of different polyadenylation sites is controlled by the interaction of the regulatory cis-acting RNA elements and the trans-acting protein factors. Among the factors, the cleavage and

103

stimulation factor CstF-64 binds to the GU-rich region downstream of the AAUAAA signal and then recruits other factors to assemble the polyadenylation complex. One possible explanation for the regulatory effect of gp37 3’ UTR is that the downstream GU- rich regions of each AAUAAA bind to CstF-64 weakly, allowing the alternative polyadenylation to take place. It was reported the sequences in the GU-rich region controls the binding (Perez Canadillas and Varani, 2003), so by controlling the sequences in the GU-rich region downstream of AAUAAA, AcMNPV can regulate the gp37 expression without requirement of additional regulatory proteins.

It is not rare for baculoviruses to have alternative polyadenylation sites in their genomes. It was reported that multiple genes have different 3’ end sites and the proportion of the various forms of these mRNAs may vary temporally, although often the shortest forms are more common in the earlier stages of transcription (Westwood et al., 1993). It has been well known that viruses have very compact genomes and use their genetic information very efficiently, so alternative polyadenylation should be commonly employed in many viral genomes. The results in this study show that alternative polyadenylation regulates the polyhedra formation without changing the total amount of the protein expressed, suggesting a possible impact on changing protein expression kinetics. This study can serve as an example of the regulatory effect of alternative polyadenylation in baculovirus and other viruses.

104

General Conclusions

Insect viruses are important in controlling insect pests in agriculture and forestry and they are good candidates for commercial production for biological control agents due to the easy production methods and stability during storage and after application in the field. Baculoviruses are well-studied insect viruses and they have been made into commercial products to control insect pests in agriculture and forestry. Ascoviruses are newly identified insect virus transmitted mechanically by endoparasitic wasps during oviposition. However the per os infectivity of ascoviruses is low which makes them difficult to use as a biological control agent with present knowledge. Never the less there are interesting features about these new viruses, such as cytopathology, gene transcription, genome organization and evolution. The findings regarding ascoviruses in this dissertation may eventually help in generating more efficient biological pesticides with wider host ranges. Research conducted in this dissertation investigated the genome organization, host range and gene transcription strategies of ascovirus as well as its comparison with the well-studied baculovirus.

Insect-specific pathogens based biopesticides are promising alternatives to chemical pesticides. Among the biopesticides, the most widely used are insect-specific toxins from Bacillus thuringiensis (Bt) (Szewczyk, et al., 2006). Bt toxins can bind to the midgut epithelium cells of the insects, disrupt the cytoplasmic membrane and eventually lyse the cells. After ingestion of the Bt toxins, insects quickly stop eating and die (Roh et al., 2007). Although different Bt toxins target different insect species, mixing different toxins or genetically modified Bt toxins can expand the host range (Tabashnik, 1992; Xue et al., 2005). There has been increasing development of resistance to Bt toxins among insects, and consequently baculoviruses are beginning to get more attention to as potential biopesticides (Szewczyk et al., 2006; Roh et al., 2007). Baculoviruses have the advantage of being effective and safe to the environment as well as to humans. Baculoviruses have been commercially used in the field including the two commercial products available in the USA and Europe: SPOD-X TM which is used to control insects on vegetable crops or flowers in greenhouses, and SpodopterinTM used to protect cotton,

105

corn and tomatoes (Szewczyk et al., 2006). As biopesticides, baculoviruses have disadvantages including narrow host range, slow killing action and high cost during production. Unlike Bt toxins that kill insects very quickly, baculoviruses must infect the insects to kill them. Therefore, it can take several days or weeks to kill an insect and during this time the insect will continue to feed (Thiem, 1997). Researchers tried to reduce the time required for baculovirus insecticides to stop insect feeding by deleting a viral gene (egt) which inhibits insect molting. The deletion results in 30% faster killing of larvae and significant reduction in food consumption (O’Reilly and Miller, 1991; Sun et al., 2004). In contrast to baculovirus, larvae infected by ascovirus decrease their feeding rate within 24h of infection. As the infection progresses, the feeding rate remains slow, and as a result, larvae gain very little weight and cannot progress in development (Federici et al., 2009). Although the mechanism responsible for this slow feeding rate is not identified, as more ascoviruses genomes are sequenced and analyzed, especially the smallest genome SfAV-1d reported in this study, it may help to determine the mechanism behind the slow feeding of larvae after ascovirus infection and improve baculovirus based biopesticides to inhibit insect consumption of crops.

Since baculovirus usually has narrow host range, research has focused on how host range is controlled and making recombinant baculoviruses to achieve broader host range (Thiem, 1997; Szewczyk et al., 2006). Advances in the understanding of baculovirus replication and the identification of genes that affect host range made it possible to construct recombinant baculoviruses for specific pest insects. In addition, the killing activity of baculoviruses may be improved by genetic modifications of the baculovirus genome with genes of other natural pathogens. The newly identified ascovirus may be one of these natural pathogens that can be used in constructing broader host range recombinant baculoviruses. In this study, two ascoviruses SfAV-1a and SfAV- 1d show difference in host range although their genomes are very closely related (Fig. 1- 1). SfAV-1a has a broader host range and it had mortality in S. exigua, S. frugerpida, T. ni as well as in P. includens while SfAV-1d has a narrow host range, it can only develop mortality in S. frugerpida (Table 1-2). When comparing the two genomes of SfAV-1a and SfAV-1d, SfAV-1d is very similar to SfAV-1a with 7 HindIII and 9 BamHI co- 106

migrating fragments (Fig. 1-1). Sequencing of the SfAV-1d HindIII fragment library end sequences showed that 4 of the 7 co-migrating HindIII REN fragments with SfAV-1a displayed 99% identity with only a few single nucleotide polymorphisms (data not shown). Even though SfAV-1d genome has not been completely assembled, the alignment of SfAV-1d shotgun library sequences to SfAV-1a genome showed that the two genomes have very similar genome organization except there is a major 14 kb deletion on SfAV-1d genome at the IR region of SfAV-1a genome (Fig. 1-2). The whole IR region is missing in the SfAV-1d genome (Fig. 1-4; Table 1-1). However, whether the 14 kb deletion is responsible for the host range difference is unknown since there is no genetic tool available to insert the fragment back to the SfAV-1d genome to restore the host range difference. In baculoviruses, the homologous repeats, which are similar to IR in ascoviurses, interspersed throughout the genome were implicated as the replication origins and the different sequences in the repeats had different replication ability (Cochran and Faulkner, 1983; Ayres et al., 1994; Xie et al., 1995; Garcia-Maruniak et al., 1996; Tillman et al., 2004). This replication ability difference may partially explain the host specificity of baculoviruses (Pearson et al., 1992; Leisy and Rohrmann, 1993; Pearson et al., 1993; Lee and Krell, 1994; Ahrens et al., 1995; Leisy et al., 1995; Pearson and Rohrmann, 1995; Xie et al., 1995). Considering the close relationship between ascoviruses and baculoviruses, the host range difference in mortality bioassay between SfAV-1a and SfAV-1d could be due to the inability of SfAV-1d to replicate in some of the hosts because of the missing sequences from SfAV-1a. It is possible that the IR region interacts with certain host proteins in the permissive cell lines allowing SfAV-1a to replicate in certain insects, which made it a possible candidate to insert into baculovirus genomes to increase baculoviruses host range. Our results showed that the deletion region on SfAV-1d and the corresponding IR region on SfAV-1a are variable regions (Fig.1-3; Fig. 1-4). It was suggested the IR region in the SfAV-1a genome is important for the switch from linear to circular configurations of the genome during virus DNA replication (Bideshi et al., 2006), which was supported by the reported variable sequences in the SfAV-1d genome in this study since the variable sequences could come from the virus replication process. However, what specific sequences from this IR region

107

could serve as the origin of the replication is not clear since this region is totally missing in SfAV-1d.

Other than repeat sequences which can affect host range in baculovirus, certain genes were identified to affect host range. Most of the baculovirus genes identified to date that influence host range are associated with late gene expression and DNA replication, such as p143 (helicase), late expression factor-7 and hcf-1 (late expression factor etc. (Maeda et al., 1993; Croizier et al., 1994; Lu and Miller, 1995; Chen and Thiem, 1997). Since ascovirus is a newly identified insect virus, understanding the transcription strategies for ascovirus including the late gene transcription may help to identify the genes which can affect host range. Three host RNA polymerase subunits homologous were identified in all the ascoviruses genomes available. While using the nuclear extract from uninfected cells, the early promoter of ascovirus SfAV-1d can be recognized but the late promoter of SfAV-1d cannot (Fig. 3-3). This means the viral RNA polymerase or other viral factors were required for late gene transcription. Since in baculovirus, the genes that affect host range are mainly related to late gene expression, further study about the viral RNA polymerase or factors in ascovirus involved in the late gene transcription may help identify the genes influencing host range in ascovirus genomes.

In baculovirus, the late gene polyhedrin is transcribed by viral RNA polymerase and has high expression. The expressed protein, polyhedrin, can form polyhedra structure which can protect the occluded virions from UV inactivation, so baculovirus pesticides can stay safe and effective in the field for years (Miller, 1997). Another late gene of baculovirus, gp37 does not always have high expression. The function of gp37 is not quiet understood, but it was suggested that GP37 is a homologue of fusolin protein from entomopoxvirus which was reported to be able to enhance baculovirus infection (Xu and Hukuhara, 1992; Li et al., 2000). During entomopoxvirus infection, the fusolin protein can form spindle shaped inclusion bodies in the cytoplasm of the infected cell (Dall et al., 1993). GP37 has different levels of expression in different baculoviruses infection. In CfDEFNPV infected cells, the expression of gp37 is high and the protein can form

108

spindle shaped inclusion bodies in the cytoplasm (Li et al., 2000). But in AcMNPV, the expression of gp37 is much less compared to gp37 in CfDEFNPV and no spindles can be observed in the cytoplasm of the infected cells by AcMNPV (Li et al., 2000 and unpublished data). Comparing the 3’ UTRs of gp37 from CfDEFNPV and AcMNPV shows that AcMNPV 3’ UTR has multiple polyadenylation sites (Fig. 4-1) and the three different polyadenylation sites were used differently at different time post-infection (Wu and Miller, 1989). This study showed that when the 3’ downstream sequence of AcMNPV gp37 was put downstream of AcMNPV polyhedrin gene, it changes the crystallization process of polyhedra (Fig. 4-3 A and D). Since it was reported that the fusolin gene, homologue of gp37, from entomopoxvirus can enhance baculovirus replication, it is worth finding out the function of gp37 and whether the 3’ UTR of gp37 has any effect on the crystallization process to ensure GP37 function properly. The research about AcMNPV gp37 and its 3’ UTR may help GP37 function determination and may allow investigators to make genetic modifications that can enhance the baculovirus replication and improve the efficiency of the baculovirus pesticide.

In conclusion, as a newly isolated insect virus, it is too early to judge the economic value of ascoviruses. Since ascovirus has very low per os infectivity and depends on endoparasitic wasps for its transmission, it is hard to put it into direct use as insecticide with the current knowledge. Considering that baculovirus and ascovirus are closely related and share similar transcription pattern, the knowledge we have learned about ascovirus can definitely help construct recombinant baculovirus to improve host range. Although the genes involved in preventing the insects from feeding are not identified, the available genomes sequences, especially the shortest SfAV-1d genome, will give more possibilities to identify the genes involved in reduced feeding. The IR region and genes responsible for the stopping the larvae from feeding during ascovirus infection are good candidates to insert into the baculovirus genome in attempt to improve host range and shorten the killing action time of baculovirus pesticides. Hopefully with the research in this study about genome organization, host range and transcription strategies of ascovirus as well as baculovirus can help build broad host range, fast acting biopesticides to solve the increasing development of resistance to Bt toxins. 109

References

Adams, J.R. and Wilcox, T.A. 1968. Histopathology of the almond moth, Cadra cautella, infected with a nuclear-prolyhedrosis virus. J. Invertebr. Pathol. 12: 269-274.

Ahrens, C.H., Pearson, M.N. and Rohrmann, G.F. 1995. Identification and characterization of a second putative origin of DNA replication in a baculovirus of Orgyia pseudotsugata. Virology 207: 572-576.

Asgari, S. 2006. Replication of Heliothis virescens ascovirus in insect cell lines. Arch. Virol. 151: 1689-1699.

Asgari, S. 2007. A caspase-like gene from Heliothis virescens ascovirus (HvAV-3e) is not involved in apoptosis but is essential for virus replication. Virus Res. 128: 99-105.

Asgari, S., Davis, J., Wood, D., Wilson, P. and McGrath, A. 2007. Sequence and organization of the Heliothis virescens ascovirus genome. J. Gen. Virol. 88: 1120-1132.

Avison, M.B., ed. 2007. Measuring gene expression. Taylor and Francis Group, New York.

Ayres, M.D., Howard, S.C., Kuzio, J., Lopez-Ferber, M. and Possee, R.D. 1994. The complete DNA sequence of Autographa californica nuclear polyhedrosis virus. Virology 202: 586-605.

Bartel, D.P. 2009. MicroRNAs: target recognition and regulatory functions. Cell 136: 215-233.

Berretta, M.F. and Passarelli, A.L. 2006. Function of Spodoptera exigua nucleopolyhedrovirus late gene expression factors in the insect cell line SF-21. Virology 355: 82-93.

Bhatia, P., Taylor, W.R., Greenberg, A.H. and Wright, J.A. 1994. Comparison of glyceraldehyde-3-phosphate dehydrogenase and 28S-ribosomal RNA gene expression as 110

RNA loading controls for northern blot analysis of cell lines of varying malignant potential. Anal. Biochem. 216: 223-226.

Bideshi, D.K., Demattei, M.V., Rouleux-Bonnin, F., Stasiak, K., Tan, Y., Bigot, S., Bigot, Y. and Federici, B.A. 2006. Genomic sequence of Spodoptera frugiperda Ascovirus 1a, an enveloped, double-stranded DNA insect virus that manipulates apoptosis for viral reproduction. J. Virol. 80: 11791-11805.

Bideshi, D.K., Tan, Y., Bigot, Y. and Federici, B.A. 2005. A viral caspase contributes to modified apoptosis for virus transmission. Genes Dev. 19: 1416-1421.

Bigot, Y., Rabouille, A., Doury, G., Sizaret, P.Y., Delbost, F., Hamelin, M.H. and Periquet, G. 1997a. Biological and molecular features of the relationships between Diadromus pulchellus ascovirus, a parasitoid hymenopteran wasp (Diadromus pulchellus) and its lepidopteran host, Acrolepiopsis assectella. J. Gen. Virol. 78: 1149- 1163.

Bigot, Y., Rabouille, A., Sizaret, P.Y., Hamelin, M.H. and Periquet, G. 1997b. Particle and genomic characteristics of a new member of the Ascoviridae: Diadromus pulchellus ascovirus. J. Gen. Virol. 78: 1139-1147.

Bigot, Y., Renault, S., Nicolas, J., Moundras, C., Demattei, M.V., Samain, S., Bideshi, D.K. and Federici, B.A. 2009. Symbiotic virus at the evolutionary intersection of three types of large DNA viruses; iridoviruses, ascoviruses, and ichnoviruses. PLoS One 4: e6397.

Bigot, Y., Stasiak, K., Rouleux-Bonnin, F. and Federici, B.A. 2000. Characterization of repetitive DNA regions and methylated DNA in ascovirus genomes. J. Gen. Virol. 81: 3073-3082.

Blanquicett, C., Johnson, M.R., Heslin, M. and Diasio, R.B. 2002. Housekeeping gene variability in normal and carcinomatous colorectal and liver tissues: applications in pharmacogenomic gene expression studies. Anal. Biochem. 303: 209-214. 111

Boehmer, P.E. and Lehman, I.R. 1997. Herpes simplex virus DNA replication. Annu. Rev. Biochem. 66: 347-384.

Bond, B.C., Virley, D.J., Cairns, N.J., Hunter, A.J., Moore, G.B., Moss, S.J., Mudge, A.W., Walsh, F.S., Jazin, E. and Preece, P. 2002. The quantification of gene expression in an animal model of brain ischaemia using TaqMan real-time RT-PCR. Brain Res. Mol. Brain Res. 106: 101-116.

Broyles, S.S. and Kremer, M. 2004. An In Vitro Transcription System for Studying Vaccinia Virus Early Genes. In Vaccinia Virus and Poxvirology. Methods and Protocols. (ed. S.N. Isaacs), pp. 135-142.

Bustin, S.A. 2002. Quantification of mRNA using real-time reverse transcription PCR (RT-PCR): trends and problems. J. Mol. Endocrinol. 29: 23-39.

Carner, G.R. and Hudson, J.S. 1983. Histopathology of virus-like particles in Heliothis spp. J. Invertebr. Pathol. 41: 238-249.

Charlton, C.A. and Volkman, L.E. 1993. Penetration of Autographa californica nuclear polyhedrosis virus nucleocapsids into IPLB Sf 21 cells induces actin cable formation. Virology 197: 245-254.

Chen, C-J. and Thiem, S.M. 1997. Differential infectivity of two Autographa californica nucleopolyhedrovirus mutants on three permissive cell lines is the result of lef-7 deletion. Virology 227: 88-95.

Cheng, X., Krell, P. and Arif, B. 2001. P34.8 (GP37) is not essential for baculovirus replication. J. Gen. Virol. 82: 299-305.

Cheng, X.W., Carner, G.R. and Arif, B.M. 2000. A new ascovirus from Spodoptera exigua and its relatedness to the isolate from Spodoptera frugiperda. J. Gen. Virol. 81: 3083-3092.

112

Cheng, X.W., Carner, G.R. and Brown, T.M. 1999. Circular configuration of the genome of ascoviruses. J. Gen. Virol. 80: 1537-1540.

Cheng, X.W., Wan, X.-F., Xue, J. and Moore, R.C. 2007. Ascovirus and its Evolution. Virologica Sinica 22: 137-147.

Cheng, X.W., Wang, L., Carner, G.R. and Arif, B.M. 2005. Characterization of three ascovirus isolates from cotton insects. J. Invertebr. Pathol. 89: 193-202.

Chomczynski, P. 1992. One-hour downward alkaline capillary transfer for blotting of DNA and RNA. Anal. Biochem. 201: 134-139.

Chomczynski, P. and Sacchi, N. 1987. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162: 156-159.

Cochran, M.A. and Faulkner, P. 1983. Location of Homologous DNA Sequences Interspersed at Five Regions in the Baculovirus AcMNPV Genome. J. Virol. 45: 961- 970.

Colgan, D.F. and Manley, J.L. 1997. Mechanism and regulation of mRNA polyadenylation. Genes Dev. 11: 2755-2766.

Croizier, G., Croizier, L., Argand, O. and Poudevigne, D. 1994. Extension of Autographa californica nuclear polyhedrosis virus host range by interspecific replacement of a short DNA sequence in the p143 helicase gene. Proc. Natl. Acad. Sci. U. S. A 91: 48-52.

Dall, D., Sriskantha, A., Vera, A., Lai-Fook, J. and Symonds, T. 1993. A gene encoding a highly expressed spindle body protein of Heliothis armigera entomopoxvirus. J. Gen. Virol. 74: 1811-1818.

D'Costa, S.M., Yao, H. and Bilimoria, S.L. 2001. Transcription and temporal cascade in Chilo iridescent virus infected cells. Arch. Virol. 146: 2165-2178.

113

Doerfler, W. 1986. The Molecular Biology of Baculoviruses. Springer-Verlag, Wien & New York.

Federici, B.A. 1983. Enveloped double-stranded DNA insect virus with novel structure and cytopathology. Proc. Natl. Acad. Sci. U. S. A. 80: 7664-7668.

Federici, B.A., Bigot, Y., Granados, R.R., Hamm, J.J., Miller, L.K. and Vlak, J.M. 2000. Family Ascoviridae. In Taxonomy of viruses. Seventh report of the International Committee on Virus Taxonomy. (eds. van Regenmortel, M. H. V., C.M. Fauquet, D.H.L. Bishop, G.B. Carstens, M.K. Estes, S.M. Lemon, J. Maniloff, M.A. Mayo, D.J. McGeoch, C.R. Pringle, et al), pp. 261-265. Academic Press, London, United Kingdom.

Federici, B.A., Bideshi, D.K., Tan, Y., Spears, T. and Bigot, Y. 2009. Ascoviruses: superb manipulators of apoptosis for viral replication and transmission. Curr. Top. Microbiol. Immunol. 328: 171-196.

Federici, B.A., Bigot, Y., Granados, R.R., Hamm, J.J., Miller, L.K., Newton, I., Stasiak, K. and Vlak, J.M. 2005. Family Ascoviridae. In In Virus Taxonomy: Eighth Report of the International Committee on Taxonomy of Viruses. (eds. C.M. Fauquet, M.A. Mayo, J. Maniloff, U. Desselberger and L.A. Ball), pp. 269-274. Elsevier Academic Press, San Diego.

Federici, B.A. and Govindarajan, R. 1990. Comparative histopathology of three ascovirus isolates in larval noctuids. J. Invertebr. Pathol. 56: 300-311.

Federici, B.A., Hamm, J.J. and Styer, E.L. 1991. Ascoviridae. In In Atlas of Invertebrate Viruses (eds. J.R. Adams and J.R. Bonami), pp. 339-349. CRC Press, Boca Raton, FL.

Federici, B.A., Vlak, J.M. and Hamm, J.J. 1990. Comparative study of virion structure, protein composition and genomic DNA of three ascovirus isolates. J. Gen. Virol. 71: 1661-1668.

114

Garcia-Maruniak, A., Pavan, O.H. and Maruniak, J.E. 1996. A variable region of Anticarsia gemmatalis nuclear polyhedrosis virus contains tandemly repeated DNA sequences. Virus Res. 41: 123-132.

Glare, E.M., Divjak, M., Bailey, M.J. and Walters, E.H. 2002. beta-Actin and GAPDH housekeeping gene expression in asthmatic airways is variable and not suitable for normalising mRNA levels. Thorax 57: 765-770.

Goidin, D., Mamessier, A., Staquet, M.J., Schmitt, D. and Berthier-Vergnes, O. 2001. Ribosomal 18S RNA prevails over glyceraldehyde-3-phosphate dehydrogenase and beta-actin genes as internal standard for quantitative comparison of mRNA levels in invasive and noninvasive human melanoma cell subpopulations. Anal. Biochem. 295: 17- 21.

Goorha, R. 1981. Frog virus 3 requires RNA polymerase II for its replication. J. Virol. 37: 496-499.

Govindarajan, R. and Federici, B.A. 1990. Ascovirus infectivity and effects of infection on the growth and development of noctuid larvae. J. Invertebr. Pathol. 56: 291- 299.

Graham, S.V. 2008. Papillomavirus 3' UTR regulatory elements. Front. Biosci. 13: 5646-5663.

Granados, R.R. 1973. Entry of an insect poxvirus by fusion of the virus envelope with the host cell membrane. Virology 52: 305-309.

Guarino, L.A. and Summers, M.D. 1986. Functional mapping of a trans-activating gene required for expression of a baculovirus delayed-early gene. J. Virol. 57: 563-571.

Guarino, L.A., Xu, B., Jin, J. and Dong, W. 1998. A virus-encoded RNA polymerase purified from baculovirus-infected cells. J. Virol. 72: 7985-7991.

115

Hamm, J.J., Nordlung, D.A. and Marti, O.G. 1985. Effects of a nonoccluded virus of Spodoptera frugiperda (Lepidoptera: Noctuidae) on the development of a parasitoid, Cotesia marginiventris (Hymenoptera: Braconidae). Environ. Entomol. 14: 258-261.

Hamm, J.J., Pair, S.D. and Marti, J.O.G. 1986. Incidence and host range of a new ascovirus isolated from fall armyworm, Spodoptera frugiperda (Lepidoptera: Noctuidae). Florida Entomologist 69: 524-541.

Hamm, J.J., Styer, E.L. and Federici, B.A. 1998. Comparison of field-collected ascovirus isolates by DNA hybridization, host range, and histopathology. J. Invertebr. Pathol. 72: 138-146.

Hardwick, J.M. 2000. Cyclin' on the viral path to destruction. Nat. Cell Biol. 2: E203-4.

Heid, C.A., Stevens, J., Livak, K.J. and Williams, P.M. 1996. Real time quantitative PCR. Genome Res. 6: 986-994.

Hoopes, R.R.,Jr and Rohrmann, G.F. 1991. In vitro transcription of baculovirus immediate early genes: accurate mRNA initiation by nuclear extracts from both insect and human cells. Proc. Natl. Acad. Sci. U. S. A. 88: 4513-4517.

Hussain, M. and Asgari, S. 2008. Inhibition of apoptosis by Heliothis virescens ascovirus (HvAV-3e): characterization of orf28 with structural similarity to inhibitor of apoptosis proteins. Apoptosis 13: 1417-1426.

Jehle, J.A., Blissard, G.W., Bonning, B.C., Cory, J.S., Herniou, E.A., Rohrmann, G.F., Theilmann, D.A., Thiem, S.M. and Vlak, J.M. 2006. On the classification and nomenclature of baculoviruses: a proposal for revision. Arch. Virol. 151: 1257-1266.

Lee, H. and Krell, P.J. 1994. Reiterated DNA fragments in defective genomes of Autographa californica nuclear polyhedrosis virus are competent for AcMNPV- dependent DNA replication. Virology 202: 418-429.

116

Leisy, D.J., Rasmussen, C., Kim, H.T. and Rohrmann, G.F. 1995. The Autographa californica nuclear polyhedrosis virus homologous region 1a: identical sequences are essential for DNA replication activity and transcriptional enhancer function. Virology 208: 742-752.

Leisy, D.J. and Rohrmann, G.F. 1993. Characterization of the replication of plasmids containing hr sequences in baculovirus-infected Spodoptera frugiperda cells. Virology 196: 722-730.

Li, X., Barrett, J., Pang, A., Klose, R.J., Krell, P.J. and Arif, B.M. 2000. Characterization of an overexpressed spindle protein during a baculovirus infection. Virology 268: 56-67.

Lopez, M., Rojas, J.C., Vandame, R. and Williams, T. 2002. Parasitoid-mediated transmission of an iridescent virus. J. Invertebr. Pathol. 80: 160-170.

Lu, A., Krell, P.J., Vlak, J.M. and Rohrmann, G.F. 1997. Baculovirus and DNA replication. In The Baculoviruses (ed. Miller, L.K.), pp. 171-191. Plenum Press, New York and London.

Lu, A. and Miller, L.K. 1995. Differential requirements for baculovirus late expression genes in two cell lines. J. Virol. 69: 6265-6272.

Lu, F., Day, L., Gao, S.J. and Lieberman, P.M. 2006. Acetylation of the latency- associated nuclear antigen regulates repression of Kaposi's sarcoma-associated herpesvirus lytic transcription. J. Virol. 80: 5273-5282.

Luo, M., Wang, Y.H., Frisch, D., Joobeur, T., Wing, R.A. and Dean, R.A. 2001. Melon bacterial artificial chromosome (BAC) library construction using improved methods and identification of clones linked to the locus conferring resistance to melon Fusarium wilt (Fom-2). Genome 44: 154-162.

117

Luque, T., Finch, R., Crook, N., O'Reilly, D.R. and Winstanley, D. 2001. The complete sequence of the Cydia pomonella granulovirus genome. J. Gen. Virol. 82: 2531- 2547.

Lutz, C.S. 2008. Alternative polyadenylation: a twist on mRNA 3' end formation. ACS Chem. Biol. 3: 609-617.

Mackay, I.M., Arden, K.E. and Nitsche, A. 2002. Real-time PCR in virology. Nucleic Acids Res. 30: 1292-1305.

Maeda, S., Kamita, S.G. and Kondo, A. 1993. Host range expansion of Autographa californica nuclear polyhdrosis (NPV) following recombination of a 0.6-kilobase-pair DNA fragment originating from a Bombx mori NPV. J. Virol. 67: 6234-6238.

Mayr, C. and Bartel, D.P. 2009. Widespread shortening of 3'UTRs by alternative cleavage and polyadenylation activates oncogenes in cancer cells. Cell 138: 673-684.

McNally, M.T. 2008. RNA processing control in avian retroviruses. Front. Biosci. 13: 3869-3883.

Miller, L.K. 1988. Baculoviruses as gene expression vectors. In Annual Review of Microbiology (Anonymous), pp. 177-199.

Miller, L.K., ed. 1997. The Baculoviruses. Plenum Press, New York and London.

Miller, L.K., Jewell, J.E. and Browne, D. 1981. Baculovirus induction of a DNA polymerase. J. Virol. 40: 305-308.

Moss, B. 2001. Poxviridae: the viruses and their replication. In Fields Virology (eds. D.M. Knipe and P.M. Howley), pp. 2849-2883. Lippincott Williams & Wilkins, Philadelphia, PA.

118

Nicot, N., Hausman, J.F., Hoffmann, L. and Evers, D. 2005. Housekeeping gene selection for real-time RT-PCR normalization in potato during biotic and abiotic stress. J. Exp. Bot. 56: 2907-2914.

Ojala, P.M., Yamamoto, K., Castanos-Velez, E., Biberfeld, P., Korsmeyer, S.J. and Makela, T.P. 2000. The apoptotic v-cyclin-CDK6 complex phosphorylates and inactivates Bcl-2. Nat. Cell Biol. 2: 819-825.

Oppenheimer, D.I. and Volkman, L.E. 1997. Evidence for rolling circle replication of Autographa californica M nucleopolyhedrovirus genomic DNA. Arch. Virol. 142: 2107- 2113.

O’Reilly, D.R. and Miller, L.K. 1991. Improvement of a baculovirus pesticide by deletion of the egt gene. Biotechnology 9: 1086-1089.

O'Reilly, D.R., Miller, L.K. and Luckow, V.A. 1992. Baculovirus Expression Vectors: A Laboratory Manual. Freeman W. H. & Co, New York, NY.

Passarelli, A.L. 2007. Baculovirus RNA Polymerase: Activities, Composition, and Evolution. Virologica Sinica: 94-107.

Pearson, M., Bjornson, R., Pearson, G. and Rohrmann, G. 1992. The Autographa californica baculovirus genome: evidence for multiple replication origins. Science 257: 1382-1384.

Pearson, M.N., Bjornson, R.M., Ahrens, C. and Rohrmann, G.F. 1993. Identification and characterization of a putative origin of DNA replication in the genome of a baculovirus pathogenic for Orgyia pseudotsugata. Virology 197: 715-725.

Pearson, M.N. and Rohrmann, G.F. 1995. Lymantria dispar nuclear polyhedrosis virus homologous regions: characterization of their ability to function as replication origins. J. Virol. 69: 213-221.

119

Perez Canadillas, J.M. and Varani, G. 2003. Recognition of GU-rich polyadenylation regulatory elements by human CstF-64 protein. EMBO J. 22: 2821-2830.

Pfaffl, M.W., Tichopad, A., Prgomet, C. and Neuvians, T.P. 2004. Determination of stable housekeeping genes, differentially regulated target genes and sample integrity: BestKeeper--Excel-based tool using pair-wise correlations. Biotechnol. Lett. 26: 509-515.

Proudfoot, N. 1991. Poly(A) signals. Cell 64: 671-674.

Radonic, A., Thulke, S., Bae, H.G., Muller, M.A., Siegert, W. and Nitsche, A. 2005. Reference gene selection for quantitative real-time PCR analysis in virus infected cells: SARS corona virus, Yellow fever virus, Human Herpesvirus-6, Camelpox virus and Cytomegalovirus infections. Virol. J. 2: 7.

Radonic, A., Thulke, S., Mackay, I.M., Landt, O., Siegert, W. and Nitsche, A. 2004. Guideline to reference gene selection for quantitative real-time PCR. Biochem. Biophys. Res. Commun. 313: 856-862.

Roh, J.Y., Choi, J.Y., Li, M.S., Jin, B.R. and Je, Y.H. 2007. Bacillus thuringiensis as a specific, safe, and effective tool for insect pest control. J. Microbiol. Biotechnol. 17: 547- 549.

Salem, T.Z., Turney, C.M., Wang, L., Xue, J., Wan, X.F. and Cheng, X.W. 2008. Transcriptional analysis of a major capsid protein gene from Spodoptera exigua ascovirus 5a. Arch. Virol. 153: 149-162.

Sambrook, J. and Russell, D.W., eds. 2001. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor,New York.

Schmit, J.D. and Dill, K.A. 2010. The stabilities of protein crystals. J. Phys. Chem. B. 114: 4020-4027.

120

Schmittgen, T.D. and Zakrajsek, B.A. 2000. Effect of experimental treatment on housekeeping gene expression: validation by real-time, quantitative RT-PCR. J. Biochem. Biophys. Methods 46: 69-81.

Schwartz, S. 2008. HPV-16 RNA processing. Front. Biosci. 13: 5880-5891.

Selvey, S., Thompson, E.W., Matthaei, K., Lea, R.A., Irving, M.G. and Griffiths, L.R. 2001. Beta-actin--an unsuitable internal control for RT-PCR. Mol. Cell. Probes 15: 307-311.

Slomovic, S., Laufer, D., Geiger, D. and Schuster, G. 2006. Polyadenylation of ribosomal RNA in human cells. Nucleic Acids Res. 34: 2966-2975.

Sonntag, K.C. and Darai, G. 1995. Evolution of viral DNA-dependent RNA polymerases. Virus Genes 11: 271-284.

Southern, E.M.D. 1975. Detection of specific sequences among DNA fragments seperated by gel electrophoresis. J. Mol. Biol. 98: 503-517.

Stark, A., Brennecke, J., Bushati, N., Russell, R.B. and Cohen, S.M. 2005. Animal MicroRNAs confer robustness to gene expression and have a significant impact on 3'UTR evolution. Cell 123: 1133-1146.

Sun, X., Wang, H., Sun, X., Chen, X., Peng, C., Pan, D., Jehle, J.A., van der Werf, W., Vlak, J.M. and Hu, Z. 2004. Biological activity and field efficacy of a genetically modified Helicoverpa armigera SNPV expressing an insect-selective toxin from a chimeric promoter. Biol Control 29: 124-137.

Szewczyk, B., Hoyos-Carvajal, L., Paluszek, M., Skrzecz, I. and de Souza, M.L. 2006. Baculoviruses-re-merging biopesticides. Biotechnol Adv 24: 143-160.

Tabashnik, B.E. 1992. Evaluation of synergism among Bacillus thuringiensis toxins. Appl. Environ. Microbiol. 58: 3343-3346.

121

Tanner, N.A., Loparo, J.J., Hamdan, S.M., Jergic, S., Dixon, N.E. and van Oijen, A.M. 2009. Real-time single-molecule observation of rolling-circle DNA replication. Nucleic Acid Res. 37: e27.

Thellin, O., Zorzi, W., Lakaye, B., De Borman, B., Coumans, B., Hennen, G., Grisar, T., Igout, A. and Heinen, E. 1999. Housekeeping genes as internal standards: use and limits. J. Biotechnol. 75: 291-295.

Thiem, S.M. 1997. Prospects for altering host range for baculovirus bioinsecticides. Curr. Opin. Biotechnol. 8: 317-332.

Tian, B., Hu, J., Zhang, H. and Lutz, C.S. 2005. A large-scale analysis of mRNA polyadenylation of human and mouse genes. Nucleic Acids Res. 33: 201-212.

Tillman, P.G., Styer, E.L. and Hamm, J.J. 2004. Transmission of an ascovirus from Heliothis virescens ( Lepidoptera, Noctuidae) and effects of the pathogen on survival of a parasitoid, Cardiochiles nigriceps (Hymenoptera, Brachonidae). Enviromental Entomology: 633-643. van Oers, M.M. and Vlak, J.M. 2007. Baculovirus genomics. Curr. Drug Targets 8: 1051-1068.

Vlak, J.M. and Rohrmann, G.F. 1985. The nature of polyherin. In Viral Insecticides for Biological Control. (eds. K. Maramorosch and K.E. Sherman), pp. 489-542. Academic Press, Orlando, Florida.

Vogelstein, B. and Gillespie, D. 1979. Preparative and analytical purification of DNA from agarose. Proc. Natl. Acad. Sci. U. S. A. 76: 615-619.

Wahle, E. and Kuhn, U. 1997. The mechanism of 3' cleavage and polyadenylation of eukaryotic pre-mRNA. Prog. Nucleic Acid Res. Mol. Biol. 57: 41-71.

122

Wang, L., Salem, T.Z., Campbell, D.J., Turney, C.M., Kumar, C.M. and Cheng, X.W. 2009. Characterization of a virion occlusion-defective Autographa californica multiple nucleopolyhedrovirus mutant lacking the p26, p10 and p74 genes. J. Gen. Virol. 90: 1641-1648.

Wang, L., Xue, J., Seaborn, C.P., Arif, B.M. and Cheng, X.W. 2006. Sequence and organization of the Trichoplusia ni ascovirus 2c (Ascoviridae) genome. Virology 354: 167-177.

Westwood, J.A., Jones, I.M. and Bishop, D.H. 1993. Analyses of alternative poly(A) signals for use in baculovirus expression vectors. Virology 195: 90-99.

Willis, L.G., Seipp, R., Stewart, T.M., Erlandson, M.A. and Theilmann, D.A. 2005. Sequence analysis of the complete genome of Trichoplusia ni single nucleopolyhedrovirus and the identification of a baculoviral photolyase gene. Virology 338: 209-226.

Wormleaton, S., Kuzio, J. and Winstanley, D. 2003. The complete sequence of the Adoxophyes orana granulovirus genome. Virology 311: 350-365.

Wu, J.G. and Miller, L.K. 1989. Sequence, transcription and translation of a late gene of the Autographa californica nuclear polyhedrosis virus encoding a 34.8K polypeptide. J. Gen. Virol. 70: 2449-2459.

Xie, W.D., Arif, B., Dobos, P. and Krell, P.J. 1995. Identification and analysis of a putative origin of DNA replication in the Choristoneura fumiferana multinucleocapsid nuclear polyhedrosis virus genome. Virology 209: 409-419.

Xu, J.H. and Hukuhara, T. 1992. Enhanced infection of a nuclear polyhedrosis virus in larvae of the armyworm, Pseudaletia separata, by a factor in the spheroids of an entomopoxvirus. J. Invetebr. Pathol. 60: 259-264.

123

Xue, J.L., Cai, Q.X., Zheng, D.S. and Yuan, Z.M. 2005. The synergistic activity between Cry1Aa and Cry1c from Bacillus thuringiensis against Spodoptera exigua and Helicoverpa armigera. Lett. Appl. Microbiol. 40: 460-465.

Xue, J.L. and Cheng, X.W. 2010. Using host 28S ribosomal RNA as a housekeeping gene for quantitative real-time reverse transcription-PCR (qRT-PCR) in virus-infected animal cells. In Curr. Protoc. Microbiol. 19: 1D. 2.1-2.13. John Wiley & Sons, Inc.

Xue, J.L., Salem, T.Z., Turney, C.M. and Cheng, X.W. 2010. Strategy of the use of 28S rRNA as a housekeeping gene in real-time quantitative PCR analysis of gene transcription in insect cells infected by viruses. J. Virol. Methods 163: 210-215.

Zhang, H., Zhang, Y.A., Qin, Q., Li, X., Miao, L., Wang, Y., Yang, Z. and Ding, C. 2006. New cell lines from larval fat bodies of Spodoptera exigua: characterization and susceptibility to baculoviruses (Lepidoptera: Noctuidae). J. Invertebr. Pathol. 91: 9-12.

Zhong, H. and Simons, J.W. 1999. Direct comparison of GAPDH, beta-actin, cyclophilin, and 28S rRNA as internal standards for quantifying RNA levels under hypoxia. Biochem. Biophys. Res. Commun. 259: 523-526.

124