ANALYSIS OF GENE EXPRESSION IN MOUSE FIBROBLAST C E L L S DURING INFECTION WITH MINUTE VIRUS OF MICE by W A R R E N PERRY WILLIAMS J B.Sc, The University College of the Cariboo, 1996 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Biochemistry and Molecular Biology) We accept this thesis as conforming to the required standard THE UNVERSITY OF BRITISH C O L U M B I A March 2003 © Warren Perry Williams, 2003 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada DE-6 (2/88) Abstract Minute Virus of Mice (MVM), a mouse parvovirus, has served as a model for understanding parvovirus infection. To further understand how parvoviruses replicate within the cell and cause disease, the effect of M V M infection on host cell gene expression in mouse fibroblast cells (LA9 cells) was studied. This was accomplished through three techniques: Differential display, Clontech macroarrays, and Affymetrix microarray analysis. The RT/PCR technique differential display was used to studied altered gene expression in unsynchronized infected mouse fibroblast L A 9 cells at 12, 24 and 36 hours post-infection. Twenty-four primer pair combinations were used, representing ~ 15% of actively transcribed RNA. Surprisingly, few genes were found to change, although the technique did detect M V M NS1 transcript. In addition, the rodent retroelements, B1 & B2 SINEs and the L l LINE were all found to be increased as a result of M V M p infection. Further gene expression studies on synchronized M V M infected LA9 cells at 12 and 24 hours post-block were undertaken with Clontechs macroarrays (~1000 genes per array). Overall, the signal from these arrays were low and there were problems with background. Again, very few transcripts appeared to change as a result of infection, with only 3 genes being altered at the 12 hour time point, and another 7 reproducibly (by independent arrays) at the 24 hour time point. These included the murine B2 SINE (2.8 fold increase, confirming the differential display results), RNA exchange factor binding protein 1 (two fold increase) and three cyclins. A preliminary study was also done with the superior Affymetrix oligonucleotide microarrays (12,000 different genes, 16 oligonucleotides/gene). Of the 5347 genes and ESTs that were detected in the screen, 74 were found to be altered in unsynchronized ii LA9 cells infected with M V M p . These included genes involved in promoting cell growth (epiregulin, 13.9 fold increase), genes that inhibited growth (nuclear protein p8, 3.3 fold decrease), immune/inflammatory genes, specifically targeting transforming growth factor (3 (2.5 fold decrease), a number of transcription factors (C/EBP, 2.5 fold decrease) and genes involved in cholesterol synthesis and transport (farnesyl diphosphate sythetase, 2 fold decrease) as well as a number of unknown ESTs. The SINE response to M V M infection was further investigated. Primer extension assays confirmed that the murine B l and B2 SfNEs are up-regulated in L A 9 cells throughout M V M infection. These studies also demonstrated that the SINE response was due to R N A polymerase III transcription and not contaminating D N A or R N A polymerase II transcription. Furthermore, expression of M V M NS1 in LA9 cells by transient transfection also leads to an increase in both murine SINEs. This is the first time that the B l and B2 SINEs have been shown to be altered by viral infection. The increase in the SINE transcripts does not appear to be due to an increase in either of the basal transcription factors TFIIIC110 or 220, as these proteins do not increase during M V M p infection. However, nuclear run-on experiments to determine i f increased SINE expression was due to increased SINE R N A transcription were inconclusive. Protein levels of some 75 different protein kinases were also investigated in synchronized, M V M p infected LA9 cells. Very few changes were observed. iii Table of Contents ABSTRACT II TABLE OF CONTENTS IV LIST OF TABLES VI LIST OF FIGURES VII ABBREVIATIONS VIII ACKNOWLEDGEMENTS IX INTRODUCTION 1 T H E PARVOVIRIDAE 1 Biological significance and applications of the Parvoviridae 3 M I N U T E V I R U S OF M I C E ( M V M ) 6 MVM Structure: 6 MVM Genome structure and replication: 5 MVM Transcription 9 MVM Proteins 12 M V M - H O S T C E L L INTERACTIONS 15 MVM, the cell cycle, and apoptosis 15 MVM and host transcription/RNA process ing/RNA transport 16 MVM and host cell translation/protein trafficking 17 MVM and host intracellular/extracellular signaling 19 MVM induced cytotoxicity 19 PROJECT G O A L S : 21 S I N E S A N D L I N E S 21 Introduction 21 LINE and SINE Structure 23 SINE replication 25 SINE transcription 27 Controlling SINE expression 27 MATERIALS AND METHODS 30 /. LA9 cell culture 30 2. MVM production 30 3. Virus titration 31 4. MVM infection 32 5. Cell synchronization 32 6. Transfection of LA9 cells 33 7. Differential Display RT/PCR 33 8. Plasmid DNA preparation 35 9. Transformations/bacterial growth 35 10. Sequence determination 35 11. Probe generation 36 12. RNA isolation 37 iv 13. DNA dot blot analysis 37 14. Northern blot hybridization 38 15. Autoradiography and band quantification 39 16. Protein isolation 39 17. Western blot hybridization 39 18. Primer extension analysis 41 19. Clontech array analysis 41 20. Affymetrix microarray analysis 42 21. Preparation of nuclei for nuclear run-on assays 44 22. Nuclear Run-on assays 45 23. Kinexus protein kinase screen 46 24. Reverse transcription and PCR amplifications 47 RESULTS 49 1. DIFFERENTIAL D I S P L A Y 49 2. C L O N T E C H M A C R O A R R A Y A N A L Y S I S 60 3. A F F Y M E T R I X M I C R O A R R A Y S 76 4. S I N E EXPERIMENTS 83 B2 SINE transcripts range from 200-600 nt 83 B2 SINE transcripts are RNA and not DNA 83 B2 and Bl SINEs levels are up-regulated throughout infection 85 B2 and Bl SINE transcripts are transcribed predominantly by RNA polymerase 111 90 The major nonstructural protein of MVM, NS1, induces increased B2 and Bl SINE levels 90 Altering SINE levels 93 TFIIIC220 and TFIIIC110protein levels do not increase during MVM infection... 93 Transcriptional up-regulation ofBl andB2 SINEs remains unresolved 97 5. K I N A S E EXPRESSION A N A L Y S I S 100 DISCUSSION: 108 DIFFERENTIAL D I S P L A Y 108 C H A N G E S IN GENE EXPRESSION IN RESPONSE TO M V M P INFECTION AS DETECTED B Y C L O N T E C H C D N A A R R A Y S : 111 S I N E EXPRESSION DURING M V M INFECTION 113 K I N A S E EXPRESSION A N A L Y S I S 118 C H A N G E S IN GENE EXPRESSION IN RESPONSE TO M V M INFECTION AS DETECTED B Y A F F Y M E T R I X OLIGONUCLEOTIDE MICROARRAYS: 118 C O M P A R I N G THE TWO TYPES OF ARRAYS 123 CONCLUSIONS 125 APPENDIX ONE: PLASMID CONSTRUCTS 127 APPENDIX TWO: OLIGONUCLEOTIDES 128 APPENDIX THREE: ANTIBODIES 129 APPENDIX FOUR: KINEXUS KINETWORKS PROTEIN KINASE SCREEN 130 REFERENCES 131 v List o f Tables T A B L E 1: PARVOVIRUS CLASSIFICATION 2 T A B L E 2: A L T E R E D BANDS D E T E C T E D BY DIFFERENTIAL DISPLAY 54 T A B L E 3: GENES A L T E R E D IN A T L E A S T ONE C L O N T E C H M A C R O A R R Y 71 T A B L E 4: A F F Y M E T R I X MICROARRAY A L T E R E D TRANSCRIPTS 81 T A B L E 5: GROUPING OF A L T E R E D GENES D E T E C T E D B Y A F F Y M E T R I X MICRO A R R A Y S BASED ON SIMILAR FUNCTION OR P A T H W A Y S 120 T A B L E 6: COMPARING M A C R O A R R A Y S A N D MICRO A R R A Y S 124 vi List of Figures FIGURE 1: M V M STRUCTURE 7 FIGURE 2: M V M TRANSCRIPTION 10 FIGURE 3: THE M V M PROTEINS 13 FIGURE 4: TOTAL LA9 PROTEIN L E V E L S DO NOT C H A N G E DURING M V M p INFECTION UNTIL L A T E IN INFECTION 18 FIGURE 5: CYTOTOXIC EFFECTS OF M V M 20 FIGURE 6: RETROELEMENT CLASSIFICATION 22 FIGURE 7: SINE A N D LINE STRUCTURE 24 FIGURE 8: CURRENT M O D E L OF SINE REPLICATION 26 FIGURE 9: STEPWISE A S S E M B L Y OF RNA P O L Y M E R A S E III TRANSCRIPTION FACTORS 28 FIGURE 10: NORMALIZING THE C L O N T E C H M A C R O A R R A Y D A T A 43 FIGURE 11: DIFFERENTIAL DISPLAY 50 FIGURE 12: DIFFERENTIAL DISPLAY P O L Y A C R Y L A M I D E GEL 52 FIGURE 13: REVERSE NORTHERN BLOTS 56 FIGURE 14: THE 7A1U B A N D RESULTS 57 FIGURE 15: A L T E R E D EXPRESSION OF SELECTED BANDS 59 FIGURE 16: LA9 C E L L S A R E SYNCHRONIZED B Y SERUM STARVATION 62 FIGURE 17: M V M INFECTED LA9 CELLS ARREST IN S/G 2 63 FIGURE 18: C L O N T E C H M A C R O A R R A Y CONTROLS 64 FIGURE 19: A TYPICAL CLONTECH A T L A S MOUSE 1.2 A R R A Y 66 FIGURE 20: GENE RATIOS (24I/24M) WITH A SIGNAL >500 COUNTS F R O M SYNCHRONIZED LA9 CELLS 24 HOURS POST B L O C K 68 FIGURE 21: GENE RATIOS (12I/12M) WITH A SIGNAL >500 COUNTS F R O M SYNCHRONIZED LA9 CELLS 12 HOURS POST B L O C K 69 FIGURE 22: GENE RATIOS (36I/36M) WITH A SIGNAL >500 COUNTS F R O M UNSYNCHRONIZED LA9 CELLS 36 HOURS POST INFECTION 70 FIGURE 23: COMARING GENE RATIOS (24I/24M) F R O M A R R A Y S I & II (SYNCHRONIZED LA9 CELLS 24 HOURS POST B L O C K ) 73 FIGURE 24: NORTHERN BLOT CONFIRMING M A C R O A R R A Y RESULTS 75 FIGURE 25A: A F F Y M E T R I X M I C R O A R R A Y GRO ONCOGENE PROBE SET 77 FIGURE 25B: SCATTER PLOT FOR THE TWO A F F Y M E T R I X A R R A Y S 77 FIGURE 26A: RESOLVING THE SIZE OF THE B2 SINE TRANSCRIPTS 84 FIGURE 26B: SINE TRANSCRIPTS A R E RNA A N D NOT D N A 84 FIGURE 27: REVERSE TRANSCRIPTION OF SINES IN P O L Y M E R A S E III A N D II TRANSCRIPTS 86 FIGURE 28: B2 SINE TRANSCRIPTS A R E UP-REGULATED THROUGHOUT M V M INFECTION A N D A R E PREDOMINATELY TRANSCRIBED B Y POL. I l l 87 FIGURE 29: B2 SINE TRANSCRIPTS A R E UP-REGULATED THROUGHOUT M V M INFECTION A N D A R E PREDOMINATELY TRANSCRIBED B Y POL. Ill 89 FIGURE 30: M V M NS1 INDUCES INCREASED B2 SINE EXPRESSION 91 FIGURE 31: M V M NS 1 INDUCES INCREASED B1 SINE EXPRESSION 94 FIGURE 32: METHODS OF INCREASING SINE A B U N D A N C E 96 FIGURE 33: TFIIIC220 PROTEIN DOES NOT INCREASE DURING M V M INFECTION 98 FIGURE 34: TFIIIC110 PROTEIN DOES NOT INCREASE DURING M V M INFECTON 99 FIGURE 35: N U C L E A R RUN-ON RESULTS 101 FIGURE 36: KINEXUS'S PROTEIN KINASE SCREEN FOR M O C K A N D M V M INFECTED SYNCHRONIZED LA9 CELLS 24 HOURS POST-BLOCK 103 FIGURE 37: CONFIRMING KINEXUS'S PROTEIN KINASE SCREEN 106 vii Abbreviations BLOTTO PBS-T with 5% Powdered skim milk DEPC Diethyl Pyrocarbonate DTT Dithiothreitol EDTA Ethylene-diaminetetra-acetic acid FBS Fetal Bovine Serum HEPES N-[2-Hydroxyethyl]piperazine-N'-[ethanesulfonic acid] M M L V Moloney Murine Leukemia Virus MOI Multiplicity of infection MOPS 3-[N-morpholino]propanesulfonic acid M V M Minute Virus of Mice PB Post Block PBS Phosphate buffered Saline PBS-T Phosphate buffered Saline with 0.05% Tween 20 PCR Polymerase Chain Reaction PI Post Infection PMSF Phenylmethylsulfonyl fluoride PVDF Polyvinylidene Fluoride RT Reverse Transcription SDS Sodium Dodecyl Sulfate SSC (20X) Sodium Chloride (3M) and Sodium Citrate (0.3M, pH 7) T A E buffer 40mM Tris-acetate, 1 mM EDTA, pH 8.0 TBE buffer 90mM Tris-borate, 0.1 mM EDTA, pH 8.3 TES N-Tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid TNE buffer 50 mM TrisCl pH 8.7, 150 mM NaCl, and 0.5mM E D T A Tris Tris(hydrixymethyl)aminomethane X-gal 5 -bromo-4-chloro-3 -indoly 1-beta-D-galactopyranoside viii Acknowledgements Completing a Ph.D. is no easy task. Luckily for me, there were many people who helped along the way. First and foremost, I would like to thank my supervisor, Dr. Caroline Astell for providing this project, all the helpful advice, and patience. I would also like to acknowledge all the past members of the Astell lab for all their input and help. Specifically, two people deserve to be singled out for further thanks: Dr. Janet St. Amand, who always had time to help a young "wet behind the ears" graduate student and who got me out of the lab and into the mountains, and Lilian Tamburic, whose exceptional trouble-shooting skills were exploited mercilessly and who introduced me to Kurosawa, Kusturica, and scotch. These three women taught me how to do good science and opened my eyes to new ideas and new horizons. My family (Joanne, Perry, Angie and Bradley) played a crucial role in the completion of this thesis with countless phone calls of encouragement, support and love. Thanks also go out to Emily Crowe, Cameron Mackereth, and Tye Teames for providing the a pitcher of beer or a bottle of wine for those experiments that did not quite make it into this work. Finally and most importantly, I would like to thank my partner Adam Lorch. Without him, this thesis would not be possible. ix Introduction The Parvoviridae The family Parvoviridae includes extremely small non-enveloped icosohedral viral particles containing a single stranded D N A genome of approximately five to seven kilobases. It is divided into two separate sub-families based on the viral host. The Densovirinae are a group of autonomously replicating viruses that infect arthropods while the Parvovirinae are a group of both autonomously and helper-dependent viruses that replicate in vertebrates. The Densovirinae consist of three genera, the Densoviruses, the Iteraviruses and the Brevidensoviruses based on genome length, presence or absence of inverted terminal repeats (ITR's), and genome organization (monosense vs ambisense) [1]. The Parvovirinae can also be further divided into three genera: the Erythroviruses, the Dependoviruses and the Parvoviruses. The Erythroviruses are autonomously replicating viruses that replicate in erythroid progenitor cells. Examples include the human pathogen B19 and simian parvovirus (SPV), which infects Cynomolgus monkeys. Dependoviruses are characterized by the requirement, though not absolute, of a helper virus for viral replication. Examples are the human adeno-associated viruses (AAV), serotypes one to five. Finally the Parvovirinae are a group of autonomously replicating viruses that replicate in a wide variety of cells. This is a diverse group of viruses with examples including minute virus of mice (MVM), Rat H-1 parvovirus, Hamster parvovius (HPV), canine parvovirus (CPV), feline panleukopenia virus (FPV), Aleutine mink disease parvovirus (ADV), and Porcine parvovirus (PPV) [2]. Table 1 shows the classification scheme of the Parvoviridae with representative viruses of each genus. 1 Family Parvoviridae: Sub-family Denosovirinae ( Athropod hosts Genus Densovirus o Large genome 5.5 -6 kb o Large ITRs o Ambisense genomic organization Examples Galleria mellonella densovirus (GraDNV) Acheta domesticus densovirus (AdDNV) Culex pipiens denosovirus (Q?DNV) Genus Itera virus o Genome size of 5 kb o ITRs of about 250 nt o Monosense genomic organization Example Bombyx mori densovirus (BmTJNV) Genus Brevidensovirus o Genome size of 4 kb o No ITRs but terminal hairpins o Monosense genomic organization Example Aedes aegypti densovirus (AaeDNY) Sub-family Parvovirinae ( Vertebrate hosts Genus Erythrovirus o Replicates in erythroid progenitor cells Examples Human Bl9 Parvovirus Simian Parvovirus (SPV) Genus Dependo virus o Requires helper virus for optimal replication Examples Adeno-associated viruses (AAV) 1-5 Genus Parvovirus o Autonomously replicating viruses that do not replicate in erythroid progenitor cells Examples Minute virus of Mice (TVTVM), Feline panleukopenia virus (FPV) Rat HI parvovirus, Aleutian mink disease parvovirus (ADV) Porcine Parvovirus (PPV), Canine parvovirus (CPV) Table 1: Parvovirus classification. 2 Biological significance and applications of the Parvoviridae The purpose of this section is to give a very broad overview of parvoviruses and their uses as reviewed in Contributions to Microbiology V o l . 4: Parvoviruses from Molecular Biology to Pathology and Therapeutic Uses, edited by S. Faisst and J. Rommelaere. A thorough review of these topics is beyond the scope of this thesis. Wherever possible, I have cited relevant review articles. The Parvoviridae are of interest for many reasons. Their most predominant features are their ability to cause disease in both humans and animals (of important economic value) and their increasing use as gene transfer vectors. However, there is also considerable interest in using them for pest control and in studying basic cell functions. The human pathogen B19 parvovirus causes hematological disease due to its tropism for replication in erythroid progenitor cells with the disease outcome depending on the age and state of the human host. The virus is extremely common in the population; 50% of children over the age of 5 have been infected and this increases to 90% in the elderly. An acute B19 infection in otherwise healthy children and adults results in fifth disease/Erythema Infectisosum. B19 infection is initially characterized by fever, chills, headache, myalgia and also by a decrease in reticulocytes due to viral replication in the progenitor cells. As the host develops antibodies and the virus is cleared from the system, the infection enters the second phase. This is characterized by the development of a rash and arthralgia and in some cases transient arthritis. These symptoms resolve in about two to six days. B19 infection is of more concern in immunodeficient/immunocompromised hosts and patients with increased erythropoesis 3 (due to hemolysis or hemorrhage) where persistent anemia (due to the inability to clear the virus) or transient aplastic crisis, respectively, can result. Pregnant women are also at risk, as a B19 infection during the second trimester of pregnancy can lead to Hydrops fetalis (a form of spontaneous abortion as a result of B19 infection of the fetus) or chronic anemia, possibly due to immunotolerance of the B19 virus [3]. Individuals with a genetic based hemolytic anemia can also undergo transient aplastic crisis as a result of B19 infection [4]. Parvoviruses also cause disease in animals of interest, with the outcome depending on the age and state of the host, with the fetus and newborn being most at risk. As parvoviruses generally require specific host cell replication factors, they infect tissues with rapidly developing cells, such as the gut epithelium, immune cells, or bone marrow, as well as many different tissues in the fetus [5]. Feline panleukopenia virus (FPV), canine parvovirus (CPV, closely related to FPV) and porcine parvovirus (PPV) cause pathological infection in cats, dogs, and pigs, respectively. In all three viruses, the fetus and newborn are most susceptible to infection, with a high risk of spontaneous abortion or re-adsorption of the fetus. Vaccines are available for FPV, CPV, and PPV [5,6]. Infection of adult mink with Aleutian mink disease parvovirus, A D V , leads to either death or chronic infection. As there is no successful vaccine for A D V , the only treatment is the destruction of the infected animals [5, 7]. Finally, infection with BmDN\ leads to the death of Bombyx mori (silk worm) larva. This caused significant economic hardship among silk growers until BmDNV resistant strains were developed [8, 9]. With the emergence of gene therapy came the interest in using parvoviruses as gene transfer vectors. The most well studied and successfully used parvovirus transfer 4 vectors are the adeno-associated viruses and there are many review articles on this subject [10, 11]. A A V is not known to cause disease in humans and, in the absence of helper virus, becomes latent, integrating into the host genome. The only major limitation with this viral vector is that only genes of less than approximately four kilobases can be used. A literature search reveals that a large number of A A V gene therapy trials are ongoing, including studies for treatment of cystic fibrosis [12] and Hemophila B [13]. Efforts are also underway to determine the potential of using B19/AAV hybrids as a viral vector to target the erythroid progenitor cells [14]. Likewise, M V M is also being studied as a possible vector where the death of the infected cell is desired or inconsequential such as in oncolytic treatment or vaccination [15]. Finally, the densoviruses JcDNV, GmDNV (greater wax moth virus), and AaeDNV (mosquito virus) are being pursued as gene transfer vectors [16]. The densoviruses are being examined as potential agents for arthropod pest control. Densoviruses are suited for this work as they are usually lethal in the host's larval stage, and are not known to infect humans or mammals. SfDNV and CeDNV have been used to control outbreaks of S. fusca and C. extranea (oil palm leaf eaters) infections [17, 18]. AaeDNV has also been used to control Aedes aegypti mosquito larvae in the former Soviet Union and a commercial formulation has been developed [19, 20]. The development of densovirus gene transfer vectors will further aid in this area. Finally, the high dependence of parvoviruses on the host cell for viral replication presents an opportunity to study host cell-virus interactions. Parvoviruses are currently being used to study apoptosis and other forms of cell death [21-23], anti-neoplastic activity [24], cell cycle [25, 26], and D N A replication [27, 28]. 5 Minute Virus of Mice ( M V M ) The parvovirus minute virus o f mice ( M V M ) has served as a prototype for the vertebrate parvoviruses. The genome has been completely sequenced [29], the viral transcripts and viral proteins identified [30-33], and the genome replication cycle worked out [27, 28, 34], although some details o f the replication mechanism remain to be elucidated. M V M can be cultured in vitro (causing an acute infection in a mouse fibroblast cell line) and grown to high titre [33]. It has extremely stable capsids [35], and importantly, infectious clones are available [36]. Furthermore, M V M is usually non-pathogenic for its murine host [37], (although some strains o f inbred neonatal mice are susceptible [38, 39]) and so can be handled with limited biosafety precautions. Finally, a variant, M V M j , [40] displays a tropism for murine lymphocytes [41] and, unlike M V M p , does not infect mouse fibroblasts providing a useful comparison between the two strains. M V M has been successfully used to model parvovirus genome replication and to characterize the functions o f the major non-structural protein. M V M Structure: M V M consists o f -26 nm non-enveloped icosohedral viral particles, wi th each particle containing one copy o f a single-stranded (negative sense) D N A genome. These particles are stable up to 56°C and through a pH range o f 3 to 9 and can survive treatment with alcohol or ethers but are susceptible to bleach [27, 35]. The structure o f M V M i determined to 3.5 A [42] resolution by x-ray diffraction o f crystals is shown in Figure la . The capsid has a T = l symmetry and is made up 60 viral capsid proteins (VP1, VP2, & VP3) with VP2 being the dominant capsid protein. 6 B Figure 1: M V M Structure. A . 3D crystal structure of M V M as viewed from the 5-fold axis of symmetry. Taken from Agbandje-McKenna et al., 1998. B. M V M genome structure showing the 3' and 5' hairpins, unpaired bubbles and flip and flop designations. 7 M V M Genome structure and replication: The M V M genome is a single-stranded linear D N A molecule of 5149 nucleotides in length [29]. It is packaged predominantly in the negative form (i.e. complementary to the transcripts) and contains hairpins at both the 3' and 5' ends (Figure lb). The 3' hairpin is "y" shaped with only one small sequence in the hairpin stem that is unpaired. The 5' hairpin sequence exists in two forms, "flip" and "flop", which differ only at three nucleotides within the hairpin turn itself and in a small unpaired sequence bubble within the stem. The unpaired sequences found in the 5' hairpin of flip are the inverted complement of those found in flop and both flip and flop are packaged equally into the capsid. Both the 3' and 5' hairpins are essential for M V M replication [43]. A molecule of the major non-structural viral protein NS1 can be found covalently linked to the 5' end of the genome, and extends outside the virus particle once the genome is encapsulated [44, 45]. This protein can be removed without affecting viral replication. Replication of the M V M genome is a complicated and not completely understood process. The virus has two origins of replication, located at the left and right hairpins, respectively [43, 46]. These two origins differ in size, primary sequence and secondary structure. The virus does not encode any polymerases, and so is dependent on a series of host cell factors for replication to occur. These include D N A polymerase 8, PCNA [47], parvovirus initiation factor (PIF) [48], the high-mobility group 1/2 proteins (HMG1/2) [49], replication protein A (RPA) and Replicative factor C [34]. Presumably, as a result of these requirements, the virus requires the host cell to be in S phase in order to replicate [50]. Furthermore, M V M cannot induce cells to enter S phase and so must wait until the cell enters S phase naturally. However, once S phase is reached, the virus causes the host 8 cell to arrest permanently in S/G2 and thus allow extensive replication of the virus. The viral non-structural protein NS1 is also required for a number of functions in replication. Replication is thought to occur through a modified form of rolling circle D N A replication [28, 51] using a single continuous D N A strand containing multiple copies of the viral genome. The processing of this molecule into viral progeny is not completely understood; however NS1, with its nickase, helicase, and D N A binding functions, is required. M V M genome replication occurs in specific bodies within the nucleus of the cell [52]. M V M Transcription M V M contains two promoters: P4 located at the left hand end of the genome, at 4 map units, and P38 located in the middle of the genome, at 38 map units [29]. A poly-adenylation signal is located on the extreme right hand end of the genome [53]. Transcription runs from left to right with the coding information being present in the complementary strand to the original viral genome. There are three major open reading frames in M V M , one in each of the three reading frames [32] (Figure 2). The P4 promoter controls transcription on the left hand side of the genome and generates the R l and R2 transcripts, which encode the non-structural proteins NS1 and NS2, respectively [31, 54]. The R2 transcript is an alternatively spliced form of the R l transcript where -1.5 kb of sequence is removed and the reading frame is shifted [55]. The resulting translation product, NS2, thus shares its N terminal sequence with NS 1 but has its own unique C terminal sequence. There is also a second smaller splice site located near the center of the genome that contains multiple splice donor and acceptor 9 R3< [y\:v:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.;.:.:.:,:.;.v.:::.:v:.^ (A)n A (A)ll A t:.:A;.-.-.-.-.-.-.-.-.-A-..,,,.-...v.-,.v...] (A)n VP1 VP2 R2S (A)n ">! (A)n (A)n NS2 -€ RH -c -c P 4 i P38 (A)n (A)n (A)n NS1 Poly A Frame 1 Frame 2 Frame 3 B 2 3 ORF1 ORF2 ORF3 Figure 2: M V M Transcription. A. The three major M V M transcripts ( R l , R2, and R3), splicing patterns and their products. The shading in the box indicates reading frame. For example, translation of the top R2 transcript starts in frame one and then changes (as a result of splicing) to frame three, and then switches back (again due to splicing) to frame one. B. The three major open reading frames in M V M . 10 sites [56]. As a result of further alternate splicing at this site, the R2 transcripts are actually a group of three slightly different transcripts. Each of these R2 transcripts encodes a NS2 protein whose protein sequence differs at its C terminal end (between 6 to 11 amino acids [57]). The significance of this heterogeneity is not known. The R l transcript also contains the small splice site, but as the NS 1 coding sequence terminates before the splice donor sites, all three R l transcripts will produce the same protein (Figure 2). The P38 promoter controls transcription on the right hand of the genome and generates the R3 transcripts, which encodes the structural proteins VP1 and VP2 [58]. As in the R l and R2 transcripts, the R3 transcript exists in three forms, again due to the presence of the small splice site. In one transcript, a translation start site is preserved upstream of the splice donor site and the VP1 protein is generated. In the other two transcripts, the same translation start site is removed, and translation does not start until much further along the mRNA, yielding the VP2 protein (Figure 2). There are several transcription factors involved in control of M V M transcription. NS1 has the ability to both bind and transactivate both promoters. It first enhances transcription at the P4 promoter (although it is not required for efficient transcription [59]), and then later as the infection progresses, it transactivates the P38 promoter (where it is essential, [60]). This ensures that the non-structural proteins are predominantly expressed early in infection and the capsid proteins later in infection. The transcription factor E2F is also needed for P4 activation [61, 62] along with the basal R N A polymerase II machinery. The P38 promoter seems to require Spl [63] and CREB binding protein (CBP, [64]) interactions with NS1 in order for transcription to occur. 11 M V M Proteins There are five viral proteins encoded by M V M . These include the non-structural proteins NS1 and NS2 and the structural capsid proteins VP1, VP2, and VP3 (derived from VP2; see below) (Figure 3a). The NS 1 protein has a number of functions. It is a multi-domain protein of 672 amino acids with an apparent molecular weight of 83 kDa [54]. It is heavily phosphorylated by at least two kinases, casein kinase II and protein kinase C X (PKC X). Phosphorylation of NS1 by PKC X controls a number of NS1 functions [65-68]. NS1 is predominantly nuclear (NS1 has a nuclear localization signal) and is present in specific M V M replication bodies within the nucleus [47, 69]. As mentioned above, NS1 contains nicking [70], helicase [71], D N A binding [72] and ATP binding/hydrolysis domains [73] that are essential for replication. NS 1 is also required for transactivation of the P4 and P38 promoters [74], cell cycle arrest [25], and is cytotoxic to the host cell. It forms dimers and oligomers through its oligomerization domain [75]. Figure 3b shows the various domains present in NS 1. Finally NS 1 has been found to interact with a large number of different proteins, including: Survival of Motor Neuron (Immunopreciptation, BIAcore biosensor) [77], T A T A Binding Protein (TBP) (GST-NS1 capture assay), Transcription Factor II A (TFIIA) (GST-NS1 capture assay), Spl (GST-NS1 capture assay) [63], and NS1 associated protein/HnRNP-Q (NSAP1) (yeast two hybrid) [78]. NS 1 also interacts with Parvovirus Initiation Factor (PIF) [76] although no evidence of binding has been shown. The function(s) of NS2 protein is much less understood. As mentioned above, it consists of three proteins that vary in their C-terminal sequence. They have apparent 12 A 83 kDa[ 25 kPalllllllllliliil NS-2 ] NS-1 83 kDa I 64 kDa 61 kDa | VP-1 VP-2 1 VP-3 1 Cytotoxicity* DNA Binding NTP binding Transactivation Nicking NLSf Helicase | NS-1 672 Oligomerization Figure 3: The M V M proteins. A : The M V M proteins and their apparent molecular weights. Common areas indicate sequence of shared amino acid sequence. B: NS1 and its associated domains (taken from Corbau et al., 2000). 13 molecular weights of 25 kDa and lengths of 188 or 197 amino acids. The NS2 proteins exist in both phosphorylated and unphosphorylated forms with the unphosphorylated forms found both in the nucleus and cytoplasm and the phosphorylated form found exclusively in the cytoplasm [79]. The NS2 proteins are needed for efficient viral replication [80, 81], and viral protein translation [82] in the murine cells. Furthermore, they are important for capsid assembly [83]. NS2 proteins bind to Crml (exportin-1) [79, 84] and, through a nuclear export signal (NES) present in NS2, are exported to the cytoplasm. Mutation of the NS2 NES leads to retardation of viral egress from the nucleus and a decrease in the production of viral progeny ssDNA [85, 86]. Finally they have been found to associate with SMN [87], and the 14-3-3 proteins [88]. VP1, which is present as 15% of the capsid, is a multifunctional protein with an apparent molecular weight of 83 kDa and a length of 718 amino acids. It contains a unique N terminal sequence and shares its C terminal sequence with VP2 and V P 3 [89]. The shared C terminal sequence of VP1 forms part of the full viral capsid, analogous to that of VP2, while the unique N terminal sequence is tucked inside the capsid. Details on the structural role of VP1 and VP2 are reported in Agbandje-McKenna et al., 1998 [42]. The unique N terminal sequence of VP1 is not required for capsid assembly but is found to be essential for viral infectivity and has a role in viral entry at some stage past the point of capsid binding [90]. More recently, it was found that this region contains a functional phospholipase A2 (PLA2) active site that is needed for infection and is conserved among many different parvoviruses [90]. The role of this viral PLA2 is not known. VP2 is the major capsid protein, forming the majority of the capsid. It has an apparent molecular weight of 64 kDa and a length of 584 amino acids. The protein 14 adopts an eight-strand (3-barrel conformation and can spontaneously form empty capsids [42]. VP3 is a cleavage product of VP2 [89]. The first approximately twenty amino acids are lost from the N-terminus, yielding a protein with an apparent molecular mass of 63 kDa and a length of 564 amino acids. This cleavage event appears to have no effect on the stability of the virus capsid. MVM-host cell interactions Viruses interact with their host cells in a number of different ways. They can alter replication, transcription and translation, RNA and protein trafficking, the cell cycle, apoptosis, the stress response, intracellular and extracellular signaling, and the immune response. This next section will examine some of these interactions observed with M V M . M V M , the cell cycle, and apoptosis M V M can arrest transformed cells in S/G2. This is accompanied by an increase in p53 and p21 c i p (a cyclin dependent kinase inhibitor that acts downstream of p53) proteins with p53 required for the block in S and both p53 and P21 c i p required for the block in G2 [25]. Initially, the block in cell cycle was thought to be due to NS-1 [91] as expression of NS-1 led to cell cycle arrest in G1/S/G2. However, NS1-expressing cells only showed an increase in p21 c i p and not p53. This is an interesting result as NS1 can induce nicks in host cell D N A [91] and D N A damage is a potent activator of p53. It is now hypothesized that MVM-induced cell cycle arrest is a combination of NS1 expression and the sudden increase in linear single stranded viral DNA, which could lead to p53 activation. Another 15 interesting observation is the accumulation of cyclin A , and not cyclin E or cdk2, in the M V M replication bodies [52]. Cyclin A is responsible for the transition from S to G2 and so sequestering cyclin A could also contribute to cell cycle arrest. Many parvoviruses can induce apoptosis in their host cells with examples including B19 [21, 22] and Rat HI parvovirus [23]. Given that M V M activates p53 and NS 1 causes nicking in host cell DNA, it would be expected that M V M infection would induce apoptosis in the host cell. So far, this remains unproven. D N A "ladders" characteristic of apoptosis have been observed in a mouse fibroblast cell line; however, effector caspases were not activated and the classic apoptotic morphology was not seen (L. Tamburic and C. Astell, unpublished results). It has been reported recently that NS 1 and NS2 interact with S M N [77, 87]. S M N has been shown to protect primary neuronal cells from virus-induced apoptosis [92]; however, what role S M N plays in M V M infection remains undefined. M V M and host transcription/RNA processing/RNA transport M V M ' s effect on host transcription has not been well characterized. There has been a report that M V M infection of Ehrlich ascities cells modifies host cell transcription and alters c-fos expression [93]. NS1 has also been shown to have a transinhibiting effect on the Rous sarcoma virus long terminal repeat promoter [74] and the C M V promoter [94] as visualized by reporter constructs. Interestingly, i f a NS-l / lexA fusion was tested on the same Rous sarcoma promoter with a LexA binding site, NS 1 exhibited a transactivating effect [95]. NS1 has also been shown to exert a transactivating effect on the c-erbAl promoter, again by a reporter gene assay, and M V M infected cells showed increased amounts of the c-erbAl transcripts in ras-transformed FREJ4 rat fibroblasts 16 [96]. Furthermore, NS1 was found to bind to FRE (a sequence that interacts with transcription factors belonging to the FTZ-F1/SF sub-group of nuclear hormone receptors) in the c-erbAl promoter [97]. The c-erbAl gene encodes thyroid hormone (T3) receptor a, which is involved in cell differentiation and proliferation, among other functions [98, 99]. There has been no further characterization of the interaction between MVM/NS1 and thyroid hormone (T3) receptor a. Finally, it has been suggested recently thatNSl oligomerization is needed for its transactivating/transrepressing functions [94]. There are no definititive studies to indicate that M V M alters host RNA processing, but there is increasing evidence that M V M proteins interact with the splicing machinery. The recently found NS1 binding protein NSAPl/hnRNP Q, [78] was identified as a component of the spliceosome [100]. Furthermore, NSAPl/hnRNP Q interacts with S M N [100] and, as already stated, S M N interacts with NS1/NS2 [77, 87]. The roles of these NS1 interactions are not understood. There have been no reports in the literature that M V M affects host cell RNA transport. The initial discovery of the NS2/Crml interaction did suggest a possible role in R N A transport as interactions of other viral proteins like HIV Rev [101] or Adenovirus E1B 55 [102] with Crml affect host or viral RNA transport. However, disrupting the NS2/Crml interaction did not affect viral protein synthesis [85] and there are no reports of changes in global protein expression during M V M infection. M V M and host cell translation/protein trafficking There are no known reports of M V M altering global host cell protein translation and trafficking. Total cellular protein levels do not change upon M V M infection until 17 hr. post infection Figure 4: Total L A 9 protein levels do not change during M V M p infection until late in infection. PVDF membrane of L A 9 fibroblast total cell protein stained with Coomassie blue. M : mock infected cells. Each lane represents 5 u,l o f total protein extract from one dish of infected LA9 cells. The decrease seen at late t ime points is presumably due to increased cell death caused by the virus. Membrane provided by Dr. Cynthia Shippam-Brett. 18 late time points, presumably due to cell lysis (Figure 4; C. Shippam-Brett & C. Astell, unpublished results). The M V M NS2 protein is needed for efficient translation of the viral proteins [82], although how this is accomplished is not known. M V M and host intracellular/extracellular signaling Several signaling pathways have been implicated in M V M infection but have not been well characterized with respect to the virus or host. As already reported, M V M infection leads to increased transcription of c-erbAl and this could alter the thyroid hormone induction pathway. M V M protein VP1 contains an active phospholipase A2 domain (PLA2). Besides being used for lipid turnover, P L A 2 is associated with inflammatory, signal transduction, and various physiological and pathological processes [103, 104]. Finally, NS2 protein interacts with the 14-3-3 proteins [88]. The 14-3-3 proteins display a variety of functions in the cell including signal transduction, checkpoint control, and apoptotic and nutrient sensing pathways. It is thought to accomplish these functions by altering the sub-cellular localization of partners [105]. The exact role of the NS2/14-3-3 interaction is not known. M V M induced cytotoxicity M V M p is cytopathic for many types of transformed cells [106, 107] including the mouse fibroblast cell line L A 9 (Figure 5). This cytopathic effect is believed to be mediated by NS1, as expression of NS1 alone in transformed cells also leads to cell death [108, 109] although NS2 can enhance this process [110]. NS-1 cytotoxicity has been mapped to the C and N terminals of the protein [74]. How NS-1 is cytotoxic to cells is not known, although there are several theories. These include the ability of NS1 to 19 60 hours mock infected 60 hours infected 72 hours infected Figure 5 : Cytotoxic effects of M V M . Transformed mouse fibroblast L A 9 cells infected or mock infected with MVMp at various times indicated. Photos taken by L. Tamburic. 2 0 transactivate or transinhibit other genes [74], to nick host cell D N A [91], to induce cell-cycle arrest [25] and to alter specific protein phosphorylation and synthesis [111]. Project Goals: The purpose of my studies was to further characterize the host cell response to infection by a small D N A virus. Specifically, I have focused on changes in gene expression in the mouse fibroblast cell line LA9 during infection with M V M p . My studies included examining changes in mRNA expression using differential display analysis and commercially available macro and microarrays. I also used a commercially available service to detect changes in kinase levels following infection. This work is to be considered as a survey of gene expression changes and does not examine the entire mouse transcriptome. SINEs and LINEs Introduction As this thesis examines Short INterspersed Elements (SINEs) and Long INterspersed Elements (LINEs) expression during M V M infection, a brief review is included to aid the reader. LINEs and SINEs are members of the retroelement family and are commonly referred to as retroposons (Figure 6a) [112]. Large numbers of retroposons are found in eukaryotic DNA, including plants, mollusks, chordates, and arthropods [113,114]. In humans, there are approximately 860,000 copies of the various LINEs, composing approximately 21% of the genome, while SLNEs comprise another 13%o of the genome with over 1.5 million copies [115]. The numbers are similar in the mouse genome, with LLNEs composing approximately 19%> of the genome with 660,000 21 0) © + LTR - LTR e O O a • o Retroviruses (infectious) Endogenous Retroviruses/ retrotransp osons (non-infectious) + RT LINEs -RT SINEs Retro-Pseudogenes B Genomic DNA RNA Integration / \ Transcription cDNA < RNA Transcript Reverse Transcription Figure 6 : A: Retroelement classification. LTR: Long terminal repeat; RT: Encoding reverse transcriptase. B: Generalized retroposon replication. 22 copies and SINEs comprising another 8.3% of genome with over a million and a half copies [116]. Retroposons replicate via a RNA intermediate that is reverse transcribed prior to re-integration into the genome (Figure 6b). A brief review of SINE and LINE structure, SINE replication, and factors that affect SINE expression will be discussed below. L I N E and SINE Structure LINEs average 6 kb in size and contain a internal R N A polymerase II promoter, one or two open reading frames (depending on the LINE), and a poly-adenylation signal at the extreme 3' end (Figure 7a). The first ORF, ORF 1, is only found in younger LINEs, for example L l LINE, and its function is unknown. The second ORF, ORF2, encodes a protein with both reverse transcriptase and endonuclease domains. In younger LINEs this is an apurinic/apyrimidinic endonuclease, while in older LINEs it is a restriction-like endonuclease. The integrated LINE D N A is flanked by direct repeats, thought to be a result of the integration and D N A repair process [117]. Human and mice share L l , L2, and L3 LINEs and only the L l LINE is considered to be active, i.e. replicating. SINEs range in size from 50-500 nt in size and do not contain any coding sequence. They contain an internal R N A polymerase III promoter (AB box) and a poly A tract at the 3' end (Figure 7a). Like LINEs, they are flanked by direct repeats of 5-20 nt in size. Many SINEs are derived from tRNAs (for example the murine B2 SINE resembles Ala tRNA) or other small RNAs [118] and can also have a LINE like sequence at the 3' end (for example the MIR SINE and L2 LINE share some of the same sequence) 23 Pol. II QRF1 ORF2 (EN/RT) pA> AAAA1 Prototypical LINE Pol. Ill AAAA Prototypical SINE B Pol III Pol III • k l [ P ^ n Pol III |tRNA-like j A Alu SINE (300 nt) Bl SINE (135 nt) B2 SINE (209 nt) Figure 7: SINE and LINE structure. A : SINE and LINE prototypical structures. Note that older LINEs lack ORF1. Pol II & Pol III: RNA polymerase II & III promoters; EN: Endonuclease; RT: Reverse transcriptase; pA: Poly-adenylation signal; Solid arrow: Target site sequence duplication (adapted from Weiner et al., 2002). B : Alu, B l & B2 SINEs. Pol III: RNA polymerase A-B box promoter; Shaded box: Shared sequence with 7SL. Size in brackets indicates the size of the core repetitive element. 24 [119]. The human Alu and the murine B l SINE differ from the fRNA-like SINEs as they contain a sequence closely related to that of the signal recognition particle 7SL RNA (Figure 7b). Termination of RNA polymerase III transcription requires an oligothymidylate tract. SINEs lack this tract, and so transcription terminates at a random oligothymidylate tract downstream of the SINE, resulting in SINEs with variable 3' ends [117]. Humans and mice share the MIR/MIR3 SINEs and have similar 7SL derived SINEs, referred to as Alu SINEs in humans and B l SINEs in mice. Mice also have three additional SINEs, the B2 SINE (which exists in low copy number in humans [120]), ID SINEs, and B4 SINEs (similar to a B l / ID SINE fusion). In humans, only the Alu SINE is considered to be active, whereas in mice, at least the B l , B2, and ID SINEs are active. The rest of this review will focus on SINEs, specifically the murine B l and B2 SINEs. SINE replication SINE replication is not fully characterized. The current model starts with RNA polymerase III transcription of the D N A SINE. The SINE R N A is then transported to the cytoplasm where it binds to a LINE reverse transcriptase/endonuclease protein. This complex is then transported back to the nucleus, where the endonuclease induces a staggered double stranded cut. The free 3' OH is used to prime reverse transcription of the SINE RNA. Host cell D N A repair mechanisms and D N A ligase repair the nicks and fill in the missing sequence [117, 121]. The replication scheme for SINEs is illustrated in Figure 8. 25 Figure 8: Current model of SINE replication. 1. SINE transcription. 2. SINE RNA transport to the cytoplasm. 3. Translation of LINE reverse transcriptase (RT) and endonuclease (EN). 4. SINE binds RT/EN. 5. SINE transport to nucleus. 6. Endonuclease induced staggered break. 7. 3 'OH prime reverse transcription. 8. DNA repair. Pol III: RNA polymerse III. 26 SINE transcription SINEs contain a type I IA-B box promoter that is recognized by R N A polymerase III. Transcription of SINEs is, therefore, dependent on the presence of a host of basal RNA polymerase III transcription factors. The process of R N A polymerase III transcriptional initiation at type II promoters and subsequent elongation is complicated and still under investigation. This section will attempt to provide a basic view of what is currently understood based on several excellent reviews [122-124]. Initiation of R N A polymerase III transcription seems to occur in a stepwise dependent manner, at least in vitro. A protein complex termed TFIIIC2 first recognizes the type II promoter. This complex is comprised of five proteins: TFIIIC220, 110, 102, 90 and 63. Together, these proteins recognize the promoter A and B boxes and recruit a second complex called TFIIIB (another complex, TFIIIC1, is also recruited but it is not well characterized). TFIIIB is composed of three proteins: TBP (TATA binding protein), BRF1 (TFIIB related protein), and Bdp (B double prime protein). This second complex functions to open the helix and, with the help of TFIIIC2, recruit the R N A polymerase complex (Figure 9). Transcription then occurs without elongation factors and the TFIIIC and TFIIIB complexes remain associated with the promoter. Termination occurs when the polymerase reaches an oligothymidylate tract. Controlling SINE expression With the large number of SINEs present in a cell's genome, it becomes essential for the host to control SINE expression or risk serious alterations to its genome. 27 iiiiiiiiiiiii M ^ ^ T A Box B Box AS TFIIIC2 •TFinC2-RNA Polymerase IE Transcription Figure 9: Stepwise assembly of RNA polymerase III transcription factors. TFIIIC2 recognizes the type II promoter and recruits TFIIIB. TFIIIB recruits R N A polymerase and opens the helix. B" : B double prime protein; TBP: T A T A binding protein; BRF: TFIIB related factor; The A and B boxes are components of the R N A polymerase III type II promoter. Dashed arrows indicate additional protein-protein interactions. 28 Currently there are several theories on how the cell accomplishes this. One possibility is that the SINE D N A becomes methylated, leading to re-arrangement of the chromatin environment and suppression of transcription [125-127]. However, not all SINEs contain large numbers of methylation sites and some of the methylation repression (of transcription) machinery doesn't seem to repress SINE transcription [128]. Another possibility is R N A interference, by which transcribed SINE R N A can be identified and destroyed before it integrates back into the genome [129]. Finally, at the individual SINE level, the presence of cis elements, chromatin context and individual promoter sequences will all affect transcription [126]. 29 Materials and Methods 1. LA9 cell culture Adherent L A 9 mouse fibroblasts [130] were cultured in low glucose D M E M (Gibco BRL) supplemented with 5% fetal bovine serum (Gibco BRL) and 25 mM HEPES (Sigma) on Falcon #3003 100mm plates at 37°C, 5% CO2 in a humid environment. Cells were routinely passaged by the removal of spent media, washed twice with 5ml PBS followed by addition of 1ml trypsin-EDTA (Gibco BRL) for 2 minutes at 37°C. The cells were resuspended in fresh media and routinely diluted 1:10 with D M E M containing 5% FBS and 25 mM HEPES. Cells were counted using an hemocytometer. 2. M V M production M V M p was cultured following a modified protocol of Tattersall et al., 1976 [33]. LA9 cells were grown in suspension in high glucose D M E M enriched with nucleotides (Gibco BRL), 5% FBS, 25mM HEPES in an environment of 5%C0 2 , 37°C to a concentration between ~ 1 x 105 to 5 x 105 cells/ml, diluting as necessary. Cells were infected at low MOI (typically 0.01 to 0.001 pfu/cell) and monitored for cytopathic effects and cessation of growth. Cells were pelleted, washed with TNE buffer, taken up in low salt buffer (50 mM Tris"Cl pH 8.7, 0.5mM EDTA, 0.1 MPMSF) and lysed by homogenization (VirTis 45") and sonication (Branson Sonifier 250). Cellular debris was pelleted and discarded. Virus particles were precipitated from the supernatant with 25 mM CaCh and resuspended in viral uptake buffer (50 m M Tris-Cl pH 8.7, 20 mM EDTA) with gentle sonication. The virus preparation was purified by centrifugation in a 30 CsCl gradient (SW-41, 28,000 rpm for 20 hr.) and full virus particles (p = 1.41 g/ml) were collected through the side o f the tube using a syringe and subsequently dialyzed against TE pH 8.7. Host cell genomic D N A was removed from the virus preparation by addition o f micrococcal nuclease (Pharmacia) to 20 ug (800 U)/ml and CaCb (to 5 mM) , and full virus particles were repurified through another round of CsCl gradient centrifugation and dialysis. Virus preparations were titred as described below. 3. Virus titration M V M virus stock was titred using a plaque assay [131]. L A 9 cells were seeded at 5 x 10 4 — 1 x 105 cells in 60 mm plates and infected with serially diluted viral stock in viral uptake buffer (50 m M Tris-Cl pH 8.7, 20 m M EDTA). After one hour to allow attachment o f the virus to cells, the virus was removed by suction and cells covered with a mixture of 7 ml of 0.75 % low melting point agarose (Gibco-BRL) / D M E M / 1 % Tryptose Phosphate (Gibco-BRL) / 5% FBS/ 20 ug/ml gentamicin (Gibco-BRL). After the agarose solidified at room temperature, the cells were grown for 5 days at 37°C, 5% CO2 at which point they were fixed by addition o f 10% formaldehyde on top o f the agarose layer for 30 minutes. The agarose/formaldehyde was removed and the cells stained with 0.3%>methylene blue in 10%> formaldehyde stain for 30 minutes. Excess dye was washed off, and the plates were allowed to dry. Plaques were then scored and the virus titre calculated. 31 4. M V M infection 100 mm plates containing unsynchronized L A 9 cells were typically infected at a MOI of 5-10 pfu/cell in 1 ml D M E M , 0.1% FBS, 10 mM HEPES at 37°C, 5% C 0 2 for 1 hour, rocking every 30 minutes. Virus containing media was then replaced with 10 ml of regular media. Serum starved synchronized cells were infected by direct addition of virus to the low serum media, followed by rocking to mix the virus and an overnight incubation at 37°C, 5% C 0 2 . 5. Cell synchronization L A 9 cells were synchronized by serum starvation [47]. 100 mm plates containing 1 x 105 cells were washed 3 times with M E M (no serum), overlayed with MEM/0.5% FBS/25 mM HEPES, and incubated at 37°C/5% C 0 2 for 4.5 days. If cells were to be infected, virus was added directly to media 12 hours before release. Cells were released from the cell cycle block by replacing the media with DMEM/20% FBS/25 m M HEPES. Cell synchrony was assessed by flow cytometry. LA9 cells were pelleted, washed with PBS, fixed with ice cold 70% ethanol for 30 minutes and then suspended in Vindelov's stain [5 mM Tris-Cl pH 8.0, 5 mM NaCl, 5 ug/ml RNase A, 0.05 mg/ml Propidium iodide (Molecular Probes, Inc), and 0.05% NP-40] at 4°C for 30 minutes. Cells were sorted on a BD FACScan instrument and analyzed using ModFitLT 2.0 software. 32 6. Transfection of L A 9 cells 60mm dishes o f LA9 cells (1.5 x 10 6 cells) were transfected with various plasmids using lipofectamine plus (Invitrogen) following the manufacturer's instructions, using a solution o f 8 ug plasmid DNA, 80 ul Plus reagent, and 20 ul o f lipofectamine in D M E M lacking FBS. After 3 hours, medium with FBS was added back to the cells and the cells were harvested at the various times indicated. 7. Differential Display RT/PCR LA9 cells were infected at an M O I o f 5 as described above and harvested at 12, 24, & 36 hr. P.I. As a control, L A 9 cells were harvested at 24 h after mock infection (same infection procedure except no virus). Cells were directly lysed on the plate by the addition o f 1ml TRIzol reagent (Gibco/BRL) per 100 mm dish. RNA was isolated according to the TRIzol procedure. RNA concentration was measured by absorbance at 260 nm in a Pharmacia Ultrospec 3000 spectrophotometer. To remove any residual DNA, RNA samples were further treated with DNase I (GenHunter) and purified by Phenol/Chloroform extraction. Differential Display RT/PCR was carried out as described in the GenHunter's RNAimage kit. Briefly, first strand D N A synthesis was carried out using 2 \ig RNA (or water to control for D N A contamination), M M L V and either H-Ti i A , H-Ti iC or H-Ti iG primers. Second strand synthesis and subsequent PCR amplification was carried out using the above reverse transcribed RNA, Taq D N A polymerase (Perkin Elmer), 25 u M dNTPs, one o f the three H-Ti ]N primers, one of the 8 arbitrary primers (H-AP1 to H-AP8, see Appendix 2) and 2 uCi a- 3 3 P dATP. This was repeated twice for every possible H-Tn/H-AP primer combination for a total o f 24 primer 33 combinations. D N A products were amplified in a Perkin Elmer 2400 Thermocycler through 40 cycles o f 94°C for 30 seconds, 40°C for 2 minutes and 72°C for 30 seconds followed by one last elongation step at 72°C for 5 minutes. The resulting PCR products were separated in a 6% denaturing polyacrylamide sequencing gel for approximately 2 hours at 32 watts in TBE buffer and visualized by exposure to Kodak Biomax X-ray film. Only bands that showed a reproducible change in two or more o f the infected samples as compared to the control sample were analyzed further. These bands were cut out from the gel and the D N A eluted. This was accomplished by soaking the gel slice in water for five minutes, pelleting the debris by centrifugation, and precipitating any D N A present in the supernatant by addition o f sodium acetate and 95% ethanol along with 50 pg glycogen (Boehringer Mannheim) as a carrier. The D N A present in the slice was re-amplified by another round of PCR using the same primer set. The PCR conditions were the same except that the radiolabel was omitted and the dNTP concentration was increased to 250 | i M . The resulting D N A was separated on a 1.5% agarose gel and imaged by ethidium bromide staining. The correct size D N A product was cut out o f the gel and purified using the MERmaid low molecular weight D N A purification kit (Bio-101) as per the manufacturer's instructions. The resulting D N A was cloned into the plasmid TOPO-TA pCR2.1 (Invitrogen), transformed into TOP 10 E. coli cells, plated on L B ampicillin (75 ug/ml), and the resultant clones were screened for blue/white colony production following the manufacturers instructions. Plasmid D N A was isolated using the mini-prep protocol as described below and digested with EcoRI (Gibco-BRL) to confirm the presence and size o f the insert. To compare clones, T ladder sequencing (described below) was conducted on inserts of the correct size using M13R-24 and 34 M13F-20 primers. To confirm the differential display results, northern blot analysis was conducted on total R N A isolated from a second infection using probes generated from the most abundant clones isolated. Clones that were confirmed to show altered expression patterns were sequenced completely. Sequence data was used to search the NCBI databases using B L A S T [132]. 8. Plasmid DNA preparation Small scale plasmid preparation from bacteria grown in L B medium was carried out by either the alkaline lysis or boiler lysis method [133]. Plasmid DNAs to be machine sequenced (PE ABI prism 310 genetic analyzer) were further purified by Qiagen miniprep columns. Large-scale plasmid preparations were carried out by alkaline lysis/polyethylene glycol precipitation [133]. Plasmids to be used in transfection experiments were further purified using a Qiagen plasmid maxi kit. A list of plasmids used in this thesis can be found in Appendix 1. 9. Transformations/bacterial growth Competent DHlOb E. coli were transformed by the heat shock method [134] and plated on to L B agar plates containing 75 u.g/ml ampicillin and incubated overnight at 37°C. 10. Sequence determination Plasmids were routinely sequenced by either manual sequencing using a T7 Sequenase 2.0 kit (Amersham) using 3 2 P label (as per the manufacturer's instructions) or 35 by automated sequencing using an ABI prism 310 sequencer. Sequence data were analyzed in EditView 1.0.1 (ABI Biosystems) and stored in D N A Strider 1.1. A list of primers used in sequencing can be found in Appendix 2. 1 1 . Probe generation D N A probes used in Northern analysis were either labeled by the random primer method or by PCR. Random primed labeling was done using a method similar to that described in the USB random primed labeling kit. 25 ng of purified D N A was added to a solution containing 25 u M each dCTP/dGTP/dTTP, 0.0175 A 2 6 o units random hexamers, 1.7mM M g C l 2 , 200 mM HEPES pH 6.6, 0.5 mM 2-mercaptoethanol, 10 u M EDTA, 400 ug/ml BSA, 16.6 mM Tris-Cl pH 8.0, 50 uCi a 3 2 P dATP and 2 units of D N A polymerase, Klenow fragment. The reaction was incubated at 37°C for thirty minutes and terminated by heating at 65°C for 10 minutes. PCR amplification labeling was used to label various D N A fragments identified by differential display. A PCR cocktail was prepared as described for the differential display reaction except that the dNTP concentration was adjusted to 20 \xM, cc- 3 2P dATP was added to a concentration of 1 uM (60ixCi), and the concentration of template D N A was set at 5 pg/jj.1. A set of duplicate reactions containing either water instead of template (negative control) or unlabelled dATP instead of a - 3 2 P dATP (positive control to confirm amplification) was also carried out. A l l probes were purified using Sephadex G-50 spin columns [133], counted using a Beckman coulter LS6000IC scintillation counter and stored at -20°C until needed. 36 12. RNA isolation Unless otherwise stated, total RNA was routinely isolated using an RNeasy mini kit (Qiagen) as per the manufacturer's instructions. An on-column treatment with DNase I (Qiagen) was also performed to remove any residual DNA. The resulting R N A was collected from the column in two 30 pi washes. R N A concentration was determined by measuring the A260 of samples, diluted in water, in an Ultraspec 3000 spectrophotometer (Pharmacia). If needed, RNA was precipitated using ethanol and taken up in a small volume of water. R N A samples were aliquoted and stored at -80°C until needed. 13. DNA dot blot analysis Plasmid D N A preparation and blotting was done according to Bio-Rad's Bio-Dot microfiltration apparatus instructions. 200 ug of each plasmid D N A was linearized by restriction digestion, extracted with phenol/chloroform, precipitated with 3 M sodium acetate/ethanol, dissolved in 100 ul water, and quantified by A260- Linearized plasmids were diluted to 62.5 ng/ul, denatured by addition of 2 M NaOH/50 m M EDTA (to 0.4 M NaOH/10 mM EDTA) followed by a 10 minute incubation at 95°C, and neutralized by addition of ice cold 2 M ammonium acetate (to 1 M). Blotting onto nitrocellose (Bio-Rad) was accomplished using Bio-Rad's Bio-Dot microfiltration (96 well) apparatus. A 9 x 12 cm nitrocellose membrane was soaked briefly in 6X SSC, assembled into the blotting apparatus and rehydrated once with 100 ul 2X SSC per well using vacuum to pull the fluid through the membrane. Plasmids, 5 ug (200 ul) per well, were blotted onto the membrane and washed once with 25 ml 2X SSC. The membrane was dried overnight 37 at room temperature, baked at 80°C for 2 hours, and finally cut into strips for use in hybridization with the nuclear run-on probes. 14. Northern blot hybridization RNA blotting and hybridization was done by a modified protocol from Molecular Cloning: A Laboratory Manual [133]. 15 ug of total R N A per sample was mixed with formamide loading buffer to a final concentration of 26 mM MOPS, 0.6 mM EDTA, 6.5 m M sodium acetate, 64% formamide, 14% formaldehyde and bromophenol blue and xylene cyanol dyes. Samples were heated at 65°C for 10 minutes, and separated on a 5.5" by 4.25" 1.25% agarose/6% formaldehyde gel in MOPS running buffer at 70V. To check the integrity of the RNA, ethidium bromide was also added to the gel to visualize rRNAs. The gel was washed 5 times with DEPC-treated water and once in 10X SSC for 30 minutes. The R N A was transferred to a Hybond N+ positively charged nylon membrane (Amersham Pharmacia) in 10X SSC via the capillary transfer method over a period of 3 days. Subsequently, the membrane was washed in 2X SSC and exposed to U V light (Biorad GS Gene Linker, 150 mJ setting) to cross-link the R N A to the membrane. Blots were dried at 80°C, briefly examined under U V to examine the efficiency of transfer and stored at 4°C until needed. Membranes were soaked briefly in 2X SSC and then pre-hybridized with 5-10 ml ExpressHyb (Clontech) solution and 100 ug denatured salmon sperm D N A for 1 h at 68°C. Radiolabeled probe (see section 11) (denatured at 95°C for 10 minutes) was mixed with an additional 5 ml ExpressHyb solution and then added to the pre-hybridization/membrane mix. Blots were hybridized with probe at 68°C for one hour and 38 then washed 2 times with 2XSSC/0.05% SDS at room temperature for 15 minutes each. Washing efficiency was assessed by a brief exposure to a storage phosphor screen and examined in a phosphorimager. If warranted, blots were washed a further 2 times with 0.1X SSC/0.1% SDS at 50°C for 20 minutes. Blots were then visualized using a phosphorimager with various exposure times. 15. Autoradiography and band quantification A l l band quantification and most autoradiography was done using a phosphorimager. The gel or membrane was exposed to a phosphorimager plate for varying amounts of time, depending on signal strength. Initially, the plates were scanned on a Molecular Dynamics phosphorimager SI and the acquired data was analyzed with IP lab gel 1.5. Later, a Molecular Dynamics Typhoon 8600 phosphorimager was used with Typhoon scanner control 1.0 to acquire the data. Data was analyzed using ImageQuant 5.2 and Microsoft Excel. 16. Protein isolation LA9 cells were collected using a cell scraper and transferred with medium to a 15 ml conical tube. Cells were pelleted and washed twice with ice cold PBS. Cell pellets were then lysed with protein sample buffer (10% glycerol, 1.5% SDS, 62.5 m M Tris-Cl pH 6.2, 2.5%) 2-mercaptoethanol, and bromophenol blue) and boiled for 5 minutes. Samples were stored at -20°C until needed. 17. Western blot hybridization 39 Protein samples were separated by electrophoresis on a SDS-polyacrylamide (6-15%) gel (8.9 cm by 10.2 cm) at 80-120V. Protein samples were then transferred to PVDF membrane (Pall) by semi-dry transfer using a Biorad trans-blot SD apparatus. Transfer was accomplished in transfer buffer (39mM glycine, 48 m M Tris-Cl, 0.037% SDS and 20%) methanol) for approximately 20-30 minutes at 10-12 volts. When transferring proteins larger than 150 kDa, transfer buffer without methanol was used and the transfer time was increased to 45-60 minutes. Blots were dried at room temperature. Antibody binding was as follows: PVDF membranes were soaked briefly in methanol, washed with PBS/0.05% Tween 20 (PBS-T), blocked for 1 hour with BLOTTO (5%> powdered milk, weight by volume, in PBS-T) at room temperature, followed by several washes with PBS-T. The primary antibody (either a rabbit or a mouse antibody) was diluted in BLOTTO and incubated with the membrane with shaking at room temperature for an hour followed by several more washes with PBS-T. Either peroxidase-conjugated donkey anti-rabbit IgG secondary antibody at a dilution of 1:50,000 or peroxidase-conjugated goat anti-mouse IgG secondary antibody at a dilution of 1:5,000 was diluted in Blotto and incubated with the membrane for one hour at room temperature, followed by another several washes of PBS-T. The secondary antibody was visualized by chemiluminescence with either ECL or E C L Plus kits (Amersham-Pharmacia) according to the manufacturer's instructions. This signal was detected either with Hyperfilm E C L film (Amersham-Pharmacia) or using a FluroS max scanner (Biorad) with a clear filter, 50 mm lens f/1.4 and no excitation (courtesy of the Steve Pelech lab). Multiple exposure times were collected for each method. Films were developed using a M35 X O M A T processor (Kodak). Data from the FluorS max scanner 40 was processed, analyzed, and quantified using Quantity one software (Biorad). A list of antibodies used in this thesis can be found in Appendix 3. 18. Primer extension analysis Primers WPW25, 26 and 27 (Appendix 2) were 5' end-labeled with y- 3 2P ATP using T4 polynucleotide kinase and the D N A precipitated by addition of 95% ethanol, ammonium acetate (to 150 mM) with glycogen as a carrier. Radiolabeled primers were hybridized to 1 pg of total RNA at 60°C for 5 minutes and chilled on ice. First strand synthesis was accomplished by the addition of M M L V (Gibco BRL) , dNTPs, DTT and buffer, followed by incubation at 37°C for 45 minutes. The resulting R N A / D N A hybrids were then denatured by heating at 95°C for five minutes and then separated on a 5% urea polyacrylamide mini-gel (Triezenberg, 1992). Gels were dried (Biorad model 583 gel drier) and bands imaged and quantified using a phosphorimager. To accurately determine the size of the primer extension products, some of the labeled D N A was loaded onto a 6%> urea polyacrylamide sequencing gel alongside a pUC19 sequencing ladder. 19. Clontech array analysis Total R N A was isolated from synchronized or unsynchronized infected or mock-infected cells and reverse transcribed in the presence of 3 2 P label using the Atlas Pure total RNA labeling system (Clontech). To check for genomic D N A contamination, PCR was used to amplify the (3-actin gene. RT/PCR was also used to detect the mRNA for NS-1 to confirm infection. Radiolabeled R N A hybridized to Clontech's Atlas mouse 1.2 arrays (a cDNA array), as per the manufacturer's instructions. Arrays were washed four 41 times with 2X SSC/1% SDS at 68°C, once in 0.1X SSC/0.5% SDS at 68°C and once in 2X SSC at room temperature. Arrays were wrapped in saran wrap and visualized using a phosphorimager. The images were saved as tiff files and analyzed using Atlaslmage 1.01a software, which can be used to locate, identify and calculate signals for each cDNA spot. Background was determined by averaging the signal in the background regions of the array. Background was subtracted from each signal. Initially, the arrays were normalized to a series of housekeeping genes, specifically cytoplasmic (3-actin and 40S ribosomal protein S29. However, subsequent re-analysis of the data suggested that this method was biased. Hence, the arrays were normalized by the following method: First, spots with signals less than 500 counts were discarded from the data set. The infected/mock-infected ratio was calculated for the remaining spots and these data were plotted versus gene identity. It is expected that the majority of spots should have a ratio of approximately one, so a single normalization factor was estimated for the control array from the above plot such that the majority of the spots fell into this range (Figure 10). A l l control signals were then multiplied by this factor and the ratios recalculated. 20. Affymetrix microarray analysis Total RNA was isolated as above (section 12) from mock infected and M V M p infected LA9 cells (unsynchronized) at 36 hours post infection. 5 ug of R N A was shipped on dry ice to the Genome Sciences Center at the British Columbia Cancer 42 CN CN CO DC 4 3.5 3 2.5 2 1.5 1 0.5 0 Array I: Raw ratios «• • • A * 20 40 60 80 Gene number 100 120 2 CN CN ro 0£ 4 3.5 3 2.5 2 1.5 1 0.5 0 Array I: Normalized ratios • •—+ o 20 40 60 80 Gene number 100 120 Figure 10: Normalizing the array data: A : Array I (24 hr post block synchronized mock vs infected) ratios versus gene number. B: Normalized array I ratios versus gene number. The two arrays were normalized by multiplying all mock-infected signals by 0.6. 43 Agency where the lab staff prepared the probes for hybridization with Affymetrix microarrays using Affymetrix's instructions. Briefly, the sample was reverse transcribed into DNA, used to generate biotinylated complementary RNA, fragmented, and then hybridized an Affymetrix microarray chip. Two types of arrays were hybridized with the above-prepared samples. First, an Affymetrix Test 3 microarray was used. This array contains multiple housekeeping genes and is used to determine i f the RNA sample is suitable for array hybridization. The remaining prepared R N A sample was hybridized to Affymetrix's U74Av2 microarray. This consists of the approximately 6000 genes found in the mouse UniGene database that have been characterized and an additional 6000 expressed sequence tags (ESTs) for a total of 12,000 genes per array. As the Affymetrix system uses a one-fluor system, each sample was hybridized to a separate array. After staining with streptavidin phycoerythrin and washing, the chip was scanned and the data analyzed in Microarray Suite 5.0. Data analysis and manipulation will be discussed further in the results section. 21. Preparation of nuclei for nuclear run-on assays Nuclei were isolated from LA9 cells using a modified detergent lysis method. LA9 cells were grown in suspension, pelleted, washed twice with PBS and then re-suspended in NP-40 lysis buffer (0.5% NP-40, 10 mM Tris-Cl pH 7.5, 10 m M NaCl, and 3 mM MgCb) with gentle vortexing. Cells were allowed to lyse on ice for 5 minutes and were then pelleted by centrifugation at 4°C. These steps were repeated two more times. The nuclei were examined under a phase contrast microscope to make sure that the cells lysed and that the nuclei remained intact. The resulting nuclei were taken up in nuclei storage 44 buffer (40% glycerol, 50 mM Tris-Cl pH 8.3, 5 mM M g C l 2 and 0.1 m M EDTA) and stored in liquid nitrogen until needed. 22. Nuclear Run-on assays Nuclear run-on assays were done using a significantly modified protocol from that described in Current Protocols in Molecular Biology [135]. An aliquot of ~ 2 x 107 to 2.5 x 107 nuclei in 100 p,l glycerol storage buffer was mixed with 100 ul of 2X reaction buffer (10 m M Tris-Cl, 5 mM M g C l 2 , and 300 mM KC1) containing 100 u M each ATP, GTP, CTP and 5 mM DTT and 250 uCi a- 3 2P UTP in a 15ml conical tube. Samples were incubated in a 30°C shaking water bath for 30 minutes. Nuclei were lysed by addition of 4 ml QRL1 (Qiagen) with 2-mercaptoethanol (Sigma) and total R N A isolated using Qiagen's R N A / D N A maxi kit using the manufacturer's instructions with the following modifications: D N A in the QRL1 lysed nuclei mixture was sheared by pipetting through a 18 gage needle 10 times and the mixture transferred to a 50 ml Sarstedt conical centrifuge tube. 4 ml of QRV1 was added and the entire mixture centrifuged at 8,000 R P M for 35 minutes in a JA-12 rotor. The supernatant was transferred to a silanized (Sigmacote, Sigma) 30 ml glass corex tube, mixed with 6.4 ml of ice cold isopropanol, centrifuged at 11,000 rpm in a JA-20 rotor for 30 minutes, and the pellet collected and washed with 4 ml 70% ethanol (11,000 rpm, JA-20, 15 minutes). The pellet was resuspended in QRL1 solution (with heating and vortexing) and loaded onto a Qiagen maxi kit column. The column was washed with 30 ml of QRW solution and followed by addition of 14 ml of QRU at 45°C to elute the R N A into another silanized 30ml corex tube. The RNA was precipitated with 14 ml isopropanol, 45 centrifuged and washed as above. The RNA pellet was dissolved in 2 ml of water. 5 ul of this mixture was counted in a scintillation counter to normalize the mock and infected samples. The remaining sample was mixed with 2 ml 2X TES solution with NaCl (20 mM TES pH 7.4, 20 m M EDTA, 0.4% SDS, 0.6 M NaCl), heated to 65°C, and hybridized to nitrocellulose dot blots at 65°C for 36 hours. The blots were washed twice in 2X SSC at 65°C for an hour each, followed by an optional 30 minute incubation with 10 ml 2X SSC with 10 ug/ml RNase A at 37°C, and a one hour wash at 37°C. Signal detection and quantification was achieved using a phosphorimager as described in 14 (above). 23. Kinexus protein kinase screen Protein samples for the Kinexus protein kinase screen were prepared following instructions from Kinexus. Briefly, pellets of synchronized adherent infected or mock-infected LA9 cells were taken up in Kinexus lysis buffer (0.5% NP-40, 20 mM MOPS, 5 mM EGTA, 2 mM EDTA, 5 mM NaF, 40 mM [3-glycerophosphate, 1 m M sodium orthovanadate, 1 mM PMSF, 3 mM benzamidine, 5uM pepstatin A, 10 \iM leupeptin), sonicated at a low setting for two rounds of 15 seconds each (Branson Sonifier 250), and centrifuged at 55,000 rpm (100,000 x g) in a T L A 100.4 rotor (Beckman Optima TL untracentrifuge) for 30 minutes at 4°C to pellet the lipid membranes. The supernatant was collected and protein concentration was determined by the B C A method (Pierce's bicinchoninic acid assay) as per the manufacturer's instructions using bovine serum albumin as a standard. Samples were mixed with protein sample buffer, boiled for 5 46 minutes and adjusted to equal concentrations. Approximately 500 ug of each sample was sent to Kinexus for analysis. From Kinexus, tiff images of the western blots and excel files containing the quantification data for each of the bands imaged were received. Initial analysis was done by visual inspection with the guidence of Dr. Steven Pelech (Kinexus). Infected sample data was normalized to control data by calculating a normalization factor based on the top ten expressed proteins in both control and infected samples. 24. Reverse transcription and PCR amplifications To test for genomic D N A contamination of R N A samples, the R N A samples were screened for the presence (3-actin DNA. 1 pg total R N A was added to a solution containing 1U Taq D N A polymerase and buffer (Perkin Elmer), 200 u M dNTPs, 3mM M g C l 2 , and 0.5 u M PW10 (skeletal B-actin, nucleotides 987 to 1004) and PW11 (skeletal R-actin, nucleotides 461 to 478) primers. If detectable D N A contamination is present this PCR assay should yield a 546 bp product. Amplification was accomplished through an initial denaturation step of 94°C for 5 min., followed by 35 cycles of 30 sec. at 94°C, 30 sec. at 55°C and 30 sec. at 72°C, and then one final extension at 72°C for 7 min. The resulting products were separated on a 1.5% agarose gel. Control tubes contained water or 260, 26, or 2.6 ng of LA9 genomic DNA. NS 1 RT/PCR: 1 ug total RNA was reverse transcribed using M M L V Reverse transcriptase (Gibco BRL), 5 u M Ti8 primer, and ImM dNTPs at 37°C for 50 minutes. An aliquot (4 uT) of the above cDNA mix was added to a solution of 200 p M dNTPs, 1U Vent D N A polymerase and buffer (New England Biolabs), 200 ug/ml BSA, and 0.17 u M 47 NS1-415 and NS 1-265 primers. Amplification was accomplished using an initial denaturation step of 94°C for 5 min., followed by 35 cycles of 40 sec. at 92°C, 40 sec. at 55°C and 40 sec. at 75°C, and then one final extension at 75°C for 5 min. The resulting products were separated on a 1.5% agarose gel in I X T A E buffer. As controls, water and p C M V N S l [43], a plasmid containing the NS1 gene, were used. A list of primers used in this thesis can be found in Appendix 2. 48 Results 1. Differential Display To begin to profiling changes in gene expression during a cytopathic infection of murine LA9 fibroblasts with M V M , differential display (DD), invented by Peng Liang [136], was used. This is an RT/PCR based technique that amplifies a set of semi-randomly chosen cDNAs (500-1000 per PCR reaction) based on primer selection. Differential display is based on semi-conservative PCR amplification. Therefore, transcripts expressed at low levels will have a corresponding lower number of PCR products whereas transcripts expressed at high levels will have a corresponding higher number of PCR products. By comparing the levels of these PCR products between uninfected and infected LA9 cells, genes that exhibit altered expression levels can be identified. As illustrated in Figure 11, in this technique R N A is first reverse transcribed to cDNA with an anchored poly T primer (VTi i where V = A , C, or G). This selects a specific subset of cDNAs to be examined in further reactions. This pool of cDNAs is then amplified through PCR using the above anchored primer and a second arbitrarily selected primer of 13 nucleotides (the arbitrary primer). Any arbitrary primer can be used provided there is no secondary structure. Using a low annealing temperature of 40°C, the arbitrary primer acts as an 8-9mer in the initial annealing step and binds to multiple cDNAs, resulting in the selection of a smaller, more manageable pool of transcripts containing approximately 50-100 products [136]. For the most part, this pool of D N A will represent the 3' ends of various cDNAs ranging in size from <100 nucleotides to ~500 nucleotides, due to the use of the anchored polyT primer and the 49 m R N A f r o m m o c k i n f e c t e d c e l l s Isolate mRNA A A A A A A Convert to cDNA Use PCR to amplify products Separate A A A A A A T T T T T T mRNA from MVM- infected cells A A A A A A 1 A A A A A A T T T T T T Figure 11: Differential Display. RNA is first reverse transcribed into cDNA through the use of reverse transcriptase and an anchored oligo dT primer. A smaller sub-set of this pool of cDNAs is amplified by PCR using an arbitrary selected 13 nt primer and a low annealing temperature in the presence of radioactive label. This set of cDNAs is then resolved on a polyacrylamide gel and quantified. 50 short elongation time. Amplified products can be separated on a denaturing polyacrylamide gel and quantified. Figure 12 shows a typical differential display gel. Differential display PCR is not always 100% reproducible (see Figure 12a) due to pipetting errors, RNA quantification errors, and inherent errors in PCR amplification. To compensate for this, three different times post-infection were examined: 12, 24, and 36 h P.I. Only bands that showed distinct trends over the three time points were selected for further analysis. To further reduce the risk of false positives, two separate PCR reactions for each sample were done. Water was used as a negative control. A total of 24 different primer pair combinations using 8 arbitrary and 3 anchored primers were screened. This detects approximately 15% of the estimated 15,000 actively transcribed genes within the cell [137]. Twenty-five bands were identified as being altered in expression level, with 10 of them being < 150 bp (Table 2). The approximate size was determined by position on the differential display gel and by the size of the resulting PCR product. As recommended by Genhunter, bands with a size below 150 nt were not studied further for two reasons. First, the sequence in these bands would be predominately 3' untranslated sequence (3' UTR). At the time of these studies, the mouse genome had not been sequenced, and it would have been difficult to identify the mRNA species. Second, the 3' UTR sequence often can contain repetitive elements, which would also prevent identification. Candidate bands were cut out of the gel from the lane giving the highest signal, re-amplified by another round of PCR (in some cases two), purified by agarose gel 51 A 7 A 1 U 7 A 2 L Figure 12: Legend on next page. 52 B Mock Infected 24 24 12 12 24 24 36 36 Hrs. post-infection Infected Infected 24 12 24 36 24 12 24 36 4C2M L l L I N E Figure 12: Differential display polyacrylamide gels. A . The PCR products from the H-AP-7 and H - A T n primed reaction were separated on a 6% denaturing polyacrylamide sequencing gel. Templates were AT, , primed cDNAs as described. B. The differential display bands 2A1, 3G3M2, 2G1U, and 4C2M. Control: Mock infected LA9 cells harvested 24 hr after mock treatment., 12,24, & 36 hr. P.I.: unsynchronized LA9 cells infected with M V M p and harvested at the times indicated. Each sample (two adjacent lanes) represents the results of two separate PCR reactions. This was repeated for each of the 24 different primer combinations. Bands that were further studied are indicated. 53 Band Id , Relative < , v Size (bp) .Intensity - 'Am pitied < Reverse Cloned\ Northern Northern ~& Change Identity- x- « ' C2B1 50o" "+/++' r" V" Can't amo .'N/A ' '"Pa':""" T * T 7A1U 500 ++++ Yes Yes Yes Yes Yes NS-1 4 C 1 U 500 ++ ' • Can't a m p - N/A ; ••;iiN/A' ;:-. . N A N A •"v"' i.; : ;' :. N/A . , ' 7 C 1 U 500 +++ Yes Yes No Yes No N/A 3G1U " 5 0 0 \ ++++ Yes Yes ' . / .- 'Yes, ' ' " / Yes • Yes NS-1 :" "'•"'.•'." 4 C 2 M 400 +++ Yes Yes Attempted Yes Yes L1 Lhe C 2 3 2 350 :•• +/++ Yes ' Yes No ; ' : ;v/"' Yes not reoro. Mt. NAD depend. D/C . 3G4M3 250 +++ Yes Yes No Yes not repro. Unknown E S T (sim to FGF receptor activating protein) 3G3M2 . 250 +++ : . Yes" ._ - Yes .. . N o Yes ' Yes B1 S INE, -4 C 3 M 250 +++ Yes Yes Attempted Yes No N/A 3G2M1 ' 250 + No No, No - No 9 2G1U 200 +++ Yes Yes No Yes Yes B 1 / B 2 S N E 5C2B1 200 +.++ •Yes Yes,,, . No ' Yes . No, ' -'. N/A .. 2A1 200 +++ Yes Yes No Yes Yes B 2 S I N E 7A2L : 200 . +++ Yes : Yes No - - Yes No Siqnal 1G1B 150 +++ - _ _ 5C1M. . . . . 1 5 0 . ++ i _ . . . 4 i •'< • 5 C 3 B 2 150 - _ _ -3A1U , . -150 - . ' - „ 1 .5^4, . . > . . . " , .... 3A2B 150 - _ 4 C 4 L 100 " -'..'.'•>¥.'l ~ f , s . , ' Attempted No ' • \ * 7G1U 100 - _ 7G2B 100 - T < . 4A2B1 100 - _ _ 4A3B2 .100 - ,'" ' "-' u J— ... I . . u .. .. Table 2: Altered bands detected by differential display. This lists the altered bands detected by differential display and the progress in confirming and identifying them. Size is an estimate from the resulting PCR product, change indicates whether the northern or reverse northern confirmed the differential display results. A l l bands that were confirmed to change were found to be up-regulated. N/A: Not applicable. Can't amp: Band could not be amplified after two rounds o f PCR amplification. 54 electrophoresis, and cloned into the pCR2.1-TOPO plasmid. Two bands could not be amplified by PCR. To confirm that the remaining cDNA fragments were altered in their expression levels, reverse northern blots were initially performed as explained in Figure 13. This technique has the advantage of being able to screen multiple clones at one time. This is crucial as a single band from the gel may contain multiple cDNA fragments of the same size and not all of them are necessarily altered in expression. Also, especially in the case of larger bands where there is less resolution between bands, there is always the risk of contamination from bands above or below the band of interest. This technique was successfully used to screen the 7A1U band, shown in Figure 14a. Sequencing analysis (both directions) identified the clone as a fragment of the M V M NS-1 transcript (viral nucleotides 439 to 899). Interestingly, this is in the 5' end of the transcript, spanning the shared sequence of NS1/NS2 and the alternatively spliced intron that is removed to generate NS2 (see Figure 2a). Further analysis of the data led to the conclusion that PCR amplification occurred with the anchored primer binding to an A-rich sequence (5'-T A 5 G A 4 T A A - 3 ' ) present in this region of the transcript. This is illustrated in Figure 14b. Although reverse northern blots were successful in identifying NS-1, they were less useful in confirming other bands. High background, low sensitivity and difficulty with relative changes in signal intensity led to a decision to use northern blots rather than reverse northern blots. To overcome the presence of multiple cDNA fragments present in one band, T ladder sequencing was used to characterize each clone. This is based on the Sanger sequencing method but only the ddTTP reaction is used, resulting in a fingerprint (T ladder) for each clone sequenced. The most abundant clone was then used to make a 55 Amplify and clone band \ Make two filter lifts Probe with radioactive cDNA from control or infected cells Control ^ cDNA Infected cell cDNA Isolate positive clones and Sequence Figure 13: Reverse northern blots. Identical colony lifts containing a clone of the band of interest were probed with radioactively labeled cDNA from either mock or M V M infected LA9 cells. Colonies that showed altered expression were grown up and sequenced. 56 Probed with LA9 cDNA Probed with MVM-infected L A 9 cDNA B 7 A 1 U amplified sequence Spliced sequence to yield R2 transcripts encoding NS2 (A)n R l Transcript 5 ' -TACAACGAGGACG -3 3 ' - A T G T T G C T C C T G C - 5 I I I I I I I ^ A A C G A G G - 3' R l cDNA, HA-AP7 3' - A T T T T T TTTT*> I I I I I I I I I I 5 ' -TTAACTAAAAAGAAAATAAG-3 ' 3 ' - A A T T G A T T T T T C T T T T A T T C - 5 ' HA-T11A Figure 14: The 7A1U band results. A : Reverse northern of clones containing 7A1U (NS-1). B : Schematic identifying the R l transcript sequence cloned from the 7A1U band. The sequence of the two primers used in the initial differential display amplification and the corresponding R l cDNA sequence is also included. 57 probe for the northern blot. Using these techniques, three bands were found to give signals that did not change upon infection and one band that gave no signal. Two more bands gave signals that appeared to show an approximately two-fold change but these changes were not always reproducible. These bands were sequenced and found to be "methylenetetrahydrofolate dehydrogenase (NAD dependent) methenyltetrahydrofolate cyclohydrolase" and an uncharacterized EST with similarity to the fibroblast growth factor receptor activating protein. Analysis of these clones was not pursued further. Northern blot analysis confirmed that the remaining bands were indeed altered upon infection. Band C2B2 (not shown) was found to contain another fragment of M V M NS-1 spanning viral nucleotides 1600 to 1660, which is a NS-1 specific sequence. Again, the anchored primer bound to an A rich sequence upstream of the poly-A tail. The remaining bands contained repetitive elements as detected by the CENSOR web server (Jurka et al., 1996). Band 3G3M2 (Figure 12b) contained a full-length copy of a B l SINE (consensus sequence nucleotides 1 to 131) with the primer binding to the A rich region in the 3' end. Band 2A1 (Figure 12b) contained a fragment of a B2 SINE (consensus sequence nucleotides 32 to 174). Northern blots probed by 3G3M2 and 2A1 are shown in Figure 15. Band 2G1U (Figure 12b) contained a fragment containing both an entire B l SINE (consensus sequence nucleotides 1 to 134) and part of a B2 SINE (consensus sequence nucleotides 19 to 174). Finally, band 4C2M (Figure 12b) contained a fragment of a L I LLNE (consensus sequence nucleotides 2520 to 2635). 58 K » i i . o s " * ^ OS I — 4.4 kb -—2.4 kb -—1.4 kb --240 b p -B l SINE Northern B2 SINE Northern 3G3M2 2A1 Figure 15: Altered expression of selected bands. Northern analysis of RNAs detectable by the most abundant clone from 3G3M2 (Bl SINE) and 2A1 (B2 SINE). Each lane contains 15 ug of total R N A Abbreviations are as follows: C: Mock infected LA9 cells at 24 hr after mock-infection. 12, 24, 36: LA9 cells 12, 24, & 36 hr after infection with M V M p . See Materials and Methods section 14. 59 2. Clontech Macroarray Analysis The high numbers of false positives, the predilection to detect SINEs, the presence of multiple clones in a single band, and the large amount of effort needed for differential display prompted an investigation of alternative means of analyzing global gene expression. One attractive method was the then emerging technique of array analysis. Arrays have been successfully used to screen host gene expression during viral infection for many other viruses, including cytomegalovirus and HIV [138, 139]. Array analysis is advantageous as there is only one known sequence per spot (hence this avoids cloning), and it is rapid and easier to carry out than differential display. Arrays, however, are limited to studying expression of genes present on the array, they often exhibit lower sensitivity (as compared to differential display) and are much more expensive (at least 4-5 years ago when these studies were initiated). Arrays are, in essence, a large series of reverse northern blots. The array itself is simply a support structure containing multiple spots of different sequence corresponding to many different genes of interest. It can be probed with labeled cDNA from either control or experimental (in my case mock or M V M infected LA9 cell cDNA). By comparing intensities of probes annealed to spots between two different arrays one can determine relative changes in gene expression for each of the genes on the array. At the time of these experiments, very few mouse arrays were available so our choices were extremely limited. We decided to use Clontech's Atlas mouse 1.2 arrays. This is a positively charged nylon macroscopic array containing 1176 different cDNA fragments (200 to 600 bp) in single spot format. These types of arrays are probed with radiolabeled cDNAs and require two separate arrays for each analysis. 60 As mentioned previously, M V M can only replicate when cells reach the S/G2 phases [91]. Therefore, at any particular time during M V M infection, individual asynchronized LA9 cells will be at different stages of infection relative to each other. Since this could potentially mask changes in gene expression, it was decided to synchronize LA9 cells by serum starvation prior to infection [47]. To confirm that cells were synchronized, they were fixed and stained with propidium iodide and analyzed by FACS analysis (Figure 16). Propidium iodide stains D N A and is used to determine the relative amounts of D N A in cells and thus the stage of the cell cycle. The majority of L A 9 cells are in G0/Gi immediately after block release and progress in an orderly fashion into S phase around 14 to 16 hours post block, and enter G2 around 20 hours post block. I also confirmed that M V M arrests cells in S/G2 by repeating the above experiment with MVM-infected cells. As can be seen in Figure 17, by 20 hours post block MVM-infected cells appear as a broad peak spanning the S/G2 region with very few cells in G0/Gi as compared to uninfected cells that have a well defined G2 peak and are beginning to cycle back into G 0 / G , . Array studies were conducted as described in the Materials and Methods section 20fs. Briefly, total R N A was isolated from cells, enriched for poly A RNAs (by using biotinylated oligo(dT) and streptavidin coated magnetic beads), and reverse transcribed in the presence of 3 2 P labeled dNTP using a primer set specific to the genes present on the array. As controls, I checked for RNA integrity by visualizing the major ribosomal RNAs, confirmed M V M infection by RT/PCR amplification of NS1 transcripts, and finally confirmed an absence of genomic D N A contamination by PCR amplification of actin D N A (Figure 18). The labeled cDNAs were then hybridized overnight to the 61 Figure 16: Synchronized LA9 cells by serum starvation. L A 9 cells were stained with propidium iodide at various times post block and analyzed by FACS analysis (Materials and Methods Section 5). M l : cells in G 0 / G , , M2: cells in S, and M3: Cells in G 2 phase. FACS analysis of unsynchronized L A 9 cells is provided for a comparison. 62 U Unsynchronized ^ 2 Cells I I | I I 'VtI l l l l 200 400 600 800 1000 FL2-H Mock-infected LA9 cells MVM-infected LA9 cells 12 hr Post Block i i i i i 1 1 1 1 1 1 i 200 400 600 800 1000 FL2-H 200 400 600 800 1000 FL2-H 16 hr Post Block H < I i i 1 1 1 1 1 1 i 400 600 800 1000 FL2-H 16 hr Post Block ' I " * ' * " 1 '• I 1 1 400 600 800 FL2-H T 100 o 00( FL2-H FL2-H Figure 17: Arrest of MVM-infected LA9 cells in S/G 2. Synchronized L A 9 cells were mock or M V M infected, stained with propidium iodide at various times post block and analyzed by FACS analysis (Materials and Methods section 5). M l : cells in G 0 / G , , M2: cells in S, and M3: Cells in G 2 phase. FACS analysis of unsynchronized LA9 cells is provided for a comparison. 63 A RNA Integrity B NS-1 RT/PCR hr post block A f >* 12M 121 24M 241 28s rRNA 18s rRNA NS-1—• hr post block A 12M 121 241 24M hr post block Genomic DNA J ^ H ^ H M M ^ H r H ( S ^ r cd CJfj CN ( N ( N 1 2 3 4 5 6 7 8 9 Actin PCR Actin Figure 18: Legend on next page 64 Figure 18: Array R N A controls. A : R N A integrity analysis. Agarose gel showing the presence of the 28S and 18S rRNAs. B: Confirmation of M V M infection. R N A was reverse transcribed with an oligo T primer and a fragment of NS-1 transcript was amplified by PCR with NS1-265 & NS1-415 primers. C: Absence of genomic D N A in R N A preparation. PCR was used to detect the presence or absence of genomic D N A using actin-specific primers PW10 & 11 (lanes 2-5). Varying amounts of LA9 genomic D N A were used as a positive control (lanes 7-9). 65 Figure 19: A typical Clontech Atlas mouse 1.2 array. A : Mock-infected, synchronized LA9 cells harvested at 24 hr P.B. B : MVM-infected, synchronized LA9 cells harvested at 24 hr P.B. The arrow indicates spot c02n (TTF, transcription termination factor/B2 SINE). 66 arrays, washed several times, and visualized using a phosphorimager. Figure 19 shows a typical array after hybridization. Data analysis was a multi-step process. Phosphorimager data was analyzed using Clontech's Atlasimage 1.01a software, which can be used to locate, identify and calculate signals for each cDNA spot. Background was determined by averaging the signal in the background regions of the array and then was subtracted from each individual signal. Initially, I normalized the arrays to a series of housekeeping genes, specifically cytoplasmic (3-actin and 40S ribosomal protein S29. However, subsequent re-analysis of the data suggested that this method was biased. Hence, the arrays were normalized by the following method. Spots with signals less than 500 counts (after background subtraction) were discarded from the data set, the infected/mock-infected ratio was calculated for the remaining spots and these data were plotted versus gene identity. It is expected that the majority of spots should have a ratio of approximately one, so a single normalization factor was estimated for the control array from the above plot such that the majority of the spots fell into this range (Figure 10, page 43). A l l control signals were then multiplied by this factor and the ratios recalculated. I examined four different array sets: R N A was isolated for two array sets (arrays I and II) comparing mock (control) and infected L A 9 cells at 24 hours post block (serum starvation. The results of one 24 h post block synchronized cells array set are shown in array I (Figure 20) and one set comparing cells at 12 hours post block synchronized cells in array III (Figure 21). An unsynchronized array set at 36 hours post infection was also examined (array IV, Figure 22) to compare differences in synchronized and unsynchronized cells. Genes that changed two-fold up or down were considered to be 67 J2 "55 o o T3 CD N CO C L CM S1 c • • • 1 Ic o CNJ O o o oo o CD a> E 3 C Qi c a; o o CNJ CD LO LO U") "si" LO LO CO (lAlfrZ/lfrZ) osiey CO LO CNJ LO T - LO O CN r-1 O SB CJ '•4—* u < a OH H • O "° c -5 CO u 06 o w Is c z I 2 -S 3 3 4- § -c > 5/5 o B co CN fi 1) S 1 2 .5" 2 0 >. BC rj 1 O > o O 8 CO (D I s-SB Q . a 3 <*j i n § U 2 CO © o A £ « •§ « "5 g o. 60 c O cO CO t CJ CJ C cj CO CO 3 £ •a M CO c d 3 CO a 55 cj CN £ ^ o 3 o -t-j S VO •3 o •-, >~l CJ * 3 T3 CJ CJ O 5 OH CJ CJ CO c X CO 2 3 CO co ••3 < 3 Z V •• I s * 8 1 cj 3 3 ^ % CJ •a o 2 fi 68 o o cu g eas CJ o CJ c TD cci CJ ed N cd O "cd T3 a 5 ,—, B V < lue ck. val o -JO JS OS 0_ m e» 3 cs C u 43 CN '—1 4) CA CJ _C~-" — cj U u CJ OJ < 4H c zed CJ £ 3 . — G c o CJ synchi indanc ein 1. e - Q o — c cd CH ,!- C - rH Ci) CJ _ E oun eas 'ZJ CJ cd oun CJ CJ CJ O ^ E c o (—1 >5 1—1 cd CJ u o — — CJ ccS cd cd C CJ ti-asig indi port - E A Ex l2M)w lue pu l2M)w rd > cd eg & -5 N • — ,—1 S—• SOT tios BP ta u cd or s CJ Gen eded. ce. .. CJ E .—i c cd CN T3 no E 3 5fl JCi •cd X ) cd E 69 A r r a y IV: 3 6 h r . P.I. u n s y n c h r o n i z e d L A 9 c e l l s 25 20 CO 2 15 CO co, •J 10 ra DC • FEFBP1 ;ALY Leuclne-iich epeatpofein • SH0O2;Ras-binding piotein3JR-8 ;pptoag) aclvatod tansiiplonteminalon • iecept>r4Gpiofein-coupled + factoM (TTF1) # Bceptor, ^ 0 20 40 60 Gene number 80 120 5 4.5 4 3.5 CO CO 3 CO CO 2.5 o 2 ro DC 1.5 1 0.5 0 A r r a y IV: 3 6 h r . P. I . u n s y n c h r o n i z e d L A 9 c e l l s PPPA ;piotea93-act\ated • iecepbr4Gpo6in-coupled receptor, In lombin leeeptor peanaplcdenaVpofein 95 (PSD-95) 4y tian x lip ton te mina f on tacbrl (TTF1) typsins-ppfeinhinaaaiyK + calnylinhinriinnpntain • Calmodulin iai pecuisoi; Wna33 «K • cyclin Dl —•• • • • ^ 1 *fflMSK3 —. inhibitor »m S 3 Btnoicacidiecepbr qamma-A 0 20 40 60 Gene number 80 100 120 Figure 22: Gene ratios (36I/36M) with a signal >500 counts from unsynchronized L A 9 cells 36 hours post infection. No normalization was needed. Ratios with a value >1 indicate an increase in abundance in the infected cells, ratios with a value <1 indicate a decrease in abundance. Both graphs plot the same data, but at different scales. mMSK: Mitogen & Stress-activated Kinase 2; R A N G A P : Ran GTPase Activating Protein 1; REFBP1: RNA and Export Factor interacting protein 1. 70 CD LL UJ _J "o3 tiL" o 1 CO •-C ,o CDh co; i : | o O X CO CD I'* I Si •it -: o .1 CP v . ' t o f l r*;-e. O 111 s D_ g CQ O UJ I .a ::s r-.Q .di 1.8 *;0 : Q ; g '-i± 2 -8 CO icbj O) in d CO CO i CM i-s CN 7.-CM CO ::;t "a p 6 CO .CO CO CO O O Q c !• o o O •ol CO • i o o I CO t co 5 CO CO CN CO o CN 'c 3 • CO I I I I , c ;cb •••o . o T J • C :(0| -2 ni l . P 0) X) i C O co in co d '.CN CO d CO i i o c o '18 c I J5 c o c CD ,2 Q oi CO CD •3 D .2 2 o JS • Q ! o X5 .. C CD UJ CQ O (/) i_ o 2 Q . C i i c o •si CO CO CN CD CN O CO I JW c -C0 O) ''.CO 1_ CD - c co J3 o 5" _l o o co 3 c JB co .3 o 00 CM CM CO O CM p. CM Qi I --Ik' . .0) CO CO c •o CO l_ O 2 Q CD ,:,C0 I Q (b c 'co p co =8 CM CO CO CO d CO CM CO CO CO CN o CO CO CO o o d CO CO CO co or CT) CO CO CO ;co co CO CM , CD « c c 2 Q |:S l l ,*-» o :co i CO CO 2 *-» CO • c 81 I o x : CD C 9\ c2 CT> CO d o d 2 Q CJ) c-• C 15 c o > CO o rep. c - 1 w ••'Q U) ! ; C TL7 C 1Q • co c2 6J • i o o 'm • CO 1! 2 cb c ' o CO o CO csi co CMI. o CO CO CN d CO CO CO CO CO col CO CO CO co 71 significantly altered in expression. This was based on the manufacturer's recommendations and my own observations. Table 3 lists all the ratios for genes that were altered in at least one array set. It should be pointed out that in arrays I I and IV, there was an area o f high background on the extreme right o f the arrays. Because of this, many genes in this region o f the array could not be analyzed. To examine reproducibility, genes with signals greater than 1000 are plotted for array I & I I (24 hours synchronized) as shown in Figure 23. There is significant overlap between the two arrays suggesting that they are fairly reproducible. Several basic observations can be made from the data. First, the signal strength was lower than expected. Approximately one hundred of the over one thousand genes on the array had signals above the threshold. Furthermore, many of these genes' signals were on the lower end, making data interpretation difficult. Second, during the first 12 hours after release from the block, expression levels o f few genes changed even though the cells were infected 24 hours previously. This is not necessarily surprising as the cells are still in G I and M V M requires cells to be in S/G2. In contrast wi th the 24 hour post block sample (36 h after infection) where the cells are now in S/G2, a number o f host cell genes now appear to be altered in expression. Finally, there are both similarities and differences between synchronized and unsynchronized infected L A 9 cells. Several genes were altered in both synchronized and unsynchronized cells, including the SINEs, PvEFBPl (RNA exchange factor binding protein one), and components o f the retinoic acid receptor. However, other proteins like cyclin D l and cdk4/cdk6 inhibitor show different expression patterns. There is yet another sub-group o f genes that shows altered 72 S3 fi V_ 1— CD CD 06 >> CD v_ s_ CD C "CD C O) "(73 C/> CD C CD G) o • H CD o r CD CD CO CD - ! c T - Ct) CNJ O h- CD CD LO CO CO LO CNJ CN LO LO O (WPZIWZ) one* 73 expression in unsynchronized cells whose signals were too low to be analyzed in the synchronized cell arrays. To confirm gene expression levels, I conducted northern blot analysis on two genes, R A N G A P (RAN GTPase Activating Protein 1) and TTF1 (Transcription Termination Factor 1). R A N G A P gave a consistently strong signal with a ratio of approximately one in the arrays analyzed. A fragment of the R A N G A P cDNA (nt 283 to 2263) containing the majority of the open reading frame and the 3' untranslated region (supplied by Dr. James Degregori, University of Colorado) was used to probe a northern blot of synchronized L A 9 cell total RNA (Figure 24a). A band of the expected size, 3kb, was detected and the expression level did not appear to change at 12 or 24 hours post block, agreeing with the array results. In contrast, TTF1 was up-regulated in response to virus infection in both the synchronized and unsynchronized later time point arrays. I therefore probed an unsynchronized total LA9 cell R N A northern blot with a full-length TTF cDNA probe (supplied by Dr. I. Grummt, German Cancer Research Centre) with the results shown in Figure 24b. However, instead of detecting the expected 4 kb transcript for the TTF mRNA, a much smaller band, <250 nt, was found that significantly increased as the infection progressed. This looked very similar to the B2 SINE northern blots (see Figure 15). Sequence analysis of the TTF cDNA confirmed the presence of a murine B2 SINE in the 3' untranslated region of the transcript. Discussions with the array manufacturer eventually confirmed that the TTF1 spot also contained B2 SINE sequence, thus, the TTF1 spot detects both TTF1 and the murine B2 SINE. Since the TTF1 transcript could not be detected by northern blot analysis, I concluded that signal coming 74 hr post block ^ ^ c n c n 9.5 kb - 7.5 kb • - 4.4 kb RAN-GAP —> mm mm mm mm - 2.4 kb - 1.35 kb post infection (hr.) ^ < ; c n ^ - ^ o o o o c n ^ c n 155 nt Figure 27: Reverse transcription of SINEs in polymerase III and II transcripts. A : Polymerase III transcripts: Primer WPW25 primes a 155 nt B2 SINE cDNA. Primer WPW27 primes a 110 nt B l SINE cDNA. In both cases, reverse transcription is terminated at the 5' end of SINE transcript. Primer WPW26 primes a 120 nt 5S R N A cDNA which is used as a loading control. B : Polymerase II transcripts: Primer 25 primes a B2 SINE cDNA larger than 155 nt as termination continues past the 5' end of the SINE. Primer sequences were based on data from Liu et al., 1995 [161]. 86 300 nt 200 nt post infection (hr.) mock 12 24 36 48 54 60 72 36 48 4 * *m m*. *m. «». B B2 SINE expression 155 nt B2 SINE cDNA 120nt5S cDNA 18000000 n 16000000 14000000 12000000 10000000 8000000 6000000 4000000 2000000 0 : 3 t CD co co o CD CN CD co 3 Infected (I) or mock-infected (M) samples at various times (hours) post-infection B2 SINE expression/5S RNA expression — , R n 121 241 361 481 541 601 721 36M 48M Infected (I) or mock-infected (M) samples at various times (hours) post-infection Figure 28: Legend on next page 87 Figure 28: Expression of B2 SINE transcripts during M V M infection. A : 5% acrylamide gel showing the products of primer extension with a B2 SINE (WPW25) and 5S RNA (WPW 26) specific primers. The B2 SPNE product (155 nt) and 5S product (120 nt) are indicated by the bold arrows. B: Radiometric quantification of the B2 SINE band in A both directly, top panel, and as a ratio, bottom panel, (B2 SINE counts/5S RNA counts) to adjust for loading. M: Mock-infected unsynchronized L A 9 cells. I: MVM-infected unsynchronized LA9 cells. 88 post infection (hr.) mock 12 24 36 48 54 60 72 36 48 300 nt 200 nt 100 nt B H O n t B l SINE B1 SINE Expression during MVM infection 2500000 2000000 1500000 1000000 500000 n 121 241 361 481 541 601 721 36M Infected (I) or mock-infected (M) samples at various times (hours) post-infection 48M Figure 29: Expression of B l SINE transcripts during M V M infection. A : 5% acrylamide gel showing the products of primer extension with a B l SINE specific primer (WPW 27). The B l SINE product (110 nt) is indicated by the bold arrow. B: Radiometric analysis of the B l SINE band (arrow) in A . M : Mock infected unsynchronized L A 9 cells. I: M V M Infected unsynchronized L A 9 cells. 5S RNA primer extension was not done simultaneously as the 120 nt 5S R N A band overlaps the 110 nt B1 SINE band. 89 B2 and Bl SINE transcripts are transcribed predominantly by RNA polymerase III To confirm the size o f the major B l and B2 SINE amplified products (Figure 28 and 29) primer extension products were separated on sequencing gels (not shown). In both cases, the major band was o f the size o f the expected RNA polymerase I I I transcript. This further suggested that the majority of SINE transcripts were o f R N A polymerase I I I origin, rather than another possibility in which the SINE is embedded in a Polymerase I I transcript. In the latter case, products larger than either the 110 nt (for B l SINEs) or 155 nt (for B2 SINEs) band would be expected (Figure 27). Interestingly, in the B1 SINE primer extension (Figure 29), a second band o f 144 nt was also seen to be up-regulated. Presumably this is due to a B1 SINE sequence embedded in another altered transcript. The major nonstructural protein of M V M , NS1, induces increased B2 and Bl SINE levels It was suspected that the major nonstructural protein o f M V M , NS1, was responsible for increased B2 and B l SINE levels. To test this hypothesis, L A 9 cells were transfected with plasmids expressing either NS1/NS2 (pCMVNSl ) , NS1 only (pCMV1989, a gift from Dr. David Pintel, University o f Missouri), or as a control (3-galactosidase (pCMV(3-gal) and RNA was harvested at various times post-transfection. The pCMV1989 plasmid contains NS-1 coding sequence with a point mutation (A to C) that alters a splice acceptor site, preventing splicing o f the NS1 ( R l ) transcript into the NS2 (R2) transcript [81]. Primer extension analysis was used to quantify B2 SINE R N A expression (Figure 30). A t 12 hours post-transfection, there was no significant difference in B2 SINE levels in the cells transfected with the NS 1 -expressing plasmids as compared to the pCMV(3-gal control. 90 A 12 hr 24 hr 36 hr hr. post-transfection ON , i _ » ON , i , ON , i O N O J D ^ O N O J O ^ O N & J ) Z i - H CQ. Z i—( GO. Z f -H CQ. S C o U U U U U U O U U C O C O O H ft ft ft ft ft ft ftft rftffr S MWk «t&: **>m* ^mimwr ^mw» im^" "vm&f * W W P * IIIMOI*? * * * «tfr 4 * « M » *Ji» M M * M K M M mm mm A «Mi tft* «•» « • * «H» * • » mm* mm mm mm, mm 155nt(B2 SINE) 120 nt (5SRNA) B B2 SINE expression in transfected f ibroblasts ^ 3 5 CO 0 0 s 3 CQ 15 2.5 CQ S 1.5 5 1 co S 0 5 m 0 M i l • Raw ratio • 5S Normalized ntf # # * ^ ^ Nc# ^ / / / / / / / ^ ^ ^ e Plasmid Figure 30: Legend on next page 91 Plasmid Expresses p C M V N S l : NS-1&NS-2 pCMV1989: NS-1 pCMVpgal: (3-galactasidase Figure 30: M V M NS1 induces increased B2 SINE expression. A : 5% acrylamide gel containing the products of B2 SINE primer extension from L A 9 cells transfected with plasmids expressing NS1, NS1/NS2 or (3-galactasidase at various times post-transfection. 36M: Mock-infected unsynchronized L A 9 cells 36 hours. 361: MVM-infected unsynchronized cells 36 hours post infection. B: Radiometric analysis o f the 155 nt band in A. Data are expressed as fold change compared to the (3-galactasidase control. 5S Normalized: Data normalized relative to the 5S RNA signal (B2 SINE/5 S) to correct for variable loading. 92 However, at 24 hours post transfection, B2 SINE levels increased in cells transfected with either p C M V N S l or pCMV1989. This trend continued at the 36-hour time point, where an approximate 3-fold increase in B2 SINE levels was seen for both p C M V N S l and pCMV1989 as compared to the pCMV(3-gal control. Interestingly, there appeared to be no difference between cells expressing NS1/NS2 or just NS1 alone. A similar trend was seen when B l SINE levels were examined (Figure 31), with an approximate 2-fold change seen by 36 hours. It should be noted that the changes in B2 and B l SINE levels were observed in spite o f the fact that transfection efficiency was only 15-20% (by X-gal staining). From these data, it was concluded that expression o f the NS1 protein alone could increase both B2 and B l SINE levels. Altering SINE levels Theoretically, SINE levels may be increased in several different ways (Figure 32). SINE transcription can be up regulated by either altering promoter accessibility (and thus removing SFNE transcriptional repression) or by increasing the amount o f basal RNA polymerase I I I transcription factors. Alternatively, transcription levels could remain unchanged, but the rate o f SINE RNA degradation could be reduced. This would also lead to an increase in concentration o f SINEs within the cell. TFIIIC220 and TFIIIC110 protein levels do not increase during M V M infection As mentioned above, one way to increase SINE transcription is by increasing the amounts o f the various basal RNA polymerase I I I transcription factors. It has been reported that murine fibroblasts (3T3 cells) transformed with SV-40 expressed higher 93 12 hr 24 hr 36 hr Hr. post-transfection > > Si 23 S CQ. * - a , H CD. ^ ^ - C ^ r -H CQ. > > > > > > > S J§ s s s s s § s O H ^ Q H CH £ ^ CH C H ^ HOnt (Bl SINE) B1 SINE expression in transfected LA9 cells UJ 2 . 5 z 55 2 CO « 1 . 5 P % 1 z w 0 . 5 OQ 0 \ # y y y y y y y y ^ ^ 0> >S> r& r& s>> < ° < ° ^ rJS <& <&> <8> J» J» >® * r * r ^ « r » r ^ ^ r ^ r ^ r ^ r ^ post-block 7 - «* 6 Figure 36: Legend on next page 103 1. p90 S6K (2.2x U) 2. S6Kp70(1.7xD) 3. Unclassified (2.3x U) 4. Unclassified (2.8x U) 5. C K I I a ' (1.9xU) 6. Cdk5 (2.3xD) 7. Mos (1.8xU) Figure 36: Kinexus's protein kinase screen for mock and MVM-infected synchronized LA9 cells 24 hours post-block. Numbers indicate bands that showed altered expression upon M V M infection. Values represent the fold change (infected/mock) as a result of infection. U: Up-regulated. D: Down-regulated. The unclassified band three was recognized by PAK-a antibody and unclassified band four was recognized by Rskl antibody. Infected values were multiplied by 1.6 to normalize to infected values. 104 appendix 3 for antibodies and expected protein sizes). The resulting western blots are shown in Figure 37. Four o f the proteins analyzed were confirmed to display altered expression during infection. These were Ribsosomal protein S6 kinase p70 (S6Kp70) and p90 (S6Kp90), unclassified band #3, and unclassified band #4. O f the remaining three proteins, one (cyclin-dependent kinase 5) could not be detected by the western blot and two, Casein kinase I I a ' and MOS kinase, did not appear to change upon infection. 105 A S6K M 116 kDa1 97 kDa' 66 kDa' 45 kDa' S6Kp90 <— S6Kp70 B CKIIa M I 45 kDa CKIIa CKIIa ' Figure 37: Confirming Kinexus's protein kinase screen. Western blots o f the same protein samples used in Figure 25 probed with: A . Anti-S6K (Santa Cruz) which detects S6Kp70 (triplet of bands around 70 kDa) and S6Kp90 (90 kDa). B. An t i -CKI I (Gift from Dr. Steve Pelech) which detects C K I I a and a ' . M : Mock-infected synchronized LA9 cells, 24 hours post-block; I: MVM-infected L A 9 cells, 24 hours post-block. 106 Unknown probed with anti-Rskl M 97 kDa 66 kDa 45 kDa «— Rskl Unknown #4 D Unknown Probed with anti- PAK a M I 97 kDa 66 kDa 45 kDa -PAK a Unknown #3 Figure 37 continued: C . Anti-Rskl (Santa Cruz) which detects Rskl (90 kDa) and the unknown band #4 in Figure 25. D. Ant i -PAK which detects PAK (70 kDa) and the unknown band #3 in figure 25. M : Mock-infected synchronized LA9 cells, 24 hours post-block; I: MVM-infected LA9 cells, 24 hours post-block. 107 Discussion: The goal o f this project was to further characterize viral/host interactions by examining changes in gene expression in the mouse fibroblast cell line L A 9 infected with M V M p . M V M p is cytotoxic to LA9 cells, resulting in cell death three days after infection (Figure 5). It is hoped that these initial studies would aid in the understanding on how M V M , and possibly other parvoviruses, causes cytopathic effects in cells. Differential Display At the beginning o f these studies, the differential display technique (see Results page 49 for an explanation o f the technique), one o f the standard methods for identifying genes that changed expression [148-151], had been used to study expression changes in cells infected with human papillomaviruses [152], cytomegalovirus [153], herpes simplex virus 1 [154], and human immunodeficiency virus [155] [156]. In most cases, viral infection led to alteration in expression levels o f large numbers o f host cell genes and it was expected that M V M p infection of mouse fibroblast LA9 cells would do the same, however, the M V M genome is much smaller than most other viral genomes. Before starting a large-scale screen examining M V M infection o f mouse fibroblasts, I first tested the differential display technique. Human embryo kidney 293 cells were infected with adenovirus and screened by differential display RT/PCR using as an arbitrary primer that could recognize an adenovirus transcript. One strongly altered band was detected (data not shown) and confirmed to correspond to the adenovirus type 5 fibre protein gene. A second test was conducted on MVM-infected L A 9 mouse fibroblasts, wi th the RT/PCR screen using an arbitrary primer specific to M V M NS1 108 (data not shown). Again the screen successfully detected the NS1 transcript as an altered RNA during infection. Together, these two tests confirmed that the differential display technique in my hands could be used to identify altered transcript abundance during viral infection. The main differential display screen examined unsynchronized mouse fibroblast cells (LA9) infected with M V M p at 12, 24, and 36 hours post-infection compared with uninfected mouse fibroblast cells. These time points were chosen to examine changes in transcription early in infection before the virus was actively replicating in most cells (12 h time point) and during the initial stages o f replication occurring as the cells reach and arrest in S phase (24 and 36 hour time points). A total o f 24 primer pairs was used, representing approximately 15% o f the actively transcribed RNA within the cell [137]. Table 2 shows the results o f this screen, summarizing the various altered bands that were detected. A l l o f the confirmed altered bands were found to be up regulated in response to infection, although some care should be taken in interpreting this statement, as these bands were comprised o f retroposons and viral transcripts. Three key observations can be made from these results. First, few ( i f any normal) genes were found to be altered as a result o f M V M p infection. Second, the M V M NS 1 transcript was detected multiple times, again confirming that the technique could detect transcripts o f altered abundance. Finally, the majority o f confirmed altered transcripts contained repetitive elements, including the B l & B2 SINEs and the L I L INE. Although the low number o f altered bands and confirmed altered genes is not unprecedented since a similar screen with Coxsackievirus-infected mouse heart tissue identified 5 altered genes [157], it was surprising given the pathogenic nature o f M V M 109 towards these cells. The low number o f detected altered genes could be due to the high stringency in selecting bands to be studied, difficulties with reproducibility, false positives and downstream cloning, or it could be that M V M infection simply does not cause large changes in host gene expression. This last possibility should be viewed with caution. Only a small portion o f the transcriptome was analyzed and only early times post-infection were examined. The repeated detection o f the NS 1 mRNA further confirmed that the screen was indeed detecting altered levels o f transcripts. It was interesting that in both cases, PCR amplification did not occur at the 3' end of the transcript, but rather towards the 5' end at polyA rich sequences. These internal amplifications during differential display are an added benefit as they further increase the number o f transcripts screened. However it does emphasize the unreliability o f the technique as the 3' end of the transcript was not detected. The majority of altered transcripts detected in the differential display screen were transposable elements comprising the B l and B2 SINEs and the L l L INE. Fragments o f SINEs and LINEs can be found in the 3' untranslated region o f some mRNAs, so the sequences detected in the screen could be o f this origin. It is clear, however, that both B l and B2 SINEs and the L l L INE transcripts increase upon infection (Figures 15, 28, & 29). This is not the first differential display screen to detect repetitive elements. Differential display has detected L l LINEs [158], Alus [159], and endogenous retroviral-like elements [160]. Furthermore, other viruses, discussed in more detail below, up-regulate SINEs, specifically the A lu element. 110 Changes in gene expression in response to M V M p infection as detected by Clontech cDNA arrays: The inherent difficulties in screening altered transcripts (in M V M p infected mouse fibroblast cells) through differential display resulted in the decision to use arrays for further screening experiments. At the time o f these experiments, there were few mouse arrays available and most were very expensive. Based on limited choices, Clontech's cDNA macroarrays were chosen. These contain approximately 1000 genes in single spot orientation. To further increase my chances o f detecting altered patterns o f expression, the mouse fibroblast cells being screened were synchronized, so that all MVM-infected cells would enter S phase and begin viral replication at the same time. Three different conditions were examined: Mock and MVM-infected synchronized mouse fibroblast cells at 12 or 24 hours post-block, or mock and MVM-infected unsynchronized mouse fibroblast cells at 36 hours post-infection (to compare with the synchronized cells). Overall, there were two weaknesses with these arrays: First, the signal strength was low, with only approximately 100 spots per array above the signal threshold cut o f f and many o f these were on the low end. Second, the arrays were prone to high background, both in the form o f small random background dots, making spot identification difficult, and in some cases, much larger sploches, covering larger sections o f the membrane. The latter was due to the two ends o f the array overlapping each other during hybridization in the hybridization oven, resulting in the loss o f data from the overlapped areas. This was corrected in later experiments. I l l Infected synchronized cells at 12 hours post-block showed very few changes in gene expression with most o f the detectable genes clustering around an infected/mock ratio o f one (Figure 21). Only three mRNA were observed to change in abundance (all increased) and none of these changes were sustained at the 24 hour time point. The low number o f altered transcripts at this point is not surprising, as virus is not actively replicating at this point. Infected synchronized cells at 24 hours post-block, by contrast, showed more changes in transcription. O f the approximately 90 detectable transcripts, 7 appeared to change reproducibly in the two replicates done. Another 7 transcripts were detected as altered in one array set but were not in the other (Table 3). These included genes involved in cell cycle, and transcription. The two replicates were fairly reproducible for transcripts with signals above 1000 (Figure 23). Interestingly, infected unsynchronized cells at 36 hours post-infection, also showed altered transcription, but in a different pattern. Many of the cell cycle genes altered in synchronized cells were not altered in unsynchronized cells or demonstrated different expression patterns (for example cyclin D l , which was down-regulated in synchronized cells but up-regulated in unsynchronized cells). This was not too surprising as synchronized cells are all in the same stage of the cell cycle, whereas unsynchronized cells are not. Transcripts from three genes (ALY/RFBP1, Transcription termination factor/TTF, and mitogen & stress activated protein kinase 2) appeared to change in abundance in both synchronized and unsynchronized cells. Northern blot analysis was done with either transcription termination factor or RAN-GAP (a transcript that did not change in expression in any o f the arrays) probes, as both had high signals in both the 112 mock and infected arrays. In both cases, the Northern blot data confirmed the changes (or lack o f changes) identified in the array; however, the transcript band detected in the northern blot for TTF was much smaller than expected, and was due to the TTF transcript containing a B2 SINE. Subsequent communication with Clontech confirmed that the TTF transcript did indeed contain B2 SINE sequence (confirming nicely the results seen in differential display). The B2 SINE transcripts was probably what was altered during transfection, as I was unable to detect the correct sized TTF transcript (Figure 24b). In conclusion, the data obtained with the Clontech array indicated that very few transcripts change in abundance early in infection, although more transcripts change as infection progresses. Problems with low signals and background made detection and confirmation difficult. These arrays also confirmed that B2 SINE transcripts increased during M V M infection. SINE expression during M V M infection The increase in SINE transcription, as detected by two different screening techniques, led me to focus my investigations on transcription o f SINEs and their role in M V M infection. Initial experiments confirmed that the apparent increase in the number of SINE transcripts was due to RNA polymerase I I I transcription and not due to contaminating D N A or fragmented SFNE-containing RNA polymerase I I transcripts. SINE transcripts are up regulated in a variety o f conditions, usually associated with cell stress. These include heat shock (Alu, B l & B2 SINES) [161, 162], ethanol treatment ( B l & B2 SINES) [162], cyclohexamide treatment (A lu SINE) [161], D N A damaging agents such as cisplatin, etoposide, and y-irradiation (Alu SINE) [159], as a result o f cell transformation by SV-40 virus (B2 SINE) [141] and by viral infection 113 including adenovirus infection (Alu SINE) [163], herpes simplex virus (HSV) infection (Alu SINE) [164], and H I V infection [146]. To my knowledge, the increased B l and B2 SINE expression during M V M p infection is the first example o f viral infection altering murine SINEs. It would be interesting to know whether M V M i infected lymphocytes or other parvovirus-mediated infections, such as B19, A A V , or CPV, can also cause an increase in SINE expression. Theoretically, SINE transcript levels can be increased in one o f two ways (Figure 32): Either SfNE transcription is increased (options one and two) or SINE transcript degradation is decreased (option 3). In all o f the viruses studied thus far, SINE induction has always been due to increased transcription [146, 147, 163]. To determine i f the increase in B l and B2 SINE transcripts was due to increased transcription, nuclear run-on (NRO) studies were undertaken to measure the amount o f newly synthesized R N A in M V M infected cells. Although this technique was shown to work, RNA polymerase I I transcription could not be blocked by addition o f a-amanitin at concentrations ranging from 5 to 25 u.g/ml. As RNA polymerase I I transcripts can contain embedded B l and B2 SINEs, this creates a large background that prevents accurate quantification o f the RNA polymerase I I I derived SINE transcripts. I t is currently unclear why a-amanitin was unable to block RNA polymerase I I transcription, as the drug was present at the right concentration (concentration was checked spectrally and increasing the concentration o f the drug to 25 ug/ml had no effect. One hypothesis is that the drug failed to enter the nuclei. This may be due to debris surrounding the nucleus, preventing uptake. Isolating nuclei for these experiments is not trivial, as the nuclei should be free from debris and yet not be damaged such that the contents can leak out. Alternatives methods o f isolating 114 nuclei (like dounce homogenization/sucrose gradient separation) could be attempted to get around this issue. It seems unlikely that the increase in SINEs in response to M V M infection is due to decreased degradation. Due to the unique location o f each integrated SINE and the lack o f a RNA polymerase I I I termination signal, each SINE transcript has a unique 3' end, which w i l l influence RNA stability. Any method of affecting SINE degradation would have to affect each individual SINE transcript. Furthermore, the decrease in degradation would have to affect B l SINEs, B2 SINEs, and L l LINEs equally and yet not global RNA levels (total cellular RNA does not increase with infection, W. Will iams, unpublished observations, data not shown). However, without data indicating an increase in transcription, decreased RNA degradation cannot be ruled out. One potential method of increasing polymerase I I I transcription is to increase the basal RNA polymerase I I I transcription factors. This occurs during HSV infection [164], H I V infection [146], and adenovirus infection [143] (although increasing RNA transcription factors only during adenovirus infection is not sufficient for SINE up-regulation [165] and in cells transformed with large T antigen [142]. Specifically, transcription factor TFII IC activity was increased and in some cases this was due to increased expression o f TFII IC 110 and TFIIIC220 component proteins. These proteins are essential for RNA polymerase I I I type I I promoter activation (see introduction for more details). I used antibodies obtained from Dr Robert White and Dr. Arnold Berk to measure expression levels o f the TFII IC 110 and TFIIIC220 proteins. Levels o f neither protein were altered as a result o f M V M p infection (Figure 33 & 34), suggesting that increasing RNA polymerase transcription factors may not play a role in M V M altered 115 SINE expression. However, this does not rule out the possibility o f altered expression o f other TFII IC or TFI I IB proteins. Chromatin rearrangement can also mediate increased SINE expression presumably by exposing new SINE promoter sites for transcription and thus releasing SINE transcriptional repression. Increased chromatin accessibility on A lu SINE, alpha satellite DNA, and L I L INE has recently been demonstrated in heat-shocked HeLa cells [166]. Increased SINE expression in adenovirus-infected HeLa cells is also thought to occur by increased chromatin re-arrangement [165]. M V M infection causes D N A damage, which would be expected to cause changes in the chromatin structure. It is tempting to suggest that this could be the reason for increased SINE numbers during M V M infection. A chromatin re-arrangement assay (nuclease digestion o f SINE containing chromatin) would be informative. M V M p nonstructural protein NS1 alone can up-regulate B l and B2 SINE (Figure 31). As NS 1 is responsible for introducing D N A nicks in both the host and viral genome [91], it is tempting to speculate that this function could be responsible for SINE induction (as suggested above). Repeating mouse fibroblast transfection/primer extension experiments with an NS-1 nickase deficient mutant would be informative. It would also be interesting to determine i f M V M structural protein V P 1 , which contains a phospholipase domain [90], also can induce increased SINE expression. The phospholipase domain could stimulate an inflammatory response which in turn could stimulate SINE expression. The role(s) o f SINEs within the cell is not ful ly understood. It is clear that SINE retroposition is a powerful mutational force. SINE integration can interrupt exons, alter 116 splicing, alter promoter activity, alter or add poly adenylation sites, and can cause sequence duplications and deletions through unequal homologous recombination [112, 167, 168]. A lu SINE integration can cause genetic disorders and cancer in humans with examples including hemophilia, B-cell lymphoma, Tay-Sachs disease, thalassemia, and Lesch-Nyhan syndrome [112]. What is not clear, however, is whether SINEs are selfish parasites or i f they convey a benefit to the host (for example in a mechanism similar to adenovirus V A I RNA). Supporters o f a beneficial role for SINEs argue that SINEs are up-regulated in a manner similar to the heat shock genes fol lowing hyperthermic shock [162], that the A lu SINE RNA can bind and antagonize PKR activation [169], and that the A lu , B1 and B2 SINEs can transiently stimulate translation o f reporter constructs in a PKR-independent manner [170]. Furthermore, SINEs are found in G/C rich (gene rich) D N A whereas LINEs are found in A /T rich (gene poor) D N A [115, 116]. As SINEs and LINEs are believed to share the same integration mechanism, this suggests that there might be some sort o f selective pressure on SINEs to remain in G/C rich D N A that is absent on the LINEs. Likewise, the role o f SINEs in M V M p infection is also unclear. The increase in B l and B2 SINEs may simply be a consequence of viral perturbations to the host cell. As such, they could act as indicators o f problems within the cell, for example chromatin re-arrangements or changes in RNA degradation. Alternatively, SINEs could be playing a protective function within the cell. SINEs could be maintaining protein synthesis by inactivating PKR. However, there have been no reports that M V M infection leads to activation o f PKR or o f host cell protein synthesis shutdown. SINEs could also potentially bind the major M V M nonstructural protein, NS1 , sequestering it away from 117 the viral replication machinery. NS 1 may not need to bind directly wi th SINE RNA as NS1 has been shown to bind cellular proteins that have RNA binding motifs [78]. Further understanding o f the roles SrNEs play within the host cell w i l l be helpful in j understanding their role in M V M infection. Kinase Expression Analysis The protein expression levels o f some 75 protein kinases were examined in synchronized mock or MVMp-infected mouse fibroblast cells at 24 hours post block: Surprisingly, expression o f only seven bands changed at least-two fold out o f the over 70 kinases screened. O f these seven candidates, levels o f only four were observed to change in a standard western blots using the same proteins samples as used in the initial screen. Two o f the candidates were unknown proteins. The remaining two proteins were variants o f S6 kinase. S6 kinase is known to phosphorylate ribosomal protein S6 amongst others. This promotes selective translation o f certain mRNAs needed for cell replication from G I to S. Down-regulation o f S6 kinase, would suggest a block in cell proliferation in G I whereas most o f the microarray data suggest an increase in proliferation. These kinase I experiments were undertaken to give an overview o f the extent o f perturbation o f cellular protein kinases expression during M V M infection. While a few changes are seen, I decided not to pursue the kinase expression changes at this time, but rather focus on : changes at the RNA expression level. > Changes in gene expression in response to M V M infection as detected by Affymetrix oligonucleotide microarrays: | Late into the writ ing o f this thesis, an opportunity developed to study changes in mouse fibroblast gene expression in response to M V M p infection with Affymetr ix 's ; 118 oligonucleotide array. These results yielded a plethora o f information and, although still preliminary, have begun to greatly enhance our understanding o f MVM/host cell interactions. This section wi l l focus on some of the preliminary findings o f these results but should in no way considered to be exhaustive. The groupings seen below were made based on information gathered from the Affymetrix website, Genespring 6.0, and various i review articles. More work wi l l be needed to confirm these initial findings and to characterize them further (e.g., biological and technical replicates as well as RT-PCR confirmation). Not all o f the altered genes are discussed below. Mouse fibroblasts seem to induce a strong overall growth signal in response to M V M p infection (Table 5a). Multiple transcripts that are involved in promoting cell proliferation are up-regulated. These include transcripts encoding growth factors (epiregulin, nerve growth factor beta, and the chemokine oncogene GRO), a transcription factor (oncogene A-myb), and a protein involved in EGF signaling (Repsl). Epiregulin, an epidermal growth factor (EGF) like growth factor that can bind and activate the EGF receptor, showed a 14-fold increase as a result o f M V M p infection. In contrast, transcripts that encode proteins that suppress cell proliferation are generally down-regulated. These include transcription factors (Mad, Nup l p8), an interferon-induced protein (Interferon activated gene 202A, p202), and a protein without a known molecular function (Tob). M V M p infection also appears to induce immune and inflammation responses in mouse fibroblast cells. Several cytokines and chemokines are up-regulated in response to infection (Table 5b). The immune response modulator transforming growth factor (3 cytokine (TGF(3) (the so called anti-cytokine) is down-regulated in response to infection. 119 A B C Cell growth/anti growth Growth epiregulin 13.9 U GR01 oncogene 4 U Myeloblastosis oncogene-like 1, A-myb 2.3 U Nerve Growth Factor beta 2.1 U RalBPI-associated EH domain protein Repsl 2 U Vascular endothelial growth factor 2 D Anti-growth Nuclear Protein 1, p8 3.3 D Max-interacting transcriptional repressor (Mad4) 2.5 D Transforming growth factor, beta induced, 68 kDa 2.5 D Bone morphogenetic receptor for Bmp2 and Bmp4 2 D Interferon activated gene 202 2 D Tob family 2 D Irnrnune/lmflammatory Lymphocyte antigen 84, Interleukin 1 receptor-like 1 6.1 U glucocortoid-regulated inflammatory prostaglandin G/H synthase 3 U GR01 oncogene 4 U CD44 antigen (also involved in cell adhesion induced signalling) 2.8 U lnterleukin-4 receptor alpha 2.5 U Small inducible cytokine A2 2.6 U Small inducible cytokine A7 2.1 U Transforming growth factor, beta induced, 68 kDa 2.5 D Bone morphogenetic receptor for Bmp2 and Bmp4 2 D Transcription factor GIF 2 D Latent TGF beta binding protein 2.3 U Transforming growth Factor Transforming growth factor, beta induced, 68 kDa 2.5 D Bone morphogenetic receptor for Bmp2 and Bmp4 2 D Transcription factor GIF 2 D Latent TGF beta binding protein 2.3 U Table 5: Grouping of altered genes detected by Affymetrix microarrays based on similar function or pathways. Table continues next page. U: Transcript is up-regulated in response to M V M p ; D: Transcript is down-regulated in response to M V M p . 120 D E F Transcription factors TEA domain family member 4 2.6 U Autosomal Zinc finger protein group 2.5 U Myeloblastosis oncogene-like 1 2.3 U Enhancer Trap locus 1 5D D site albumin promoter binding protein (Dbp) 3.3 D Nuclear Protein 1, p8 (probable TF) 3.3 D CCAAT/enhancer binding protein (C/EBP) 2.5 D Max-interacting transcriptional repressor (Mad4) 2.5 D PPAR gamma coactivator (PGC-1) 2.5 D basic-helix-loop-helix protein class B2 2.5 D Chromobox homolog 4 (Drosophila Pc class) 2.5 D Transcription factor GIF 2 D Zn-15 transcription factor (zfp 292) 2 D Ets-2 2 D C/EBP & Dbp TFs CCAAT/enhancer binding protein (C/EBP) 2.5 D D site albumin promoter binding protein (Dbp) 3.3 D Nuclear Protein 1, p8 3.3 D c-Cbl associated protein CAP 2 D Stearoyl-coenzyme A desaturase 2 2 D Cholesterol Farnesyl diphosphate sythetase 2 D Mevalonate (diphospho) decarboxylase 2.5 D ATP binding cassette sub-family A (ABC1) member 1 3.3 D Table 5 continued. U: Transcript is up-regulated in response to MVMp; D: Transcript is down-regulated in response to MVMp. 121 Furthermore, its receptor (bone morphogenetic receptor for Bmp2 and Bmp4) and a transcription factor (GIF) that is activated by and enhances the effects o f TGF(3 are also down-regulated. In contrast, a protein (latent TGF beta binding protein) that binds the latent form of TGFP and down-regulates its activity is up-regulated. Together, these results suggest a multi-pronged approach to down-regulate TGFP activity (Table 5c). TGF(3 is also involved in suppressing cell proliferation, so down-regulation o f TGFP agrees with the observations made in the previous paragraph. Expression o f many transcription factors is altered during M V M infection o f mouse fibroblasts (Table 5d). The majority o f the transcription factors appear to be down regulated. In the case of the down-regulated transcription factor C/EBP p, and the related transcription factor Dbl , three genes (Nuclear protein 1 p8, c-cbl associated protein CAP, and Stearoyl-coenzyme A desaturase 2) under the control o f these factors were also down regulated (table 5e). C/EBP P and Dbl are involved in cell differentiation, cell proliferation, inflammation, and metabolism specifically in hepatocytes, adipocytes and haematopoietic cells [171]. Interestingly, M V M infection also seemed to alter genes involved in both cholesterol synthesis and transport (ATP binding cassette sub-family A member 1) as listed in Table 5f. What potential role cholesterol synthesis and transport plays in infection remains unclear. Finally, there are 24 EST sequences that remain to be identified and characterized (Table 4) with respect to M V M infection. Future work would also focus on both confirming altered gene expression by northern blot analysis, semi-quantitative PCR and primer extension analysis. The arrays 122 must be repeated, possibly with flanking time points (perhaps 24 and 48 hours post M V M infection). It would also be interesting to repeat the array studies with M V M i in a lymphocytic cell line to try and determine key transcription alterations. Array studies on cells fibroblast cells expressing just M V M ' s non-structural protein NS1 or just the capsid protein VP1 (to determine what role the phospholipase active site has on inflammation) would also be informative. Comparing the two types of arrays Table 6 lists the gene that were altered by M V M infection in unsynchronized LA9 cells at 36 hr P.I. as detected by Clontech macroarray. Also listed are the values found for the same genes in the Affymetrix oligonucleotide array. Interestingly, there is not perfect agreement between the two. Genes that did not show altered expression in response to M V M infection in the macroarray generally demonstrated the same trend in the oligonucleotide microarray (for example cyclin-depndent kinase regulatory subunit 2). Genes that were altered in the cDNA macroarray, however, often were not predicted to be altered in the oligonucleotide array (for example calmodulin and RNA exchange factor binding protein one). The possible reasons for these discrepancies may lie in the different nature o f the two arrays. The Clontech macroarray has each gene represented by one cDNA spot whereas the Affymetrix oligonucleotide array has each gene represented by at least 16 different oligonucleotides scattered throughout the chip. Therefore the oligonucleotide array system allows a better averaging o f the signal from any one specific gene. Furthermore the oligonucleotide arrays also has a mis-match oligonucleotide for every perfect oligonucleotide, allowing a measurement o f the specific background to each sequence. The macroarray system lacks mis-match probes and as 123 i CD CD LL. CN LU GO —i o 8 ,-O Q Z o o Q c o o it— '8 CO CO 5 CN LO Q co D 1 CO f-J • o ' o 4^ :"» Q J CO CO T— CO CO CO o CN CO -f-» 3 CO CO c -+—» c 5 TJ C "a o 52 co 8Vl co 1 c o CD o CH i •"Lid" ; CD CO •CO I c Q < Q 8 CO c o CD CD TJ Q J2. I. a £r CD CO ' CO E 1 a i c '8 CO CN CN 8 CO CD CO. CN SI' CO CO 51 CO CD CO c t>0 o OH & KJ CO o < T3 T3 CD N 'S O o c C 5 0 a «s o u. o T3 C cj i O I - I o S o U H 124 such, certain spots may have higher background. Alternatively, the signal cut-off values in the macroarrays could have been set too low, resulting in slightly skewed ratios. Northern blot confirmations and repeating the microarray experiments to confirm these results w i l l resolve these discrepancies. Conclusions Through the use o f differential displays and arrays, altered gene expression in M V M p infected mouse fibroblasts was investigated. Overall, M V M p infection seemed to alter transcription in a small number o f genes. This was true for each experimental approach: differential display, Clontech macroarrays (approximately 10 changes out o f 1200 genes assessed), and Affymetrix microarrays (74 changes out o f 12,000 genes assessed). However, from these altered expression profiles several interesting results were seen: One, that M V M p infection appears to promote cell proliferation through the up-regulation o f several growth factors and the down-regulation o f suppressors o f cell proliferation. Two, infection also appears to stimulate an inflammatory and immune response, by activating cytokines and by specifically targeting transforming growth factor P. Three, a small number o f different transcription factors are altered during infection suggesting that transcription would further change as the infection progressed. Finally, cholesterol metabolism and transport may be altered as a result o f infection. Differential display and macroarray analysis also indicate that the retroposons B1 & B2 SrNEs and the L I L INE are also up-regulated as a result o f M V M p infection. This was confirmed by northern blots and primer extension assays. Furthermore, expression o f just M V M p ' s major non-structural protein NS1 could cause up-regulation o f B l and B2 SINE, suggesting that this multi-functional protein may play a major role in SINE up 125 regulation. Two critical basal RNA polymerase I I I transcription factors (TFIIIC220 and 110) were not altered during M V M p infection, suggesting that altered SINE expression many not be due to increased levels o f RNA polymerase I I I transcription factors. Whether the increase in SINE transcription is due to increased transcription or decreased degradation remains to be determined. 126 Appendix One: Plasmid Constructs Plasmid Contains p l B 6 - l p 2 A l p 7 A l U pCMV1989p p C M V N S l pCR3G3M2 p M 0 5 S l pRSETmTTF pT l -3 p T l T 3 m p l 8 RAN-GAP cDNA fragment (nt 283 to 2263). A gift fromDr. James Degregori. DD fragment o f B2 SINE (nt 32-174) in Topo-pCR2.1 DD fragment of NS1 (nt 439-687) in Topo-pCR2.1 Expressible NS1 containing a A to C point mutation at nt 1989. This alters a splice acceptor site and prevents splicing to the NS2 transcript. A gift from Dr. David Pintel. Expressible NS1/NS2 D D fragment o f B l STNE (nt 1-131) in Topo-pCR2.1 Murine 5S cDNA. A gift from Dr. G. Frederiksen. Murine transcription termination factor cDNA (TTF). A gift from Dr. I. Grummt. Rat tRNA cluster. A gift from Dr. T. Sekiya. Expressible Chicken (3-Actin. 127 CA S ff> 2 o ! 2 c '•4-* o c 3 HH CN GO —< CQ PQ oo CN r o Qi O =5 = O DX) O H '-3 a a f } m m m m cu o C cu gf s c CM o u a H H < O H H U o a < o H O o a u H u u H t—1 o y o o o < o < H H 3 u H O H H U U H H co r n r o cn i O o < o u < H H H H H H H H H H H u o u o H U H H H H H H H H H H o a u o < < < < m i H O < U u o o u < o u u o o H < H u a H f - H < o O H U U P U o u a H u < u u a o o o o o < H a < O u o < u o u u < H a o H o a H < < H O O O t—1 u u a u < < < u H H U H o u a <; H a H U < u < H < H a a u u u < u a < < H U U a r o i a H u < u H u < < o o H u u H o H U a o H U u - ) i n » n i n > n i n i n i n i n i n i n i n > n i n i n i n i r i ^ io >o 3f o CN r o PH -< P H < m r -PH PH PH < < < oo < CJ O & PH ~ ~ - PH HH < H H H c o r o