|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 FACTORS INVOLVED IN THE REGULATION OF GENE EXPRESSION OF THE LATENT HERPES SIMPLEX VIRUS TYPE 1 GENOME By DACIA L. K WIATKOWSKI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009
2 2009 Dacia L. Kwiatkowski
3 ACKNOWLEDGMENTS I would like to thank my Grama for her attention and support in whatever I am involved in She is truly a person of inspirational character that I constantly look up to I would also like to thank Ben ners and Jerri Bear for providing much needed stress relief and comedy over the past several years. Along with them, I would thank my Buddy, Jess, for moving down to this swampland of a state with me. It was amazing having family in the same city as opposed to 1200 miles away. To Dagny Taggart and Hank Rearden, thank you for your clarity and confidence. I would like to thank Dave for our endless miscommunications resulting in countless hours of arguing, when we are in fact saying the same thing. And lastly, I would like to thank my other mentor Nicole for always chall enging me and being supportive throughout my time in the Bloom lab
4 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 3 LIST OF TABLES ................................................................................................................................ 6 LIST OF FIGURES .............................................................................................................................. 7 ABSTRACT .......................................................................................................................................... 9 CHAPTER 1 INTRODUCTIO N ....................................................................................................................... 12 Herpesviruses .............................................................................................................................. 12 Herpes Simplex Virus Type 1 .................................................................................................... 12 Lytic Infection ...................................................................................................................... 13 Latent Infection .................................................................................................................... 15 Functions of the LAT .......................................................................................................... 16 Latent Transcription ............................................................................................................ 17 Animal Models ..................................................................................................................... 18 Epigenetic Regulation ................................................................................................................. 20 DNA Methylation ................................................................................................................ 21 Chromatin ............................................................................................................................. 21 Types of Heterochromatin. .................................................................................................. 23 Chromatin on the Lytic HSV 1 Genome. ........................................................................... 24 Chromatin on the Latent HSV 1 Genome. ......................................................................... 25 Reactivation and Chromatin. ............................................................................................... 26 Insulators and CTCF ............................................................................................................ 27 Insulators in HSV 1. ............................................................................................................ 29 Aim of Dissertation ..................................................................................................................... 29 2 MATERIALS AND METHODS ............................................................................................... 37 Viruses and Cells ......................................................................................................................... 37 Mouse Infections ......................................................................................................................... 37 Chromatin Immunoprecipitation (ChIP) .................................................................................... 38 Data Analysis Methods ............................................................................................................... 38 Guanidine Thiocyanate (GTC) RNA Isolation ......................................................................... 39 Reverse Transcription ................................................................................................................. 39 Semi quantitative PCR Analysis ................................................................................................ 40 3 POLYCOMB PROTEINS BIND THE HSV1 GENOME ...................................................... 41 Introduction ................................................................................................................................. 41 Results .......................................................................................................................................... 43
5 Discussion .................................................................................................................................... 49 4 DIFFERENCES IN TRANSCRIPT LEVELS IN THE LAT REGION OF HSV1 STRAINS 17 SYN + AND KOS CORRESPOND TO DIFFERENCES IN EUCHROMATIN DEPOSITION .............................................................................................. 62 Introduction ................................................................................................................................. 62 Results .......................................................................................................................................... 64 Discussion .................................................................................................................................... 67 5 INITIAL DOSE OF INFECTION AFFECTS THE LAT AND THE TAL TRANSCRIPT LEVELS BUT NOT CHROMATIN PROFILES DURING LATENCY ..... 78 Introduction ................................................................................................................................. 78 Results .......................................................................................................................................... 79 Discussion .................................................................................................................................... 85 6 DISCUSSION .............................................................................................................................. 94 Heterochromatin on the Latent Genome .................................................................................... 94 The LAT Region an d HSV 1 Strains ......................................................................................... 95 Effect of Dose on the Latent Transcriptional Profile ................................................................ 97 The Battle for Survival: Neuron vs. HSV 1 .............................................................................. 98 LIST OF REF ERENCES ................................................................................................................. 103 BIOGRAPHICAL SKETCH ........................................................................................................... 112
6 LIST OF TABLES Table page 3 1 Mean Normalized Values of triMe H3K27 Enrichment for 17 syn ............... 61 3 2 Mean Normalized Values of macroH2A Enrichment for 17 syn ................... 61 4 1 Mean Normalized Values of triMe H3K27 Enrichment for 17 syn + and KOS .................. 76 4 2 Mean Normalized Values of diMe H3K4 Enrichment for 17 syn + and KOS ..................... 77 4 3 Mean RNA molecules per genome for 17 syn + and KOS .................................................... 77
7 LIST OF FIGURES Figure page 1 1 Map of the LAT region including the LAT and TAL transcripts. ...................................... 31 1 2 Gene expression profile of HSV 1 life cy cle ....................................................................... 32 1 3 Mouse Footpad model. ......................................................................................................... 33 1 4 Types of chromatin and PTMs .............................................................................................. 34 1 5 Chromatin profile of the latent genome ................................................................................ 35 1 6 Insulator function and presence in the latent HSV 1 genome ............................................. 36 3 1 TriMe H3K27 is enriched on the 17syn + latent genome ..................................................... 54 3 2 Histone variant macroH2A is incorporated into the viral chromatin on the 17 syn + latent genome .......................................................................................................................... 55 3 3 triMe H3K9 is enriched on the 17syn + latent genome ........................................................ 56 3 4 Viral transcript abundance in 17 syn + latently infected mouse DRG ................................. 57 3 5 ................................. 58 3 6 genome .................................................................................................................................... 59 3 7 Bmi1 is enriched on the 17syn + latent genome .................................................................... 60 4 1 triMe H3K27 is enriched on the KOS latent genome .......................................................... 74 4 2 diMe H3K4 enrichment varies depending on the HSV 1 virus strain ................................ 75 4 3 Viral transcript abundance in 17 syn + and KOS latently -infected DRG ............................. 76 5 1 Viral transcript abundance in KOS/1 latently infected mouse DRG .................................. 89 5 2 LAT primary transcript abundance in KOS/1 latently infected mouse DRG .................... 90 5 3 TAL transcript abundance in KOS/1 latently -infected mouse DRG .................................. 90 5 4 ICP27 transcript abundance in KOS/1 latentlyinfected mouse DRG ................................ 91 5 5 T k transcript abundance in KOS/1 latentlyinfected mouse DRG ...................................... 91 5 6 Viral transcript abundance in KOS/M latently infected mouse DRG ................................ 92
8 5 7 Genomes per KOS/1 latent ly infected mouse DRG ............................................................ 92 5 8 ChIP of acetyl H3 K9, K14 on the KOS/1 genome ............................................................. 93 6 1 Mechanism of repression by heterochromatin formation in HSV 1 ................................. 101 6 2 Heterochromatin deposition has an effect on the reactivation potential of HSV 1 virus strains and mutants ...................................................................................................... 102
9 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy FACTORS INVOLVED IN THE REGULATION OF GENE EXPRESSION OF THE LATENT HERPES SIMP LEX VIRUS TYPE 1 GENOME By Dacia L. Kwiatkowski December 2009 Chair: David Bloom Major: Medical Sciences Immunology and Microbiology Herpes Simplex Virus type 1 (HSV 1) establishes a lifelong latent infection in the sensory ganglia of its host. During latency, no viral proteins are produced and the genome is for the most part transcriptionally repressed with the exception of the latency associat ed region. This region gives rise to two latent transcripts, the latency assoc iated transcript (LAT) and t he Transcript A ntisense to the LAT (TAL). The overall focus of this dissertation was to shed light on the mechanism of transcriptional repression during latency and the role that the latent transcripts play in this process. Specifically, we investigated 1) the formation of heterochromatin on the latent genome, 2) the cellular and viral factors that play a role in deposition of heterochromatin and, 3) epigenetic and/or transcription profile differences that might explain strain -specific effects on virulenc e and reactivation. Given recent epigenetic studies that correlated transcriptional stat us of the latent HSV 1 genome with histone posttranslational modifications, we investigated the hypothesis that establishment of heterochromatin i s a mechanism of transcriptional repression The wild type latent genome was found to harbor the heterochromatic marks trimethyl H3K27, trimethyl H3K9 and the histone variant macroH2A The s e data indicated that latent HSV 1 gene repression in volves an epigenetic means. The heterochromatin mark trimethyl H3K27 indicates the
10 interaction of repressive cellular proteins, Polycomb group (PcG) proteins, with the HSV 1 genome. ChIP analysis on the latent genome revealed binding of Bmi1, a member of the polycomb repressive comp lex 1, was present on many regions of the latent genome but was most enriched at the LAT enhancer region of the viral genome. Analysis of a LAT negative virus revealed increased deposition of trimethyl H3K27 signifying the importance of the LAT region in proper heterochromatin deposition. Thus, a novel mechanism of repression in HSV 1 was identified. HSV1 strains 17syn + and KOS differ in biological phenotypes such as virulence, reactivation and in the number of viral genomes per cell that establish latency in neurons. The d ifferences observed may be a result of altered c hromatin deposition, specifically of dimethyl H3K4, which may be due to variatio ns in the amount of latent transcripts derived from the LAT region. The LAT promoter in 17syn + is more dimethylated than the LAT enhancer ( and also the TAL promoter), while the converse is true for strain KOS. The amount of LAT accumulation during latenc y was also dependent on the strain of virus, with 100 -fold more LAT in 17syn + infected ganglia than in KOS infected ganglia Differences in transcriptional control of the LAT region may provide mechanistic insight to the strain -specific phenotypes in virulence and reactivation Overall, t he result s of this dissertation provide a potential model for the establishment of latency. As the genome is entering latency, cellular repressive proteins associate with HSV 1 catalyzing transcriptionally repressiv e histone posttranslational modifications. To ensure transcriptional repression remains during latency, PcG proteins associate with the latent genome to maintain trimethyl H3K27. Though it is not fully understood, the LAT region plays a role in establish ing repressiv e trimethyl H3K27 marks and can limit deposition on the genome. Future
11 studies will be required to determine if the LAT acts through cis elements within the region, the act of transcribing through the LAT region, or through the non-coding RNA s t hat are transcribed from within the region.
12 CHAPTER 1 INTRODUCTION Herpesv iruses The Herpesviridae family consists of large double -stranded DNA viruses that infect a broad range of hosts yet are species -specific. Uniquely, these viruses establish a life long infection wi thin the host by silencing viral gene transcription and replication existing in a state referred to as latency These do rmant viral genomes do not integrate into the host genome but rather exist in multiple copies as extrachromos o mal circular episomes within the nucleus of the infected host cell This latent state, however, is not permanent and is subject to periodic insta nces of reactivation allowing for infectious virus to be produced. Human herpesviruses can be divided into three sub families: and These subf amilies are differentiated by such characteristics as host range, genome organization, length of replicat ion cycle and the site of latency herpes viruses establish latency in sensory neurons; -herpesviruses can establish latency in several cell types including but not limited to myeloid cells, secretory glands, and lymphoreticular cells ; an d -herpes viruses establish latency in lymphocytes. This dissertation will focus on latency of the human Herpes Simplex Virus Type 1 (HSV -herpesvirus family Herpes Simplex Virus Type -1 HSV1 is a human herpesvirus t hat typically causes lesions, more commonly known as cold sores or fever blisters in the orofacial region, resulting in the disease known as Herpes Labialis. These lesions can be quite painful and are often associated with a social stigma. Besides infection around th e mouth region, i t is possible to have a primary infectio n on any mucosal surface such as on the surface of the eye or the genitals Recurrent r eactivations in the eye result in the disease Herpe s Kera titis which is the leading cause of blindness due to a n
13 infectious agent in the U.S. (100) Even though there are drugs to treat the reactivating virus, they are not always effective. The only drugs currently available to treat the clinical symptoms of an HSV 1 infection are in the acycloguanosine c lass of nucleotide analogs and function only when the virus is actively replicating. There are no known therapies that can target the latent genome and deplete the reservoir of virus. Therefore, having a better understanding of how the virus enters laten cy could advance our knowledge on molecular control of gene expression and may ultimately help devise a strategy to attack the virus. The scope of this dissertation will focus on epigenetic factors involved in the repression of lytic gene transcription during the establishment and maintenance of the latent HSV 1 genome. Infection with HSV 1 is ubiquitous throughout the population, with 68% of Americans over the age of 12 having detectable antibody in their serum (86) ; this increases to 90% by age 70 (89, 101) Though the majority of people are infected with HSV 1, it is estimated that only 1 6 % actually experience symptoms of clinica l reactivation resulting in the above stated disease s (55) However, clinical symptoms do not have to be present to signify shedding of infectious virus. In fact, in a clinical study 98% of the subjects (50 total) were found to shed HSV 1 DNA at least once in their saliva or tears over a 30 -day study (41) Lytic Infection During an initial infection with HSV 1 in the oral mucosa, the virus actively replicates in e pithelial cells. Entry into the cell involves binding of viral glycoproteins to receptors on the cell surface followed by fusion of the viral envelope with the cell membrane. This stimulates release of the viral capsid into the cell cytosol. The capsid travels by use of the microtubule cytoskeleton and docks at the nuclear pores to allow release of the viral DNA into the nucleus of the cell. Replication of the HSV genome results in new virions that will then spread to
14 surrounding sensory neurons termini that innervate the epithelial region and travel via fast axonal transport to the cell body of the neuron (88) Upon entering the nucleus of the neuron the viral DNA can either assume an actively replicating, lytic, state or enter latency by repres si ng the lytic genes (Figure 1 2 ). The factors that determine whether a genome ensues an actively transcribing and replicating track (lytic) vs a transcriptionally silent and non -replicating track (latent ) are unknown. If the genome takes the lytic path of active transcription, it does so with the use of viral and cellular factors. VP16, a virion protein that travels with the capsid as a part of the tegument recruits the cellular factors host cell factor (HCF) and the transcription factor, Oct 1 T his complex binds to the promoters of immediate early (IE) genes through the DNA sequence TAATGARAT recognized by Oct 1 while the VP16 activation domain stimulates transcription (58) Lytic replicati on is dependent on a temporal cascade of gene expression in which IE genes are required to be transcribed and translated in order to stimulate expression of the early (E) genes which are then necessary for expression of the late (L) gene class. IE genes in general are transcriptional transactivators needed for transcription of some IE genes but also for expression of all E gene s The five immediate early genes are ICP0, ICP4, ICP22, ICP27, and ICP47. Most important is the IE gene ICP4, without which all viral transcription and replication is inhibited. ICP4 along with ICP0 ICP22 and ICP27, function s in modulating viral gene expression. Additionally, ICP0 functions in altering the cellular nuclear compartment by disrupting ND10 bodies including the promyelocytic leukemia (PML) protein through the use of its E3 ubiquitin ligase RING finger domain (24, 35) This disruption may facilitate formation of viral replication co mpartments and i s thought to play a role in evading the immune system by disabling the host cell s interferon response (8, 9, 80, 93)
15 E genes encode proteins generally needed for vira l DNA replication while L genes represent capsid and structural proteins needed for assembly of new virions. Some key E genes include the viral DNA polymerase (UL 30), the origin binding protein (UL 9), the single stranded DNA binding protein (ICP8 ) and t hymidine kinase (tk) a phophotransferase involved in maintaining nucleotide pools for DNA replication. Examples of late genes include structural proteins such as the many glycoprotein s ( i.e. gB, gC, gD, ) and some tegument proteins such as VP16. All together, the viral genes act in coordination to produce proteins necessary for DNA replication and packaging into viral progeny Latent Infection After viral entry into some neurons, latency is established instead of a lytic infection. I t is thought th at in these cells, the linear genome circul ar izes immediate ly after entry into the nucleus and exists as an extrachromosomal episome that is associated with nucleosomes (81) A ll transcription is repressed with the exception of the Latency -Assoc iated Transcript (LAT) (Figure 1 1, 1 2 ). This abun dant latent phase transcript consists of a n 8.3 8.5 kb non-coding RNA from which a 2.0 kb stable intron with a half life of 24 hours can be spliced (Figure 1 1) (25, 94) Further, splic ing of the 2.0 kb intron can occur to yield a 1.5 kb intron, present in a subset of neurons. The reason for accumulation of the 2.0 kb intron is not known, nor is a phenotype associated with HSV1 mutants that do not acc umulate the intron (66) To assay if the LAT intron cont rols lytic gene transcription and therefore reactivation, LAT intron mutants were generated and assayed for reactivation in the rabbit ocular model (described below) The LAT intron mutants were found to have the same levels of reactivation as the wild type virus (39, 73, 75) Additionally, LAT mutants that do not reactivate s till express the 2.0 kb intron (5, 36, 59, 73)
16 Recently, a latent T ranscript A ntisense to the L AT called TAL was identified The promoter of this transcript is located around the LAT promoter/5 exon and produces a 1.7 kb polyadenylated RNA (Figure 1 1) (Giordani et al. in prep). Though it is still being investigated, the TAL might be involved in reactivation since a reactiv ation defective mutant which has a portion of the 5 exon deleted, diminishes expression of the TAL but not the LAT (Giordani et al. in prep). Functions of the LAT The LAT region and transcript have been implicated in several viral functions. Pe rhaps the strongest evidence of a phenotype is that of reactivation. When LAT promoter deletion mutants are assayed for in vivo reactivation within the rabbit and mouse models, lower levels relative to wild type HSV 1 are observed. Studies aimed at deter mining the region of the LAT that plays a role in efficient reactivation determined that only the first 1.5 kb of the primary transcript i s required for wild type levels of reactivation (5, 73) Further analyses narrowed this region down to the first 699 bp of LAT which i s denoted the reactivation critical region (39) However, it is not resolved if this reactivation phenotype is a result of cis elements within the region, a function of the actual transcript or perhaps a combination of both. Additionally, the same 5 end of LAT responsible for wild type reactivation levels has also been shown to have an apoptotic function. A LAT mutant, dLAT2903 which deletes the LAT promoter through the 5 end of the intron was found to have large numbers of TUNEL positive neurons in the rabbit 7 days post -infection while wild type had none detected (74) Therefore, it can be concluded that a region of LAT is able to protect neurons from virus -induced death. Recently, the LAT RN A was shown to encode microRNAs (miRNAs) (96) These little RNAs (21 23 nucleotides) are thought to play a role in regulating gene expression post transcriptionally. miRNAs, in general, function by binding to target sequences in the 3 UTR of
17 a primary transcript and either inhibit translation or direct mRNA degradation (78) Of th e four total miRNAs predicted to be processed from the LAT primary transcript, one has a demonstrated HSV 1 lytic transcript target. The miRNA antisense to the end of ICP0, miR H2 3P, is able to reduce ICP0 protein levels when a miRNA -containing plasmid a nd an ICP0 containing plasmid are co -transfected into 293 cells. A similar phenomenon was observed with the miRNA miR -H6, derived from a transcript upstream of the LAT, which can reduce ICP4 protein levels in vitro (96) To date, however, hairp in deletions of these miRNAs have not been generated in the context of HSV 1 recombinants in order to determine whether their absence results in altered acute or latent infections Overall, it would seem that the LAT region serves not one function but ma ny that are intricately connected therefore making it difficult to tease each function apart. Latent Transcription As mentioned above, the only abundant transcription detected during latency is that of the LAT. However, by sensitive RT PCR analyses, some lytic gene transcripts can also be detected at very low levels. Since these transcripts are detected when performing whole ganglia RNA analysis, it is difficult to determine if those lytic gene transcripts are derived from a few genomes that are undergoin g spontaneous reactivation or if all the HSV 1 genomes exhibit a general leakiness in lytic gene expression. When latently infected mouse TG were analyzed by immunohistochemistry for the presence of viral proteins, six neurons out of a total 38 TG assay ed were positive (61) Therefore, this data would support the model for spontaneous reactivation but does not exclude the possibility that all of the latently -infected neurons may be undergoing low level lytic transcription. It has also been shown that the LAT transcript regulates the abundance of low level lytic transcripts during latency and is strain -specific In the mouse model, transcript abundance from
18 lytic genes was compared between HSV 1 strain KOS and a KOS LAT mutant, dl LAT1.8. In the absence of the LAT, an increased number of lytic gene transcripts were observed for the IE gene ICP4 as well as the E gene tk suggesting that the LAT acts as a repressor on lytic gene transcription during latency (11) This was not the same tre nd observed in the rabbit model using HSV1 strain 17syn a tighter repression of lytic gene expression was reported as IE gene ICP 4 and E gene tk had a decreased transcript abundance lev syn + (31) Therefore, it would seem that in the rabbit, the L AT acts to de repress lytic genes while in the mouse it acts to repress lytic genes. The reason behind this difference in transcriptional control is not yet known. It is however possible that the LAT effect is not due to a difference in model, but a diff erence between strains 17 syn + and KOS. This issue is discussed in Chapter 4. Even though the LAT is known to be abundantly transcribed during latency, studies have shown that this transcript is not expressed in every infected cell. In Mehta et al las er capture of individual cells along with in situ hybridization for the LAT and PCR indicated that only 1/3 of the cells in which viral DNA could be detected also had signal for the LAT intron (62) Furthermore, LAT expression does not seem to be dependent on the number of genomes in the cell as Chen et al. demonstrated that there is considera ble overlap in the number of genomes between LAT positive and LAT negative latent cells (12) This would suggest that transcriptional control within all cells is not the same but that currently unknown factors, perhaps some neuronal -specific ones, contribute to regulation of L AT expression. Animal Models As HSV 1 latency has not been reliably demonstrated in cell culture models, researchers have relied primarily on animal models It should be noted, however that it is possible to generate a quiescent infection by applying inhibitors o f HSV replication immediately after infection of
19 virus in certain cell lines or primary cultures Quiescent models have been generated using rat pheochromocytoma (PC12) cells (16, 92) human foreskin fibroblasts (38) and primary dorsal root neurons (3, 38, 99) Quiescent infections do not completely mimic latent infections in vivo but these systems do have similar features such as transcriptional repression of lytic genes and the express ion of the LAT. Therefore, these quiescent infectio ns may represent model s for studying biochemical and molecular aspects of HSV 1 latency in a more homogeneous cell system. However, since these cell culture models still need to be proven as reliable models of latency, the mouse footpad and rabbit ocular mode l s are commonly u tilized for laten t studies. Although there are other animal models used by research groups, the work in this dissertation has relied on the mouse footpad and rabbit eye models, which are introduced below Mouse footpad model. Th is model involves infection of the rear footpads of the mouse allowing the virus to enter the sensory nerve termini that innervate the footpad epithelial layer (Figure 1 3 ). V irus can then travel up the nerve to the dorsal root gang lia (DRG) and establish l atency. In both animal models discussed here, the virus is considered to be latent by day 28 post infection. At this point, neither infectious virus nor lytic proteins can be detected at the original site of infection or in the DRG. Additionally, infect ious virus is not present at the periphery (original site of infection) even when the virus is stimulated to reactivate which differenti ates the mouse model from human infection. In human infection, virus can be detected at the periphery after reactivation While the reactivation seen in the mouse model differs from that of the human, the mouse model is an ideal system for studying molecular reactivation rather than clinical reactivation. Molecular reactivation refers to production of lytic transcri pts, proteins or virions that can be detected in sensory neurons after latency has been
20 established. Clinical reactivation is the appearance of symptoms at the original site of infection in the mucosal layer that is a result of virions produced by reactiv ation. A classic method for studying reactivation in the mouse is by a procedure called exp lantation induced reactivation. This involves removing and incubating the ganglia in a dish containing tissue culture medium. The first infectious virus can be detected by 12 hours post explant thereby providing a system to study effects of reactivation immediately after stimulation (76) This has enabled researchers to study various aspects of reactivation at early time points post -explant. Rabbit ocular model. To study in vivo clinical reactivation, the rabbit ocular model is utilized. Virus is applied to the eye after scarification of the cornea which allows the virus to travel up the nerve termini to the trigeminal ganglia (TG) where latency is established. As in human infections, the rabbit also exhibits spontaneous reactivation, during which virions can be d etected at the periphery. The virus may also be induced to reactivate by io n tophoretic application of epinephrine to the eye resulting in virus that is detectable in the tears of rabbits as early as two days post -treatment Therefore this model serves a s a great tool to assay mutant viruses for spontaneous induced reactiv ation events. Epigenetic Regulation Control of gene expression is a complex system in which many layers of regulation exist to ensure proper control of expression. Besides DNA cis elements, other layers of control have been identified such as protein binding, DNA methylation chromatin formation and deposition of post translational modifications (PTMs) (79) It is important to note that these other layers do not affect gene expression by alteri ng the DNA sequence but rather alter ing expression through proteins interacting with the DNA.
21 DNA M ethylation The covalent modification of DNA by the transfer of a methyl group to a cytosine at CpG dinucleotides is common in vertebrates This action results in gene silencing due to the physical blockage of trans criptional machinery, transcriptional activators or repressors and other proteins that bind to DNA in a methylation -dependent manner Using techniques such as bisulfi te sequencing, researcher s can determine if methylation is present on the DNA as any unmethylated cytosines are converted to the nucleotide uracil. This procedure along with a technique utilizing methylation -sensitive restriction endonucleases followed by Southern blot hybridiz ation, was used to look at the methylation status of the latent HSV 1 genome. No significant methylation exists on HSV1 DNA, specifically on the ICP4 and LAT locus during either acute or latent infection (22, 48) Therefore, gene expression of HSV 1 is not controlled by DNA methylation Chromatin The nucleoprotein complex that is called chromatin is formed when an octamer of histone proteins consisting of two molecules each of H2A, H2B, H3, and H4 is wrap ped within 147 bp of DNA. This repeating unit across the DNA resembles a beads -ona -string structure Further compaction of the chromatin occurs when l inker H1 histones bind to and promote folding into 30 nm fibers therefore generating a higher -order chromatin structure. This action serves to help compact the DNA to fit into the physical constraints of the nucleus. In addition to providing the structure and architectu re of chromatin, all histone proteins play a role in modulating gene expression. There are also variants of the above stated histone proteins such as H3.3, H2A.X, H2A.Z, and macroH2A to name a few, that may replace their counterpart in the octamer to exert a specific function on gene expression, as well as altering chromatin structure. Though the DNA
22 is wrapped around histone proteins, it is accessible to transcription factors and the replication complex. Histone proteins can be moved around on the DNA and can also be covalently marked by posttranslational modifications ( PTMs ) to differentiate areas of active transcription vs areas of repression. PTMs are present on the N and C terminal tails o f the unprotected histone proteins that protrude out of the histone octamer. Examples of PTMs are acetylation, methylation, phosphorylation, ubiquitination and sumoylation. The comb ination of PTMs on histones act in coordination to determine the gene status and results in what is known as the histone code (90) This simply states that the possible combinations of PTMs greatly increase the potential coding information of the genome. This effectively separates chromatin based on PTMs and the transcriptional status of the gene into two categories, 1) euchromatin, or transcriptionally permissive region s, and 2) heterochromatin, or transcriptionally repressed regions (Figure 1 4 ). Euchromatic regions are in a structurally decondensed, open state allowing transcription factors access to the DNA while heterochromatin is in a condensed structure deny ing ac cess of those same factors to the DNA. The PTMs that are deposited on the chromatin of a gene can be used to classify the transcriptional status based on previous observation that they accumulate in either euchromatic or heterochromatic regions. It is no t known if the transcriptional status is responsible for the PTMs observed on the chromatin of a gene or if it is the PTMs that dictate the state of a gene. Studying PTMs on chromatin not only extends our knowledge of epigenetic regulation but can provide insight into what factors are ultimately important for gene control Given that gene expressi on is most likely a multifactor al process, it cannot be fully understood until all regulatory factors involved are revealed.
23 Types of Heterochromatin. H etero chromatin may be further divided into constitutive and facultative heterochromatin. While both represent transcriptionally inactive genes, they differ in their ge nome location, scale, PTMs, chromatin binding proteins and nature Constitutive heterochrom atin is generally found in relatively small gene poor regions as well as on repetitive sequences that maintain the tight 30 nm chromatin structure. This includes regions such as centromeres and telomer e s. Constitutive heterochromatin can be associated with a number of different PTMs but is most commonly defined by trimethyl H3 lysine 9 (H3 K9) and trimethyl H4 lysine 20 (H4 K20) Chromodomaincontaining proteins, such as heterochromatin protein 1 (HP1), can then dock at these PTMs, sp ecifically trimethyl H3 K9, and induce transcriptional repression and further the spread of heterochromatin on the gene. The important characteristic that separates the two types of heterochromatin is the nature of its reversibility. Constitutive heteroc hromatin is historically considered to be irreversible and is maintained that way throughout the life of the cell However with the discovery of H3 K9 demethylase enzymes, this description may not be completely accurate Facultative heterochromatin is in a reversible form that would allow it to revert back to euchr omatin if stimulated to do so. This type of repression is often seen over large regions, typically on developmentally regulated genes Examples of facultative heterochromatin can be found at t he inactive X chromosome and the developmental hox locus. Here, the PTM trimethyl H3 lysine 27 (H3 K27) as well as the H2A histone variant, macroH2A, are greatly enriched. Repression through trimethyl H3 K27 is established and reinforced through the binding of polycomb group proteins (PcG) therefore blocking transcriptional machinery access to the DNA. A more detailed explanation of the mechanism of PcGs will be provided in Chapter 3. The mechanism by which the histone variant macroH2A contributes to re pression can be broken down into three parts. MacroH2A interferes with histone
24 acetylation, prevents RNA polymerase II transcription initiation, and blocks nucleosome remodeling and sliding of the octamer (21) Often times, facultative he terochromatin can be mistaken for euchromatin when observed by immunofluorescence due to its varied condensed state as it does not always reta in the tight 30 nm structure (95) Overall, these types of repression have many distinct characterist ics illustrating that there is more than one way to silence a gene therefore reinforcing the idea tha t gene control is a multifactor al process. Chromatin on the Lytic HSV -1 G enome. Studying epigenetic regulation on the lytic viral genome has not been straight forward. Viral DNA does not associate with nucleosomes in the capsid (70, 77) and it is contro versial how many genomes become nucleosomal after release into the nucleus. Initial reports indicated that the DNA was largely nonnucleosomal during lytic infection as assayed by nuclease digestions (53, 54, 64, 65) but later studies using MNase digestion and ChIP indicated that nucleoso mal structures could be detected (43) Chromatin modifying enzymes and PTMs can be found on the lytic genome as well. A lot of focus has been on the role of viral proteins such as VP16, VP22 and ICP0 and their effect on histone occupancy. How ever, the role that chromatin, chromatin modifying enzymes and viral proteins have on the lytic infection is unclear In one study knocking down histone acetyltransferases p300, CBP, PCAF, and GCN5 as well as chromatin remodeling complexes BRM and Brg1 yielded no alteration in IE gene expression (49) This suggests that lytic infection is not dependent on these histone modifying enzymes. While epigenetics l ikely plays a role in the switch from the lytic cycle to latency, it is not fully understood what its purpose is during a productive infection in cell culture, or the mucosal epithelia during initial infection or reactivation events.
25 Chromatin on the La tent HSV -1 G enome. When nucleosomes were discovered wrapped around the latent genome (17) it became clear that chromatin likely play s a rol e in control of gene expression during latency. As only one transcript is abundantly expressed during latency, it is possible that the chromatin on HSV1 promotes expression of that gene while repressing the res t. To determine if the LAT locus was indeed in a euchromatic state, chromatin immunoprecipitation (ChIP) was utilized. Two common PTMs of euchromatin are acetyl H3 lysine 9, 14 (H3 K9, K14) and dimethyl H3 lysine 4 (H3 K4). Latent HSV 1 KOS genomes from mice were assayed for the presence of acety l H3 K9, K14 and found to contain the highest enrichment of this PTM on the LAT enhancer in the 5 exon. The LAT promoter was also enriched in this mark relative to the lytic genes ICP0, ICP4, and ICP27 (47, 48) This pattern was not altered in a LAT promoter mutant indicating that the active transcription of the gene was not necessary for a euchromatic state. While this PTM represents a general transcriptional permissiveness, others more directly reflect movement of RNA polymerase II across a gene such as the dimethyl H3 K4 PTM. The latent 17syn +genome from rabbits was investigated for this active transcription marker. The LAT enhancer region was greatly enriched in dimethyl H3 K4 relative to the ICP0 and ICP27 promoters (31) Like the previous study, a LAT promoter mutant was also investigated and found to still maintain an enriched euchromatin status relative to ICP0 and ICP27 but the relative amount was lower than in the wild type virus. This signifies that while active transcription does relate to the degree of PTM enrichment, there is some oth er factor or element in the LAT region that is keeping the gene euchromatic even when no transcript is being produced. (Figure 1 5 ) Since the LAT region is maintained in a euchromatic state, it was hypothesized that the lytic genes would be maintained in a heterochromatic state. The viral genome over the course of infection through the establishment of latency was monitored for the presence of dimethyl H3 K4
26 and a pseudo-heterochromatic marker dimethyl H3 lysine 9 (H3 K9) This PTM is referred to as pse udo -heterochromatic due to data indicating that this mark can be present in both types of chromatin (97) Additionally, antibodies to detecting this PTM are notorious for cross reacting with other various H3 K9 methyl states making interpretation of the data difficult. It was found that o ver time, the genome ex periences a decrease in dimethyl H3 K4 and increase in dimethyl H3 K9 on lytic genes (98) Establishment of latency seems to be correlated to chromatin status as ac cumulation of heterochromatin and decrease in euchromatin occurs at the sam e time as decrease in lytic gene transcripts. The type of heterochromatic repression, though, has yet to be determined. Since these PTMs found on the HSV 1 genome reflect what is seen on the mammalian host cells it suggests that the virus utilizes chro matin for transcriptional control in a similar manner as the host. This may be a mechanism the virus adapted from the host to help control gene expression during latency. U nderstanding the PTMs present on viral genes will shed light on chromatin as a mechanism of transcriptional control Reactivation and Chromatin If chromatin is responsible for maintaining gene expression patterns during latency, then one would expect to see a change in the overall status occurring when the virus is stimulated to reactivate. DRG were explanted for various times and changes in the chromatin were monitored by the use of acetyl H3 K9, K14 antibody The LAT enhancer experienced a decrease in acetylation that correlated with a decrease in LAT abundance following ex plant Additionally, following this event, the enrichment of acetyl H3 K9, K14 increased at the ICP0 promoter ; however no increase in ICP0 transcript abundance was observed (1). These data indicate that the chromatin profile o f the LAT enha ncer and ICP0 promoter respond to stress stimuli inducing reactivation. It is possible that this dynamic relationship between LAT and ICP0 acts as a switch
27 between the latent and lytic infection and determine s the outcome of gene expression. This response to stress was recapitulated in the mouse model using a histone deacetylase inhibitor compound called sodium butyrate. Similar to the Amelio et al paper, this system also saw a decrease in acetylation on the LAT enhancer and a simultaneou s increase in acetylation on lytic genes ICP0 and ICP4 (68) Therefore, reactivation in mouse DRG stimulated b y two different sources result in the same chromatin changes. Whether or not these alterations would have produced infectious virions is unclear as these were not longitudinal studies, but it is likely that a shift in euchromatin would be necessary to activate the lytic genes. Insulators and CTCF Due to the o verwhelming need for ordered gene expression in the cell a precise control mechanism(s) must be in place to provide. One level of regulation exists through the prese nce of insulators which are DNA cis -elements that can influence gene expression. In sit uations where actively transcribed genes are adjacent to inactive genes, a barrier must be in place to ensure that the heterochromatin does not spread into the region of euchromatin. These barriers are known as chromatin insulators and are present in all eukaryotic cells Besides acting as a chromatin boundary, insulators also have the capability of blocking an enhancer from acting on a promoter when positioned between the two (15, 30, 42) (Figure 1 6 ). The mechanism by which insulators block en hancer s is not fully understood but three models are proposed to explain how this may occur. These models are explained in this dissertation since they may explain the function of the noted insulators of the HSV 1 genome, to be discussed below. The first is known as the promoter decoy model in which the insulator mim ics a promoter by recruiting transcriptional machinery thereby creating a competition between this ps eudo-promoter and the original promoter The second model acts simply by providing a physi cal barrier therefore blocking the transcriptional machinery from reaching the promoter. When -globin 5 HS4
28 insulator was positioned between the HS4 enhancer and the target promoter, transcriptional machinery (RNA polymerase II) accumulated at the insulator (103) Whether this data supports the promoter decoy model or rather is acting as a barrier is unclear but again it is likely that the insulator is acting by both mechanisms thereby providing repetition in the con trol of gene expression if one of those functions is lost The third model which has stimulated lots of interest recently involves a looping mechanism. Insulator proteins that bind the insulator DNA element can form protein -protein interactions with oth er insul ator or nuclear proteins forming a loop and allowing those two distant elements to interact It is most likely that these models are not mutually exclusive but act in coordination in different combinations to provide their enhancer blocking functi on. A cellular protein known to bind all vertebrate insulators is called CCCTC -binding factor (CTCF). This eleven -zinc finger protein can have many different roles depending on the context of the situation such as transcriptional repression or activation and insulator function When CTCF is bound to the insulator it can bind other insulator proteins thereby assisting in looping as well as anchoring itself to proteins e nriched at nuclear structures. In forming these loops or chromatin hubs insulators binding CTCF are proposed to have a more global role in nuclear organization of the cellular genome. This action may be most important in providing cells with cell specificity as gene expression patterns define the overall function of the cell. When ChIP was performed on two different human cell types for CTCF, it was found that there was a small difference in the CTCF binding pattern suggesting CTCF binding may be regulated in a cell specific manner (4, 45) It is more likely, though, that cell specificity is due to a difference in the interaction of other insulator proteins besides CTCF In Drosophila, binding of the insulato r protein Su(HW) did not change in response to heat shock yet there was a massive change in gene
29 expression (29) Therefore, it would seem that it is the protein -protein interactions of the insulator proteins that have changed and are responsible for the alteration in gene expression patterns rather than a change in the binding of insulator proteins to the insulator region Insulators in HSV -1. T he differential enrichment of euchromatic PTMs between the LAT region and lytic genes such as ICP0, suggest s that there is a chromatin insulator acting to prevent spread of eu chromatin into the ICP0 locus. Since all vertebrate insulators are known to bind CTCF, an in silico search was performed on the HSV 1 genome lo oki ng for CTCF binding motifs. Seven cluster sites were found within HSV1 and ChIP was performed to determine if this protein was bound at the identified sites (Figure 1 6 ). All identified CTCF sites have CTCF bound to them during latency (2). Additionally, a 1.5 kb region surrounding the CTCF cluster in the LAT intron demonstrated insulator activity by enhancer -blocking and silencing in a luciferase assay (2). Furthermore, insertion of the B2 insulator (located in the intron of LAT) int o Drosophila melanogaster embryos resulted in position effects in the eye indicating that this HSV 1 insulator is functional as a barrier (10) Binding of C TCF to HSV1 insulator sites may act to organize the genome into distinct chromatin regions as well as provide a nuclear structure to the virus. It is also possible that binding of CTCF plays a role in the neuronal specificity of the LAT promoter. Aim o f Dissertation The HSV 1 lytic genes during latency have been shown to be repressed while the LAT region remains active in 1/3 of the HSV 1 DNA-containing neurons. The mechanism by which this repression occurs, is maintained and then is ultimately revers ed during reactivation is unknown. This dissertation aimed to shed light on how repression occurs through chromatin transcriptional control by identifying cellular protein interactions with the latent genome. Additionally, the effects of viral strain and dose of the initial inoculum on transcriptional control
30 were investigated by studying the abundance of transcripts and presence of chromatin markers Ultimately, the work in this dissertation will demonstrate that the LAT region not only plays a key role in the establishment and maintenance of repressive chromat in marks, but may explain the differences in transcriptional control observed between HSV 1 viral strains.
31 Figure 1 1. Map of the LAT region including the LAT and TAL transcripts.
32 Figure 1 2 Gene expression profile of HSV 1 life cycle. During the lytic phase greater than 70 gene products are expressed while in the latent phase only one gene is abundantly expressed.
33 Figure 1 3 Mouse Footpad model. Mice are infected at an LD50 on the rear footpads allowing the virus to enter the DRG. At this dose, half of the group will succumb to the infection while the other half survives. The remaining half is sacrificed after 28 days post infection.
34 Figure 1 4 T ypes of chromatin and PTMs. Nucleosomes (circles) consisting of an octamer of histone proteins, assemble on a gene. A) An open organization of nucleosomes (gree n circles) permit transcription of the gene. PTMs such as diMe H3K4 and acetyl H3K9, K14 associate onto the nucleosome tails forming e uchromatin B) A tightly close assembly of nucleosomes (red circles) deny access of transcription factors to the DNA re sulting in transcriptional repression. H eterochromatin is marked by PTMs such as triMe H3K9 (constitutive) and triMe H3K27 (facultative).
35 Figure 1 5 Chromatin profile of the latent genome. Euchromatic PTMs are marked by the green star while the heter ochromatic PTM is marked by the red star. Euchromatin is abundant on the LAT region of HSV 1 during latency.
36 Figure 1 6 Insulator function and presence in the latent HSV 1 genome. A) Barrier activity of insulators; heterochromatin (red circles) cannot spread into the region of euchromatin (green circles) due to the occupancy of insulator proteins and histone acetyl transferases (HAT) at the insulator B) Enhancer blocking activity, t he presence of the insulator between the enhancer and promoter prevents stimulation of transcription at the promoter by the enhancer. C) Location of CTCF -binding insulators within the latent HSV 1 genome are denoted by blue lines.
37 CHAPTER 2 MATERIALS AND METHODS Viruses and Cells Infections were performed by using various virus strains. Low Passage stocks of HSV 1 strain 17 syn + ( 202 bp deletion that removes the LAT transcriptional start site (18) ), -gal gene under the control of the LAT promoter inserted in the gC locus (26, 84) ), KOS/M ( H. Marsden ) KOS/29(L Feldman, 200 bp deletion that removes the LAT transcriptional start site (20, 84) ) and KD6 (19) were used for mouse infections Viruses were amplified and titrated on rab bit skin cells using Eagles minimal essential medium (Life Technologies) supplemented with 5% calf -serum, 250 U of penicillin/ml, 250 g of streptomycin/ml and 292 g of L-glutamine/ml (Life Technologies). Mouse Infections Four to six -week -old outbred ND4 Swiss mice (Harlan Sprague Dawley, Inc) were anesthetized by isoflurane inhalation and pretreated with 0.05 ml of a 10% (wt/vol) sterile saline solution injected into each rear footpad. At three hours post treatment, mice were anesthetized by intramus cular injection of 0.020 ml of a cocktail consisting of acepromazine (2.5 to 3.75mg/kg of body weight), xylazine (7.5 to 11.5 mg/kg) and ketamine (30 to 45 mg/kg). The rear footpads were lightly abraded with an emery board to remove the keratiniz ed epithe lia. A virus inoculum of 500 PFU/0. 05 ml per mouse was then applied to the feet. The virus was allowed to adsorb for 30 45 minutes while the mice remained on their backs. Mice were sacrificed at >28 days post infection and care was taken to ensure that the ganglia were removed and processed as quickly as possible postmortem (3 5 minutes per mouse).
38 Chromatin Immunoprecipitation (ChIP) ChIPs were performed as previously described (47) Briefly, DRG from three mice (6 8 from L4 6) were re moved at a minimum of 28 days post infection. Tissue was homogenized and the chromatin was cross -linked using formaldehyde (final concentration of 1%) and washed with PBS as previously described. Cells were first lysed and then treated with sonications t o shear the chromatin into fragments ranging from 300 to 800 bp, as determined by gel elecrtrophoresis. The chromatin was then incubated with an antibody and allowed to incubate overnight at 4 C while shaking. The following antibodies and concentrations were used: 5 g of Anti -trimethyl H3K27 (Millipore 07449), 5 g of Anti MacroH2A (Millipore 07 219), 5 g of Anti -Bmi1 (Millipore 05 637), 4 g of Anti trimethyl H3K9 (Millipore 07442), 10 g of Anti acetyl H3 K9,K14, and 3 g dimethyl H3K4. Antibody c omplexes were captured using salmon sperm DNA-protein A agarose beads and washed in accordance with Millipores protocol. Samples were decrosslinked, digested with RNase A and Proteinase K, and the DNA was purified (QIAGEN). Unbound and bound fractions w ere analyzed by Taqman real -time PCR in triplicate on a Step One Plus Thermocycler (Applied Biosystems) (see below) Samples were analyzed as a bound/( unbound+bound) ratio [B/(U+B)] to represent the bound/input ratio and then normalized to a cellular control B/(U+B). All ChIPS were validated using cellular targets enriched and depleted in the protein of interest. Additionally, a negative control ChIP w as performed using rabbit Control IgG (Abcam Ab46540) at a concentration of 2 g per sample. Data Analysis Methods Corrected B/(U+B) ratios normalized to a cellular control B/(U+B) for the various primers were analyzed within each immunoprecipitation usi ng a nested design in the ANOVA (63) Differences between primers among the virus strains by their respective immunoprecipitations were evaluated by a similar nested ANOVA using two factor s, immunoprecipitation antibody
39 and primer, and the interaction of these factors. Comparisons of primer means in both analyses were conducted using alpha level correct ion with a simulation method (23) Guanidine Thiocyanate (GTC) RNA Isolation RNA was is olated by guanidine thiocyanate (GTC) extraction following the protocol described by Chirgwin et al (1979) with minor modification (13) DRG were homogenized in 1 ml of GTC and brought up to a volume of 2.5 ml after washing the gr inder with GTC. An aliquot of 100 l was removed and saved for DNA extraction. The remaining homogenate was layered onto a 5.7 M cesium chloride cushion in an SW41 polycarbonate tube, and centrifuged at 30,000 x g for 12 16 hours at 15 C. The RNA pell et was recovered and dissolved in nuclease free water. This was then precipitated and treated with DNase (TURBO DNase, Ambion). The aliquot for DNA preparation was precipitated and resuspended in a solution containing Proteinase K. The enzyme was inacti vated and diluted to a volume of 500 l and was ready for PCR analysis. Reverse Transcription After DNase treatment, the RNA was divided into seven aliquots on which reverse transcription was performed. Random decamers (Ambion) primers were used in conj unction with Omniscript reverse transcriptase (Qiagen) according to the manufacturers protocol. Reactions were carried out in 30 l volume containing 10 l RNA, 1 l Omniscript reverse transcriptase, 1 M random decamer primer, 1 U Superase In Rnase inhi bitor (Ambion) and Qiagen supplied buffer and 5mM dNTP mix. A reverse transcriptase negative control containing 15 l water and 5l RNA was run side -by -side with the RT reactions for 1 hour at 37 C. The reactions were pooled and the cDNA was concentrate d using 0.1 vol 3 M sodium acetate, 0.02 vol linear acrylamide (Ambion), and 2.5 vol 100% ethanol. The cDNA samples were incubated
40 overnight at 20 C, pelleted at 14,000 for 15 minutes and resuspended in 65 L nuclease -free water. Semi quantitative PCR A nalysis DNA was quantitated using the fast run on the Step One Plus Thermocycler. Briefly, reactions were carried out in 10 l volumes consisting of 5 l 2x TaqMan Fast Universal PCR Master Mix (App lied Biosystems), 3 l water, 0.5 l primer/probe and 1. 5 l sample. Cy cle conditions were as follows: 95 C for 20 seconds (1 cycle), followed by 95 C for 1 second and 60 C for 20 seconds (40 cycles). Threshold values were set within the linear range of PCR target amplification. Cycle thresholds were deter mined after thresholds were set for each primer set using the Applied Biosystems StepOne Plus analysis software.
41 CHAPTER 3 POLYCOMB PROTEINS BIND THE HSV1 GENOME Introduction During latency, the genome is maintained as a nucleosomeassociated extrachromosomal episome in which only one gene product is abundantly expressed (81) The mechanism by which the genome establishes and maintains this latent state is currently unknown, although it is hypothesized that the reversible nature of the latent phase involves the constant interplay between cellular and viral factors that either keep the viral genome repressed or allow for reactivation. Recent studies have shed light on epigenetic modifications to the HSV 1 genome during the establish ment and maintenance of latency. PTMs representative of euchromatic regions were found enriched on the LAT region, while a pseudoheterochromatic mark, dimethyl H3 K9, was found to increase in enrichment on lytic genes as the genome established latency (98) These studies suggest that the virus can associate with cellular enzymes responsible for catalyzing histone PTMs and that these modifications on HSV 1 genes can reflect the transcriptional status. The studies describ ed above and in Ch apter 1 analyzed only euchromatic marks on the genome, with the exception of Wang et al (98) In Vakoc et al., the examined dimethyl H3 K9 PTM has been shown to be present in both types of chromatin, euchromatin and heterochromatin, hence the pseudo -heterochromatic name (97) In the present stud y, we sought to determine whether there was an association between the HSV 1 lytic genes and heterochromatin during latency. Determining the composition of the heterochromatin would not only indicate the nature of repression of the HSV1 lytic genes during latency but could also elucidate the cellular and viral factors that mediate this repression.
42 Heterochromatin can be divided into two types, constitutive and facultative (95) Though both are transcriptionally silent, each has distinct characteristics, as described above. Both types of heterochromatin are deposited and maintained by different proteins, making each method of repression unique. This helps explain the important functional difference between facultative and constitutive ; that facultative retains the ability to convert between heterochromatin and euchromatin while constitutive remains repre ssed through the life cycle of the cell. Because HSV1 can reactivate in response to stress stimuli it seems likely that the repression controlling latency is reversible. Therefore, we hypothesized that the latent HSV 1 genome is repressed, at least in part, by the deposition of facultative heterochromatin. To investigate what type of heterochromatin is present on the genome during latency, we performed chromatin immunoprecipitation (ChIP) assays on dorsal root ganglia of mice latently -infected with HSV 1 strain 17 syn +. We examined the presence of both trimethyl H3 K27 (triMe H3K27), a PTM representative of facultative heterochromatin, as well as the presence of histone variant macroH2A, which is enriched in facultative heterochromatin and can directly inhibit transcription (21) In addition, we examined the latent genome for the presence of a constitutive heterochromatin PTM, trimethyl H3 K9 (triMe H3K9). Finally, since HSV 1 recombinants lacking the LAT promoter have previously been shown to possess a leakier lytic transcription profile during latency (11) we examined the presence of facultative marks on the LAT promoter dele absence of the LAT, both triMe H3K27 and macroH2A are dramatically increased. This indicates that the LAT plays an important role in directing the chromatinization of the HSV 1 genome during latency and may explain why LAT mutants exhibit a decreased ability to reactivate from latency.
43 Results HSV -1 Lytic Genes are Enriched in triMe H3K27 To investigate if the latent HSV 1 genome was enriched in triMe H3K27, a marker of facultative heterochromatin, we performed ChIP on DRG dissected from 17 syn + latently infected mice. Relative quantities determined by real time Taqman PCR were analyzed as bound/(unbound + bound) and normalized to a cellular control, APRT, a house keeping gene. In order to verify that the immunoprecipitation worked efficiently, all ChIP samples were validated by comparing an area enriched in triMe H3K27, the intergeneic region upstream of the hoxa5 locus (denoted as upHoxa5) (44) to an underenriched cellular gene (APRT); there is a 2.8 -fold enrichment of the upHoxa5 locus relative to APRT (Figure 3 1A) Once the ChIP samples were validated, representative genes from all three HSV 1 gene classes were examined. Specifically, these genes were ICP0, ICP27 and ICP4 for immediate early, tk for early, and gC for late gene classes (Figure 3 1B). Our results indicate that triMe H3K27 PTMs are significantly enriched across the HSV 1 genome (Figure 3 1B). Interestingly, the degree of enrichment varied among the HSV 1 genes examined, with no clear correlations betwee n gene classes and accumulation of facultative marks. Immediate -early genes ICP0 and ICP27 had similar relative quantities of triMe H3K27 while ICP4 had significantly less (p<0.05, relative to ICP0, ANOVA), suggesting that ICP4 might not be regulated by f acultative heterochromatin to the same extent. This also demonstrates that heterochromatin deposition is not dependent on the gene class but is unique to each gene. Furthermore, the early gene tk and late gene gC have similar relative quantities of triMe H3K27 as ICP0 and ICP27. To verify that the triMe H3K27 antibody was acting in a specific manner, a control was performed using rabbit anti -mouse IgG antibody; the results of this precipitation confirmed that the triMe H3K27 precipitations were specific (Figure 3 1C). In summary, these data indicate that triMe H3K27
44 accumulation on the latent genome is not limited to one gene class or region of the genome but is widespread. MacroH2A is Incorporated into HSV Chromatin on Lytic Genes During Latency Because triMe H3K27 was found to be associated with the latent HSV 1 genome, we sought to determine whether another characteristic component of facultative heterochromatin was present on the latent genomes. MacroH2A is a histone H2A variant enriched in areas of facultative heterochromatin. ChIP assays for macroH2A were successfully validated by comparing a gene on the X chromosome, midline, that is known to be enriched in macroH2A to an under -enriched cellular control, upHoxa5 (7) (Figur e 3 2A) As with the triMe H3K27 ChIP assays, macroH2A was enriched on chromatin associated with all HSV 1 lytic genes and the LAT region (Figure 3 2B). Immediate early gene ICP4 appeared to be the most enriched in this histone variant relative to the ot her lytic genes but was not statistically significant suggesting that there is equal deposition of macroH2A over the lytic genes. Interestingly, a trend was observed in which the LAT enhancer was more enriched relative to the LAT promoter and several othe r lytic targets. In summary, our results demonstrate that macroH2A, a measure of facultative heterochromatin, has replaced at least some of the H2A present in the histone octamer in the chromatin on all HSV 1 gene targets. The relative amounts of macroH2 A are also independent of gene class and also are unique from the patterns of triMe H3K27 enrichment. Overall, the results of the triMe H3K27 and macroH2A ChIP experiments indicated that all regions of the HSV1 latent genome examined are enriched in these marks, and this finding suggests that facultative heterochromatin may play a role in gene silencing during latency.
45 HSV -1 Latent Genome is Enriched in the Constitutive H eterochromatin PT M triMe H3K9 The previous findings suggest that the HSV 1 genome is significantly enriched in facultative heteroch romatin. To determine if the latent genome is only enriched in facultative heterochromatin marks, we also investigated a PTM representative of constitutive (irreversible) heterochromatin. ChIPs were performed with antibody to triMe H3K9 and were validated by comparing enrich ment levels between a classic constitutively heterochromatic region (mouse centromere) to the house keeping gene APRT (67) This validation indicated that the mouse centrom ere was 3.75 -fold more enriched in triMe H3K9 relative to APRT (Figure 3 3A) Surprisingly, we found that triMe H3K9 was also enriched on the latent HSV 1 genome (Figure 3 3B). Like both facultative heterochromatin marks examined, this PTM was present ov er most gene targets examined including the LAT region (promoter and enhancer), and immediate early and early genes (Figure 3 3B). Though late gene gC did not seem to correlate with the other genes in accumulation of triMe H3K9, it was not significantly different to other viral targets examined. In summary, these data suggest that the HSV 1 genome is significantly enriched in both constitutive and facultative heterochromatin marks during latency. The implications of this are discussed in the Discussion. Transcript Levels During Latency Do Not Correlate with Heterochromatin Enrichment Since some lytic genes that were examined had higher levels of enrichment in heterochromatic marks than others (ICP4), we sought to determine if genes with more heterochroma tin were more transcriptionally repressed. Total RNA and DNA were isolated from latently infected mouse DRG using guanidine thiocyanate extraction. The samples were processed into cDNA by reverse transcription and quantitative real -time Taqman PCR was pe rformed. RNA molecules per genome were calculated for the LAT, ICP27, ICP4, tk and gC as
46 shown in Figure 3 4 As expected, the LAT primary transcript was the most abundant RNA detected, with a mean of 1.5 molecules per genome (Figure 3 4A) To visualize the differences between the lytic genes examined, we removed the LAT data set and re -graphed the lytic genes separately on a different scale (Figure 3 4B) This manner of graphing the data revealed that ICP27 had significantly higher levels of transcript s than ICP4, tk and gC (p<0.05). Furthermore, ICP4 had more molecules per genome than tk and gC, which showed similar levels of transcript accumulation (p<0.05). This suggests that not all lytic transcripts are expressed or repressed during latency to th e same degree, in agreement with previous studies (11, 31, 46) When the transcription data is compared to the heterochromatin enrichment levels described earlier, some of the investigated genes show a correlation between transcript levels and heterochromatin enrichment, in which higher enrichment of heterochromatin corresponds to lower transcript abundance. Specifically, ICP4 had the lowest level of enrichment of triMe H3K27 (Figure 3 1B) relative to other genes and a higher transcript abundance during latency relative to tk and gC suggesting a relationship between facultative heterochromatin deposition and transcript abundance levels. ICP27, however, did not show a clear correlation between trans cript abundance and triMe H3K27 enrichment as it had relatively greater amounts of heterochromatin than ICP4 but also a 4-fold enrichment in transcript abundance. A similar pattern was observed on the LAT enhancer. In summary, this indicates that the tra nscription status of only some genes is reflective of its facultative heterochromatin status. It is also possible that these transcript levels are so low that they are not able to be compared for an accurate correlation to enrichment of heterochromatin. Other mechanisms may be in place for the transcriptional repression of ICP4 and the enriched heterochromatin deposition of LAT.
47 The LAT Regulates Enrichment of Facultative Heterochromatin Marks Since LAT mutants display leakier lytic transcript abundance in the mouse, we investigated whether this is due to an altered repression through facultative heterochromatin presence of triMe H3K27 and macroH2A. I n the absen ce of the LAT, a dramatically different pattern of enrichment is seen on the latent genome than was observed for 17syn + An increase in genomes (Figure 3 5B, Tabl e 3 1) While lytic genes gC and tk showed roughly the same amount of triMe H3K27 enrichment as was observed for 17 syn +, the LAT enhancer, ICP0, ICP27 and ICP4 al showed increases in enrichment (3.0 -, 2.0 1.7 and 5.0 -fold, respectively). However, o nl y the increases noted on the L AT enhancer, ICP0 and ICP4 were significantly different between 17 syn (Table 3 1) Similarly, in the case of enrichment when compared to 17syn + (Figure 3 6B, Table 3 2) The most dramatic increases seen were enrichment on ICP0 (6 -fold), ICP27 (3.5 -fold) and tk (3 -fold), however only the lytic gene ICP0 displayed a significantly different enrichment (Table 3 2) Taken together, these differences in incorporat ion suggests that the LAT plays some role in regulating the amount of both triMe H3K27 and macroH2A that is incorporated into the viral chromatin in a gene -specific manner. PRC1 Protein Bmi1 Associates with the Latent Genome It was found that triMe H3K27 is present on the genome during latency. The only mammalian histone methyltransferase known to catalyze triMe H3K27 in vivo is EZH2, a member of the polycomb repressive complex, PRC2, that acts in accordance with another
48 multiprotein complex, PRC1 to est ablish and maintain gene repression (87) PRC2, also known as the establishment complex, acts to initially repress a gene by establishing triMe H3K27. PRC2 is later replaced by PRC1, which acts to maintain the repressive marks on the gene and is referred to as the maintenance complex. Since triMe H3K27 was enriched on the HSV 1 latent genome, this indicated that the polycomb group proteins might be binding on the genome. Therefore, we decided to investigate if one set of these cellular pro teins was interacting with the latent HSV 1 genome. Since we demonstrated that the triMe H3K27 marks are already established at latency, we hypothesized that the maintenance PRC1 complex would be the most likely PRC bound to the HSV 1 genome at this time. To determine if repression of HSV 1 lytic genes during latency might be regulated by polycomb-mediated repression, we performed ChIP assays with an antibody to Bmi1 to look for association of this protein with the latent genome. Validation of this ChIP was performed by comparing a cellular gene known to bind PRC1 (HOTAIR) to a cellular gene that does not bind this protein (myo1D) (44) (Figure 3 7A ) Comparison to negative controls included ChIP using mouse IgG serum (Figure 3 1C) as well as an internal viral control, the intron of ICP0 (Figure 3 7B). Analysis of the latent HSV 1 genome showed enrichment for all regions examined except for the intro of ICP0 (Figure 3 7B) with the LAT enhancer region having the greatest enrichment of Bmi1 (p<0.05). The LAT enhancer was also more greatly enriched than the LAT promoter and ICP27 (p<0.05). This indicates that while Bmi1 binds to the latent viral genome, it may be binding to specific regions, such as the LAT enhancer. In summary, this data indicates that PRC1, the maintenance P RC complex, binds to the HSV 1 genome during latency, consistent with the presence of the triMe H3K27 facultative heterochromatin mark.
49 Discussion The mechanism behind repression of the HSV 1 lytic genes during latency is presently unclear. We sought to shed light on this issue by investigating the epigenetic histone marks associated with the lytic genes. We hypothesized that the lytic genes would be enriched in facultative heter ochromatin due to its reversible nature. By performing ChIP, we found that not only was the genome enriched in facultative marks (triMe H3K27 and macroH2A) but also in constitutive PTMs (triMe H3K9). The presence of heterochromatin confirms that cellular enzymes interact with the viral genome to confer repression during latency. Since both kinds of heterochromatin are present on the genome, it indicates interaction with two different kinds of histone methyltransferases (HMTs). Usually the HMT EZH2 (trim ethylates H3K27) of the polycomb repressive complex and SUV39H (trimethylates H3K9) catalyze methyl marks in facultative and constitutive heterochromatin, respectively. Our data therefore indicates that there must be two independent mechanisms acting to r epress HSV 1 genomes: one t hat promotes facultative heterochromatin and on that promotes constitutive heterochromatin. It is not clear, however, how these marks on the viral genome are divided. It is possible that these marks are present on different gen omes within the same latently -infected cell or that all of the marks on the genomes within a given cell are on the same, resulting in cells with facultative marks and cells with constitutive marks. We also noted that the LAT enhancer, previously characte rized as euchromatic, also had enrichment of heterochromatic marks. While this could indicate that both types of chromatin marks are on the same region of DNA, it seems more likely that the presence of heterochromatin indicates that there are populations of genomes, some associated with euchromatin while some are associated with heterochromatin It has been estimated that only one third of latent genomes express the LAT meaning that two thirds of the genomes do not express LAT (33, 62) The
50 difference in transcription patterns between these two populations of genomes (LAT positive and LAT negative) may explain the difference in chromatin deposition. Genomes expressing LAT would be associated with euchromatic marks while those not expressing LAT would be associated with heterochromatin. These data support previous papers indicating that there are two populations of latent genomes within sensory ganglia that exhibit differing LAT expression (12, 62) The presence of both types of heterochromatin on the LAT enhancer region raises several questions. One hypothesis to explain this observation is that the nonLAT expressing genomes can be fur ther divided into two sub -categories: those associated with constitutive PTMs and those associated with facultative PTMs. The genomes associated with the latter may be more prone to reactivation, even without transcription of the LAT, due to the reversibl e nature of this type of chromatin. Alternatively, both marks may be present on the same histone protein. In Arabidopsis, both methyl H3K9 and H3K27 can be found on the same H3 protein and are required for binding of certain proteins such as the DNA methyltransferase CHROMOMETHYLASE3 (56) Further experiments would need to be completed at the single cell level to help shed light on this issue. Even then, it might be difficult to interpret the results if genomes within the same cell harbor differing PTMs. It is also possible that in terms of the LAT enhancer, one copy of the LAT gene within the genome may be associated with one PTM while the other copy is as sociated with a different type. Another less likely explanation for the presence of both marks is that trimethyl H3K9 incorporation into the genome assumes a different function than in mammalian cells. C onstitutive heterochromatin in plants is marked by methyl and dimethyl H3K9, not trimethyl H3K9. The virus may have adapted its own definition of histone PTMs in the viral chromatin meaning that trimethyl H3K9 may be representative of
5 1 something other than constitutive heterochromatin (i.e. euchromatin). However, this seems unlikely because the host cell is mammalian and the relative quantities between the LAT enhancer and all the lytic genes examined are similar. To determine if the LAT, the abundant non-coding RNA expressed during latency, affected enric hment in heterochromatic marks we performed ChIP assays on the LAT mutant trime H3K9 between 17syn (31) the accumulation of triMe H3K27 and macroH2A incorporation. We found a signif icant increase in enrichment of both facultative marks in the LAT promoter mutant. The increased amount of facultative heterochromatin in the absence of LAT suggests that the lytic genes are in a state more ready to revert back to euchromatin than they ar e in 17 syn +. This correlates with transcription studies in which LAT was shown to have a repressive function and in the absence of LAT, a leakier transcription profile during latency was observed (11) While it is clear that the LAT has an effect on the for mation of chromatin on the genome, it is not clear if it the LAT RNA itself, the act of transcribing the LAT, or the DNA -binding balance of facultative heteroch romatin markers is maintained by interaction of polycomb proteins with trithorax proteins (34) It seems plausible that deletion of the 202 bp segment of the LAT promoter may have ablated binding of a component of the trithorax protein complex therefore causing misregulation of the facultative heterochr omatin formation on the latent HSV 1 genome. Based on the evidence of triMe H3K27 enrichment of the HSV 1 genome we investigated if polycomb group proteins were involved in the maintenance of these marks during
52 latency. To demonstrate that the PRC1 compl ex interacts with the HSV 1 genome during latency, we performed ChIP on latentlyinfected mouse DRG using an antibody to a component of the PRC1 complex, Bmi1. Our data indicated that distinct regions of the genome exhibit enrichment of this polycomb prot ein. These experiments suggest that PRC1 binds to the genome and maintains the repressive triMe H3K27 marks present on the genome. Overall, this displays a new mechanism for the establishment and maintenance of latency through the use of the cellular pol ycomb group proteins. Repression of lytic genes on the HSV 1 latent genome seems to be of a complex nature that is partly controlled through epigenetic regulation. It is possible that heterochromatin formation on the latent episome is a default host defense mechanism to silence exogenous DNA, and the virus has evolved a mechanism to alter this process and promote preferential deposition of heterochromatic marks that favor reversible silencing of the lytic genes. Control of facultative heterochromatin may be due in part to LAT -binding components of the trithorax protein group to regulate chromatin deposition. By introducing a certain heterochromatic marker onto the genome such as triMe H3K27, the virus partakes in the cellular mechanism of repression and maintenance of the repressive marks on the genome by interacting with polycomb complexes. The virus is then able to use the PRC1 complex to maintain the repressive marks on the genome, which would be advantageous in controlling reactivation from laten cy. However, the genes are not locked into a repressive state, since the PRC1 proteins can be removed, allowing transcription factors to access the DNA. To our knowledge, this is the first example of a virus interacting with cellular proteins of the poly comb group to repress transcription on its genome. At the time that this work was published online in JVI, another paper by Cliffe et al. also came out showing similar results (14) This paper also exam ined triMe H3K27 on the latent
53 HSV1 genome and found accumulation of this facultative heterochromatic marker. However, upon examining a LAT promoter mutant, a different effect on triMe H3K27 accumulation was identified In the absence of LAT transcripti on, a decrease in enrichment of this PTM was observed indicating that the LAT plays a role in depositing this mark on the genome. While this is in contrast to the increase of this mark that our data revealed with our LAT promoter mutant, it is clear that the LAT is indeed playing a role in the heterochromatinization of the latent HSV 1 genome. It is possible that the differences we observed were due to the virus strain as we used 17syn + while Cliffe et al. used a KOS background. The implications of this are discussed in Chapter 4 in which we did in fact document a difference in chromatin between virus strains. Additionally, it could be due to the site of latency since our model involves the latent genome in the DRG while Cliffe et al. analyzed the trigem enial ganglia (TG). Based on preliminary data, it would seem that the site of latency also plays a role in heterochromatin accumulation. Lastly, since Cliffe et al. did not show positive to negative enrichment values for their trimethyl H3K27 ChIPs, it i s possible that they were not efficient or specific to the mark they investigated.
54 Figure 3 1. TriMe H3K27 is enriched on the 17syn + latent genome. A) Validation of the ChIPs using anti triMe H3K27 was performed by analyzing bound and unbound fractions of ChIP by real -time PCR with positive control PCR primers/probe to upHoxa5 compared to a negative control PCR primers/probe, APRT. Data are graphed as the B/(U+B) ratio. Average fold enrichment between upHoxa5 and APRT is denoted by the brack et B) ChIPs using anti triMe H3K27 were subjected to real time PCR using primers specific for the HSV 1 target genes indicated and the results were graphed as B/(U+B) ratio normalized to APRT B/(U+B). Mean values are displayed for each gene ; data from six independent ChIPs are graphed (n=6) C) ChIPs of latently infected mouse DRG using control IgG was performed and analyzed as described for panel B. Mean values for each gene are shown ( n=3 ). L. pro, LAT promoter; L. enh, LAT enhancer.
55 Fig ure 3 2. Histone variant macroH2A is incorporated into the viral chromatin on the 17syn + latent genome. ChIP of histone variant macroH2A on 17syn + latently -infected mouse DRG. A) Validation of the ChIPs using anti -macroH2A were performed by analyzing bound an d unbound fractions of ChIP by real -time PCR with positive control PCR primers/probe to midline compared to negative control PCR primers/probe, upHoxa5. Data were graphed as the B/(U+B) ratio. Average fold enrichment between midline and upHoxa5 is denote d by the bracket. B) ChIPs using anti -macroH2A were subjected to real time PCR using primers specific for the HSV 1 target genes indicated, and the results were graphed as B/(U+B) normalized to APRT B/(U+B). Mean values for each gene are shown; data from six independent ChIPs are graphed (n=6). L pro, LAT promoter; L. enh, LAT enhancer.
56 Figure 3 3. triMe H3K9 is enriched on the 17syn + latent genome. A) Validation of the ChIPs using anti triMe H3K9 were performed by analyzing bound and unbound fr actions of ChIP by real -time PCRs with positive control PCR primers/probe to mouse centromere compared to negative control PCR primers/probe, APRT. Data were graphed as the B/(U+B) ratio. Average fold enrichment between mouse centromere and APRT is denot ed by the bracket. B) ChIPs using anti -triMe H3K9 were subjected to real -time PCR using primers specific for the HSV 1 target genes indicated, and the results were graphed as B/(U+B) normalized to APRT B/(U+B) Mean values for each gene are shown; data f rom five independent ChIPs are graphed (n=5). L pro, LAT promoter; L. enh, LAT enhancer.
57 Figure 3 4. Viral transcript abundance in 17 syn + latently infected mouse DRG. RNA was isolated from latently infected DRG, reverse transcribed into cDNA, an d analyzed by real time PCR as described in Materials and Methods. Data are shown as RNA molecules per genome. A) Results of PCR analysis for the LAT (5 exon) and lytic genes. B) Results of PCR analysis for the lytic genes (note different scale).
58 Fi gure 3 of the ChIPs using anti triMe H3K27 were performed by analyzing bound and unbound fractions of ChIP by real -time PCR with positive control PCR primers/probe to upHoxa5 compared to negative control PCR primers/probe, APRT. Data were analyzed as the B/(U+B) ratio. Average fold enrichment be t ween upHoxa5 and APRT is denoted by the bracket. B) ChIPs using anti triMe H3K27 were subjected to real time PCR using primers s pecific for the HSV 1 target genes indicated, and the results were graphed as B/(U+B) normalized to APRT B/(U+B). Mean values for each gene are shown; data from five independent ChIPs are graphed (n=5). ChIPs of latently infected mouse DRG using control IgG was performed and analyzed as described for panel B. Mean values for each gene are shown (n=2). L. enh, LAT enhancer.
59 Figure 3 genome. A) Validation of the macroH2A ChIPs were performed by analyzing bound and unbound fractions of ChIPs by real time PCR with positive control PCR primers/probe to midline compared to negative control PCR primers/probe, APRT. Data were analyzed as the B/(U+B) ratio. Average fold enrich ment between midline and APRT is denoted by the bracket. B) ChIPs using anti -macroH2A were subjected to real -time PCR using primers specific for the HSV 1 target genes indicated, and the results were graphed as B/(U+B) normalized to APRT B/(U+B). Mean va lues for each gene are shown; data from five independent ChIPs are graphed (n=5). L. enh, LAT enhancer.
60 Figure 3 7. Bmi1 is enriched on the 17syn + latent genome. A) Validation of the ChIPs using anti Bmi1 were performed by analyzing bound and un bound fractions of ChIPs by real time PCR with positive primers/probe to HOTAIR compared to negative control PCR primers/probe, myoD1. Data were graphed as the B/(U+B) ratio. Average fold enrichment between HOTAIR and myoD1 is denoted by the bracket. B) ChIPs using anti Bmi1 were subjected to real time PCR using primers specific for the HSV 1 target genes indicated, and the results were graphed as B/(U+B) normalized to APRT B/(U+B). Mean values for each gene are shown; data from eight independent ChIPs are graphed (n=8). L. pro, LAT promoter, L. enh, LAT enhancer.
61 Table 3 1. Mean Normalized Values of triMe H3K27 Enrichment for 17 syn Primer Target B/(U+B) 17 syn + a B/(U+B) b Fold Difference c syn + P valued LAT promoter 9.84 NA LAT enhancer 9.94 29.44 2.96 <0.0001 ICP0 12.80 25.95 2.03 <0.0001 ICP27 8.96 15.72 1.75 0.35 ICP4 4.70 23.55 5.00 <0.0001 tk 9.66 13.82 1.43 0.95 gC 11.51 9.23 0.80 1.00 a Mean values of B/(U+B) values normalized to APRT B/(U+B) are listed from ChIPs using anti triMe H3K27 in 17syn + latently infected mouse DRG (Figure 3 1). b Mean values of B/(U+B) values normalized to APRT B/(U+B) are listed from ChIPs using anti triMe H infected mouse DRG (Figure 3 5). NA, not applicable. c syn + B/(U+B) values. d P values between the values in columns 2 and 3 were determined by ANOVA (see Materials and Me thods Table 3 2 Mean Normalized Values of macroH2A Enrichment for 17 syn Primer Target B/(U+B) 17 syn + a B/(U+B) b Fold Difference c syn + P valued LAT promoter 5.96 NA LAT enhancer 8.39 9.50 1.13 1.00 ICP0 1.83 11.32 6.19 0.016 ICP27 2.16 7.70 3.57 0. 85 ICP4 4.22 4.60 1.09 1.00 tk 1.67 5.14 3.08 0.9 9 gC 2.11 3.58 1.70 1.00 a Mean values of B/(U+B) values normalized to APRT B/(U+B) are listed from ChIPs using anti -macroH2A in 17 syn + latently infected mouse DRG (Figure 3 1). b Mean values of B/(U+B) values normalized to APRT B/(U+B) are listed from ChIPs using anti -infected mouse DRG (Figure 3 5). NA, not applicable. c syn + B /(U+B) values. d P values between the values in columns 2 and 3 were determined by ANOVA (see Materials and Methods
62 CHAPTER 4 DIFFERENCES IN TRANSCRIPT LEVELS IN THE LAT REGION OF HSV1 STRAINS 17SYN + AND KOS CORRESPOND TO DIFFERENCES IN EUCHROMATIN DEPOSITION Introduction HSV1 establishes a lifelong infection in sensory neurons. During this latent stage, the only abundant transcript is the latencyassociate d transcript (LAT). Lytic transcripts can be detected, though by use of sensitive RT -PCR techniques. Whether this is due to a few neurons undergoing reactivation or if all genomes are experiencing low levels of lytic transcription is unknown (11, 27, 46) The LAT has been shown to have an effect on t ranscript abundance from lytic genes during latency depending on the animal model it is studied in. In the mouse ocular model, the LAT was found to have a repressive role over the lytic genes as transcript abundance for those genes increased in a LAT prom oter mutant (11) Contrary to this finding, a LAT promoter mutant in the rabbit ocular model was found to have decreased lytic transcript abundance suggesting that LAT is needed to keep the lytic genes in a de repressed state (31) Differences between the two models are actively being studied to determine which specific factors may play a role in the differing mechanism s of LAT. Another difference between the two LAT promoter mutant studies is that they each utilized a different strain of HSV 1. The studies in the Chen et al. paper used the KOS strain and LAT promoter mutant KOS/29 while Giorda ni et al used the HSV1 strain 17 syn + and the corresponding (11, 31) There are wide variations in the pathogenic properties of HSV1 strains that have been isolated including differences in virulence and reactivation phenotypes (37) HSV 1 strain 17syn + exhibits high levels of reactivation while HSV 1 strain KOS does not (37) Additionally, differences in establishment of latency have been noted. While the amounts of total viral DNA per latently infected ganglion from mice infected with both viruses are similar, th ere is a statistical difference in the distributions of genomes per cell
63 (83) KOS infected neurons had significantly fewer genome copies per cell than 17syn + infected neurons suggesting a link between establishing latency with a high number of genomes and a higher reactivation potential (83) The above mentioned data suggest that phenotypic differences between HSV 1 strains 17syn + and KOS may be controlled at the transcriptional level due to either sequence -specific affects (altered cis -factors) between the strains or by the distribution of genomes per cell. In an attempt to understand the factor(s) that may be involved in the transcriptional control of latent HSV 1 gene e xpression we analyzed transcript levels as RNA molecules per genome for both KOS and 17syn + infected mouse DRG The results of these analyses indicated that the only difference in transcript abundance between the two strains originated within the LAT region. This suggested that transcription from the LAT region, known to play a role in facilitating efficient reactivation from latency, was regulated differently between these two strains. Gene expression is regulated on many levels incl uding the complex system of chromatin formation. Previous studies have shown that the latent genome i s associated with nucleosomes and that th e transcriptional status of a gene can be assessed by examining histone posttranslational modifications (PTMs) (79) Therefore in order to assess whether there were fundamental differences in the PTMs associated with the LAT region between strains KOS and 17syn +, we looked at a marker of euchromatin, or transcriptional permissiveness, dimethyl H3K4 (diMe H3K4) as well as a marke r of facultative heterochromatin, or transcriptional repression trimethyl H3K27 (triMe H3K27). Similar to the transcript analysis, we saw that the main difference between the two viruses was the enrichment in the LAT region. Furthermore, our analyses of heterochromatin revealed a greater enrichment of the triMe H3K27 mark on KOS
64 (5 0) suggesting that KOS lytic genes are more tightly repressed than 17syn + during latency. In summary, this study on the chromatin between two phenotypically different strains has revealed that differences in gene regulation are apparent at the level o f chromatin. These al tered enrichments in euchromatic marks in the LAT region and heterochromatin marks over the entire genome may reflect how strain KOS had a reduced ability to reactivate from latency. Results Facultative H eterochromatin PTMs are Increased on Lytic Genes in KOS Relative to 17syn + Since HSV 1 strain KOS has a low reactivat ion phenotype, we hypothesized that this may be directly related to the enrichment of facultative heterochromatin on lytic genes. ChIP s were performed on murine DRG latently -infected with KOS f or triMe H3K27 on six independent samples Relative quantities determined by real time Taqman PCR were analyzed as B/(U+B) and normalized to the cellular control APRT. These ChIP experiments were validated using a PCR prim ers/probe positive control for the cellular gene myoD1 relative to a negati ve control APRT resulting in a 3. 2 1 fold enrichment indicating that the precipitation was efficient (Figure 4 1A) Consistent with the analysis of heterochromatin on 17syn + (Chapte r 3, (50) ) the KOS latent genome was enriched in triMe H3K27 on all the targets examined (14, 50) (Figure 4 1 B) What wa s striking about the results is the relative amount of enrichment on KOS lytic genes relative to 17 syn +. Lytic genes ICP27, ICP4, tk and gC had greater amount of triMe H3K27 than their 17syn + counterparts (1.74, 2.71, 2.12and 2 .25 -fold enrichment, respectively) (Table 4 1) Immediate early gene ICP0 however, did not follow this trend and was 1.71 -fold enriched on the 17syn + genome relative to KOS. The targets in the LAT region, the LAT promoter and enhancer, had similar amounts of triMe H3K27 deposition between both
65 strains suggesting that heterochromatin regulation of the LAT region is not responsible for the reactivation phenotype. Overall, HSV 1 strain KOS exhibited an increase in heterochromatin marks relative to strain 1 7 syn +. Th e LAT Enhancer in KOS is Most Enriched in dimethyl H3K4 as Opposed to the LAT Promoter in 17 syn + To obtain a full understanding of the epigenetic PTM status of these two strains, we also examined a marker o f euchromatin. DRG from mice l atently -i nfected with either KOS or 17syn + were analyzed for the presence of dimethyl H3K4 on viral genes using ChIP. These ChIPs were validated by comparison of the B/(U+B) relative quantity of the positive control XIST to the negative control mouse centromere (m cent) and fold enrichments of XIST/m. cent were 4.36 and 2.94 -fold for 17syn + and KOS, respectively (Figure 4 2A, C) In the KOS ChIP s the LAT enhancer wa s significantly dimethylated relative to the lytic genes examined as well as to the LAT promoter (Figure 4 2B). These data were consistent with a previous study that found acetyl H3 K9, K14 enriched in the 5 exon of KOS (47) Though the LAT promoter appears as enriched in this mark as the lytic genes, it is actually more enriched than ICP0 and ICP27, similar to the previous findings, and was slightly enriched over tk (1.99, 2.6 and 1.51fold, respectively). No difference was observed between the LAT promoter and lytic genes ICP4 and gC. In contrast, this pattern of euchromati c marks was not observed for 17syn +. In this strain, t he LAT promoter was greatly dimethylated relative to lytic genes ICP27, ICP4, tk and gC (Figure 4 2D) Furthermore in 17syn + the LAT promoter was greatly enriched relative to the 5 exon the convers e of result s from KOS. Additionally, immediate early gene ICP0 was also enriched in dimethyl H3K4, suggesting that it is poised for or currently undergoing, transcription. In comparing the LAT regions of these two strains, the LAT promoter was 2.25-
66 fold more enriched in 17syn + while the LAT enhancer was 26.37-fo ld more enriched in strain KOS (Table 4 2). Based on the comparisons between lytic genes for this euchromatic mark, the 17syn + genome seems to be more repressed than KOS since it has consiste ntly lower levels of diMe H3K4 on the lytic genes than KOS with the exception of ICP0 (Table 42). These findings seem ed contradictory to previous transcriptional analyses that suggested the lytic genes in 17syn + are de -repressed as opposed to lytic genes in the KOS strain (31) In terms of euchromatin, the data presented here suggest that o verall KOS has the more relaxed chromatin structure on its lytic genes during latency Altered Transcript Abundance of the LAT and the TAL between strain s KOS and 17 syn + Because of the observed differences in the euchromatic marks in the LAT region betw een KOS and 17syn +, we examined the accumulation of LAT transcripts in latently -infected mice with either strain. Due to the recent discovery of an antisense transcript to the LAT, TAL (Giordani et al in prep) we also investigated the transcript abundance fr om the TAL as well as several other lytic genes As in all previous studies, LAT was the most abundant transcript relative to the lytic genes (Figure 4 3) In 17 syn +, the LAT transcript was more abundant by 71.63fold when compared to TAL. Yet the TAL transcript was still present at higher levels than the lytic genes. The TAL transcript was detected at greater levels than ICP27, ICP4, tk and gC (2.32, 9.12 30.96and 22.77-fold, respectively). These results d emonstrated that the t ranscript levels of both the LAT and the TAL correlate with the observed enrichment of euchromatic marks of 17syn + data in Figure 4 2B ; the LAT promoter has the highest levels dimethyl H3K3 deposition and the highest amount of transcr ipt while the enhancer, the putative location of the TAL promoter, is not as dimethylated and has lower transcript levels.
67 Upon comparing these transcript levels in 17syn + to those of KOS, a different profile of the LAT region was observed. Though in KO S the LAT was still abundant on a per genome basis, the LAT wa s more than 100-fold reduced relative to 17syn + The amount of TAL between the two strains remained the same (1.0 fold). This difference in transcript levels may be explained by an altered chromatin status in the LAT region in that the LAT enhancer/TAL promoter is highly dimethylated relative to the LAT promoter (Figure 4 2) Both the LAT and TAL transcripts are more abundant than all the lytic transcripts investigated here with the greates t enrichment being between TAL/LAT and gC with an 18 -fold enrichment, followed by 5 fold increase over ICP27, a 15 -fold increase over tk, a nd a 4 -fold increase over ICP4 Therefore, it seems that the main difference in transcript abundance between these t wo strains lies with the LAT region. The only lytic transcript that displayed increased abundance between 17 syn + and KOS was the IE gene ICP27. This does not correlate to the euchromatin data as ICP27 was under -enriched in dimethylation relative to ICP4, tk and g C, yet ICP27 had a similar or increased transcript abundance relative to ICP4, tk and gC (0.88-, 3.02 and 3.75-fold, respectively). Discussion A number of fundamental biological differences exist between HSV 1 strains 17syn + and KOS KOS is a low phenotypic reactivator while 17 syn + efficiently reactivates following stress induction in both the mouse and rabbit models Studies by Sawtell and Thompson suggest that this difference may be due in part to the genome copy number per cell. Cells inf ected with KOS have a lower mean viral genome copy number than strains 17 syn + and McKrae (83) Though the LAT promoter has been analyzed extensively for strain -specific differences in cis -elements no evidence exists that reactivation differences are due to the LAT promoter sequence. Therefore, our study sought to identify if other mechan isms of gene regulation play a role in differentiating
68 between virus strains as well as extend our knowledge of s train -s pecific latent -phase transcript patterns. Gene regu lation is a complex multifactor al process that can be affected on many l evels An area of active study involves determining how epigenetics, specifically how histone PTMs, may play a role in gene expression. Upon surveying the laten t genome for the heterochromatic marker triMe H3K27, we found an altered dep osition of this marker in KOS relative to 17 syn +. Several lytic genes had an increased amount of enrichment of this repressive marker relativ e to 17syn +. This suggested that those genes were in a more repressive state that may be inhibiting transcription. This repressed state of transcription may in turn result in fewer molecular reactivations. Furthermore, this same pattern of increased het erochromatin deposition was noted (50) Since this mutant also exhibits a reduced re activation phenotype, similar to KOS, perhaps these defects in reactivation are due to the increased amount of heterochromatin. Heterochromatin along with other unidentified factors could play a r ole in determining how latent the viral genome really i s. Recently, we and others demonstrated that facultative PTMs were present on the HSV 1 genome during latency (14, 50) However, our data differe d when we examined the effect of LAT on facultative heterochromatin deposition. We observed an increase of triMe H3K27 deposition in the absence of LAT transcription while the other group saw a decrease Based on the data presented here, we strongly sugg est that there may be strain -dependent difference s for repress ion through heterochromatin. This would explain the altered latent triMe H3K27 profiles between the two viruses and potentially explain the lack of reactivation observed in KOS. The se altered profiles are likely the result of different regulation of the Polycomb binding proteins responsible for methylating H3K27. It is unknown what factor(s) may be contributing
69 to this different function between strains but recruitment of PcGs would probably depend o n cis elements or possibly the LAT noncoding RNA Though there is a high sequence similarity between the strains, there are some regions that are more highly conserved than others which may play a role in epigenetic regulation. If these contradi cting data are in fact due to a strain specific function, it would provide a model system for studying how gene regulation through PcGs directly affects the biology of infection and latency. Further support of altered epigenetic regulation between stra i ns came from data investigating euchr omatin differences between KOS and 17syn +. Previously, it had been observed that the LAT region was enriched in the euchromatic marker acetyl H3K9, K14 specifically on the LAT enhancer in the KOS background. Though the LAT promoter was also acetylated relative to lytic genes, it was not as enriched in this mark as the LAT enhancer. The only other euchromatic data available on strain 17 syn + i s from the rabbit ocular model and only showed the LAT enhancer relative to lytic genes (31) But as with the KOS mouse data, it was found that the enhancer was also enriched relative to the ICP0 and ICP27 promoters (31) While the se studies agreed that euchromatin was present, we were initially unsure if the discrepancies in LATs influence on transcription was due to the strains investigated or the animal system studied. G iven that the transcr iptional status would affect chromati n status we decided to take a closer look at the similarities of one euchromatic marker, diMe H3K4, in the mouse footpad model. In the current study, we saw dramatic differences in di methylation patterns between the two strains. In con cordance with Kub at et al. (47) we observed that the LAT enhancer had the greatest amount of enrichment of diMe H3K4 while the LAT promoter wa s not as methylated yet wa s more so than the lytic genes examined in KOS However, in the 17syn + strain, it was the
70 LAT promoter that was greater than the enhancer signifying that perhaps there are different transcriptional profiles of the LAT region between the two virus strains. The differences in enrichment pattern s are further supported by a 2. 25-fold enrichment in the 17syn + LAT promoter relative to KOS while the LAT enhancer if 26.37-fold enriched in KOS relative to 17syn +. In fact, the areas of euchromatin enrichment correlate to the promoters of LAT and TAL. The LAT promoter primers obviously detect the LAT promoter while the LAT enhancer primers detect the promoter region of TAL. This may also indicate which promoters are most active in the se virus strains. In 17 syn +, the euchromatin data would suggest that the LAT promot er is more acti ve while in KOS the TAL promoter seems more transcriptionally active. Additionally, the ICP0 promoter was found to have an enriched amount of diMe H3K4 relative to the LAT enhancer and other lytic genes. It is increased relative to ICP0 in KOS (2.57fold). As this PTM is deposited as the transcription machinery travels across the gene, it would seem to indicate that this area is actively being transcribed during latency. ICP0 and LAT expression are hypothesized to be incompatible meaning that LAT would be transcribed in the latent phase while ICP0 would be in the lytic phase. Euchromatin on the ICP0 locus may suggest that some genomes are currently ex periencing active transcription of ICP0, a sign that could signify molecular reactivation. On the other hand, it may represent a population of genomes that are poised and ready to reactivate. The idea that populations of genomes exist, rather than all the viral genomes in one cell, is not unprecedented as it is already known that 1/3 of latently infecte d neurons express the LAT (62) Therefore, some genomes have euchromatin on the ICP0 promoter in anticipation of a stimulus that would allow transcription to proceed. Often promoters controlled by Polycomb/Trithorax groups contain both triMe H3K27 and diMe H3K4 and may act as a switch between whether the gene is activated or repressed. Since ICP0
71 in 17syn + also had more enrichment of triMe H3K27 than in KOS, it might signif y that the ICP0 promoter plays a role in Polycomb silencing. Kwiatkowski et al. (Chapter 3) showed that the PRC1 protein Bmi1 was also bound to the ICP0 gene, even though it did not have the h ighest amount of accumulation observed (50) The presence of the facultative marks may prevent transcription machinery from advancing along the gene resulting in elongati on of transcripts. Perhaps it is necessary for ICP0 in 17 syn + to contain a higher facultative heterochromatin deposition amount than KOS because ICP0 is in a leakier state in 17syn + and the opposing marks serve as a mechanism of gene control and remain in a repressed state until something stimulates the PcGs to clear out. A third option is that the euchromatin marks are not signifying transcription of ICP0 but rather the class of L/S junction-spanning transcripts (L/STs) that are sense to LAT These tran scripts can vary in size anywhere from 2.3 to >9.0 kb and are transcribed with late kinetics possibly even constituting a new class of transcripts, likely originating from the bidirectional ICP0 promoter and extending over genes ICP34.5 and ICP4 (6, 102) These transcripts can only be detected abundantly in the absence of ICP4 as there may be a RNA -mediated silencing mechanism or by the ICP4 protein binding and repressing expression of transcripts (51, 102) While preliminary results had not detected the presence of these transcripts during latency, it is possible that more sensitive PCR techniques may detect their low expression levels. Overall, euchromatin on the ICP0 gene represents one of t he layers of gene control and may signify a new layer of gene regulation that was not previously identified based off of tr anscript abundance during latency As chromatin profiles between KOS and 17syn + proved to be different, especially in the LAT region, we decided to take a closer look at laten t -phase transcripts between these strains. RNA and DNA isolated from latently in fected mice were analyzed as RNA molecules per
72 genome for comparison. We found that the major differen ce between strains is the LAT transcript abundance There was over a 100-f old reduction of LAT in KOS relative to 17 syn + while synthesis o f TAL did not change significantly (1.63-fold) between strains I n KOS the quantities of RNA molecules of TAL were similar to LAT. This suggests a difference in gene expression regulation between the two strains. This is further supported by the euchromatic data described above which showed the highest enrichment of diMe H3K4 in 17syn + to be the LAT promoter. The shift in euchroma tin to the LAT enhancer in KOS does not prevent LAT expression but rather lowers the abundance of transcripts from that gene. It would seem that s ome epigenetic mechanism or perhaps a cis element is responsible for the shift in the peak of diMe H3K4 from the promoter to the enhancer from 17syn + to KOS This may ultimately be due to populations of cells as we do not know if LAT and TAL e xpression occur in the same cell. If the virus strains are establishing latency in different types of neurons or satellite cells, it may generate a different transcript profile. Though HSV 1 viral strains 17syn + and KOS are highly related in sequence, t hey possess different capabilities in terms of acute infection, latency and reactivation. Further differences in the nature of their latent stages have been identified in this study in terms of chromatin profile and latent -phase transcript abundance. The factors that contribute to the altered epigenetic profile are unknown but it is becoming increasingly clear the role the chromatin plays in gene regulation of the latent HSV 1 genome. We also provide further evidence of the newly identified TAL that show s altered expression levels depending on the virus strain which is reflected in the euchromatin profile of the LAT region After having identified a molecular difference between 17syn + and KOS the mechanism of gene regulation between viral strains become s a much more
73 addressable question, one that is most likely modulated by several factors. Therefore, data obtained about the molecular control of repression may not be applicable to all HSV 1 strains.
74 Figure 4 1. triM e H3K27 is enriched on the KOS latent genome. A) Validation of the ChIPs using anti triMe H3K27 was performed by analyzing bound and unbound fractions of ChIP by real -time PCR with positive control PCR primers/probe to myoD1 compared to a negative control PCR primers/probe, APRT. Data are graphed as the B/(U+B) ratio. Average fold enrichment between myoD1 and APRT is denoted by the bracket. B) ChIPs using anti triMe H3K27 were subjected to real -time PCR using primers specific for the HSV 1 target genes i ndicated, and the results were graphed as B/(U+B) ratio normalized to APRT B/(U+B). Mean values are displayed for each gene, data from six independent ChIPs are graphed (n=6). L. pro, LAT promoter, L. enh, LAT enhancer.
75 Figure 4 2. diMe H3K4 enrichm ent varies depending on the HSV 1 virus str ain. A) Validation of the KOS ChIPs using anti diMe H3K4 was performed by analyzing bound and unbound fractions of ChIP by real -time PCR with positive control PCR primers/probe to XIST compared to a negative cont rol PCR primers/probe, mouse centromere (m. cent). Data are graphed as the B/(U+B) ratio. Average fold enrichment between XIST and m. cent is denoted by the bracket. B) ChIPs on K OS latently infected mouse DRG using anti diMe H3K4 were subjected to real time PCR using primers specific for the HSV 1 target genes indicated, and the results were graphed as B/(U+B) ratio normalized to XIST B/(U+B). Mean values are displayed for each gene, data from five independent ChIPs are graphed (n=5). C) Validation of the 17syn + ChIPs using anti -diMe H3K4 were analyzed and graphed similar to panel A. D) ChIPs on 17 syn + latently infected mouse DRG using anti -diMe H3K4 were analyzed and graphed similar to panel B (n=4). L. pro, LAT promoter, L. enh, LAT enhancer.
76 Figure 4 3. Viral transcript abundance in 17 syn + and KOS latently infected DRG. RNA was isolated from latently infected DRG, reverse transcribed into cDNA, and analyzed by real time PCR as described in Material and Methods. Data are shown as RNA molecul es per genome. A) Results of the PCR analysis for the LAT (5 exon), TAL and lytic genes. B) Results of PCR analysis minus the data points from LAT 17 syn +. (note different scale). Table 4 1. Mean Normalized Values of triMe H3K27 Enrichment for 17 syn + and KOS Primer Target B/(U+B) 17 syn + a B/(U+B) KOS b Fold Difference c KOS/17 syn + LAT promoter 9.84 7.53 0.77 LAT enhancer 9.94 10.56 1.06 ICP0 12.80 7.47 0.58 ICP27 8.96 15.59 1.74 ICP4 4.70 12.74 2.71 tk 9.66 19.23 2.12 gC 11.51 25.86 2.25 a Mean values of B/(U+B) values normalized to APRT B/(U+B) are listed from ChIPs using anti triMe H3K27 in 17syn + latently infected mouse DRG. (50) b Mean values of B/(U+B) values normalized to APRT B/(U+B) are listed from ChIPs using anti triMe H3K27 in KOS laten tly infected mouse DRG (Figure 4 1 ). c Fold enrichment of KOS B/(U+B) values relative to 17 syn + B/(U+B) values.
77 Table 4 2 Mean Normalized Values of diMe H3K4 Enrichment for 17 syn + and KOS Primer Target B/(U+B) 17 syn + a B/(U+B) KOS b Fold Difference c KOS/17 syn + LAT promoter 4.83 2.15 0.44 LAT enhancer 0.75 19.66 26.37 ICP0 2.90 1.13 0.39 ICP27 0.34 0.73 2.16 ICP4 0.17 2.48 14.43 tk 0.35 1.29 3.69 gC 0.41 1.88 4.61 a Mean values of B/(U+B) values normalized to APRT B/(U+B) are listed from ChIPs using anti -diMe H3K4 in 17syn + laten tly infected mouse DRG (Figure 4 2 ). b Mean values of B/(U+B) values normalized to APRT B/(U+B) are listed from ChI Ps using anti -diMe H3K4 in KOS laten tly infected mouse DRG (Figure 4 2 ). c Fold enrichment of KOS B/(U+B) values relative to 17 syn + B/(U+B) values. d P values between the values in columns 2 and 3 were determined by ANOVA (see Materials and Methods Table 4 3 Mean RNA molecules per genome for 17syn + and KOS Primer Target 17syn + a KOSb Fold Difference c 17 syn +/KOS LAT 1.7338 0.0153 113.57 TAL 0.0242 0.0148 1.63 ICP4 0.0027 0.0034 0.77 ICP27 0.0104 0.0030 3.42 tk 0.0008 0.0010 0.78 gC 0.0011 0.0008 1.31 a Mean values of RNA molecules per genome of the indicated transcripts in 17 syn + laten tly infected mouse DRG (Figure 4 3 ). b Mean values of RNA molecules per genome of the indicated transcripts in KOS latentl y infected mouse DRG (Figure 4 3 ). c Fold enrichment of 17syn + RNA molecules relative to KOS RNA molecules d P values between the values in columns 2 and 3 were determined by ANOVA (see Materials and Methods
78 CHAPTER 5 INITIAL DOSE OF INFECTION AFFECTS THE LAT AND THE TAL TRANSCRIPT LEVELS BUT NOT CHROMATIN PROFILES DURING LATENCY Introduction HSV1 establishes lat ency in sensory neurons where it assumes a silent transcriptional p rofile with the exception of abundant transcription from the LAT and TAL However, a stress stimulus can promote the virus to exit the latent state and enter the lytic cycle. The re is evidence in the literature that, i n some HSV 1 strains, the ability to recover infectious virus from a reactivation stimulus is dependent on the amount of viral DNA in the ganglia; increased DNA levels results in increased infectious virus recovered (40, 52, 84) For instance, i f the initi al dose of infection of an ICP34.5 mutant which spontaneously reactivates at a 10 -fold lower rate than wild type at 2x105 pfu/eye, is increased to 1x108 pfu/eye in the rabbit ocular model, it raises the rate of spontaneous reac tivation to levels comparable with wild type (72) It is important to note that this mutant was made in the HSV 1 strain McKrae, which has a hi gh spontaneous reactivation phenotype in vivo (71) This suggests that one can alter the reactivation phenotype by increasing the amount of input virus. The increase in infection dose also causes an increase in the number of neurons that harbor latent HSV 1 genomes (82) It was previously shown that HSV 1 strains with high reactivation phenotypes contain a high copy number of genomes per neuron than strai ns with a low reactivation phenotype (83) This suggested that the ability of HSV 1 to reactivate may be partly dependent on the number of genomes per cell. Since the re is evidence that the dose of virus during a primary infection clearly has a role on the establishment and reactivation of HSV1 we wanted to investigate whether there is a dose -dependent effect on regulation of viral gene suppression It is known that a n increase in the dose of virus can affect LA T expression (72) High doses of infection result in a greater number
79 of LAT positive cells as determined by in situ hybridiz ation (72) Furthermore, as the amount of detectable LAT increases, so does the frequency of reactivation (32, 60, 91) Therefore, it would seem that the frequency of reactivation is dependent on increased LAT transcription and possibly other viral genes. For this reason, we investigated five different doses of KOS/1 to see if there were altered latent transcript levels or an altered enrichment in the euchromatin deposited on the latent genome This strain of virus is avirulent allowing us to increase the dose with out compromising the mouse. We d etermined that the LAT and the TAL transcript levels were most changed by the initial dose of infection but that the euchromatin deposition pattern on the HSV 1 genome remained similar over all doses. Results The LAT is th e most abundant transcript relative to lytic transcripts and peaks in abundance between the 5x103 and 5x104 dose s To determine if the initial dose of infection can alter the transcription profile of the latent genome, the transcript abundance of KOS/1 la tently -infected mice was compared between five different inoculums. The KOS/1 mutant contains the lacZ reporter gene under control of the LAT promoter in place of the gC capsid gene, which can be used to monitor latent gene expression (26, 84) These initial doses included 500 pfu, 5x103 pfu, 5x104 pfu, 5x105 pfu and 5x106 pfu. After 28 d.p.i. the DRG were harvested and RNA and DNA were processed and analyzed by real -time PCR An n of 8 was obtained for all doses except 500 pfu, which had an n of 4, due to viral DNA levels being below the limit of detection. As shown in Figure 5 1 and 5 2 the transcript levels of the LAT 5 exon appear ed higher with increase d dose and peak at the 5x104 pfu dose before returning to levels s imilar to the 500 pfu dose by the 5x105 and 5x106 pfu dose s While the transcript levels between the doses appeared to be different, none were significantly different from to the dose lower or higher than it (p>0.08). The only noted
80 statistical differences in LAT levels were between the 5x103 pfu dose and the high pfu doses, 5x105 and 5x106 (p=0.02 for both). With these two high doses accumulating lower levels of the LAT than the 5x103 pfu dose, it suggest s that once a threshold level of incoming virus is achieved the LAT transcript begins to decrease in transcript abundance. Though the LAT levels decrease with the increased inoculums after the 5x104 pfu dose they still remained to be more abundant than the lytic tra nscript tk (p=0.009 and p=0.007 for 5x105 and 5x106 pfu doses, respectively) and ICP27 (p=0.04 for 5x106 pfu) (Figure 5 1 ). Abundance of the TAL transcript is dependent on the initial dose of virus The same samples mentioned above w ere also analyzed for the presence of the transcript antisense to LAT, the TAL. This has previously been shown to be present in similar levels as the LAT in the KOS strain at a dose of 1000 pfu (Chapter 4) We observed from the present study that the ratio of the LAT to TAL is dependent on the dose of infection. There were similar amounts of the LAT and the TAL in the 500, 5x105 and 5x106 pfu infections while there was more LAT than TAL in the 5x103 (p=0.01) and 5x104 (p=0.0 5) pfu doses (Figure 5 1 ). When th e TAL abundance for all doses was compiled in the same graph, the pattern of enrichment going from low dose to high dose resembles the shape of a n upside -down bell curve (Figure 53). This is particularly interesting as the opposite pattern was ob served when LAT is graphed the same way (Figure 5 2) Between the 500 and 5x103 pfu infections there is a 4 -fold decrease in TAL transcripts (p=4.5E 5) The TAL levels present in the 5x103 pfu infection are also lower than the levels of the 5x105 and 5x 106 doses (p=0.025 and p=0.007, respectively). This pattern of TAL abundance was the reverse of what was observed for the LAT between the 5x103 dose and the high pfu doses of 5x105 and 5x106 suggesting that the LAT may be suppressing TAL transcription at certain doses.
81 At both the lowest (500) and highest (5x106) dose, the TAL wa s significantly more abundant than the lytic transcript s ICP27 and tk (p<0 .003 and p<0.0025, respectively). In addition to the decrease in TAL relative to LAT at the 5x103 pfu dose, the TAL was now at higher levels than lytic transcript tk (p=0.015) but not ICP27. As the dose was increased to 5x104 pfu the TAL wa s present in similar amounts to lytic genes ICP27 and tk (Figure 5 1 ). TAL transcript level s began to increase in the 5x105 p fu dose and were now once again more enriched than tk (p=0.005) but not ICP27. This upside down bell curve pattern indicates that th e abundance of the transcript was influenced by the initial dose of infe ction as indicated in Figure 5 3 Ly tic viral transcript lev els are not affected by the inoculum dose Immediate early gene ICP27 and early gene tk were examined at the five doses described above The ICP27 transcript abundance did not change as the initial dose of infection was increased from 500 pfu to 5x106 pfu (p>0.24) (Figure 5 4 ). Detection of the tk transcript was more difficult as the levels for the samples (n=8) were at or near the limit of detection indicating that this transcript is present at very low levels (Figure 5 5) In Figure 5 1, 54, and 5 5 th e filled grey lines represent the limit of detection for a particular dose and primer With that said, the only statistical difference observed for tk between the doses was between the 5x103 pfu and 5x105 pfu dose s (p=0.04). G iven that the majority of points for 5x103 pfu and 5x105 pfu were below the limit of detection, it is possible that these data sets are not different from each other, just too low to accurately distinguish the 5x103 from the 5x105 pfu dose. Therefore, the ICP27 and tk transcript levels do not appear to be influenced by the initial dose of inoculum.
82 The latent viral transcription profile is not altered due to the deletion of gC in KOS/1 T he previous studies were all performed using a gC null HSV mutant KOS/1 Therefore, we wanted to see if the pattern of transcript abundance was not simply an effect of study ing a mutant virus. M ice were infected with wild type KOS at doses of 5x 103 and 5x 105 pfu and DRG were harvested at 28 d ays p.i. for RNA and DNA analysis. Abundance of the viral transcripts LAT, TAL, ICP27 and tk were measured by real time PCR analysis and the data was analyze d as RNA molecules per genome. The pattern of transcript abundance was similar to that of KOS/1 in th at the LAT region transcripts were detectable at higher levels than the lytic transcripts for both doses indicating that the gC mutation does not affect the latent transcriptional profile of HSV1 (compare Figure 5 6 to Figure 5 1 ). We did note an increas e in the amount of LAT and ICP27 transcripts between KOS/1 and KOS however at the 5x103 pfu dose The abundance of the LAT in KOS/1 was on average 4.5 -fold increased relative to KOS, while ICP27 was 13.6fold increased in KOS/1 relative to KOS The d ifference in LAT transcripts between the two viruses was significant (p=0.029) while the difference for ICP27 was not suggesting that the only transcript affected by the gC mutation was the LAT at the 5x103 pfu dose For the 5x105 pfu dose, there was more LAT (2 -fold), ICP27 (2.4 -fold) and tk (4.28 -fold) in KOS/1 relative to KOS but the only significant difference was for tk (p=0.009). As with most of the tk data points, the transcript levels for the samples for both KOS/1 and KOS at the 5x105 pfu dose were at or around the limit of detection, meaning that there is most likely similar amounts of the transcript between the two viruses. No difference in transcript abundance was observed between the two different doses for the LAT, TAL and ICP27 for KOS However, there was an increased amount of the viral transcript tk in the 5 x 103 pfu dose relative to 5x 105 (p=0.053) Since half of the data points from these sets were below the limit of detection they cannot be reliably used to m easure a difference. Therefore, it would seem that t ranscript abundance at these two doses are
83 likely equivalent Overall, the gC mutation in KOS/1 did not affect the transcript abundance during latency with the exception of the LAT at the 5x103 pfu dose Increas ing the amount of inoculum above 5x103 pfu does not increase the amount of genomes in the DRG One of the concerns of increasing the infection inoculum was that there might be more seeding of genomes in the DRG. Therefore, we compared jus t the DNA molecules from all five doses and found that the only difference in the amount of DNA establishment was from the lowest dose, 500 pfu. On avera ge, the 500 pfu dose had 12,548 DNA molecules per DRG while doses 5x103 pfu through 5x106 pfu had on averag e 62,545, 45,534, 50,225, and 47, 534, re spectively as shown in F igure 5 7 The genomes in the 500 pfu dose were significantly lower than that of doses 5x103, 5x104 and 5x106 pfu (p=0.007, p=0.045, p=0.036, respectively). In fact, half of the original samples from the 500 pfu dose had to be excluded from the data set for the lack of detectable HSV 1 indicating that the infections at this low dose were not efficient resulting in an n of 4 usable samples instead of 8. However, once the threshold of 5x103 pfu had been reached, there was not an increased presence of genomes in the DRG. While the total number of genomes is similar between the higher doses, it is not certain if the number of genomes per cell wa s equivalent. For example, it is unclear whet her a small number of cells had a higher number of genomes or if a large number of cells had a small number of genomes Fur ther experiments would need to be done to investigate this phenotype. Dose of infection does not affect the euchromatin profile o f the latent genome Transcription is controlled on many levels and is partly influenced by the presence of PTMs on chromatin. Previous studies have shown that the LAT region is euchromatic relative to the lytic genes during latency (31, 47, 48, 69) To examine if t his aspect of transcriptional control might be influenced by the initial dose of infection, DRG of l atently -infected mice of the
84 five previously mentioned doses of KOS/1 were harvested and analyzed by ChIP for the presence of acetyl H3 K9, K14. Results we re analyzed as the bound/input ratio (B/U+B) normali zed to the cellular control Xist B/(U+B). The graphs for the five doses are shown in Figure 5 8 A similar pattern of euchromatin deposition as previously shown in C hapter 4 of this dissertation as wel l as by Kubat et al (47, 48) was present for all of the KOS/1 doses The LAT 5 exon/enhancer wa s more enriched in acetyl H3 K9, K14 than the LAT promoter as well as the lytic genes ICP0 and ICP4. Relative to the LAT promoter, the LAT enhancer wa s more enriched by 2.36, 7.27-, 10.39, 18.94, and 21.33-fold in the 500, 5x103, 5x104, 5x105 and 5x106 pfu doses, respectively. The only dose that was significantly more enriched at the LAT enhancer was 5x104 pfu (p=0.04) while the others were approaching significance (p=0.06 0.08) a nd might be significant if the number of experimental samples were increased. Overall, the fold enrichment of the euchromatin on the LAT enhancer relative to the LAT promoter increases with the dose but is for the most part not significantly different. In fact, it is the LAT enhancer values that vary between doses 5x103 through 5x106 pfu Values for the LAT promoter, ICP0 and ICP4 remain the same between the four highest doses. This indicates that the LAT enhancer is influenced by the dose of infectio n. The lowest dose, 500 pfu, is set apart from the higher doses by the relative quantity values of the LAT promoter, ICP0 and ICP4. All of the targets examined from the 500 pfu latently infected DRG appeared to be in a greater euchromatic state than the higher doses. Since all the higher doses are very simila r, the following comparison was done between 500 pfu and 5x103 pfu, which was represent ative of the h igher doses. The 500 pfu dose wa s 3.48 -fold more enriched in acetyl H3 K9, K14 on the LAT enhancer, 10.7 -fold more on the LAT promoter, 7.69 fold more on ICP0 and 6.76 -fold more on ICP4. Of these targets, o nly the LAT enhancer was
85 significantly more enriched in the 500 pfu dose relative to 5x103 pfu (p=0.017). Even though the LAT is more enriched in euchromatin, it actually has lower levels of transcript abundance than the 5x103 pfu dose (Figure 5 2 ). However the TAL transcript, represented in the ChIP assays by the LAT promoter primers, is increased in transcript abundance in 50 0 pfu relative to the 5x103 dose which correlates to the acetyl H3 K9, K14 data. Without the transcript levels for ICP0 and ICP4, it cannot be determined if the 500 pfu abundance matches the pattern of euchromatin since the TAL transcript levels correlat e to euchromatin while the LAT transcript levels do not. Perhaps a better marker of euchromatin could be utilized to investigate this relationship of transcript abundance and euchromatin. Discussion It is becoming clear that many factors can affect a n HSV 1 infection. The number of genomes establishing latency in the neurons is most likely a major determinant of the biolo gy of infection. I ncreasing the initial virus dose can not only increase the viral DNA in the neurons (by genome number per cell as well as the number of latently -inf ected neurons per ganglion ) but it also can affect LAT expression and reactivation (32, 60, 72, 82, 83, 91) Though ther e is no direct evidence for the following hypothesis, it seems likely that there is a mechanism used by the virus to detect the extra genomes that may cause the above noted differences. This mechanism may be as simple as a threshold value meaning that with the more genomes that enter the host a higher number of sensory neurons become seeded with viral genomes resulting in an efficient establishment of latency. At that point, the increased number of genomes ma y be more optimal for responding to stress stimuli to result in reactivation of the virus. Even induced reactivation in animal models does not seem to cause every latentlyinfected neuron to respond (85) Therefore, by increasing the number of genomes per sensory ganglion, the virus can increase its chances of reactivation and therefore survival b y transmission to a new host.
86 If the above proposed model is accurate, we would expect to see a noticeable change in the transcript abundance or transcriptional permissiveness through euchromatin as we increased the inoculum dose from 500 pfu to 5x106 p fu. While the pattern of acetylation of H3 K9, K14 remained the same over all doses, we noted that t he LAT 5 enhancer increased in acetylation relative to the LAT promoter as the dose increased suggesting that the LAT region is more transcriptionally pe rmissive This could thereby allow for increased LAT transcription (Figure 5 8 ). The s e data seems to be in concordance with previous findings that described a greater number of LAT -expressing neurons when the dose of infection is increased (72) Ho wever the interpretation of the result is made difficult by the presence of the TAL promoter. Though the promoter has not yet been fine mapped for TAL, it was hypothesized in Chapter 4 that th e euchromatin on the LAT 5 enhancer in KOS may be representative of the permissiveness of the TAL promoter. Therefore, by this reasoning, it wou ld seem that the TAL promoter be comes more permissive at higher doses. Determining which gene the TAL or the LAT, is becoming more transcriptionally permissive with dose can be answered by our transcript abundance analysis during latency Our latent transcript analysis of 500 to 5x 106 PFU inoculum doses examined the RNA molecules per genome of the LAT, the TAL, ICP27 and tk transcripts No differences in the two lytic transcripts were observed, but we did not e dose dependent changes in the LAT and the TAL transcripts In fact, the abundance of these two transcripts were inversely proportional to the 5x103 and 5x104 pfu doses One attractive hypothesis for this result is that the LAT is downregulating the TAL transcript resulting in its decrease. However, it does not explain why both transcripts are present in similar amounts at the 5x105 and 5x106 pfu dose s Perhaps at the highest inoculum doses, the numbers of incoming genomes have an inhibitory effect on
87 tran script abundance. This may be accomplished through a mechanism by which the transcription machinery is saturated out due to the high number of viral genomes or perhaps to a dosage compensation mechanism, similar to that of the silent X chromosome in female mamma ls. In X chromosome inactivation, one X chromosome must be silenced to compensate for the increased number of that chromosome relative to in male cells. The number of latently -infected neurons can be increased by rai sing the infecting inoculum (82) However, when we analyzed the DRG for the amount of viral genomes, we observed no significant difference for the four higher doses. The only dose that was different was 500 pfu, which seemed to be an inefficient dosage level. The amount of genomes recovered from mice infected with that dose was variable and much lower than with the higher doses. Since our viral DNA levels did not change with dosage amount, we believe that instead of changing the total number of genomes present in the ganglia, we might have changed their distribution pattern. This means that we might have more latently -infected neurons with lower genome numbers or just an increased amount of genomes per cell. The only way to determine what is happening at these doses wou ld be to perform laser capture microscopy followed by PCR. Though previous studies have documented changes in genome distribution patterns, establishment and reactivation, we could not detect any further characteristic changes in the transcript abundanc e as a function of dose. We were able to show differences in the LAT and the TAL abundance at some doses, but there was no t a positive correlation effect as with previous studies (72) It is possible that other mechanisms of transcriptional control are present to prevent aberrant gene expression if the number of viral genomes per neuron gets too high. We did, however notice that the transcriptional p ermissiveness of the LAT 5 enhancer relative to
88 the LAT promoter increased with the dose of infection which may be reflectiv e of an increased number of LAT -expressing neurons as previously documented (72) Therefore, these described changes in transcriptional permissiveness, as well as characteristics of infection impacted by the infection in oculum, indic ate that viral load can indeed a ffect many aspects of an HSV 1 infection. Overall, the amount of infection inoculum is a factor that researchers should consider when setting up experiments as their results may be unique to a particular dose of infection rather than reflective of how all HSV 1 infections occur.
89 Figure 5 1. Viral transcript abundance in KOS/1 latently infected mouse DRG. RNA was isolated from latently infected DRG, reverse transcribed into cDNA, and analyzed by rea l time PCR as described in Materials and Methods. Data are shown as RNA molecules per genome. A) RNA from a 500 pfu infection. B) RNA from a 5,000 pfu infection. C) RNA from a 50,000 pfu infection. D) RNA from a 500,000 pfu infection. E) RNA from a 5 ,000,000 pfu infection. Note the different scales. The filled grey bars represent the limit of sensitivity for that primer.
90 Figure 5 2 LAT primary transcript abundance in KOS/1 latently infected mouse DRG. RNA was isolated from latently -infected DRG, reverse transcribed into cDNA, and analyzed by real -time PCR as described in Materials and Methods. Data are shown as RNA molecules per genome. The samples were from mouse infections of 500, 5,000, 50,000, 500,000, and 5,000,000 pfu. Figure 5 3. TAL transcript abundance in KOS/1 latently infected mouse DRG. RNA was isolated from latently infected DRG, reverse transcribed into cDNA, and analyzed by real time PCR as described in Materials and Methods. Data are shown as RNA molecules per genome. The samples were from mouse infections of 500, 5,000, 50,000, 500,000, and 5,000,000 pfu.
91 Figure 5 4. ICP27 transcript abundance in KOS/1 latently-infected mouse DRG. RNA was isolated from latently infected DR G, reverse transcribed into cDNA, and analyzed by real time PCR as described in Materials and Methods. Data are shown as RNA molecules per genome. The samples were from mouse infections of 500, 5,000, 50,000, 500,000, and 5,000,000 pfu. The filled grey bars represent the limit of sensitivity for that primer. Figure 5 5 T k transcript abundance in KOS/1 latently -infected mouse DRG. RNA was isolated from latently infected DRG, reverse transcribed into cDNA, and analyzed by real time PCR as described in Materials and Methods. Data are shown as RNA molecules per genome. The samples were from mouse infections of 500, 5,000, 50,000, 5 00,000, and 5,000,000 pfu. The filled grey bars represent the limit of sensitivity for that primer. The filled grey bars represent the limit of sensitivity for that primer.
92 Figure 5 6. Viral transcript abundance in KOS/M latently infected mouse DRG. RNA was isolated from latently infected DRG, reverse transcribed into cDNA, and analyzed by real time PCR as described in Materials and Methods. Data are shown as RNA molecules per genome. A) RNA from a 5,000 pfu infection. B) RNA from a 500,000 pfu inf ection. Note the different scales. Figure 5 7. Genomes per KOS/1 latently infected mouse DRG. DNA was back -extracted from an RNA preparation and analyzed by real -time PCR as described in Materials and Methods. Data are shown as the number of HSV 1 genomes per whole mouse ganglion. The samples were from mouse infections of 500, 5,000, 50,000, 500,000, and 5,000,000 pfu.
93 Figure 5 8. ChIP of acetyl H3 K9, K14 on the KOS/1 genome. Data were graphed as the B/(U+B) ratio normalized to the cellular control XIST B/(U+B). A) 500 pfu infection. B) 5,000 pfu infection. C) 50,000 pfu infection. D) 500,000 pfu infection. E) 5,000,000 pfu infection. Data from four independent ChIPs are graphed (n=4). L pro, LAT promoter; L. enh, LAT enhancer/exon.
94 CHAPTER 6 DISCUSSION HSV1 has co -existed with humans for thousands of years and is well adapted to life -long persistence in its host. This virus has devised a survival plan in which it establishes a latent reservoir of virus within the sensory ganglion that can reactivate to produce more virions which can then transmit HSV1 to other hosts. The ability of the virus to establish laten cy also prevents the host from clearing the infection since no foreign antigens (viral proteins) are produced during that phase of infection Therefore the establishment of a latent infection is vital for HSV 1s long -term survival. The most baffling aspec t of HSV 1 is how it suppresses lytic gene transcription to promote a latent infection. B y day 28 p.i. in animal models, the virus is essentially transcriptionally silent except for the LAT and the TAL transcripts The mechanism for gene repression is poorly understood but is likely a multifactoral process This dissertation was aim ed at illuminating some of the key factors that could explain the suppression of lytic genes The results presented here illustrate the complexity of this process suggesting that the establishment of latency is propagated by redundant mechanisms to ensure seeding of virus in the sensory ganglion. This is supported by the fact that to date, no HSV 1 muta t ion has been identified that conf ers the ability to establish a latent infection While the virus has evolved many ways to ensure establishment in the host, it does not guarantee that the latent virus is capable of reactivation. Heteroc hromatin on the Latent Genome Upon entry into the nucleus, the viral genome is immediately chromatinized and remains as an extrachromosomal episome (81) After nucleosomes are assembled on the viral DNA, repression occurs not through the presence of DNA methylation but by modifications to the histon e tails (PTMs) (22, 48) Initially only the lack of euchromatin on lytic genes provided us
95 with evidence of repress ion through heterochromatin (31, 47, 48, 68) We now know t hat repressive PTMs are present over the entire genome including the LAT region, once considered to be only euchro matic (14, 50) The likely reason heterochromatin is enriched on the LAT region is that 2/3 of latently -infected neurons do not express LAT (62 ). Heterochromatin on repressed lytic genes, and on some of the LAT genes, further enforces that chromatin PTMs can either serve as another layer of transcriptional control or can be used to reflect the transcriptional status of a gene. The mechanism by which the viral genome becomes associated with heterochromatin is partl y influenced by the LAT region (14, 50) In HSV 1 strain 17syn +, our data suggests tha t either the act of transcription of the LAT transcript or cis elements within the LAT promo ter provide control over triMe H3K27 deposition by reducing t he enrichment on the genome (50) Additionally t he deposition of the triMe H3K27 mark is mediated by the cellular Polycomb group (PcG) proteins which bind to the latent HSV 1 genome (87) (Figure 6 1) Cellular genes that are controlled by PcGs interact with Trithorax proteins which promot e euchromatic marks. Usually there is a switch in the gene promoter that decides if the PcGs are going to repress the gene or if the Trithorax proteins are going to express the gene (57) Therefore, control of PcG proteins on the latent HSV 1 genome is most likely a result of their interaction with Trithorax proteins. It seems likely that deleti on of the LAT promoter abr ogates binding of the Trithorax proteins resulting in the misregulated deposition of triMe H3K27 in the HSV (50) The L AT Region and HSV -1 Strains It has been documented that many differences about the nature of infection between KOS and 17 syn + exist. What is most striking is that a KOS infection is avirulent, while 17 syn + causes clinical disease (37) Additionally, strain KOS does not reactivate in rabbits upon stimulation
96 while 17 syn + does so efficiently (37) These differences do not stem from inefficient establishment as both are capable of a latent infection in the sensory gangl ion. The amount of genomes per cell though, does vary with strain 17 syn + having a higher mean of genomes per cell than KOS (83) W e sought to find out if these results were related to differences in either chrom atin deposition during latency or the latent transcriptional profile. We first observed that strain KOS had increased levels of triMe H3K27 relative to the 17syn + genome. Since the increased amount of heterochromatin was similar to results from a nonrea ctivating LAT mutant, 17, which c ould explain w hy KOS does not reactivate. Upon examining the two strains for differences in transcriptional permissiveness, we did not obser ve less euchromatin (diMe H3K4) on the lytic genes in 17syn + relative to KOS. The only dramatic change between the two strains was in the LAT region. The LAT promoter in 17syn + was hyperdimethylated relative to the LAT 5 exon while the converse was true for KOS. We interpreted this to re flect the fact that this region contains the promoters for the LAT as well as the novel transcript, TAL, antisense to the LAT and originates in the LAT prom o ter/5 exon region. If both of these transcripts play a role i n reactivation of the virus, altered permissiveness of their promoters might explain differences in viral phenotypes. The transcripts in the LAT region, the LAT and the TAL, were expressed more abundantly than the lytic genes during latency in both strai ns While there is more of the LAT (71 -fold) than TAL in 17syn +, there are equal amounts of the two transcripts in KOS. What causes the difference is a 100 -fold drop in LAT abundance in KOS relative to 17syn +. The abundance of lytic transcripts does not change between the two strains. Therefore, the transcription of the LAT region, particularly the interplay between the LAT and the TAL transcripts, probably plays
97 a role in the establishment of a genome that is capable of reactivating and may vary on a cell -to cell basis Effect of Dose on the Latent Transcriptional Profile To investigate the effect of increasing inoculum dose on the latent transcriptional profile, we looked at five doses of the KOS/1 virus a nd its effect on euchromatin and transcript abundance. We observed no change in the transcriptional permissivene ss pattern of the latent genome, alt hough the enrichment of acetyl H3 K9, K14 on the LAT 5 exon did increase relative to the LAT promoter as t he dose was increased. This correlate s with results from previous studies which found that the number of LAT -expressing latently infected neurons increased when the dose of infection was increased (72) However, we identified a novel dose -dependent pattern for LAT abundance. At lower doses, the LATs abundance increased with dose while it decreased at the higher doses. The exact opposite phenotype was observed with the TAL transcript. This once again suggests an important interaction between these two transcripts which may be regulation of each other. Due to the antisense orientation and the possible regulation of each ot her, the relationship between the LAT and the TAL may be similar to that of XIST and TSIX These two transcripts are involved in X chromosome inactivation in which XIST is expressed from the silent X and coats the chromosome while recruiting repressive histone modifying proteins to ensure the X remains transcriptionally silent. TSIX is expressed from the active X and is incompat ible with XIST expression from the same chromosome Many similarities ex ist between the LAT in latent HSV 1 and XIST in the X inactivation model the first of which was identified when a BLAST search revealed sequence similarity between the two transcripts (5). Within the LAT transcript, sequence similarity was identified in the enhancer region as well as several other regions encoding the downstream LAT. Both XIST and LAT are expressed from re gions designated to be silent and w hile these RNAs do contain introns
98 and may be differentially spliced or polyadenylated, there has been no evidence that either RNA codes for proteins. Other similarities in the organization of these RNAs exist as well. Promoters of both transcripts contain CpG islands, which are commonly methylated to influence the transcriptional permissivity of the gene, however, neither are methylated (28, 48) Another feature of these transcripts is the fact that they contain clusters of C TCF binding motifs which are occupied by CTCF in both systems (2, 28) Analyses presented in this disse rtation indicate that the dose of the infection inoculum is an important determinant of the pattern of latent gene expression. Given the similarities between the LAT and XIST, the dosage sensing mechanisms utilized by X inactivation is an attractive model for HSV 1 and could ultimately affect efficiency of establishment by heterochromatin deposition, LAT expression, and reactivation. The Battle for Survival: Neuron vs. HSV -1 Establishment of the latent HSV 1 genome is dependent on the silencing of lytic genes. Control of gene expression during latency on the viral genome is similar to that in cellular systems. The formation of chromatin on the genome as well as the presence of DNA elements such as insulators allow the genomes to be partitioned into area s of active and repressed transcription. It is important for HSV 1 to have these mechanisms in place to control gene expression so it is not just at the mercy of the cell. As the viral DNA enters the nucleus of the cell, the cells response is likely to quickly cover it with nucleosomes and lay down repressive marks so that the DNA does not threaten survival of the neuron. In this race of the cell to shut down the genome and the virus to maintain transcriptional control, there are factors which can give one of the two the advantage. During the process of chromatinization, HSV1 has factors of its own that can control deposition of heterochromatin, one of which is the LAT. It is not a coincidence that one of the abundant transcripts produced during late ncy would be linked to control over heterochromatin. By expressing this non-coding RNA, the virus arms
99 itself with a weapon against too much heterochromatin. For this reason, the LAT may be protecting survival of the HSV 1 genome during latency so that i t may reactivate in the future. After all, HSV 1 LAT mutants have reduced reactivation phenotypes indicating that these mutants are disabled by the cell. It seems that the way in which LAT provides this function is through the presence of cis elements in the LAT promoter that recruit cellular Trithorax proteins. Polycombregulated genes often have both kinds of proteins bound at or upstream of the promoter of a gene and act as a switch for gene expression (57) If binding of the Trithorax proteins is lost at the LAT promoter in the LAT mutant, it would explain the increased amount of triMe H3K27 observed This increase of heterochromatin on key lytic genes such as ICP0 and ICP4 may be too much for the virus to overcome during a reactivation stimulus event resulting in the reduced reactivation phenotype (Figure 6 2) Strain KOS displays a greater heterochromatic profile indicating that it is more shutdown than 17syn + and we know that in the rabbit ocular model KOS does not reactivate with the same efficiency as 17 syn +. Therefore, whatever difference exists between these tw o viruses is likely allowing for increased heterochromatin deposition. When we examined the transcript profile during latency, we noted that the main difference between these strains was LAT expression. In KOS, the LAT transcript is reduced 100 fold rela tive to 17 syn + making it equal to the TAL. Perhaps this reduction in the LAT transcript is the reason for increased facultative heterochromatin deposition since reinforces the role of LAT in controlling triMe H3K27 on HSV 1 and may partly explain why differences in infections between viral strains exist. The importance of regulation of the LAT region transcripts, the TAL and the LAT has emerged from our studies. First we saw how decreasing the LA T has an effect on
100 heterochromatin deposition in KOS. Then we showed that the expression patterns of these two transcripts are dependent on t he amount of input genomes Though w e do not understand how these two transcripts are interacting, it would seem that one might be regulating expression of the other at some viral doses since LAT expression appeared to be incompatible with TAL. This area is currently being investigated and may be another tool for the virus to use in the establishment of the latent genome and maintaining the ability to reactivate from latency. The establishment of the latent genome is clearly a complex multifactoral process that we are just beginning to understand. Even as of yet, we do not know how the LAT prevents heterochromati n formation on some regions of the viral genome Investigating this process further may give us insight into the other factors that are involved in the regulation of chromatin on the genome. If the cell and virus really are fighting a battle, discovering the weapons the virus has evolved to stay ahead in the game would be an advantage for the development of potential therapies. Perhaps we could change the dynamics of the fight to allow the genome to become completely shut down by the cell thereby attain ing a permanent latent infection.
101 Figure 6 1. Mechanism of repression by heterochromatin formation in HSV 1. A) The viral capsid (red hexagon) releases the viral DNA into the cell nucleus where it circularizes. Nucleosomes (blue circles) associate with the viral DNA before Polycomb repressive complex 2 ( PRC2 ) binds to the chromatin and stimulates trimethylation of H3K27 (yellow triangles) B) Trimethylation of H3K27 by PRC2 is kept in balance by the trithorax (Trx) cellular proteins (left) and/or the viral transcript, LAT (red line) (right).
102 Figure 6 2 H eterochromatin deposition has an effect on the reactivation potential of HSV 1 virus strains and mutants. In the wild type 17syn + genome, a proper amount of H3K27 trimethylation is established allowing for efficient reactivation producing virions upon different HSV 1 wild type strain KOS, too much trimethylation of H3K27 is present on the latent genome resulting in a lack of virions being produced after a reactivation stimulus is appl ied to the cell.
103 LIST OF REFERENCES 1. Amelio, A. L., N. V. Giordani, N. J. Kubat, E. O'Neil J, and D. C. Bloom. 2006. Deacetylation of the herpes simplex virus type 1 latencyassociated transcript (LAT) enhancer and a decrease in LAT abundance precede an increase in ICP0 transcriptional permissiveness at early times postexplant. J Virol 80: 20638. 2. Amelio, A. L., P. K. McAnany, and D. C. Bloom. 2006. A chromatin insulator like element in the herpes simplex virus type 1 latencyassociated transcript region binds CCCTC-binding factor and displays enhancer -blocking and silencing activities. J Virol 80: 235868. 3. Arthur, J. L., C. G. Scarpini, V. Connor, R. H. Lachmann, A. M. Tolkovsky, and S. Efstathiou. 2001. Herpes simplex virus type 1 promoter activity during latency establishment, maintenance, and reactivation in primary dorsal root neurons in vitro. J Virol 75: 388595. 4. Barski, A., S. Cuddapah, K. Cui, T. Y. Roh, D. E. Schones, Z. Wang, G. Wei, I. Chepelev, and K. Zhao. 2007. High-resolution profiling of histone methylations in the human genome. Cell 129: 82337. 5. Bloom, D. C., J. M. Hill, G. Devi -Rao, E. K. Wagner, L. T. Feldman, and J. G. Stevens. 1996. A 348-base -pair region in the latency associated transcript facilitates herpes simplex virus type 1 reactivation. J Virol 70: 244959. 6. Bohenzky, R. A., A. G. Papavassiliou, I. H. Gelman, and S. Silverstein. 1993. Identification of a promoter mapping within the reiterated sequences that flank the herpes simplex virus type 1 UL region. J Virol 67: 63242. 7. Changolkar, L. N., and J. R. Pehrson. 2006. macroH2A1 histone variants are depleted on active genes but concentrated on the inactive X chromosome. Mol Cell Biol 26: 441020. 8. Chee, A. V., P. Lopez, P. P. Pandolfi, and B. Roizman. 2003. Promyelocytic leukemia protein mediates interferon -base d anti -herpes simplex virus 1 effects. J Virol 77: 71015. 9. Chelbi Alix, M. K., and H. de The. 1999. Herpes virus induced proteasome -dependent degradation of the nuclear bodies associated PML and Sp100 proteins. Oncogene 18: 935 41. 10. Chen, Q., L. Lin, S. Smith, J. Huang, S. L. Berger, and J. Zhou. 2007. CTCF dependent chromatin boundary element between the latency associated transcript and ICP0 promoters in the herpes simplex virus type 1 genome. J Virol 81: 5192201. 11. Chen, S. H., M. F. Kramer, P. A. Schaffer, and D. M. Coen. 1997. A viral function represses accumulation of transcripts from productive -cycle genes in mouse ganglia latently infected with herpes simplex virus. J Virol 71: 587884.
104 12. Chen, X. P., M. Mata, M. Kelley, J. C. Glorioso, and D. J. Fink. 2002. The relationship of herpes simplex virus latency associated transcript expression to genome copy number: a quantitative study using laser capture microdissection. J Neurovirol 8: 20410. 13. Chirgwin, J. M., A. E. Przybyla, R. J. MacDona ld, and W. J. Rutter. 1979. Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 18: 52949. 14. Cliffe, A. R., D. A. Garber, and D. M. Knipe. 2009. Transcription of the Herpes Simplex Virus Latency -Associat ed Transcript Promotes the Formation of Facultative Heterochromatin on Lytic Promoters. J Virol. 15. Conte, C., B. Dastugue, and C. Vaury. 2002. Coupling of enhancer and insulator properties identified in two retrotransposons modulates their mutagenic impact on nearby genes. Mol Cell Biol 22: 176777. 16. Danaher, R. J., R. J. Jacob, and C. S. Miller. 1999. Establishment of a quiescent herpes simplex virus type 1 infection in neurally -differentiated PC12 cells. J Neurovirol 5: 25867. 17. Deshmane, S. L., and N. W. Fraser. 1989. During latency, herpes simplex virus type 1 DNA is associated with nucleosomes in a chromatin structure. J Virol 63: 943 7. 18. Devi -Rao, G. B., D. C. Bloom, J. G. Stevens, and E. K. Wagner. 1994. Herpes simplex virus type 1 DNA rep lication and gene expression during explant induced reactivation of latently infected murine sensory ganglia. J Virol 68: 127182. 19. Dobson, A. T., T. P. Margolis, F. Sedarati, J. G. Stevens, and L. T. Feldman. 1990. A latent, nonpathogenic HSV 1 derived vector stably expresses beta -galactosidase in mouse neurons. Neuron 5: 35360. 20. Dobson, A. T., F. Sederati, G. Devi -Rao, W. M. Flanagan, M. J. Farrell, J. G. Stevens, E. K. Wagner, and L. T. Feldman. 1989. Identification of the latencyassociated trans cript promoter by expression of rabbit beta -globin mRNA in mouse sensory nerve ganglia latently infected with a recombinant herpes simplex virus. J Virol 63: 384451. 21. Doyen, C. M., W. An, D. Angelov, V. Bondarenko, F. Mietton, V. M. Studitsky, A. Hamiche, R. G. Roeder, P. Bouvet, and S. Dimitrov. 2006. Mechanism of polymerase II transcription repression by the histone variant macroH2A. Mol Cell Biol 26: 115664. 22. Dressler, G. R., D. L. Rock, and N. W. Fraser. 1987. Latent herpes simplex virus type 1 DNA is not extensively methylated in vivo. J Gen Virol 68 ( Pt 6):17615.
105 23. Edwards, D., and J. J. Berry. 1987. The efficiency of simulation-based multiple comparisons. Biometrics 43: 91328. 24. Everett, R. D. 2000. ICP0, a regulator of herpes simplex virus during lytic and latent infection. Bioessays 22: 76170. 25. Farrell, M. J., A. T. Dobson, and L. T. Feldman. 1991. Herpes simplex virus latencyassociated transcript is a stable intron. Proc Natl Acad Sci U S A 88: 7904. 26. Farrell, M. J., T. P. Margolis, W. A. Gomes, and L. T. Feldman. 1994. Effect of the transcription start region of the herpes simplex virus type 1 latencyassociated transcript promoter on expression of productively infected neurons in vi vo. J Virol 68: 533743. 27. Feldman, L. T., A. R. Ellison, C. C. Voytek, L. Yang, P. Krause, and T. P. Margolis. 2002. Spontaneous molecular reactivation of herpes simplex virus type 1 latency in mice. Proc Natl Acad Sci U S A 99: 97883. 28. Filippova, G N., M. K. Cheng, J. M. Moore, J. P. Truong, Y. J. Hu, D. K. Nguyen, K. D. Tsuchiya, and C. M. Disteche. 2005. Boundaries between chromosomal domains of X inactivation and escape bind CTCF and lack CpG methylation during early development. Dev Cell 8: 314 2. 29. Gerasimova, T. I., K. Byrd, and V. G. Corces. 2000. A chromatin insulator determines the nuclear localization of DNA. Mol Cell 6: 102535. 30. Geyer, P. K., and V. G. Corces. 1992. DNA position-specific repression of transcription by a Drosophila z inc finger protein. Genes Dev 6: 186573. 31. Giordani, N. V., D. M. Neumann, D. L. Kwiatkowski, P. S. Bhattacharjee, P. K. McAnany, J. M. Hill, and D. C. Bloom. 2008. During herpes simplex virus type 1 infection of rabbits, the ability to express the late ncy associated transcript increases latent -phase transcription of lytic genes. J Virol 82: 605660. 32. Gordon, Y. J., E. G. Romanowski, T. Araullo -Cruz, and P. R. Kinchington. 1995. The proportion of trigeminal ganglionic neurons expressing herpes simplex virus type 1 latency associated transcripts correlates to reactivation in the New Zealand rabbit ocular model. Graefes Arch Clin Exp Ophthalmol 233: 64954. 33. Gressens, P., and J. R. Martin. 1994. In situ polymerase chain reaction: localization of HSV2 DNA sequences in infections of the nervous system. J Virol Methods 46: 6183. 34. Grimaud, C., N. Negre, and G. Cavalli. 2006. From genetics to epigenetics: the tale of Polycomb group and trithorax group genes. Chromosome Res 14: 36375. 35. Hagglund, R., and B. Roizman. 2004. Role of ICP0 in the strategy of conquest of the host cell by herpes simplex virus 1. J Virol 78: 216978.
106 36. Hill, J. M., J. B. Maggioncalda, H. H. Garza, Jr., Y. H. Su, N. W. Fraser, and T. M. Block. 1996. In vivo epinephrine reacti vation of ocular herpes simplex virus type 1 in the rabbit is correlated to a 370 base -pair region located between the promoter and the 5' end of the 2.0 kilobase latencyassociated transcript. J Virol 70: 72704. 37. Hill, J. M., M. A. Rayfield, and Y. Haruta. 1987. Strain specificity of spontaneous and adrenergically induced HSV 1 ocular reactivation in latently infected rabbits. Curr Eye Res 6: 91 7. 38. Jamieson, D. R., L. H. Robinson, J. I. Daksis, M. J. Nicholl, and C. M. Preston. 1995. Quiescent vi ral genomes in human fibroblasts after infection with herpes simplex virus type 1 Vmw65 mutants. J Gen Virol 76 ( Pt 6): 141731. 39. Jarman, R. G., J. M. Loutsch, G. B. Devi -Rao, M. E. Marquart, M. P. Banaszak, X. Zheng, J. M. Hill, E. K. Wagner, and D. C Bloom. 2002. The region of the HSV 1 latency associated transcript required for epinephrine induced reactivation in the rabbit does not include the 2.0 kb intron. Virology 292: 5969. 40. Katz, J. P., E. T. Bodin, and D. M. Coen. 1990. Quantitative polym erase chain reaction analysis of herpes simplex virus DNA in ganglia of mice infected with replication incompetent mutants. J Virol 64: 428895. 41. Kaufman, H. E., A. M. Azcuy, E. D. Varnell, G. D. Sloop, H. W. Thompson, and J. M. Hill. 2005. HSV 1 DNA in tears and saliva of normal adults. Invest Ophthalmol Vis Sci 46: 2417. 42. Kellum, R., and P. Schedl. 1992. A group of scs elements function as domain boundaries in an enhancer -blocking assay. Mol Cell Biol 12: 242431. 43. Kent, J. R., P. Y. Zeng, D. At anasiu, J. Gardner, N. W. Fraser, and S. L. Berger. 2004. During lytic infection herpes simplex virus type 1 is associated with histones bearing modifications that correlate with active transcription. J Virol 78: 1017886. 44. Kim, S. Y., S. W. Paylor, T. Magnuson, and A. Schumacher. 2006. Juxtaposed Polycomb complexes co regulate vertebral identity. Development 133: 4957 68. 45. Kim, T. H., Z. K. Abdullaev, A. D. Smith, K. A. Ching, D. I. Loukinov, R. D. Green, M. Q. Zhang, V. V. Lobanenkov, and B. Ren. 2007. Analysis of the vertebrate insulator protein CTCF -binding sites in the human genome. Cell 128: 123145. 46. Kramer, M. F., and D. M. Coen. 1995. Quantification of transcripts from the ICP4 and thymidine kinase genes in mouse ganglia latently infected w ith herpes simplex virus. J Virol 69: 138999.
107 47. Kubat, N. J., A. L. Amelio, N. V. Giordani, and D. C. Bloom. 2004. The herpes simplex virus type 1 latency associated transcript (LAT) enhancer/rcr is hyperacetylated during latency independently of LAT tr anscription. J Virol 78: 1250818. 48. Kubat, N. J., R. K. Tran, P. McAnany, and D. C. Bloom. 2004. Specific histone tail modification and not DNA methylation is a determinant of herpes simplex virus type 1 latent gene expression. J Virol 78: 113949. 49. Kutluay, S. B., S. L. DeVos, J. E. Klomp, and S. J. Triezenberg. 2009. Transcriptional coactivators are not required for herpes simplex virus type 1 immediate -early gene expression in vitro. J Virol 83: 343649. 50. Kwiatkowski, D. L., H. W. Thompson, and D. C. Bloom. 2009. Polycomb group protein Bmi1 binds to the HSV 1 latent genome and maintains repressive histone marks during latency. J Virol. 51. Lee, L. Y., and P. A. Schaffer. 1998. A virus with a mutation in the ICP4-binding site in the L/ST promoter of herpes simplex virus type 1, but not a virus with a mutation in open reading frame P, exhibits cell -type -specific expression of gamma(1)34.5 transcripts and latency associated transcripts. J Virol 72: 425064. 52. Leib, D. A., D. M. Coen, C. L. Bogard, K. A. Hicks, D. R. Yager, D. M. Knipe, K. L. Tyler, and P. A. Schaffer. 1989. Immediate -early regulatory gene mutants define different stages in the establishment and reactivation of herpes simplex virus latency. J Virol 63: 75968. 53. Leinbach, S. S., and W. C. Summers. 1980. The structure of herpes simplex virus type 1 DNA as probed by micrococcal nuclease digestion. J Gen Virol 51: 4559. 54. Lentine, A. F., and S. L. Bachenheimer. 1990. Intracellular organization of herpes simplex virus type 1 DNA as sayed by staphylococcal nuclease sensitivity. Virus Res 16: 275 92. 55. Liesegang, T. J. 2001. Herpes simplex virus epidemiology and ocular importance. Cornea 20: 1 13. 56. Lindroth, A. M., D. Shultis, Z. Jasencakova, J. Fuchs, L. Johnson, D. Schubert, D. Patnaik, S. Pradhan, J. Goodrich, I. Schubert, T. Jenuwein, S. Khorasanizadeh, and S. E. Jacobsen. 2004. Dual histone H3 methylation marks at lysines 9 and 27 required for interaction with CHROMOMETHYLASE3. EMBO J 23: 428696. 57. Lund, A. H., and M. van L ohuizen. 2004. Polycomb complexes and silencing mechanisms. Curr Opin Cell Biol 16: 23946.
108 58. Mackem, S., and B. Roizman. 1982. Differentiation between alpha promoter and regulator regions of herpes simplex virus 1: the functional domains and sequence of a movable alpha regulator. Proc Natl Acad Sci U S A 79: 491721. 59. Maggioncalda, J., A. Mehta, N. W. Fraser, and T. M. Block. 1994. Analysis of a herpes simplex virus type 1 LAT mutant with a deletion between the putative promoter and the 5' end of the 2.0 kilobase transcript. J Virol 68: 781624. 60. Maggioncalda, J., A. Mehta, Y. H. Su, N. W. Fraser, and T. M. Block. 1996. Correlation between herpes simplex virus type 1 rate of reactivation from latent infection and the number of infected neurons in tr igeminal ganglia. Virology 225: 72-81. 61. Margolis, T. P., F. L. Elfman, D. Leib, N. Pakpour, K. Apakupakul, Y. Imai, and C. Voytek. 2007. Spontaneous reactivation of herpes simplex virus type 1 in latently infected murine sensory ganglia. J Virol 81: 11069 74. 62. Mehta, A., J. Maggioncalda, O. Bagasra, S. Thikkavarapu, P. Saikumari, T. Valyi Nagy, N. W. Fraser, and T. M. Block. 1995. In situ DNA PCR and RNA hybridization detection of herpes simplex virus sequences in trigeminal ganglia of latently infect ed mice. Virology 206: 63340. 63. Milliken, G. A., and D. E. Johnson. 1984. Analysis of messy data. Lifetime Learning Publications, Belmont, Calif. 64. Mouttet, M. E., D. Guetard, and J. M. Bechet. 1979. Random cleavage of intranuclear herpes simplex virus DNA by micrococcal nuclease. FEBS Lett 100: 1079. 65. Muggeridge, M. I., and N. W. Fraser. 1986. Chromosomal organization of the herpes simplex virus genome during acute infection of the mouse ce ntral nervous system. J Virol 59: 764 7. 66. Mukerjee, R., W. Kang, V. Suri, and N. W. Fraser. 2004. A non-consensus branch point plays an important role in determining the stability of the 2 kb LAT intron during acute and latent infections of herpes simpl ex virus type 1. Virology 324: 340 9. 67. Nakayama, J., J. C. Rice, B. D. Strahl, C. D. Allis, and S. I. Grewal. 2001. Role of histone H3 lysine 9 methylation in epigenetic control of heterochromatin assembly. Science 292: 1103. 68. Neumann, D. M., P. S. Bhattacharjee, N. V. Giordani, D. C. Bloom, and J. M. Hill. 2007. In vivo changes in the patterns of chromatin structure associated with the latent herpes simplex virus type 1 genome in mouse trigeminal ganglia can be detected at early times after butyrate treatment. J Virol 81: 1324853.
109 69. Neumann, D. M., P. S. Bhattacharjee, and J. M. Hill. 2007. Sodium butyrate: a chemical inducer of in vivo reactivation of herpes simplex virus type 1 in the ocular mouse model. J Virol 81: 610610. 70. Oh, J., and N. W Fraser. 2008. Temporal association of the herpes simplex virus genome with histone proteins during a lytic infection. J Virol 82: 35307. 71. Perng, G. C., K. Chokephaibulkit, R. L. Thompson, N. M. Sawtell, S. M. Slanina, H. Ghiasi, A. B. Nesburn, and S. L. Wechsler. 1996. The region of the herpes simplex virus type 1 LAT gene that is colinear with the ICP34.5 gene is not involved in spontaneous reactivation. J Virol 70: 28291. 72. Perng, G. C., H. Ghiasi, S. M. Slanina, A. B. Nesburn, and S. L. Wechsler 1996. High dose ocular infection with a herpes simplex virus type 1 ICP34.5 deletion mutant produces no corneal disease or neurovirulence yet results in wild -type levels of spontaneous reactivation. J Virol 70: 288393. 73. Perng, G. C., H. Ghiasi, S. M. Slanina, A. B. Nesburn, and S. L. Wechsler. 1996. The spontaneous reactivation function of the herpes simplex virus type 1 LAT gene resides completely within the first 1.5 kilobases of the 8.3 kilobase primary transcript. J Virol 70: 976 84. 74. Perng, G. C., C. Jones, J. Ciacci -Zanella, M. Stone, G. Henderson, A. Yukht, S. M. Slanina, F. M. Hofman, H. Ghiasi, A. B. Nesburn, and S. L. Wechsler. 2000. Virus induced neuronal apoptosis blocked by the herpes simplex virus latencyassociated transcript. Science 287: 15003. 75. Perng, G. C., S. M. Slanina, H. Ghiasi, A. B. Nesburn, and S. L. Wechsler. 1996. A 371nucleotide region between the herpes simplex virus type 1 (HSV 1) LAT promoter and the 2 kilobase LAT is not essential for efficient spontaneous reacti vation of latent HSV1. J Virol 70: 20148. 76. Pesola, J. M., J. Zhu, D. M. Knipe, and D. M. Coen. 2005. Herpes simplex virus 1 immediate -early and early gene expression during reactivation from latency under conditions that prevent infectious virus production. J Virol 79: 1451625. 77. Pignatti, P. F., and E. Cassai. 1980. Analysis of herpes simplex virus nucleoprotein complexes extracted from infected cells. J Virol 36: 81628. 78. Pillai, R. S., S. N. Bhattacharyya, and W. Filipowicz. 2007. Repression of protein synthesis by miRNAs: how many mechanisms? Trends Cell Biol 17: 11826. 79. Quina, A. S., M. Buschbeck, and L. Di Croce. 2006. Chromatin structure and epigenetics. Biochem Pharmacol 72: 15639.
110 80. Regad, T., and M. K. Chelbi Al ix. 2001. Role and fate of PML nuclear bodies in response to interferon and viral infections. Oncogene 20: 727486. 81. Rock, D. L., and N. W. Fraser. 1985. Latent herpes simplex virus type 1 DNA contains two copies of the virion DNA joint region. J Virol 55: 84952. 82. Sawtell, N. M. 1997. Comprehensive quantification of herpes simplex virus latency at the single -cell level. J Virol 71: 542331. 83. Sawtell, N. M., D. K. Poon, C. S. Tansky, and R. L. Thompson. 1998. The latent herpes simplex virus type 1 genome copy number in individual neurons is virus strain specific and correlates with reactivation. J Virol 72: 534350. 84. Sawtell, N. M., and R. L. Thompson. 1992. Herpes simplex virus type 1 latencyassocia ted transcription unit promotes anatomical site -dependent establishment and reactivation from latency. J Virol 66: 215769. 85. Sawtell, N. M., and R. L. Thompson. 1992. Rapid in vivo reactivation of herpes simplex virus in latently infected murine ganglionic neurons after transient hyperthermia. J Virol 66: 21506. 86. Schillinger, J. A., F. Xu, M. R. Sternberg, G. L. Armstrong, F. K. Lee, A. J. Nahmias, G. M. McQuillan, M. E. Louis, and L. E. Markowitz. 2004. National seroprevalence and trends in herpes s implex virus type 1 in the United States, 19761994. Sex Transm Dis 31: 75360. 87. Schwartz, Y. B., and V. Pirrotta. 2008. Polycomb complexes and epigenetic states. Curr Opin Cell Biol 20: 26673. 88. Smith, G. A., S. P. Gross, and L. W. Enquist. 2001. He rpesviruses use bidirectional fast axonal transport to spread in sensory neurons. Proc Natl Acad Sci U S A 98: 346670. 89. Smith, J. S., and N. J. Robinson. 2002. Age -specific prevalence of infection with herpes simplex virus types 2 and 1: a global revie w. J Infect Dis 186 Suppl 1: S3 28. 90. Strahl, B. D., and C. D. Allis. 2000. The language of covalent histone modifications. Nature 403: 415. 91. Stroop, W. G., and M. C. Banks. 1994. Herpes simplex virus type 1 strain KOS 63 does not cause acute or recu rrent ocular disease and does not reactivate ganglionic latency in vivo. Acta Neuropathol 87: 1422. 92. Su, Y. H., R. L. Meegalla, R. Chowhan, C. Cubitt, J. E. Oakes, R. N. Lausch, N. W. Fraser, and T. M. Block. 1999. Human corneal cells and other fibroblasts can stimulate the appearance of herpes simplex virus from quiescently infected PC12 cells. J Virol 73: 417180.
111 93. Taylor, J. L., D. Unverrich, W. J. O'Brien, and K. W. Wilcox. 2000. Interferon coordinately inhibits the disruption of PML -positive ND10 and immediate -early gene expression by herpes simplex virus. J Interferon Cytokine Res 20: 80515. 94. Thomas, D. L., M. Lock, J. M. Zabolotny, B. R. Mohan, and N. W. Fraser. 2002. The 2 kilobase intron of the h erpes simplex virus type 1 latencyassociated transcript has a half life of approximately 24 hours in SY5Y and COS 1 cells. J Virol 76: 532 40. 95. Trojer, P., and D. Reinberg. 2007. Facultative heterochromatin: is there a distinctive molecular signature? Mol Cell 28: 1 13. 96. Umbach, J. L., M. F. Kramer, I. Jurak, H. W. Karnowski, D. M. Coen, and B. R. Cullen. 2008. MicroRNAs expressed by herpes simplex virus 1 during latent infection regulate viral mRNAs. Nature 454: 7803. 97. Vakoc, C. R., M. M. Sachde va, H. Wang, and G. A. Blobel. 2006. Profile of histone lysine methylation across transcribed mammalian chromatin. Mol Cell Biol 26: 918595. 98. Wang, Q. Y., C. Zhou, K. E. Johnson, R. C. Colgrove, D. M. Coen, and D. M. Knipe. 2005. Herpesviral latencyassociated transcript gene promotes assembly of heterochromatin on viral lytic -gene promoters in latent infection. Proc Natl Acad Sci U S A 102: 160559. 99. Wilcox, C. L., and E. M. Johnson, Jr. 1988. Characterization of nerve g rowth factor dependent herpes simplex virus latency in neurons in vitro. J Virol 62: 3939. 100. Xu, F., J. A. Schillinger, M. R. Sternberg, R. E. Johnson, F. K. Lee, A. J. Nahmias, and L. E. Markowitz. 2002. Seroprevalence and coinfection with herpes simp lex virus type 1 and type 2 in the United States, 19881994. J Infect Dis 185: 101924. 101. Xu, F., M. R. Sternberg, B. J. Kottiri, G. M. McQuillan, F. K. Lee, A. J. Nahmias, S. M. Berman, and L. E. Markowitz. 2006. Trends in herpes simplex virus type 1 a nd type 2 seroprevalence in the United States. JAMA 296: 96473. 102. Yeh, L., and P. A. Schaffer. 1993. A novel class of transcripts expressed with late kinetics in the absence of ICP4 spans the junction between the long and short segments of the herpes s implex virus type 1 genome. J Virol 67: 737382. 103. Zhao, H., and A. Dean. 2004. An insulator blocks spreading of histone acetylation and interferes with RNA polymerase II transfer between an enhancer and gene. Nucleic Acids Res 32: 490319.
112 BIOGRAPHICAL SKETCH Dacia Kwiatkowski grew up in Albany, NY and attended Guilderland High School. Afterwards, she majored in Biology at Marist College in Poughkeepsie, NY while participating in crew and band. All four years were spent perfor ming undergraduate research under the direction of Raymond Kepner in ecological microbiology. The summer of 2003 was spent at an internship at the David Axelrod Institute, a part of the Wadsworth Health Center in Albany, NY under the direction of Gary Wi nslow studying the effect of E rlichea on the immune system. After graduating from Marist College she attended the University of Florida s Interdisciplinary Program (IDP) in the College of Medicine where she performed her graduate work under the direction of David Bloom on HSV 1 latency. Upon earning her Ph.D., Dacia performed post doctoral work at the Cornell Medical Hospital studying antigenic variation in malaria in New York, New York.