Role of herpes simplex vinis type-1 latency associated transcript (lat) on establishment and reactivation


Material Information

Role of herpes simplex vinis type-1 latency associated transcript (lat) on establishment and reactivation
Physical Description:
x, 91 leaves : ill. ; 29 cm.
O'Neil, Jerome E., 1972-
Publication Date:


Subjects / Keywords:
Medical Sciences thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Medical Sciences -- UF   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 2004.
Includes bibliographical references.
Statement of Responsibility:
by Jerome E. O'Neil.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 003090233
System ID:

Full Text








I would like to thank a number of people for their assistance, encouragement, and

friendship over the past several years. While the specifics of individual projects will

fade, I will always remember the people with whom my time was spent.

My sincerest gratitude goes out to Dave and Linda Bloom for their endless

kindness. You will not find two nicer people, and I wish them all the best.

I extend my appreciation to the original crew from Arizona: Robert, Niki, Anne,

and Lee. We can all agree it has been an interesting ride. I wish them good luck

wherever the next road leads (and suggest that they avoid I-10).

It is time to move on when the new guys have been around for 3 years. I would

like to thank the second-generation students (Tony, Zane, and Nicole) for their friendship

and enthusiasm. It is comforting to know that the lab is being left in good hands. Thanks

also go to Peterjon for his various sorts of expertise. I will miss his lengthy stories.

Thanks to the entire U.F. faculty for allowing our transition to proceed so

smoothly. I especially thank my committee members (Dr. Maurice Swanson, Dr. Sankar

Swaminathan, and Dr. Jirg Bungert) for their guidance. I was fortunate to have

surrounded myself with such approachable individuals.

Finally, and most importantly, I would like to thank my family for all of their

support. My parents probably want this more than I do. I thank them and apologize for

being so long.



ACKNOW LEDGM ENTS ............................................................................................ ii

LIST OF TABLES....................................................................................................... vi

LIST OF FIGURES .......................................................................................................... vii

ABSTRACT....................................................................................................................... ix

INTRODUCTION .......................................................................................................1

Acute Infection ....................................................................................................... 3
Genome Structure and Organization ............................................... ..............3
Primary Infection............................................................................................. 4
Replication Cascade ......................................................... ............................. 5
Immediate Early Genes .............................................................. ................... 5
Infected cell polypeptide-4 (ICP4) ................................... .............................7
Infected cell polypeptide-0 (ICPO) ................................... ............................8
Infected cell polypeptide-27 (ICP27) ........................................... .............. 10
Early Genes .................................................................................................... 1
Late Genes ..................................................................................................... 12
Virulence protein- 16 (VP 16)..................................................... ..................13
Virion host shut-off (VHS) protein ............................................... .............. 13
Latency .................................................................................................................14
Establishment ................................................................................................ 15
M maintenance .................................................................................................. 16
Reactivation................................................................................................... 17
Role of Neurons in Latency............................................................................18
Octamer binding proteins ................................................... ......................... 20
Latency Associated Transcript (LAT) .....................................................................22
Cyclic-AMP response element binding (CREB) proteins.................................25
Alternate M odel for LAT ................................................... ......................... 27
Animal M models ..................................................................................................... 29
M urine M models .............................................................................................. 29
Rabbit M models ...............................................................................................31
Cell Culture M odels .......................................... ............... ...........................32

ESTABLISHMENT OF LATENCY.................................................................. 39

M materials and M methods ......................................................................................... 40
C ells and V iruses...........................................................................................40
Infections and Reactivation ................................... .........................................41
D N A Extraction............................................................................................. 41
Analysis of the Relative Amounts of Viral DNA by PCR................................41
PCR Analysis to Determine Relative Levels of Latent Viral DNA and Wild
Type Revertants .......................................................................................42
R esults................................................................................................................. 43
Acute Replication in Rabbit Corneas and Trigeminal Ganglia Following High
T iter Infection ............................................................................................ 43
Analysis of Viral DNA Levels in Corneas and Trigeminal Ganglia During the
A cute Infection........................................................................................... 44
The Relative Amount of Latent Viral DNA in Trigeminal Ganglia of Rabbits
Infected With Wild Type or LAT Mutants are Similar Regardless of Infecting
D ose ........................................................................................................ 45
A Non-replicating HSV-1 Recombinant Establishes a Latent Infection in the
Trigeminal Ganglia, but at Lower Levels Than Wild-Type Virus .................45
D discussion ............................................................................................................. 46


M materials and M methods ......................................................................................... 58
Cells and V iruses...........................................................................................58
M house Infections ........................................................................................... 59
Generation of LAT Transgenic ........................................................................61
Explant Co-cultivation.................................................................................61
RN A Processing ............................................................................................ 61
Reverse Transcription..................................................................................62
D N A Processing ............................................................................................ 63
PCR A nalysis.............................................................................................. 64
Real-Time PCR Analysis ............................................................................ 65
R esults................................................................................................................. 66
A Severe Decrease in LAT Abundance Occurs During the First Hours of
Explant Co-cultivation of Latently Infected Dorsal Root Ganglia..................68
Reduced Abundance of LAT Also Occurs During Co-cultivation of Uninfected
Transgenic Mouse Dorsal Root Ganglia......................................................69
Detection of Immediate Early Transcripts During Explant Co-cultivation.........70
Does Lytic Gene Activation Require a State of LAT-Repression ....................71
Which is the First Immediate Early Gene Activated During Reactivation .........72
D discussion .............................................................................................................73

Decreased LAT Abundance During Explant Co-cultivation.............................74
Lytic Gene Activation During Explant Co-cultivation, and the Requirement of
LAT-Repression................................................... ....................................76
Proposed Mechanism for LAT's Role in Reactivation .....................................77

REFEREN CE LIST ..................................................................................................... 87

BIOGRAPHICAL SKETCH ....................................................................................... 91


Table page

Table 1.1. The Eight Human Herpes Viruses ............................................................... 37

Table 1.2. The five immediate early (IE) genes of HSV-1 .............................................37

Table 1.3. Non-essential early genes ........................................................................ 38

Table 1.4. Seven essential early genes...................................................................... 38

Table 1.5. CREB signal transduction pathway ............................................................38

Table 2.1. PCR Prim ers .............................................................................................. 52

Table 2.2. Relative amount of viral DNA (VP5/actin) at high dose of inoculation
(500,000 pfu)* .................................................................................................... 52

Table 2.3. Relative amounts of viral DNA in cornea and trigeminal ganglia during acute
infections post low dose inoculation with viruses of different LAT genotypes*.....53

Table 2.4. Relative amounts of viral DNA present in TG during latency in rabbits
infected with different doses of virus* .................................................................54


Figure page

Figure 1.1. Schem atic ofHSV-1 virion. ........................................................................34

Figure 1.2. Organization of linear HSV-1 genome......................................................34

Figure 1.3. Latent H SV-1 episom e. ................................... ............................................34

Figure 1.4. Genomic location of Immediate Early genes and LAT................................35

Figure 1.5. Splicing of primary LAT to produce a stable 2.0kb intron lariat. A) The 3'-
OH group of a non-consensus branch point (guanine) attacks the 5' exon/intron
border, B) The free 3'-OH of the 5' exon attacks the 3' exon/intron border to
connect exons and C) The intron remains as a lariat (5'-2' linkage), non-consensus
branch point inhibits debranching. ........................................................................35

Figure 1.6. Schematic representation of core LAT promoter elements............................36

Figure 1.7. Viral mutants used to map reactivation critical region (RCR).......................36

Figure 2.1. Viral titers recovered from eye swabs....................................................49

Figure 2.2. Viral titers recovered from corneas...........................................................49

Figure 2.3. Viral titers recovered from trigeminal ganglia.............................................50

Figure 2.4. Levels of establishment using high and low does 17APst (and Rescue)........50

Figure 2.5. Reduced establishment using KD6 versus 17syn+............................... ..51

Figure 3.1. 17syn+ co-cultivation experiment number one............................................81

Figure 3.2. 17syn+ co-cultivation experiment number two............................................81

Figure 3.3. KOS co-cultivation experiment number one........................................ ..82

Figure 3.4. KOS co-cultivation experiment number two..........................................82

Figure 3.5. Standard PCR analysis of 17syn+ co-cultivation experiment one and two. ..83

Figure 3.6. LAT-Transgenic co-cultivation experiment number one.............................83

Figure 3.7. LAT-Transgenic co-cultivation experiment number two.............................84

Figure 3.8. LAT-Transgenic co-cultivation experiment number three...........................84

Figure 3.9. LAT-Transgenic co-cultivation experiment number four............................85

Figure 3.10. Standard PCR analysis of LAT-Transgenic experiments 1-4 ....................85

Figure 3.11. Detection of ICPO and ICP4 from 17syn+ experiment number one............86

Figure 3.12. Detection of ICPO and ICP4 from 17APst and KOS infected mice.............86

Figure 3.13. Levels of establishment from ICPO(-), ICP4(-) and KOS viruses................86

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



Jerome E. O'Neil

May 2004

Chair: David C. Bloom
Major Department: Medical Sciences-Immunology and Microbiology

Our study examined the latency-associated transcript (LAT) of herpes simplex

virus type-1 (HSV-1) and its contribution to both establishments of latency and

reactivation. Experimental analysis includes the two predominant animal models for

HSV-1 research: rabbits for studying acute replication kinetics and establishment of

latency; and mice for examining LAT's involvement in reactivation.

Results from the rabbit experiments demonstrate that, within our limits of

sensitivity, neither acute replication kinetics nor levels of latent genomes are detectably

altered by either inoculating dose or LAT genotype. Analyses are extended to indicate

the extent to which peripheral replication contributes to levels of latent viral genomes in

the trigeminal ganglia.

Results from murine models demonstrate a dramatic decrease in LAT abundance

during the first hours of explant co-cultivation of latently infected dorsal root ganglia

(DRG). The fact that this decrease is reproducible in both latently infected mice and in

an uninfected LAT-transgenic mouse line indicates that the LAT locus is responsible for

relaying cellular reactivation signals independent of exogenous viral factors. These

results also offer tentative insight regarding the role of chromatin boundaries in

regulating HSV- 's lytic and latent stages. Finally, while results are unable to

demonstrate a direct correlation with LAT kinetics, Immediate Early (IE) transcripts are

reliably detected during the early hours post-explant.

Results reported herein address current research topics, revealing molecular events

of both establishment of latency and reactivation, while also demonstrating the variability

inherent to both model systems.


Experimental approaches to address the fundamental questions of Herpes Simplex

Type-1 (HSV-1) should begin by recognizing that herpes viruses represent an

evolutionarily ancient virus family, and have undergone co-speciation with their hosts for

millennia'. One should not expect the intimate relationship between an infected cell and

its lifelong viral companion to be simple. Herpes viruses' large and complex genomes

clearly indicate an ability to obtain cellular machinery and mechanisms that contribute to

their efficient and unique biology. As a virus whose host serves as its only reservoir,

HSV-1 must take great care to minimize deleterious effects while maintaining a low

immunologic profile. In my opinion, the secrets of HSV-1 will ultimately be revealed by

applying advancements in cellular gene regulation to HSV-1 systems. This benefit will

ultimately be repaid to cellular biologists when HSV-1 is realized as an ideal 'mini-

chromosome' for studying an array of eukaryotic gene functions from a much simpler


Herpes viruses infect members of virtually all groups of vertebrates. At least 80

distinct isolates have been identified from a variety of species2. Because of the relatively

benign effects of an infection in its native host, many isolates tend to be overlooked

unless specifically sought; almost all vertebrate species that have been carefully

examined are found to support at least one host-specific herpesvirus3'4. Humans are

clearly no exception to this rule. Since transmission is most efficient in "herd

populations," the social nature of early humans (or their predecessors) provided an ideal

environment for colonization. Descriptions of human infections date to ancient Greece

5,6, and can be found in the writings of Hippocrates and Shakespeare. The word herpes

actually derives from the Greek "to creep," referring to the ability of lesions to

periodically reappear over the course of an individual's life. Eight discrete human herpes

viruses have been identified to date and are mentioned briefly in (Table 1.1).2

Two basic features define all herpes viruses regardless of genome size and

structure, cell tropism, or replication kinetics. The first feature is their patterns of

productive cycle gene expression. As discussed in detail later, acute replication occurs in

a well ordered cascade beginning with Immediate Early (IE/o), then Early (E/P), and

finally Late (L/y) transcripts. The second feature is the ability to establish and maintain a

lifelong latent infection in their host, from which periodic reactivations may occur in

response to a variety of stress-stimuli. Discussion hereafter focuses on the prototypical

alphaherpesvirus, HSV-1.

HSV-1 is a nuclear replicating, double-stranded DNA virus, enclosed in both an

icosadeltahedral capsid and an envelope. Thel52-kilobase (kb) genome is promoter rich,

generally using one promoter per protein. Studying a specific gene of interest is typically

feasible because of the relative absence of overlapping reading frames. The large

genome allows considerable flexibility in accepting transgenes without disrupting

neighboring genes; this extensive "payload capacity" makes the construction of ectopic

reporter constructs and promoter mutants relatively straightforward. Primary HSV-1

infections in immunocompetent individuals are generally unremarkable, often occurring

without production of a detectable peripheral lesion. While periodic reactivations have

the potential to produce a recurrent lesion at the initial site of infection, many (possibly

even most) reactivations result in no obvious clinical symptoms. The ability to detect

virus from oral swabs by PCR, in the absence of clinical symptoms, underscores the

potential for person-to-person spread. While lesions are uncomfortable, HSV-1 is of little

medical concern to healthy individuals, even those prone to severe and common

reactivations. Immunosuppressed individuals (elderly, neonates, transplant patients,

HIV+ patients, etc.) are much more inclined to suffer from a disseminated infection,

which may include extreme peripheral lesions and possibly fatal encephalitis.

Acute Infection

Genome Structure and Organization

HSV-1 is a large, double-stranded DNA virus that, like all herpes viruses, consists

of four basic structural elements: 1) a core, containing the 152 kb genome wrapped in a

torroid form, 2) a 100-110 nm icosadeltahedral capsid, composed of 162 capsomer

components (each component consisting of at least 6 proteins), 3) an amorphous layer of

tegument surrounding the capsid, containing 15-20 proteins, and 4) an envelope derived

from the nuclear membrane, containing at least 11 different viral encoded glycoproteins7

(Figure 1.1). The HSV-1 genome has a high GC content (68%) and is organized into two

unique sequences (unique long, UL, and unique short, Us), each of which is flanked by

inverted repeats (repeat long, RL, and repeat short, Rs). The architecture of the genome

permits four individual isoforms to exist based on the conformation of the Rs segments

relative to the UL region8(Figure 1.2).

The HSV-1 genome encodes 75-80 proteins, most of which are controlled by their

own specific promoter. While genes located within the repeat regions are present in two

copies, no distinct location is required for a gene to be active. In general, as long as the

entire promoter and ORF are included, individual genes can be moved great distances

without affecting their kinetics or the viruses' overall biology. HSV-l's unique existence

is directly linked to its ability to acquire homologues of critical cellular genes that permit

its existence in the non-permissive environment of the nervous system.

Primary Infection

Since HSV-1 is a non-aerosolized, enveloped virus, it is extremely sensitive to

desiccation and requires direct contact of moist mucosal surfaces for its spread. An

HSV-1 infection typically initiates within the oral-facial area. The virus contacts

mucosal surfaces and replicates within cornified epithelial cells; eventually gaining

access to, and replicating within, cells of the basal epithelial membrane. The primary

infection is generally unremarkable and resolves within 2-3 weeks after seroconversion.

While the virus is ultimately cleared, and the individual makes a full recovery, its true

mission has long since been achieved. During replication, the virus gains access to axons

of sensory and motor neurons of the peripheral nervous system (PNS) innervating the

infected region. Fusion of viral and cellular membranes directed by specific glycoprotein

receptors delivers the capsid into the cell. By a process called retrograde fast axonal

transport, the viral capsid ascends the axon at a rate of approximately 1 cm/hr. The virion

arrives at the soma and delivers its nucleocapsid to the nuclear pores, releasing the viral

DNA into the nucleus. Once in the nucleus the virus follows one of two fates: acute

genes may be activated to initiate replication and produce progeny virus; or genomes

circularize to yield a transcriptionally silent (latent) episome (Figure 1.3). Factors

involved in this process, both cellular and viral, are only partially understood and

continue to be the focus of research and controversy. For organizational purposes, our

discussion continues with the population that "chose" the acute pathway, followed by a

detailed discussion of events associated with latency.

Replication Cascade

Alphaherpesviruses have an extremely rapid replication cycle, compared to other

herpesvirus family members. In some cell culture systems, infectious progeny can be

produced in as little as 8 hours. Replication in vivo may not necessarily proceed this

rapidly, it can take 18-20 hrs in neuronal cells7. Though the specifics of the HSV-1

replication cascade have been determined largely in cell culture, the use of neuronal

derived cell lines makes it reasonable to assume that in vivo kinetics would not differ

extensively. The replication of all herpes viruses follows a progressive cascade of

increasing complexity divided into 3 (general) categories. The first viral genes to be

expressed are the Immediate Early (IE/a) genes. These IE genes mobilize transcriptional

machinery and prime the cell for further viral gene expression. The second wave of viral

genes expressed is the Early (E/P) genes. The E genes are specifically involved in viral

DNA replication. Finally, Late (L/y) gene expression provides structural proteins for

virion assembly and egress. Each gene class is discussed in detail in the next sections.

Immediate Early Genes

HSV-1 encodes five IE genes, all of which are located in or around the repeat

regions of the genome. The fact that these important genes lie very near (and in one case

anti-sense to) the major locus responsible for reactivation from latency (the latency

associated transcript, discussed later) emphasizes the likelihood of this region playing a

significant regulatory role. The IE genes are classically identified as capable of being

expressed in the absence of prior viral protein synthesis. Ribosomal inhibitors such as

cycloheximide will therefore prevent expression of all but these five viral gene products.

An interesting point is that while initiation of the acute cascade in infected neurons

requires IE genes for priming, the very first events are actually directed by L gene

products carried into the nucleus from the tegument. Two tegument proteins, virion host-

shutoff (VHS) protein and virulence protein-16 (VP16), play a crucial role in the early

events of acute replication. The VHS is retained in the cytoplasm where it begins the

process of shutting down host transcription to allow the virus access to cellular resources,

primarily by disrupting polyribosome complexes and degrading viral and cellular RNAs.

The VP16 enters the nucleus along with the genome and forms a trimeric complex with

cellular proteins Oct-1 and host cell factor (HCF). This complex interacts with specific

(TAATGARAT) elements present in the promoters of all five IE genes, activating their

transcription and initiating acute replication. A more thorough explanation of VHS and

VP16 is given in the discussion of L genes.

Because of their importance in initiating the replication cascade, IE genes tend to

possess relatively extensive core promoters that contain more regulatory elements than

their E and L counterparts. Of all HSV-1 acute phase promoters, those of IE genes are

most similar to cellular RNA pol II promoters. The unifying factor common to all five

promoters is the presence of multiple TAATGARAT motifs; the most conserved portion

of an extended consensus IE enhancer upstream of the core promoters. This motif is

related to the cellular promoter octamer motif (ATGCAAAT), which is specific for POU

family transcription factors.

The IE genes are regulated at both transcriptional and posttranscriptional levels,

and perform distinct functions in the newly infected cell. Only 2 of the 5 IE genes are

essential for viral replication in cultured cells, though mutation of nonessential genes may

significantly attenuate viral replication and virulence. A brief description of these genes

and their proposed functions is given in (Table 1.2).

Since the IE genes are key regulators of both acute replication and reactivation

from latency, and are the focus of much of the research discussed herein, significant

detail is provided regarding their structure and function.

Infected cell polypeptide-4 (ICP4)

ICP4 is a multifunctional phosphoprotein essential for both viral replication9 and E

gene transactivation. The inability of ICP4 (-) viruses to replicate renders them incapable

of reactivating from latency. The 350 kilodalton (kDa), dimeric, DNA-binding protein9'10

is found primarily in the nucleus shortly after infection11,12 and is considered HSV-1 's

major regulatory factor. Depending on the promoter, ICP4 can act as either a repressor or

activator of viral gene expression. In general, ICP4 increases the transcription rate of all

viral gene classes, but actually represses its own promoter12-14. Transcriptional auto-

regulation is (most likely) controlled by an ICP4 binding site within its promoter that

spans the transcriptional start site1, inhibits transcription factor complex formation or

blocks access of upstream activators from Spl binding sites16. Transient assays have

shown that the repressive effects of ICP4 on its promoter are dominant over VP16

activation14, and out compete Spl for binding during viral infections.

In addition to auto-regulation, the ICP4 promoter is positively regulated by both

viral and cellular proteins17. Numerous cellular promoter-like elements exist, including 2

cis-acting TAATGARATs, at least 4 Spl sites, a CAAT box, and a TATA box7,17. The

ICP4-directed activation of gene expression occurs by enhancing recruitment of TFIID

(the TBP subunit responsible for TATA recognition) to the promoter, an early step in pre-

initiation complex (PIC) formation"8' Most E gene promoters as well as the LAT

promoter (described later) contain ICP4 binding sites. It is widely believed that ICP4

simultaneously turns IE genes off and turns E genes on to facilitate a smooth transition

between kinetic classes.

Infected cell polypeptide-0 (ICPO)

The ICPO is a multifunctional, 775 amino acid, nuclear-phosphoprotein that plays

an important role in stimulating both lytic replication and reactivation. Regulatory

elements within the ICPO promoter indicate an ability to respond to both viral and cellular

factors. In addition to multiple overlapping composite TAATGARAT motifs for

activation by VP16/HCF/Oct-1 complexes, the promoter contains a cyclic-AMP response

element (CRE) site. The CREs are common cellular promoter elements that serve as a

major target for signal transduction pathways. Though ICPO is described as dispensable

for replication in cell culture, this is only true for high-dose infections, which somehow

compensate for its absence. The ICPO is essential for replication after a low multiplicity

of infection (M.O.I), typically in the range of 0.01 to 0.001 plaque-forming units (pfu)

per cell.

The ICPO protein serves two distinct functions: to reorganize the transcriptional

machinery of (quiescent) neuronal cells, such that viral gene expression is favored; and to

act as a general enhancer of all viral gene classes. Like the tegument protein VHS, ICPO

plays a critical role in re-aligning the transcriptional status of newly infected neurons to

favor lytic gene expression. Immunofluorescence analysis of recently infected neurons

reveals a punctate localization of ICPO protein within the nucleus. This distribution is

due to ICPO's affinity for distinct nuclear substructures, called ND10 bodies (also called

promyelocytic oncogenic domains or PML bodies), via a specific localizing element.

These substructures, associated with the nuclear matrix and important for cellular gene

regulation and transcription, are believed to be the site where IE gene transcription

occurs. Proteosome-dependent degradation of ubiquitinylated proteins plays a key role in

many cellular processes; and the presence of ICPO at ND10s has been shown to co-

localize with enhanced levels ofpolyubiquitin. Ubiquitination requires an ubiquitin

activating enzyme (El), an ubiquitin conjugating enzyme (E2), and frequently a

substrate-specific ubiquitin-protein ligase (E3). Being one of only five intron-containing

viral messages, ICPO's second and third exons encode distinct domains that can serve as

an E3 ubiquitin ligase. Exon 2 contains a zinc-binding RING finger domain (common to

ubiquitin ligases), while exon 3 contains a separate ubiquitin ligase domain; both of

which cause increased levels of conjugated ubiquitin at ND10 substructures for their

degradation. Disrupting ICPO's RING finger destroys its ability to associate with

conjugated ubiquitin and therefore proteosomes. Indeed, a human neuronal

teratocarcinoma cell line (NT2) with inherently low levels ofND Os is not readily

infected by HSV-1. Proteosome recruitment to ND10 substructures results in their

disruption; and is believed to permit global increases in transcriptional activity by

relieving transcriptional repression. The ICPO has also been reported to deliver cyclin-

D3 to NDlOs for degradation, possibly resetting the cell cycle to favor viral gene


The second major influence attributed to ICPO is the ability to function as a

promiscuous transactivator of all gene classes, including cellular genes and genes of other

viruses. The ICPO acts at the transcriptional level; and unlike ICP4, does not require

specific promoter elements to exert its influence. Though ICPO actually binds DNA

inefficiently, it is capable of activating any promoter that exhibits basal level activity

without requiring any particular cis-acting element. The ICPO can, therefore, positively

influence cellular-like promoters of IE genes (along with the more rudimentary E and L

gene promoters, including TATA-less promoters9). It has been proposed that ICPO's

ability to induce IE promoters, considered with ICP4's ability to inhibit them, mediates

the early "switch" by which individual neurons follow lytic or latent pathways14. Since

ICPO does not transactivate E genes as readily as ICP4 does, it seems likely that ICPO

provides a boost to the productive cascade while ICP4 controls its temporal progression.

The nuclear reorganizing and transactivating functions of ICPO are clearly critical

for acute replication and reactivation. Even though it is possible to overcome the

replication restriction of an ICPO (-) mutant by infecting at a high M.O.I., the mutation

has a much greater impact on the ability of the virus to reactivate. Since reactivation is

(somewhat) analogous to a very low M.O.I. infection, it is not surprising that ICPO

mutants are deficient in this respect.

Infected cell polypeptide-27 (ICP27)

As shown in Table 1.2, the ICP27 gene is one of the two IE genes essential for viral

replication. The ICP27 gene acts as both a positive and negative post-transcriptional

regulator; and like ICP4, represses IE genes. One mechanism by which HSV-1 ensures

that its mRNAs receive priority over cellular transcripts is via ICP27's ability to prevent

intron-containing mRNAs from being exported from the nucleus. Of HSV-l's 75-80

genes, only five contain introns and three of these are IEs (Figure 1.4). While ICP27

positively regulates viral genes by promoting 3'end maturation and nucleocytoplasmic

transport, it also seems to be involved in down-regulating IE genes when necessary. It is

likely that the activity of ICP4 and ICP27 is crucial for sharpening the IE-to-E gene

transition once the stage has been properly set for full-scale (E) gene expression.

Early Genes

Early gene synthesis begins once IE genes have sufficiently mobilized the cellular

transcriptional machinery. The primary role of E genes is to provide viral replication

machinery; and provide enzymes for increasing deoxyribonucleotide (dNTP) pools and

repairing newly synthesized genomes. Once sufficient levels of these essential

components are produced, E genes are shutoff and L gene transcription begins. Since E

gene synthesis terminates at the start of replication, their classic definition is sensitivity to

translation inhibitors without being affected by inhibitors of replication, such as

phosphonoacetic acid (PAA).

All E gene promoters have been shown to contain a set of core elements that

include both a CAAT and TATA box20. As a rule, E gene promoters are simpler than IE

gene promoters but more complex than L gene promoters. For example, while an IE

promoter may contain many Spl sites, an E promoter might contain one or two Spl's,

while L promoters have zero.

A number of E genes are dispensable for replication in cell culture, most likely

because cellular homologues are capable of serving as substitutes. In fact, roughly half of

HSV-1 's genes are not required for replication in dividing cells7. The absence of non-

essential genes may still have a profound effect on pathogenesis, cell tropism, immune

evasion, and replication in specific cells, especially (non-dividing) neurons. Included

among the non-essential E genes are those encoding enzymes that provide DNA

metabolism precursor genes such as thymidine kinase (TK), ribonucleotide reductase,

uracil-DNA glycosylase, and dUTPase (Table 1.3).

Only seven E genes are absolutely necessary and sufficient for viral replication in

cell culture (Table 1.4). Replication can ensue once levels of these seven proteins reach

sufficient levels. Replication proceeds via a rolling circle mechanism21 from three

separate origins of replication (Ori), one in the UL and two in the Us, to produce

concatameric copies of the viral genome which must resolved into individual linear

molecules for packaging.

Late Genes

With the replication machinery in place and viral genomes being amplified, the

third and final gene class begins to be expressed. Due to their dependence on the

initiation of replication, L genes are classified by their sensitivity to replication inhibitors.

While it is convenient to place all HSV-1 genes into one of three categories, the

delineation between E and L genes is not as clear-cut as it is between IE and E genes. Of

the nearly 30 L gene products, some are detectable at low levels prior to the onset DNA

replication but do no reach maximum levels until post-replication. These genes are

therefore referred to as "leaky-late" (py) genes, while L genes that do not exhibit such

leakiness are termed "strict-late" (y) genes. Since the previously diffuse pattern of viral

proteins (i.e. ICP4 and ICP8) in the nucleus switches to a punctate pattern upon initiation

of replication, it has been proposed that replication occurs within discrete nuclear

compartments22'23 and that E and L gene transcription occur in distinct environments.

This has been proposed to be involved in regulating the switch from E to L gene


As mentioned previously, HSV-1 promoters tend to become progressively less

complicated over the course of lytic replication. A leaky-late promoter, such as the one

controlling the major structural component VP5, contains an Spl site, TATA box, and

initiator element (INR); while a typical strict-late promoter may consist of only a TATA

box and an INR or a downstream activating sequence (DAS). L gene products

predominantly encode structural components involved in capsid assembly and packaging

of viral DNA. Capsids assemble in the nucleus and bud through the nuclear membrane,

thereby, acquiring an envelope while incorporating tegument proteins (such as VHS and

VP16) critical for re-initiating the acute cascade in a new cell. Progeny may be released

as free viral particles but are most often transmitted by direct cell-to-cell contact.

Virulence protein-16 (VP16)

VP16 is a tegument phosphoprotein that forms complexes with the cellular

transcription factor Oct-1 and HCF upon entering the nucleus to activate IE genes24 at

TAATGARAT and GCGCGAA (Spl) motifs. While the POU family member

transcription factor Oct-1 provides the actual DNA binding, VP16 provides an acidic

domain that enhances PIC formation25. Unlike ICP4, which recruits TFIID, VP16 acts at

a much later step by enhancing the binding of general transcription factors (GTFs) TFIIB

and/or TFIIA26'27. While many similarities exist between lytic and reactivation cascades,

an important difference is that during reactivation VP16 is not present to enhance IE

promoters. The manner in which IE genes are activated in the absence of VP16

represents a fundamental difference between lytic replication and reactivation that

remains to be addressed.

Virion host shut-off (VHS) protein

As mentioned earlier, VHS remains in the cytoplasm during the early stages of the

primary infection, disrupting polysomes and rapidly degrading pre-existing mRNAs.

VHS also degrades HSV-1 mRNAs, most likely to hasten the transition between kinetic

classes of genes. Interestingly, at late times during infection VHS is directly bound by

VP16 to prevent it from unnecessarily degrading viral mRNAs. In the absence of VP16 it

has been demonstrated that VHS will ultimately terminate viral protein synthesis.


While lytic replication of HSV-1 requires an intricately organized cascade of highly

specialized genes, the true evolutionary marvel of this virus is most evident when it does

(almost) nothing at all. By far, the majority of HSV- 's existence is as a latent episome;

even during the exception to this rule, very few of the thousands of latent molecules are

ever called to action. Since HSV-1 has only a small chance of infecting a new host

during any short time interval 28, the ability to maintain latent genomes provides a

reservoir for the eventual transmission to other individuals. Latency is often described as

consisting of three separate phases: 1) establishment of the latent infection, 2)

maintenance of latency, and 3) reactivation from latency.

While there are many differences between HSV-1 's lytic and latent cycles, latent

genomes are typified by two distinguishing characteristics. First is circularization of the

linear genome, which likely prevents degradation due to availability of free DNA ends.

The absence of linear genomes in latent tissue was first reported by Southern and

restriction enzyme analysis showing that terminal sequences could not be detected and

terminal fragments were an uncharacteristic size, respectively 29,30. This was followed by

the observation that junction specific probes detected terminal fragments but not junction

fragments 31. The second characteristic unique to latency is a profound restriction of viral

gene expression. The only exception to this state of complete quiescence is transcription

from a distinct region of the R L. The transcript produced during latency is termed the

latency-associated transcript (LAT). LAT does not belong to any of the lytic gene classes

and is the focus of considerable research. For organizational purposes, LAT and its

effects on all aspects of the viruses' biology will be discussed later.

A central problem to studying latency is the lack of a reliable cell culture model.

This requires all latent analysis to be performed using animal models, which re-create

many (but not all) of the aspects of this state. Later sections discuss different animal

models of latency, including their attributes and shortcomings.


As mentioned earlier, upon entering the neuronal nucleus viral genomes follow one

of two fates, either initiating acute replication or entering latency. The "decision" to

enter latency is considered by some to result from a failure of the acute cascade. Others

propose that the genome copy number (high vs. low copy) entering an individual neuron

dictates the path followed. The complexity of neuronal cell types within ganglia makes it

very possible that the cell directs the process, with specific populations of neurons

permissive for lytic replication and others more prone to harbor latent genomes.

Establishment of the latent infection is a completely passive event; no gene product

is involved, and even replication defective viruses cause latent infection when inoculated

at peripheral sites. To date, no viral mutant(s) have been constructed which are unable to

establish latency. Further indication that a very early decision determines which path is

followed is evident in latent ganglia cross-sections from early times post-infection.

Infecting mice (eyes or footpads) with an HSV-1 recombinant containing an ectopic LAT

promoter driving expression of the P-galactosidase gene (LAT-LacZ) followed by in situ

hybridization for LacZ (latent population) or acute viral antigen (replicating population)

shows that latent neurons appear as rapidly as productively infected neurons, and that the

populations do not overlap 32. While latency may represent a failure of the productive

cascade, it is unlikely that neurons that support replication contribute to the latent

population; replication is actually believed to result in cell lysis and death. Since

significant impairment to the host is rare, it seems unlikely that extensive loss of neurons

occurs during this phase. By releasing progeny virus, neurons that die may ultimately

enhance establishment by re-seeding the periphery (anterograde fast axonal transport) to

increase exposure of axonal termini so that other neurons become infected. The inability

of replication defective viruses to establish levels of latent genomes equivalent to wild

type virus is probably due to less than maximal exposure of axonal termini at the

periphery. The importance of extensive exposure to nerve termini for efficient

establishment is underscored by the fact that HSV-1 does not spread directly form

neuronal soma to adjacent neurons or support cells gliall cells), which are relatively

resistant to HSV-1 infection 33. Sensory ganglia lack inter-neuronal connections, thus,

the ability to migrate to other ganglia is due to connections of sensory and motor neurons

of the PNS to the spinal cord.


Since levels of replication defective and LAT (-) mutants are not diluted over time,

maintenance of latent genomes for the life of an infected individual is (like

establishment) considered a passive event. Rather than promoting large-scale

transcriptional activity and cellular re-organization, genomes entering latency are best

served by maintaining as low a profile as necessary. A major factor contributing to the

stability of latent HSV-1 is the tightly wound and extensively chromatinized state of the

episome 34. Genomes are so well repressed that even though an average latent neuron

contains 10-100 viral copies 29,31,35 they are undetectable by in situ hybridization. To

date, the only method for detecting latent HSV-1 from tissue sections is by in situ PCR.

A second example of the repressed state that occurs during latency is the rapid shut

down of viral reporter constructs. A striking illustration of the tight repression of latent

HSV-1 is that heterologous reporters such as the cytomegalovirus (CMV) IE or cellular

phosphoglycerine kinase (PGK) inserted into the genome are quickly repressed as the

virus enters latency. Since unmodified cellular transcription systems have been shown to

be capable of activating all kinetic classes of HSV-1 promoters, the absence of

transcription from all but the LAT promoter clearly indicates that the status of the latent

genome as a transcriptional template is extensively modified. Current progress in our lab

has demonstrated that specific histone modifications, rather than DNA-methylation,

direct this repression.


HSV-1 faces one problem from the strategy that allows existence in a non-dividing,

immuno-privileged site, for up to 80 years: how does such a highly repressed molecule

covertly re-enter the lytic cycle, especially without the help of tegument proteins to

potentiate the IE genes? Reactivation is the process by which latent genomes re-initiate

productive cycle transcription in response to a variety of cellular signals. Numerous

sporadic stress-mediated events in the host have been attributed to HSV-1 reactivation,

such as, stress, exposure to UV light, menses, lactation, malnutrition, fatigue, anxiety,

and immunosuppression. Infected individuals are capable of undergoing repeated bouts

of reactivation throughout their lives, with each event representing only a minor fraction

of the total latent population. Even in the most dramatic reactivation model, explant co-

cultivation of latent murine DRG, less than one percent of latent genomes reactivate.

Anterograde fast axonal transport along the original axon(s) involved in the primary

infection deliver viral progeny to the initial site of infection, potentially causing a

renewed lesion. One reason HSV-1 is so common within the general population (>90%

of people are seropositive) is that most reactivations are sub-clinical; an individual may

not be aware they are even infected, let alone undergoing recrudescence. In fact, only

approximately 10% of infected individuals ever develop recurring oral-facial lesions.

The severity and duration of lesions due to episodic reactivations tend to progressively

diminish and exhibit shorter recovery times; most likely due to a primed immune

response's ability to rapidly bring the infection under control. Though a given

reactivation event occurs from only a small percentage of the total latent population, their

relatively frequent occurrence over an individual's life without causing any obvious

neurological defects indicates that neurons survive the process. Indeed, it seems illogical

for the virus to both reduce its latent pool and draw undue attention by causing cell


Role of Neurons in Latency

HSV-1, like all alphaherpesviruses, is neurotropic with neurons serving as the

nominal site of latency. HSV-1 has an overall affinity for the entire nervous system,

through which it will readily migrate if presented the opportunity. Human autopsy

analysis has shown HSV-1 to be readily detectable in a significant percentage of human

brain samples. Interestingly, HSV-2 is also detectable, albeit less frequently, in human


The nervous system is comprised of two basic components: 1) the central nervous

system (CNS), consisting of the brain and spinal cord, and 2) the peripheral nervous

system (PNS), consisting of neurons and their projections (axons and dendrites). The

role of the PNS is to link peripheral stimuli to the CNS. Two types of neurons comprise

the PNS: afferent neurons, and efferent neurons. Afferent (sensory) neurons carry

information towards the CNS, linking sensory receptors within the body to processing

circuits in the CNS. Efferent (motor) neurons carry information from the CNS back to

the periphery in response to sensory input. There are two distinct classes of efferentt)

motor neurons: 1) somatic neurons connect the CNS and skeletal muscles, and therefore

control voluntary movement, 2) autonomic neurons innervate smooth muscle, cardiac

muscle, and glands, and therefore control involuntary processes. A ganglion (ganglia

plural) is a local accumulation of neuronal cell bodies and support cells in the PNS.

Nerves tracts extending from the ganglia to the periphery are simply bundles of axons

from individual neurons gathered together. While motor neuron ganglia reside within the

spinal cord, sensory neuron ganglia are found on both sides of the spinal cord in distinct

out-pockets called lacunae. Sensory ganglia are located at distinct intervals along the

entire length of the spinal cord and are named in accordance with the vertebral sections

they innervate. These include the trigeminal ganglia, thoracic ganglia, lumbosacral

ganglia, and dorsal root ganglia.

The primary sites of latency in humans are the sensory neurons of the trigeminal

ganglia. While somatic ganglia can harbor latent virus, they simply are not as readily

infected as the trigeminal ganglia. Trigeminal ganglia have a complex architecture

divided into three separate regions, due in large part to their formation by the fusion of

two nerve branches (ophthalmic and maxillomandibular) during embryonic

development*. Since these neurons supply distinct oral-facial regions (ophthalmic,

maxillary, and mandibular) 36 it should not be too surprising that primary infection of any

oropharynx area can result in latency within the trigeminal ganglia.

A typical murine ganglion has been shown to consist of approximately 20,000

cells, of which 18,000 are support cells and 2,000 are neurons. Following inoculation,

approximately one-third of these neurons will become latently infected, harboring an

average of 10-100 viral genomes per cell 37. In situ PCR analysis has indicated that only

10-30% of neurons harboring latent virus produce detectable amounts of the LAT. Since

LAT is often used as a marker for latent infection, many researchers describe latent

neurons as falling into two classes: LAT (+) or LAT (-).

Due to the vast array of functions ascribed to a given ganglion, neuronal

populations should not be expected to be homogeneous. Considering the many types

(and sub-types) of neurons and the extensive co-speciation that has occurred between

HSV-1 and its host, it seems reasonable for the virus to have acquired tropisms for

specific neuronal subsets. Though the lack of a cell culture model for latency makes

these issues exceedingly difficult to address, certain neurons have been shown to be more

permissive for acute replication while others tend towards latency. In situ analysis of the

ophthalmic division of murine trigeminal ganglia reveals that of four classes of neurons

examined (Substance P+, CGRP+, LD2+ and SSEA-3+), the SSEA-3 population

represents the major latent group. Stage-specific embryonic antigen-3 (SSEA-3) is the

marker for a major population of neurons and is present on 40% of mature trigeminal

ganglia neurons in the ophthalmic division.

Octamer binding proteins

A major determinant of a neurons' response to the presence of viral genomes is

dictated by its individual transcription factor profile. Neurons have been shown to

express different isoforms of POU family transcription factor octamer-binding proteins

(Oct). This is the transcription factor pivotal in forming complexes with VP16 and HCF

during lytic replication to activate IE genes. Octamer binding proteins, like Oct-1 and

Oct-2.1, bind a variety of DNA elements (including TAATGARATs) to facilitate

transcriptional activation. Neurons are capable of expressing a variety of Oct-isoforms

that may render them more or less susceptible to acute replication and reactivation.

While Oct-1 and Oct-2.1 are the predominant isoforms and have been shown to increase

transcriptional activity at TAATGARAT elements, they are not the major isoform of all

neurons. Some neurons express Oct-2.4 and Oct-2.5 isoforms that lack the strong C-

terminal activation domain (AD) of Oct-1 and Oct-2.1. Neurons expressing these Oct-

variants have been shown to actually repress IE genes. Co-transfection studies have

shown that Oct-2.4 and Oct-2.5 actually repress VP16 transactivation, while BHK cells

expressing Oct-2.4 or Oct-2.5 are much less permissive for lytic replication than control

BHK cells*. It is possible that these factors determine which fate the virus chooses

following entry, and therefore which population serves as the reservoir for reactivation.

Neurons have also been shown to be capable of differentially regulating IE

promoters in the absence of viral proteins. Construction of transgenic mice containing

viral promoters (ICPO, ICP4, ICP27, or gC) controlling expression of the LacZ gene

demonstrate that patterns of P-gal expression vary based on the class of promoter,

neuronal phenotype, and even between the same neurons at different ages 38,39. While the

ICP4 promoter is inactive in many neuronal and non-neuronal tissues, it is very active in

the trigeminal ganglia, actually 100-fold more so in newborns than adults. The ICPO

promoter is also negative in non-neuronal cells, but is active in adult rather than newborn

trigeminal ganglia; while the gC (L) promoter is negative in both neuronal and non-

neuronal tissues. These results indicate the capacity of latent virus to respond to cellular

signals, further emphasizing the importance of maintaining a highly repressive state to

prevent untimely reactivations. It also raises the possibility that neuronal factors require

only a brief release from HSV-l's repressive state to activate IE promoters; independent

of viral accessory factors such as VP16.

Latency Associated Transcript (LAT)

The 1987 discovery of a promoter, not involved in lytic replication, capable of

subverting global transcriptional repression40 led many to believe the secrets of latency

and reactivation would soon be revealed. Sixteen years, and numerous phenotypic

descriptions later, the exact function of LAT remains unresolved.

LAT was originally identified by in situ analysis of latent tissue showing positive

signal only within neuronal nuclei, and only with RL region probes. This result was in

sharp contrast to acutely infected ganglia in which probes for any genomic region provide

cytoplasmic and nuclear signals from neuronal and non-neuronal cells. The source of the

latent hybridization signal was mapped to a nuclear-localized 8.5-kb primary transcript,

containing a 5' cap, intron, poly-A tail, and several putative open reading frames (ORFs)

which is partially anti-sense to RL genes ICPO and y34.5. Further examination reveals

that the primary LAT does not encode a protein and has a very short half-life, in fact full-

length poly-A+ message is nearly impossible to isolate from latent tissue. The unstable

primary transcript undergoes autocatalytic splicing to produce a highly stable 2.0-kb

intron. The introns' stability (half-life of approximately 24 hours) explains why in situ

analysis yields stronger positive signals with probes to this region, compared to LAT

probes 5' and 3' of the intron. Since the 2.0 kb intron is readily detectable in latent tissue

it is often used as a marker of latent neurons, and is termed the "major" LAT. In certain

neurons the major LAT undergoes a second splicing event to produce 1.45 to 1.5-kb

introns termed the "minor" LAT. The validity of using the major LAT as a marker of

latency is not without problems since, as mentioned previously, only 10-30% of latent

neurons produce detectable levels of the 2.0-kb intron41-43

In vitro transcription of the primary LAT demonstrates that autocatalytic splicing

results in a poly-A (-) molecule44. S1 nuclease and RNase protection assays of latent

tissue have precisely mapped the 5' and 3' ends of this RNA to canonical splice signals,

while asymmetric PCR analysis demonstrates that the intron exists as an uncapped

lariat45. The presence of a non-consensus splice-acceptor site results in an atypically

branched intron lariat which is not recognized by the cellular de-branching machinery,

thereby accounting for its stability46(Figure 1.5).

The size and complexity of the LAT promoter further emphasizes the likelihood

that this region is involved in sensing and responding to cellular stimuli. LAT's 202-bp

core promoter is unique compared to even the most complex IE promoters. Numerous

canonical RNA pol II like elements render this promoter more typical of cellular rather

than HSV-1 promoters. The location of LAT core promoter elements are shown below

(Figure 1.6), though it is important to note that factors 5' and 3' of this region have been

proposed to also affect its activity47-49

While the LAT promoter exhibits its greatest activity in neurons; transient and

reporter assays, both in vivo and in cell culture, indicate basal activity in most cell types.

Considering its basal activity and presence of an ICP4 binding site, the fact that the LAT

promoter does not behave like an IE promoter further emphasizes the likelihood of other

factors being involved in promoter regulation during latency.


Three observations have led to the general consensus that LAT's primary function

is in reactivation from latency, rather than establishment or maintenance. First, LAT (-)

mutants establish latent infections equivalent to, and as readily as, wild type virus.

Second, the number of latent genomes in neurons infected with LAT (-) mutants does not

progressively diminish over time; eliminating the likelihood that LAT perpetuates

maintenance. Finally, multiple animal models demonstrate that LAT (-) mutants exhibit

severely restricted reactivation phenotypes compared to their parental strains; making it

very likely that LAT primarily functions during reactivation.

It should be made clear that this is a general statement intended to provide a

unifying theme that most herpesvirus researchers agree with. A number of caveats exist

regarding every aspect of LAT's "known" role in reactivation. Conclusions vary based

on viral strain, animal model, and even animal strain, not to mention variability between

research groups. This section concludes with an alternative explanation (to which we do

not subscribe) for LAT's role in latency and reactivation.

Mice and rabbits comprise the two major systems for studying HSV-1 latency and

reactivation; while a brief description is given below, they are explained in greater detail

later. While murine (eye and footpad) models are the most cost efficient, they do not

mimic human reactivation as accurately as rabbits (ocular) models, which exhibit both

spontaneous and induced clinical reactivation. Animal models have allowed the portion

of LAT responsible for reactivation to be mapped to an 800-bp region termed the

reactivation critical region (RCR)50-54. The RCR includes the 202bp LAT core promoter

and entire (603-bp) 5'exon. The fact that reactivation maps upstream of the intron is

fortuitous since it allows mutational analysis to exclude the anti-sense ICPO transcript

extending approximately half way through the intron. The RCR and the individual

constructs involved in mapping the reactivation phenotype are shown below (Figure 1.7).

Note that even the most severely restricted viral mutants exhibit basal levels of

reactivation, indicating that LAT is not the only factor involved.

Cyclic-AMP response element binding (CREB) proteins

The complexity, unique transcriptional status, and involvement in reactivation

render the LAT promoter a prime candidate for being the site through which the cell and

virus communicate stress-stimuli. Of the numerous elements within the LAT promoter,

the most appealing from a regulatory standpoint are its two CRE sites. Located within

the core promoter defined by the 17APst construct, eight-bp CRE motifs serve as the

target of cyclic-AMP response element binding (CREB) proteins, a major member of the

POU (CREM/CREB/ATF) family of transcription factors. CREB is a ubiquitous

transactivator present in the nucleus in an inactive form. A variety of signal transduction

cascades set in motion a chain of events (Table 1.5) leading to increased intracellular

concentrations of the second messenger cyclic-AMP, and ultimately CREB activation.

One such pathway initiates when adrenergic molecules (such as the hormone

epinephrine) bind P-adrenergic receptors on the cell surface. CREB's conformational

change (described in Table 1.5) results from phosphorylation occurs at a "P box" within a

basic-domain leucine zipper (bZIP) that allows CREB homodimers to efficiently bind

CRE sites. CREB's N-terminal region is glutamine rich and provides strong interaction

with the basal transcriptional apparatus55'56, specifically CREB binding protein (CBP)

and p300. Activated CREB homodimers can also interact with AP-1 family (bZIP)

transcription factors such as Jun and Fos.

The cyclic-AMP response element modulator (CREM) gene contains numerous

introns, allowing alternate splicing to produce a variety of CREM/CREB/ATF isoforms.

The ubiquitous nature of these factors allows specific and rapid response to external

stimuli. It is, therefore, equally important that these transcription factors are able to be

quickly down regulated or shut-off. Specific de-phosphorylation of active CREB

homodimers is one way to halt its activity. A second interesting mechanism by which

down regulation could occur is through the activation of a specific CREB/ATF/CREM

inhibitor. This CREB regulator results from alternate splicing of the CREM gene to

include an intronic promoter (P2) not present in other isoforms. Unlike other family

members that are pre-made and awaiting activation, this protein is inducible and is

therefore termed the inducible cyclic-AMP early repressor (ICER). The repressive

capacity of ICER derives from the lack of an activation domain (AD) present in its

ubiquitous/activator relatives. ICER forms inactive heterodimers with CREM family

members to either prevent formation of active homodimers or physically obstruct

promoter CRE sites. Since ICER and CREB both require Ras-dependent signaling

pathways, the time required to express ICER may allow CREB a window during which to

activate specific promoters before ICER is present at high enough levels to reverse the

process. CRE sites within ICER's promoter allow for auto-regulation, indicating that it

may turn itself off once CREB has been inactivated.

As previously mentioned, the presence of CRE sites within the LAT promoter

suggests that cyclic-AMP modulates LAT promoter activity, and thus, stress-induced

reactivation. Mutating the -43 CRE to a non-consensus binding site results in a partial

reduction in reactivation from the rabbit eye model57. Mutation of the other (-83) CRE

both individually and in tandem with the -43 CRE was the focus of my research for a

number of years. While initial results indicated very interesting reactivation and

virulence phenotypes for these constructs, factors outside my control unfortunately put

this project on hold. Other researchers have, none-the-less, provided evidence of cyclic-

AMP modulation of the LAT promoter. Cell culture analysis shows the LAT promoter to

be responsive to cyclic-AMP*; and cyclic-AMP antagonists (such as propranolol) inhibit

reactivation of latent marine ganglia following explant co-cultivation and heat-shock8.

It is important to note the similarity between HSV-l's ability to respond to a

variety of environmental stress-stimuli, and the CREM family of transcription factors

ability to serve as the final point of action for a variety of signal transduction pathways.

Alternate Model for LAT

Herpesvirus researchers do not universally accept the model that LAT acts

primarily during reactivation. While we may not fully subscribe to these observations

and proposed mechanisms, the work is not without merit and certainly deserves mention.

The model system employed for much of this work comes from the labs of Nancy

Sawtell and Rick Thompson (University of Cincinnati). Briefly, murine ocular infections

allow establishment of latency in the trigeminal ganglia. Reactivation is induced by

hyperthermic stress (3-minute swim in a 430C H20 bath, 24-hours prior to sacrifice),

followed by single-cell PCR of dissociated neurons. This PCR process has been coined

"contextual analysis of DNA" (CXA-D). LAT (-) mutants analyzed by this method

exhibit a decreased ability to reactivate similar to previously described. This restriction,

however, is said to result from measurably reduced (2-fold) numbers of latent genomes in

LAT (-) infected ganglia5961. Their results indicate that when infecting with wild type

virus, a greater inoculating dose results in greater numbers of latent genomes, allowing

more frequent reactivations. The LAT (-) viruses reduced frequency of reactivation is,

therefore, simply the result of a smaller latent reservoir. Our own results indicate that a

replication-competent virus reaches a level of saturation in terms of the number of latent

genomes, regardless of infecting dose.

The model that LAT's primary role is in establishment of latency is taken further

by the observation that LAT (-) mutants exhibit increased neurovirulence; presumably

due to the absence of an anti-apoptotic function that protects infected neurons. This

effect is manifest in "rare" neurons that harbor thousands of viral genomes rather than the

typical 10-100. Since this high M.O.I. neuronal population is proposed to represent the

primary source for reactivation, the loss of these neurons removes the most relevant latent

population62. It has also been proposed that high M.O.I. neurons result from both lytic

and latent phases occurring simultaneously, and LAT's job is to repress the productive

cycle63. Separate groups have reported that LAT (-) mutants exhibit 70% more apoptotic

neurons than their wild type counterparts64, and over-expression of LAT in transformed

cell culture inhibits IE gene expression65

We agree that dissociating establishment and reactivation is difficult, and efficient

establishment is a pre-requisite for reactivation. Our analyses have, however, been

unable to detect significant differences in levels of latent genomes based on either dose or

status of the LAT region. Animal models exhibit inherent fluctuations that render 2-fold

differences statistically insignificant. We also fail to observe increased neurovirulence

with LAT (-) mutants and find it difficult to believe that large-scale loss of neurons

would not produce a readily observable defect in our animals.

Animal Models

The broad cell and tissue tropism of HSV-1 permits its study in a variety of animal

models. As might be expected of a virus that has undergone co-speciation with its host

for millennia, finding a model system capable of replicating all aspects of HSV-1 's

biology is simply impractical. Potentially interesting viral strains are sometimes avoided

due to extreme neurovirulence and pathogenesis in non-native hosts; analogous to the

high mortality rate seen in humans infected with simian B virus, an innocuous

herpesvirus of primates. Strain-specific glycoprotein alterations can profoundly affect

viral dissemination and latency in specific models, leading to very different conclusions

about the viruses' biology. Identical strains of virus, likewise, may replicate and spread

very differently depending on the animal model used, as well as the animal strain and

age. Despite their drawbacks, the lack of reliable cell culture models of latency

emphasizes the value of animal models in our current understanding of in vivo

herpesvirus biology. Our laboratory focuses exclusively on the two the most prevalent

animal models for studying latency and reactivation: mice and rabbits.

Murine Models

Due to their cost-effectiveness, relative ease of handling, and wide availability of

reagents, the mouse represents the most common species for herpesvirus research. The

very fact that HSV-1 establishes latency in neuronal ganglia was first demonstrated in the

murine system4o. The most practical means of infecting mice is on their rear footpads,

allowing the virus access to the major axons projecting to this area; those of the dorsal

root ganglia (DRG). Pre-treatment of the footpad with an injection of 10% saline four

hours prior to infection allows the epidermis to be efficiently removed with an emery

board just prior to infection. While under sedation, this large, moist, surface allows the

virus to enter and replicate within basal epithelial cells to provide efficient and uniform

access to axonal termini. This in turn results in a reliable and homogeneous population of

latently infected DRGs. Failure to saline pre-treat footpads results in a much less

uniform, and 100-fold less efficient, infection. Footpad infections with virulent strains of

virus will often present clear evidence of CNS involvement. Within five to seven days

post-infection (d.p.i.) some mice appear lethargic, develop scruffy coats, exhibit hind

limb paralysis, and may eventually die. Individuals that recover from the primary

infection appear physiologically normal and are typically allowed at least 28 days for the

acute infection to completely subside. DRGs can then be removed for analysis of latent

genomes. One of the major benefits ofHSV-1's ability to traffic through the PNS and

CNS is that it allows the effect of specific mutations to be examined for alterations in

pathogenesis, neuroinvasiveness, and neurovirulence.

Reactivation from latent murine DRG is only possible by a process termed explant

co-cultivation, in which ganglia are removed from the animal and maintained in media in

a CO2 incubator. Dissection from the animal and propagation in media is clearly a

stressful event that may not fully reflect more subtle signals involved in human

reactivations. Though correlation with human reactivation may not be perfect,

transcription of acute phase genes is evident within four hours post-explantation, viral

replication is detected within approximately 12 hours, and infectious virus may appear by

three days (using wt or LAT (-) virus). Explant co-cultivation of LAT (-) mutants (strain

17syn+) from both trigeminal and dorsal root ganglia will reactivate to produce viral

progeny with slightly (but measurable) delayed kinetics compared to their wild type

parents. A less virulent HSV-1 strain (KOS-M), however, demonstrates no difference in

kinetics between LAT (-) and wild type66, whether this is due to strain or investigator

differences is unclear. Unfortunately, only molecular reactivation can be examined in

murine systems, neither clinical reactivation (detection of virus at periphery) nor

spontaneous reactivation occurs.

An alternative to footpad infections performed in our laboratory, are ocular

infections that allow the virus to establish latency in the trigeminal ganglia. While

neurons of the trigeminal ganglia provide a more accurate representation of the latent site

in humans, infections are inefficient and less uniform than the footpad model. Latent

murine trigeminal ganglia also use the explant co-cultivation model for reactivation.

Some researchers induce reactivation prior to dissection by hyperthermic stress 24 hours

before sacrifice. Priming reactivation in this way supposedly allows more efficient and

synchronous reactivation.

Rabbit Models

Animal models of HSV-1 reactivation should re-create three main criteria to be

considered analogous to the human condition: 1) the ability to occur spontaneously, 2)

the ability to respond to stress, and 3) allow complete recovery. While rabbits are

sometimes avoided due to difficulty in handling and expense, they meet all of these

requirements. Inoculation of scarified rabbit corneas with HSV-1 results in peripheral

replication that allows the virus access to axons of trigeminal ganglia neurons. Latent

rabbit trigeminal ganglia can be analyzed at the molecular level following dissection, as

in the murine system, but provide the added benefit of exhibiting both spontaneous and

induced clinical reactivation. After allowing 21 days for acute replication to subside, tear

film swabs can detect the presence of infectious virus due to spontaneous reactivations67.

Efficient reactivation can also be induced by direct iontophoresis of P-adrenergic

reagents (such as epinephrine) into the cornea68-72. This results in a high frequency of

virus in the tear ducts within 16 to 96 hours post-induction. Iontophoresed rabbits exhibit

no adverse effects from this procedure and recover completely. The ability to recover

virus from the TG in this manner provides valuable insight into the course of human


Due to the more subtle nature of reactivation stimuli between rabbit and murine

systems (iontophoresis vs. explant co-cultivation) the rabbit eye model represents a more

discrete reactivation event, similar to the small percentage of neurons responsible for

human reactivation. Large scale DNA replication is, therefore, not evident at post-

induction time points (48 hours) as with murine co-cultivations. LAT (-) mutants are

much more severely restricted in their reactivation frequency and levels of productive

cycle transcripts than in murine systems73.

Cell Culture Models

Numerous references have been made throughout this text regarding the lack of

viable (or widely accepted) cell culture models of latency and reactivation. As with

many of the basic tenants of HSV-l's biology, this convention is not without caveats.

Most attempts to establish latent infections in cell culture have been met with partial

success; generally more closely mimicking a smoldering/persistent type of infection

rather than latency. Genomes, for example, tend not to: form episodes, transcribe LAT,

or remain "latent" without the assistance of replication inhibitors.

There is a cell culture system that does not suffer from these restraints. Christine

Wilcox (University of Colorado) has developed a primary cell culture system using DRG

from 15-day old Sprague-Dawley rats. Briefly, ganglia are dissociated and cultured in

5% newborn bovine serum and 2.5S nerve growth factor (NGF). The presence of

fluorodeoxyuridine inhibits non-neuronal (dividing) cells, allowing for a relatively pure

(1-3x105 cells) neuronal culture. This primary culture is maintained for two weeks before

infecting with HSV-1. Twenty-four hours post-infection the cells are treated with

acyclovir, followed by inoculation with an M.O.I. of 0.01 pfu/cell. The infection

proceeds for seven days in the presence of acyclovir, after which time the acyclovir can

be removed without the virus undergoing lytic replication. The virus not only remains

"latent" at this point, but it genomes are circular and actively transcribe LAT.

Interestingly, all neurons harboring viral genomes in this system accumulate LAT; there

is no LAT (-) population as seen in vivo. The homogeneous, LAT (+), latent virus

population can then be induced to reactivate following a number of stimuli; including

NGF-withdrawal or addition of the membrane permeable cyclic-AMP analog forskolin.

An interesting observation of the post-induction reactivation process is a direct

correlation between decreased LAT accumulations concomitant with elevated

intracellular levels of the CREB antagonist ICER, mentioned previously and discussed in

more detail in chapter two.


Figure 1.1. Schematic of HSV-1 virion.

R Rl Rs U Rs

Figure 1.2. Organization of linear HSV-1 genome.

Repeat Short Repeat Long

Unique Short

Repeat Short

Repeat Long

HSV-1 Episome
152260 bp

Unique Long

Figure 1.3. Latent HSV-1 episome.




Figure 1.4. Genomic location of Immediate Early genes and LAT.

0 5'exon

Branch Point-- OH
I 3'exon

0 5'exon


3' exon OH

o 5'exon 3'exon +

S2.0kb intron

Figure 1.5. Splicing of primary LAT to produce a stable 2.0kb intron lariat. A) The 3'-
OH group of a non-consensus branch point (guanine) attacks the 5'
exon/intron border, B) The free 3'-OH of the 5' exon attacks the 3'
exon/intron border to connect exons and C) The intron remains as a lariat (5'-
2' linkage), non-consensus branch point inhibits debranching.



Figure 1.6. Schematic representation of core LAT promoter elements.

RI. R1R Rs Rs
U' U i

RCR ^"Latency Associated Transcript

Promoter 5' Exon Intron

A202bp A348bp A480bp
.l A313bp A489bp

Figure 1.7_ A969bp

Figure 1.7. Viral mutants used to map reactivation critical region (RCR).

Table 1.1. The Eight Human Herpes Viruses


and Host Cell Range
Rapid replication, broad
host/cell range

Slow replication, restricted
host/cell range

Site of Latency

Glandular or
lymphatic cells

Table 1.2. The five immediate early (IE) genes of HSV-1
IE Gene Replication Location and Modifications Proposed Function
ICP4 Essential Nucleus Master switch for
(phosphorylated) coordinated gene expression
ICP27 Essential Nucleus Inhibits export of spliced
(phosphorylated) mRNA
ICPO Dispensable Nucleus Promiscuous transactivator
(phosphorylated) Restructures cellular
(spliced) transcriptional machinery
ICP22 Dispensable Nucleus Maintains broad host-cell
(phosphorylated) range (possibly)
ICP47 Dispensable Cytoplasm Interferes with antigen
(spliced) presentation by MHC-1

Virus and Associated Disease
HSV-1 (Herpes simplex type-1)
Oral-facial infection
HSV-2 (Herpes simplex type-2)
Genital infection
VZV (Varicella-zoster virus)
Chicken pox/Shingles
CMV (Cytomegalovirus)
Infectious mononucleosis
HHV-6 (Humanherpesvirus-6)
Mild childhood disease
HHV-7 (Humanherpesvirus-7)
Mild childhood disease

EBV (Epstein Barr Virus)
Mononucleosis and cancer
HHV-8 Kaposi Sarcoma

Slow replication, restricted T or B cells
host/cell range

Table 1.3. Non-essential early genes
Non-essential E Gene (examples) Function
Thymidine Kinase (UL23) Deoxypyrimidine kinase far greater substrate range
than its cellular counterpart. Phosphorylates a wide
variety of nucleoside analogues (responsible for
HSV-ls susceptibility to acyclovir).
Ribonucelotide Reductase (UL6) Reduces ribonucleotides to deoxyribonucleotides to
create dNTP substrates for replication.
Uracil-DNA glycosylase (UL2) DNA repair and proof-reading. Corrects insertion of
dUTP and deamination of cytosine residues
(important due to HSV-ls high GC content).

dUTPase (UL50)

Hydrolyzes dUTP to dUMP. Prevents dUTP
incorporation and provides dUMP pool for
conversion to dTMP by thymidylate synthetase.

Table 1.4. Seven essential early genes
Essential Early Gene Number of Genes
DNA Polymerase (UL30) 1
DNA Binding Proteins 2
Origin Binding Protein 1
Components of Helicase/Primase Complex 3

Table 1.5. CREB signal transduction pathway
Order Signal
1 Epinephrine (adrenaline), tyrosine derived "fight for flight" hormone. Made
from adrenal glands of CNS and neurons of CNS and PNS.
2 Binds G-coupled receptors (p-adrenergic receptor) on plasma membrane. 7-
transmembrane receptor, associated with GTP binding proteins on inner surface.
3 RAS (GTP binding protein), composed of a, p and y subunits. A-subunit binds
GDP when inactive. Receptor activation causes a-subunit to bind GTP, a can no
longer bind P and y subunits. Free P and y subunits are effector molecules.
4 Effector molecules activate the enzyme adenyl cyclase. Catalyzes the formation
of 3'5'-cyclic AMP from ATP.
5 Second messenger (cyclic-AMP) diffuses through the cytosol to activate protein
kinase A (PKA).
6 PKA (2 inhibitory and 2 catalytic subunits) releases catalytic subunits (kinase A)
upon activation. They migrate to nucleus tophosphorylate specific targets.
7 CREB contains a single kinase A recognition site (Serine 133), phosphorylation
causes a conformational change that efficient binding of CRE sites.



HSV-1 establishes latency in neurons of sensory ganglia innervating the site of

initial infection. The virus can reactivate spontaneously, or under conditions of stress, to

cause a recurrent infection. During latency the genome forms an episome in neuronal

nuclei from which no viral replication occurs. Approximately 1/3 of the latent infected

neurons express high levels of a single transcript, termed the latency-associated transcript

(LAT). This transcript is important for reactivation, despite the fact that LAT does not

encode a protein.

While LAT is clearly required for efficient reactivation in animal models, its

mechanism is not well understood. One factor that complicates these analyses is that

observations vary depending on the animal model and viral strain used. In the rabbit eye

model, for example, latency is established in trigeminal ganglia by corneal inoculation,

with reactivation (either spontaneous or induced by iontophoresis of epinephrine) scored

by recovery of infectious virus in the tear film. In the mouse model, latency is

established in the trigeminal or dorsal root ganglia by inoculation corneas or rear

footpads, respectively. Viral reactivation from ganglia can be induced by thermal stress

(as demonstrated by expression of reporter genes from neuronal tissue), or by explant co-

cultivation of dissected ganglia on cultured cells.

Large LAT deletion mutants exhibit reduced numbers of latent viral genomes in

neurons of both mice and rabbits. Smaller LAT deletion mutants (such as the promoter

mutant 17APst or 17A348, a 348 bp 5' exon deletion) do not, however, demonstrate

significant differences in total numbers of latent HSV-1 genomes.

It is possible that our previous studies, which use relatively large inoculating doses

(1xl05 pfu), may mask subtle replication or establishment deficits inherent to these LAT

mutants. The course of the acute infection in the rabbit eye model was, therefore,

examined using a 1000-fold range of 17APst (and its rescue) inoculation doses.

Differences in acute infection kinetics and establishment of latency were not detected by

this method. The observation that peak establishment occurs with even low dose

inoculums suggests that saturation of latent sites occurs relatively early. To determine

the contribution of the initial inoculum to establishment, rabbits were infected with a non-

replicating HSV-1 recombinant. While this recombinant is capable of establishing

latency, the total number of latent genomes is much lower than wild type virus, indicating

that peripheral replication contributes to maximal establishment of latency.

Materials and Methods

Cells and Viruses

Virus was propagated on cultured rabbit skin (RS) cells. Virus titers were

determined on Vero cells grown in minimal essential medium supplemented with 5%

fetal calf serum and antibiotics. The following HSV type-1 (HSV-1) genotypes

previously described were used in these experiments: Wild-type strain 17syn+, 17APst (a

203-bp portion of the LAT promoter deleted), 17A348, (a LAT mutant with bases

119,007-119355 deleted) and its rescuant (17A348R), RHA-6 (a construct expressing the

5' portion of LAT by virtue of nucleotides 120,290-120,467 removed and replaced with a

442-bp fragment of the SV40 encoding the cleavage/polyadenylation signal site).

Infections and Reactivation

Unscarified rabbit eyes are inoculated with the indicated number ofpfu in 25 tl

aliquots. Animals are sacrificed between 1 and 7 days post-infection for acute studies,

and their corneas and ganglia harvested. Latently infected trigeminal ganglia are

recovered from rabbits at least 40 days post-infection.

DNA Extraction

Dissected corneas or ganglia are incubated with 0.6 ml of extraction buffer (25 mM

EDTA, 100 mM NaCI, 1% SDS, 10mM Tris pH 7.5) and 50 gl ofproteinase K solution

(15 mg/ml) overnight at 48 C. DNA is extracted three times with phenol-chloroform

(1:1) and once with chloroform. DNA is precipitated with ethanol overnight and pelleted

by centrifugation. The pellet was washed once with 70% ethanol, air-dried and dissolved

in 200 l of water.

Analysis of the Relative Amounts of Viral DNA by PCR

Semi-quantitative PCR analysis incorporating 32P a-dCTP is able to detect 1 pg of

purified HSV-1 DNA compared to a control plasmid containing a sub-cloned fragment of

the VP5 gene. When purified viral DNA is mixed with uninfected ganglia it is possible

to detect less than 1000 viral genomes. This PCR method is also able to detect the viral

DNA and cDNA- from a single infected cell. Actin primer sets are used to amplify DNA

corresponding to cellular genomes to normalize product intensities. The ratio of both

signals is determined by densitometry.

Amplification by PCR was carried out as previously described, using the following

primer sets (sense/anti-sense strand): VP5 (149-bp product), 5'-


rabbit actin (110-bp product) 5'-AAGATCTGGCACCACACCTT-3'/5'-

CGAACATGATCTGGGTCATC-3'. The reactions were carried out in M.J. Research

thermal cycler as follows: denaturation, 94 C for 30 sec; annealing, 55 oC for 30 sec;

and extension, 60 s at 72 C. The final cycle was terminated with a 10-min extension

step. The products were made radioactive for autoradiography and image quantitation by

addition of 0.2 pC of a -32P-dCTP. For each reaction we used 20 pl (10%) of the DNA

sample, and the final volume of the reaction was 100 Pl. One fifth of the amplified

product (corresponding to 2% of the original material) was fractionated on 6%

polyacrylamide gels in Tris-borate-EDTA. The PCR signals were visualized by scanning

an appropriately exposed autoradiogram by use of a Deskcan II scanner (Hewlett-

Packard). The signals were quantified by densitometry using IP Lab Gel software (Signal

Analysis Corporation) in accordance with operational instructions.

PCR Analysis to Determine Relative Levels of Latent Viral DNA and Wild Type

PCR primers specific for the cellular actin gene serve as an internal standard for

normalizing levels of latent viral DNA between samples. Actin primers (sense 5' AAG


TC 3') yield a 110-bp product. PCR primers specific for the HSV-1 polymerase gene are

used to detect both wild type and KD6 viral DNA. Polymerase primers (sense: 5' CAT


TGT A 3') yield a 92-bp product. PCR primers specific for the HSV-1 ICP4 gene are

used to confirm that levels of establishment are not due to wild type revertants when

analyzing the KD6 viral recombinant. ICP4 primers (sense: 5' CTG ATC ACG CGG


3') yield a 144-bp product. PCR reactions are performed in a 50pl final volume,

consisting of 40.5 tl sterile H20, 1 pl of both forward and reverse primers (600 ng/tl), 1 ll

dNTPs (1.25 mM each), 5pl 10X AS buffer (Qiagen; Tris-C1, KC1, (NH4)2SO4, 15mM

MgCl2, pH 8.7), 1 ll respective DNA/cDNA sample, and 0.55tl HotStar Taq DNA

Polymerase (Qiagen; 5 U/pl). The amplification profile consists of 15 minutes at 950C to

activate the Taq, followed by one three-minute cycle of 940C, 550C, and 720C; this is

followed by 30 identical cycles of one minute each (Ericomp TwinblockTMSystem, Easy

Cycler). PCR products are resolved on 5% polyacrylamide gels, stained with SYBR

Green (Molecular Probes), and scanned with a Storm Phosphorimager (Molecular

Dynamics) using a 450 nm wavelength laser. Relative levels of latent genomes are

determined by establishing the ratio of HSV-1 polymerase product to cellular actin levels

within each sample. This accounts for fluctuations in signal intensity due to inconsistent

sample processing. Viral polymerase specific PCR products are compared to a plasmid

titration containing the sub-cloned target sequence spiked into processed uninfected

rabbit trigeminal ganglia tissue. The signal intensity of each sample is compared to this

titration to determine the relative number of latent HSV-1 molecules in each sample.


Acute Replication in Rabbit Corneas and Trigeminal Ganglia Following High Titer

The contribution of both LAT expression and inoculation dose is analyzed over the

course of acute ocular infection of rabbits with either 500 or 500,000 pfu / eye of 17APst

or 17APstR (rescue). Infectious virus yields during the acute infection were measured in

tear swabs, corneas, and trigeminal ganglia (Figure 2.1-2.3). At high viral doses, recovery

is greatest from tears and corneas on the first day post-infection (d.p.i.). These levels

tend to reach a lower stage plateau by days 2 through 8, after which continual decreases

result in undetectable levels of virus by day 14. Viral titers from trigeminal ganglia

increase during the first 3 days of infection, followed by 3 days (d.p.i. 4-6) of maximal

virus titers, and finally a steady decrease. While dose clearly affects infection kinetics,

no significant effect is observed relative to LAT status. PCR analysis to determine the

relative amount of viral DNA present in corneas and trigeminal ganglia following high

dose (5x105 pflu) infection also fail to detect significant differences based on LAT

genotype at all time points (Table 2.1).

Analysis of Viral DNA Levels in Corneas and Trigeminal Ganglia During the Acute

The course of infection is then examined following a much lower dose infection

(500 plaque forming units per eye). As with high titer infections, relative amounts of

HSV-1 DNA in corneas are greater than trigeminal ganglia during the entire acute

infection course. A variety of LAT mutants, in addition to the prototypical LAT (-) virus

(17APst), that differ in LAT expression and reactivation phenotypes are included in this

analysis. Mutant 17A348 expresses LAT, but exhibits a significant reactivation

impairment following epinephrine induction. Mutant RHA-6, which contains an SV40

cleavage/polyadenylation sequence in the middle of the 2.0 kb LAT intron, expresses

LAT and reactivates normally.

Rabbits inoculated with 500 pfu of reactivation-impaired viral recombinants

(17A348 and 17APst) demonstrate significantly decreased levels of virus in trigeminal

ganglia during the acute phase of infection, as compared to wild type and RHA-6 viruses

(Table 2.2). At day five post-infection, the average value for the low reactivation

mutants (0.35 0.19) is a marginally significant difference (p-0.068, t--test) compared to

normal reactivators (0.56 +0.38). Average values at day 7 (0.29 0.18 for low

reactivating viruses, and 0.63 0.31 for normal reactivating viruses) are again

significantly different (p=0.006), but by the time latency is established (21 days) all

corneal infections are statistically indistinguishable using any construct.

The Relative Amount of Latent Viral DNA in Trigeminal Ganglia of Rabbits
Infected With Wild Type or LAT Mutants are Similar Regardless of Infecting Dose

Levels of viral DNA in ganglia following clearance of the acute infection suggests

that viral DNA levels in the ganglia are independent of LAT genotype and infecting dose.

This is confirmed using semi-quantitative PCR to compare relative amounts of latent

viral DNA over a range of infecting doses (Figure 2.4 and Table 2.3). Rabbits corneas

inoculated with 500 to 50,000 pfu / eye are sacrificed 30 days post infection to determine

levels of latent HSV-1 (Table 2.4). No statistical difference in amount of viral genomes

is detected as a function of either LAT genotype or initial virus dose. As with high titer

infections, neither dose nor LAT genotype effect DNA levels in latently infected

trigeminal ganglia.

A Non-replicating HSV-1 Recombinant Establishes a Latent Infection in the
Trigeminal Ganglia, but at Lower Levels Than Wild-Type Virus

To assess the contribution of input inoculum on establishment of latency, a non-

replicating (ICP4 (-)) HSV-1 recombinant (KD6) is used. The amount of HSV-1 DNA in

trigeminal ganglia is determined by PCR from rabbits inoculated with xl 05 or x106 pfu

of this virus at 14 d.p.i. (Figure 2.5). While trigeminal ganglia of rabbits inoculated with

KD6 contain detectable HSV genomes, overall levels are lower than those observed using

replication-competent LAT (-) or LAT (+) viruses. PCR analysis of these ganglia (using

primers specific for the ICP4 gene) indicates that the DNA present is not due ICP4

revertants. These results demonstrate that while non-replicating HSV-1 recombinants

can establish a latent infection, replication is required to achieve wild type levels of



It has been previously suggested that LAT plays a role in protecting neurons from

death or apoptosis during the initial stages of establishment. These observations have

been made with deletion mutants that extend from the entire LAT promoter into the 2.0

kb intron, and often display altered virulence. While we have never observed such

effects with the 202 bp LAT promoter mutant (17APst), the statistical power required for

discerning 4-fold (or less) establishment or virulence defects are difficult to achieve in

the rabbit model. The goal of this study was to determine if subtle deficits in replication

or establishment are detectable using 10 to 1000-fold lower than normal inoculi of

17APst in the rabbit eye model. Significant differences in the amount of infectious virus

produced during the acute infection in cornea and ganglia, or on the level of latent

genomes in trigeminal ganglia of the rabbit, are not detected. These results suggest that

the primary defect of 17APst in the rabbit eye model occurs at the level of reactivation.

While our levels of sensitivity cannot rule out the possibility that 17APst affects quality

of establishment, we were unable to detect any differences during the acute infection

even when allowing additional replication cycles to occur.

This study provides the additional opportunity to monitor the course of an ocular

infection as a function of dose. Not surprisingly, peak acute titers in the tears, cornea and

trigeminal ganglia are delayed by several days when lower inoculi are used.

Interestingly, peak levels of viral DNA in the trigeminal ganglia are reached slightly

earlier, suggesting that maximum establishment occurs fairly early, and at relatively low

inoculation doses. This in turn suggests that corneas provide a limited number of entry

sites into the nervous system (or number of available neurons), which become saturated

relatively quickly. To address this question more directly a non-replicating virus (KD6)

is used. Since this virus cannot undergo additional rounds of replication in the cornea, it

allows assessment of the amount of viral DNA delivered to the trigeminal ganglia as a

direct function of input. Results indicate that while significant establishment of latency is

achieved, even doses of 1x106 pfu yield approximately 10-fold lower than wild type

levels of establishment. This indicates that while a non-replicating virus can establish

latency in rabbit eyes, replication is required to establish maximal latent infections. This

requirement is likely due to mechanical barriers that must be overcome to efficiently gain

access to the nerve termini projecting to the trigeminal ganglia. While infecting the

corneal surface (even with scarification) provides access to many nerve termini,

replication and cell-to-cell spread are much more important factors.

A final conclusion of this study is that lower (and probably more physiologically

relevant) doses of viruses are sufficient to efficiently establish latency in the rabbit

trigeminal ganglia. It is interesting to note that increasing inoculum does not decrease the

scatter in total levels of establishment over a range of doses. This adds additional support


to the notion that efficient establishment of latency requires a significant cellular

component, which may be regulated at the individual ganglia level.

^ 5

1 2 3 5 7 14
Post Inoculation Day

Figure 2.1. Viral titers recovered from eye swabs.

.-- -- APst(500)
-- -APst-r (500)
SA--*--apst (5xlop)

2f 4 --- ----y- ^l't--,
--- '
o 3

1 2 3 5 7 14
Post Inoculation Day

Figure 2.2. Viral titers recovered from corneas.

A-- APft(50D)
--S-- APs-r (500)
-*-- Aft (5x W)
M t-r(Uxld)

1 2 3 5 7 14
Post Inoculation Day

Figure 2.3. Viral titers recovered from

500 ofu

500 Dfu

trigeminal ganglia.

500 ofu

17APst 17APstR
500 ofu 500 Dfu

17APst 17APstR
500 ofu 500 ofu

HSV Polymerase genomic equivalents

Figure 2.4. Levels of establishment using high and low does 17APst (and Rescue).








500 pfu

V Pol


Left and Right Trigeminal Ganglia
(Polymerase DNA)
KD6 (IxlOs pfu) KD6 (1x10' pfu)
IL IR 2L 2R 3L 3R IL IR 2L 2R 3L 3R


Left and Right Trigeminal Ganglia
KD6 (lxlO pfu)
IL 1R 2L 2R 3L 3R +
I- vim SM\

17I(1x105pfu) 17(lxi0'pfu)
IL IR 2L 2R 3L 3R 4L 4R 5L 5R IL IR 2L 2R ML 3R 4L 4R 5L 5R

|**as*** **|t**

Figure 2.5. Reduced establishment using KD6 versus 17syn+.


Table 2.1. PCR
Gene Target



Rabbit actin

Primer pair

Table 2.2. Relative amount of viral DNA (VP5/actin) at high dose of inoculation
(500,000 pfu)*
DPI Cornea Ganglia
17APst 17APstR 17APst 17APstR
Mean SEM Mean + SEM Mean SEM Mean SEM
1 1.510.54 1.19+0.99 0.120.12 0.230.19
2 2.290.76 1.400.94 0.650.26 0.370.28
3 2.110.32 2.380.59 1.100.26 1.860.66
5 2.310.64 1.590.18 1.800.36 1.740.39
7 2.161.30 2.010.27 0.800.20 0.540.40
14 0.440.14 0.360.34 0.430.30 0.210.23
*Rabbits eyes were inoculated with 500,000 PFU of 17APst or 17APstR (rescue). At the
indicated times post infection (DPI), the rabbits (2 rabbits per virus per time point) were
sacrificed and corneas (4 per virus per time point) and TG (4 per virus per time point)
dissected. Total DNA was isolated from the tissue and amplified with VP5 and actin
primer sets in combination. The relative amounts of viral DNA (VP5/Actin) were
determined by densitometry.

Product size (bp)




Table 2.3. Relative amounts of viral DNA in cornea and trigeminal ganglia during acute
infections post low dose inoculation with viruses of different LAT genotypes*
Virus Days p.i. Cornea Mean+SEM Ganglia Mean+SEM
17 syn+ 1 0.21+0.12 0.030.01
2 0.82+0.55 0.030.02
3 0.880.46 0.070.04
5 0.79+0.87 0.490.48
7 1.420.49 0.500.31
21 0.220.09 0.220.13
17APst 1 0.27+0.16 0.030.30
2 0.370.24 0.080.30
3 0.510.36 0.050.30
5 1.440.56 0.390.30
7 0.830.79 0.160.30
21 0.300.23 0.250.31
17A348 1 0.400.28 0.04+0.04
2 0.230.22 0.030.04
3 0.360.27 0.030.04
5 0.800.55 0.310.04
7 0.830.70 0.210.21
21 0.230.12 0.280.21
17A348R 1 0.300.33 0.030.01
2 0.590.47 0.030.01
3 0.92+0.67 0.17+0.35
5 1.83+0.69 0.610.42
7 1.831.45 0.740.70
21 0.220.10 0.220.01
RHA-6 1 0.100.13 0.040.03
2 0.070.09 0.03+0.02
3 0.670.35 0.07+0.04
5 1.20+ 0.35 0.460.34
7 0.7910.65 0.570.34
21 0.150.11 0.330.15
*Rabbit eyes were inoculated with 500 PFU of 17syn+, 17APst, 17A348, 17A348R and
RHA-6. At the indicated days post infection, cornea and TG (4 each per virus per time
point) were dissected and the relative amounts of viral DNA determined.
Relative amounts of viral DNA presented as the ratio of the HSV VP5 gene to cellular
actin ratio as determined by PCR (see Materials and Methods). Means and standard error
of the mean (SEM) are presented as Least Squares Means values and were calculated as
described in the Materials and Methods.

Table 2.4. Relative amounts of viral DNA present in TG during latency in rabbits
infected with different doses of virus*
Rabbit Tattoo #
Virus/dose (left or right TG) HSV-1 DNA (genome equivalents) Mean (SEM)*
17APst A3(L) 30,000 18,300 7888
500 PFU A3(R) 2,000
A5(L) 40,000
A5(R) 1,200
17APstR A9(L) 800 12,200 7888
(rescue) A9(R) 8,000
500 PFU A10(L) 30,000
A10(R) 10,000
17APst A26(L) 1,200 10,750 7888
50,000 A26(R) 1,800
PFU A30(L) 3,000
A30(R) 11,000
17APstR A31(L) 8,000 16,500 + 7888
(rescue) A31(R) 3,000
50,000 A32(L) 15,000
PFU A32(R) 40,000
*Rabbits were inoculated with the indicated doses of 17APstR or 17APst in both eyes.
Total DNA was isolated from latent ganglia (40 d.p.i.) and analyzed by PCR
amplification with actin and VP5 primer sets. Data are from 4 TGs per dose per virus per
time point.
Relative amounts of viral DNA expressed as genome equivalents of HSV determined
following semi-quantitative PCR for the HSV DNA polymerase gene and standardized to
the amount of cellular actin present in each sample. Standard curves were generated
using known amounts of HSV polymerase target DNA were used in order to calculate the
number of genomes present in each sample (see Materials and Methods).
IMeans and standard error of the mean (SEM) were calculated as described in the
Materials and Methods section.



Examination of the molecular processes involved in HSV-1 latency and

reactivation represent a major portion of our laboratory's resources. Projects consist of

defining chromatin boundaries during latency, identifying specific cellular factors that

signal reactivation, and examining LAT's role as a liaison between the cell and the virus.

A fundamental issue central to accomplishing these goals is to characterize the very early,

if not first, events of the reactivation process. While acute replication kinetics in cell

culture are well established, the events of in vivo reactivation are partially understood at

best. A complete picture of the molecular processes controlling reactivation is not only

interesting in its own right, but can also serve as a template for examining specific viral


Our models of latency and reactivation center on the theory that the factors)

responsible for maintaining the delicate balance between the repressive state of latency

and the ability to sense and respond to subtle cellular stimuli are located within the LAT

region. Despite its unique transcriptional status, location, promoter complexity, and

mutational analysis indicating an involvement in reactivation, a clear demonstration of

how (or if) LAT responds during reactivation has yet to be shown. Transcription of LAT

during latency provides an ideal marker by which to monitor changes in relative

abundance during the early events of explant co-cultivation, to determine if (and indeed

how) LAT responds to reactivation stimuli. Considerable effort optimizing reverse

transcription and PCR conditions now allows sensitive and quantitative analysis of the

early molecular events of explant-induced reactivation. These results will provide

valuable insight towards our reactivation models, specifically regarding the degree to

which LAT is involved and whether it plays an active or passive role.

Explant co-cultivation of latently infected murine DRG is well suited for this

analysis because the stress of explanation permits more extensive transcriptional activity

than subtler reactivation models. Using this model, LAT's abundance is examined over

an early range of time points using wild type virus strains 17syn+ and KOS. A separate

analysis of ganglia from a partially characterized LAT-transgenic mouse is also included

as a means of monitoring LAT in a more homogeneous environment independent of

exogenous viral factors. These results will indicate the extent to which this locus is able

to respond to reactivation stimuli, and also provide insight regarding LAT's location in

the signaling cascade and whether it behaves as an activator or repressor.

The second goal of these experiments is to establish a correlation between changes

in LAT abundance and activation of lytic cycle transcripts during reactivation. Reliably

detecting such a discrete event will be more difficult than monitoring LAT levels. Since

LAT is actively transcribed during latency (up to 30% of latent neurons produce

abundant levels) a readily detectable baseline exists for comparison to later time points.

Lytic genes must, however, reach minimal thresholds before they can be reliably

detected. This sensitivity requirement is further complicated by well-documented basal

lytic gene leakinesss" during latency. Reliable determination that a given transcript is

present at significantly higher levels than background is difficult, especially when the

goal is to monitor the very first events of reactivation.

Our final goal is to determine which of the key IE genes (ICPO or ICP4) is the first

to respond during reactivation. In addition to wild type (17syn+ and KOS) infected

ganglia, viral mutants deleted for either ICPO or ICP4 (KD6) are also be examined. The

experimental design is based on the logic that the first transcript activated during

reactivation will not be effected by the others absence. If, for example, the promiscuous

transactivator ICPO is the key IE gene, then its appearance should not be affected by the

absence of ICP4. Likewise, if the master regulatory gene ICP4 is the major determinant

then it should be detected even in ICPO's absence. In addition to providing a genetic

basis for deciphering the events of reactivation, these mutants provide the additional

benefit of being replication-deficient, thereby reducing the transcriptional leakiness

associated with latency. Besides the sensitivity requirements, a potential drawback

associated with these mutants is that their replication restriction prevents them from

establishing nearly as robust of a latent infection as wild type virus. The possibility exists

that the resolution required for this analysis will be impossible from less than maximal

levels of latent genomes.

In the event that a clear distinction between ICPO and ICP4 activation can not be

established, this work should at least address the issue of whether IE genes are activated

during the early hours of explant co-cultivation. Previous reports claim that reactivation

does not follow the standard IE, E, L, gene kinetics determined in cell culture. These

results state that E transcripts precede IE's by as much as 24 hours. While we feel these

experiments are not nearly as well controlled or optimized as ours, the observation

deserves to be addressed.

Materials and Methods

Cells and Viruses

Wild type HSV-1 strain 17syn+ (syncytia forming) is amplified and titered on rabbit

skin (RS) cells. Briefly, RS cells are maintained in minimal essential media (MEM,

Gibco Life Technologies, Gaithersburg, MD) with 5% calf serum and antibiotics (250 U

penicillin, 250 l.g/ml streptomycin, 2.5 [tg/ml amphotericin B, and 292 Ljg/ml L-

Glutamine) at 370C in a humidified 5% carbon dioxide atmosphere. Sub-confluent RS

cell monolayers are infected with an M.O.I. of 0.01 pfu/cell and allowed to replicate 3-4

days or until 100% cytopathic effect (C.P.E) is visible. Cells are detached from the flasks

by shaking, and the media transferred to 250 ml Sorvall bottles. Virus is concentrated by

centrifugation at 10,000K at 40C for 40 minutes. The virus containing pellet is

resuspended in 1 ml of media, freeze thawed twice, and stored at -800C for titering.

The 17syn+-based ICP4 (-) virus (KD6) has the entire ICP4 coding sequence

removed from both Rs copies and requires the E5 helper cell line for amplification due to

its replication restriction. E5 cells are Vero (African Green Monkey Kidney) cells stably

transfected with the HSV-1 ICP4 gene. Cells are maintained as previously described,

using media supplemented with 10% fetal bovine serum (FBS) rather than 5% calf serum.

Viral infection and harvesting is identical to the method described above. The KD6 virus

is capable of spontaneously reacquiring the ICP4 gene during amplification on E5 cells.

Revertancy to wild type status occurs with a frequency of approximately one in every ten

thousand infectious particles (lx104). To ensure that KD6 viral stocks exhibit as low a

revertant rate as possible, the virus is plaque purified and titered on both E5 and Vero

cells. Comparing viral titers on permissive (E5) and non-permissive (Vero) cells allows

the percentage of wild type virus present in the stock to be determined.

Since a frank deletion of its coding sequence would disrupt the anti-parallel LAT

region, the KOS-based ICPO (-) mutant (provided by Priscilla Schaffer, Harvard

University) contains translational stop codons of all three ORFs. This construct does not

require a helper cell line for amplification due to an ability to overcome its replication

deficit by infecting cells at a high M.O.I. (5 pfu/cell). Since ICPO does not require a

helper cell line, there is no chance of reverting to wild type status. A Vero based helper

cell line is, however, available for titering ICPO (-) stocks since viral plaques cannot arise

from the low M.O.I. infections required (single viral particles) to determine the number

of infectious particles/ml.

Wild type HSV-1 strain KOS is a much less virulent isolate than 17syn but is fully

capable of establishing a latent infection and reactivating. It is included in this analysis to

serve as a second wild type strain and because the ICPO (-) mutant is KOS-based.

Mouse Infections

Four to six week old female Swiss-Webster mice are anesthetized with halothane

and pre-treated with 0.05 ml of 10% saline solution injected under each rear footpad.

Four hours after pre-treatment the mice are anesthetized (ketamine, acepromazine,

xylazine), and the foot's keratinized epithelial layer is lightly abraded with an emery

board. The moist surface is infected with 500 pfu of virus (17syn+ or 17APst) in a 50 l1

volume applied to the entire surface with the side of the pipet tip.

Since wild type revertants of KD6 are fully capable of replication, the presence of

even low-level revertants may overwhelm the replication-defective KD6 virus, making

analysis impossible. This requirement is further complicated by the extremely high dose

inoculation of KD6 required to establish sufficient levels of latent genomes in the DRG.

The replication-deficient status of KD6 requires infecting with the maximal allowable

titer (105-106 pfu/mouse) that contributes the fewest (1-10 pfu/mouse) wild type

revertants. To ensure that KD6 infected ganglia do not contain significant levels of

revertants, back-extracted DNA is examined with PCR primers for deleted (ICP4) and

non-deleted (i.e. polymerase gene) regions. If ICP4 is detectable at levels equivalent to

the polymerase specific primers, the tissues cannot be used. In general, mice exhibiting

CNS involvement and death following infection strongly indicate the presence of an

unacceptable revertant rate. If mice show none of the symptoms of a wild type infection,

they are most likely infected with KD6 alone.

KOS infections (wild type and ICPO (-) mutant) are both performed at high M.O.I.s

due to their extremely low neurovirulence. Even though both viruses are replication-

competent, the absence of specific viral surface glycoprotein entry mediators renders this

HSV-1 strain avirulent. Since wild type revertants of the ICPO (-) mutant are not a

concern, the ability to infect with high doses allows efficient establishment of latency.

Animals are placed on their backs while the anesthetic wears off (30-45 minutes) to

allow efficient viral adsorption. Mice infected with replication competent viruses may

exhibit lethargy, scruffy coats, and hind limb paralysis associated with CNS involvement

within 5-7 days post-infection. Fifty percent of these mice typically succumb to viral

dissemination; those that survive exhibit complete recovery and are allowed 28 days for

acute replication to subside. Mice are considered latently infected by 28 days post-

infection (d.p.i.) and are suitable for examination of reactivation.

Generation of LAT Transgenic

Transgenic mice (c57b6) were constructed (Fox Chase Cancer Center) using a 3.0

kb sub-cloned fragment of the HSV-1 LAT region extending from the core promoter to

the first 500 bp of the 3'exon. Though not yet fully characterized, transgenic mice

contain an intact promoter and at least the first 2.0 kb of the primary transcript.

Explant Co-cultivation

Mice are sacrificed by cervical dislocation and dorsal root ganglia (DRG) L4 and

L5 are removed under aseptic conditions. Dissected ganglia are immediately placed in

500 tl of media (MEM plus 10% FBS) and maintained at 370C in a humidified CO2

incubator. At specific intervals post-explant (0, 1, 2, 3, or 4 hours), ganglia are transferred

to 1.5 ml tubes and snap frozen in liquid nitrogen (LN2). Since approximately three

minutes are required to remove the DRG from a single mouse, zero hour time points are

removed and frozen individually. Co-cultivation experiments of latently infected DRG

represent the pooled tissue from three mice (approximately 12 ganglia), while the

transgenic mouse experiments require DRG of a single animal per time point (four


RNA Processing

Frozen DRGs representing various co-cultivation time points are homogenized on

ice in 600gl Trizol (Gibco/BRL) using glass dounces. Following complete tissue

disruption the dounce is rinsed with an additional 600ul Trizol, the homogenate is

transferred to 1.5 ml tubes, followed by incubation at room temperature for 5 minutes.

240pl of chloroform is added, followed by 15 seconds of vigorous vortexing, a 15-minute

incubation at room temperature, and centrifugation at 9,000 x g for 15 minutes at 40C.

The aqueous phase (approximately 600pl) is removed to a separate tube and precipitated

with 0.7 volumes isopropanol. While maximal RNA recovery is important, the

likelihood of DNA contamination makes it wise to avoid getting too close to the

interface. RNA is precipitated by centrifugation at 14,000xG for 15 minutes at 4C. The

resulting pellet is rinsed briefly with 70% ethanol, taking care to remove all excess liquid.

Contaminating DNA is removed by resuspending the RNA in 90Pl DEPC-treated water

(Qiagen), 9gl 10X DNase buffer (Ambion), 2[l DNaseI (Ambion), and 1 ll RNasin

(Ambion), followed by incubation in a 370C H20 bath for one hour. The reaction is

stopped by addition of 0.2 volumes of DNase Inactivation Solution (Ambion) for 2-3

minutes, mixing periodically. Following a 15 second centrifugation at 9,000 x g, the

aqueous RNA supernatant is removed to a sterile tube.

Reverse Transcription

Reverse transcription reactions are performed using the Omniscript-RT kit

provided by Qiagen. The range of total RNA for optimal reverse transcription efficiency

using the enzyme provided in this kit is 50ng-2jlg. Since the pooled ganglia from three

mice (12-16 ganglia) yields approximately 15 g of total RNA, RNA aliquots must be

divided into 15-18 separate reactions of 20l for optimal cDNA synthesis. An individual

20[l reverse transcription reaction includes 2pl 10X buffer, 2l1 5mM dNTPs, 2ul

random decamer primers (2.5mM each)(Ambion), 0.25ul RNasin (Ambion), 6.75gl

DEPC H20, and 1 gl (4U/pl) Omniscript-RT. Reverse transcription reactions are

incubated for 1 hour at 370C, followed by heat inactivation at 930C for 5 minutes and

rapid cooling on ice. Individual reactions are combined into a single tube and

precipitated with 0.1 volumes 3M-sodium acetate (NaAc), 0.02 volumes linear

acrylamide, and 2.5 volumes 100% ethanol. Following centrifugation at 14,000 x g for

15 minutes, the pellet is dried and resuspended in 20[1i sterile H20. A minus reverse

transcription control (-RT), to account for DNA carry-over, is essential since very few

HSV-1 genes contain introns for cDNA specific analysis. An approximately 20pl RNA

aliquot not used in the reverse transcription reaction is put through the same regimen as

the other samples to ensure that it is concentrated in a manner equivalent to the cDNA

samples. To avoid wasting reagents, reaction components are replaced with sterile water.

LAT is not nearly as difficult to detect from the transgenic mice since the DRGs

contain many more copies of target DNA than even the most efficiently infected ganglia.

This allows the DRGs from a single mouse to be used per time point; which is important

since they are often in short supply. LAT is abundant enough that only one-third of the

total RNA isolated from each mouse is sufficient for analysis. The remaining sample is

frozen for future use if necessary.

DNA Processing

DNA is purified from the aqueous/organic interface to ensure that tissues are

latently infected (or positive transgenics). The interface is back-extracted with 3

successive volumes (150pl each) of 0.1M Tris (pH 8.0) plus 0.1% Sarkosyl. After each

back-extraction the sample is vortexed 15 seconds, briefly centrifuged, and the aqueous

phase transferred to a fresh tube. The final sample (approximately 450-500Pl) is

incubated with 20mg/ml proteinase-K at 370C overnight. DNA is phenol/sevag extracted

and ethanol precipitated as before. The resulting pellet is resuspended in 10Opl sterile

H20. While recovery of DNA by this method is not as efficient as from a straight DNA

isolation, it is sufficient for this application.

PCR Analysis

Analysis of cDNA (or back-extracted DNA) requires comparing time points within

the experiment as well as results between experiments. Primers to the single copy

cellular APRT gene allow samples to be normalized for efficiency of RNA isolation and

reverse transcription. LAT specific PCR primers are available for the 5'exon, and 2.0 kb

intron. PCR reactions are performed in a 50il final volume, consisting of 40.5gl sterile

H20, 1l of both forward and reverse primers (600 ng/ul), lpl dNTPs (1.25 mM each),

5pl 10X AS buffer (Qiagen; Tris-Cl, KC1, (NH4)2SO4, 15mM MgCl2, pH 8.7), 1 l

respective DNA/cDNA sample, and 0.5pl HotStar Taq DNA Polymerase (Qiagen; 5

U/pl). The amplification profile consists of 15 minutes at 950C to activate the Taq,

followed by one three-minute cycle of 940C, 550C, and 720C; this is followed by 30

identical cycles of one minute each (Ericomp TwinblockTMSystem, Easy Cycler). PCR

products are resolved on 5% polyacrylamide gels, stained with SYBR Green (Molecular

Probes), and scanned with a Storm Phosphorimager (Molecular Dynamics) using a 450

nm wavelength laser. LAT 5'exon primers (sense: 5'CGG CGA CAT CCT CCC CCT

AAG C 3', and anti-sense: 5' GAC AGA CGA ACG AAA CAT TCC G 3') yield a 149-

bp product. LAT intron primers (sense: 5' GAC ACG GAT TGG CTG GTG TAG TGG

G 3', and anti-sense: 5' ACG AGG GAA AAC AAT AAG GGA CGC 3') yield a 102-bp

product. APRT primers (sense: ACT CCA GGG GCT TCC TGT TTG 3', and anti-

sense: ATC CAC AAT GAC CAC TCT CTG 3') yield 373-bp product DNA, and a 186-

bp cDNA, amplification products, respectively. ICPO primers (sense: 5' GGG CGG

GCG GTA CGT AGT CT 3', and anti-sense: 5' GAC GGG CAA TCA GCG GTT CG

3') yield 275-bp DNA and 138-bp cDNA products respectively. ICP4 primers (sense: 5'

CTG ATC ACG CGG CTG CTG TAC ACC 3', and anti-sense: 5' GGT GAT GAA

GGA GCT GCT GTT GCG 3') yield a 144-bp product.

Real-Time PCR Analysis

To achieve higher resolution and more easily quantifiable data, real-time

primer/probe oligos were constructed for the LAT 5'exon as well as cellular APRT and

XIST genes. The "Assays-by-Design" program of Applied Biosystems guarantees

optimized primer/probe sets for any DNA sequence submitted to them. The primer/probe

sequences for the XIST gene are (forward: 5' GCT CTT AAA CTG AGT GGG TGT

TCA 3', reverse: 5' GTA TCA CGC AGA AGC CAT AAT GG 3', probe: 5' FAM-ACG

CGG GCT CTC CA 3'). The primer/probe sequences for APRT are (forward: 5'CTC


3', probe: 5' FAM-CCC CAC ACA CAC CTC 3'). The primer/probe sequences for the

5'LAT exon are (forward: 5' GGC TCC ATC GCC TTT CCT 3', reverse: 5' AAG GGA


(cDNA) and ICP4 specific real-time primers and probes are used for detection of lytic

genes. The ICPO prime/probe sets are (sense: 5' CAC CAC GGA CGA GGA TGA C 3',

anti-sense: 5' GGC GGG CGG TAC GT 3', and probe: 5' FAM-ACC TGG ACG AAG

CAG ACT 3'). The ICP4 primer/probe sets are (sense: 5' GAC GGG CCG CTT CAC

3', anti-sense: 5' GCG ATA GCG CGC GTA GA 3', and probe: 5' FAM- CCG ACG

CGA CCT CC 3'). All reactions are run in triplicate in a 20pl final volume consisting

of: 1 l of sample, 10 Ol of Mastermix (Applied Biosystems: TaqMan Universal PCR

Master Mix, No AmpErase UNG), 1 l1 of Assay Mix (containing primers and probe),

and 8pl of sterile H20. Reactions are run on the Applied Biosystems Model 7000

Taqman machine under the following conditions: one two-minute cycle at 500C, one ten-

minute cycle at 950C, and forty cycles of fifteen seconds at 950C followed by one minute

at 600C.


The first aim of this project was to monitor the primary LAT's relative abundance

over a range of explant co-cultivation time points, to determine the degree to which the

LAT locus responds during the early hours of explant-induced reactivation. The murine

explant co-cultivation model was chosen for a number of reasons. First, mice are

relatively inexpensive and easy to handle. Since this work requires replicate analysis of

multiple time points using different viral strains and mutants, more cumbersome systems

(such as the rabbit eye model) are far too inflexible. Second, mice allow uniform and

efficient establishment of latency. Comparisons between time points and experiments are

impossible if dramatic fluctuations in total numbers of latent genomes occur. Finally,

murine dorsal root ganglia respond well to the stress of explanation, thus providing the

most likely environment for determining if the LAT promoter responds from a

distinguishable percentage of the total latent genome population.

Wild type virus reactivation analyses are based on mice latently infected with

HSV-1 strain 17syn or KOS. Time points (0, .5, 1. 2, 3, and 4 hour) represent the pooled

dorsal root ganglia of three mice (approximately 12 ganglia total). For each time point

mice are sacrificed and their ganglia harvested over no more than a fifteen-minute period,

thus, actual culture times may vary plus or minus seven minutes. The potential for

regulatory events to occur during the three to five minute dissecting process requires

zero-hour ganglia to be individually removed and snap frozen in liquid nitrogen. Pooling

ganglia is intended to increase consistency between time points by averaging out

variations of infection efficiency, while also maximizing the amount of total RNA for


To ensure consistent handling, tissues are processed at the same time. Sample

preparation is divided into DNA and RNA components. The majority of the isolated

RNA (90%) is reverse transcribed into cDNA using random decamer primers, while a

minimal amount (approximately 10%) is set aside as a minus reverse transcription (-RT)

control. DNA is purified by back-extraction of the aqueous/organic interface following

RNA removal. Though done as almost an afterthought this step provides sufficient

material for PCR examination. DNA analysis is included to prove that all tissues

harbored equivalent levels of latent genomes, while -RT controls show that cDNA PCR

products are not due to contaminating DNA. The -RT samples importance is

underscored by the lack, in most cases, of spliced HSV-1 message.

Initial sample analysis was to be performed via quantitative (quantitative-

competitive) PCR. This technique requires extensive sample dilution and competitor-

oligo titrations to ensure PCR reactions reflect a linear range of detection, such that signal

intensity reflects the amount of target. For this reason we switched to the much higher

resolution technique of Real-Time PCR. While data from both methods are shown,

quantitation is based on Real-Time PCR results. Two cellular controls are included to

demonstrate consistent RNA isolation and reverse transcription. The single copy APRT

(adenine phosphoribosyltransferase) gene is a standard cellular control representing an

abundant cytoplasmic mRNA. The second cellular control is the mammalian dosage

compensation gene XIST (X-inactive specific transcript), which, like LAT, is an

alternately spliced, nuclear-localized RNA encoding no known protein that associates

with repressed DNA. Because we are trying to quantitatively assess the relative levels of

LAT over a range of time points, it is important to demonstrate that changes in abundance

are not due to general transcriptional activation or RNA degradation. Cellular controls

from nuclear and cytoplasmic regions should address these concerns.

A Severe Decrease in LAT Abundance Occurs During the First Hours of Explant
Co-cultivation of Latently Infected Dorsal Root Ganglia

Results of Figure 3.1-3.4 demonstrate that cellular APRT and XIST controls exhibit

only minor fluctuations (a maximum of 6-fold) during explant co-cultivation. The results

of multiple experiments demonstrate that, relative to zero hour levels, severe reduction in

LAT abundance occur during explant co-cultivation. Although the timing is variable,

LAT levels decrease 10 to 1000-fold relative to cellular controls. Figure 3.1 represents

Real-Time PCR analysis of DRG latently infected with HSV-1 strain 17syn+. Results

demonstrate a dramatic (500-fold) decrease in LAT abundance at only the one-hour time

point. Cellular controls indicate that the decrease at one hour is LAT specific; rather than

overall RNA degradation during co-cultivation, the total amount of RNA seems to

increase slightly at later time points. Figure 3.2 indicates a more gradual pattern of LAT

reduction, reaching a maximum of 10-fold. Decreased LAT levels are slightly more

dramatic (5-fold) than cellular controls, which also tend to decrease slightly.

Results are similar when DRG latently infected with HSV-1 strain KOS (Figure 3.3

and 2.4) are examined. As shown in Figure 3.3, a dramatic decrease (reaching 1000-fold)

in LAT occurs at the three and four-hour time points. Cellular controls again fluctuate no

greater than 5-fold over the same time points. Figure 3.4 demonstrates steadier and more

gradual LAT reductions of approximately 9-fold, similar to the results of Figure 3.2.

Though LAT kinetics are not as dramatic as in the previous experiment, cellular controls

indicate that this reduction is LAT-specific and not simply the product of general RNA

degradation or promoter repression. Control PCRs (DNA and -RT) determine that

tissues are latently infected with equivalent levels of HSV-1, and RNA is devoid of

contaminating DNA. Standard PCR analysis of 17syn+ experiments (2.1 and 2.2) is

shown in Figure 3.5. Primers specific for both the 5'LAT exon and 2.0 kb intron both

show dramatic decreases in LAT abundance at the previously reported time points.

Control PCRs to detect viral and cellular DNA (HSV-1 polymerase and APRT,

respectively) indicate equivalent infections and sample processing. The -RT control

indicates that DNA was successfully removed prior to cDNA synthesis.

Reduced Abundance of LAT Also Occurs During Co-cultivation of Uninfected
Transgenic Mouse Dorsal Root Ganglia

The variability associated with explant-induced decreases in LAT abundance from

wild type infected DRG led to analysis of LAT-transgenic mice to provide a more

uniform LAT environment. Transgenic mice express sufficient LAT that only a portion

of the RNA (20%) from a single animal (four ganglia) is adequate for monitoring LAT's

abundance. Since these mice are not infected, there is no variation in the amount of LAT

within each tissue. The availability of LAT-transgenic mice provide an independent

system to validate observations from latently infected ganglia, while also providing

insight into LAT's ability to respond to stress in the absence of other viral factors.

Though not yet fully characterized, the inserted transgene is known to contain the portion

of LAT (the core promoter extending into the 2.0 kb intron) to which reactivation has

been mapped. Since these experiments employ both male and female animals, only

APRT controls are valid for this analysis.

Results of Figure3.6-3.9 demonstrate that decreased LAT abundance also occurs

following explant co-cultivation of uninfected LAT-transgenic mouse DRG. The

decreases in LAT levels are, however, as variable as in latent infections. Figure 3.6

demonstrates a dramatic (1000-fold) LAT reduction at 0.5 hours only. Cellular APRT

controls fluctuate only 2-3 fold over these same time points. Results of Figure 3.7 exhibit

decreased LAT-abundance at all time points, reaching a maximum of 1000-fold by four

hours. While cellular APRT controls tend to mirror LAT kinetics, they vary only 4-5

fold. Figure 3.8 displays severely reduced LAT levels (500-fold) at only the three hour

time point, while Figure 3.9 exhibits dramatic decreases in LAT abundance (800-fold) at

both one and four hour time points relative to the cellular control. The -RT controls (not

shown) show the absence of contaminating DNA. Standard PCR analysis using PCR

primers to the 5' LAT exon are again included (Figure 3.10) to demonstrate the LAT

decrease, and to emphasize the value of Real-Time PCR for quantitative analysis. As

before, APRT and -RT controls indicate that samples were handled consistently.

Detection of Immediate Early Transcripts During Explant Co-cultivation

Given the dramatic LAT kinetics observed, we wished to determine whether

decreased LAT levels during co-cultivation correlate with the appearance of immediate

early (IE) genes, as reported from the rat primary DRG cell culture model (Wilcox et al).

Since Real-Time PCR primer / probe oligos provided unsuitable sensitivity, standard

PCR was used to detect the IE gene ICPO with primers spanning the genes first intron.

The ability to detect spliced product allows specific detection of ICPO cDNA, eliminating

possible contributions from both DNA and anti-sense LAT. Select cDNA samples were

also analyzed with PCR primers specific for the IE gene ICP4. Since ICP4 is unspliced,

it is not as easily distinguished from background DNA. Preliminary results using Real-

Time PCR primer and probe sets specific for ICP4 failed to detect ICP4 cDNA at any co-

cultivation time point (not shown). Unfortunately, significant portions of the cDNA

samples were used for the Real-Time analysis. Select remaining samples were re-

examined by standard PCR using ICPO and ICP4 specific primers. As shown in Figure

3.11, examination of cDNA from 17syn4 experiment 2.1 demonstrates the presence of

both ICPO and ICP4 during the early hours of explant co-cultivation.

Does Lytic Gene Activation Require a State of LAT-Repression

The fact that both lytic (ICPO and ICP4) and latent (LAT) phase transcripts exhibit

variable kinetics during explant co-cultivation emphasizes the complex nature of

reactivation, and indicates that numerous factors are involved. Our longstanding model

has been that the LAT locus functions (at the DNA or RNA level) to repress lytic gene

expression during latency until proper reactivation signals) are received. We have

proposed a tentative model to explain the decreased LAT abundance during explant co-

cultivation. If LAT repression is the first step of reactivation, then the variable nature of

its kinetics will make reliable monitoring of lytic gene activation nearly impossible. The

optimal environment for monitoring lytic gene activation during explant-induced

reactivation is not from wild type virus but rather a LAT mutant that provides a pre-set

state of LAT repression. The absence of LAT should, therefore, provide a synchronous

population of latent genomes, which already have the "safety" off. A uniformly LAT-

negative environment should provide much more reliable detection of lytic cycle

transcripts. The LAT promoter mutant 17APst is a 17syn+-based construct in which the

202-bp core promoter (including the TATA box) has been removed. Though this

construct does not make LAT, it establishes latency as readily as its wild type parent.

While 17APst is almost completely null for reactivation in the rabbit eye model, it

reactivates with only a slight delay in kinetics during co-cultivation. Interestingly, this

construct is also known to exhibit greater lytic gene leakiness during latency*, possibly

fitting with our model that it lacks a layer of regulation.

The results of two separate 17APst co-cultivation experiments are shown in Figure

3.12. The first experiment demonstrates the presence of both ICPO and ICP4 at 30

minutes, one, and two hours. The second experiment detects ICPO by one hour and ICP4

by two hours. The second half of Figure 3.12 examines KOS tissues from experiment 2.3

and 2.4 to determine if time points demonstrating dramatic LAT decreases are positive

for ICPO. As before, ICPO's detection is sporadic and only partially corresponds with

decreased LAT abundance.

Which is the First Immediate Early Gene Activated During Reactivation

Having proven the ability to detect IE genes within the first four hours of co-

cultivation, our final goal was determine which IE gene (ICPO or ICP4) responds first

during reactivation. The experimental design employs viruses deleted for each respective

gene (ICPO and ICP4 / KD6). Since the gene that responds first should not be affected by

the others absence, these mutants should allow us to determine the order of their

appearance. While the replication impairment of both viral mutants provides much lower

transcriptional background noise during latency, it also results in many fewer latent

genomes within each dorsal root ganglion. As evident in Figure 3.13, infectious doses as

high as 2x106 pfu/mouse fail to achieve levels of establishment equivalent to wild type

virus (KOS). The fact that the ICPO- mutant seems to have established latency more

efficiently than its KD6 counterpart (despite infecting with half as many infectious

virions) indicates that it is capable of limited replication in vivo. Since detection of ICPO

proved to be less than straightforward in wild type infected tissue, this reduction in total

latent genomes only makes our job more difficult. Indeed, these fears proved valid, as

lytic transcripts were not reliably detected from multiple experiments. Two separate

ICPO (-) co-cultivation experiments spanning 0, 1, 2, 3, and 4-hour time points failed to

detect either ICP4 or ICPO transcriptional activity. Four separate KD6 experiments

spanning similar time courses also failed to produce a consistent result. In some cases

ICPO cDNA signal was detected from KD6 experiments, but it was unclear whether this

was due to wild type revertants. The fact that the KD6 virus fails to accumulate enough

LAT to be detected by Real-Time PCR underscores the disparity in overall levels of

latent genomes. While the ICPO (-) virus does accumulate enough LAT to be detected by

Real-Time PCR, it unfortunately fails to provide any insight regarding acute gene

activation. In summary, the level of sensitivity required to reliably detect acute

transcripts, coupled with less than maximal establishment of latency, prevents this

analysis from being readily accomplished. This result is not wholly surprising

considering the difficulty of detecting IE transcripts from wild type infected tissues. The

failure to detect lytic genes from either mutant may result from the failure of LAT to

respond appropriately, as well as sub-optimal establishment of latency.


Final assessment of the goals outlined at the beginning of this chapter can be

summarized as a collectively partial success. While changes in LAT abundance were

detected during reactivation, they were sporadic; while lytic phase transcripts were

identified, their appearance demonstrated neither reproducible kinetics nor a correlation

with decreased LAT abundance. Arguably the most important result of this project has

been the thorough demonstration of the inherent variability associated with explant co-

cultivations. Rather than viewing the situation as bleak, at a certain level these

inconsistencies should not only be accepted, but also expected. Why should detailed

examinations of animal models behave perfectly when reactivation in the natural host

exhibits extreme variation? Some infected people reactivate frequently over their entire

lives, some sporadically, and others not at all. Regulation of reactivation is clearly a

complicated event, in which LAT's involvement is important but not absolute. In the

rabbit eye model, which most closely reproduces the strict regulation of human

infections, even the prototypical LAT reactivation mutant (17APst) exhibits induced-

reactivation frequencies approaching 10%. The intrinsic variability associated with

animal models, combined with HSV- 's penchant for redundant back-up mechanisms do

not logically lend themselves to refined molecular analyses. My experience has been that

variability compromises reproducibility, and the more carefully controlled and optimized

an experiment is, the more apparent these inconsistencies become.

Decreased LAT Abundance During Explant Co-cultivation

Real-Time PCR results from both wild type infected and LAT-transgenic mice

provide direct evidence that the molecular switch through which cellular signals are

relayed to latent HSV-1 genomes resides within the LAT locus. This argument is based

primarily on two points: 1) decreased LAT abundance represents (within our limits of

sensitivity) the first quantifiable molecular event following the reactivation stimuli, and

2) observations of the LAT-transgenic show this effect to be reproducible in the absence

of other viral factors.

The major concern regarding with our observations of LAT's dramatic decrease in

abundance is the erratic nature of its timing and magnitude. One reason latently infected

DRG were pooled for each time point was to prevent aberrant infection efficiencies from

skewing results. Even in the event that one of the three mice within a given time point

was completely uninfected; this should contribute a mere 3-fold change in LAT

abundance. Standard PCR analysis of back-extracted DNA clearly demonstrates that all

time points harbor roughly equivalent levels of latent genomes. The use of two (very

different) cellular controls further strengthens the argument that individual samples are

more similar than different. Considering the number of processing steps involved in

these experiments, error is certain to be introduced by the investigator. Cellular controls,

however, would have indicated if wholesale mishandling occurred in a specific sample.

While APRT and XIST levels fluctuate to a certain extent, they tend not to do so over

greater than 2-3 fold, indicating that neither RNA degradation nor general promoter

activation occur during explant co-cultivation.

The use of lines to connect time points when graphing each experiment may be

potentially misleading. Though difficult to prove, I do not believe that LAT-levels

necessarily rise and fall hundreds of fold over the four-hour time course. Given LAT's

unreliable nature, tissues of certain time points simply do not respond as dramatically as

others. Whether LAT abundance would eventually decrease if given more time is

impossible to know. Experiments demonstrating less dramatic (10-fold) decreases in

LAT abundance may be due to dilution caused by asynchronous kinetics of individual

ganglia. The most dramatic time points would, likewise, reflect more perfectly timed

LAT kinetics. Finally, if decreased LAT abundance represents a random artifact, then of

the eight individual experiments performed, the effect should occasionally occur in time

zero samples.

Analysis of the LAT-transgenic was included to eliminate averaging effects of

latently infected DRG. The fact that results were as variable when using only 20% of the

total RNA from one mouse, indicates that pooling of tissues was not the cause. Despite

the availability of a single cellular control, Real-Time PCR results again indicate that

LAT abundances are accurately reported. While some of these experiments compare

DRG of males and females, previous work has shown that sex does not influence LAT-

expression profiles. Since LAT is capable of responding to explant-induced stress

independent of exogenous viral factors, it seems logical to state that this represents the

first event of sensing cellular stimuli to initiate the reactivation cascade.

Lytic Gene Activation During Explant Co-cultivation, and the Requirement of LAT-

The original goal of monitoring lytic gene activation during the early hours of

explant co-cultivation proved to be more easily planned than accomplished. The

unpredictable nature of molecular reactivation coupled with the extreme sensitivity

requirements made textbook completion of our goals nearly impossible. Valuable

observations were, none-the-less, obtained from IE transcript analysis. Though sporadic

in nature, ICPO (and to a lesser extent ICP4) was detected during many of the co-

cultivation time points examined. The fact that lytic gene activation does not precisely

correlate with decreased LAT abundance is understandable given LAT's inherent

variability, and the pooling of ganglia for each time point. The observation that a LAT-

null (17APst) mutant may exhibit more consistent lytic gene activation fits our current

model and is encouraging, but requires more detailed analysis.

The major benefit of the acute portion of the reactivation study is that the original

question responsible for this entire project is answered. We wanted to examine the

validity of observations claiming that E genes preceded IE genes (by up to 24 hours)

during explant co-cultivation. Our data clearly show that IE genes can be detected well

before twenty-four hours. We were not able, however, to prove or refute claims of E

genes preceding IE genes, since our results detect lytic transcripts representative of all

gene classes at very early times post-explant (data not shown). While this result may not

be earth shattering, we feel well-controlled experiments demonstrating a systems

complexity are more valuable than poorly controlled experiments that diagram complex

mechanisms. If this work contributes anything to the HSV field, it should be that explant

co-cultivations (as with human reactivation) are not static, and should not rely on a clock

for molecular analysis.

Proposed Mechanism for LAT's Role in Reactivation

A unifying model for LAT's involvement in reactivation has eluded HSV-1

researchers since its discovery. LAT has long been known to play a major role in

reactivation; the effect of specific promoter and 5'exon deletions have been carefully

examined in a variety of animal models. Cell culture systems capable of re-creating

certain aspects of HSV-1 latency and reactivation have, likewise, demonstrated LAT's

capacity to sense and respond to induction stimuli. While previous in situ analysis

detected 2-fold decreases in LAT abundance during co-cultivation of marine trigeminal

ganglia, this is the first evidence of a dramatic and global decrease in LAT abundance in

vivo. The most exciting extension of this result is the correlation it establishes with

Christine Wilcox's cell culture model of latency and reactivation (see introduction). This

is one of the few in vitro systems capable of reproducing the three hallmarks of latency:

genome circularization, LAT accumulation, and an ability to respond to induction stimuli.

Using this model, it was reported that six hours post-induction and continuing through

twenty-four hours, in situ analysis reveals a dramatic (10-fold) decrease in LAT

abundance, concomitant with activation of the CREB antagonist ICER. This represents a

potentially powerful relationship whereby observations in vivo can be examined in more

rigorously controlled settings. Likewise, observations from cell culture can be applied to

more biologically relevant animal models. Future work will, for example, certainly

include determining whether reduced LAT abundance correlates with increased levels of


HSV-1 faces the unique problem of having to balance two very different

requirements; while it expends considerable effort creating a state of extreme repression

that may last forever, it must retain the ability to respond to stressors as subtle as

ultraviolet light. The mechanism controlling this process must require layers of

regulation, which begin at a general (global) level, but become increasingly stringent.

Here is the problem: reactivation cannot be too easy (the viruses' evolutionary strategy is

based on subversion), or so difficult that it requires a lottery of events. With this in mind,

the difficulty associated with detecting ICPO during reactivation should not be terribly

surprising. Even in the explant co-cultivation model, less than one percent of latently

infected neurons reactivate. The dramatic decreases in LAT abundance reported here

clearly occur on a much greater scale than this! I believe that LAT directs the first step of

the reactivation process, and while it is important, it is not absolute. The mechanism

controlling decreased LAT abundance serves as a global "safety" to provide specific

neurons the opportunity to fire. More stringent downstream factors (neuronal phenotype,

transcription factor profile, latent genome copy number etc.) then decide the viruses'

ultimate fate.

A number of interesting models have been invoked to explain the actual

mechanism by which LAT functions during latency and reactivation. These include roles

as a functional RNA to potentiate genome silencing (similar to XIST's inactivation of X

chromosomes), or as a ribozyme or RNAi to selectively degrade unwelcome lytic

transcripts during latency. Others propose that transcription from this region maintains a

window of accessibility in an otherwise highly chromatinized molecule to allow select

transcription factors a place to initiate reactivation. Evidence has also suggested that

LAT's reactivation critical region (RCR) contributes a bi-modal enhancer that alternately

modulates lytic (most likely ICPO) and latent (LAT) promoters. While promoter mutants,

and in vitro correlations with ICER induction, indicate transcription level LAT

regulation, the sudden and dramatic decreases in LAT abundance cannot be solely

explained by promoter activity. The fact that the stable 2.0 kb intron's (24 hour half-life)

abundance also decreases, indicates that specific factors (such as RNAi) are actively

facilitating this degradation. The most recent and compelling data from our laboratory

indicates that specific histone modifications play an important role in establishing

chromatin domains that render certain regions transcriptionally active (euchromatin) and

others inactive (heterochromatin). Preliminary analysis of latent genomes indicates that

such a boundary resides around the LAT region, with acetylated histones associated with

the LAT promoter, and methylated histones associated with nearby lytic gene promoters.

If the RCR acts an enhancer capable of activating ICPO during reactivation, chromatin

boundaries might prevent this from occurring during latency. Our co-cultivation analysis

will have, therefore, identified the window during which chromatin boundaries are

absolved. Results (from our lab and others) indicate that the acetylation profile of histone

tails in the LAT region change during the early hours of reactivation. Nancy Sawtell and

Rick Thompson (University of Cincinnati) have recently reported that the LAT promoter

becomes more highly associated with acetylated histones following explant co-cultivation

of murine trigeminal ganglia. I believe that continued transcription of the LAT locus

during latency allows the RNA polymerase to maintain the proper chromatin state.

Transcriptional repression of the LAT locus would allow insulator boundaries to

restructure so that the RCR enhancer could activate lytic genes. This work has already

begun, but requires a more thorough understanding of histone profiles during latency

before comparisons to reactivation can be made. A fifth LAT-transgenic co-cultivation

experiment was performed to determine if decreased LAT abundance correlated with the

promoter's transition from acetylated to methylated histones. While a significant

decrease in LAT cDNA levels was not observed at either time point (2 or 4 hours), an

approximate two-fold reduction of acetylated histone-4 (histone 4 acetylated at lysine 9)

was detected by two hours. While other results in our laboratory have shown more

dramatic effects, two-fold differences determined by Real-Time PCR are considered

reliable in the chromatin field.


-__ LAT I


& lo.l .. ... .
-1 0 1 2 3 4 5

Tie hersr)

Figure 3.1. 17syn+ co-cultivation experiment number one.


Al A



-0.5 0 0.5 1 1.5 2 2.5
Time (hours)

Figure 3.2. 17syn+ co-cultivation experiment number two.


. .."" m m





-1 0 1 2
Time (hours)

3 4 5

Figure 3.3. KOS co-cultivation experiment number one.




' t I .


0 0.5 1
Time (hours)

1.5 2 2.5

Figure 3.4. KOS co-cultivation experiment number two.

. . .' .

0 1 2 3 4 H20

nub -a?*;

0 0.5 1 2 HO

& i



Figure 3.5. Standard PCR analysis of 17syn+ co-cultivation experiment one and two.

0 0.5 1
Time (hours)

1.5 2 2.5

Figure 3.6. LAT-Transgenic co-cultivation experiment number one.

Time (hrs)


- LAT k









T...; ;




...* ..1 *1...I...l.....I *

-1 0 1 2 3 4 5

Time (hours)

Figure 3.7. LAT-Transgenic co-cultivation experiment number two.




o- LAT
--0- PR

-1 0 1 2 3 4 5
Time (hours)

Figure 3.8. LAT-Transgenic co-cultivation experiment number three.





...3.... p.. .3... p....3.

-1 0 1 2
Time (hours)

3 4 5

Figure 3.9. LAT-Transgenic

co-cultivation experiment number four.

Time(hrs) 0.5 1 2 0 .5 1 2 3 4 0 .5 1 2 3 4 0 .5 1 234

5'LATcDNA 00


APRTcDNA 4 c6 ** *- AD do a aJ w. Ow

Figure 3.10. Standard PCR analysis of LAT-Transgenic experiments 1-4.


A im i


Time (hrs)

0 1 2 3 4


Figure 3.11. Detection of ICPO and ICP4 from 17syn+ experiment number one.

Inw (hI)

Figure 3.12. Detection of ICPO and ICP4 from 17APst and KOS infected mice.

'0 1 2 .3 4.. 1-. 2., '4 4.... 3::: : '..

^ii, & i*----gs
ICPO (-)-lxl6pfiumouse KD6-2x106pfu/mouse KOS-lxl05pfu/mouse

Figure 3.13. Levels of establishment from ICPO(-), ICP4(-) and KOS viruses.


1 White, D.O. and Fenner, F.J. (1994) Medical Virology, Academic Press

2 Roizman, B., ed. (1982) The Herpes viruses, Plenum Publishing Corporation

3 McGeoch, D.J. and Cook, S. (1994) JMol Biol 238, 9-22

4 McGeoch, D.J. et al. (1995) JMol Biol 247, 443-58

5 Wildy, P. (1973) Cancer Res 33, 1465-8

6 Roizman, B. et al. (1970) Nature 227, 1253-4

7 Wagner, E.K. and Bloom, D.C. (1997) Clin. Micro. Reviews 10, 419-443

8 Buchman, T.G. et al. (1980) Ann N YAcad Sci 354, 279-90

9 DeLuca, N.A., McCarthy, A.M. and Schaffer, P.A. (1985) J. Virol. 56, 558-570

10 Metzler, D.W. and Wilcox, K.W. (1985) J Virol 55, 329-37

11 Preston, C.M. (1979) J Virol 32, 357-69

12 Knipe, D.M. and Smith, J.L. (1986) Mol Cell Biol 6, 2371-81

13 Zhu, Z. and Schaffer, P.A. (1995) J Virol 69, 49-59

14 O'Hare, P. and Hayward, G.S. (1985) J. Virol 53, 751-760

15 Kristie, T.M. and Roizman, B. (1986) Proc Natl Acad Sci US A 83, 4700-4

16 Smith, C.A. et al. (1993) J Virol 67, 4676-4687

17 Post, L.E., Mackem, S. and Roizman, B. (1981) Cell 24, 555-65

18 Everett, R.D., Earnshaw, W.C., Findlay, J. and Lomonte, P. (1999) EMBO 18,

19 Nabel, G.J., Rice, S.A., Knipe, D.M. and Baltimore, D. (1988) Science 239, 1299-

20 Wagner, E.K., Guzowski, J.F. and Singh, J. (1995) in Progress in Nucleic Acid
Research and Molecular Biology (Cohen, W.E. and Moldave, K., eds), pp. 123-
168, Academic Press

21 Jacob, R.J. and Roizman, B. (1977) J Virol 23, 394-411

22 de, B.K.A. and Knipe, D.M. (1988) Cell 55, 857-68

23 Knipe, D.M., Senechek, D., Rice, S.A. and Smith, J.L. (1987) J Virol 61, 276-84

24 Campbell, M.E., Palfreyman, J.W. and Preston, C.M. (1984) JMol Biol 180, 1-19

25 Grondin, B. and DeLuca, N. (2000) J Virol 74, 11504-10

26 Roberts, S.G. et al. (1993) Nature 363, 741-4

27 Kobayashi, N., Boyer, T.G. and Berk, A.J. (1995) Mol Cell Biol 15, 6465-73

28 Villarreal, L.P., Defilippis, V.R. and Gottlieb, K.A. (2000) Virology 272, 1-6

29 Rock, D.L. and Fraser, N.W. (1983) Nature 302, 523-5

30 Puga, A. et al. (1984) J Gen Virol 65 (Pt 2), 437-44

31 Efstathiou, S. et al. (1986) J Virol 57, 446-55

32 Margolis, T.P., Dawson, C.R. and LaVail, J.H. (1992) Invest. Ophthalmol. Vis. Sci.

33 Cook, M.L. and Stevens, J.G. (1973) Infect Immun 7, 272-88

34 Deshmane, S.L. and Fraser, N.W. (1989) J Virol 63, 943-7

35 Cabrera, C.V. et al. (1980) Nature 288, 288-290

36 Margolis, T.P. et al. (1993) Virology 197, 585-592

37 Sawtell, N.M., Poon, D.K., Tansky, C.S. and Thompson, R.L. (1998) J. Virol. 72,

38 Mitchell, W.J. (1995) J Virol 69, 7942-50

39 Loiacono, C.M., Myers, R. and Mitchell, W.J. (2002) J Virol 76, 2449-59

40 Stevens, J.G. et al. (1987) Science 235, 1056-9

41 Ecob-Prince, M.S., Hassan, K., Denheen, M.T. and Preston, C.M. (1995) J. Gen.
Virol. 76, 1527-1532

42 Gressens, P. and Martin, J.R. (1994) J. Virol. Methods 46, 61-83

43 Mehta, A. et al. (1995) Virology 206, 633-640

44 Farrell, M.J., Dobson, A.T. and Feldman, L.T. (1991) Proc. Natl. Acad. Sci. U.S.A.
88, 790-794

45 Wu, T.-T., Su, Y.-H., Block, T.M. and Taylor, J.M. (1996) J. Virol. 70, 5962-5967

46 Spivack, J.G., Woods, G.M. and Fraser, N.W. (1991) J Virol 65, 6800-10

47 Batchelor, A.H. and O'Hare, P. (1992) J. Virol. 75, 3573-3582

48 Kenny, J.J. et al. (1994) Virology 200, 220-235

49 Zwaagstra, J.C., Ghiasi, H., Nesburn, A.B. and Wechsler, S.L. (1991) Virology
182, 287-297

50 Jarman, R.G. et al. (2002) Virology 292, 59-69

51 Bloom, D.C., Hill, J.M. and Wagner, E. (1996) in preparation

52 Bloom, D.C. et al. (1996) J. Virol. 70, 2449-2459

53 Hill, J.M. et al. (1997) J. Virol. 71

54 Loutsch, J.M. et al. (1998) J Virol., submitted

55 Laoide, B.M., Foulkes, N.S., Schlotter, F. and Sassone-Corsi, P. (1993) Embo J 12,

56 Nakajima, T. et al. (1997) Genes Dev 11, 738-47

57 Bloom, D.C., Stevens, J.G., Hill, J.M. and Tran, R.K. (1996) Virology 236, 202-

58 Gebhardt, B.M. and Kaufman, H.E. (1995) Antiviral Res 27, 255-61

59 Sawtell, N.M., Poon, D.K., Tansky, C.S. and Thompson, R.L. (1998) J Virol 72,

60 Sawtell, N.M. and Thompson, R.L. (1992) J Virol 66, 2150-6

61 Sawtell, N.M. and Thompson, R.L. (1992) J Virol. 66, 2157-2169

62 Thompson, R.L. and Sawtell, N.M. (2001) J Virol 75, 6660-6675

63 Margolis, T.P. et al. (1992) Virology 189, 150-60

64 Pemg, G.C. et al. (2000) Science 287, 1500-1503

65 Mador, N. et al. (1998) J Virol 72, 5067-75

66 Devi-Rao, G.B., Bloom, D.C., Stevens, J.G. and Wagner, E.K. (1994) J. Virol. 68,

67 Nesbur, A.B., Elliot, J.H. and Leibowitz, H.M. (1967) Arch. Ophthalmol. 78, 523-

68 Berman, E.J. and Hill, J.M. (1985) Invest. Opthalmol. Vis. Sci. 26, 587-590

69 Hill, J.M., Haruta, Y. and Rootman, D.S. (1987) Curr. Eye Res. 6, 1065-1071

70 Hill, J., Rayfield, M. and Haruta, Y. (1987) Current Eye Res. 6, 91-97

71 Kwon, B.S. et al. (1981) Invest. Ophthalmol. Vis. Sci. 21, 442

72 Kwon, B.S., Gangarosa, L.P., Green, K. and Hill, J.M. (1982) Invest. Ophthalmol.
Visual Sci., 818-821

73 Bloom, D.C. et al. (1994) J. Virol. 68, 1283-1292