3%
(hum.
:

v.

“5.555.:

9.3:.-

a. .
.N. a. m. . .. , . )flhfi .5.an .
s. . .25 1 . s...

5.
Iv II
at?

..
III a

.Wfififi

 

. urn-“TA.
LL...

 

 

 

 

 

. . 3P
1.1».

Wit-Sho- w.” .5.
E. 3116s.}-

..rl). I)».

 

 

 

 

NIEV EIIRRSTYLBA

Illllllllllllllllllllllll ll lllllll

3 1293 015550480

lllllllll

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled
The effect of deletion mutations of the
activation domain of VP16 upon herpes simplex virus
type 1 lytic infection

presented by

Rath Pichyangkura

has been accepted towards fulfillment
of the requirements for

Ph.D. . Biochemistry
degree In

 

 

 

Date 71/72 61/?6

012771

MSU is an Affirmative Action/Equal Opportunity Institution

 

  
    
   

 
 

LIBRARY
MIChIQan State
University

PLACE N RETURN BOX to romovo thIo chookout from your rooord.
TO AVOID FINES return on or bdoro dot. duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Afflmotlvo Action/Equal Opportunlty InItItution
m

 

THE EFFECT OF DELETION MUTATIONS OF THE ACTIVATION
DOMAIN OF VP16 UPON HERPES SIMPLEX VIRUS TYPE 1 LYTIC
INFECTION

By

Rath Pichyangkura

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Biochemistry

1996

ABSTRACT

THE EFFECT OF DELETION MUTATIONS OF THE ACTIVATION
DOMAIN OF VP16 UPON HERPES SIMPLEX VIRUS TYPE 1 LYTIC
INFECITON

By

Rath Pichyangkura

VP16 is an essential structural protein in the virion of herpes simplex virus
type 1 which functions in the activation of the viral immediately early genes and
in maturation of the virion particle. The transcriptional activation domain of the
protein can be separated into two subdomains, the N-subdomain and the C-
subdomain. Both of these subdomains are necessary for the full transcription
activation function of the protein, but either one is able to activate transcription
independently of the other. VP16 truncation mutations with deletion of either
one or both subdomains were constructed in the context of the viral genome and
tested for their effect upon viral lytic infection in cell culture.

Recombinant viruses with deletions in the activation domain of VP16 are
viable in tissue culture and produce intact particles containing the truncated
form of VP16 protein. However, these viruses demonstrated defects in their

plaque characteristics, single step growth, and the number of particles per plaque

forming unit. A recombinant virus bearing a deletion of the C-subdomain gave a
slightly small plaque with no other defective phenotype. Deletion of the N or
both subdomains gave much smaller plaques and grew poorly in single step
growth assays. These viruses also exhibit higher particle per plaque forming unit
ratios, comparing to the wild type virus. The particle per plaque forming unit
ratios of the viruses that bear deletions of the N subdomain or of both
subdomains were approximately 30-fold and loo-fold greater than of the wild
type virus. Despite their defects, all recombinant viruses with truncated VP16
genes were able to deliver their DNA to the nucleus of the host cell. Therefore,
these defects involve the activation of immediate early gene expression.

Homologues of VP16 have been identified in herpes simplex virus type 2,
bovine herpes virus type 1, equine herpes virus types 1 and 4, varicella zoster
virus, and Marek's disease virus. Hydrophobic cluster analysis of deduced
protein sequences was used to predict the location and critical amino acids of the
activation domain of the varicella zoster virus VP16 homologue. This prediction
was tested by mutational analysis which further emphasized the importance of
large hydrophobic residues as an essential component of the activation domain
of VP16 and its homologues. Hydrophobic cluster analysis also suggested the

location and critical residues in other VP16 homologues.

ACKNOWLEDGMENTS

I wish to thank Dr. Steven Triezenberg for being an excellent graduate
advisor and mentor. I also thank him for his patience, understanding and
encouragements through the years. I also wish to thank the members of the
Triezenberg laboratory, both past and present, namely Doug and Andrea Cress,
Dave Pepperl, Jeff and Marty Regier, Lisa Ortquist, Fan Shen, Peter Horn, John
Stebbins, Lee Alexander , Jaya Reddy, Susan Sullivan, Eugene Chung and Soren
Jaglo-Ottosen. I thank them for their great friendship and creative discussions.

I acknowledge the contributions of the members of my guidance
committee Drs. Laurie Kaguni, Lee Kroos, Lee Velicer and Pamela Green. I
acknowledge the contributions of my collaborators Drs. Hiroyuki Moriuchi,
Masako Moriyuchi and Jeff Cohen (NIAID), for the VZV ORF 10 work, and Dr.
David Koelle (University of Washington) for the CD4+ T—cell work. I also wish
to acknowledge the Anantamahidol foundation for their financial support.

I thank the members of the Thai community as well as the international
community that kept my life interesting during my stay at MSU. I also thank my
close friends Darunee Buripakdi and Takeshi Ueda for their encouragement and
support during the final period of my dissertation. Finally, I thank my parents,

brother and sister for their support, encouragement and patience over the years.

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

PAGE
LIST OF TABLES viii
LIST OF FIGURES ix
LIST OF ABBREVIATIONS xi
CHAPTER I: INTRODUCTION 1
Eukaryotic transcriptinn 1
Regulated transcription 4
Transcription activators 6
Structure of activation domains 9
VP16 11
VP16 homologue: 18
Overview 19

 

CHAPTER II: EFFECTS OF C-TERMINAL DELETION MUTATIONS
IN VP16 ON THE LYTTC INFECTION OF HSV-l KOS ....... 22

 

 

 

 

Infrndiir'tinn 22
Mpthndo 27
Isolation and identification of recombinant HSV-l ................................ 37
Plaque and growth characteristic of recombinant HSV-l ...................... 44
Recombinant HSV-l are defective in infecti an 52
Discussion 58
Future studies 62

 

TABLE OF CONTENTS (cont'd)
PAGE
CHAPTER III: IDENTIFYING THE TRANSCRIPTION
ACTIVATION DOMAINS IN VP16 HOMOLOGUES ...... 64

 

 

Intrnducfinn 64
Mathnde 66
Hydrophobic cluster analysis 73

 

Site directed oligonucleotide mutagenesis of the proposed
VZV ORFIO activation domain 82

Secondary structure prediction by Chou-Fasman and

 

 

 

 

Garnier-Osguthorpe-Robson method 93

Secondary structure prediction by Sequery and PHD ........................... 98
merlrceinn 103
Future studies 112
APPENDICES 114

APPENDIX A: Hydrophobic cluster analysis predicts an amino
-terminal domain of varicella-zoster virus open reading
frame 10 required for transcriptional activation.
Hiroyuki Moriuchi, Masako Moriuchi, Rath Pichyangkura,
Steven J. Triezenberg, Stephen E. Straus, and Jeffrey I. Cohen,
Proc. Natl. Acad. Sci. USA Vol. 92, pp. 9333-9337, 1995 ............. 114
APPENDIX B: List of protein from sequery search 127
APPENDIX C: List of publications 128

 

vi

TABLE OF CONTENTS (cont'd)
PAGE

LIST OF REFERENCES 129

 

vii

LIST OF TABLES

 

 

 

 

 

 

PAGE
CHAPTER II
Table 1. Particle count and particle per plaque forming unit ratio ............... 48
CHAPTER III
Table 1. Mutagenesis olignnur'lmtide 68
Table 2. Relative activity of GAL4-ORF10(5-79) bearing amino
acid substitution at position 28 88
Table 3. Phe28 is required for transactivation of the VZV IE62
promoter by ORF 10 protein 90
Table 4. Relative activity of GAL4-ORF10(5-79) bearing amino
acid substitution at hydrophobic residues flanking Phe28 .............. 92
Table 5. Secondary structure prediction by Chou-Fasman and
Garnier-Osguthorpe-Robinson, and Emini surface
probability of the N-subdomain of HSV-1 VP16,
HSV-2 VP16 and VZV ORF1 n 95
Table 6. PHD secondary structure prediction of the
N-subdomain of HSV-1 VP16, HSV-Z VP16, BHV-l
UL48, and EHV-l Genel 7 100
Table 7. PHD secondary structure prediction of the
N—subdomain of EHV-4 GeneBS, MDV UL48,
MDV UL48, and VZV ORF1 n 102
Table 8. Secondary structure prediction by Sequery ....................................... 108
APPENDIX B
Table 1. List of proteins from Sequery search 127

 

viii

LIST OF FIGURES

 

 

 

 

 

 

 

 

 

PAGE
CHAPTER I

Figure 1. HSV-l KOS particle 12
Figure 2. Schematic representation of the replication cycle

of HSV-1 in susceptible cells 15

CHAPTER 11

Figure 1. Plasmid map of pKOSVP1 6 7 29
Figure 2. Fragments containing the VP16 and derivatives

used to generate recombinant Vin no: 32
Figure 3. Southern blot analysis of recombinant viral DNA ......................... 39
Figure 4. Western blot analysis of recombinant viral particles ..................... 41
Figure 5. Western blot analysis of RPS particles generated

in Vero 16-8 cells 43
Figure 6. Plaque characteristic of RP1, RP3, RP4, and RPS

on Vero cells 46
Figure 7. Southern blot analysis of revertant viral strain: 49
Figure 8. Single step growth curve of recombinant virus

strains RP1, RP3, and RP4 (Standard condition) .............................. 51
Figure 9. Penetration curve 53
Figure 10. Single step growth curve of recombinant virus

strains RP1, RP3, and RP4 (Acid wash condition) ........................... 54
Figure 11. Titer of recombinant viral stocks in Vero and

Vero 16-8 56
Figure 12. Southern blot analysis of nuclear localized

recombinant HSV-l DNA 57

 

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

LIST OF FIGURES (cont'd)

 

 

 

 

PAGE

CHAPTER III
Residues allowed at each position for the Sequery
program search and the six window used in the search .............. 71
Linear alignment of the amino acid sequence of
VP16 and its hnmnlngupe 74
HCA of the complete sequence of VP16 and other
VP16 hnmnlngnpc 78
HCA of the N-subdomain in VP16, ORFIO, and other
VP16 hnmnlngnpc 81
HCA of the C-subdomain in VP16 and other VP16
hnmnlngupc 84
Locations of the N and C subdomain in VP16 and
its hnmnlngnpc 85

 

BHV-l

CD
CPE

EHV-l
EHV-4
FCS
GTFs
HCA
HMG
HSV-1
HSV-2
hpi.
ICP

IE

MDV
moi

NMR

LIST OF ABBREVIATIONS

bovine herpes virus type 1
chloramphenicol acetyltransferase
circular dichroism
cytopathic effect
carboxy-terminal domain
early

equine herpes virus type 1
equine herpes virus type 4
fetal calf serum

general transcription factors
hydrophobic cluster analysis
high mobility group proteins
herpes simplex virus type 1
herpes simplex virus type 2
hours post infection

infected cell protein
immediately early

late

Marek's disease virus
multiplicity of infection

nuclear magnetic resonance

PHD
pfu
RNAPII
SDS
SDS PAGE
TAFs
TBP

TF
VP16
VPIBII
VZV

LIST OF ABBREVIATIONS (cont'd)

profile network system from Heidelberg

plaque forming unit

RNA polymerase II

sodium dodecyl sulfate

sodium dodecyl sulfate polyacrylamide gel electrophoresis
TBP associated factors

TATA binding protein

transcription factor

viral protein 16

viral protein 16 from herpes simplex virus type 2

varicella zoster virus

Single letter abbreviation for the amino acids: A, Ala; C, Cys; D, Asp; E, Glu; F,
Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S,
Ser; T, T'hr; V, Val; W, Trp; and Y, Tyr

CHAPTER I
INTRODUCTION

Eukaryotic Transcription

The regulation of gene expression in eukaryotes is one of the most
interesting and extensively studied aspects in biology. Gene expression is
regulated at multiple steps from transcription initiation, elongation, and
termination; mRNA processing, transport, and stability; translation,
modification, and degradation of gene products. This multi-level control of
regulation is essential for cell growth, development, maintenance, and the ability
to respond to both internal and external stimuli.

Transcription of protein-encoding genes by RNA polymerase II (RN APII)
is the initial and major regulated step in gene expression. In contrast to
prokaryotic RNA polymerase which is composed of 3 core peptides (a , [3, B') and
a factor, that confer its specificity to the promoter, eukaryotic RNAPII is
composed of 8-14 peptides, ranging from 10-220 kDa, and requires a variety of
additional factors for selective promoter recognition and regulation of
transcription initiation (Young, 1991). These factors are roughly divided into two
major categories, general transcription factors (GTFs) and transcription
regulatory factors.

The largest subunit of RNAPII has a unique heptapeptide repeat sequence,
Tyr-Ser-Pro-Thr-Ser-Pro-Ser, at the carboxy-terminal domain (CTD) of the

2
protein (Cadena & Dahmus, 1987). In cells, the largest subunit of RNAPII is

found in two forms, no and Ila, in which the CTD is highly phosphorylated and
unphosphorylated, respectively. The CTD is essential for cell viability, and
appears to be involved in the response to transcriptional regulation, and pre-
initiation complex formation of some promoters (Young, 1991). The highly
phosphorylated CI'D may also play a role in transcription initiation as a switch
from pre-initiation complex to open complex formation and promoter clearance
(Young, 1991; Zawel & Reinberg, 1993; Zawel & Reinberg, 1995).

The promoter structure of protein-encoding genes varies considerably
among different gene families, nevertheless each of them share common core
structures. RNAPII promoters have two possible core elements which can
mediate the nucleation of the transcription initiation complex, either
independently or synergistically. One of these elements, the TATA box, in
mammals is located approximately 25-30 nucleotides upstream of the
transcription start site. The other, the initiator (Inr) element is normally found
encompassing the transcription start site (Smale 8: Baltimore, 1989; Weis &
Reinberg, 1992; Zawel 8: Reinberg, 1995).

Most GTFs for RN APII which are essential for a basal level of
transcription have been purified, characterized, and cloned. The schematic order
of assembly of these factors into a transcription preinitiation complex has been
studied and laid out, however there are still some ambiguities concerning the
precise order of assembly, as well as the basal factors needed for different
promoters. Transcription initiation begins with the template commitment step in
which transcription factor (TF) IID, a large multiple protein complex comprising
the TATA binding protein (TBP), and the TBP associated factors (T AFs), binds to
the TATA motif of the promoter. The formation of the TFIID-DNA complex can
be further facilitated and stabilized by TFIIA. TFIIB then joins the complex

3
followed by the recruitment of RN APII and TFIIF to the pre-initiation complex.

TFIIE and TFIIH then join the complex. In the presence of ATP and dNTP's,
extensive phosphorylation of the CT D by TFIIH occurs followed by open
complex formation and the synthesis of the nascent mRN A chain (Zawel 8:
Reinberg, 1995).

There is increasing evidence suggesting that these factors exist in a large
multi-protein complex and the schematic order of assembly might not reflect the
natural phenomena of complex assembly within a living cell. RNAPII was found
to be associated with multiple transcription machinery SRB, SWI/SNF and GTFs
in a RNAPII holoenzyme (Koleske & Young, 1994; Koleske 8: Young, 1995;
Ossipow, et al., 1995; Chao, et al., 1996; Wilson, et al., 1996). TFIIF was initially .
found to be associated with RNAPII, and recent studies have established an
interaction between TFIIB, TFIIF and RNAPII in solution (Fang & Burton, 1995;
Ha, et al., 1993). It is plausible to assume that in the process of initiation complex
assembly, RNAPII, TFIIB, and TFIIF could be recruited to the TFIID-TFIIA-DNA
complex as a TFIIB-RNAPII-TFIIF, multi-protein complex instead of as individual
factors. This could be the case for other assembly steps in the initiation complex
as well.

On TATA-less promoters, initiation of complex assembly usually
commences with the association of initiator binding factors such as YYI (U sheva
& Shenk, 1994) or TFII-I (Roy, etal., 1993) to the Inr element nucleating
preinitiation complex assembly, followed by the recruitment of GTFs, RNAPII
and other components of the transcription machinery. Moreover some TAFs
such as TAF 150 have also been found to specifically recognize promoters
containing an Inr element (Verrijzer, et al., 1994; Verrijzer, et al., 1995).

Regulated Transcription

Regulated transcription is accomplished by the binding of additional
factors either upstream or downstream of the promoter element, which interact
either directly or indirectly with the transcription initiation complex to regulate
its activity. These factors could either stimulate or repress transcription from the
promoter which they regulate. Transcriptional regulatory proteins consist of at
least two functional domains. One of these domains confers specificity to the
promoter, through specific DNA binding or protein-protein interactions. The
other domain is the regulatory domain which functions to regulate the level of
transcription (Johnson & McKnight, 1989; Triezenberg, 1995; Zawel & Reinberg,
1995). These regulatory factors are simply characterized by their functional role
as transcription activators or repressors.

Despite the regulation by promoter specific factors, more general modes of
gene regulation, such as those mediated by histone proteins, high mobility group
proteins (HMG) and the methylation state of the DNA template could greatly
contribute to the precise regulation of gene expression. It is well documented
that packaging of a DNA template into chromatin by histone proteins renders the
DNA template inert to transcription (Paranjape, et al., 1994). Both biochemical
and genetic studies have demonstrated the significant role of histone proteins in
transcriptional regulation (Paranjape, et al., 1994). Histone H1, in a complex with
template DNA, has been shown to have a potent inhibitory effect on RNAPII
(Croston, et al., 1991 ; Laybourn 8: Kadonaga, 1991). The N—terminal domains of
histone H3 and H4 have been shown to play a role in the regulation of gene
expression. The N-terminal domain of H4 is necessary for activation of the GAL1

promoter, as well as silencing of the yeast mating loci and other yeast promoters,

5
while H3 is required for the repression of most genes which require H4 for

activation (Wolffe, 1994). Lysine residues in the N-terminal domain of histone
H3 and H4 are also subjected to reversible acetylation, which has been found to
have a positive effect upon transcription (Paranjape, et al., 1994).

HMG proteins are a group of non-histone proteins which were initially
defined as proteins that could be extracted from nuclei with 0.35 M NaCl and
were soluble in 2% trichloroacetic acid. Three groups of HMG proteins have
been defined, I-IMG1 / 2, HMG14/ 17, and HMGI/ Y (Paranjape, et al., 1994).
HMG 1/ 2 was found to be able to stimulate RNAPII transcription and the
binding of upstream stimulating factors in vitro (Watt 8: Molly, 1988). HMG17
can stimulate activation of transcription by VP16 on in vitro assembled chromatin
(Paranjape, Krumm, 8: Kadonaga, 1995). HMGI/ Y was also found to stimulate
the binding of NF-KB to the PDRII promoter element (Thanos 8: Maniatis, 1992).
In spite of these minor clues, the biological role of these proteins in transcription
regulation remains to be clarified.

Methylation of the DNA, at the 5 position of cytosine residues in the
sequence dedG, plays an important role in higher eukaryotic transcription
regulation. There is a strong correlation between the presence of 5-
methylcytosine (m5C) in the DNA template and the repression of transcription
(Cedar, 1988; Doerfler, 1983; Doerfler, et al., 1990; Dynan, 1989; Guentte, et al.,
1992; Lewis a: Bird, 1991 ; Peek, et al., 1991). Methylation of the DNA template
can effect transcription presumably at two levels, by affecting the packaging of
the DNA, and by interacting either directly or indirectly with the transcription
machinery of the cell.

Methylated cytosine residues in chromatin are less accessible to nudeases
such as micrococcal nuclease, suggesting that they occupy a portion of DNA that
is tightly packed (Lewis 8: Bird, 1991). When chromatin is fractionated, the

6
nucleosome fraction which is tightly associated with histone H1 was found to

contain the majority of the methylated cytosine (Lewis 8: Bird, 1991). There is
also evidence indicating that histone H1 can associate more readily with
methylated than unmethylated DNA (Johnson, Goddard, 8: Adams, 1995).

Methylation of specific sites in genes or flanking sequences, such as a
promoter sequence (Guentte, et al., 1992; Peek, et al., 1991) or an intron
(Graessmann, et al., 1994), can also repress transcription. It is believed that
methylation inhibits activation by interfering with the binding of positive
regulators to their recognition sites. There are also proteins that bind specifically
methylated DNA sequences (Boyes 8: Bird, 1991) that might block transcription
by RN APII. Despite these inhibitory effects, methylation of some promoters or
genes does not seem to effect transcription or the function of certain activators
(Harrington, et al., 1988; wolfl, et al., 1991).

Although these results do not imply that methylation is only involved in
the inhibition of gene expression, it suggests an undeniable role of methylation
on gene expression by RNAPII. The mechanism of how methylation effects
transcription by RN APII is very intriguing and still remains to be studied.

Transcription Activators

Transcription activators are characterized roughly according to the most
prominent species of amino acid found in the primary structure of the activation
domains. Four groups of activators have been initially defined as acidic-rich ,
proline-rich, glutamine-rich and serine and threonine-rich activators (Johnson, et
al., 1993). Recent studies have demonstrated that the most apparent amino acids
in these activation domains may not be the most critical residues that contribute

to activation ability (Triezenberg, 1995). For example, in the case of the acidic-

7
rich activators, VP16 (Cress 8: Triezenberg, 1991b; Regier, Shen, 8: Triezenberg,

1993) and GCN4 (Drysdale, et al., 1995) or the glutamine-rich activators Spl (Gill,
et al., 1994) and Oct-1 (Tanaka 8: Herr, 1994), the most critical residues for their
fimction appears to be the bulky hydrophobic residues,

Transcriptional activators could function in multiple distinct fashions at
multiple steps in the transcription process; however, the mechanism of how these
proteins work is not well established. To understand the mechanism of how
these activators work, attempts have been made to identify the "targets" of
activation domains. Biochemical and biophysical studies have shown
interactions between GTFs and transcription activators. The acidic activation
domain of VP16 has been shown to interact with TBP (Ingles, et al., 1991 ;
Stringer, Ingles, 8: Greenblatt, 1990), and TFIIB (Lin, et al., 1991). The 62 kDa
subunit of TFIIH was also shown to interact with the acidic activators VP16 and
p53 (Xiao, et al., 1994). Mutations in the activation domain that effect the
transcription activation ability also effect the interaction between these factors
suggesting a significant role for these interactions in transcription activation.

In some cases the activation domains of activators do not directly interact
with their targets, but rather through factor(s) which are termed adaptors,
coactivators or mediators. A number of factors which serve as a bridge or link
between activators and GTFs have been characterized. ADA2 in complex with
ADA3 and GCNS was found to be necessary for the acidic activators GAL4 and
the chimeric GALA-VP16 to work properly in yeast (Berger, et al., 1992; Horiuchi,
Silverman, Marcus, 8: Guarente, 1995; Marcus, et al., 1994). Previously, several
TAFs were also found to interact with transcription activators. Drosophila
TAFII40, homologous to human TAFII32, was found to interact with the
activation domains of VP16 (Goodrich, et al., 1993; Klemm, et al., 1995) and p53
(Thut, et al., 1995). Drosophila TAF110 interacts with the glutamine rich Spl

8
(Hoey, et al., 1993), human TAF55 interacts with CI'F (Zawel 8: Reinberg, 1995)

and TAF150 has been shown to interact with NTF-l (Chen, et al., 1994).

Furthermore, Hengartner et al. have demonstrated that the multi-protein
RNAPII holoenzyme, comprising RN APII, SRB, SWI/SNF and GTFs (Koleske 8:
Young, 1994; Koleske 8: Young, 1995; Ossipow, et al., 1995; Chao, et al., 1996;
Wilson, et al., 1996), was capable of responding to transcription activators. The
acidic activation domain of VP16 was shown to specifically bind to the RNAPII
holoenzyme, suggesting that transcription activators may function, in part,
through direct interaction with the RNAPII holoenzyme (Hengartner, et al.,
1995).

The mechanism by which these interactions could lead to the activation of
transcription has been studied for some activators. Transcription activators were
found to have the ability to disrupt or reconfigure chromatin to allow access of
GTFs as well as other molecules of activators (Axelrod, Reagan, 8: Majors, 1993;
Pazin, Kamakaka, 8: Kadonaga, 1995) to the promoter.

Recent studies by Wilson et al. demonstrated that SWI/SNF is a
component of the RNAPH holoenzyme (Wilson, et al., 1996). This would provide
the holoenzyme with the capability to disrupt nucleosomal DNA. Therefore, the
recruitment of RNAPII holoenzyme by transcription activators, such as VP16,
would facilitate the disruption of nucleosome and stable binding of the various
transcription components of the initiation complex to the promoter.

Some activators are able to promote assembly of the preinitiation complex.
For example the formation of the TFIID-TFIIB-TFIIA complex is enhanced by the
activation domain of Zta (Lieberman, 1994; Lieberman 8: Berk, 1994; Pazin, et al.,
1995). Moreover, in some cases activators can promote elongation of the mRNA
chain (Yankulov, et al., 1994) or alleviate inhibitory effects of some repressors
such as Drl or Dr2 (Kraus, et al., 1994; Merino, et al., 1993).

9
Previously it has been shown that the largest subunit of RNAPII has a

highly conserved acidic domain that shares similarity with the acidic activation
domain of VP16 (Xiao, et al., 1994). This domain could also act as a potent
activator of transcription when fused to the GAL4 DNA binding domain. These
findings together with others suggest an interesting model in which transcription
activators could work at both steps, by recruiting the G'I'Fs to the promoter and
also displacing the RNAPII to promote transcription initiation and promoter
clearance. This is a very plausible model, since the VP16 activation domain has
been shown to work at steps following transcription initiation (T iley, et al., 1992).

Another interesting aspect of how activators might work is by altering the
conformation of a transcription factor. TFIIB undergoes a conformational shift
upon assembly into the preinitiation complex (Roberts, et al., 1993). The
activation domain of VP16 can induce that same conformational change in T'FIIB
(Roberts, 1994). Therefore, VP16 could promote preinitiation complex assembly
by providing TFIIB with the proper conformation to enter the preinitiation

complex.

Structure of activation domains

Presently there are no well defined molecular models for the structure of
transcriptional activation domains, which prevents us from fully understanding
the mechanism of how activators work. Biophysical data, using circular
dichroism (CD) and nuclear magnetic resonance (NMR) , found that activation
domains are mostly unstructured, though some activation domains can be
induced to adopt an a-helical or B-sheet structure (Donaldson 8: Capone, 1992;
Van Hoy, et al., 1993) by altering the pH of the solution. In other words,

activation domains do not possess a defined conformation in solution at

10
physiological conditions. Nevertheless, emerging data from mutational analysis

of various activators has provided some clues as to their structure (T riezenberg,
1995) and the function of some activators depends on specific residues (Cress 8:
Triezenberg, 1991b), which indicates some degree of the structural requirement
for function.

Recent biophysical studies on the activation domain of VP16 using
fluorescence spectroscopy by Shen et al. was performed in the context of a GAL4
fusion protein. These studies demonstrated that the activation domain of VP16 is
mobile and unstructured in solution (Shen, et al., 1996a). Shen et al. also showed
that when the target protein, TBP or TFIIB, was added, the VP16 activation
domain was restricted from segmental motion. The environment of the
fluorescence probe used also changed to a more hydrophobic environment
suggesting that there are interactions between the activation domain and its
target (Shen, et al., 1996b). Thus transcription activators might form a distinct
conformation once they interact with their targets. Co-crystallized complexes of
activation domains with their targets could provide information concerning their
structure.

A majority of transcription activators are found to comprise multiple
domains or subdomains that can function as transcription activators
(Triezenberg, 1995). These domains could work completely independently,
dependently, or synergistically with each other. Different subdomains may also
be functionally distinct from each other. In other words they are not always
simply a duplication of one another. This is clearly shown in the case of VP16,
which is composed of two distinct activation subdomains. These two
subdomains can interact with different components of the transcription
machinery (Goodrich, et al., 1993;1ngles, et al., 1991; Lin, et al., 1991; Shen, et al.,
1996b) suggesting different mechanisms of activation. Therefore, a single

11
activator could function at multiple steps in the regulation of transcription.

VP16

Herpes simplex virus type 1 (HSV-l), is one of the most extensively
studied viruses in the herpes family. The genome of HSV-1 is a large double-
stranded DNA virus of approximately 150 kbp. The HSV-l genome has been
sequenced entirely (McGeoch, et al., 1986; McGeoch, Dalrymple, 8: Davison,
1988) and most of the genes and the origin of replication have been defined. The
viral DNA is packed within an icosahedral capsid surrounded by an amorphous
tegument and a spiked envelope derived from the host nuclear membrane
(Figure 1). The virion contains at least 33 viral polypeptides which represent
approximately one half of the 70 or so genes estimated to be encoded by the
HSV-1 genome (Roizman 8: Sears, 1990).

HSV-l has been used as a model system to study gene regulation in
mammalian cells. HSV-1 gene expression is coordinately regulated in a
sequentially ordered manner. Three major groups of genes, Immediate Early (IE)
genes, Delayed Early (DE) genes and Late (L) genes, have been characterized
based primarily on the order in which they are expressed. The IE genes are
expressed early in infection, and reach a peak expression rate at 2-4 hours post
infection. Their expression is independent of viral protein synthesis. Five IE
genes have been identified in HSV-l; ICPO, ICP4, ICP22, ICP27 and ICP47. These
gene products (except ICP47) were found to have regulatory functions
influencing the expression of another group of genes. The DE genes are
expressed later in infection. They are not expressed in the absence of IE gene
expression but can be induced in the absence of viral DNA replication. The

expression of DE genes peaks at about 5-7 hours post infection and they have

12

 

 

 

Figure 1. Transmission electron micrograph of HSV-1 (KOS) virion. Virions
were purified from infected cells by ultracentrifugation. This sample on a 200
mesh, gilder grid, was stained with 1% ammonium molybdate and observed on a
JEM-IOOCX II transmission electron microscope (JOEL). The arrows point to; (1)
viral envelope, (2) viral tegument, (3) viral capsid and (4) viral core, containing
the viral genome. Magnification: 58,000X

13
been found to encode primarily enzymes and proteins needed for viral

replication. The last group of genes, L genes, are expressed last in the cascade of
HSV-1 gene expression. The true L genes, such as the gC and L42 genes (Homa,
et al., 1991), are completely dependent upon viral DNA replication, while
expression of other viral L genes is reduced when viral DNA synthesis is
inhibited. These genes mostly encode the viral structural proteins and essential
proteins needed for viral packaging. Late gene expression is followed by the
assembly of the virion and shedding of the virion particles (Roizman 8: Sears,
1990; Whitley, 1990).

In the process of virion assembly, one of the L gene products termed VP16
[also known as Vmw65, ICP25, and a-Trans Inducing Factor or a-TIF (Campbell,
Palfreyman, 8: Preston, 1984; Post, Mackem, 8: Roizman, 1981)] that specifically
activates the viral IE genes is packed into the tegument of the virion. VP16, the
product of the HSV-l UL48 gene functions as a transcription activator . VP16
proteins are present at approximately 500-1000 molecules per virion particle
(Heine, et al., 1974; Spear 8: Roizman, 1972). These VP16 molecules will activate
the viral IE genes and start the cascade of viral gene expression in an infection.
VP16 was found to be essential for virion replication. Viral particles bearing
genomes that lack the gene encoding VP16 can infect and replicate their DNA in
the host cell, but fail to produce intact Virions (Weinheimer, et al., 1992).
Therefore, VP16 seems to serves two distinct functional roles, as a transcription
activator for the viral IE genes and also as a component of the virion essential for
particle assembly. The viral infection cycle as well as the order of viral gene
expression are summarized in Figure 2.

Deletion mutagenesis has been used to roughly map VP16, a 490 amino
acid residue protein, into two domains. The specificity domain which targets

VP16 to the IE promoters maps in the N-terminal residues 1410 (Ace, et al., 1988;

    

I” Became:

mad: mil
,. wqti Ind
III," ’Mfl‘f—
inns um 1:14 '1
“yd-alimony, .1:

u?’“ ' «.[xn '8‘"

mm [db 5“... a ehmg “+1 fLII I‘J '

:51: new abet: L11)...

11V :(7.
“.I M5".-

w

 

 

 

 

.ang’lnfi I!" '1‘1'1’...'A“.1', “LIN ."flr '(i'l‘t'qfl

F41“ brunt-swam-

TIL. _... ,

{at rm .5 L 5‘7“ '
7'1: “I"? in rim;

l

-..-IIdA tam elii‘

.-€‘.
IV ."J.
. 5-- .
511’
-- r - - 1‘
.>\ ' '
. v. I ‘j
., I
3...? 2 l
\_ a"
'.
‘ I
l
' I

-J

at r\'«"'.ri'.)':13vl r! 11.x x . m. r.) .u: ‘-. nwecqfifl Idsfba .9 ”-395

...;....e:: ' .. . ~ rim, 1.. Mt lather-tartan“;

$415!" M3 {d ”by?

1 ".4 Fifty“ "117' w" 1’9: 1" .1. a: 3.181?an smash

.: .1 v. :- . um ..21..s:;-.wr tardy-'gmtvzselnmssm

v. .2‘ ' -.-.. .. "I in. ' 1w .5:qu 0.2%!thth tad")
.m L“: : . ”'17‘3“? ”admin-{kg mic- Mauser

5 1112"! ".2; - f." 31111:"“1’3 .0 Edit-”gush 55mins

f\/1’.~II-" - it " by». ' ~

    
  
 
   
  
   
     
     

“. 1;” "

at?

 
 

It.
.5
graft“ I .
filo‘x’. .ng ‘ I
«napalm (ft-w

 
  

 

14

Figure 2. Schematic representation of the replication cycle of HSV-1 in
permissive cells. (1): The virus initiates infection by attachment to the cell
surface by the viral glycoproteins followed by the fusion of the envelope with the
plasma membrane. The fusion of the viral envelope with the cell membrane
releases at least two viral proteins into the cytoplasm, viral host shut off protein
(vhs) which inhibits protein synthesis in the host cell by promoting the
degradation of the polyribosomes, and VP16 which will migrate to the nucleus to
activate the expression of HSV-1 IE genes. (2): The viral capsid containing the
DNA is then transported to the nucleus, where the viral DNA is released into the
nucleus and circularizes. (3): The transcription of the IE gene is induced by VP16.
The IE transcripts are translated and the IE proteins are transported back into the
nucleus. (4): The presence of the IE gene products induces transcription and
translation of the DE genes. The viral DNA is then replicated. (5): L genes are
then transcribed and translated. The L proteins consist primarily of structural
proteins. (6): The capsid proteins form the viral capsids. The viral DNA is then
packed into the empty capsid with other structural proteins. (7): Viral
glycoproteins and tegument proteins accumulate and form patches in the nuclear
membranes. The capsid containing the DNA and additional proteins attach to
the underside of the membrane patch containing the viral protein, and are
enveloped. (8): The envelope capsids accumulate in the endoplasmic reticulum
and are transported into the extracellular space.

15

 

 

 

 

 

16
Friedman, Triezenberg, 8: McKnight, 1988; Triezenberg, Kingsbury, 8: McKnight,

1988a; Werstuck 8: Capone, 1989b). The activation domain which is both
necessary and sufficient for VP16 to activate transcription maps at the C-terminal
residues 410-490 (Sadowski, et al., 1988; Triezenberg, et al., 1988a; Werstuck 8:
Capone, 1989b; Cousens, et al., 1989b). The activation domain of VP16 is one of
the most potent activation domains tested (Carey, Leatherwood, 8: Ptashne, 1990;
Cousens, et al., 1989; Sadowski, et al., 1988). VP16 does not directly bind DNA
(Kristie 8: Sharp, 1990), but works through interactions with Oct-1 and at least
one other host cell protein called HCF, CFF, C1 or VCAF-l (Katan, et al., 1990;
Kristie, LeBowitz, 8: Sharp, 1989; Xiao 8: Capone, 1990). VP16 specifically
interacts with the POU domain of Oct-1 which binds to a recognition site, the
TAATGARAT (where R is a purine) (Kristie 8: Roizman, 1984; Triezenberg,
LaMarco, 8: McKnight, 1988b) element or the overlapping ATGCTAATGARAT
in the HSV-l IE gene promoters (Cleary 8: Herr, 1995; Preston, Frame, 8:
Campbell, 1988; Stem 8: Herr, 1991).

VP16 was categorized as a member of the acidic activator group due to the
prominence of acidic residues found in its activation domain. Its activation
domain, the 80 C-terminal amino acids of VP16, have been divided into two
subdomains, the N-subdomain (residues 410-456) and the C-subdomain
(residues 456-490). These two subdomains have distinct primary structures, and
evidence suggests that they can function independently and mechanistically
differently from each other (Goodrich, et al., 1993; Ingles, et al., 1991; Lin, et al.,
1991; Shen, et al., 1996b). Studies on the N-subdomain have shown that acidic
residues are necessary, but that hydrophobic residues, especially a phenylalanine
residue at position 442 (Phe 442), seem to be the key elements of this activation
subdomain (Cress 8: Triezenberg, 1991b; Regier, et al., 1993). Despite the
differences in the function and primary structure of the N and C subdomains,

17
random polymerase chain reaction and alanine scanning mutagenesis

experiments of the C-subdomain, residues 455-487, suggested that bulky
hydrophobic residues are also important for activity of the C-subdomain rather
than the abundant acidic residues (P. Horn 8: S. Sullivan, unpublished results).
As mentioned earlier, the mechanism of VP16 activation domain function is not
totally understood. Nonetheless, data from numerous biochemical, genetic and
biophysical studies on VP16 have provided important clues suggesting that the
N-subdomain of VP16 activation domain is able to interact with TBP, TFIIB and
TFIIH, and the C-subdomain with TAFII40 and TFHB (Ingles, et al., 1991 ; Lin, et
al., 1991; Goodrich, et al., 1993, Shen, eta1., 1996b).

Some structural studies on the activation domain of VP16 have been done.
Preliminary studies using ultraviolet spectroscopy, Fourier-transform infrared
spectroscopy, and circular dichroism done by Doug Cress using the full length
activation domain or N—subdomain of VP16, either in the form of a fusion protein
with GAL4 residues 1-147, or the peptide fragment of the activation domain
itself, suggest that the activation domain is likely to be a mixture of coil and a-
helical structure. Initial hypotheses suggesting that these acidic activators might
fold as an amphipathic (it-helix or the negative noodle model have been refuted
by mutational studies (Cress 8: Triezenberg, 1991b; Regier, et al., 1993).

Recent work using fluorescence spectroscopy techniques such as dynamic
quenching, time-resolved fluorescence decay, and time-dependent anisotropy
has been performed by Fan Shen. Chimeric proteins comprising the GAL4 DNA
binding domain residues 1-147, fused to VP16 activation domain, in which Trp
residues were substituted for Phe at either position 442 or 473 of VP16, providing
a unique fluorescence probe, were used. The results suggest that both
phenylalanine residues are well exposed and highly mobile in solution (Shen, et
al., 1996a). A more ordered structure of the activation domain could be induced

k

18
upon addition of TBP and TFIIB (Shen, et al., 1996b).

Data from both Doug Cress' mutagenesis and Fan Shen's fluorescence
spectroscopy studies suggest that the activation domain of VP16 is poorly
structured in solution and that a more structured form of the activation domain

is likely to be induced upon interaction with its target protein.
VP16 Homologues

HSV-l is a member of the alphaherpesvirinae or the a herpes virus
subfamily. The a. herpes are classified on the basis of a variable host range,
relatively short reproduction cycle, rapid spread in cell culture, efficient
destruction of infected cells and finally their ability to establish latent infections
primarily in the sensory ganglia (Roizman, 1990). Homologues of VP16 have
been identified in other a herpes viruses such as varicella-zoster virus (VZV)
(Davison, 1991; McKee, et al., 1990), bovine herpes virus type 1 (BHV-l)
(Carpenter 8: Misra, 1992), equine herpes virus type 1 and 4 (EHV-l and EHV-4)
(Purewal, et al., 1992; Purewal, et al., 1994), and the Marek's disease virus (MDV)

(Yanagida, et al., 1993; Koptidesova, et al., 1995; M. Boussaha, Ph.D. thesis, 1996).

These VP16 homologue proteins were found to activate transcription of reporter
genes linked to a gene promoters (Misra, 1994; Moriuchi, et al., 1993; Purewal, et
al., 1994; M. Boussaha, Ph.D. thesis, 1996) similarly to HSV-l VP16. This finding
indicates that all these homologues of VP16 must contain domains responsible
for targeting to the promoter and for activation. Significant sequence homology
is shared among these VP16 homologues. Attempts were made to identify the
activation domains of these proteins by comparing their primary structure with
the well characterized HSV-l VP16 protein. Unfortunately, the amino acid

sequence of these proteins varies considerably at the C-terminus. Moreover, the

19
VZV gene 10 homologue of HSV-1 VP16 lacks the 80 amino acid C-terminus

altogether. Deletion mutagenesis has been performed on the different VP16
homologues to map their activation domains. The activation domain of BHV-l
was mapped to the C-terminus of the protein, but the protein retains some
transcription activation ability despite missing the C-terminal 87 amino acid
residues (Misra, 1994). Activity of EHV-l gene 12 seems to lie at the C-terminus
as well, despite poor homology with HSV-l VP16. Deletion of the last 7 amino
acids abolishes its transcription activation ability and the methionine residue at
position 476 is likely to be a critical element for its activation function (Elliott,
1994). These findings suggest that though these proteins share significant

homology, their means of transcriptional activation might be different.

Overview

Despite considerable information emerging from mutational, biochemical
and biophysical studies on VP16 and its activation domain, very little is known
regarding the function of VP16 in viral lytic infection. VP16 plays two major
roles in viral infection, as a transcription activator for the viral IE genes and a
structural protein essential for assembly and maturation of the viral particle (Ace,
et al., 1988; Ace, McKee, Ryan, Cameron, 8: Preston, 1989; Campbell, et al., 1984;
Weinheimer, et al., 1992). VP16 has recently been found to interact with other
components in the tegument, the viral host shut off protein (Sirnibert, et al., 1994)
and VP22 (Elliott, et al., 1995), which might be important in the maturation or
disassembly of the particle during infection. Therefore it is becoming
increasingly interesting to know how different mutations in VP16 would effect
HSV-l replication within the context of the viral genome.

 

20
The significance of the transcriptional activation function of VP16 in lytic

infection was assessed by constructing recombinant viruses bearing mutant VP16
genes, truncated in their activation domains. Viruses bearing either of the two
subdomains (N or C subdomains) are weakened in lytic growth. Viruses lacking
the activation domain altogether are viable in culture, but significantly
debilitated. These results are described in chapter II.

Despite the available amino acid sequences of VP16 homologues from
other a herpes viruses, the activation domains of other VP16 homologues remain
poorly defined. Linear sequence comparison does not show any obvious
homology between the activation domain of VP16 (C-terminal 80 amino acids)
and the C-terminal portion of the other VP16 homologues. However, these
homologues were found to be able to activate transcription from their on
promoters, which indicates that they should possess a transcription activation
domain(s).

In chapter III, I have examined these sequences using a technique termed
Hydrophobic Cluster Analysis (HCA) (Lemesle-Varloot, et al., 1990; Woodcock,
Momon, 8: Henrissat, 1992). This analysis presents the amino acid sequence of
proteins in a two dimensional a helical representation. The similarity is based on
comparing conserved clusters of hydrophobic residues presumed to constitute
the hydrophobic core of globular proteins rather than the similarities in the
amino acid sequences (Lemesle-Varloot, et al., 1990; Woodcock, et al., 1992). I
propose that these herpes viruses do share similar transcription activation
domains. Interestingly, with the exception of herpes simplex virus type 2 (HSV-
2), the position of the N-subdomain in the homologue protein from BHV -1, VZV,
EHV-1, EHV-4 and MDV all reside in the N ~terminal portion of the protein. The
C-subdomain, however, still residues at the C-terminus, with the exception of

VZV gene 10 and MDV UL48 which lack the C-subdomain altogether. This

 

21
hypothesis is supported by introducing single amino acid substitutions in the

proposed activation domain of VZV gene 10, replacing the critical phenylalanine
residue at position 28 (Phe28) and other hydrophobic residues in the vicinity
thereof. This result has been published in the W31
WA (1995, Vol. 92, pp. 9333-9337). At the end of chapter

III, I have used several other comparison and prediction methods recently
become available to predict the secondary structure of the N -subdomain of VP16.
The profile network system from Heidelberg program (PHD) (Rost 8: Sander,
1994) and Sequery (Collawn, et al., 1990; Collawn, et al., 1991), were used in this
study.

 

CHAPTER II

EFFECTS OF C-TERMINAL DELETION MUTATIONS IN VP16 ON THE
LYTIC INFECTION OF HSV-l KOS

Introduction

To understand the structural and functional role of VP16 in the lytic
infection of HSV-1, the gene encoding VP16 has been identified and cloned from
the viral genome (Campbell, et al., 1984). VP16 binds to its responsive element
through interactions with at least two host cell factors including Oct-1 and HCF
(Katan, et al., 1990; Kristie, et al., 1989; Xiao 8: Capone, 1990). The DNA element
necessary for binding of VP16 and host cell factors of VP16 has been defined
(Bzik 8: Preston, 1986; Cordingley, Campbell, 8: Preston, 1983; Kristie 8:
Roizman, 1984; Mackem 8: Roizman, 1982a; Mackem 8: Roizman, 1982b). The
functional dissection of VP16 has been performed using short-term transfection
assays. In these assays plasmids encoding the native or mutated derivative of
VP16 were co-transfected into host cells together with a reporter plasmid
containing the HSV-l a promoter linked to a CAT or thimidine kinase gene (Ace,
et al., 1988; Ace, et al., 1989; Campbell, et al., 1984; Triezenberg, et al., 1988a).

Deletion mutations, linker insertions, and point mutations of VP16 have
been generated and studied (Ace, et al., 1988; Cress 8: Triezenberg, 1991b; Regier,
et al., 1993; Triezenberg, et al., 1988a; Peter Horn and Susan Sullivan unpublished

22

23
results). Based on these data, the activation domain of VP16, residues 419-490, is

divided into two subdomains: the N-subdomain, residues 410 -456; and the C-
subdomain, residues 456-490. Either of these domains is sufficient to activate
transcription in the context of a GAL4-VP16 fusion protein, but both domains are
necessary for full activation activity (Cress 8: Triezenberg, 1991b; Regier, et al.,
1993; Triezenberg, et al., 1988a; P. Horn and S. Sullivan unpublished results).
Walker et al. demonstrated that the C-subdomain of the VP16 activation domain
in the context of a GAL4-VP16 fusion protein is critical for its activity on a
promoter containing a single GAL4 binding site. Deletion of the C-subdomain or
point mutations of bulky hydrophobic residues in the C-subdomain abolish
activity of the GAL4-VP16 fusion protein. The reduction of the activity could be
complemented by using a promoter with multiple GAL4 binding sites and
duplication of the N-subdomain, but not the C-subdomain (Walker, et al., 1993).
This indicates that the N and C subdomains are not simply redundant
subdomains. The N and C subdomains seem to have distinct primary structures
and biochemical properties and may serve different roles in transcription
activation (Goodrich, et al., 1993;1ngles, et al., 1991 ; Shen, et al., 1996b; Xiao, et
al., 1994; P. Horn and S. Berger unpublished results).

Very little is known about how mutations in VP16 would affect its
function as a component of the virion and the overall lytic infection. Earlier work
demonstrated that VP16 has at least two functional roles in the virus life cycle,
activation of IE genes and a structural role as a component of the virion particles
(Ace, et al., 1988; Ace, et al., 1989; Campbell, et al., 1984; Spear 8: Roizman, 1972).
To study the effect of VP16 mutations on HSV-l infection, recombinant HSV—l
bearing mutations of the VP16 gene were generated.

Ace et al. introduced a twelve nucleotide insertion in the VP16 gene,
producing a HSV-l bearing VP16 with a four amino acid insertion at residue 379.

24
This HSV-l strain, in1814, is defective in infection and expression of IE genes

(Ace, et al., 1989). In the Ace et al. study, #11814 contains the entire VP16 coding
sequence, but bears an insertion which renders the VP16 protein defective in
complex formation on IE promoters. This was illustrated by a gel mobility shift
using a IE promoter responsive DNA sequence. The in 1814 virus was defective
in infection and produced viruses with a high particle per pfu ratio. However, at
high multiplicity of infection (100-1000 pfu/ cell) the expression of HSV-1 genes
was not affected. This was determined by measuring the level of thymidine
kinase gene expression.

Care should be taken in interpretation of these data. Though the in1814
mutant VP16 was inactive in the transient expression assay at low moi,
stimulation by wild type VP16 was only 5—10 fold in these experiments .
Residual activity of the mutant VP16 could easily be below the detectable levels
under such circumstances. Detection of mutant VP16 participating in complex
formation on IE promoters is also limited by the sensitivity of the assay. At a
high multiplicity of infection, approximately 105-106 molecules of VP16 are being
introduced into cells. In this case the cumulative effect of residual activity might
be sufficient to allow normal levels of gene expression in the in1814 infection.

Poon et al. constructed point mutations substituting various cysteine
residues with glycine residues at different positions in VP16, generating
temperature sensitive mutations of VP16 as well as recombinant viruses bearing
those mutations (Poon 8: Roizman, 1995). The temperature sensitive viruses
bearing the mutant VP16 generated by Poon et al. yielded very low titers at non-
permissive temperatures. This effect could either be due to the inability of
mutated VP16 to initiate infection by activating IE genes or a defect in the
maturation of the virus. The particle per pfu ratio of these viruses was not

determined under non-permissive conditions, but it is likely that the overall

25
production of viral particles would dramatically decrease. At non-permissive

temperatures the mutated VP16 proteins are likely to be improperly folded,
resulting in proteins that could not function in the maturation process. Results
from Poon et al. indicated that these viruses are affected late in infection implying
that these viruses are defective in maturation of the particles. The effect of these
VP16 mutations on IE gene expression was only tested on the ICP4 promoter. A
slight reduction of transcription was observed for the virus bearing mutant VP16
containing substitutions at Cyslzo and Cysl76 (Poon 8: Roizman, 1995).

An experiment where the gene encoding VP16 was deleted altogether was
also performed (Weinheimer, et al., 1992). The 8MA virus was incapable of
replicating in normal cell lines. Instead, a complementing cell line constitutively
producing VP16, Verol6-8, was used to replicate this virus. This demonstrates
that although these viruses carry the VP16 (produced from the Verol6-8 cell they
were replicating in) they cannot complete their lytic infection cycle since they
lack the capability to produce new VP16 protein during the late stages of
infection. This result reinforced the essential role of VP16 in maturation of the
viral particle and as a structural component of the virion.

Changes made in VP16 could affect either the maturation or IE activation
of the virus. To fully understand the role of VP16 in gene expression within the
context of the viral genome, we must be able to separate the role of VP16 in
transactivation of the viral IE genes from its role in maturation as a structural
component of the virus. Though Ace et al. attempted to use insertion
mutagenesis to prevent VP16 from activating IE genes, there are still concerns
about the residual activity of mutant VP16. Work by Poon et al. did not provide
any new insights concerning the role of VP16 in activation of the IE genes or
maturation of the virus, though, the temperature sensitive viruses may prove to

be very useful in future experiments.

26
Earlier work in our laboratory showed that the last 80 amino acids of VP16

are necessary and sufficient for VP16 activation of IE genes. The truncated VP16
protein, 1-410, though unable to activate transcription, can form a complex with
host proteins on the IE promoters (Triezenberg et al., 1988a; Werstuck 8: Capone,
1989a; Greaves 8: O'Hare, 1989). If the truncated VP16 could also support the
structural role of VP16 in the virion, it would be possible to produce mutant
HSV-l viruses bearing VP16 proteins that are only defective in their activation
function. These viruses could then be used to study the specific effect of IE gene
expression on replication of the virus. Moreover, the function of the different
subdomains of the VP16 activation domain could be determined by using
different VP16 deletions.

In this chapter of my thesis I have shown that recombinant viruses
carrying deletion mutations in the activation domain of VP16 could be generated,
and demonstrated that these viruses are viable in cell culture. I have used the
viral strain 8MA, in which the gene encoding VP16 has been replaced by the gene
encoding for B-galactosidase (W einheimer, et al., 1992), as a parental strain to
generate recombinant viruses via homologous recombination (Figure 2). The
8MA strain can only replicate in a complementing cell line (Verol6-8) that
constitutively expresses VP16, Verol6-8 cells (Weinheimer, et al., 1992). The
8MA system provides a powerful screen and selection for the recombinant
viruses.

I have studied the effects of a set of deletion mutations of the VP16
activation domain which lack either the N or C subdomain, or the whole
activation domain, upon HSV-l lytic infection The ability to form plaques,
plaque characteristics, single step replication, and some aspects of the effect of
mutant VP16 on structure and IE gene activation were studied.

27
Methods

Cell line and viral strains

Vero cells were grown in Dulbecco's modified Eagle's medium (GIBCO
BRL) (DMEM) supplemented with 10% Fetal calf serum unless indicated
otherwise. Verol6-8 cells (Weinheimer, et al., 1992) were grown in DMEM
supplemented with 10% Fetal calf serum and 450 ug/ ml G418.

8MA virus, kindly provided by Steve Weinheimer, was propagated on
Ver016-8 cells. HSV-l K05 and recombinant viruses were propagated on Vero
cells unless indicated otherwise. Viral stocks were prepared by infecting host
cells at 0.01 pfu / cell and harvesting cells and media after an extensive cytopathic
effect was apparent. The cells were sonicated in a TC75 tissue culture flask
(Coming) to disrupt the cell membrane and release the Virions. The
cell debris were removed by centrifugation and the media containing Virions

were aliquoted and frozen at -80° C.

Plasmids

Plasmid pKOS-VP16 (Weinheimer, et al., 1992) was digested with
restriction endonuclease EcoRI and XbaI, then re-ligated, thus removing SacI,
KpnI, SmaI, and BamHI restriction endonuclease sites in the multiple cloning
region of the plasmid. A BamI-II site was created 19 base pair downstream from
the VP16 TAG stop codon, using site directed mutagenesis with the
oligonucleotide S'CCGGACCCGGATCCCCCGT-B' (ST35). Double-stranded
oligonucleotide encoding stop codons in all three reading frame, generated by
annealing the oligonucleotide 5'-GATCCTAGTTAATTGA-3' and 5'-

28
GATCTCAAT'TAACTAG—3', was inserted at the BamI-II site, regenerating the

BamHI site at the open reading frame proximal end of the oligonucleotide,
resulting in pKOS-VP16.2 (Figure 1).

A series of pKOS-VP16.2 derivatives were constructed as follows. pKOS-
VP16.2 was digested with EcoRV and BamHI to remove the wild type VP16 gene.
Truncated VP16 gene fragments were isolated from VP16 expression plasmids,
pMSVP16A456 (VP16 with deletion of codons 456-490), pMSVP16ASmaI (VP16
with deletion of codons 413-452), and pMSVP163850N (VP16 with deletion of
codons 413-490) with EcoRV-BamI-II digestion and subcloned into pKOS—VP16.2.
Fragments were subcloned into the prepared vector resulting in pKOS-
VP16A456, pKOS-VP16ASmaI, and pKOS-VP163650N.

A derivative of pKOS-VP16.2 containing the VP16 gene from HSV-2
(VP16II) (Cress 8: Triezenberg, 1991a) was constructed. pMSVP16-2(AE11), an
expression plasmid containing the VP16II gene (A. Cress unpublished results),
was digested with XhoI, treated with T4 DNA polymerase to create blunt ends,
and digested with BamI-II. The XhoI-BamHI fragment from pMSVP16-2(AE11),
containing the VP16II gene fragment with 33 nucleotides upstream of the start
codon and 45 nucleotides from the stop codon, was then subcloned into

pKOSVP16.2 replacing the EcoRV-BamI-II VP16 gene.

Viral DNA extraction and purification

Vero cells were grown in TC120 tissue culture flasks (Coming) to 80%
confluency and inoculated with Virions at 0.01 multiplicity of infection (moi).
Cells were incubated at 37° C until 90% of the cells showed cytopathic effects
(CPE). The cells were scraped, transfered to a TC75 tissue culture flask and

sonicated, with 30 second bursts three times using a cup horn, to release the

29

 

 

 

 

Hind Ill
Sph l
Pst I
Dra III 5394
Nae I 5291 n ' 449
ern I 4842 "--.‘::~-...
Aha II 4783 3' Flanking sequence
Sea I 4725 m l-l 1074
Stop codon in 3 RF Sty I 1069
Sma I 1159
pKOS-VP16.Z Sac ' 1251
5522 b" Sal I 1288
VP16 gene
Ale I 3765 5' flanking sequence
. Kpn I 1976
Afl III 3354
Sal I-Xhol Sal l 2505
Xba I
Eco RI

 

Figure 1. Plasmid map of pKOSVP16.2. The pKOSVP16 (Weinheimer, et al.,
1992) was modified by removing restriction endonuclease sites between the
EcoRI and KM site. Mutation of a single nucleotide in the 3' untranslated region
creates a unique BamHI site 19 base pairs from the stop codon of the VP16 gene.
An oligonucleotide containing stop codons in all three reading frames was then
inserted at the BamHI site providing a stop codon for the carboxy-terminal
deletion derivatives of VP16, which were subcloned into pKOSVP16.2 replacing
the wild type VP16 gene (described in detail in the methods section).

30
virion particles. The sample was separated by centrifugation at 12K rpm in 5534

rotor, at 4° C for 30 minutes. The supernatant which contains the virion particles
was removed, overlayed on 10 ml of 20% glycerol in PBSa (137 mM NaCl, 10 mM
Nazi-IP04, 1.5 mM KHzPO4, and 2.7 mM KCl) cushion and centrifuged at 25K
rpm in SW28 rotor, at 4° C for 90 min. The supernatant was discarded and the
virion particle pellet was resuspended in 2 ml of virion lysis buffer (100 mM
NaCl, 10 mM EDTA, 50 mM Tris-HCI pH 8.0, and 0.5% lauryl sarcosinate). The
sample was treated with 250 1.11 proteinase K (2 mg/ ml) and incubated at 37° C
for 1 hour twice. The DNA was then isolated by ultracentrifugation Equal
volumes of sample and cesium trifluoroacetate (2 g / ml, CsTFATM, Pharmacia
LKB) were combined and mixed well, and 25 ul of ethidium bromide (10
mg/ml) was added. The sample was transferred into 5 ml ultracentrifuge tube,
sealed and centrifuged at 45K rpm in VTi 65.2 rotor, at 15° C overnight. The
purified DNA sample was removed, extracted with isoamyl alcohol to remove
ethidium bromide, and dialyzed in TE (10 mM Tris-HCl, 1 mM EDTA) overnight.

The DNA concentration was determined by absorbance at 260 nm.

Generating Recombinant viruses

pKOS-VP16.2, pKOS—VP16II, pKOS-VP16A456, pKOS-VPIéASmaI, and

pKOS-VP163850N were digested with EcoRI and PstI to liberate the VP16 gene
fragment (Figure 2). Vero cells were seeded in P60 tissue culture plates at 5.0x105
cells per plate. Each plate was transfected with a mixture of 5 pg of 8MA viral
genomic DNA and 5 ug of digested pKOS-VP16.2 or one of its derivatives. The
DNA mixture was transfected via CaPO. DNA co-precipitation (Graham & Eb.,
1973). The cells were then overlaid with DMEM (GIBCO, BRL) supplemented
with 5% fetal calf serum (FCS, Hyclone Laboratories) and 1% low melting-point

31

Figure 2. Fragments containing the VP16 gene and derivatives used to
generate recombinant viruses.. The viral strain and the fragments used for the
construction of the recombinant viruses are shown. The HSV-l genome and the
BamHI fragment containing the VP16 gene and the approximate location of the
VP16 gene in the HSV-l genome are shown at the top of the panel. The
fragments used to generate the recombinant viral strains RP1, RP2, RP3, RP4 and

RPS from pKOSVP16.2 (wild type VP16), pKOSVP16A456 (VP16A456-490),

pKOSVP16ASmaI(VP16A413-452), pKOSVP163550N (VP16A413-490) and
pKOSVP16II (VP16 gene from HSV-2) are shown in the lower half of the panel.
The viral strains produced are listed at the right hand side of each DNA
fragment.

 

32

HSV-1 Genomic DNA

 

 

§\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\m

 

 

 

 

 

 

 

 

 

 

ECORI EOOR
fl-Galactoeldaso
Alia T”
1 j—
VP1 a wild-type
A70 7“
all“ HI P! I
5”” VP16.2(1-490 wt)
Am
I
F _IH
M HI Pd I
5”” vmomse-«o
Am
I
F J\)_
Bani HI Pd I
5”” mums-452
A70
I
F W
all! HI P. I
5”” means-490
m 1'“
Soon: ",1” aunt-II Pstl

Figure 2

. n
................
av -. . . .-. . .~.an:v...-.'.-.-.-.-:-:'3-.~f‘.'2-3-:- ..............

..........................................

8MA

HSV-1 KOS

RP1

RP3

RP4

RPS

RP2

3
agarose (Sea Plaque, FMC), and incubated at 37° C for 3-4 days or until plaques

formed. At least three plaques from each plate were picked and subsequently
plaque purified twice. DNA preparations of each were analyzed by Southern
blot to verify that the isolated viruses contain the proper form of the VP16 gene.
Primary stocks were prepared and titered to determine the plaque forming unit
(pfu)-

Revertant virus strains were generated by co-transfecting Vero cells with
DNA from each mutant virus and a wild type VP16 DNA fragment from pKOS-
VP16.2. Three plates were transfected for each DNA mixture. Plates were
overlaid with DMEM supplemented with 5% FCS without agarose. After
plaques formed (3-4 days) 500 ul of media was removed from each plate, diluted
100 fold, and inoculated on Vero cells at 80-90% confluency. These plates were
overlaid with DMEM supplemented with 5% FCS and 1% agarose and incubated
at 37° C. Plaques which showed a revertant phenotype (plaque formation within
2 days and large plaque size) were picked and plaque purified twice. The
presence of the wild type VP16 gene in the revertants was verified by Southern
bolt analysis. Viral stocks were generated and titered.

Southern blot

Recombinant viral DNA (approximately 1 ug) was digested to completion
with BamHI and separated by gel electrophoresis in 1% agarose using TAB
buffer. The DNA was stained with ethidium bromide, denatured, and transfered
to nitrocellulose membrane using standard methods. The EcoRI-Pstl VP16 gene
fragment from pKOS-VP16.2 was used as a probe. Random primed DNA
labeling kit (Boehringer Mannheim Biochemica) and a-PZP] dCTP (Amersham
Life Science, 3000Ci/mmol) was used. The nitrocellulose blot was pre-

34
hybridized for two hours in 5X Denhardt's solution (Sambrook, et al., 1989), 6X

SSC (Sambrook, et al., 1989), 0.5% SDS, and 100 ug/ ml salmon sperm DNA at 65°
C. The synthesized probe was added and incubated overnight at 65° C. The blot
was subsequently washed twice in 2X SSC, 0.5% SDS at 67° C, twice in 0.5x SSC,
0.5% SDS at 67° C, and twice in 0.2x SSC, 0.5% SDS at 67° C or until the
background was eliminated. After the blot was exposed to X-ray (X-OMAT, AR,
Kodak) film for 3—4 hours, the film was developed using standard methods.

Western blot

Virions were inoculated at 0.1 moi on Vero or Ver016-8 cells grown in
TC120 tissue culture flasks to 80-90% confluency. The infected cells were
incubated until 90% of the cells showed CPE. The Virions were harvested by
sonication and centrifugation as described above. Intact Virions were purified by
centrifuging the particles through a 20% glycerol PBSa cushion at 25K rpm in a
SW28 rotor, at 4° C for 90 minutes. Pelleted Virions were resuspended in 100 pl
PBSa. 10 pl of each purified virion sample was solubilized in SDS sample
loading buffer by heating the samples to 100° C for 10 minutes, and separated by
SDS PAGE in a 12% gel. The proteins were transferred to a nitrocellulose
membrane using standard protocols. Western blot analysis was performed using
the biotin/ avidin system, Vectastain® ABC kit, Vector laboratories, and the anti-
VP16 polyclonal antibody, C-8.

Single step replication assay

Vero cells were seeded into a six-well tissue culture plate at 3.5x105 cells
per well and incubated overnight at 37° C. Each plate was washed with DMEM

35
without fetal calf serum. The number of cells per plate was determined, and the

cells were subsequently infected with 5-10 pfu/ cell of RP1, RPS or RP4 and
overlayed with 3 mls of DMEM supplemented with 5% PCS. Infected cells from
each well were scraped and collected from each plate. At 4, 8, 12, 18, 24, and 36
hours post infection (hpi.) the infected cells were scraped and transfered into 15
ml tubes, and frozen at ~70° C for later titering. The infected cells were thawed
and sonicated with three 30 second bursts to release the virion particles. The titer
of each sample was then determined on Vero cells. The titer of the virus from
each strain was plotted against time, hpi.. The single step replication curve of

each strain was determined at least three times.

Penetration curve

Each strain of recombinant virus, RP1, RP3, RP4 and RPS, was inoculated,
100 pfu per plate, on fifteen IP60 plates of Vero cells at approximately 90%
confluency. Three plates from each strain were removed at 15 minutes intervals
for a period of 1 hour and washed once with 2 mls of DMEM without fetal calf
serum followed by 2 mls of Glycine saline buffer (8 g of NaCl, 0.38 g of KCl, 0.10
g of MgClz ~6H20, 0.10 g of CaClz'2H20 and 7.5 g of glycine per liter, pH adjusted
to 3.0), followed by two subsequent washes with DMEM without PCS. The
plates were overlayed with DMEM supplemented with 5% PCS and 1% low
melting point agarose, and incubated until plaques were formed. The agarose
was then removed from each plate and the cells were stained with 1% methylene
blue in 70% isopropanol. The number of plaques in each plate was counted,
recorded and plotted against the time allowed for absorption.

36
Determination of particle per pfu/ml ratio

Virus stocks were generated by inoculating Vero cells, approximately 80%
confluent, in TC120 tissue culture flasks at 0.01 pfu/ cell. The inoculum was
removed after 1 hour. Cells were washed with DMEM without FCS then
overlayed with DMEM supplemented with 2% FCS. The cells were incubated at
37° C until 90% of the cells show CPE. The cells were scraped, spun down, and
resuspended in 10 mls of DMEM. The cell suspension was sonicated to release
the virion particles. Cell debris were removed by centrifugation at 4,000 rpm in a
table top centrifuge. The supernatant was frozen in 1 ml aliquots and used for
titering and particle count. The number of particles per ml was determined by
transmission electron microscopy, using standard latex bead (Latex spheres of
known concentration, 0.176 um, Ernest F. Fullam, Inc.), and negative staining
with 1% ammonium molybdate. The number of latex beads and Virions from 20
viewing screen were recorded from each quarter of the grid (total of 80
samplings per stock). The numbers of latex beads and Virions were used to
calculate the number of virion particles in each viral stock. The titer and number
of particles in each stock was then used to calculate the particle per plaque

forming unit ratio.

Southern blot of infected cell nuclei

Vero cells in P100 plates were infected, in triplicate, with the recombinant
viruses. Cells were infected with equivalent amounts of particle, equal to the
amount of particles present in 10 moi of wild type RP] strain. At 3 hours post
infection, nuclei of the infected cells were harvested. The cells were washed once
with 5 mls PBSa, removed from the plates by adding 5 ml of ice cold RBS .

37
solution (10 mM NaCl, 10 mM Tris-HCl pH 7.4, and 3 mM MgC12) supplemented

with 0.1% (v/v) Nonidet P-40 and harvested by scraping the cells with a rubber
policeman. The cell suspension was collected and homogenized, 3 strokes, in a
ice-chilled glass homogenizer. The samples were transferred into 10 ml tubes
and chilled on ice. Nuclei were collected by centrifugation in a table top
centrifuge at 1,000 rpm for 10 minutes. The integrity of the nuclei was checked
using a phase-contrast microsc0pe.

The nuclei were then resuspended in 500 pl of a buffer containing 100 mM
Tris-HCl, 10 mM EDTA, and 0.1% SDS, and were treated with proteinase K (100
pg per ml, pH 7.5) at 37° C for 2 hours. The sample was treated with 50 ul of 3 M
sodium acetate pH 5.3 and extracted twice with phenolzchloroform (1:1)
saturated with TB. The aqueous phase was transferred to a new tube and 2.5
volumes of ethanol were added. The DNA was then precipitated by
centrifugation and resuspended in 100 pl of TE. Approximately 1 ug of DNA (25
ul) of each sample was cut with 5 units of BamI-II and separated on a 1% agarose
gel and subjected to Southern blot analysis, probing with the EcoRI-Pstl fragment
of pKOSVP16.2 containing the VP16 gene.

Results
Isolation and identification of recombinant HSV—l

Homologous recombination using 8MA virus in Vero cells was performed
to generate recombinant HSV-l. Plaques were isolated and analyzed by

Southern blot analysis to verify that the recombinant viruses had acquired the
appropriate wild type or mutated VP16 gene.

38
The wild type HSV-l KOS parental strain cut with Baml-II and probed

with the VP16 gene fragment produces a single 8.5 kbp band. In contrast the
recombinant viral DNA cut with BamI-II and probed in the same manner would
yield two distinct bands, since all the recombinant viruses contain an additional
BamHI site within the fragment containing the VP16 gene.

Two distinct bands were observed when the BamHI digested recombinant
viral DNA was probed with the EcoRI-Pstl VP16 gene fragment from
pKOSVP16.2 (one of approximately 4.5 kbp and another of 4.0-3.7 kbp band
reflecting the different deletion mutations in the VP16 gene) replacing the single
11.8 kbp fragment of 8MA parental strain, as shown in Figure 3. The slightly
larger size of the BamI-II fragment from 8MA is due to the larger size of the lad
gene compared to the VP16 gene in HSV-l K05.

Because the difference in size of the fragments from strains RH and RP4
was not clearly resolved, a second Southern blot was performed using SacI
digestion (data not shown). This method clearly showed that RP4 contains the
appropriate VP16 mutation.

Five recombinant viral strains were isolated. RP] and RP2 express wild
type VP16 from HSV-l and HSV-2, respectively. Viral strains RP3, RP4 and RP5
express truncated mutant VP16 with deletions of residues 456-490, residues 413-
452, and residues 413-490, respectively.

To confirm that these recombinant viruses express and contain the mutant
VP16 molecules in their particle, purified Virions were subjected to Western blot
analysis. Polyclonal antibody against VP16 was used to probe for VP16. An
approximately 65 kD polypeptide corresponding to full length VP16 protein was
detected for the control HSV-l KOS as well as RH and RP2 (Figure 4). RP3, RP4
and RP5 particles contained a slightly faster migrating VP16 protein as expected
from the truncated genes.

39

HSV-1 KOS
HSV-1 8MA
Malker DNA

 

\- (‘0 st to a
Q. 0.. 0. Q
Kbp a: a: a: n: m
23.13 _ _ (p
9.42 — m
6 56 .
' u-
n
w “I “
2.32 —
2.03 — .m.

 

 

 

Figure 3. Southern blot analysis of recombinant viral DNA. The recombinant
viruses were isolated, purified, their DNA extracted, and analde by Southern
blot analysis to confirm that they have acquired the proper VP16 gene construct.
The viral DNA was cut with BamHI to completion and separated on a 1% agarose
gel. The DNA was then transferred to a nitrocellulose membrane and probed for
the VP16 gene as described in detail in the methods section

40
Southern and Western blot analyses clearly confirm that each of the

recombinant viruses contain and express the appropriate mutant VP16 gene and
protein in their particles. These results indicate that virus strains carrying a
deletion mutation in the activation domain of VP16 could be generated. The
complementing cell line was not required for generation or isolation of the
viruses, suggesting that the truncated VP16 was sufficient to support maturation
and the structural role of VP16 in the virion particle.

Another interesting question is whether or not the virus has a preference
for packaging of wild type over truncated versions of VP16 into the particles. If
the truncated VP16 was unable to fully function in particle assembly or
maturation, we might see some preference for packaging the wild type protein
over the truncated form. I have chosen RP5 for this experiment since it lacks the
entire 80 amino acid portion of the activation domain. If there is any defect in the
packaging of truncated VP16 into the particle at all, RP5 would be the most likely
candidate to reveal the defect.

A Western blot of RP5 virions produced in Vero 16-8 cells was performed.
Two bands of VP16 were observed, one migrating similarly to wild type VP16
protein, and the other at a lower molecular weight, similar to the truncated VP16
protein carried by RP5 particles (Figure 4). This result demonstrates that when
RP5 was grown in Verol6-8 cells, the virions contain comparable amounts of
both the wild type and the truncated versions of VP16.

To be confident that the signals from VP16 protein seen on the Western
blot reflects the VP16 proteins that were packaged in the virion, not
contamination from host cells, RP5 particles produced from Verol6-8 cells were
also purified from the media and subjected to Western blot analysis. This would
eliminate any contamination of the wild type VP16 protein from the host cells.
When the signals from the Western blot were compared between both virion

41

 

fig.
i
§§ it?
E EWHHMI
97—

68- ‘ “_.-

 

 

 

Figure 4. Western blot analysis of recombinant viral particles. Viral particles
were purified from recombinant viral strains RP1, RP2, RP3, RP4 and RP5. RP5
was grown both in Vero and Verol6-8 cells to study the packaging preference of
full length versus truncated VP16 protein into the viral particle, as described in
the methods section. The viral particles were subjected to SDS PAGE and
Western blot analysis. The results demonstrate that the recombinant viral
particles contain the proper form of VP16.

42
preparations, similar ratios of the wild type and truncated form of VP16 were

found (Figure 4).

To further confirm that the ratio of the signals observed from the Western
blot reflects the accurate ratio of the two forms of VP16 protein, lower amounts
(10 5, and 2.5 ul) of purified virions were loaded on the SDS PAGE, 10, 5 and 2.5
ul, for Western blot analysis, as shown in Figure 5. The results demonstrated that
the signal we observed on the previous gel was in linear range, reflecting the
accurate ratio of the two proteins in the particles.

These results demonstrated that there is no preference for packaging of
the wild type VP16 protein over the truncated form in the process of virion
assembly. They also suggest that the truncated form of VP16 is sufficient for
viral maturation and structural requirement. Therefore any effects of these
mutations on viral lytic infection should reflect defects in early events of
infection.

The VP16 homologue from HSV-2 was previously characterized in our
laboratory. Since VP16 from HSV-l and HSV-2 share very high homology, it is
interesting whether or not VP16 from HSV-2 can fully complement HSV-l
growth. To answer this question, I have recombined the gene encoding VP16
from HSV-2 (VP16II) into 8MA, resulting in viral strain RP2. RP2 forms plaques
within two days of inoculation and gives rise to similar size plaques as RP1.
Slightly different plaque characteristics can be observed under the phase-contrast
microscope in cells infected with RP2. The cells in RP2 plaques do not show as
extensive CPE as in RPl plaques. These data demonstrated that VP16II can
complement HSV-l for growth on cell culture. This indicates that VP16II shares
similar functional properties to VP16 of HSV-1 and that these functions are
conserved between HSV-l and HSV-2.

 

 

M 8 2
1o_ _5 .2-2. 5 1‘. g
RP5/16—8free particle + + + i 5', E
RP5/16-8 cell asso. + + + ‘1 I kd
— 97
— 68
I ”uw”"""’ ’
..... a... - ‘— at";
\. hmv.‘”M,__w My .,. , M _
~ 29

 

 

 

Figure 5. Western blot analysis of RP5 particles generated in Verol6-8 cells.
RP5 particles were purified either from the media (free particle) or cell associated
viruses (cell asso.) as described in the methods section. The purified particles
were then subjected to SDS PAGE followed by Western blot analysis. The
number above the bar indicates the volume (in microliters) of sample loaded in
the lanes below. The ”+” indicates the sample loaded, ”cell asso.” or ”cell free”.

44
Plaque and growth characteristics of recombinant HSV-l

We expected that recombinant viruses carrying the different truncated
derivatives of the VP16 activation domain would demonstrate different growth
characteristics, since activation of the viral IE genes is one of the most important
steps in the replication of HSV-1. RPl which bears wild type VP16 should
replicate similarly to the wild type virus, HSV-l KOS. RP3 and RP4 should
replicate slightly slower since they contain only one of the activation
subdomains. I expected RP3 to grow slightly better than RP4 since it contains the
N-subdomain which was shown to be a stronger activator than the C-subdomain,
in mammalian cells (Walker, Greaves, & O'Hare, 1993). I expected RP5 to grow
very slowly since RP5 carries the VP16 gene that lacks the entire activation
domain, a mutant which should abolish activation of IE genes entirely.

These recombinant viral strains produced different plaque characteristics
on Vero cells. RP1, RP3 and RP4 formed visible plaques within 2-3 days post
infection, varying in size from the larger RPl plaques to smaller plaques for RP3
and RP4. RP5 did not form visible plaques until 5-7 post infection and the
plaques were still very small compared to the other recombinant viruses, as
shown in Figure 6.

The titer of the virus is dramatically affected by the VP16 mutations. For
RP4 the titer was reduced approximately 10 fold and RP5 approximately 100
fold, as shown in Table 1. To determine whether the lower titers of RP4 and RP5
are due to inability to produce intact particles, the particles per pfu ratio of each
strain was determined. The results in Table 1 suggest that RP4 and RP5 were
able to produce as many particles as RP] and RP3. However the particle per pfu
ratio of RP4 and RP5 were approximately 36 fold and 120 fold higher,
respectively, than that of RH and RP3. This suggested that the low titers

45

Figure 6. Plaque characteristics of RP1, RP3, RP4, and RP5 on Vero cells.
Viruses were inoculated on Vero cells to perform a standard plaque assay, as
described in the methods section. The plates were incubated for 3 days at 37° C
then the agarose was removed and the plates stained with 1% methylene blue. A
set of RP5-infected cell plates was also allowed to incubate further to 5 days. The
plaque characteristics of RP1, RP3, RP4 at 3 days post infection and RP5 at 3 days
and 5 days post infection are shown

46

RP1 RP3

RP5

 

RP5
(5 days post infection)

 

Figure 6

47
observed for RP4 and RP5 were a result of a failure of the particles to infect cells

productively.

The phenotypes observed for RP3, RP4 and RP5 could result from other
mutations inadvertently introduced into the HSV-l genome during the
recombination event or subsequent growth of these viruses. Thus, the
phenotypes observed could result from a cumulative effect of more than one
mutation. To verify that the phenotype observed was only caused by the
truncation in the VP16 gene, I generated revertants by introducing the wild type
VP16 gene back into RP3, RP4 and RP5. If the defective phenotype was only due
to the truncation in VP16, these viruses should exhibit the wild type phenotype.
On the other hand, if the virus has picked up other mutations, they should
continue to show the defective phenotype.

Wild type VP16 gene was recombined back into RP3, RP4 and RP5 to
create revertant strains, RP3R, RP4R and RP5R. Southern blot analysis was
performed to verify that the revertants acquired the wild type VP16 gene (Figure
7). Each revertant gave a single band when probed with the VP16 gene fragment,
demonstrating that all the revertants have acquired the wild type VP16 gene.
Titer, particle per pfu ratio and plaque characteristics (data not shown) of RP4R
and RPSR were determined (Table 1). RP3 had a similar particle per pfu ratio as
' wild type, therefore, the particle per pfu of RP3R was not determined. All
revertants exhibited the phenotype of wild type viruses which demonstrates that
the defective phenotype of the parental recombinant viruses, RP3, RP4 and RP5,
were due to the truncations in the VP16 activation domain.

To further characterize these viruses, single step replication kinetics of
RP1, RP3 and RP4 were determined. The single step replication curve of RP5
could not be determined because the titer of the virus was too low for this
experiment. Figure 8 shows that within the first 4 hours of infection the infected

Table 1
Particle count and particle per pfu ratio

 

 

 

 

 

 

 

 

Recombinant Titer of stock Number of particle / pfu
HSV-l Strain (pfu/ ml) particles per ml Ratio
RP1 3.9x108 3.4x1o9 8.6
RP3 5.8x103 5.3x109 9.1
RP4 3.3x107 1.2.1010 360
RP5 3.1x1o6 3.6x109 1200
RP4R 1.1x109 6.1x109 5.7
RPSR 1.1x109 6.3x109 5.5

 

 

 

 

 

Table 1. Particle count and particle per pfu ratio. Recombinant virus stocks
were generated, titered and intact particles in each stock were counted as
described in the methods section The particle per pfu ratio of each stock was

- calculated. These results demonstrated that all the recombinant viruses were
able to produce comparative amount of intact particle despite their low ability to
initiate viral infection.

49

 

       

     

5
5 E2”
m I E? m a: a: a or
23.13 — w
9.42—
6.56—
4.37— I... ” m “
2.32— m
2.03— m

 

 

 

Figure 7. Southern blot analysis of revertant viral strains. The purified viral
DNA from RP3R, RP4R and RPSR was subjected to Southern blot analysis. The
purified DNA was digested with BamI-II and probed with the Pst-EcoRI VP16
gene fragment from pKOSVP16.2, described previously. Viral DNA from RP3,
RP4 and RP5 was used as a control in this experiment.

50
cells yielded 105-106 pfu/ ml for all three strains tested. RP1 and RP3 titers rose

to approximately 107 pfu/ ml at 16 hpi.. The RP4 titer was slightly lower than
that of RP1 and RP3. The titer of RP4 was about 5x106 pfu/ ml at 16 hpi.. These
results suggested that there might be differences in the growth curves, but there
was a very high initial titer of the recombinant viruses at 4 hpi.. It was surprising
to see such high titers at 4 hpi., since it was unlikely that any new particles were
produced by that time. I reasoned that there might be viruses which adsorbed to
the cells but did not penetrate and infect the cells at the 4 hour time point and
would then have been carried along when the cells and viruses were collected.
This would result in the high titer observed at 4 hpi.. These viruses would mask
differences in the number of progeny viruses being produced at early as well as
at later time points during infection. Viruses that have penetrated the host cell
loose their envelope and cannot reinfect.

To resolve this problem I have applied an additional washing step with
glycine-saline buffer pH 3.0. This removes the cell-adsorbed viruses which have
not yet penetrated into the host cells. First, I had to determine whether or not
there would be a sufficient amount of virus entering the host cell to start the
infection once the cells have been washed. A viral penetration assay was
performed where cells were inoculated with 100 pfu per plate of recombinant or
wild type virus, and subsequently washed with the glycine-saline buffer pH 3.0
at different time intervals. The number of plaques on each set of plates at
different times was determined and compared with the unwashed plates, with
the results shown in Figure 9. At the end of the first hour approximately 50
percent of the RP1, RP3 and RP4 particles have absorbed and were able to start
infection Therefore in the single step replication experiments where viruses
were inoculated at 5-10 pfu per cell, there should be at least 2—5 pfu per cell after
the wash.

51

 

Titer (piulml)

—O— RP1
MD-.. RP3

-—A—— RP4

102

 

‘0 I I i'vrTrTrv'v'I1I'

O 4 8121620242832 3640
TImo(hpI.)

Figure 8. Single step growth curve of recombinant virus strains RP1, RP3, and
RP4 (Standard condition). Vero cells were inoculated with recombinant virus
RP1, RP3 or RP4 at 5-10 moi. At 4,8,12,16,24 and 36 hpi, the cells were harvested
and the titer at each time point was determined and plotted against time, hpi., as
described in the methods section

52
Single step replication experiments with the additional wash with glycine-

saline pH 3.0 after absorption were then performed, as shown in Figure 10. The
titer at 4 hpi. dropped dramatically and a difference between RP1 and RP4 could
be seen clearly. Despite the difference in the plaque forming characteristic there
seems to be no difference in the single step replication curve of RP1 and RP3. In
contrast, RP4 replicated poorly and did not produce an outburst of the particles
at 8-12 hpi., as seen for RP1 and RP3, but rather replicated slowly throughout the

course of infection.

Recombinant HSV-l are defective in infection

Together, the plaque characteristics, particle per pfu ratio and single step
replication experiments show that RP3, RP4 and RP5 are defective in the overall
viral infection Although there were no significant differences in the single step
replication experiment or the particle per pfu of RP3, the smaller plaque size of
RP3 is clearly distinguishable from RP1.

To roughly determine the steps during infection that might be defective in
these viruses, I used Verol6-8 cells which constitutively express wild type VP16
protein If the viruses are defective in the activation of their IE genes, Verol6-8
cells should rescue their defective phenotype by providing the wild type protein
which could function in IE activation.

The results in Figure 11 indicate that when RP3, RP4 and RP5 were titered
on Verol6-8 cells, VP16 present in the Verol6-8 cells did not complement the
infection of RP4 and RP5. A significantly higher number of plaques was not
found on Verol6-8 compared to Vero cells (panel A). On the other hand, RP5
produced in Verol6-8 cells gave a titer as high as the RP1 strain (panel B).

Further more, the plaque characteristics of all the recombinant viruses strains are

53

 

60q e

SOq
.5 . -
E 40-
‘6 . .
5i '-{1- R94
0. 30- ‘ —°— RP‘3
E - —a— RN
8 --o-- RP—S
a? 20-

10-

0" I T fl ' I t V Tfi I

 

10 20 3O 40 SO 60 70

Figure 9. Penetration curve. Vero cells in p60 tissue culture plates were
inoculated at 100 pfu per plate. At 15, 30, 45, and 60 minutes after inoculation the
plates were washed with glycine-saline, pH 3.0, solution to remove the viruses
that have adsorbed but not penetrated into the host cells. The plates were
overlaid with media containing agarose and incubated until plaques formed. The
number of plaques formed on the plates (expressed as a percentage compared to
unwashed plates) was plotted against the time allowed for penetration, as
described in the methods section

   

Titer (piulmi)

—O— RP1

-.-g-.. RP3
—n— RP4

 

 

1
1O 'I'I'l'l'l'l'lfiI'I'I

0 4 8 12 16 20242832 3640

Time (hpi.)

Figure 10. Single step growth curve of recombinant virus strains RP1, RP3 and
RP4 (Acid wash condition). Recombinant viruses, RP1, RP3, and RP4 were
inoculated on Vero cells in a 6-well plate at 5-10 moi. After the viruses were
allowed to adsorb and penetrate for 1 hr. the cells were washed with glycine-
saline buffer pH 3.0 to remove the excess viral particles. The infection was
allowed to proceed and the cells were collected at different time points and
assayed for the presence of infectious virus as described in Figure 7.

55
similar on Verol6-8 cells. The similar plaque characteristics of all the viral strains

on Verol6-8 cells suggests that the progeny viruses produced were able to spread
and infect with similar kinetics. These results indicate that wild type VP16 can
rescue the defective strains only when the VP16 protein is present in the particle,
not when VP16 is supplied by the infected host cell.

The absence of the whole activation domain or one of the subdomains
might also affect the structure of the virion in such a way that the uncoating or
migration of the virus is blocked. Therefore the growth defect we observed
could result from problems in these steps, not from a defect in transcription
activation. To test this hypothesis I examined how well viral DNA reaches the
cell nucleus. Equal amounts of particles of the recombinant viruses, equivalent
to 10 pfu/ cell of RP1, were used to infect Vero cells. At 3 hpi. the nuclei were
harvested, DNA was extracted and subjected to Southern blot analysis.

Figure 12 shows the Southern blot of DNA extract from nuclei of cells
infected with the recombinant viruses. The results clearly indicate that RP3, RP4,
and RP5 can deliver their DNA into the host cell nucleus within the first 3 hours
of infection. This demonstrated that RP3, RP4 and RP5 do not have any
detectable defect in their ability to deliver their DNA to the host nucleus. These
results imply that the steps prior to DNA delivery, such as adsorption, .
penetration, uncoating and migration of the particle, should also be normal in

these recombinant viruses.

56

A. 109
D Titer on Vero cells
Tlteron16-8 cells
E 108
5
g 107
10°

 

 

 

 

HP‘I RP3 RP4 RPS

Viral Strain

D Titer on Vere cell
Titer on 16-8 oelis

Titer (pfu/mi)

 

 

Cell line used to generate RP5 stock

Figure 11. Titer of recombinant viral stocks in Vero and Vera 16-8.
Recombinant viral stocks RP1, RP3, RP4 were generated in Vero cells, and RP5
was generated in Vero and Vera 16-8. (A) The titer of viral stocks titered on Vero
or Vero 16-8 are compared. (B) The titer of RP5 stocks generated in Vero or Vero
16-8 cells was determined on Vero and Vero 16-8 cells.

57

 

 

 

5 infected cell
1‘! Virion DNA nuclear DNA
gr II I
v- 0‘) v 10 v- v l0
5 EL 0. 0. 0. E 0. 0.
Kim a m cr 0: n: 3’ n: n: a:
9.42 — H
6.56 — .
use“...
4.37 — 'llll
n...
2.32 - Q
2.03 — ‘

 

 

 

 

Figure 12. Southern blot analysis of nuclear localized recombinant HSV-l
DNA. Vero cells were infected with an equal number of particles of RP1, RP3,
RP4, and RP5. At 3 hpi. the nuclei from the infected cells were harvested and
DNA extracted. The DNA was digested with BamI-II to completion and Southern
blot analysis was performed, as described in the methods section. The purified
DNA from virions was used as a positive control in this experiment.

58
Discussion

Previous work suggested that the VP16 gene is essential for viral
replication (Weinheimer, et al., 1992). Mutation of the VP16 gene within the
context of the viral genome results in defective particles (Ace, et al., 1989; Poon 8:
Roizman, 1995). Attempts have been made without success to generate deletion
mutations of the VP16 gene within the context of the viral genome (Ace, et al.,
1989).

I have demonstrated that viruses with deletions in the VP16 activation
domain can be generated and are viable in cell culture. Five viral strains were
generated: RP1, RP2, RP3, RP4, and RP5. These derivatives of HSV-1 contain the
full length VP16 gene, VP16 from HSV-2, and VP16 with a deletion of amino acid
residue 456-490, residues 413-452, and residues 413-490, respectively. Similar
frequencies of generation of RP1 (recombining wild type VP16) and RP2, RP3,
RP4 and RP5 were observed (data not shown), indicating that introduction of
HSV-2 VP16 or the truncated VP16 genes is not a rare recombination event. A
complementing cell line, Verol6-8, was not necessary for either the construction
or isolation of these recombinant viruses. Moreover, Western blot analysis
demonstrates that the appropriate derivative of VP16 is produced and assembled
into the virion in each of the recombinant viral strains. These results suggested
that truncated VP16 was able to support viral maturation and produce infectious
virions.

Recent studies have shown that VP16 interacts with other viral proteins
present in the tegument such as the viral host shutoff protein (Smibert, et al.,
1994) and VP22 (Elliott, et al., 1995). VP16 interacts with VP22 via its activation
domain, and deletions and point mutations in the VP16 activation domain
greatly affect the interaction between VP22 and VP16. Mutations in the

59
activation domain resulted in alteration of the localization of the two proteins in

the host cell during viral maturation (Elliott, et al., 1995). Our results suggest
that this interaction must not be essential for viral maturation or particle
assembly, since recombinant viruses bearing deletions of the activation domain
were able to grow and produce intact particles.

Though viable, these viruses exhibit some defects in replication RP1
(carrying the full length VP16 gene) has an identical phenotype to wild type
HSV-l KOS. RP3 produced significantly smaller plaques when compared to RP1,
though single step replication and particle per pfu ratio were similar to that of
RP1. The smaller plaques of RP3 could reflect the cumulative effect of slightly
slower replication and cell to cell spreading of the virus observable only upon
multiple rounds of infection RP4 and RP5 are greatly defective in their ability to
form plaques. Both gave higher particle per pfu ratio and slower replication as
exhibited by their plaque characteristics. Single step replication kinetics studies
also demonstrated that RP4 replicates much slower than RP1 and RP3 (not
determined for RP5 ).

Particle count by electron microscopy demonstrated that these viruses are
normal in their ability to produce intact particles. The next reasonable step to
take was to test whether or not their phenotype was caused by the loss of
transcription activation activity of truncated VP16. Therefore, the ability of
Verol6-8 cells to rescue the phenotype of these recombinant viruses was tested.
Verol6-8 cells are expected to rescue the defects in IE gene expression, by
providing wild type VP16 in trans for the incoming virus. We would expect RP4
and RP5 to give a comparable titer with RP1 when titered on Verol6-8 cells.
Results from Friedman et al. demonstrated that a cell line constitutively
expressing a truncated form of VP16 which lacks the C-terminal 78 amino acid
was shown to be resistant to HSV-l infection (Friedman, et al., 1988). This result

60
clearly demonstrates that a VP16 derivative provided in trans by the host cell can

interact with incoming virus during early infection.

However, our results show no significant differences in the titer of RP4
and RP5 on Ver016-8 compared to Vero cells (Figure 11). There were no
differences in the plaque characteristics of RP1, RP3, RP4 and RP5 on Verol6-8
cells (data not shown), indicating that Verol6-8 cells were able to complement
the growth defect of the progeny viruses. This was also observed when RP5 was
grown in Verol6-8 cells. We observed almost a 1000 fold increase in the titer of
the RP5 stock produced in Verol6-8 versus Vero cells, shown in Figure 118. The
presence of wild type VP16 protein in the RP5 particle was able to complement
the low titer phenotype.

The failure of Verol6-8 cells to complement these viruses in plaque assays
could result from an inadequate level of VP16 in these cells early in infection.
Western blot analysis using a whole cell extract from Verol6-8 cells did not
detect VP16 protein, suggesting that there is not a high level of VP16 protein
expressed in this cell line (data not shown). It could also be the result of an
aberrant localization of the VP16 in Verol6-8 cells or different post-translational
modification of VP16 in Verol6-8 cells. The localization of VP16 could be
verified by indirect immuno fluorescence, and the mobility of VP16 protein from
the virion and Vero16-8 cells could be compared by isoelectric focusing to test for
differences in phosphorylation states.

The above considerations did not rule out the possibility that Verol6-8 can
complement for the activation of IE transcription, but that the recombinant
viruses failed to penetrate, uncoat and deliver the viral DNA properly into the
host cell nucleus.

I have shown by Southern blot analysis of the nuclear localized viral DNA
that RP3, RP4 and RP5 are not defective in their ability to deliver DNA into the

61
host nucleus. This implies that the absorption, penetration and uncoating of

these viruses is normal or not greatly effected by the mutations in VP16. Since
the viral DNA was delivered into the host nucleus, the failure of RP4 and RP5 to
initiate infection properly is most likely a result of failure of truncated VP16 to
activate transcription of IE genes.

Preliminary experiments in our laboratory done recently by Eugene
Chung, repeating the titering experiment of RP4 and RP5 using a BHK3378#2 cell
line constitutively expressing VP16 protein (kindly provided by Dr. Paul
Hippenmeyer, Monsanto) demonstrated that RP4 and RP5 can be complemented
by this new cell line. These results confirm that the defective phenotype
observed in RP3, RP4 and RP5 is most likely to be a result of aberrant IE gene
expression

The reason why Verol6-8 cannot complement RP4 and RP5 while
BHI<3378#2 can is still obscure. The expression level of VP16 protein in both cell
lines is low and cannot be detected by Western blot analysis (data not shown).
The differences between these two cell lines could be a result of different
localization or modification of the VP16 protein.

Eugene has also shown by transient expression assay that RP4 and RP5
were unable to activate transcription from a reporter plasmid containing the
HSV-l ICP4 promoter. Curiously, RP3 was able to activate transcription from
the reporter plasmid similarly or even better than RP1, which carries the wild
type VP16. Preliminary quantitative reverse transcriptase polymerase chain
reaction (RTPCR) experiments by Eugene, on the IE genes, ICPO, ICP4, ICPZZ and
ICPZ7 also demonstrated that RP4 and RP5 are defective in their ability to
activate these genes. These results support the hypothesis that RP4 and RP5 are
defective in the activation of IE genes which could greatly contribute to the
growth defect.

62
We have clearly demonstrated that the activation domain of VP16 is

necessary for normal viral growth. The recombinant viruses were able to
produce intact particles, but the particles are defective in infection as shown by
the high particle per pfu ratio. The migration of the viral DNA into the nucleus
of the host cell is normal, suggesting that adsorption, penetration, uncoating and
migration of the viral particle are normal for RP3, RP4 and RP5. A transient
expression assay using the ICP4 promoter as well as RT'PCR results
demonstrated that these viruses are defective in IE gene expression. Therefore,
the growth defects observed in these recombinant viruses are most likely to be a
result of the defect in IE gene activation and overall gene expression Our results
also suggest that the interaction between VP22 and VP16 during maturation is

not essential for viral replication.

Future studies

Since we have recombinant viruses bearing the different deletions in the
VP16 protein we should be able to measure the activity of these mutant VP16
proteins in the context of the viral genome. IE gene mRN A products could be
measured and the activity of the derivative VP16 proteins could be determined.
Transient expression assays performed by transfecting reporter plasmids, with a
reporter gene linked to IE promoters or their derivatives, followed by infection
with these recombinant viruses would also be interesting. It could give us
information on activation by virion delivered VP16 and derivatives and their
ability to activate IE promoters. Mutations in the promoters of the IE genes could
also be construct and tested in this fashion to investigate the effects of different

deletion mutations on the different promoters.

63
Other mutations in VP16 that have been tested in transient expression

assays could also be transferred in to the virus to study the effect upon viral
replication Especially single point mutations that greatly decreased VP16
activity, such as the Phe 422 to Ala change in the N-subdomain and substitutions
of other important residues in the C-subdomain, should be tested in the viral
context, both to determine the effects on the overall lytic infection of the virus

and the effect on viral gene expression.

CHAPTER III

IDENTIFYING THE ACTIVATION DOMAINS IN VP16 HOMOLOGUES

Introduction

VP16 of HSV-1 is known for its ability to activate IE transcription and
plays an essential role in HSV-l maturation and particle assembly. Homologues
of VP16, HSV-2 VP16 (Cress & Triezenberg, 1991a), BHV-l UL48 (Carpenter &
Misra, 1991), VZV ORF10 (Davison, 1991; McKee, et al., 1990), MDV UL48
(Koptidesova, et al., 1995; Yanagida, et al., 1993), EHV-l Gene12 and EHV-4
GeneB5 (Purewal, et al., 1994; Purewal, et al., 1992) have been identified and the
amino acid sequences have been deduced.

Despite the availability of the deduced protein sequences, the activation
domains of these homologues of VP16 have not yet been identified (with the
exception of VP16 from HSV-2). Linear sequence comparisons of the
homologous proteins do not reveal any obvious similarity between the VP16
activation domain and the sequence at the C-terminus of other homologues. In
some VP16 homologues such as VZV ORF10 and MDV UL48 the entire 80 amino
acid C-terminal portion of the protein is missing.

VP16 was characterized as an acidic activator because the activation
domain of the protein is rich in acidic residues. However, Cress et al. 1991 and
Regier et a1 . 1993 have shown that hydrophobic residues, especially Phe residue

64

65
442 (Phe442) and others in its vicinity, are critical for its function Considering

the essential role of hydrophobic residues for activity in mind, we have
employed a sequence analysis program termed hydrophobic cluster analysis
(HCA) to analyze VP16 and its homologues, and search for their activation
domains. HCA presents the protein sequence in a two dimensional 01 helical
representation A contour is then drawn grouping the adjacent hydrophobic
residues, which presumably form the hydrophobic core of globular proteins,
producing clusters of hydrophobic residues. These clusters were then visually
compared to the patterns obtained for putative homologues (Lemesle-Varloot, et
al., 1990; Woodcock, et al., 1992).

This analysis has revealed activation domains in VP16 homologues that
were not obvious when the linear amino acid sequences were compared. In this
study I and my collaborators have demonstrated, by site directed oligonucleotide

mutagenesis substitutions of predicted critical residues, that the activation
. domain in VZV ORF 10 is indeed homologous to the N-subdomain of the VP16
activation domain. These results were published in the Emedingsgflhe
W (1995, Vol. 92, pp. 9333-9337).

We have identified the activation domain of at least three homologues of
VP16 from HSV-l (Triezenberg, et al., 1988a), HSV-2 (A. Cress unpublished
results) and VZV (Moriuchi, et al., 1995). The activation domains of BHV-l,
MDV, and EHV-l & 4 are also being mapped and identified in other laboratories,
with results supporting my predictions (Elliott, 1994; Misra, 1994; Mekki
Boussaha, manuscript in preparation). These results demonstrated that we have
successfully predicted the location and critical residues for the activity of the
VP16 homologous.

To study structural aspects of the VP16 activation domain, we have used
several different algorithms to predict the secondary structure of the N-

66
subdomain of the activation domain of VP16 and its homologues. We expected

that the secondary structural analysis of the N-subdomain from the different
homologues would demonstrate a common preferred secondary structure and
would thus provide insights on the structure of the activation N-subdomain of
VP16. The Chou-Fasman (Chou & Fasman, 1978), Garnier-Osguthorpe-Robson
(Garnier 8r Osguthorpe, 8: Robson, 1978) and two new methods, Sequery
(Collawn, et al., 1991; Collawn, et al., 1990) and profile network system from
Heidelberg (PHD) (Rost & Sander, 1994) were applied in these studies. The
predictions by Chou-Fasman, Garnier-Osguthorpe-Robinson and PHD suggested
that the region around the predicted orifice] Phe residue does not have a strong
preference for any secondary structure. However, results from predictions by
Sequery suggested that the secondary structure of the segment is likely to be loop
or extended structure.

Methods

DNA and Plasmids

Plasmid GAL4-ORF10(5-79), expressing the chimeric protein GAL4-
ORF 10, which contains the DNA binding domain GAL4 residue 1-147 fused to
the N-terminal amino acid residues 5-79 of VZV open reading frame 10 (VZV-
ORF 10) was provided by]. Cohen

pGAL4-ORF10(5-79) was digested to completion with EcoRI. The EcoRI
fragment containing the sequence encoding amino acid residues 5-79 was then
subcloned into M13mp18. The fragment orientations of several of the subclones
were identified by standard complementation test and complementary M13

dc

sil

67
clones were sequenced to obtain the phage with the correct orientation, resulting

in M13mp18-ORF10.
The reporter plasmid, p5GAL4-E1b-CAT, containing five GAL4 binding
sites linked to the CAT gene was described previously (Lillie 8: Green, 1989).

Site directed oligonucleotide mutagenesis

Oligonucleotides used for site-directed mutagenesis are shown in Table 1.
The Oligonucleotides were provided by]. Cohen's laboratory. M13mp18-ORF10
was mutated using standard protocols (Zoller and Smith 1982; Kunkel, 1985; Bio-
Rad catalog #170-3571). M13mp18—ORF10 was grown in E. coli host C] 236
[(dutl, ungl, thil, relAl; pC]105(Cmr)] which incorporates uridine nucleotides at
approximately 1% of the normal thymidine residues. Uridine-containing phage
DNA purified by precipitation with PEGSOOO, extraction with phenol and
chloroform, and then ethanol precipitation. The mutagenic oligonucleotide
primers were annealed to template DNA in 20 mM Tris-HCl (pH 7.4), 2 mM
MgC12, 50 mM NaCl by heating to 65 °C and cooling, 1 °C / minute to 4 °C. A
standard primer / template ratio of 20:1 was used (Bio-Rad catalog #170-3571).
Synthesis of the complementary strand DNA was performed in 23 mM Tris-HCl
(pH 7.4), 5 mM MgC12, 35 mM NaCl, 1.5 mM dithiothreitol, 0.4 mM
deoxynucleotide triphosphates, 0.75 mM ATP, 0.2 unit/v.1 T4 DNA ligase, and 0.1
unit/ pl T4 DNA polymerase. The synthesis was carried out for 5 minutes at 4 °C,
5 minutes at 25 °C and 60 minutes at 37 °C. The reaction was terminated by
heating at 65 °C for 10 minutes. The synthesis reactions were then used to
transform competent dut+ ung+ MV1193 cells {(Alac-pro AB), thi, supE, A(sr1-
recA)306::Tn10(tet') [F':traD36,proAB lac IQZAM151}. Transformation of the
synthesized duplex DNA into MV1193 cells results in the selection against the

Table 1
Mutagenic Oligonucleotides

 

F28S 5'-GTT GT G GAC GCA TCT GAT GAA TCG TIC-3'
F28Y 5'-GTT GTG GAC GCA TAT GAT GAA TCG 'ITG-3'
F28L 5'—G GTT GT G GAC GCA CIT GAT GAA TCG TT-3'
F281 5'-G GTT GT G GAC GCA ATT GAT GAA TCG 'IT-3'
F28V 5'-G GTT GTG GAC GCA GIT GAT GAA TCG TT-3'
F28A 5'-G GTT GT G GAC GCA QCT GAT GAA TCG TT-3'
F28P 5'-G GTT GT G GAC GCA CCT GAT GAA TCG TT-3'
F28T 5'-G GTT GTG GAC GCA ACT GAT GAA TCG TT-3'
V24A 5'-ACG GAA CAA GCG GET GTG GAC GCA TIT-3'
V25A 5'-GAA CAA GCG GTT GCG GAC GCA TIT CAT-3'
[32A 5'-A T'TT GAT GAA TCG EEG T'IT GGT GAT GT -3'
F338 5'-GAT GAA TCG TTG TCT GGT GAT GT A GCA-3'
F33Y 5'-GAT GAA TCG TTG TAT GGT GAT GTA GCA-3'

 

Table 1. Mutagenesis Oligonucleotides. These Oligonucleotides were
synthesized in I. Cohen's laboratory. They were used in the mutagenesis of
Phe28 and the flanking hydrophobic residues, Va124, Va125, Leu32, and Phe33, in
the activation domain of VZV ORF10 in the context of M13mp18-ORF10(5-79).
The mutated fragment was then subcloned into pGAL4-ORF10(5-79). The bold-
underlined letters indicates the nucleotides that have been substituted to create
the amino acid changes.

69
uridine-containing wild type DNA strand. Transformant MV1 193 plagues were

screened for mutations by dideoxy sequencing (Sanger, Nicklen, 8: Coulson,
1977). Replicative form DNAs of the mutant phage were prepared using
standard protocols. The double stranded phage DNA was cut with EcoRI
releasing the fragment encoding the VZV ORF 10 residues 5-79. The EcoRI
fragment was then subcloned back into pGAL4-ORF10(5-79) and the correct
orientation of the fragment was confirmed by StyI digestion The mutated
derivatives of pGAL4-ORF10(5-79) were amplified in DH5a cells. The plasmid
DNA was extracted and purified twice by centrifugation in CsCl gradient and
used in transient expression assays to determine the activity of the mutant
relative to wild type GAL4-ORF10(5-79).

Transient expression assays were performed by Hiroyuki 8: Masako
Moriuchi in the laboratory of]. Cohen. Plasmids encoding mutant derivatives of
GAL4-ORF10(5-79) were co-transfected with a reporter plasmid p5GAL4-E 1b-
CAT (Lillie 8: Green, 1989) into Vero cells. CAT activity was determined for wild
type GAL4-ORF10(5—79) and each of the derivatives. The relative CAT activity of

the mutant relative to the wild type protein was determined.

HCA

The peptide sequences of VP16 homologues were acquired from the
GenBank Genetic Sequence Data Bank (Access numbers: HSV -1 KOS VP16, S. I.
Triezenberg's unpublished results; HSV-2 VP16, M60050; BHV-l UL48, 211610;
VZV ORF10, X04370; EHV-l Gene12, M86664; EHV-4 GeneB5, L16590; MDV
UL48, X73370). The primary sequences were then processed through the
hydrophobic cluster analysis program, HCA-plot-V2, kindly provided by Dr. B.
Henrissat (Centre National de la Recherche Scientifique, Université Joseph

70
Fourier, Grenoble France). The HCA patterns of each of the homologues were

converted to postscript format and printed. The hydrophobic cluster patterns of
all the homologous were then aligned and compared by eye. The linear sequence
alignment of the homologues was also compared using Pileup, a multiple
sequence alignment program from the Genetic Computer Group, Inc. (GCG),
1994, and used in conjunction with the data from HCA to align and compare the
primary sequences. The algorithm used for Pileup was according to the work of
Feng 8: Doolittle, 1987.

Sequery

The computer program termed Sequery searches a database of proteins
with known structures (the Brookhaven protein data bank) for a specific amino
acid sequence (the query sequence) from a protein of interest (Collawn, et al.,
1991; Collawn, et al., 1990). Sequery searches the data base for similarity in the
sequence independent of the protein origin or function. The program output is a
list of the proteins in the database which contain a sequence matching the query
sequence. The amino acid portion of the matching amino acids is also listed. The
protein structures of the matching proteins were then retrieved and displayed.
The structure of the protein from the database containing the query sequence
represents a predicted protein structure for the query sequence.

The sequence from residues 439-447 of HSV-1 VP16 was used as a query
sequence using a sliding window of four residues. The amino acids designed as
matching at each position were the original residues presented in HSV-l VP16
and other conserved residues found at that position from the linear amino acid
alignment of the N-subdomain (Figure 4B). The amino acid residues allowed at

each position are shown in Figure 1. Some residues were not allowed and some

71

 

 

 

 

 

 

 

 

437 442 449
DALDDFDLDMLGD
IEEYNDEL

V E
< 1 >
< 2 >
< 3—-->
<-----4 >
< 5—-—>
< 5 >

Figure 1. Residues allowed at each position for the Sequery program search
and the six windows used in the search. A sliding window of four amino acid
residues, starting from amino acid 439 to 447, was used to search for matching
sequences in the Brookhaven protein data bank. The residues allowed at each
position were selected based on the conserved residues found in other VP16
homologues. Some residues were avoided and some were added to minimize
picking up unrelated sequence and yield a workable amount of structures from
each search The position of each window is indicated by "<->".

72
residues (known to retain the function of VP16 activation ability) were added to

avoiding picking up irrelevant sequences and to yield a workable sample size.

PHD

The computer program termed PHD (Rost 8: Sander, 1994), predicts the
secondary structure of proteins by using evolutionary information of protein
structures in combination with computer neural networks. The program takes
into account not only the local information present in the primary sequence of
the protein, but also picks up similar sequences from the data base and uses the
information from other sequences to assist in the prediction This method has
improved the accuracy of the protein structural prediction to above 70% for the
three-state prediction, 0: helix, 8 sheet and loop, of a given protein sequence.

The sequences of the VP16 homologues from the linear alignment of the
N-subdomain in Figure 3 were analyzed by the PHD program. Sequences were
sent for analysis via world wide web (http:/www.embl-
heidelberg.de/predictprotein /predictprotein.html) The PHD results were
returned via electronic mail. The output shows the prediction of the three-state
structure as well as the similar sequences picked out from the data base.

Chou-Fasman and Garnier-Osguthorpe-Robson

Secondary structure predictions by the Chou-Fasman and Garnier-
Osguthorpe-Robson algorithms use the information from the statistical survey of
a selected set of proteins from the protein data base previously described (Chou
8: Fasman, 1978; Garnier, Osguthorpe, 8: Robson, 1978). The sequence of VP16
homologues were subjected to secondary structure prediction by the
Peptidestructure program in the GCG, which employs both the Chou-Fasman

73
and Garnier-Osguthorpe-Robson methods. The surface probability was

calculated according to the formula of Emini et al., 1985.

Results

HCA

The activation domain of VP16 was mapped to the C-terminal 80 amino
acid residues of the protein (Triezenberg, et al., 1988a). Homologues of the VP16
protein in other 01 herpes viruses have been identified and their amino acid
sequences have been deduced from cloned gene sequences. Sequence
comparison of the deduced amino acid sequences of VP16 homologues has
demonstrated that the amino acid sequences are mostly conserved in the middle
portion of the protein which corresponds to residues 50-410 of HSV-1 VP16. The
N and C terminus of the protein, however, were divergent in the amino acid
sequence as well as length, except for HSV-1 and HSV-2 VP16 which have a high
degree of similarity throughout the molecule (Figure 2). The sequence of VZV
ORF10 and MDV UL48 are shorter than the rest of the homologues and lack the
residues corresponding to the activation domain of VP16, residues beyond 410 of
HSV-1 VP16. EHV-l Gene12, EHV-4 GeneBS, and BHV-1 UL48 contain residues
corresponding to residues beyond residues 410 in HSV-l VP16, but these
residues were not conserved in sequence when compared by conventional linear
alignment methods (Figure 2). Therefore, comparison of these proteins using
conventional linear sequence alignment to identify activation domains in these
VP16 homologues was unfruitful.

Though the activation domain of VP16 was characterized as an acidic
activation domain, Cress et al. 1991 and Regier et al. 1993 have shown that in

HSV-lVP16
HSV-ZVP16
EHV-lgenelz
EHV-lgeneSB
BHV—lUL48
VZVORFlO
MDVUL48

HSV-lVP16
HSV-ZVPIG
EHV-lgenelz
EHV—lgeneSB
BHV-lUL48
VZVORFlO
MDVUL48

HSV-lVPlG
HSV-2VP16
EHV-lgenelz
EHV-lgeneSB
BHV-IUL48
VZVORFlO
MDVUL48

HSV—lVP16
HSV-ZVP16
EHV-lgenelz
EHV—lgeneSB
BHV-IUL48
VZVORFlO
MDVUL48

HSV—lVPlG
HSV-ZVP16
EHV-lgenelz
EHV—lgeneSB
BHV-IUL48
VZVORFlO
MDVUL48

HSV-lVP16
HSV—ZVP16
EHV-lgenelz
EHV-lgeneSB
BHV-lUL48
VZVORFIO
MDVUL48

HSV-lVPlG
HSV-ZVP16
EHV-lgenelz
EHV-lgeneSB
BHv-IUL48
VZVORFlO
MDVUL48

0000000000

oooooooooo

AGGPKNTPAA
AGGPKNTPAA
TCELMDMDGA
TCELMDVDGV
AAA'IMDPYDA
S...KTEQAV
KSPDLDILRT

PAALFNRLLD
PAALFNRLLD
PGALYQRLQG
PTALYQRLQA
LAALLERMQA
PKILYQQLIR
PNSLYTRLLH

LPSDVVEWG

LPSDVIDWGD
SVDEVVRAGL
SVNEVVKAGL
DADAVVGAMY
DPDDVI....
SPSEV....L

SRFFHAELRA
AQFFRGELRA
QRFYLSELRT
QRFYLSELRA
QAHFLAELRA
QDSFTVELRA
QNFYLGELEA

GEMLRATIAD
REMLRTTIAD
MHKFKQVVRD
MHKFKQVVRD
ADRLRQLVAA
YTQLRQSILL
CSKARQYIAE

DCLCCDLESW
DGLCCDLESW
VSLHYFWAQR
VSLHYTWPQR
VSLYYAWPQR
AALKFTWTER
VYLKCEWLQE

0000000000

..........

PPLYATGRLS
PPLYATGRLS
VASFDEGMLS
VASFDEGMLS
IEAFDDSLLG
VDAFDESLFG
IEEFDETLLS

.3.:
DLGPSAGPAL
DLGFSAGPAL
ELGFPEGQTL
ELGFPEGQAM
ELGFPDGPAL
DLDFSEGPRL
ELDFVEGPSI

AYVPERA...
AHVPERS...
DSLPTPSHYS
DSLPIPTNYI
LAVPGDAE..
.STVSTKDHV
EELSKNTWTY

33. 3
REESYRTVLA
REESYRTVLA
REEHYARLLR
REEHYSRLLR
REERYAGLFL
REEAYTKLLV
REKSYATMFY

3333.....x

RYYRETARLA
RYYRBTARLA
RYYREAANLA
RYYRETANLA
RYYREASRLA
RYYREVASLA
RYYREAARFA

RQLAGLFQPF
RQLACLFQPL
RKFECLFHPV
RKFECLFHPV
RQFTCLFHPV
RQFTCAFHPV
RHFHCLFQPV

...... MDLL
FDARPAASIV
....MAANIA

SPLA....AG
DVASDIGFET
EIEVRTQSIP

CTMLDTWNED
CTMLDTWNED
LSAMEKWNED
LFAMEKWNED
LRAMERWNED
LSCLETWNED
LARLEKINVD

.QIDIRAHGD
.PIDIRAHGD
PEVDLNAHGD
PEVDLNAHGS
.RLDLNAHAN
EMFNLTTRGS
TALNLNEHGE

..3
NFCSALYRYL
NFCSALYRYL
GYCVALLHYL
GYCVALLHYL
GYCRALLQHL
TYCKSIIRYL
GYCRALAEYI

0.0...

RVLFLHLYLF
RVLFLHLYLF
RLLYLHLYVS
RLLYLHLYIS
RLAFAHMYVA
RLLYLHLYLT
KLLYVHLYLS

0.00:00000

MFVNGALTVR
MFINGSLTVR
LFNHGVVILE
LFNHGVVILE
LFNHGVVALE
LCNHGIVLLE
IFNHGVVIVE

VDELFADMDA
VDDLFADR..
MFAAAEENDD
MFADIEDYDD
MSGRIKTAGR
...MECNLGT
LFAEIEAYAN

....... QAQ
....... QAQ
SVYSIPTKKR
SIYSSPAQKR
PLYDGPSPAR
SLYSHAVKT.
SPLVAPSVTK

.33 .3
LFSALPTNAD
LFSGFPTNAD
MFSALPGHVD
MFSAIPVHVD
LFSCLPTNAD
LFSCFPINED
LFSCFPHNKH

.3 .3.
VAFPTLPATR
VAFPTLPATR
EPFPEVPALE
EPFPEVPALE
QPLPAPPASE
VRLPSPPKQP
MALPMPPTTK

RASVRQLHRQ
RASVRQLHRQ
YGSAKR...Q
YGSAKR...Q
RATAAR...G
QGTAKRTTIG
RQSAIKDLRD

LTREILWAAY
LSREILWAAY
VTREVSWRLH
VTREVSWRLH
TAREVSWRLH
VTREFSWRLY
TTRDVSQRLE

GVPIEARRLR
GVPVEARRLR
NDPLEFHDLQ
NDPLEFNDLQ
DGFLDAAELR
GKPLTASALR
GRVLTAPELR

DGASPPPPRP
DGVSPPPPRP
PYPGKSGYND
TRSCEYGY.G
ALASQCG.GA
EHPSTDTWNR
TM ....... D

LMPSPPMPVP
LMPSPPMPVP
LALPPPKAAS
LALPPPKATS
FALPPPRPAP
.APSPPWVAS
MSLPSPSPAP

.3 . ..33.
LYRECKFLST
MYRECKFLST
LYTEIALLST
LYTEIALLST
LYADAALLSA
LYSDMMVLSP
LYEHAKILSV

..3 3. .
DGLGLYYEAL
DELPSYYEAM
DDLEIYVISA
DELETYVISA
EGLPEYVAGV
TGLPAYVQEV
ADLPSYVDDI

AHMRGRDRDL
AHMRGRNRDL
LRGSGSDASL
LRGAGSDSAL
.RGAAGAGAQ
LNIQNPDQKA
ARVEDKNIGA

AEQMMRPDLF
AEQMMRPDLF
ASQVINQGVF
ASQVVNQGIF
SQQSQAQGVF
ASQSAHPDVF
ASQMGRQNIF

ELNHIREHLN
ELNHIREHLN
RINYRRRELG
RINYRRRELG
RLNYRRRELG
EINYRRRELG
AQNYIRSEFG

HSV-lVP16
HSV-ZVP16
EHV-lgenelz
EHV—lgeneSB
BHV-lUL48
VZVORFlO
MDVUL48

HSV-IVP16
HSV-ZVPlG
EHV-lgenelz
EHV-lgeneSB
BHV-lUL48
VZVORFIO
MDVUL48

HSV-lVP16
HSV-ZVP16
EHV-lgenelz
EHV-lgeneSB
BHV-lUL48
VZVORFlO
MDVUL48

HSV-lVP16
HSV-ZVP16
EHV-lgenelz
BHV—lgeneSB
BHV-lUL48
VZVORFlO
MDVUL48

HSV-IVP16
HSV-ZVP16
EHV-lgenelz
EHV-lgeneSB
BHV—lUL48
VZVORFlO
MDVUL48

333.3 ...3
LPLVRSAATE
LPLVRSAAAE
LPLIRAGLIE
LPLIRAGLIE
LPLVRAGLVE
LPLVRCGLVE
LPLIRCKLVE

SPSEAVMREH
SEGESVMREH
TP.LFPLAEH
TP.LFPLAEH
GR.LAPEREH
EP.RHVRADH
NPQLHVHKEH

SFLPAGH.TR
SFLSAGQRPR
GVRQTAATLA
GIRQTAETLA
RVASPATHLA
GFLTR .....
TL* .......

GDGDSPGPGF
GDVESPSPGM

0000000000

oooooooooo

75

. :3 3
EPGAPLTTPP
EPGAPLTTPP
EENSPLEABP
EENLPLESEP
VEVGPLVEEP
ENKSPLVQQP
EPDMPLISPP

AYSRAR.TKN
AYSRGR.TRN
SYSKRIGGRL
SYSKRIDGRL
SYARPRGA.I
PYAKVVENR.
VHVQKLESPP

RLS.TAPPTD
RLSTTAPITD
.IPSNLTLQS
.LPSNLTLQS
QAPSAKGAAP

0000000000

TPHDSAPYGA
T.HDPVSYGA

0000000000

..........

TLHGNQARAS
VLQGNQARSS
LFSGKLPRTI
TFSGKLPRTI
PFSGSLPRAL
SFSVHLPRSV
PFSGDAPRAS

.33...3..
NYGSTIEGL.
NYGSTIEGL.
SYGTTTEAM.
SYGTTAEAM.
NYGTTPEAM.
NYGSSIEAMI
NYGTTVEALL

VSLGDELHLD
VSLGDELRLD
METDGLD...
METDVLD...
AEFAALAGLA

LDMADFEFEQ
LDVDDFEFEQ

GYFMVLIRAK
GYFMLLIRAK
GFLTHQIRTK
GFLTHQIRTK
GFLNYQVRAK
GFLTHHIKRK
VYLLQCIRSK

LDLP.DDDAP
LDLPDDDDAP
MDPPSPSAVL
MDPPSPSAVL
LRPPSPSEVL
LAPPSPSEIL
MDSSDRNSIS

GEDVAMAHAD
GEEVDMTPAD

0000000000

MFTDALGIDE
MFTDAMGIDD
....... YSS
....... YSS
ANPFGGTYDA

0.... O.

LDSYS.SFTT
LDSYS.SVAT
MEAYSDAHPA
MEAYSNAHPS
MGA..PAEAG
LDAYAVKHPQ
LEVYSLSHPP

EEAGLAAPRL
AEAGLVAPRM
PGDPVPPLTV
PGDPVPPLTV
PCDPAPAATV
PGDPPRPPTC
PGDPVATTIS

ALDDFDLDML
ALDDFDLEML

0000000000

YGG* ......
FGG* ......
MTGDELNQMF
ISGDELNQMF
LLGDRLNQLL

IIIIIIIIII

0000000000

Figure 2. Linear alignment of the amino acid sequence of VP16 and its
homologues. The sequence of all the proteins were analyzed using the pileup
program in the GCG. Gap weight was set at 3.0 and Gap length weight of 0.1.
Both N and C-termini of the homologues are not conserved in contrast to the
central part of the proteins. A dot (.) indicates that the residue at that position is
identical in at least four members of the homologues, a colon (z) indicates that the
residue at that position is identical in all homologues.

76
addition to the net negative charge of the activation domain, several key aromatic

and bulky hydrophobic residues play a critical role in the function of VP16.
Therefore, to identify activation domains we have employed the HCA program
which specifically identifies the conserved hydrophobic residues and clusters, to
analyze and compare the homologues of VP16. We hypothesized that if these
key aromatic and bulky hydrophobic residues are essential for activity, the HCA
pattern of these residues should be conserved in VP16 homologues.

The question of whether or not these homologues of VP16 work as
transcription activators still remains. Moriuchi et al. (1993) demonstrated that
VZV ORF 10 despite lacking the 80 amino acid C-terminal residues, was able to
activate transcription of the viral IE promoter. This result suggests that despite
the differences in the primary structure of this VP16 homologue, it is able to
activate transcription. It is possible that it might also retain similar activation
domains.

The sequences of HSV-1 VP16, HSV-2 VP16, BHV-l UL48, VZV ORF10,
MDV UL48, EHV-l Gene12, and EHV-4 GeneB5 were subjected to HCA. The
results are shown in Figure 3. The proteins were aligned by the conserved
clusters of the homologue proteins rather than the position of amino acid
residues in the primary structure. The HCA plot demonstrated that the most
conserved portion among the homologues is the central portion of the protein
corresponding to residues 50-390 of HSV-1 VP16. The terminal portions, both N-
terminal and C-terminal, are less conserved.

The activation domains of HSV-1 VP16, N-subdomain (residues 410—456)
and C-subdomain (residues 456-490) have distinct HCA patterns as shown in
Figure 4 and 5. The pattern seen in the N-subdomain of VP16 from HSV-l is a
ladder like cluster of Leu residues, between residues 410-430, and a horseshoe-
like shape with the critical Phe residue at the apex of the horseshoe cluster,

Figure 3. HCA of the complete sequence of VP16 and other VP16 homologues.
The complete deduced sequences of HSV-1 VP16, HSV-2 VP16, ORF10, equine
herpes virus I (EHV-1)Gene12, EHV-4 GeneBS, bovine herpes virus 1 (BHV-l)
UL48, and Marek's disease virus (MDV) UL48 proteins were analyzed by HCA.
The amino acid sequences are represented (in duplicate) along diagonals from
upper left to lower right with the one-letter amino acid code except that black
diamonds indicate glycine, squares indicate threonine, squares with central dots
indicate serine, and stars indicate proline residues. Sets of adjacent hydrophobic
residues are encompassed by contour lines. The HCA patterns were aligned by
the conserved hydrophobic clusters rather than the position of the residues in
their linear sequence. The vertical lines show the boundary of each conserved
region.

 

... bass...
gmwwwdaha 8.5 >:m

 

manna T>Im

      

9*
. ii-
.344?

3250 TEN

78

mg >92
3:0 >N>

3g N.>m=

I
II I I I
I I I I I I I
I I I I I
I I II I II I I I I I III I
I I I I I I I I I II I I I
.I.I I II I II I I I I I I
II II III. I I 3 IIII I I
I I I
I I. I I I I
I
I I . I I I I I I . I I . I o I II
.I II . I... II. II I I I. I v I .I.,.: II I
II. o .I .. I I . .I II a. I I I . I I. I2... I I I
. . I I I . I o I
I I . . I I.. I I
I I I I I I I I . I
. m»... .... .. . ., . . . .. .. .. ,. 1... :.. . IE>I>mm
I .o u II . I IIII.". .- I o .cIIt 1. II II I I
II. I I _. . .. .. I I I . I
IIIIiioM I Ila-“I I ... I II I. I . I. .I I . I I I It I II I I

I.- _ _ _ _ . _ _ fin _ _ _ _

Figure 3

 

79
between residues 430-456 (Figure 4). Since this cluster contains the critical Phe

residue and was also shown to be able to activate transcription (Emami & Carey,
1992), I have used this horseshoe shape pattern to identify the N-subdomain in
other homologues. The most obvious homology was found in the BHV-l UL48
N-terminus between residues 22-46, with the Phe residue predicted to be critical
for the activity at Phe residue 33. Though the horseshoe cluster in VZV ORF10
was not as obvious, in conjunction with the linear alignment one can clearly see
the conserved Phe residue, residue 28, as well as other conserved hydrophobic
and acidic residues. Careful observation of the sequences of other homologues
(except HSV-Z) revealed that all homologues contain a fragment in the N-
terminus of the protein that is homologous to the N-subdomain of HSV-1 VP16,
shown in Figure 4A.

Linear sequence comparison of all the VP16 homologues in Figure 48
clearly shows the conserved critical Phe residue as well as hydrophobic and
acidic residues in its vicinity. The critical Phe found at the apex of the horseshoe
like shape of the cluster was conserved and used as the core of the alignment.
Bulky hydrophobic residues corresponding to Leu residue 439, Met residue 446
and Leu residue 447 of HSV-1 VP16 were conserved in all the homologues except
for the residue corresponding to Leu residue 444. This position in other
homologues was found to be a conserved acidic residue, either Asp or Glu.
Though the position of most of the acidic residues is not conserved among all the
homologues, a single Asp residue following the critical Phe was found to be
conserved. Therefore this Asp residue was also used, together with the critical
Phe residue, to identify and align the N-subdomain in VP16 homologues.

The pattern of hydrophobic clusters in the C-subdomain is distinct from

the N-subdomain and appears as a pair of ladder like clusters followed by a

triangular cluster, shown in Figure 5A. Though both the N and C subdomain

80

Figure 4. HCA of the N-subdomain in VP16, ORF10, and other VP16
homologues. The complete deduced sequence of HSV-1 VP16, HSV-2 VP16,
ORF10, equine herpes virus I (EHV-l) Gene12, EHV-4 GeneBS, bovine herpes
virus 1 (BHV-l) UL48, and Marek's disease virus (MDV) UL48 proteins were
analyzed by HCA, and locations of activation domains were predicted by using
VP16 as a standard. (A) HCA profiles of subdomains resembling VP16N show
the characteristic horseshoe-shaped cluster and the centrally located Phe
residues. (B) Linear alignment of the subdomains shown in A, with shading
indicating conserved residues, boldface letters indicating conserved bulky
hydrophobic residues, dark shading denoting conserved acidic residues, and
open ovals representing residues conserved among homologues but not present
in HSV—l VP16.

81

 

mmmzmw .VIE

 

NHMZMU HIE

 

 

 

 

Figure 4

82
seem to have ladder like clusters, they are formed by different adjacent residues

in the HCA plot giving different cluster orientations. Careful observation of the
VP16 homologues has also revealed similar clusters located in the C-terminal
portion of HSV-2 VP16, BHV-l UL48, EHV-l Gene12, and EHV-4 GeneB5 shown
in Figure 5A. Linear alignment of the primary sequence also shows conserved
bulky hydrophobic residue corresponding to residues Met470, Phe473, Met474,
Phe475, as well as acidic and polar residues in the vicinity (Figure SB). Many of
the conserved hydrophobic residues in this alignment have been shown by
alanine scanning and random mutagenesis of this C-subdomain in HSV-l VP16
to be important for activity of the C-subdomain (Susan Sullivan and Peter Horn
unpublished results).

HCA predicts that all the VP16 homologues contain a conserved N-
subdomain. The N-subdomain is located at the N-terminus of the VP16
homologue in BHV-l, EHV-l, EHV-4, VZV and MDV, and at the C-terminus in
HSV-l and HSV-Z. The C-subdomain, however, is not conserved in all
homologues. It is only present in HSV-1, HSV-2, BHV-l, EHV-l and EHV-4
VP16 homologues, and is located at the C-terminus of all the proteins. The
location of the predicted N and C subdomain of each homologue is shown in

Figure 6.

Site directed oligonucleotide mutagenesis of the proposed VZV ORF10

activation domain

Results from the HCA of the VP16 homologues suggest that these
proteins, except for HSV-2 VP16, contain the N-subdomain located near the N-
terminus of the protein. I tested this hypothesis using VZV ORF 10, already
known to activate transcription from its IE promoter as well as the HSV-l ICPO

Figure 5. HCA of the C-subdomain in VP16 and other VP16 homologues. (A)
HCA profiles of subdomains resembling VP16C, drawn as in Figure 4A. (B)
Linear sequence alignment of the subdomains shown in A, with shading, ovals,
and boldface type as described in the legend of Figure 4B.

m ouflwmm

H I III 2
IS www.mw....w,m.mm.I.“

vmv

 

 

 

 

 

 

 

om

om

 

 

VP16 Type 1&2

VZV ORPIO & MDV-UL48

[I Jcl

BHV-l UL48

[- IL]

EHV-l GENEIZ 8r. EHV-4 GENEBS

 

 

 

 

Figure 6. Locations of the N and C subdomains in VP16 and its homologues.
HCA was used to analyze the deduced protein sequence of VP16 and its
homologues. We predict the location of both activation subdomains of VP16
type 1 and 2 to be at the C-terminus of the protein. In other homologues the N-
subdomain was predicted to be located at the N-terminus of the protein while the
C-subdomain remains at the C-terminus of the protein (except MDV UL48 and
VZV ORF10 where the C-subdomain is not present in the protein).

86
and ICP4 promoters (Moriuchi, et al., 1993). I proposed that if the activation

domain of VZV ORF10 is indeed located near the N-terminus of the protein and
Phe residue 28 (Phe28) is the critical residue for its activity, mutations of the
Phe28 and other hydrophobic residues in the vicinity would reduce or abolish
the ability of VZV ORF10 to activate transcription.

The activation domain of VZV ORF10 was initially mapped to the N-
terminal portion of the protein, between residues 5-79, by H. Moriuchi. This was
accomplished by fusing different deleted fragments of VZV ORF 10 to the DNA
binding portion of the GAL4 protein (residue 1-147) and testing these fusion
proteins for the ability to activate expression of CAT from the reporter plasmid
pSGAIA-Elb-CAT. VZV ORF10 fragments containing residues 267-410 and 113-
351 did not show any significant activity. When a fragment containing residues
5—175 was tested, it showed approximately 100 fold activation over the control
GAL4 DNA binding domain (residues 1-147). Further deletion of the VZV
ORF10 fragment to residues 5-79 resulted in an increase in activity to
approximately 300 fold. Dissection of residues 5-79 to residues 5-41 and 38-79
suggested that the activation domain lies within residues 38-79 (results shown in
Appendix A). However, substitution of large bulky hydrophobic residues in the
38-79 region, Ile65, Leu66, Tyr67, Leu70 and Leu74 to Ala, by H. & M. Moriuchi,
did not reduce the activity of GAL4-ORF10(5-79). Therefore the exact location
and critical residues involved in transcription activation were unknown. This
was not surprising since some random fragments from E.coli DNA fused to the
GAL4 DNA binding domain were found to be able to activate transcription (Ma
& Ptashne, 1987). Therefore interpretation of these results must be made with
caution.

Although these results did not clearly identify essential residues or the
exact location of the activation domain, they were sufficient to further strengthen

87
my hypothesis that the activation domain of VZV ORF 10 is indeed located in the

N -terminal portion of the protein.

I mutated the proposed critical residue for transcription function Phe28 by
site directed mutagenesis. Phe28 was changed to other bulky hydrophobic
residues, Tyr, Leu, Ile, smaller hydrophobic residues Val and Ala; hydrophilic
residues Ser and Thr; and Pro which may disrupt secondary structures. I also
designed oligonucleotides to make mutations of the flanking hydrophobic
residues Va124, Va125, Leu32, and Phe33. Since a bulky hydrophobic residue at
Va124 is not conserved at that position (Figure 4) I predicted that changes at
Va124 should not affect the ability of GAL4-ORF10(5-79) to activate transcription.

Mutations of Phe28 generated in M13 phage were subcloned into the
GAL4 fusion vector. The constructs were tested by H. Moriuchi using transient
expression assays. A reporter plasmid p5GAL4-E1B-CAT was cotransfected into
Vero cells with the wild type or mutated versions of the plasmid expressing
GAL4-ORF 10 fusion protein, and CAT activities were determined (details in
Appendix A).

The results shown in Table 2 clearly demonstrated the importance of
Phe28 as a critical residue in the VZV ORF 10 activation domain. Mutation of
Phe28 to other bulky hydrophobic residues, Tyr, Leu, and Ile reduced the activity
of GAL4-ORF 10 to approximately 50% of the wild type fusion protein.
Substitution by the smaller hydrophobic residue Val further reduced the activity
of GAL4-ORE 10 to approximately 20% of wild type. Substitution by Ala and
small hydrophilic residues, Ser and flu reduce the activity to less than 15% of
wild type chimeric protein. Curiously, the Pro substitution did not abolish the
activity of the fusion protein, but gave similar activity to Ala and small
hydrophilic residues. This shows that the structural constraint imposed by the
substitution of a Pro residue at Phe28 did not have any further effect than

Table 2
Relative activity of GAL4-0RF10(5-79) bearing amino acid substitutions at

 

 

position 28
Residue substituted at Relative activity
position 28 (% of wild-type)
Tyr 52 i 17
Leu 54 :l: 17
Ile 50 :t 12
Val 23 d: 4
Ala 11 :t 2
Pro 13 :t 3
Ser 9 j: 4
Thr 7 :t 2

 

Table 2. Relative activity of GAL4-ORF10(5-79) bearing amino acid
substitutions at position 28. Vero cells were cotransfected with 5 ug of plasmid

expressing an amino acid substitution mutant of GAL4—ORF10(5-79) and 5 ug of
a reporter plasmid containing 5 GAL4 binding sites. The relative activity is
represented as the percent CAT activity of the mutant relative to that obtained
from the wild type GAL4-ORF10(5-79). Means and standard deviations were
calculated from at least five independent transfections.

89
substitution by a small hydrophobic residue such as Ala (Table 2).

I have shown that in the fusion protein GAL40RF10(5—79) the residue
responsible for the activation activity lie within amino acid residues 5-79 from
VZV ORF10 and that Phe28 is the critical residue for its activity. Mutation of this
residue to small hydrophobic residues or hydrophilic residues greatly reduces
the activity of this domain. These results lead to another very important
question. Does this domain in the context of VZV ORF 10 (rather than as a GAL4
fusion) function as an activation domain with Phe28 as a critical residue for its
function.

To demonstrate that this domain is essential for VZV ORF10 activation,

a mutation was constructed in which Phe28 was substituted with Ala in the
context of VZV ORF10 protein. The wild type VZV ORF 10 gene and VZV ORF 10
bearing the substitution of Phe28 to Ala were subcloned into the pCMV
expression vector and co-transfected into Vero cells with a reporter plasmid,
containing the VZV ORF62 promoter (IE promoter) linked to a CAT gene. A
plasmid derivative of pCMV bearing the HSV-l VP16 gene with truncations of
residues 413-490 as well as the parental pCMV vector’were used as controls. The
substitution of Phe28 to Ala in VZV ORF 10 rendered the protein inactive in the
transient expression assay. The activity of the mutant was at the same level as
HSV-l VP16 with deletions of amino acid residues 410 to 490 (Table 3). This
clearly demonstrates the significance of Phe 28 in the ability of VZV ORF10 to
activate transcription, and confirms our hypothesis that this domain is the VZV
ORF 10 activation domain.

Our laboratory has demonstrated the importance of hydrophobic residue
Phe442 for the activation function of the HSV-l VP16 N-subdomain, and has
successfully used patterns of hydrophobic residues to identify the activation
domain of its homologue VZV ORF 10. In addition to the critical Phe residue

90

Table 3
Phe28 is required for transactivation of the VZV 11362 promoter by ORF10

 

 

protein
Activator Fold induction
pCMVlO 24:3
PCMV10F28A 1.0 :t 0.1
pCMV16 290150
pCMV16d413-490 0.8 :l: 0.1

 

Table 3 Phe28 is required for transactivation of the VZV IE62 promoter by
ORF10 protein. Cells were cotransfected with plasmids expressing native
ORF10(pCMV10), an amino acid substitution mutant of ORF10 (pCMV10F28A),
full length VP16 (pCMV16) or a carboxy terminal truncation of VP16
(pMCV16d413-490) together with p62CAT (the VZV ORF62 promoter followed
by a CAT gene). Fold induction is the CAT activity in transfected cell extracts
relative to that obtained when using pCMV (vector control). Means and
standard deviations were calculated from at least five independent transfections.

91
there are also other flanking hydrophobic residues present in the horseshoe like

cluster (Figure 4). These residues are conserved and may play a role in the
activation function. I have generated mutations of these flanking hydrophobic
residues, Va125, Leu32, Phe 33 and Va124, substituting them with small
hydrophobic and hydrophilic residues, Ala and Ser. Substitution at Va124 serves
as a control, since bulky hydrophobic residues are not conserved at this position.
A structurally conserved aromatic but slightly polar residue, Tyr, was also
substituted at Phe33. These derivatives were then tested for their ability to
activate transcription as GAL4-ORB 10 fusion proteins, as previously described.

Substitution of Leu32 and Phe33 to Ala and Ser reduced the activity of the
fusion protein to approximately 50% and 10%, respectively (Table 4). These
results suggest bulky hydrophobic residues at position 32 and 33 are both
important for its activity . When Tyr, another bulky hydrophobic residue, was
substituted at Phe33 the activity of the GAL4-ORE 10 fusion protein dropped only
25%. The reduction in activity could be a result of the slightly polar character of
Tyr compared to Phe (Table 4). Curiously, substitution of Va125 to Ala had no
significant effect on the activity of the fusion protein. This result was unexpected
since bulky hydrophobic residues were found to be conserved at this position in
the alignment shown in Figure 4. This result suggests that there is no
requirement for a bulky hydrophobic residue at this position, at least in this
GAL4-ORF10(5-79) chimeric construct. Substitution at Va124 to Ala did not
diminished activation as expected.

These results clearly demonstrated that the activation domain of VZV
ORF 10 is indeed near the N-terminus of VZV ORF 10 protein. The results also
clearly demonstrated the importance of Phe28 for the activation function, both in
the context of a chimeric GAL4-ORF 10 fusion protein and the wild type VZV
ORF10 protein. Flanking hydrophobic residues, at position 32 and 33, were also

 

92

Table 4
Relative activity of GAL4-ORF10(5-79) bearing amino acid substitutions at
hydrophobic residues flanking Phe28

 

 

 

Amino acid substitution Relative activity
(% of wild-type)
Va124 to Ala 140 i 40
Va125 to Ala 130 :1: 50
Leu32 to Ala 45 i 18
Phe33 to Ser 12 i 4
Phe33 to Tyr 76 :t 18

 

Table 4. Relative activity of GAL4-ORF10(5-79) bearing amino acid
substitutions at hydrophobic residues flanking Phe28. Vero cells were
cotransfected with 5 ug of plasmid expressing an amino acid substitution mutant,
of hydrophobic residues flanking Phe28 of GAL4-ORF10(5—79), and 5 ug of a
reporter plasmid containing 5 GAL4 binding sites. The relative activity is
represented as the percent CAT activity of the mutant relative to that of the wild
type GAL4-ORF10(5-79). Means and standard deviations were calculated from
at least five independent transfections.

93
shown to be important for activation by VZV ORF10, at least in the GAL4-ORF10

chimeric protein. These results support the hypothesis that the activation
domain of VZV ORF10 is homologous to the activation domain of the HSV-l
VP16, N-subdomain. This strongly suggests that my predictions of the activation
domain of the other VP16 homologues are likely to be correct.

Secondary structure prediction by Chou-Fasman and Garnier-Osguthorpe-

Robson method

Previously, the Chou-Fasman and Garnier-Osguthorpe-Robson methods
were used to predict the structure of HSV—1 VP16 alone (W. D. Cress
unpublished results). In this study, the Chou-Fasman and Garnier-Osguthorpe—
Robson methods were used to predict the secondary structure of the N-
subdomain of all the homologues. We expected that introduction of the N-
subdomain homologues to the prediction might result in some other predictions
that agree with other current mutational and biophysical studies on the
activation domain of VP16.

The prediction results from Chou-Fasman and Garnier-Osguthorpe-
Robson method, shown in Table 5, suggest that the residues around the critical
Phe residue lie within a region of the protein that is likely to form a helical
structure followed by a turn. The only exception to this prediction is EHV-4
GeneB5, where the Chou-Fasman method predicted the sequence around Phe39
to be B sheet. Interestingly the critical Phe residue in all the predictions is found
to be highly exposed to solvent, with a surface probability according to Emini et
al. of higher than 1.0, with the exception of the EHV-l and EHV-4 homologues
for which the value is slightly less, at 0.77. This result suggests that the critical
Phe residue is likely to be solvent exposed.

94

Table 5. Secondary structure prediction by Chou-Fasman and Garnier-
Osguthorpe-Robson, and Emini surface probability prediction of the N-
subdomain of HSV-1 VP16, HSV-2 VP16, VZV ORF10, BHV-l UL48, El-IV-l
Gene12, EHV-4 GeneBS and MDV UL48. The sequences of the N-subdomain of
HSV-1 VP16 (residues 341-455), HSV-2 VP16 (residues 432-456), VZV ORF10
(residues 17-41), BHV-l (residues 22-46), EHV-l Gene12 (residues 53-77), EHV-4
GeneB5 (residues 28-52) and MDV UL48 (residues 32-56) were subjected to
secondary structure prediction by the peptidestructure program from GCG. The
calculations were based on the Chou-Fasman and Garnier-Osguthorpe-Robson
method. The surface probability of each residue was also calculated by the
method of Emini et al. Abbreviations: Position, Position of amino acid; Amino
acid, Amino acid residue; Surf Pr., Surface probability; CF, Chou-Fasman
prediction; GOR, Garnier-Osguthorpe-Robson prediction. The output of the

prediction is reported as: "H", for strong a helical prediction; "h", for weak helical
prediction; "B", for strong 5 sheet prediction; "b", for weak [3 sheet prediction; "'1'",
for strong turn prediction; "t", for weak turn prediction; for residues that no
prediction was made.

95

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I. one 0 II. I? I In. I: I RI.
.I. I IE I II .I. a: I me. I: o E
.I. .I 28 n In FEL I? I. m3 Pllhmr
. _I E I IF I RI m E. I oplmllmmwl
II I mg I R I PS m 3. I. one a E.
m I $8 > em I III P I? I. I. o? o cu.
m II and 0 mm m I. Inc 0 RI .I. .I. $5 a I:
m I one o In I I one o as. .I. I Inc m III.
II I Ed "I R m m 98 .I I3. I I. Inc .I El
m I. 8.0 I R m m 33 2 III I II “no 2 e3.
m I 3.0 m Im II II 33 m III II E SE a III
II .I mg m 8 II II mg I an. m II “no .I I:
m I $.Plbl1lbm m m mg n III II : mg n IF
m : EMILILNT : II Ellmllms. : : EIPILF
: II who I RH m E 3.0 a as. m m I: o :4
m I Imp P 8 m : Ed I III m m E 0 III
II I as > mu II F PS I Li. I m be .I 9:.
II I Inc > 37 m II 93 < 3. m = Id I F
m I Ed I. R m m 3.0 0 LE. m II who 0 F
II II me o an m m I; < 5. m m 38 I. III.
II I I: m Ln II 3 mad I RN F II «no m L?
m I 8.“ .I. cu m m m: .I. 3. m II Inc I. III
. I Hm I. II I m 85 2 In. I m Imp E 3.
I E...“ m I: m . 23 a 3 m II and I Lmnl ,
I 3N m E m 85 > In. P F I3 > I?
28 28 28
kbp. LP E. ass... 828a. kbp. FF Elm: 05.5. 828m. rubb- [bl E cane gamma.
ciao >N> 03> ~->m: or; $6:

 

559:5...— oIIIIuIIIII rue—I38 IIomI—oméI—IofiauIOLIIIIIIQ _..—I finishes—5

m 03:9

 

96

 

5:23.:— oIIIIoIIIIm raw—.88 IISI—cméeofiaumOLoIEaG IE: =aEuaI—éoIIU

8.253 m u-Aah

97

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

II II 3H.“ d owl
II II 23 I. RIM-II
II II 3: Bl R
E I I9. w > Lmlu
II II SH m “MI
E II 36 I Im
II II 9.6 m bMI
II II 36 I m av
II II Re I In
E II cod ‘I F7
II II Ed .I. wv

, I I. omd m m».
II I Bulb 3.
II I EIFIIFI
II II INI m Fl
II II II: m 2.
II .— mII I 3
II II “II .I. on
II II wad NI am
I II 3.0 A El!
II II Nmb I on
II II cod 0 mm?
I II CEO II 3“
II II who D Mm
. II mmI II mm I

Eon
MIDD mm mam ocIE< seamen-
wSD >Q$I

 

 

3.258 m 03:8

5523.5 352:3 rue-I88 :cmnc¢-2I.85=me..—£an IE: seamicao

98
Secondary structure prediction by Sequery and PHD

I have employed two new methods of secondary structure prediction
which have a higher degree of accuracy to analyze the N-subdomain, expecting
to find structural predictions compatible to our mutational and biophysical
analysis. Recently a new method of prediction, the PHD program, which uses a
computer neural network and yields a higher degree of accuracy was introduced
(Rost & Sander, 1994). The prediction by the PHD program was used in this
study, since it has a degree of accuracy (from 70% to 88%) for three-state
prediction, a helix, B sheet and loop (as high as the overall average for
homologous modeling). PHD uses the statistical information and the structural
information available in the present data base in conjunction with the
evolutionary information contained in multiple sequence alignments as input to
the neural network.

PHD prediction results show the input amino acid sequence, the overall
structural prediction for each amino acid residue and the detailed structure
predictions for each of the amino acids in the given sequence. PHD results are
shown in Tables 6 and 7. PHD did not suggest any strong prediction of a
particular secondary structure around the critical Phe residue. Predictions of the
secondary structure in the different homologues were made around the critical
Phe , but the reliability was equal or less than 5 from a scale of 0-9. In most cases
no predictions were made in this region. Both the N and C-terminus of the N -
subdomain in all homologues were strongly predicted to be a loop structure with
a reliability of 7-9 in most cases. A strong prediction of a helical structure was
found in a stretch of 6-7 residues N-terminal to the critical Phe in VZV ORF 10,
MDV UL48, and BHV-l UL48, but the degree of reliability drops dramatically at
the critical Phe residue and following residues (Tables 6 and 7).

Table 6. PHD secondary structure predictions of the N-subdomain of HSV-1
VP16, HSV-Z VP16, BHV-l UL48 and EHV-l Gene12. The amino acid
sequences of the N-subdomain of these homologues were subjected to the PHD
secondary structure prediction program (Rost et. al., 1994). Abbreviations:
secondary structure; H=helix, E=extended (sheet), blank=rest (loop); AA, amino
acid sequence; PHD, Profile network prediction Heidelberg; Rel, Reliability index
of prediction (0-9); detail: er, 'probability' for assigning helix; prE, 'probability'
for assigning strand; prL, 'probability' for assigning loop (note: the 'probabilites'
are scaled to the interval 0-9); subset: SUB, a subset of the prediction, for this
subset the following symbols are used, L, is loop (for which above " " (space) is
used), ".", means that no prediction is made for this residue, as the reliability is
Rel < 5

100

 

 

 

 

 

 

3355:............I.I_.I_.I...I_ mom Humansm _..IAAAAAAA.....EII.IIIIIITIII:I1I1I.I_ mom "ummnsm
mmmmImmmmmmmsmmmmmbmbmmmm_-IIQ . IIIIIIIIIIIIIIOCOOIImmIII_-Iua
OmmIIommmIoomsmvmNOOIONIO_-mua OOOOOOOOOOIoooooooooooooo_-mua
OIMIIIIIMIIIINNNNNIIIoooo_-mua OGOIIIOONImmmbmmmmmmmoooo_-mua

”IImImo "IImImo
IIIIIINIIMIIINIMIMIIIIIII_ Imm IIIIIIIIIINImpmmmmmwsmmam_ Imm
_ IIIIIIIIII IIEIII mmmm . DIE _ IIIImImImIIIIIIIIIII _ DIE
_mgmmHflImflIwmommfiEwnznqum. (d _mwfiqmmGAQmQQaHdgmgs 5m

Moooosoooomoooosoooonfioooo~oooo Moos.\oeooNoooosooooHoooosoooo
mm 5»an $80332 .386 3%. aIseI km 593 asaczai .3: Sim ”522?

_AAAAAAAAAA. .Aq..I.I..I..I...I..I..I.I..I_ mom "umeSm _AAAAAAAA. . .II. . ...I.I.I..I.I..I..I..I..I1I_ mom "ummnam
mmammmmmmsmIIIIIIIIIIIIII_-Isa mammmmIIIIINIIMIIIIIIIIII_-Iua
OOOOGOOOOOOOOOOOOOIOOOIIo_-msa oooooooooooooooooOOOOOOIo_-muq
ooooooooommmmmmmIIImmoooo_-mua COOOOImommIIIIINIIImmoooo_-mca

"IImumo "IImumc
Immmmmmmmmmmmmsmmwmsmmhsm_ Imm mmmmsmmmMIIIIINIIIIIIIIII_ Imm
_ _ QIInI _ 555$ _ DIE
_mmmmmkuogqnmnnfladmfizg. «In. _mommnonmgnanmnnngz <4

Mucous-000N000.sooooHoooosoooo Moo-osoonoNoooosooooHooooQOooo
_ mm 593 _.IsaoBaII 63> I->mII ”595 In 523 $523-2 SE, I.>mII ”52°F
22.8 I

.53”.— 65 «vs—D I.>IIII .3.—Ir ”arm: .3me I.>mI.I I: 52.3—55.2 05 I: 55260.5 9:523; hunt-I83. GE.—

9 «II—ah.

 

101

Table 7. PHD secondary structure predictions of the N-subdomain of EHV-4
GeneBS, MDV UL48 and VZV ORF10. The amino acid sequence of the N-
subdomain of these homologues were subjected to the PHD secondary structure
prediction program (Rost et. al., 1994). Abbreviations: secondary structure;
=he1ix, E=extended (sheet), blank=rest (loop); AA, amino acid sequence; PHD,
Profile network prediction Heidelberg; Rel, Reliability index of prediction (0-9);
detail: er, 'probability' for assigning helix; prE, 'probability' for assigning
strand; prL, 'probability' for assigning loop (note: the 'probabilites' are scaled to
the interval 0-9); subset: SUB, a subset of the prediction, for this subset the
following symbols are used, L, is loop (for which above " " (space) is used), ".",
means that no prediction is made for this residue, as the reliability is Rel < 5

102

 

_AQ.....A......mmmmmmmmng_ mDm ”uwmflfim
mbvmmMIIIIIIIIIIOOOOOIIII_-Iua
OIIII«Ioooooooooooooooooo_-mud
ooIIImmmmvmvammmmmmmmmoo_-mca

"IImImo
IIINNImmvmmIImsmmmmmmsmmm_ Imm
_ mum mmmmmmmm: . Gum
_0H0m<>DOmAmmOm<Q>>4OMBXmm_ <¢

Mose.sooooNoooosooooH0000§oooo

m 523 fiafigiéE ”59E.

 

 

 

_A.....3.mmmmmmmmm..mm.flqd_ mDm Iuwmflfim _Amm...m.m.........qaaqd.q_ mDm “ummflflm
mmNHHHHNHHNNNNHOOHNHHmmmm _ LIHQ mmflmamammvmvvmmmmobmbmbmm _ LIHQ
ommmwvaflmoooooooomaoooooo _ tuna thmwmmmwmfiamqvvvmooaoamo _ umHQ
OHNNNmmwwbhhmhmmmmohhmooo_anQ OOONNNHOHHNMNNNNNNHHHoooo_umhn

”Hflmummu "Hflmuwfl
mvmmvmmvmmmmmmbmwglbmmmm_ Hmm mmmvvmmvmamNHNNNNthwwwwm_ Hmm

_ mmmmmmmmmmmmmmmmmmm _ 9mm _ mmmmmmmm mmmm _ Gum

_09m>mHQMJJBNOmmmHBmAHDQDm_ <4 _WNHmmmdmflzwmflmm<>>00>nzfim. (d

MoooosoooomooooQOOOOHOOOO§OOOO Moo00\OOOONOOOO~0000H0000§0000
_ mm 523 fiaéfirvmmfib E aIseI mm 523 fiaczaz E ”5929

 

 

3.5—0 >N> can @943 >92 fine—59 vu>=m Io cfificgsmIL 2: I0 55260.3 0.5333 56:83 a:

b 039—.

 

103
A method termed Sequery was also used. Secondary structure prediction

by Sequery was done using a sliding window of four residues from residues 439
to 447 in HSV-l VP16. The residues allowed at each position were selected based
on the conserved residues from the homologue proteins. The residues allowed at
each position are-shown in Figure 1. The results are shown in Table 8. The
sample size indicates the number of proteins with matching sequence resulting
from the search (the list of the proteins is shown in the Appendix B). The
structures of the query sequence observed from matching proteins for region 439-
442 shows equal probability of a helix and loop / extended structure. This
suggests that the segment probably does not have strong preference for any
secondary structure. This region was followed by an extended loop or strand
structure, encompassing the critical Phe residue, residues 440-445, followed by a
strong a helical prediction at residues 444-447. These results suggest that the
residues around Phe 442 are likely to form an extended structure, either a loop or
strand, since the extended structure found did not have the classical hydrogen
bonding pattern of [3 sheet structures.

Discussion

Using HCA to compare the sequence of VP16 and its homologues, we
predicted the location of the activation domains of these homologues in a way
that was not obvious by the classical linear alignment method (Figure 4—5). We
tested our hypothesis using the VP16 homologue VZV ORF10. VZV ORF 10 was
the first homologue that was found to have the ability to activate transcription
(Moriuchi, et al., 1993). It functions as an activator despite the lack of the 80
amino acids at the C-terminus that were found to be necessary and sufficient for
transcription activation in HSV-l VP16 (Sadowski, et al., 1988; Triezenberg, et al.,

104

Table 8

Secondary structure prediction by Sequery

 

 

 

 

 

 

 

 

 

Search Amino Sample Secondary structure (%)
window acid size
residues
a helix [3 sheet Loop/ Turn
Extend
1 439-442 9 45 0 55 O
2 440-443 7 14 0 86 O
3 441-444 12 8 17 67 8
4 442-445 9 0 33 56 11
5 443-446 12 58 0 42 0
6 444-447 20 95 0 0 5

 

 

 

 

 

 

 

 

Table 8. Secondary structure prediction by Sequery. The favorable structure of
the string of sequence searched by Sequery was calculated. The sample size is
the number of similar strings of sequence that Sequery found from the search.
For each string, the corresponding protein structure was called up and viewed
using the computer program Insight 11 (Version 2.3.0, Biosym Technologies,
1993). The structure of the particular sequence string was recorded and used to
calculate the most favorable structure of the string of sequence. The percentage
of each structure found at each window of four residues is shown.

105
1988a). VZV ORF10 was predicted by HCA to contain a region homologous to

the N-subdomain of HSV-1 VP16. Deletion mutagenesis and transient
expression assays performed by Moriuchi et al. confirmed our hypothesis
concerning the location of the VZV ORF 10 activation domain.

Linear alignment, using HCA as a guide, suggested the critical residues in
VZV ORF10 that could be involved in the activation function of VZV ORF10.
Phe28 as well as conserved flanking hydrophobic residues were predicted to be
important for activation. Site directed oligonucleotide mutagenesis changing
Phe28 into other bulky hydrophobic residues, Tyr, Leu, and Ile, medium size
hydrophobic residue Val, or small hydrophilic residues Ser and Thr, decreased
the activity of GAL4-ORF10(5-79) from 50%-10%, respectively (Table 2).
Changing the flanking hydrophobic residues, Leu33 and Phe34 to smaller
hydrophobic or hydrophilic residue also decreased the activity of the fusion
protein (Table 5).

These results clearly demonstrate that Phe28 is indeed a critical residue
essential for VZV ORF 10 transcription activity. They also show the importance
of the flanking hydrophobic residues which were predicted by HCA and linear
alignment. These results confirm our hypothesis that the activation domain of
VZV ORF 10 lies at the N-terminus of the protein and is homologous to the N -
subdomain of HSV-1 VP16.

Since HCA proved to work well to identify the activation domain of VZV
ORF 10, I extended my predictions further to other VP16 homologues, BHV-l
UL48, EHV-l Gene12, EHV-4 GeneB5 and MDV UL48. I predicted the location as
well as critical residues that are likely to be involved in the activation function of
these proteins, shown in Figure 4-5. We predicted that the N and C activation
subdomains of BHV-l UL48 are separated, in contrast to HSV-l VP16, as shown
in Figure 6. The N-subdomain of BHV-l UL48 is predicted to lie at the N-

106
terminus of the protein (residues 22-46). Phe33 is predicted to be the critical Phe

residue homologous to Phe442 of HSV-1 VP16, with essential flanking
hydrophobic residues Ile30, Leu37 and Leu38 (Figure 4B). The C-subdomain of
BHV-l UL48 is located at the C-terminus of the protein (residues 486-505). Leu
residues at position 494, 499, 502 and 503, are predicted to be important for its
activity (Figure SB).

We also predicted the activation domains of EHV-l Gene12 and EHV-4
GeneBS to be separated, similar to BHV -1 UL48. The N-subdomains of EHV-l
Gene12 and EHV-4 GeneB5 are located at the N-termini of the protein. EHV-l
Gene12 N-subdomain (residues 53-77) has the critical Phe residue at position 64
and EHV—4 GeneB5 N-subdomain (residues 28-52) has the critical Phe residue at
position 39 (Figure 4B). The C-subdomain of EHV-l Gene12 and EHV-4 GeneBS
are located at the C-terminus of the protein. EHV-l Gene12 C-subdomain
(residues 460—479) have conserved bulky hydrophobic residues Met468, Leu473,
Met476 and Phe477 and EHV-4 GeneB5 C-subdomain (residues 435-454) have
conserved bulky hydrophobic residues Ile443, Leu448, Met451 and Phe452
(Figure 5B).

Previous studies by Peter Horn and Susan Sullivan indicated the
importance of the Glu residue 474 in HSV-l VP16 for activity, therefore, Gln
residue 500 in BHV-l UL48, residue 474 in EHV-l Gene12 and residue 449 in
EHV-4 GeneBS, could also be essential for the activity of these C-subdomains.
There are also other conserved hydrophobic, acidic and polar residues within
this C-subdomain that might also be important.

Recent studies in BHV-l UL48 demonstrated that the C-terminal portion
of BHV-l UL48 was able to activate transcription in a GAL4 fusion context.
However, deletion of these amino acids from the C-terminus of BHV-l UL48 did
not totally eliminate the ability of the protein to transactivate IE-l, a major BHV-l

107
a gene promoter (Misra, 1994). These data suggest that BHV-l UL48 may not be

functionally collinear with HSV-l VP16, since transactivation of BHV-l UL48
requires both the C-terminus and another region present within the N-terminal
portion of the protein. Recently Misra et al. has shown that the N—terminus of
BHV-l UL48 indeed contains another activation domain (personal
communication), consistent with my prediction.

Studies on the EHV-l Gene12 homologue of VP16 were also performed.
The activation domain of EHV-l Gene12 was mapped to the C-terminal end of
the protein (Elliott, 1994). A Met residue at position 476 seems to be essential for
its activity.

A prediction of the MDV UL48 activation domain location and critical
residues was also made. MDV UL48 is similar to VZV ORF10 in that it bears
only one of the two subdomains, the N-subdomain, which is located at the N-
terminus of the protein. The critical Phe residue is located at position 43 with
flanking hydrophobic residues Ile40 and Leu48 and Leu49.

This hypothesis has been tested by Mekki Boussaha in Dr. Paul Coussens
group (Molecular Virology Laboratory, Department of Animal Science, Michigan
State University). The predicted critical Phe residue 43 was substituted with Tyr,
Ala or Pro in the context of wild type MDV UL48 protein. The mutant protein
was tested in a transient expression assay. The expression of wild type or
derivatives of MDV UL48 was driven by a CMV promoter and the MDV ICP4
promoter linked to the CAT gene was used as a reporter plasmid. When Phe43
was substituted with Ala or Pro, the activity of MDV UL48 dropped to less than
10% of the wild type protein. The activity of the mutant carrying the Tyr
substitution, on the other hand, was comparable to that of wild type MDV UL48
protein (M. Boussaha, manuscript in preparation).

108
Our results together with the findings from other laboratories

investigating homologues of VP16 generally supported the predictions of
activation subdomains made by HCA, except for the N subdomain of EHV-l
Gene12 predicted by HCA, which did not have detectable activity in a study by
Elliott and co-workers. Interestingly, the activation domain of VP16 homologues
appear separated in all homologues other than HSV-l and HSV-Z. The N-
subdomain is located at the N -terminus of the protein while the C-subdomain
remains at the C-terminus, as summarized in Figure 6. VZV ORF10 and MDV
UL48 are different from other homologues in that they lack the C-subdomain.

These findings raise interesting ideas in terms of the structure and
function of VP16 and is homologues. It suggests that transcription activators
may be equipped with more than a single activation domain. These domains
could be located either in tandem or at separate locations in the protein; even at
different termini of the protein We may also perceive the nature of these
activation domains as "activation modules" that are self-sufficient and work at
different sites within the protein. The results from HCA and mutagenesis of
VZV ORF 10, also suggest that the N subdomains from VP16 homologues are
likely to have a similar mechanism of transcription activation. However, the
mechanism by which these domains work when they are placed in tandem or
when they are at opposite ends of the protein may or may not be identical. This
is also a very interesting aspect for further investigation.

Knowledge of the structural aspects of a protein is always needed to fully
understand the function and mechanism of how a protein works. Activation
domains from transcription activator proteins have been an interesting specimen
for protein chemists, because of their essential role in gene regulation of living
organisms. However, most efforts to determine the basic secondary structure of

an activation domain have been unsuccessful. NMR or X-ray crystallographic

109
analysis has not determined the structure for any activation domain. The

domain is believed to be very mobile and unstructured in solution (Shen, et al.,
1996a; Shen, et al., 1996b; O'Hare 8: Williams, 1992; Donaldson 8: Capone, 1992).
Since biophysical means have not yielded any structural information for
transcription activation domains, structural predictions based on a group of
known transcription activation domains might give insight into this question.

Earlier attempts have been made to predict the secondary structure of the
activation domain of VP16. The activation domain of HSV-1 VP16 between
residues 421-444 was predicted to be a helical by the classical Chou-Fasman and
Garnier-OsguthorpeoRobson method (W. D. Cress unpublished results). The a
helical structure was predicted to be amphipathic and this structural feature was
proposed to be essential for the activity of the VP16 N subdomain. In the process
of testing the hypothesis Doug Cress introduced Pro residues, via site directed
oligonucleotide mutagenesis, substituting His residue 425, and Ala residues 432
and 435, simultaneously. The double substitution, Ala residue 432 and Ala
residue 435, was constructed because this region was strongly predicted to
presume on helical structure. Neither the single or double substitution had any
detectable effect on activation by the VP16 N-subdomain. This demonstrates that
a helical structure of the N-subdomain is not likely to be essential for its function.
Therefore, it is very unlikely that the N-subdomain assumes a helical structure.
This structural prediction for HSV-l VP16 was contradictory to the mutational
analysis results (Cress 8: Triezenberg, 1991b).

Taking advantage of the collection of N subdomain homologues identified
by HCA, we subjected the sequences of the N-subdomain of all the VP16
homologues to several currently available secondary structure prediction
programs, including the classical Chou-Fasman and Garnier-Osguthorpe-Robson
method.

110
The Chou-Fasman method established the statistical conformational

potential of all the amino acid residues to form a helix, B sheet and turns , from a
set of 15 to 29 proteins. These values are arranged in their hierarchical order and
used for the prediction of the structural conformation of a primary sequence
given. The structure determination by this method is made by scoring these
assigned statistical values, according to a specific set of rules, on a given segment
of the protein (Chou 8:: Fasman, 1978).

The Garnier-Osguthorpe-Robson method assigns a window of 17 residues,
encompassing 8 residues N -terminal (-8) and 8 residues C-terminal (+8) of the
position to be predicted. The statistical conformation potential was calculated for
each residue at all of the 17 position within the window (from a set of 26
proteins). The statistical values of a particular residue are different depending
on the position of the residue within the window. These values are used in the
Garnier-Osguthorpe-Robson predictions (Garnier, et al., 1978). Therefore the
prediction by Garnier-Osguthorpe-Robson takes into account the effects of the
position of the residues from -8 to +8 upon the conformational potential of each
residue, where the Chou-Fasman prediction has a fixed conformational potential
for each residue. However, both methods have a comparable degree of accuracy.

The Chou-Fasman and Garnier-Osguthorpe-Robson methods predicted
the sequence around the critical Phe residue in most of the homologues to be on
helical follow by a turn The only exception was EHV-4 GeneB5 where the Chou-
Fasman method predicted the sequence around the critical Phe residue to be B
sheet. Though the Chou-Fasman and Garnier-Osguthorpe-Robinson methods
are attractive in terms of their simplicity and acceptable degree of accuracy, the
statistical values were based on the calculation of only 15-29 proteins and
predictions were also biased by the algorithm used.

111
An interesting prediction of surface probability according to the Emini et

a1 . method was made, suggesting the critical Phe residue in all the homologue
proteins is highly exposed to solvent. This result does agree with results from
fluorescence studies by Fan Shen (Shen, et al., 1996a).

The two currently available programs, PHD and Sequery, gave some
interesting data but did not give a common prediction. The PHD program did
not give any reliable secondary structure prediction within the region of the
critical Phe residue. In most cases the program did not assign a specific structure
to the residue, since it found that particular sequence to have probabilities for
either a helical, B sheet or loop. Moreover, PHD could not find similar sequence
from the available data base that matches with our sequence. This lowers the
degree of accuracy of the prediction by PHD to less than 72% for the three-state
prediction. We conclude from the results given by PHD that in most cases the
region surrounding the critical Phe residue does not have a strong preference to
fold into any particular secondary structure.

Sequery gave a unique and interesting prediction of the domain in the
region of the critical Phe residue. It suggested an extended structure or loop
encompassing the critical Phe residue followed by a strong a helical prediction
(Table 8). This data fits well with the biophysical data currently available.
Instead of converting the conformational information of the protein into
statistical values using a limited set of proteins, Sequery searches for the exact
match of the query sequence from the current data base of protein of know
structure (more than 1300 proteins). This method would eliminate the bias
imposed on the predictions caused by the limited set of proteins used to generate
the statistical values as well as potential flaws in the prediction methods.

These predictions imply that the primary structure of the N-subdomain
might not have a strong preference for any specific secondary structure. In this

112
respect these data agree with the previous biophysical studies (Shen, et al., 1996a)

as well as mutational analysis (Cress 8: Triezenberg, 1991b). It is very plausible
that the N subdomain of VP16 and its homologues assumes a loop or extended
structure.

From all the information available currently, it is most likely that the N-
subdomain of VP16 and its homologues are flexible in their structure. The
critical Phe residues are likely to be exposed to solvent. The structure of this
domain might be expected to be induced upon interaction with its target
proteins. Fan Shen has shown by anisotropy decay experiments that upon
addition of TBP to the VP16 activation domain, a more ordered structure of both
the N and C subdomain could be induced. Additionof TFIIB, another target of
the VP16 activation domain, could also induce a slight change in the C-
subdomain (Shen, et al., 1996b).

Future studies

Though mutational analysis and preliminary work from other laboratories
strongly suggest that my predictions are correct, thorough mutational analysis of
the proposed activation domains both in the context of fusion proteins and wild
type protein should be investigated for the other homologues of VP16. In
particular the homologues in EHV-l and EHV-4 where initial results have not
identified the N-subdomain of the protein need further investigation. The
mechanism by which these homologues activate transcription could be
elucidated, for instance, by identification of their target proteins, structural
determination of the activation domain by itself or in a complex with its target
protein in biophysical studies, or by crystallographic studies. The results from
one homologue might help explain or raise a testable hypothesis for the others.

113
Swapping of the activation domains or subdomains of the homologues, with

careful construction, would also be interesting to show that these domains are
homologous and could be exchanged.

It would also be very interesting to expand the search for similar
activation domains in other transcription activators using patterns of
hydrophobic residues from HCA. My preliminary attempts have identified
portions of the activation domain of GCN4, residues 107-136, and the p53
activation domain, residues 48-66, which have striking similarity to the C-
subdomain of VP16 (data not shown). Recent mutational analysis of this GCN4
activation domain has shown that all residues within the hydrophobic clusters
predicted by HCA to be the core of this activation domain are indeed important
for its activation function (Drysdale, et al., 1995). Some of the critical
hydrophobic residues in the region between residues 90-120 of GCN4 have also
been correctly predicted by Doug Cress (Cress 8: Triezenberg, 1991b). However
for p53, the hydrophobic residues that I have predicted to be important for
activation have not yet been studied, although some hydrophobic residues in the
N-terminus , Leu residue 22 and Trp residue 23, were shown by Lin et al. to be
important for the activation function of p53 (Lin, Chen, Elenbaas, 8: Levine, 1994;
Lin, Teresky, 8: Levine, 1995).

I have shown that HCA is a powerful tool for recognition and
differentiation of the two different groups of activation domains in herpes
homologues of VP16. The pattern of hydrophobic residues of the activation
domain might also prove to be a very useful criterion to characterize and group
the activation domains of different transcription activator proteins discovered in

the future.

APPENDICES

APPENDIX A

114
APPENDIX A

Hydrophobic cluster analysis predicts an amino-terminal domain
of varicella-zoster virus open reading frame 10 required for
transcriptional activation

(herpesvirus / VP16/ transcription)

HIROYUKI MORIUCHI", MASAKO MORIUCHI", RATH PICHYANGKURA“,
STEVEN J. TRIEZENBERG",
STEPHEN E. STRAUS‘, AND JEFFREY I. COHEN"

*Medical Virology Section, Laboratory of Clinical Investigation, National
Institute of Allergy and Infectious Diseases, Bethesda, MD 20892; and
*Department of Biochemistry, Michigan State University, East Lansing, MI 48824-
1319

Communicated by Steven L. McKnight, Tularik, Inc., South San Francisco, CA,
July 5, 1995

ABSTRACT Varicella-zoster virus open reading frame 10 (ORF10) protein, the
homolog of the herpes simplex virus protein VP16, can transactivate
immediate-early promoters from both viruses. A protein sequence comparison
procedure termed hydrophobic cluster analysis was used to identify a motif
centered at Phe-28, near the amino terminus of ORF10, that strongly resembles
the sequence of the activating domain surrounding Phe-442 or VP16. With a
series of GAL4-ORF10 fusion proteins, we mapped the ORF10 transcriptional.
activation domain to the amino-terminal region (aa 5-79). Extensive
mutagenesis of Phe-28 in GAL4-ORF10 fusion proteins demonstrated the
importance of an aromatic or bulky hydrophobic amino acid at this position, as
shown previously for Phe-442 of VP16. Transactivation by the native ORF10
protein was abolished when Phe-28 was replaced by Ala. Similar amino-
terminal domains were identified in the VP16 homologs of other
alphaherpesviruses. Hydrophobic cluster analysis correctly predicted
activation domains of ORF10 and VP16 that share critical characteristics of a
distinctive subclass of acidic activation domains.

 

115

Transcriptional activators play an important role in the regulation of
eukaryotic genes and typically contain two domains: one conferring specific
association with promoter sequences, usually a DNA-binding domain, and a
second domain governing transcriptional activation. Several different types of
activation domains have been identified and classified on structural grounds as
acidic, glutamine-rich, or proline-rich The functional significance of these
classifications, however, is unclear. Indeed, recent studies ( 1, 2) demonstrated
that not all activation domains of a given class interact with the same target;
therefore, each of these classes may include several functionally and structurally
distinct subclasses.

Genetic studies demonstrated that a net negative charge is not sufficient
for transcriptional activation by acidic activation domains of proteins like VP16
(3) and GAL4 (4). Alignment of amino acid sequences of several transcriptional-
activation domains also suggested the importance of bulky hydrophobic residues
(3). Detailed mutagenesis studies demonstrated that the presence of bulky
hydrophobic amino acids interspersed among acidic amino acids is critical for
the transcriptional-activation ability of VP16 (5, 6).

VP16, a prototypical acidic activator, is a component of the tegument of
herpes simplex virus type 1 (HSV-1) and type 2 (HSV-2) particles and stimulates
expression of viral immediate-early (IE) genes (7). Varicella-zoster virus (VZV)
open reading frame 10 (ORF10) encodes a homolog of VP16. The ORF 10 product
is also a tegument protein (8) and transactivates the VZV and HSV IE promoter
in transient expression assays (9). These proteins share considerable amino acid
sequence homology; however, the ORF10 protein is 80 amino acids shorter,
lacking the carboxyl-temiinal acidic tail known to be critical for transactivation
by VP16 (10, 11).

To define the structural basis of activation by ORF10, we used a
computerized protein sequence comparison procedure termed hydrophobic
cluster analysis (HCA) (12, 13). Rather than selecting sequences according to
optimal amino acid similarities, the HCA method compares clusters of
hydrophobic residues presumed to constitute hydrophobic cores in globular
proteins. Using HCA, we identified a motif centered at Phe-28 near the amino
terminus of ORF 10 protein that strongly resembles the motif surrounding Phe-
442 of VP16, a residue known to be critical for the transactivating ability of the
carboxyl-terminal region of VP16. Using fusion proteins bearing the DNA-
binding domain of GAL4, we mapped the ORF 10 transcriptional-activation
domain to the amino terminal region (aa 5—79). Detailed mutagenesis studies of
ORF 10 support the HCA-based hypothesis that the activation domain of ORF10,
like that of VP16, depends upon a specific pattern of bulky hydrophobic residues
for activation of transcription. The HCA method also predicts the presence of
similar domains in the amino termini of the VP16 homologs (including VP16
itself) from a number of other alphaherpesviruses. Our results demonstrate that
the activation domains of VZV ORF10 and HSV-1 VP16 share critical structural
features characteristic of a distinctive subclass of acidic activation domains.

MATERIALS AND METHODS

116

Plasmids. Plasmids expressing GAL4 ORF 10 fusion proteins were
constructed by ligating fragments from pMTPORFlOR (9) to plasmid pGAL4
(14) or pGAL4E (15) to maintain an open reading frame. Single amino acid
substitution mutants in the activation domain of ORF 10 were constructed by
oligonucleotide-directed mutagenesis (as in ref. 3) of the EcoRI restriction
fragment of GAL4-ORF10(5-79). Wild-type and amino acid substitution mutants
of VP16 residues 1-25 were constructed by ligating double-stranded synthetic
oligonucleotides into plasmid pGAL4.

An ORF10 amino acid substitution mutant (pCMV10F28A) was
constructed by using PCR-oriented mutagenesis (16). The VP16 deletion mutant
was constructed by digesting pCMV 16 (17) completely with BamHI and partially
with SalI, isolating a 5.9-kb fragment, blunt-ending the DNA fragment with DNA
polymerase I (Klenow fragment), and ligating it to a double-stranded
oligonucleotide (CTAGTCTAGACTAG). An expression vector for the GAL4-
VP16 fusion protein (18) and reporter plasmids p62CAT (19), pSGAL4-E1b-CAT,
and pIGAL4-E1b-CAT (20) were described previously.

Preparation of Antiserum to ORF10 Protein. 'Plasmid pGEXlO(41 410)
was constructed to express a glutathione S-transferase (GST) fusion protein
bearing amino acids 41410 of ORF10. A BamHI—EcoRV restriction fragment from
pGORFlO [ref. 21; nt 12,280-13,636 of VZV (22)] was inserted into pGEX-2T
(Pharmacia) that had been previously digested with BamHI and SmaI. The GST-
ORF 10 fusion protein, expressed in and purified from Escherichia coli strain DH5
cells, was used to raise anti-ORF10 antisera in rabbits.

Transfections, Immunoprecipitations, and Immunoblotting. To assess
the transcriptional function of various activators, Vero cells were transfected, and
chloramphenicol acetyltransferase (CAT) assays were performed, as described
(23). To test expressed levels of the activating proteins, COS cells were
transfected with plasmids encoding GAL4 fusion proteins and cell extracts were
immunoprecipitated with anti-GAL4(1-147) antibody, as described (24). For
immunoblotting, extracts of MeWo cells transfected with pCMV, pCMV 10, or
pCMV10F28A were electrophoresed, blotted, and then incubated with anti-
GST10(41-410) antibody.

HCA. The sequences of VP16 and its homologs were processed by using
an HCA program provided by B. Henrissat (Centre National de la Recherche
Scientifique, Université Joseph Fourier, Grenoble, France).

RESULTS

The Amino-Terminal Region of ORF10 Protein Contains the
Transcriptional-Activation Domain. To identify amino acid sequences of ORF10
that can function as transcriptional-activation domains, portions of the ORF 10
protein were fused to the GAL4 DNA—binding domain (aa 1-147). These
constructs (activator plasmids) were assessed in transient expression assays for

117

their ability to activate transcription from a reporter plasmid containing five
tandem GAL4-binding sites upstream of the adenovirus ElB TATA box and the
CAT gene.

Analysis of a series of GAL4-ORF 10 fusion constructs showed that VZV
ORF 10 protein contains a potent transcriptional-activation domain near its amino
terminus (Fig. 1).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MIN .3“
1 410
VZV 9““ —
Activabn: Fold Adlvalon
GAL4 F'EIBW - l " 'l 1
alumina (287-410) CG—i—'Au (M4 ) 4 w_"0 1.3
GAL4-0mm 113351) am My) *3“ 1.5
GAL4-ORF10( 5-175) Tau HIT—i. ) 120 s 20
GAL4-ORF10(5-79) m" 310::90
GAL4-ORF10(5-61) ‘ ALAN-M7) 5 ‘ mu
GAL4-ORF10 (61-175) "TAU <—'7"_1-u ) _7‘ 11: 4
GAL4-ORF10( 5-41) W“ 15 a s
GAL4-emu (as-7s) mu l 1.147) 3" _7“ 150 1 so
GAL4-VP16 (us-490) TAU—(TM ) "3 m 090: so
Roporbt: 56AM
kind at.
soauewcm <7 I )
m3 i C"

 

 

 

 

 

FIG. 1. Localization of the ORF10 transcriptional-activation domain. Vero cells were
cotransfected with 5 ug of plasmid expressing a GAL4-ORF10 fusion protein and 5 ug of reporter
plasmid containing five GAL4-binding sites, and cell lysates were assayed for CAT activity. Fold
activation is the CAT activity relative to that obtained for GAL4 alone. Means and standard
deviations were calculated from at least five independent transfections.

When fused to the GAL4 DNA-binding domain, an amino-terminal fragment (aa
5-175) of the VZV ORF 10 protein activated transcription, while the middle or
carboxyl-terminal portions of the protein failed to activate transcription.
Successive carboxyl-terminal deletions within the activating fragment of ORF 10
indicated that maximal activity was conferred by amino acids 5-79. Further
deletion from the carboxyl terminus (aa 42-79) of GAL4-ORF10(5-79) abolished
most of its activity, while deletion from the amino terminus (aa 5 37) reduced
activity by 50%. While the carboxyl portion (aa 38-79) of GAL4-ORF10(5-79) had
substantial activity, a single amino acid substitution within the amino-terminal
portion (aa 28) abolished activity of the GAL4 construct and the native ORF10
protein (see below). The level of transcriptional activation seen with GAL4-
ORF10(5-79) was about 50% of that obtained with the potent activator GAL4-
VP16. GAL40RF10(5-79) did not transactivate a reporter plasmid lacking
GAL4-binding sites, indicating that the activation was specific (data not shown).
All of the fusion proteins were expressed at comparable levels in transfected cells

118

when assayed by immunoprecipitation with anti-GAL4 antibody (data not
shown).

The HCA Method Detects a Motif Centered at Phe-28 of ORF10 That
Resembles the Motif Surrounding Phe-442 of VP16. A previous alignment of
the amino acid sequences of ORF10 and VP16 noted a number of highly similar
regions, but failed to identify a domain similar to the VP16 activation domain
(25). Given the recent evidence that the ORF10 protein is indeed a transcriptional
activator (9), the complete amino acid sequences of ORF 10 and VP16 were
analyzed by using the HCA method in which the size, shape, and orientation of
clusters of hydrophobic residues were compared. We identified a motif
surrounding Phe-28 near the amino terminus of ORF 10 that strongly resembles
the motif centered at Phe-442 within the VP16 carboxyl-terminal activation
domain (Fig. 2,4). The HCA plot (Fig. 2B) indicates that both of these motifs have
horseshoe-shaped arrays of hydrophobic amino acids with Phe residues at their
apices. As shown clearly in the linear sequence alignment (Fig. 2C), in both
proteins, the hydrophobic residues are interspersed with numerous acidic (Asp
and Glu) and potentially phosphorylated (Ser and Thr) residues.

Importance of Aromatic or Bulky Hydrophobic Amino Acids at Position
28 of ORF10 Protein. Previous studies demonstrated that mutations of Phe-442
of VP16 markedly influence transactivation level (3, 5). Substitution of aromatic
(Tyr or Trp) or bulky hydrophobic (Leu, Ile, or Met) amino acids for Phe-442
decreased but did not abolish function, while substitution with most other amino
acids reduced; the transactivating activity of a truncated form of VP16 by 290%
(5). To assess whether an aromatic or bulky hydrophobic amino acid at position
28 of ORF10 (which aligns with VP16 Phe-442 by HCA) is also important for
transactivating activity, Phe-28 was replaced with other amino acids in GAL4-
ORF10(5-79) fusion proteins, and these proteins were assessed for their ability to
transactivate p5GAL4-E1b-CAT. All constructs tested were shown to be
expressed in transfected cells by immunoprecipitation with anti-GAL4 antibody
(data not shown).

Substitution of Phe-28 with Tyr, another aromatic amino acid, or with
either Leu or Ile, two bulky hydrophobic amino acids, reduced the activity of the
wild-type construct by 50% (Table 1). Substitution with smaller hydrophobic
amino acids (Val, Ala, or Pro) reduced the activity by 80 90%, while substitution
with hydrophilic amino acids (Ser or Thr) reduced the activity by >90%. These
findings implicate an aromatic or bulky hydrophobic amino acid at position 28 as
being important for the activation function of ORF 10.

119

a. I HIE

VP16 Type 1&2

[II I

VZV ORF10 8': MDV-UL48

[If [a

BHV-l UL48

 

 

 

 

 

CITE 107

EHV-l GENElZ 8: EHV-4 GENRES

 

VP16-1 VAMAHADALDDFDLDMLGDGDSPGP

455
VP16-2 XZDMFFPADALDDPDLEMLGD VESPSPS

VZVORPIO 11378K;_1‘JIEZQAV'VDAF'DElSI:I"GD ASDIG:

Riv-101.48 ZPSTMSPYDAIEAPDDSLLGS P LAAGP:

MDVUL48 3IDZDLDILRTIEIszli'mla'rlu. """""" ' ”EMVRT

BIN-1 mu SELMDMDGAVASEDEGMLSA "Eisvys

xiv-4 mszréLxdvocvvasginscunsas ______ SI vs:
FIG. 2. HCA of VP16, ORF10, and other VP16 homologs. The complete sequenc: of HSV-1
VP16 (25), HSV-2 VP16 (26), ORF10 (22), equine herpesvirus I (EHV-l) gene 12 (27), EHV-4 gene
B5 (28), bovine herpesvirus 1 (BHV- -1) UL48 (29), and Marek disease virus (MDV) U148 (30)
proteins were analyzed by HCA, and locations of activation domains were predicted by using
VP16 as a standard, (A) Bars represent the approximate relative sizes of VP16 and its homologs.
The VP16 activation domain comprises two subdomains designated N (black box) and C (open
box). The amino-terminal black box of VP16 represents a cryptic activation domain. (B) HCA
profiles (I 2) of subdomains resembling VP16N show the characteristic horseshoe-shaped cluster
and the centrally located Phe residues. The amino acid sequences are represented (in duplicate)

along diagonals from upper left to lower right with the one—letter amino acid code except that
black diamonds indicate glycine, squares indicate threonine, squares with central dots indicate

 

 

 

120

serine, and stars indicate proline residues. Sets of adjacent hydrophobic residues are
encompassed by contour lines. (C) Linear alignment of the subdomains shown in B, with shading
indicating conserved residues, boldface letters indicating conserved bulky hydrophobic residues,
dark shading denoting con served acidic residues, and open ovals representing residues
conserved among homologs but not present in HSV—1 VP16.

Table 1. Relative activity‘of GAL4-ORF10(549) bearing amino acid substitutions at position 28

 

 

Residue at position 28
Relative activity, %

Tyr 52 t 17
Leu 54 :l: 17

Ile 50 :t: 12

Val 23 i 4

Ala 11 :t 2

Pro 13 i 3

Ser 9 t 4

Thr 7 r: 2

 

Vero cells were cotransfected with 5 mg of plasmid expressing an amino acid substitution mutant
of GAL4 ORF10(5-79) and 5 mg of a reporter plasmid containing five GAL4-binding sites. Data
are presented as the percent CAT activity of the mutant relative to that obtained for wild-type
GAL4-ORF10(549). Means and standard deviations were calculated from at least five
independent transfections.

Residues Flanking Phe-28 Also Influence Activity. Several hydrophobic
residues (V a1-25, Len-32, and Phe-33) flank Phe-28 in the horseshoe-like
hydrophobic cluster (Fig. 2). Some of the corresponding amino acids in VP16
(Len-439 and Leu-444) have previously been shown to influence its function (5).
To determine whether these residues also contribute to the activity of ORF10,
site-specific mutations were made in ORF10(5-79). While substitution of Val-24
(not in the horseshoe-like cluster) by Ala did not reduce the activity, substitution
of Leu—32 (in the horseshoe-like cluster) by Ala reduced activity to 45% of wild
type (Table 2). Substitution of Phe-33 (in the horseshoe-like cluster) by Ser
markedly reduced activity, while substitution of Phe-33 by Tyr, another aromatic
amino acid, retained 80% of the wild-type activity. Contrary to our expectation,
substitution of Val-25 (in the horseshoe—like cluster) by Ala did not reduce
activity. Each of the constructs was expressed at comparable levels in transfected
cells when assayed by immunoprecipitation (data not shown). Thus, while some
of the hydrophobic residues flanking Phe-28 contribute to the transactivating
ability of GAL4-ORF10(5-79), other residues may not be essential for activity.

The carboxyl-terminal portion (aa 62-79) of the ORF10 transcriptional-
activation domain (aa 5-79) also contains any hydrophobic residues but does not
form a characteristic horseshoe-like cluster. In contrast to the hydrophobic
residues that flank Phe-28, substitution of any of the hydrophobic residues in the
carboxyl-terminal portion of the activation domain (Ile-65, Len-66, Tyr-67, Leu-
70, Ile-71, or Len-74) by Ala did not reduce the activity of GAL4-ORF10(S-79)
(data Dot shown). These results suggest that none of these hydrophobic residues
is critical for the transactivating function of the domain.

121

Table 2. Relative activity of GAIA-ORFIO(5-79) bearing amino acid substitutions at hydrophobic
residues flanking Phe-28

 

 

Amino acid substitution Relative activity, %
Val-24 -> Ala 140 i 40
Val-25 -> Ala 130 t 50
Len-32 --> Ala 45 i 18
Phe-33 -> Set 12 :t 4
Phe-33 -> Tyr 76 :l: 18

 

Mutant activation domains were tested as described in the legend to Table 1.

A Motif Surrounding Phe-28 Is Essential for Transactivation by Native
ORF10 Protein. To determine if the region surrounding Phe-28 is required for
transactivation by native ORF10 protein, we mutated the ORF10 gene at this site
and assessed its ability to transactivate the VZV ORF62 target promoter.
Substitution of Phe-28 with Ala abolished transactivation by ORF10 (Table 3).
An immunoblot analysis indicated that the ORF 10 substitution mutant construct
(pCMV10F28A) was expressed in transfected cells at levels comparable to that of
native ORF 10 (pCMV10) (data not shown). As previously shown (10, 11),
deletion of the carboxyl-terminal region of VP16 abolished transactivation. These
results indicate that a motif centered at Phe-28 is critical for transactivation by
native ORF 10 protein.

Table 3. Phe-28 is required for transactivation of the VZV IE62

 

 

promoter by ORF10 protein
Activator Fold induction
pCMV10 24 j: 3
pCMV10F28A 1.0 i 0.1
pCMV 16 290 :t: 50
pCMV16d413-490 0.8 1- 0.1

 

Cells were cotransfected with plasmids expressing native Of (pCMV10), an amino acid
substitution mutant of ORF10 (pCMV10F28A), full-length VP16 (pCMV16), or a carboxyl-
terminal truncation of VP16 (pCMV16d4l3-490) together with p62CAT (the VZV ORP62 promoter
followed by the CAT gene). Fold induction is the CAT activity in transfected cell extracts relative
to that obtained when using pCMV (vector control). Means and standard deviations were
calculated from at least five independent transfections.

The HCA Method Predicts That VP16 Homologs of Other
Alphaherpesviruses have Transcriptional-Activation Do. mains in Their
Amino Termini. Using HCA, we analyzed the structures of diverse VP16
homologs from other alphaherpes-viruses. VP16 has two distinct transcriptional-
activation subdomains near its carboxyl end (Fig. 2A). The first of these
subdomains surrounds Phe-442 and has been designated HI (6) or VP16N (2).
The HCA method predicts the presence of similar subdomains in the amino-
terminal region of each of the alphaherpesvirus VP16 homologs examined (Fig.

122

2A and B). Each amino-terminal subdomain has the characteristic horseshoe-like
shape of hydrophobic amino acids with a Phe residue at the apex and
interspersed with acidic amino acids (Fig. 2B). The most conserved features of
this subdomain are the Phe, a subsequent Asp, and bulky hydrophobic residues
three positions to the amino-terminal side and four and five positions to the
carboxyl-terminal side (Fig. 2C).

At the extreme carboxyl-terminal region (aa 452-490), VP16 has a second
activation subdomain, previously designated H2 (6) or VP16C (2). The HCA
method predicts the presence of similar carboxyl-terminal domains in BHV-l
UL48 and EHV—l gene 12 proteins but not in VZV ORF10 or MDV UL48 proteins
(Figs. 2,4 and 3 A and B). The five similar domains include a pattern of
hydrophobic residues interspersed with acidic or polar amino acids. Both
alanine-scanning and random mutagenesis of the VP16C subdomain indicate
that the positions in this pattern that are most conserved among the homologs
are also the most critical residues in this region (P. Horn, 5. Sullivan, V. Olson,
and S. T., unpublished data). The extreme carboxyl-terminal domains of BHV-l
UL48 and EHV-l gene 12 proteins are important for transactivation (31, 32).

   

HNdGan EN4GNHS IMWUL“

""" 490

TDVLD SSIS

PFGGT DALL
«as

 

Toonogssnr

 

FIG 3. (A) HCA profiles of subdomains resembling VP16C, drawn as in Fig. 2B. (B) linear
sequence alignment of the subdomains shown in A, with shading, ovals, and boldface type as
described in the legend of Fig. 2C.

The VP16 Amino-Terminal Region Contains a Weak Transcriptional-
Activation Domain. The previous analysis suggests that among the
alphaherpesviruses, the HSV VP16 proteins are distinct in having a VP16N-like
subdomain as part of the larger carboxyl-terminal activation domain, whereas in
all other VP16 homologs, this subdomain is found near the amino terminus.
Deletion mutations removing the carboxyl-terminal 80 amino acids of VP16

123

abolished transcriptional activation, suggesting that amino-terminal sequences of
VP16 are not strong activators in the native protein (10, 11). While HCA profiles
of residues 1-40 of VP16 show little obvious similarity to the amino-terminal
regions of ORF 10, direct alignment of the sequences indicated some potential
similarities (Fig. 4). To test the hypothesis that residues 1-25 of VP16 make up a
cryptic activation domain, this region was fused to the GAL4 DNA-binding
domain, and transcriptional activation was assessed in transient expression
assays. GAL4-VP16(1-25) activated transcription up to 90-fold, compared with a
690-fold activation by GAL4-VP16(413-490) (data not shown).

14
N A

...A S

38

FIG. 4. Comparison of amino terminal sequences of VP16 and ORF10. Shading indicates identical
or conservatively changed amino acids.

 

To determine whether the VP16 amino-terminal domain has structural
properties similar to those seen with the VP16 carboxyl-terminal or ORF10
amino-terminal domains, amino acid substitutions were made at Leu-4.
Substitution of Len-4 by the aromatic amino acids Trp or Phe augmented activity,
while substitution by the smaller hydrophobic amino acid Ala or the hydrophilic
amino acid Thr virtually abolished activity (Table 4). Replacement of Met-1, Leu-
8, and Phe-9, which flank Len-4 and make up a characteristic horseshoe-like
cluster in this domain, with Ala nearly abolished activity. Thus, the properties of
the cryptic VP16 amino-terminal domain somewhat resemble those of the VP16
carboxyl-terminal and ORF 10 amino-terminal domains.

Table 4. Relative activity of GAL4-VP16(1-25) mutants bearing various amino acid substitutions

 

 

Mutation in VP16 (1-25) Relative activity, %
Len-4 -> Trp 210 i 20
Len-4 -> Phe 150 i 40
Len-4 ~> Ala 55 :t 1.7
Len-4 -> Thr 6.8 i 1.0
Met-1 -> Ala 8.4 i 2.8
Len-8 -> Ala 1.6 :t 0.4
Phe-9 -> Ala 1.3 :t 0.4

 

Relative activity is the CAT activity in transfected cell extracts relative to that obtained when
using wild-type GAL4-VP16(1-25). Means and standard deviations were calculated from at least
five independent transfections.

DISCUSSION

124

Using a series of GAL4-ORF 10 fusion proteins, we mapped the ORF 10
transcriptional-activation domain to the amino terminus of the protein. Using
HCA, we detected a motif surrounding Phe-28 near the amino terminus of
ORF10 that strongly resembles the motif centered at Phe-442 near the

carboxyl terminus of VP16. These two motifs are composed of characteristic
horseshoe-like clusters of hydrophobic residues surrounded by numerous acidic
residues. Amino acid substitutions for Phe-28 in GAL4-ORB 10 fusion proteins
confirmed the importance of aromatic or bulky hydrophobic residues at this
position, as shown previously for Phe-442 of VP16 (5), although VP16 shows a
more marked preference for aromatic amino acids at this position. The critical
role of Phe-28 was verified by the complete absence of activating ability by the
intact ORF 10 protein when Ala was substituted for Phe-28. Certain hydrophobic
residues flanking Phe-28 of ORF 10 also contributed to the transactivating
activity, similar to the results seen with mutagenesis of residues flanking Phe-442
of VP16 (5). These data conclusively support the HCA-based prediction that the
ORF 10 activation domain shares critical structural features with the VP16
carboxyl-terminal activation domain.

Other members of the alphaherpesvirus subfamily, BHV-l and EHV-l,
have VP16 homologs that transactivate IE promoters of their respective viruses
(31, 32). Using HCA, we located activation domains similar to VP16 and ORF10
near the amino termini of BHV-l UL48, MDV UL48, EHV-l gene 12, and EHV-4
gene B5 proteins. Each of these domains is characterized by a horseshoe-like
hydrophobic cluster interspersed with numerous acidic residues. Although the
activation domains of these proteins have not yet been thoroughly characterized,
our results predict that these amino-terminal domains are important; moreover,
key roles for specific amino acids can be predicted.

We identified a domain at the amino terminus of VP16 that shares some
features of the VP16 carboxyl-terminal and ORF 10 amino-terminal domains.
This VP16 amino-terminal domain proved to be a weak transcriptional activator
when fused to a heterologous DNA-binding domain; however, it is not
noticeably active in the context of the native VP16 molecule. A VP16 carboxyl-
terminal truncation mutant that retains the amino-terminal domain does not
have transactivating activity (10, 11). We hypothesize that the amino-terminal
domains in VP16 and its homologs evolved from common ancestral domains.
Unlike its homologs, however, the VP16 amino-terminal domain lost or failed to
develop transcriptional-activation function during evolution because VP16
acquired a potent carboxyl-terminal activation domain.

VP16 has a second domain at its extreme carboxyl terminus that is
important for transcriptional activation (2, 5, 6) and that consists of two clusters
of hydrophobic amino acids with adjacent acidic residues (Fig. 3A and B).
Similar domains that participate in transcriptional activation are located at the
extreme carboxyl termini of BHV-l UL48 and EHV-l gene 12 proteins (31, 32).
The VZV ORF 10 and MDV UL48 proteins apparently lack this domain.

HCA has previously been used to compare the sequences of numerous
enzyme families, as well as DNA-binding domains (33) and dimerization
domains (34) of eukaryotic transcription factors. The results reported here

125

demonstrate the value of HCA for predicting and comparing transcriptional-
activation domains among groups of related proteins.

We thank B. Henrissat for discussions regarding HCA, D. Last for providing anti-
GAL4 antibody, L. P. Perera for plasmids pCMV and p62CAT, and 1. Lillie for
plasmids pElb-CAT, p1GAL4-Blb-CAT, and p5GAL4-E1b-CAT. This work was
supported in part by grants from the National Institutes of Health (ROS-A127323,
K04-A101284) and the American Cancer Society (1FRA-328) to S.1.T.

 

10.
11.
12.
13.
14.
15.
16.
17.
18.
19.

20.

Gill, G., Pascal, E., Tseng, Z. H. 8: Tjian, R. (1994) Proc, Natl. Acad.
Sci. USA 91, 192 196.

Goodrich, 1. A, Hoey, T., Thut,1., Admon, A. 8: TJian, R. (1993) Cell
75, 519 530.

Cress, W. D. 8: Triezenberg, S. 1. (1991) Science 251, 87-90.

Leuther, K K, Salmeron, 1. M. 8: Iohnston, S. A. (1993) Cell 72, 575-
585.

Regier, 1. L., Shen, F. 8: Triezenberg, 5.1. (1993) Proc. Natl. Acad. Sci.
USA 90, 883-887.

Walker, 5., Greaves, R. 8: O'Hare, P. (1993) Mol. Cell. Biol. 13, 5233-
5244.

Batterson, W. 8: Roizman, B. (1983) 1. Virol. 46, 371-377.
Kinchington, P. R., Hougland, 1. K., Arvin, A. M., Ruyechan, W. T.
8: Hay, 1. (1992) 1. Virol. 66, 359-366.

Moriuchi, H., Moriuchi, M., Straus, S. E. 8: Cohen, 1. I. (1993) 1.
Virol. 64, 2739-2746.

Triezenberg, S. 1., Kingsbury, R. C. 8: McKnight, S. L (1988) Genes
Dev. 2, 718-729.

Cousens, D. 1., Greaves, R., Goding, C. R. 8: O'Hare, P. (1989)
EMBO 1. 8, 2337-2342.

Lemesle-Varloot, L., Henrissat, B., Gaborlaud, C., Bissery, V.,
Morgat, A. 8: Mornon,1.-P. (1990) Biochimie 72, 555-574.

Woodcock, S., Mornon,1.-P. 8: Henrissat, B. (1992) Protein Eng. 5,
629-635.

Keegan, L., Gill, G. 8: Ptashne, M. (1986) Science 231, 699-704,
Cohen, 1. 1., Heffel, D. 8: Seidel, K. (1993)] Virol. 67, 4246-4251.
Ho, S. N., Hunt, H. D., Horton, R. M., Pullen, 1. K. 8: Pease, R. L.
(1989) Gene 77, 51-59.

Moriuchi, M., Moriuchi, H, Straus, S, E. 8: Cohen, 1. I. (1994)
Virology 200, 297-300.

Sadowski, I., Ma, 1., Triezenberg, S. 1. 8: Ptashne, M. (1988) Nature
(London) 335, 563-564.

Perera, L. P., Mosca, 1. D., Sadeghi-Zadeb, M., Ruyechan, W. T. 8:
Hay, 1. (1992) Virology 191, 346-354,

Lillie, 1. W. 8: Green, M. R. (1989) Nature (London) 335, 39-44.

21.

24.

26.
27.
28.

29.

31.

126

Meier, 1. L. 8: Straus, S. E. (1993) 1. Virol. 67, 7573-7581.

Davison, A. 1. 8: Scott, 1. E. (1986) 1. Gen. Virol. 67,1759-1816.
Moriuchi, H., Moriuchi, M., Straus, S. E. 8: Cohen, 1. I., (1993) 1.
Virol. 67, 4290-4295.

Cohen, 1. 1. 8: Kieff, E. (1991) 1. Virol. 65, 5880-5885.

Dalrymple, M. A., McGeoch, D. 1., Davison, A. 1. 8: Preston, C. M.
(1985) Nucleic Acids Res. 13, 7865-7979.

Cress, W. D. 8: Triezenberg, S. 1. (1991) Gene 103, 235-238.

Telford, E. A. R, Watson, M. S., McBride, K, 8: Davison, A. 1. (1992)
Virology 189, 304-316.

Purewal, A. 5., Allsopp, R., Riggio, M., Telford, E. A. R., Azam, S.,
Davison, A. 1. 8: Edington, N. (1994) Virology 198, 385-389.
Carpenter, D. E. 8: Misra, V. (1991) Gene 119, 259-263.

Yanagida, N ., Yoshida, S., Nazerian, K 8: Lee, L. F. (1993) 1. Gen.
Virol. 74,1837-1845.

Misra, V., Bratanich, A. C., Carpenter, D. 8: O'Hare, P. (1994)

1. Virol. 68, 4898-4909.

Elliot, G. D. (1994) 1. Virol. 68, 4890-4897.

Laget, M., Callebaut, I., delaunoit, Y., Stehelin, D., 8: Mornon,1.-P.
(1993) Nucleic Acids Res, 21, 5987-5996.

Mornon,1.-P., Thoreau, E., Rowlands, D., Callebaut, I. 8: Moreau, G.
(1994) C. R. Acad. Sci. Ser. III 317, 597-606.

APPENDIX B

127
APPENDIX B

LIST OF PROTEINS FROM SEQUERY SEARCH

Protein Hatching Protein Matching

4 h ' n 'd
4 l -2 1
14-17

4- 7

-58
1 -1
2 -2 2
142-14
-72
1 -l71
27- 0
1 -l
-2

42 -42
-1
41-44
2 -211
441-444

60-63

47—
41-44
174-177

7-
-1 4

 

APPENDIX C

128
APPENDIX C

LIST OF PUBLICATIONS

1. Koelle, D. M., Corey, L., Burke, R. L., Eisenberg, R. 1., Cohen G. H,
Pichyangkura, R., Triezenberg, S. 1. (1994). Antigenic specificities of human
CD4+ T-cell clones recovered from recurrent genital herpes simplex virus
type 2 lesions. Journal of Virology, Volume 68, pages 2803-2810

2. Moriuchi, H., Moriuchi, M., Pichyangkura, R., Triezenberg, S. 1., Straus, S. E.,
8: Cohen, 1. I. (1995). Hydrophobic cluster analysis predicts an amino-
terminal domain of varicella-zoster ORF 10 required for transcriptional
activation. Proceedings of the National Academy of Science USA, Volume 92,
pages 9333-9337

LIST OF REFERENCES

LIST OF REFERENCES

Ace, C. I., Dalrymple, M. A., Ramsay, F. H., Preston, V. G., 8: Preston, C. M.
(1988). Mutational analysis of the herpes simplex virus type 1 trans-inducing
factor VP16. W62. 2595-2605.

Ace, C. I., McKee, T. A., Ryan, 1. M., Cameron, 1. M., 8: Preston, C. M. (1989).
Construction and characterization of a herpes simplex virus type 1 mutant
unable to transinduce immediate-early gene expression. |. yn Q1. 63, 2260-2269.

Axelrod, 1. D., Reagan, M. S., 8: Majors, 1. (1993). GAL4 disrupts a repressing
nucleosome during activation of CAM transcription in vivo. Gem-1m Z, 857-
869.

Berger, S. L., Pina, B., Silverman, N., Marcus, G. A., Agapite,1., Regier, 1. L.,
Triezenberg, S. 1., 8: Guarente, L. (1992). Genetic isolation of ADA2 - a potential
transcriptional adaptor required for function of certain acidic activation domains.

Cell ZQ, 251-265.

Boyes,1., 8: Bird, A. (1991). DNA methylation inhibits transcription indirectly via
a methyl-CpG binding protein. £911 fl, 1123-1134.

Bzik, D. 1., 8: Preston, C. M. (1986). Analysis of DNA sequences which regulate
the transcription of herpes simplex virus immediate early gene 3: DNA
sequences required for enhancer-like activity and response to trans-activation

by a virion polypeptide. WM 14, 929-943.

Cadena, D. L., 8: Dahmus, M. E. (1987). Messenger RNA synthesis in mammalian
cells is catalyzed by the phosphorylated form of RNA polymerase II. L_BiQL
Chem 262, 12468-12474.

129

130

Campbell, E. N. ,Palfreyman,1. W., 8: Preston, C. M. (1984). Identification of
herpes simplex virus DNA sequences which encode a trans-acting polypeptide
responsible for stimulation of immediate early transcription. M180,

19.

Carey, M., Leatherwood, 1., 8: Ptashne, M. (1990). A potent GAL4 derivative
activates transcription at a distance in vitro. Sgigmg 24Z, 710-711.

Carpenter, D. E., 8: Misra, V. (1991). The most abundant protein in bovine herpes
1 virions is a homologue of herpes simplex virus type 1 UL47. M 22,
3077-3084.

Carpenter, D. E., 8: Misra, V. (1992). Sequences of the bovine herpesvirus 1
homologue of herpes simplex virus type-1 alpha-trans-inducing factor (UL48).
Cane 112, 259-263-

Cedar, H. (1988). DNA methylation and gene activity. Cell 53, 3-4.

Chao, D. M., Gadbois, E. L., Murray, P. 1., Anderson, S. P., Sonu, M. S., Parvin, 1.
D., 8: A., Y. R. (1996). A mammalian SRB protein associated with an RNA

polymerase 11 holoenzyme. mm 380, 82-85.

Chen, 1., Attardi, L. D., Verrijzer, C. P., Yokomori, K., 8: Tjian, R. (1994).
Assembly of recombinant TFIID reveals differential coactivator requirements for
distinct transcriptional activators. gen Z9, 93-105.

Chou, P. Y., 8: Fasman, G. D. (1978). Prediction of the secondary structure of
proteins from their amino acid sequence. W AZ, 45-148.

Cleary, M. A., 8: Herr, W. (1995). Mechanisms for flexibility in DNA sequence
recognition and VP16-induced complex formation by the Oct-1 POU domain.

W 15, 2090-2100.

Collawn, 1. F., Kuhn, M. A., Liu, L.-F. S., Tainer, 1. A., 8: Trowbridge, I. S. (1991).
Transplanted LDL and mannose-6-phosphate receptor internalization signal
promote high -efficiency endocytosis of the transferrin receptor. BM], 10,
3247-3253.

131
Collawn, 1. F., Stangel, M., Kuhn, L. A, Esekogwu, V., 1ing, S., Trowbridge, I. S.,
8: Tainer, 1. A. (1990). Transferrin receptor internalization sequence YXRF
implicates a tight turn as a structural recognition motif for endocytosis. C911 63,
1061-1072.

Cordingley, M. G. ,,Campbell M. E. M, 8: Preston, C. M. (1983). Functional
analysis of a herpes simplex virus type I promoter: Identification of far-upstream

regulatory sequences. W 11, 2347-2365.

Cousens, D. 1., Greaves, R., Goding, C. R, 8: O'Hare, P. (1989). The C-terminal 79
amino acids of the herpes simplex virus regulatory protein, Vmw65, efficiently
activate transcription in yeast and mammalian cells in chimeric DNA-binding
proteins. W 8, 2337-2342.

Cress, A., 8: Triezenberg, S. 1. (1991a). Nucleotide and deduced amino acid
sequences of the gene encoding virion protein 16 from herpes simplex virus type

2. Gene 193,, 235-238.

Cress, W. D., 8: Triezenberg, S. 1. (1991b). Critical structural elements of the VP16
transcriptional activation domain. 59191192 231, 87-90.

Croston, G. E., Kerrigan, L. A., Lira, L M., Marshak, D. R., 8: Kadonaga, 1. T.
(1991). Sequence-specific antirepression of histone Hl-mediated inhibition of
basal RNA polymerase II transcription. m 231, 643-649.

Davison, A. 1. (1991). Varicella-zoster virus. 12mm 22, 475-486.

Doerfler, W. (1983). DNA methylation and gene activity. mm 52,
93-124.

Doerfler, W., Toth, M., Kochanek, S., Achten, S., Freisemrabien, U., Behnkrappa,
A., 8: Orend, G. (1990). Eukaryotic DNA methylation - facts and problems. EEBS
Len. 263, 329-333.

Donaldson, L., 8: Capone, 1. P. (1992). Purification and characterization of the
carboxyl-terminal transactivation domain of Vmw65 from herpes simplex virus

type 1. W 262, 1411-1414.

132

Drysdale, C. M., Duenas, E. Jackson, 8. M., Reusser, U. Braus, G. H, 8:
Hinnebusch, A. G. (1995). The transcriptional activator GCN4 contains multiple
activation domains that are dependent on hydrophobic amino acids. MoLCelL
m 13, 1220-1233.

Dynan, W. S. (1989). Understanding the molecular mechanism by which
methylation influences gene expression. W 3, 35-36.

Elliott, G., Mouzakitis, G., 8: O'hare, P. (1995). VP16 interacts via its activation
domain with VP22, a tegument protein of herpes simplex virus, and is relocated
to a novel macromolecular assembly in coexpressing cells. Lin-91. 62, 7932-7941.

Elliott, G. D. (1994). The extreme carboxyl terminus of the equine herpesvirus 1
homolog of herpes simplex virus VP16 is essential for immediate-early gene
activation. 1.111191: 63, 4890-4897.

Emami, K. H, 8: Carey, M. (1992). A synergistic increase in potency of a
multimerized VP16 transcriptional activation domain. EMBQL 11, 5005-5012.

Emini, E. A., Hughes, 1. V., Perlow, D. 5., 8: Boger, 1. (1985). Induction of hepatitis
A virus-neutralizing antibody by a virus-specific synthetic peptide. 1.111291. 33(3),
836-839.

Fang, S. M., 8: Burton, 2. F. (1995). RNA polymerase II-assosiated protein (RAP)
74 binds transcription factor (TF) IIB and blocks TFIIB-RAP30 binding. LBiQL
Chem. 221, 11703-11709.

Feng, D.-F., 8: Doolittle, R. F. (1987). Progressive sequence alignment as a
prerequisite to correct phylogenitic tree. LW 23, 351-360.

Friedman, A. D., Triezenberg, S. 1., 8: McKnight, S. L. (1988). Expression of a
truncated viral trans-activator selectively impedes lytic infection by its cognate
virus. Nam 333, 452-454.

Garnier,1., Osguthorpe, D. 1., 8: Robson, B. (1978). Analysis of the accuracy and
implications of simple methods for predicting the secondary structure of
globular proteins. M 129, 97-120.

133
Gibson, W., 8: Roizman, B. (1974). Proteins specified by herpes simplex virus. X.

Staining and radiolabelling properties of B capsid and virion proteins in
polyacrylamide gels. M13, 155-165.

Gill, G., Pascal, E., Tseng, Z. H, 8: 'ljian, R. (1994). A glutamine-rich hydrophobic
patch in transcription factor Sp1 contacts the dTAFII110 component of the
Drosophila TFIID complex and mediates transcriptional activation. mm

M 21. 192-196-

Goodrich, 1. A., Hoey, T., Thut, C. 1., Admon, A., 8: Tjian, R. (1993). Drosophila
TAF(II)40 interacts with both a VP16 activation domain and the basal
transcription factor TFIIB. £611 23, 519-530.

Graessmann, A., Sandberg, G., Guhl, E., 8: Graessmann, M. (1994). Methylation
of single sites within the herpes simplex virus tk coding region and the simian
virus 40 T-antigen intron causes gene inactivation. W 14, 2004-2010.

Graham, F. L., 8: Eb., A. 1. V. D. (1973). A new technique for the assay of
infectivity of human adenovirus 5 DNA. W _32, 456-467.

Greaves, R., 8: O'Hare, P. (1989). Separation of requirements for protein-DNA
complex assembly from those for functional activity in the herpes simplex virus
regulatory protein Vmw65. 1.11391, 63, 1641-1650.

Guentte, D. K, Ritzenthaler, 1. D., Foley, 1.,1ackson,1. D., 8: Smith, B. D. (1992).
DNA methylation inhibits transcription of procollagen alpha 2(I) promoters.

mm 283. 699-703.

Ha, L, Roberts, 5., Maldonado, E., Sun, X. 0., Kim, L. U., Green, M., 8: Reinberg,
D. (1993). Multiple functional domains of human transcription factor-IIB -
distinct interactions with 2 general transcription factors and RNA polymerase-II.

Gem 2. 1021-1082.

Harrington, M. A., Jones, P. A., Imagawa, M., 8: Karin, M. (1988). Cytosine
methylation does not affect binding of transcription factor Spl. W
3d, 33, 2066-2070.

134
Heine, J. W., Honess, R. W., Cassain, E. N., 8: Roizman, B. (1974). Proteins

specified by herpes simplex virus. XII. The virion polypeptides of type 1 strains.
1, ya 61.14, 640-651.

Hengartner, C. 1., Thompson, C. M., Zhang, 1., Chao, D. M., Liao, S. M., Koleske,
A. 1., Okamura, S., 8: Young, R. A. (1995). Association of an activator with an
RNA polymerase 11 holoenzyme. Gem 2, 897-910.

Hoey, T., Weinzierl, R. O. 1., Gill, G., Chen, J. L., Dynlacht, B. D., 8: Tjian, R.
(1993). Molecular cloning and functional analysis of drosophila TAF110 reveal
properties expected of coactivators. Cell 22, 247-260.

Homa, F. L., Krikos, A, Glorioso,1. C., 8: Levine, M. (1991). Functional analysis
of regulatory regions controlling strict late HSV gene expression. In E. K. Wagner

(Eds ). WW (pp 207-222) CRC Press. Inc

Horiuchi,1., Silverman, N., Marcus, G. A., 8: Guarente, L. (1995). ADA3, a
putative transcriptional adaptor, consists of two separable domains and interacts
with ADA2 and GCN5 in a trimeric complex. W 13, 1203-1209.

Ingles, C. 1., Shales, M., Cress, W. D., Triezenberg, S. 1., 8: Greenblatt, J. (1991).
Reduced binding of TFIID to transcriptionally compromised mutants of VP16.
Nam: 35L 588-590-

Johnson, C. A., Goddard, J. P., 8: Adams, R L. (1995). The effect of histone H1
and DNA methylation on transcription. W 3113, 791-798.

Johnson, P. F. 8: McKnight, S. L (1989). Eukaryotic transcriptional regulatory
proteins W 5.8, 799-839

Johnson, P. P., Sterneck, E. 8: Williams, S. C. (1993). Activation domains of
transcriptional regulatory proteins. W 4, 386-398.

Katan, M. ,,Haigh A., Verrijzer, C. P, Vandervliet, P. C, 8: O'Hare, P. (1990).
Characterization of a cellular factor which interacts functionally with Oct-1 1n the
assembly of a multicomponent transcription complex. W 18,
6871-6880.

135

Klemm, R., Goodrich, 1., Zhou, S., 8: Tjian, R. (1995). Molecular cloning and
expression of the 32-kDa subunit of human TFIID reveals interactions with VP16
and TFIIB that mediate transcriptional activation. W 22,
5788-5792.

Koleske, A. 1., 8: Young, R. A (1994). An RNA polymerase II holoenzyme
responsive to activators. Nam 363, 466-469.

Koleske, A. 1., 8: Young, R A (1995). The RNA polymerase 11 holoenzyme and its
implications of gene regulation. MW 22, 113-116.

Koptidesova, D., Kopacek,1., Zelnik, V., Ross, N. L. 1., Pastorekova, S., 8:
Pastorek, 1. (1995). Identification and characterization of a cDNA clone derived
from the Marek's disease tumour cell line RPLl encoding a homologue of alpha-
transinducing factor (VP16) of HSV-1. mm 142, 355-362.

Kraus, V. B., Inostroza, J. A., Yeung, K, Reinberg, D., 8: Nevins, J. R (1994).
Interaction of the Dr] inhibitory factor with the TATA binding protein is
disrupted by adenovirus E1A. W 21, 6279-6282.

Kristie, T. M., LeBowitz, 1. H, 8: Sharp, P. A. (1989). The octamer-binding
proteins form multi-protein - DNA complexes with the HSV alpha TIF
regulatory protein. EMELL 8, 4229-4238.

Kristie, T. M, 8: Roizman, B. (1984). Separation of sequences defining basal
expression from those conferring alpha gene recognition within the regulatory

domains of herpes simplex virus 1 alpha genes. MW 31,
4065-4069.

Kristie, T. M., 8: Sharp, P. A. (1990). Interactions of the Oct-1 POU subdomains
with specific DNA sequences and with the HSV alpha-trans-activator protein.

Genesllflel. 4. 2383-2396.

Laybourn, P. 1., 8: Kadonaga, 1. T. (1991). Role of nucleosomal cores and histone
H1 in regulation of transcription by RNA polymerase II. Science 234, 238-245.

136

Lemaster, 5., 8: Roizman, B. (1980). Herpes simplex virus phosphoproteins. II.
Characterization of the virion protein kinase and of the polypeptides
phosphorylated in the virion. mm 33, 798-811.

Lemesle-Varloot, L., Henrissat, B., Gaboriaud, C., Bissery, V., Morgat, A., 8:
Mornon, J.-P. (1990). Hydrophobic cluster analysis: procedures to derive
structural and functional information from 2-D-representation of protein

sequences. mm R 555-574.

Lewis, 1., 8: Bird, A. (1991). DNA methylation and chromatin structure. EEBS
L211. 233, 155-159.

Lieberman, P. (1994). Identification of functional targets of the Zta transcriptional
activator by formation of stable preinition complex intermediates. MQLCEILBIQL
A, 8365-8375.

Lieberman, P. M., 8: Berk, A. 1. (1994). A mechanism for TAFs in transcriptional
activation: activation domain enhancement of TFIID-TFIIA-promoter DNA
complex formation. Gem 3, 995-1006.

Lillie, J. W., 8: Green, M. R. (1989). Transcriptional activation by the adenovirus
E1A protein. Nam 333, 39-44.

Lin, 1., Chen, 1., Elenbaas, B., 8: Levine, A. J. (1994). Several hydrophobic amino
acids in the p53 amino-terminal domain are required for transcriptional
activation, binding to mdm-2 and the adenovirus 5 E1B 55-kD protein. Genes
DEL 3, 1235-1246.

Lin, 1., Teresky, A., 8: Levine, A (1995). Two critical hydrophobic amino acids in
the N-terminal domain of the p53 protein are required for the gain of function
phenotypes of human p53 mutants. Qnmggng 10, 2387-2390.

Lin, Y. 8., Ha, 1., Maldonado, E., Reinberg, D., 8: Green, M. R. (1991). Binding of
general transcription factor TFIIB to an acidic activating region. Nature 333, 569-
571.

137

Ma, 1., 8: Ptashne, M. (1987). A new class of yeast transcriptional activators. £911
31, 113-119.

Mackem, S., 8: Roizman, B. (1982a). Differentiation between alpha promoter and
regulator regions of herpes simplex virus type 1: The functional domains and

sequence of a movable alpha regulator. MW 22, 4917-4921.

Mackem, S., 8: Roizman, B. (1982b). Structural features of the herpes simplex
virus alpha gene 4, 0, and 27 promoter-regulatory sequences which confer alpha
regulation on chimeric thymidine kinase genes. um 44, 939-949.

Marcus, G. A., Silverman, N ., Berger, S. L, Horiuchi,1., 8: Guarente, L (1994).
Functional similarity and physical association between GCN5 and ADA2:
putative transcriptional adaptors. EMBQL 13, 4807-4815.

McGeoch, D. 1., Dalrymple, M. A., 8: Davison, A. J. (1988). The complete DNA
sequence of the long unique region in the genome of herpes simplex virus type 1.

L_Gen._¥iml. 62. 1531-1547.

McGeoch, D. 1., Dolan, A., Donald, S., 8: Brauer, D. H. K. (1986). The complete
sequence of the short repeat region in the genome of herpes simplex virus type 1.

W14. 1727-1745

McKee, T. A. Disney, G. H, Everett, R. D., 8: Preston, C. M. (1990). Control of
expression of the varicella-zoster virus major immediate early gene. mm
21, 897-906.

Merino, A., Madden, K R., Lane, W. S., Champoux,1.1., 8: Reinberg, D. (1993).
DNA topoisomerase I is involved in both repression and activation of
transcription. Name x3, 227-234.

Misra, V., Bratanich, A.C., Carpenter, D., O'Hare, P. (1994). Protein and DNA
elements involved in transactivation of the promoter of the bovine herpesvirus
(BHV) 1 IE-1 transcription unit by the BHV a gene trans-inducing factor. 1.11191.
63, 4898-4909.

138
Moriuchi, H, Moriuchi, M., Pichyangkura, R., Triezenberg, S. 1., Straus, S. E., 8:
Cohen, J. I. (1995). Hydrophobic cluster analysis predicts an amino-terminal
domain of varicella-zoster ORF 10 required for transcriptional activation. 13m

W 22. 9333-9337.

Moriuchi, H, Moriuchi, M., Straus, S. E., 8: Cohen, 1. I. (1993). Varicella-zoster
virus open reading frame 10 protein, the herpes simplex virus VP16 homolog,
transactivates herpesvirus immediate-early gene promoters. I, 21m]. 62, 2739-
2746.

O'Hare, P. 8: Williams, G. (1992). Structural studies of the acidic transactivation
domain of the Vmw65 protein of herpes simplex virus using 1H NMR.

flinchemistrx 31. 4150-4156

Ossipow, V., Tassan, 1. P., Nigg, E. A., 8: Schibler, U. (1995). A mammalian RNA
polymerase II holoenzyme containing all components required for promoter-
specific transcription initiation. Cell 33, 137-146.

Paranjape, S. M., Kamakaka, R. T., 8: Kadonaga, J. T. (1994). Role of chromatin
structure in the regulation of transcription by RNA polymerase II. M
m 63, 265-297.

Paranjape, S. M., Krumm, A., 8: Kadonaga, J. T. (1995). HMG 17 is a chromatin-
specific transcriptional coactivator that increases the efficency of transcription
initiation. M 2, 1987-1991.

Pazin, M. J., Kamakaka, R. T., 8: Kadonaga, 1. T. (1995). ATP-dependent
nuclepsome reconfiguration and transcriptional activation from preassembled
chromatin templates. $394192 266, 2007-2011.

Peek, R., Niessen,R. W. L. M. ,Schoe.nmakers,1 G. G., 8:Lubsen,N. H (1991).
DNA methylation as a regulatory mechanism 1n rat gamma-crystallin gene
expression NudeicAcidsRes. 12. 77-83

Poon, A. P. W., 8: Roizman, B. (1995). The phenotype in vitro and in infected cells
of herpes simplex virus 1 alpha-trans-inducing factor (VP16) carrying
temperature sensitive mutations introduced by substitutions of cysteines. 1.3111121.
62, 7658-7667.

139
Post, L. E., Mackem, 5., 8: Roizman, B. (1981). Regulation of alpha genes of herpes
simplex virus: Expression of chimeric genes produced by fusion of thymidine
kinase with alpha gene promoters. C211 24, 555-565.

Preston, C. M., Frame, M. C., 8: Campbell, M. E. M. (1988). A complex formed
between cell components and an HSV structural polypeptide binds to a viral
immediate early gene regulatory DNA sequence. Cell 32, 425-434.

Purewal, A. 5., Allsop, R., Riggio, M., Telford, E. A R., Azam, 5., Davison, A. J., 8:
Edington, N. (1994). Equid herpesvirus 1 and 4 encode functional homologs of
the herpes simplex virus type 1 virion transactivator protein, VP16. 351-91981 123,
385-389.

Purewal, A. 5., Smallwood, A. V., Kaushal, A., Adegboye, D., 8: Edington, N.
(1992). Identification and control of the cis-acting elements of the immediate-
early genes of equid herpesvirus type 1. W 23, 513-519.

Regier, J. L., Shen, F., 8: Triezenberg, S. 1. (1993). Pattern of aromatic and
hydrophobic amino acids critical for one of two subdomains of the VP16

transcriptional activator. W 20, 883-887.

Roberts, 5. G. E., Green, M. R. (1994). Activator-induced conformational change
in general transcription factor TFIIB. Nature 321, 717-720.

Roberts, 5. G. E., Ha, 1., Maldonado, E., Reinberg, D., 8: Green, M. R. (1993).
Interaction between an acidic activator and transcription factor IIB is required for
transcriptional activation. m 363, 741-744.

Roizman, B. (1990). Herpesviridae: A brief introduction. In W
Fields) (pp. 1787-1794).

Roizman, B., 8: Sears, A. E. (1990). Herpes simplex viruses and their replication.
In W (pp- 17951842).

Rost, B., 8: Sander, C. (1994). Combining evolutionary information and neural
networks to predict protein secondary structure. 2119131118 13, 55-72.

140

Roy, A. L., Malik, 5., Meisterernst, M., 8: Roeder, R. G. (1993). An alternative
pathway for transcription initiation involving TFII-I. Nature 363, 355-359.

Sadowski, 1., Ma, 1., Triezenberg, 5., 8: Ptashne, M. (1988). GAL4-VP16 is an
unusually potent transcriptional activator. Nature 333, 563-564.

5ambrook,1., Fritsch, E. F., 8: Maniatis, T. (1989). In mm
W (2nd ed.). Cold Spring Harbor, NY: Cold Spring Harbor
Laboratory Press.

Sanger, F. N icklen 5., 8: Coulson, S. A. (1977). DNA sequencing with chain-
terminating inhibitors. W 24, 5463-5467.

Shen, F., Triezenberg, S. 1., Hensley, P., Porter, D., 8: Knutson, J. R. (1996a).
Critical amino acids in the transcription activation domain of the herpesvirus
protein VP16 are solvent-exposed in highly mobile protein segments. LfiigL.

Chan. 221 . 4819-4826.

Shen, F., Triezenberg, 5. 1., Hensley, P., Porter, D., 8: Knutson, 1. R. (1996b).
Transcriptional activation domain of the herpesvirus protein VP16 becomes
conformationally constrained upon interaction with basal transcription factors. L

BioLChem. 2Z1 . 4827-2837

Smale, S. T., 8: Baltimore, D. (1989). The "initiator" as a transcription control
element. Cell 32, 103-113.

Smibert, C. A., Popova, B., Xiao, P., Capone, J. P., 8: Smiley, 1. R. (1994). Herpes
simplex virus VP16 forms a complex with the virion host shutoff protein vhs. L

Mlml. 63, 2339-2346.

Spear, P. G., 8: Roizman, B. (1972). Proteins specified by herpes simplex virus. V.
Purification and structural proteins of the herpesvirion. 1.111.011 2, 431.

Stern, 5., 8: Herr, W. (1991). The herpes simplex virus trans-activator VP16
recognizes the Oct-1 homeo domain: evidence for a homeo domain recognition
subdomain. W1. 3, 2555-2566.

141

Stringer, K. F., Ingles, C. 1., 8: Greenblatt, J. (1990). Direct and selective binding of
an acidic transcriptional activation domain to the TATA-box factor TFIID. Man“:
343, 783-786.

Tanaka, M., 8: Herr, W. (1994). Reconstitution of transcriptional activation
domains by reiteration of short peptide segments reveals the modular
organization of a glutamine-rich activation domain. W 14, 6056-6067.

Thanos, D., 8: Maniatis, T. (1992). The high mobility group protein HMG I(Y) is
required for NF-kappa B-dependent virus induction of the human IFN-beta gene.
C611 21, 777-7789.

Thut, C. 1., Chen, 1., Klemm, R., 8: Tjian, R. (1995). p53 transcriptional activation
mediated by coactivators TAFH40 and TAFH60. Science 262, 100-104.

Tiley, L. 5., Madore, 5. 1., Malim, M. H., 8: Cullen, B. R. (1992). The VP16
transcription activation domain is functional when targeted to a promoter-
proximal RNA sequence. 921162.221. 6, 2077-2087.

Triezenberg, 5. 1. (1995). Structure and function of transcriptional activation

domains W 5.. 190-196

Triezenberg, S. J., Kingsbury, R. C., 8: McKnight, S. L. (1988a). Functional
dissection of VP16, the trans-activator of herpes simplex virus immediate early
gene expression. mm 2, 718-729.

Triezenberg, S. 1., LaMarco, K L., 8: McKnight, S. L. (1988b). Evidence of
DNAzprotein interactions that mediate HSV-1 immediate early gene activation by

VP16. Gem 2, 730-743.

Usheva, A., 8: Shenk, T. (1994). TATA-binding protein-independent initiation:
YY1, TFIIB, and RNA polymerase 11 direct basal transcription on supercoiled
template DNA. (.611 26, 1115-1121.

142
Van Hoy, M., Leuther, K. K, Kodadek, T., 8: Johnston, S. A. (1993). The acidic

activation domains of the GCN4 and GAL4 proteins are not a helical but form [3
sheets. Cell 22, 587-594.

Verrijzer, C. P., Chen, J.-L., Yokomori, K, 8: 'Ijian, R. (1995). Binding of TAFs to
core elements directs promoter selectivity by RNA polymerase 11. Call 31, 11 15-
1125.

Verrijzer, C. P., Yokomori, K, Chen, J. L., 8: Tjian, R. (1994). Drosophila TAFII
150: similarity to yeast gene T‘SM-l and specific binding to core promoter DNA.
Science 2%. 933-941.

Walker, 5., Greaves, R., 8: O'Hare, P. (1993). Transcriptional activation by the
acidic domain of Vmw65 requires the integrity of the domain and involves
additional determinants distinct from those necessary for TFIIB binding. Mg],

W 13, 5233-5244.

Watt, F., 8: Molly, P. L. (1988). High mobility group proteins 1 and 2 stimulate
binding of a specific transcription factor to the adenovirus major late promoter.

W 16. 1471- 1486

Weinheimer, 5. P., Boyd, B. A., Durham, 5. K., Resnick, J. L., 8: O'Boyle, D. R.
(1992). Deletion of the VP16 open reading frame of herpes simplex virus type 1.
ij'ml. 68, 258-269.

Weis, L, 8: Reinberg, D. (1992). Transcription by RNA polymerase-II - initiator-
directed formation of transcription-competent complexes. EA3E3L 6, 3300-3309.

Werstuck, G., 8: Capone, J. P. (1989a). Mutational analysis of the herpes simplex
virus trans-inducing factor Vmw65. figng 23, 213-224.

Werstuck, G., 8: Capone, J. P. (1989b). Identification of a domain of the herpes
simplex virus trans-activator Vmw65 required for protein-DNA complex
formation through the use of protein A fusion proteins. 1.36.191, 63, 5509-5513.

Whitley, R 1. (1990). Herpes simplex viruses. In W (pp. 1843-
1888).

143

Wilson, C. 1., Chao, D. M., Imbalzano, A. N ., Schnitzler, G. R, Kingston, R. E., 8:
Young, R. A. (1996). RNA polymerase 11 holoenzyme contains SWI/SNF
regulators involved in chromatin remodeling. Call a, 235-244.

Wolffe, A. P. (1994). Transcription: in tune with the histones. Can 22, 13-16.

Wiilfl, 5., Schrfider, M., 8: Wittig, B. (1991). Lack of correlation between DNA
methylation and transcriptional inactivation: The chicken lysozyme gene. 12:55,
W 38, 271-275.

Woodcock, 5., Mornon,1.-P., 8: Henrissat, B. (1992). Detection of secondary
structure elements in proteins by hydrophobic cluster analysis. W 3,
629-635.

Xiao, H., Pearson, A., Coulombe, B., Truant, R., Zhang, 5., Regier, 1. L,
Triezenberg, S. 1., Reinberg, D., Flores, O., Ingles, C. 1., 8: Greenblatt, J. (1994).
Binding of basal transcription factor TFIIH to the acidic activation domains of
VP16 and p53. W14, 7013-7024.

Xiao, P., 8: Capone, 1. P. (1990). A cellular factor binds to the herpes simplex virus
type-1 transactivator Vmw65 and is required for Vmw65-dependent protein-
DNA complex assembly with Oct-1. W 10, 4974-4977.

Yanagida, N ., Yoshida, 5., Nazerian, K, 8: Lee, L. F. (1993). Nucleotide and
predicted amino acid sequences of Marek's disease virus homologues of herpes
simplex virus major tegument proteins. W 24, 1837-1845.

Yankulov, K, Blau,1., Purton, T., Roberts, 5., 8: Bentley, D. L. (1994).
Transcriptional elongation by RNA polymerase II is stimulated by
transactivators. Cell 22, 749-759.

Young, R. A. (1991). RNA polymerase-II. W 31, 689-715.

Zawel, L., 8: Reinberg, D. (1993). Initiation of transcription by RNA polymerase
II: a multi-step process WM 44. 67-108

144

Zawel, L., 8: Reinberg, D. (1995). Common themes 1n assembly of eukaryotic
transcription complexes. W M, 533-561.