Epstein-Barr virus glycoprotein gp85

MISSING IMAGE

Material Information

Title:
Epstein-Barr virus glycoprotein gp85 biochemical characteristics, purification and mapping to viral genome by Douglas Earl Oba
Physical Description:
vi, 115 leaves : ill. ; 29 cm.
Language:
English
Creator:
Oba, Douglas Earl
Publication Date:

Subjects

Subjects / Keywords:
Glycoproteins   ( mesh )
Herpesvirus 4, Human   ( mesh )
Genes, Viral   ( mesh )
Pathology thesis Ph.D   ( mesh )
Dissertations, Academic -- Pathology -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1988.
Bibliography:
Bibliography: leaves 98-114.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001019232
oclc - 20674845
notis - AFA0568
System ID:
AA00009086:00001


This item is only available as the following downloads:


Full Text









EPSTEIN-BARR VIRUS GLYCOPROTEIN gp85:
BIOCHEMICAL CHARACTERISTICS, PURIFICATION
AND MAPPING TO VIRAL GENOME



















By

DOUGLAS EARL OBA


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1988














ACKNOWLEDGMENTS


Appreciation is in order for all those who have helped in completing this work. My

committee members, Drs. Howard Johnson, Carlo Moscovici, Guy Palmer and Ben Dunn,

as well as other faculty members, Drs. Alfred Esser and John Dankert, have been

extremely helpful and encouraging. Of course, the principal committee member, general

overseer, patient teacher, and excellent mentor, Dr. Lindsey Hutt-Fletcher, has been

invaluable in seeing me through this work. The many hours in the lab have definitely not

been in solitude with Susan, Steve, Nancimae, Soman, Ruwaida, Ann, Barbara, Cindy

and everyone else who has inhabited the lab during my sojourn and kept me on my toes.

Dr. Bala set the standard (and the pace) in the first weeks which we all strive to match. My

work in the lab was enduringly supported by my wonderful family. Thanks are extended

to my parents, who have assisted in more ways than parents are required; to my wife,

Ellen, for putting us through it; and to my children, Lynn and Eric, who only see the best

in all that we go through.













TABLE OF CONTENTS



ACKNOWLEDGMENTS ....................................................................... ii

ABSTRACT ..................................................................................... v

CHAPTERS

I INTRODUCTION

Historical Perspective .................................................................... 1
Clinical Manifestations ................................................................... 1
Characteristics of EBV.................................................................. 4
Molecular Nature of the Virus ........................................................ 5
Biologic Significance of EBV Proteins............................................... 16

II BIOCHEMICAL CHARACTERIZATION OF gp85

Introduction.............................................................................. 19
Materials and Methods.................................................................. 20
Results.................................................................................... 26
Discussion ................................................................................ 37

III PURIFICATION OF gp85

Introduction............................................................................... 39
Materials and Methods.................................................................. 39
Results.................................................................................... 42
Discussion................................................................................ 51

IV MAPPING OF gp85 TO THE GENOME

Introduction.............................................................................. 55
Materials and Methods.................................................................. 56
Results.................................................................................... 58
Discussion................................................................................ 68

V ANTIBODIES TO gp85

Introduction.............................................................................. 70
Materials and Methods.................................................................. 71
Results.................................................................................... 76
Discussion................................................................................ 89









VI SUMMARY AND CONCLUSIONS

Recapitulation ............................................................................. 94
Importance of Present Studies and Future Directions .............................. 97

REFERENCES.............................................................................. 98

BIOGRAPHICAL SKETCH .............................................................. 115











































iv














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

EPSTEIN-BARR VIRUS GLYCOPROTEIN gp85:
BIOCHEMICAL CHARACTERISTICS, PURIFICATION
AND MAPPING TO VIRAL GENOME

By

Douglas Earl Oba

April 1988

Chairman: Lindsey M. Hutt-Fletcher
Major Department: Pathology and Laboratory Medicine


Epstein-Barr virus is a human herpesvirus that induces a variety of lymphopro-

liferative disorders and is implicated in the etiology of two cancers, nasopharyngeal carci-

noma and Burkitt's lymphoma. The virus contains at least three glycoproteins in its outer

envelope, gp350, gp250, and gp85. When this work began, gp85 had been partially

characterized using a monoclonal antibody F.2.1 and shown to contain epitopes that

induced neutralizing antibodies; little else was known about the molecule. This dissertation

describes additional characterization of gp85 and identification of the gene that encodes it.

A new monoclonal antibody, E1D1, was produced and shown to react with gp85. E1D1

was used for purifying gp85 and for preparation of monospecific rabbit antibodies. The

BXLF2 open reading frame of the EBV genome, which has homology with the glycopro-

tein gH of herpes simplex virus, was identified as a candidate gene for gp85. A seventeen

residue synthetic peptide derived from the BXLF2 sequence was used to raise rabbit anti-

serum. Anti-peptide antibody was purified from the serum by chromatography on Protein-

A agarose and synthetic peptide coupled to Affigel. It was then used to immunoprecipitate

a molecule with the same electrophoretic mobility as gp85. Purified anti-peptide antibodies

v









reacted with gp85 in Western blots. Comparison of protease digests of the protein

immunoprecipitated by antipeptide antibodies with that recognized by the monoclonal anti-

body F.2.1, indicated that the two were identical. The polyclonal monospecific rabbit anti-

serum raised against purified gp85 had a weak, but specific, reaction with the BXLF2

derived peptide in an ELISA. These data indicated that the BXLF2 open reading frame

encodes gp85 and thus suggest that gp85 is the EBV equivalent of glycoprotein gH. Two

additional synthetic peptides derived from the BXLF2 sequence were used for antibody

production in rabbits and additional monoclonal antibodies to gp85 were produced by a

variety of protocols including immunization with whole virus, coupled peptide, or peptide

priming followed by challenge with whole virus. These antibodies will be valuable for

determining the epitopes of gp85 crucial for neutralization of virus and for study of the

functional characteristics of this important surface glycoprotein.













CHAPTER I

INTRODUCTION


Historical Perspective

During the past two and a half decades much research has been devoted to under-

standing the nature of Epstein-Barr virus (EBV) and its association with human disease.

The virus was originally isolated in 1964 by Epstein, Achong, and Barr from an African

Burkitt's lymphoma biopsy that had been cultured in vitro {55). Since that time a con-

siderable amount of information has been amassed concerning the nature and infectivity of

the virus. Much of it relates to diseases associated with the virus and the response of the

human body to infection. More recently, an understanding has emerged regarding some of

the more basic biology of the virus. Work described here contributes to this effort by

expanding our knowledge of virion proteins that are crucial to virus infectivity. Although a

detailed chronicle cannot be given of previous EBV research, a brief historical summary of

key discoveries is important to provide an appropriate background to the new work

described herein.

Clinical Manifestations

Burkitt's Lvmphoma

It was the British surgeon, Dennis Burkitt, who first asked the scientific questions

which led to the discovery of the Epstein-Barr virus. As a surgeon in Africa during the

1950s he became interested in what was "not only the most common children's tumor in

tropical Africa, but was more common than all other children's tumors added together" (21

p. xxiv). He described a tumor with a distribution demarcated by geographic features

affecting climate (21). In search for a human cancer which could be caused by a virus,










M.A. Epstein took interest in the tumor described by Dr. Burkidtt. This interest led to the

initial discovery of the virus in 1964 {55).

Although the virus, later named Epstein-Barr virus (53), was isolated from cell cul-

ture derived from Burkitt's lymphoma (BL), a sole causal relationship was not shown to

exist (146, 150). A prospective epidemiologic study in Uganda (41) provided evidence

that supported the hypothesis that EBV plays at least the role of cofactor in development of

African BL (54). However, unlike the tumor, the virus was found to exist in all human

populations with a frequency of 90-100% {56, 163).

The presence of chromosomal translocations in all cell lines derived from BL (110)

is an additional factor in understanding BL. One unique case of African B cell lymphoma,

BJAB, has been described which does not have a chromosomal anomaly (230); however,

questions have arisen regarding the diagnosis of BL in this case. The cytogenetic anomaly

which is characteristic of BL is a translocation of chromosome 8 most often with

chromosome 14 but also with chromosomes 2 or 22 ( 14, 137). The c-myc oncogene has

been localized to the human chromosome 8, and translocations involve the juxtaposition of

this gene with that of the immunoglobulin heavy chain gene cluster on chromosome 14, the

kappa light-chain genes on chromosome 2, and the lambda light-chain gene on chomosome

22 {119). The exact nature of the activation or deregulation of the c-myc oncogene in

these tumor lines varies because of variable rearrangements but all occur in regions of the

genome which are actively transcribed. A role for this oncogene in the development of BL

seems fairly likely. However, since EBV is capable of transforming B lymphocytes

without involving a c-myc translocation, it is probable that the virus is not required for

this oncogenesis. The frequency with which malignant clones affected by an altered myc

regulation may arise have increased when EBV, in cooperation with immunosuppression

resulting from persistent malaria infections, results in unrestricted proliferation of B cells.

This induction of B lymphocyte proliferation by EBV is best seen in infectious mono-

nucleosis.









Infectious Mononucleosis

Epstein-Barr virus infects almost everyone by adulthood, establishes itself as a latent
infection, and induces permanent seroconversion. It infects children with appearance of

little or no disease whereas infection of adolescents results in clinically manifest infectious

mononucleosis (IM) in approximately 67% of cases (81, 85, 165). The clinical mani-

festations of an EBV-induced IM are caused primarily by a rapid polyclonal T and B cell

proliferation (180). Primary replication of the virus probably occurs in pharyngeal

epithelium (185) through which circulating B lymphocytes become infected and

transformed by EBV causing the rapid proliferation (173, 197). This number of infected
B cells is eventually reduced by an increasing number of T lymphocytes which proliferate

to check the abnormal numbers of infected B cells { 182). The symptoms characteristic of

IM are a result of the immune conflict within the body. A majority of individuals continue

to harbor the virus for life in an apparently latent form {54). This asymptomatic carrier

state was originally thought to be a result of a small population of B cells which remain
infected (165); however, it has been suggested that B cells may be infected by virus shed

from undifferentiated epithelial cells of the oropharnyx in which there is chronic, but low-

grade replication of EBV { 146, 226, 227).

Since an infection with EBV requires a healthy immune response to restrain the

abnormal proliferation of B lymphocytes, a variety of lymphoproliferative disorders can

arise in immunocompromised hosts. Lymphoproliferation can be life-threatening and may

be a result of either a primary infection or failure to control an existing infection. Immuno-

deficient patients who experience EBV induced lymphoproliferative disorders include those
with the X-linked lymphoproliferative syndrome (XLP), ataxia-telagiectasia, Wiskott-
Aldrich syndrome, severe combined immune deficiency after bone marrow transplantation

with T-depleted non-HLA-matched bone marrow, Chediak-Higashi syndrome and, more

recently, acquired immunodeficiency syndrome (AIDS) (29, 74, 92, 164).









Nasopharvngeal Carcinoma

Epstein-Barr virus has also been implicated in nasopharyngeal carcinoma (NPC).

The association of EBV with NPC was first suggested when high antibody titers to EBV

were found in all NPC patients examined {40, 42, 111). Although low-levels of anti-

bodies can be detected in everyone previously infected with EBV, unusually high levels of

IgA antibody specific for EBV viral capsid antigen (VCA), increased the likelihood that

EBV was involved with the carcinoma [84). Additional evidence for a role of EBV in the

induction of NPC was provided by a study in which all undifferentiated carcinomas of the

nasopharynx were shown to contain DNA from the EBV genome (3). Unlike BL which

is generally restricted geographically, NPC occurs throughout the world. It does, how-

ever, occur with a high frequency among certain populations of southern Asia, particularly

the Kwantung Province of southern China (42) which suggests evidence for other factors

acting with EBV in the development of the tumor ( 111, 193).

Characteristics of EBV

MorholoQg

The first description of Epstein-Barr virus isolated from a Burkitt's lymphoma cul-

ture was by electron microscopy and indicated that the virus has the morphology of a her-

pesvirus (55). It has since been given the provisional designation as human herpesvirus 4

and because of its host range specificity and characteristics of latent infection is classified in

the subfamily of gammaherpesvirinae ( 176). The innermost component of the virus par-

ticle is the core which consists of core proteins and deoxyribonucleic acid (DNA). The

core is surrounded by an icosahedral nucleocapsid composed of capsomeres arranged in

hexagonal and pentameric array. The nucleocapsid is surrounded by a lipid envelope with

spikes on its outer surface (107), and an amorphous tegument fills the cavity between the

nucleocapsid and the envelope. The viral envelope is thought to be acquired as the virus

capsid buds through the nuclear membrane of the cell. The lipid content of the envelope is









derived from host cell membranes in which cellular proteins have been replaced by those

encoded by virus { 192).
Tropism

Even though EBV was classified as a herpesvirus, it became apparent immediately
that the tropism of this virus differed from that of other known human herpesviruses in that

the virus was unable to infect established fibroblast cell lines {53). Until recently it was

generally accepted that EBV infected only B cells. Initial work implicating that the virus

replicated in oropharyngeal epithelium was discounted on the grounds that infiltrating

lymphocytes infected with virus could account for its presence in oropharyngeal secretions
[ 135, 172). Although it has recently been established that the virus replicates in vivo in

epithelial cells as well as in lymphocytes ( 118, 185, 225), the only cell currently available

for study of replication in vitro is the lymphocyte, and all virus thus far studied is

lymphocyte derived.

The ability of the EBV to infect lymphocytes is initiated by attachment of virus to the
cell receptor, CR2, which binds the C3dg fragment of complement as well as EBV ( 57,

138). Recently it has been shown that epithelial cells also express a receptor which facili-

tates attachment of EBV (185). Like the B cell CR2 which is lost during the differentiation

to a plasma cell (204), the epithelial receptor is present on cells of a less-differentiated

state. Infection of epithelial cells expressing the receptor early in differentiation, followed

by active EBV replication in fully differentiated desquamating cells,provides a favorable

condition for survival of the virus without detriment to the host.

Molecular Nature of the Virus
Genome

The genome of EBV is a double-stranded DNA molecule of over 172 kilobases
which contains non-random single-stranded breaks ( 162). It is found in the virus particle

as a linear molecule (76, 178) but exists as a circular episome inside the nucleus of the









infected cell (4, 186). Knowledge of the genome was derived from restriction endonu-

clease maps (64), cloning {34, 169) and eventually genomic sequencing (5).

The genome consists of five large regions of unique DNA domains, U1-U5, which
are separated by four regions of internal repeats, 11-14, and flanked on both ends with tan-

dem repeats (25, 36, 65). The genomic organization of EBV isolates from different
populations is similar (169) except in the U2 domain which contains considerable variation
( 108). The virus genome is replicated early during the S phase of infected cells {75) by a
cellular DNA polymerase and a DNA polymerase encoded by the viral genome. The virus

encoded DNA polymerase is sensitive to phosphonoacetic acid (PAA) and acyclovir
whereas the cellular polymerase is resistant to the two drugs { 191). Replication of the
genome has been separated into early and late phases using sensitivity to PAA as a stan-

dard.

Infection of B lymphocytes with EBV results in a latent infection without production
of new virus. The latently infected cells usually contain more than one copy of the com-

plete EBV genome, most of which is not integrated with cell DNA and is found covalently
closed in circular episomes {122). The episodes are formed by covalent binding of the
tandem repeats at each end of the DNA (34). Replication of this episomal DNA has been

proposed to occur in a manner similar to that of SV40 DNA (72). Some defective EBV
DNA has also been found in latently infected cells (78). A limited number of EBV genes

are expressed in latently infected, growth-transformed lymphocytes and are most likely
involved in growth transformation or maintaining latent infection.

Proteins

Although the total number of EBV proteins expressed from the EBV genome is
unknown, analysis of the sequence data indicates that there are 84 major open reading

frames. It is suggested that most of these reading frames are expressed as proteins because
of size, positioning with promoters and polyadenylation/RNA processing sites (5). Most

of the earlier work on viral proteins was described in terms of serologically defined anti-









gens which were grouped into three categories: the transformation antigens, including the
nuclear antigen complex (EBNAs) and lymphocyte-defined membrane antigen (LYDMA);

the early antigens (EA), made prior to viral DNA synthesis; and the late antigens which

include the structural virion membrane antigens (MA) and viral capsid antigens (VCA)

{ 142, 152). Information about the virus proteins has increased dramatically in the past

few years resulting in a number of newly described proteins, some of which are classified
in Table 1 as latent proteins, early antigens, internal proteins and membrane proteins. The

brief listing of these proteins which has been compiled from the literature indicates the

names of these various proteins, the genes which encode them, and the location within the
infected cells where the protein is detected. A number of these proteins, particularly the

latent proteins, have been described by various groups using different nomenclature for
identification; therefore, gene mapping has been essential for correlating them.

Latent Proteins

Cells which contain latent virus produce a small number of EBV specific proteins.
Four major mRNAs have been identified in transformed cell lines, three of which encode
EBV nuclear antigens (EBNAs) and the other encodes a protein which is located in the

plasma membrane {49, 89).

Nuclear antigens. Analysis using human immune sera indicated EBNA to be a group
of highly variable molecular weight proteins located in the nucleus (83). Recent evidence

indicates that the EBNA family is comprised of at least five or six proteins. The single

grouping of EBNA was first expanded to two distinct EBNAs, EBNA-1 and EBNA-2
when the proteins were mapped to the genome. EBNA-1 is a protein of approximately 75-
78 kDa which binds to metaphase chromosomes {77, 171) and is responsible for episome
maintenance {228). The protein is encoded by the Barnm HI-K1 open reading frame of the

genome as was shown by antibodies raised against a fusion protein from the gene (59, 87)

and synthetic peptides (43) which react with the same protein as was originally defined
with human sera.










Table 1. Epstein-Barr virus proteins.

Name MW (x 103) Gene Comment


Latent
EBNA-1
EBNA-2
EBNA-3a
EBNA-4/3b
EBNA-5/LP
EBNA-6/3c
LMP



Zebra
?


DNAase
DNA polymerase
Thymidine kinase
M-EA (D)
MS-EA (D)
EA (R)
Capsid associated
VCA/gB
VCA
VCA
Membrane proteins
gp350/300
gp250/200
gp85
p140
pl105
gp69-85
gG


75-78
78-87
142
166
20-130
180
63



43
17


52
110
67
48-52
60
85


125/110
152
160


300-350
200-250
80-90
140
105
28, 69-89
80-90, 205


BamHI-K (BKRFl)
BamHI YH
BamHI-E (BERFI)
BamHI-E (BERF2a/2b)
BamHI-WY
BamHI-E (BERF3/4)
BamHI-NJ (BNLF1)



BamHI-Z (BZLFl)
BamHI-H (BHRFl)


Bamk-G (BGFLFS)
BamHI-A (BALF5)
BamHI-X (BXLF1)
BamHI-M (BMRF1)
BamHI-M/S (BMLF1)
BamHI-O (BORF2)


Barn HI-A (BALF4)


BamHI-C (BCLF1)


BamHI-L (BLLFl)
BamHI-L (BLLFl)
BamHI-X (BXLF2)
BamHI-N(BNRFl)


Required for maintenance of episome
Increases CD23 expression, absent in P3HRI



Absent in P3HR1


Transforms established rodent cells;
associates with vimentin


Switch to lytic cycle
Hydrophic, possibly latent; homolog of bcl-2



HSV homolog


DNA binding protein; polymerase activation
trans-activator, switch to lytic cycle
Large subunit of ribonucleotide reductase


Glycosylated


gC?
gC?
gH
homolog of HVS tegument protein
gB?


BILF2 or BDLF3?
BILF2 or BDLF3?









EBNA-2 is a heterogenous protein ranging from 78-87 kDa encoded by the BamHI
fragments Y and H, in the BHRF1 open reading frame (86,179). The mRNA encoding
EBNA-2 is actually bicistronic and encodes two separate proteins, ENBA-2 and EBNA

leader protein (EBNA-LP) {219). EBNA-LP is a protein of approximately 40 kDa and is

encoded by the leader portion of the EBNA-2 gene from the BamHI W (a region of
multiple repeat exons) and Y fragments. Another independently described nuclear antigen

complex called EBNA-5 has been located to the same Barnm HI W,Y regions of the genome.

EBNA-5 is encoded by highly spliced mRNA resulting in a complex, broad-range of

polypeptides ranging from 20 kDa to as large as 130 kDa depending on the number of W

repeat exons used for generating message {58, 219).

The region encoding the EBNA-2 and EBNA-5/LP is of particular interest because
virus strains which have deletions in these regions are unable to transform lymphocytes

{174). Transformation-competent recombinants between these strains and other EBV
isolates have restored the wild-type EBNA-5/LP and EBNA-2 domain {35, 187).
Importance of EBNA-2 in B cell growth was demonstrated by expression of the EBNA-2

gene using a retrovirus vector. Expression of this gene in EBV-negative BL cells resulted
in the induction of the B-cell surface protein Blast 2 or CD23. This B cell surface protein
may function as a receptor for a B cell growth factor {73) and the secreted form of CD23

has been identified with autocrine growth-promoting activity (198).

The third major mRNA identified in transformed cells actually appears to be mRNAs
from a family of three related tandem genes encoded by the Barnm HI-E region of the

genome {88,105,177). These messages encode the nuclear antigens, EBNA-3, EBNA-4,
and EBNA-6, also called EBNA 3a, 3b, and 3c. These three proteins of approximately
142 kDa, 166 kDa, and 180 kDa, respectively, are all located in the nuclear fraction of the

cell. Their role in the transformation process is currently being investigated.

Latent membrane protein. At least one other protein is made during latent infection,
the latent membrane protein (LMP). Original evidence for a protein which was present on







10

the plasma membrane of latently infected cells transformed by EBV came from studies of

cell-mediated immunity { 189, 202) in which cytotoxic T-cell clones were shown to recog-

nize an HLA-restricted (215) target on the surface of these cells. This protein, LYDMA,

was consistently present by immunologic assay; however the biochemical nature of the

antigen remained a mystery. Since the major mRNAs present in transformed cells indicated

that at least one other protein was made during latent infection, more recent work centered

around the gene, BNLF1, corresponding to this mRNA. Monoclonal antibodies and rabbit

antisera have been raised against a fusion protein encoded by this gene which react with a

63 kDa protein found in plasma membranes of transformed cells (128, 217). This protein,

named LMP or p63, bears a target for cytotoxic T-cells specific for EBV infected cells

{208), therefore providing the mechanism for T-cell recognition obtained with the original

LYDMA. LMP is phosphorylated and is isolated from the plasma membrane in an

insoluble form associated with the cytoskeletal component, vimentin {6, 121, 127). The

gene encoding LMP has been expressed in rodent cells resulting in tumorigenic

characteristics such as altered cell morphology, loss of contact inhibition and anchorage-

independent growth. The cells expressing the gene were similarly tumorigenic in nude

mice (217). LMP may assist in the transformation process as do membrane proteins of

other tumor viruses { 15), but the exact role will become more evident as the protein is

associated with changes in infected cells.

Lytic cycle proteins.

The viral lytic cycle can be activated by chemical inducers, such as phorbol esters

and sodium butyrate, leading to the production of early antigens, viral capsid antigens and

membrane antigens (231). A virus-encoded protein has also recently been implicated in

this latent to lytic switch. This protein of 43 kDa is encoded by the BZLF1 open reading

frame of het DNA and is called Zebra protein (Barn HI fragment Z Epstein-Barr replication

activator). Using deletional, site-directed, and chimeric mutagenesis and gene transfer, the

Zebra protein was demonstrated to be essential for the virus cycle switch (32).







11

Early antigens. Following disruption of latency several EBV-encoded proteins, the
early antigens (EA), which are insensitive to viral DNA synthesis inhibitors such as phos-

phonoacetic acid (PAA) and acyclovir, are initially produced ( 145, 152). They were

originally described by indirect immunofluorescence of virus-producing cell lines and were

divided into two components, diffuse (D) and restricted (R) antigens, according to their

distribution and sensitivity to methanol fixation (52, 82). The EA-D antigen is composed

of a complex of polypeptides ranging from 47 to 56 kDa and 60 kDa while the EA-R

component is a single protein of 85 kDa. The EA-D complex of polypeptides of 47-56 kDa

has been mapped to the Barn HI-M fragment, BMRF1 open reading frame (27, 159, 201)

and has been further named M-EA-D. Two proteins, identified as components M-EA-D are

essential for EBV DNA polymerase activity in conjunction with a protein of 110 kDa

(120). The 60 kDa polypeptide is encoded by the BMLF1 open reading frame from the

same BamHI-M fragment with a portion of the neighboring S fragment and has been

designated by some as MS-EA-D (26). The function of this early antigen may betrans-

activation of viral and eukaryotic promoters { 149).

The restricted component of EA is associated with the formation of cytoplasmic fila-

mentous proteins which undergo structural changes during the lytic cycle (66, 126, 153)

and more recently the protein has been mapped to a rightward reading frame of the Barnm HI

O fragment (BORF2) by amino acid sequencing. This component and a putative gene

product, BARF1, are probably associated as components of the ribonucleotide reductase

since the amino acid sequences of the ribonucleotide reductase of herpes simplex virus

(HSV) and the two EBV open reading frames share significant homology. Another puta-

tive protein which has only been identified by comparison of the DNA sequence with other

herpesviruses is encoded by the BALF5 open reading frame which shares significant

homology with HSV DNA polymerase.

Viral capsid antigens. A serologic assay which detects a cytoplasmic stain has been

the classic assay for viral capsid antigen. However, both membrane antigens as well as







12
capsid antigens are present in the cytoplasm of infected cells. Therefore, the classic test for

VCA is not specific for capsid proteins alone. With the specificity of monoclonal anti-

bodies to the major membrane antigens as well as capsid antigens, proteins specifically

located in the viral capsid have been identified. Besides the membrane antigens which are

detected by the VCA assay, three proteins with molecular weights of 125 kDa, 152 kDa,

and 160 kDa have been identified which appear to be specific for the viral capsid and

monoclonal antibodies have been used to study the 125 kDa and 160 kDa proteins ( 109,

153, 200, 206). The larger protein of 160 kDa is apparently the major component of the

capsid (210); however, the 125 kDa protein appears to be the major immunogen since all

human sera from EBV-infected individuals react with it, but not always the larger protein.

Both proteins are located in the nucleus and the 125 kDa protein can also be found in the

cytoplasm (213). Only the 125 kDa protein is glycosylated as determined both by radio-

labeling with sugars and inhibition of glycosylation, but both are late proteins made after
viral DNA replication occurs. Recent work indicates, however, that the 125 kDa protein is

probably the same protein recently described as the EBV homologue of the HSV gB (68,
156, 157). The gB homologue is encoded by the BALF4 open reading frame and is a

glycosylated 110 kDa protein not found on the virus membrane as is HSV gB but located

intracellularly. Expression of the BALF4 gene which had been fused with an E. coli

expression vector resulted in the presence of a 93 kDa protein which could be immuno-

precipitated by human immune sera. Antibodies to this protein reacted with an EBV

specified protein in infected cells of 110 kDa.

Membrane antigens. The membrane antigen complex like other EBV proteins was
first described by surface immunofluorescence of virus producing cells using human

immune sera. The complex was further resolved by analysis of the virion and the infected

cell membrane into four major envelope proteins of approximately 300-350 kDa, 200-250

kDa, 140 kDa and 85 kDa (47, 205). Three of these proteins, gp350/300, gp250/200,

and gp85, are glycosylated, while the fourth, p140, is not (195, 210). Recently a protein









of 105 kDa has been described which is membrane bound but differs from the other
membrane antigens in that it is an early protein and is not glycosylated. The possibility

exists that additional membrane proteins remain to be identified since there are a number of

unassigned open reading frames which have the characteristics of those encoding

membrane proteins.

gp35Q/200. Because they are immunodominant, more has been described about
gp350/300 and gp250/200 than the other membrane proteins. The estimated molecular
weights of the two glycoproteins vary widely, partly because of varied glycosylation, but
also because of the difficulty in precisely assigning molecular weights in polyacrylamide
gel electrophoresis to glycoproteins with such a high molecular weight. However, the use
of monoclonal antibodies has helped unify the understanding we now have of these two
glycoproteins. Principal emphasis was put on understanding their function as many of the

antibodies produced against them were capable of neutralizing viral infectivity (93, 209,
206).

Glycoproteins gp350/300 and gp250/200 are antigenically similar (93) and have
been mapped to the same gene in the Barn HI L fragment (BLLF1) { 12, 95, 97). An

intron is spliced out of the gene without change in reading frame to yield a smaller mRNA

which encodes gp250/200 (12). Glycoprotein gp350 has been purified by electrophoresis
(140) and by affinity chromatography {170) and incorporated into liposomes for

inoculation into mice and rabbits for antibody production and into cotton-top tamarins with

resulting protection from viral infection (141). The gene encoding gp350/220 has been
successfully expressed in a variety of cell lines {143, 181,222). Some neutralizing
antibodies that recognize gp350/220 are capable of inhibiting binding of the virus to its
receptor (146). The role of gp350 as the EBV attachment protein that binds to CR2 has

been further elucidated with the purified glycoprotein incorporated into liposomes and on

beads (203). Earlier work suggested that gp220 as well as gp350 plays a role in
adsorption of EBV to the receptor (220).









gES5. One of the three glycoproteins originally described, gp85, is the subject of
this work. The protein was identified first with human and rabbit sera (149, 194) and then

with monoclonal antibodies which had been produced against the viral membrane antigens

expressed on EBV producing cells (196). The basic characteristics of the glycoprotein
were defined by using two monoclonal antibodies F.2.1 and G.3.1 {195). It was

estimated to have a molecular weight of approximately 86,000 and to contain sugar
residues that could be enzymatically removed to reveal a precursor of 69 kDa. The two

monoclonal antibodies were able to neutralize infectivity of the virus implicating the

functional importance of gp85. Relatively little was added to the initial characterization
since a majority of research centered around the larger, more abundant, and
immunodominant gp350/220. Therefore, the studies described here were undertaken to

shed more light on the nature of gp85.

p.40.. The protein, p 140, is assumed to be encoded by the Barn HI N region of the
genome which codes for a protein of 142 kDa (5, 96). It shares significant homology
with a 160 kDa protein of herpesvirus saimiri which is likewise a non-glycosylated or

poorly glycosylated protein and is part of the tegument of the virus (23). Little more is
known of the nature or function of the molecule. It is difficult to extract from the virus and

antibodies to it do not react with the surface of virions or virus producing cells. Therefore,
like its HVS homologue, it may be a tegument protein.

1.5.. Another major envelope protein of 105 kDa has recently been described
which also lacks glycosylation. This protein, p105, is insensitive to PAA and is a major

component of the viral membrane antigens purified by detergent extraction {9). Mono-
clonal antibodies which were used to identify p105 demonstrate a strong cross-reactivity
with herpes simplex virus (HSV) gB (8, 51); however, this protein of 105 kDa differs
from the glycoprotein of 110 kDa demonstrated to have sequence homology with HSV gB.

Therefore, it is probable that the functional homologue of a glycoprotein essential for
infectivity by HSV, gB, is different than the genetic homologue.









Minor glycoproteins. An additional EBV-induced membrane glycoprotein is cur-

rently being characterized in our laboratory. This novel glycoprotein can be labeled with

[3H]glucosamine and runs as a diffuse band of 69-85 kDa in reducing and non-reducing

polyacrylamide gels. Digestion of the molecule with N-glycanase results in a protein of

approximately 28 kDa. The synthesis of gp69/85 is sensitive to PAA indicating it likely to

be encoded by the virus. Glycoprotein gp69/85 is of similar molecular weight but

characteristically different than another EBV-induced molecule identified with monospecific
antisera to HSV gG2. The EBV gG protein appeared under reducing conditions in SDS-

PAGE to be a diffuse protein of 80-90 kDa {8). Non-reduced samples of the protein are

slightly greater than 205 kDa, a major difference from gp69/85.

In addition to gp350/220, gp85, and p140, several other proteins were initially

described as part of the virion, gpl05/115, p56 and several minor glycoprotein in the range

of 120-190 kDa {24, 166, 207). Several of these may be breakdown products or pre-

cursors to some of the major envelope glycoproteins. A protein of approximately 53 kDa is

a major degradation product of p105 which may be the same as p56 originally described,

and the precursor for gp350/220 is a protein of approximately 120 kDa {210). Additional

membrane bound proteins are most likely encoded by EBV since many of the 84 open

reading frames of the genome encode potential proteins possessing the necessary charac-

teristics of membrane association determined by computer analysis (112), and cDNA

clones have been isolated which correspond to transcripts from unassigned open reading

frames indicating additional proteins not yet identified { 158). Since monoclonal antibodies

have been extremely useful in clearly identifying the major envelope glycoproteins and as

more monospecific antibodies made against native viral proteins, synthetic peptides, or

gene fusion products are produced, the identity and function of these proteins may become

more evident.









Biologic Significance of EBV Proteins

An understanding of the biology of EBV is important, not only to ameliorate diseases

in which it is implicated, but also to comprehend the delicate balance which the human

body has developed to cope with such a common virus infection.

B Cell Activation
It is evident that proteins encoded during latent infection of lymphocytes are involved

in alteration of B cell growth and transformation. Adsorption of EBV to the CR2 molecule

on B cells may initiate B cell activation by inducing the cells to produce growth factors

necessary for proliferation. Irradiated EBV which is capable of binding CR2, but unable to

function as a T-independent B cell activator, is able to synergize with B cell growth factor
and induce B cell activation as detected by [3H]thymidine uptake {98). Therefore,

infective virus maintains the ability of B cell activation in the absence of external growth

factors. The evidence that EBV-transformed B cells secrete growth factors { 16, 130, 175)

is supported by reports eluded to previously that EBNA-2 can upregulate the production of

CD23. This molecule has been identified as a growth-factor receptor and an autocrine

growth enhancer (73, 198, 218). Apparent from this information is that the virus has

adequately adapted to establish itself in a replicating and potentially long-lived cell.

Virus Entry

Differing from the latent proteins which are vital for transformation and latency, and

from the structural capsid proteins, the envelope proteins are most likely involved in steps
which enable the virus to enter and exit cells during infection (44). The primary step in

infecting host cells is that of adsorption of the virion to the cell surface. This step has been
clearly defined by elegant work which demonstrated that the large membrane glycoproteins,

gp350/220, bind to the cell receptor CR2 (57, 146, 203,220). Once the virion has been

brought in close proximity to the cell surface, overcoming the charge repulsion which

occurs when lipid bilayers are juxtaposed, the genomes of enveloped viruses are usually

introduced into host cells by membrane fusion { 129). Paramyxoviruses and Sendai








17
viruses are capable of fusing with the membrane of the cell at the surface independent of a

pH change {223). Other enveloped viruses, however, introduce their genomes into the

cell via receptor-mediated endocytosis in a pH dependent fashion. Acidification of the

endosome activates the fusion protein by causing a conformational change resulting in

fusion of the two membranes ({ 37, 45, 63, 229). Upon entry, via fusion, the nucleocapsid

is probably transported to the nucleus at a nuclear pore resulting in release of the viral DNA

into the nucleoplasm which has been shown for HSV using temperature sensitive mutants

(11). The transport may involve interactions with the cellular cytoskeleton as demon-

strated with adenovirus (33).
Virus Egress

Egress of the virus from infected cells is also an essential function of envelope pro-

teins. Mature Herpes simplex viruses are assembled by budding at the inner nuclear

membrane ( 38). The glycoproteins which are incorporated in the virion are transported to

the inner nuclear membrane. Both gp 110 of EBV and its genetic homologue HSV- 1 gB

contain stretches of basic amino acids within the predicted inner nuclear membrane domain

(68) which are similar to sequences of other nucleus targeted proteins { 104). Association

of viral glycoproteins with nucleocapsid and other structural proteins may cause an

accumulation of the glycoproteins at the inner nuclear membrane (107). The nucleocapsid

containing viral DNA is enveloped as the virus particle buds into the perinuclear space.

Additional modification of the glycoprotein probably occurs in the Golgi complex {101,

102, 103) after which the mature virion is transported via vesicles to the cell surface.

Fusion of the transport vesicle with the cell plasmalemma is probably the means whereby

the mature virion is released into the extracellular space { 180). Although these

mechanisms have been described for Herpes simplex virus, they provide a model whereby

the role of EBV glycoproteins may be investigated.









Purpose of This Work

It was with the purpose of elucidating the role that gp85 plays in initiation of

infection with EBV that this work was begun. Since antibodies to gp85 neutralized virus

infectivity, we hypothesized that the glycoprotein, a constitutive part of the viral envelope,

was essential for one of the steps by which the virus enters the cells. Also, although

binding and fusion are the two functions characteristically prescribed for envelope glyco-

proteins, they are also important to the later steps of transport and release of virions. In

order to characterize the functional properties of gp85, it was first necessary that its bio-

chemical and structural characteristics be understood.

The initial phase of the work described herein involved characterization of a protein

reactive with a monoclonal antibody which we felt reacted with gp85. The remainder of the

dissertation describes the purification of gp85, mapping of the glycoprotein to the genome,

and production of polyclonal and monoclonal antibodies to gp85.













CHAPTER II

BIOCHEMICAL CHARACTERIZATION OF gp85

Introduction

Analysis of membrane soluble proteins which were recognized by human sera con-

taining high antibody titers to membrane antigens provided evidence for the existence of

four major membrane proteins induced by EBV, one of which was gp85 {144, 167, 195).

Since the antibodies which recognized membrane antigens were capable of neutralizing

infectivity of the virus { 154), much interest devolved on these proteins as important to

induction of immunity. Specific antibodies to the major membrane proteins were first

experimentally produced by immunizing rabbits with purified B95-8 strain virus (47).

This work confirmed the identification of the major proteins made with human sera and

also furnished additional biochemical characterization. Stmad and colleagues were able to

produce monoclonal antibodies to the membrane proteins, two of which, F.2.1 and G.3.1

reacted specifically with gp85 (195). Using these monoclonal antibodies they defined

some of the biochemical properties of the protein {194) which combined with the

information gained from biosynthetic studies with rabbit antiserum (48) provided a broad

understanding of the glycoprotein.

Pulse-chase experiments by Edson and Thorley-Lawson using rabbit antiserum

identified a precursor to gp85 of 83 kDa which was present only in the first hour of chase

and rapidly underwent post-translational processing of the glycosyl chains to reach the

mature molecule of 85 kDa. The protein was shown to possess both high-mannose and

complex type oligosaccharides by Endo H digestion of the glycoprotein which resulted in a

75 kDa polypeptide replaced during the chase period by an intermediate protein of 81 kDa

(48). This shift from 75 kDa to 81 kDa was assumed to be processing of some of the

19







20

high-mannose to complex oligosaccharides during the chase period. Stmad and colleagues

similarly showed with pulse-labeling immunoprecipitates that Endo H was able to remove

the high mannose type N-linked carbohydrate (214) resulting in a series of five proteins

from 71 kDa to 86 kDa; however, treatment with Endo D did not alter the 86 kDa protein

indicating that complex-type N-linked carbohydrate may not be present (193). Growth of

the cells in tunicamycin which inhibits the transfer of core sugar to asparagine residues

{ 50) resulted in a protein of 69 kDa, which was the smallest protein they were able to

identify using the monoclonal antibodies. Treatment with monensin which blocks some

processing of glycoproteins { 102) resulted in an 84 kDa protein. This information sug-

gested, therefore, that gp85 was a polypeptide of approximately 70 kDa which contained

perhaps five N-linked sugars, resulting in a final glycoprotein of approximately 85 kDa.

Although additional monoclonal antibodies were produced against EBV membrane

antigens, the two developed by Strnad and colleagues were the only ones described in the

literature which reacted with gp85. A small amount of these two antibodies, but not the

hybridomas which produce them, was available in our laboratory; the limited supply cur-

tailed the number and type of experiments which could be done to further the characteri-

zation of gp85. N. Balachandran and Jeanne Pittari in our laboratory then produced a

number of monoclonal antibodies against the membrane antigens, one of which, E1D1,

reacted with a protein of 85 kDa. The work presented in this dissertation began by ana-

lyzing the protein recognized by E1D1 to determine whether it reacted with the same 85

kDa protein as that immunoprecipitated by the antibodies originally used to characterize

gp85.

Materials and Methods

Cells and Virus

The cell lines used in these studies were the EBV (B95-8 strain) transformed mar-

moset cell line, MCUV5, and the human Burkitt's lymphoma isolate, P3HR-1 Clone-13







21

(Cl-13) {90) a lymphoblastoid cell line infected with EBV. Several other lymphoblastoid

cell lines were used as EBV-negative controls including Raji cells which carry EBV DNA

but do not produce virus and BJAB cells which are EBV negative. B lymphoblastoid cell

lines were grown as suspension cultures in RPMI 1640 medium (Gibco) containing 100
IU/ml penicillin and 100 gg/ml streptomycin (RPMI P/S) and 5% (for Raji and MCUV5)

or 10% (for CI-13 and BJAB) heat-inactivated fetal bovine serum (FBS). Both the EBV-
producing cells and non-producing cells were fed by approximately doubling the culture

volume every two to three days.
One of the major difficulties in working with EBV is that the virus preferentially
establishes latent rather than productive infections in lymphocytes. Virus was obtained

from several of the lymphoblastoid cell lines in which a low percentage, usually around

5%, of cells spontaneously produces virus. Virus yields were often low, always unpre-

dictable and assays for virus infectivity are cumbersome. However, 30 ng/ml of the tumor
promoter 12-0- tetradecanoyl phorbol-13-acetate (TPA) were used to increase the number
of cells entering the replicative cycle (231). TPA induction resulted in approximately 20-

30% of the cells producing virus. Virus pools were made from 4.5 liters of cells grown to
confluency and then induced with TPA. Seven days following induction, the cells were

removed by centrifugation at 4,000 g for 10 min after which the virus was pelleted by a 90

min centrifugation at 16,000 g with 0.1 mg/ml Bacitracin added to the media to reduce

virus aggregation. The pelleted virus was then resuspended in RPMI P/S supplemented

with Bacitracin at a 250-fold concentration and stored frozen at -70C {183).
Assays of Viral Antigens and Virus Binding

Viral antigen expression on the surface of TPA-induced cells and binding of EBV to

EBV receptor positive and EBV receptor negative cells were both measured by indirect
immunofluorescence on cells fixed with 0.1% para-formaldehyde. The antiviral antibodies

were bound to the viral proteins expressed on the surface of TPA-induced cells or virus
bound to the receptor and visualized with fluorescein-conjugated F(ab')2 sheep anti-mouse









immunoglobulin (Cappel Laboratories, Cochranville, Pa.). Fluorescence intensity and

distribution were assessed visually on a Nikon epifluorescence microscope. Virus concen-
trations in concentrated samples were determined using this indirect immunofluorescence

assay to determine the end-point dilution at which the virus was detectable.
Radiolabeling of Proteins
Proteins were radiolabeled intrinsically with [35S]methionine, [3H]glucosamine, and

[3H]galactose and extrinsically with 125Iodine. Metabolic labeling was accomplished by

inducing Cl-13 cells with TPA as previously described for two days, resuspending the cells

in RPMI media deficient in the amino acid or sugar to be labeled and incubating for 15 min.
The radiolabeled amino acid or sugar was then added to the medium and incubated for 18
hours in the presence of TPA. For treatment with phosphonoacetic acid (PAA) the cells
were grown for two days in normal media supplemented with 200 gpg/ml PAA and labeled

as usual with the same concentration of PAA in the labeling media (190). Labeling in the

presence of tunicamycin was accomplished by incubating the cells in medium containing 5
p.g of tunicamycin per ml for 3 hours before labeling as usual in the presence of

tunicamycin (99). Labeled cells were washed by centrifugation two times and stored at
-200C as a pellet.
Extrinsic labeling of 1 x 107 cells resuspended in phosphate buffered saline (PBS)

was accomplished with 0.5 mCi of 1251 by the use of Iodogen@ (tetrachlorodiphenylgly-
couril; Pierce Chemical Co.) {60) followed by five washes in PBS to remove free iodine.

Concentrated virus was similarly labeled by resuspending 4.5 ml of concentrated virus
(250X) in 300 pl of PBS, labeling with 125I by the use of lodogen and separating the free
iodine by sieving on a 5 ml bed volume desalting column (P6-DG; Bio-Rad).
Efficiency of labeling was determined by comparing 10 Wd samples which were pre-
cipitated and washed with two changes of cold 5% trichloroacetic acid (TCA) and one wash

of cold 95% ethanol and then counted on a gamma counter for 125I and B-scintillation
counter in scintillant fluid for [35S] and [3HI. The ratio of TCA precipitable to







23

non-precipitable counts was used to determine the amount of free radioactivity remaining in

the sample.

Antibodies

The monoclonal antibodies E1D1 and E8D2 were produced by N. Balachandran in

our laboratory {8,9). The E1D1 hybridoma was monocloned by limiting dilution. Cells

were diluted to one cell per well in a flat-bottom 96-well tissue culture plate. Culture

supernatant from the wells exhibiting growth was screened by indirect immuno-

fluorescence. Monocloning by limiting dilution was repeated one more time to assure the

monoclone nature of the hybridoma. Concentrated culture supernatant was used to deter-

mine the isotype of the immunoglobulin produced using double immunodiffusion in

agarose. High-titered ascitic fluid was raised by injecting the -106 cells into BALB/c mice

which had been primed with 0.5 ml of 2,6,10,14-tetramethylpentadecane (Pristane){ 10).

The monoclonal antibody F.2.1 {196) was a gift from David Thorley-Lawson

(Tufts Medical School, Boston, MA) received after most of the preliminary characterization

using EID1 had been completed. The hybridoma cells were monocloned as described

above and used for the production of high titered ascitic fluid or concentrated culture

supernatant. The rabbit anti-EBV serum (R2) was made previously in our laboratory by
immunization with TPA-induced Cl-13 cells and then extensive absorption with Raji cell

lysates to remove cell-specific antibodies. The rabbit serum R77 was made against the

EBV gB fusion protein and was a gift of Dr. Elliott Kieff.

All monoclonal antibodies and R2 rabbit serum were purified by affinity chroma-

tography on protein-A coupled to Sepharose 4B (Sigma). Approximately 8 ml of ascitic
fluid or serum which had been centrifuged at 400 g or 50-100 ml concentrated culture

supernatant was filtered through a 0.8 pm filter and placed on a 15 ml bed-volume column

of protein-A sepharose which had been extensively washed with PBS. The antibody was

adsorbed to the column, eluted with 0.1 M acetic acid in PBS, the pH equilibrated with 1.5

M Tris pH 8.8 and the purified antibody dialyzed with four changes of PBS for 12 hours.








Dialyzed antibody was concentrated by dialysis against polyethylene glycol (PEG) to the

original volume of ascitic fluid or a ten-fold concentration of culture supernatant.
Immunoprecipitation

Radiolabeled proteins were solubilized from cells and virus by resuspending the pel-
let of virus or cells in radioimmunoprecipitation buffer (RIPA) containing 0.05 M Tris
hydrochloride pH 7.2, 0.15 M NaCl, 1.0% deoxycholate, 1.0% Triton X-100, 0.1%
sodium dodecyl sulfate (SDS), 100 U aprotinin per ml and 0.1 mM phenylmethylsul-
fonylfluoride (PMSF). The sample was then sonicated for 30 sec and centrifuged at
100,000 g for 1 hour to remove the non-soluble portion of the lysate. Samples of lysates
were TCA precipitated as described previously and in experiments in which two or more
lysates were used, the counts were equilibrated by dilution in RIPA buffer. A volume of
400 gl of the supernatant was added to 125 p1 of protein A-agarose beads (Genzyme
Corp., Boston, Massachusetts) and 25 pIl of purified antibody and rocked for 2-4 hours at
4'C. Lysates containing proteins labeled with [35S]methionine were preabsorbed with 100
p1 protein A-agarose prior to precipitation with antibody. The samples were washed by
centrifuging in an Eppendorf centrifuge for 4 min, removing the supernatant, and
resuspending in fresh RIPA buffer, this was repeated three or four times. The final
precipitates were dissociated by boiling for 3 min in 100 p1 sample buffer containing 0.65
M Tris, 1% SDS, 10% glycerol, and 1% 2-mercaptoethanol. A five pl aliquot of each
sample was counted to determine the amount of radiolabeled protein immunoprecipitated.
The samples were used for electrophoresis or stored frozen at -20*C.
Electrophoresis
Immunoprecipitated samples were analyzed by SDS-polyacrylamide gel electro-
phoresis (SDS-PAGE) (116). Polyacrylamide gel separating buffer with 0.65 M Tris pH
8.8, 9% or 12.5% acrylamide, 0.1% SDS, 0.1% glycerol, 0.83% ammonium persulfate,
and 0.28% N,N'diallyltartardiamide (DATD) or bis-acrylamide (BIS) overlayed with 1-
butanol was polymerized for at least 5 hours. After polymerization, the overlay was







25
washed with 0.65 M Tris pH 6.8 and stacking buffer with 0.65 M Tris pH 6.8, 4% acryl-

amide, 0.1% SDS, 0.05% glycerol, 0.83% ammonium persulfate, and 0.28% DATD was

poured and polymerized for 45 min. The samples were loaded into individual wells and the

gel submerged into reservoir buffer of 25 mM Tris and 200 mM glycine with 1% SDS and

electrophoresed for approximately 6 hours under a constant current of 26 mAmps or

overnight at a constant voltage of 50 V. Gels were stained with 0.1% coomassie blue stain

for 15 min and destined with several changes of destain (7% methanol and 7% acetic acid)

over a 24 hour period to visualize the molecular weight standards. Gels containing proteins

labeled with 1251 were dried on filter paper and placed in contact with XAR film (Eastman

Kodak Co., Rochester, N.Y.) at -70"C for fluorography. Gels containing [35S] and [3H]-

labeled proteins were infused with 2,5-diphenyloxazole (PPO) by treating the gel in two 30

min changes of dimethyl sulfoxide (DMSO) followed by 22% PPO in DMSO for 3 hours.

The PPO was precipitated in the gel by soaking in dH20 for at least 1 hour and dried on

filter paper. The gels were then placed in contact with XAR film at -70'C for fluorography

{17).

Cross-adsorption

The identities of the two proteins immunoprecipitated by E1D1 and F.2.1 were con-

firmed by cross-adsorption with the two antibodies. The 1251 labeled Cl-13 lysate was

immunoprecipitated for 2 hours at 4*C with one of the antibodies. After centrifugation, the

supernatant was again reacted with the same antibody, and the procedure was repeated

three to four times. The supernatant collected after the last immunoprecipitation, as well as

a sample of the original unabsorbed antigen, was then reacted with the cross-reactive anti-

bodies. All of the immunoprecipitates were analyzed by SDS-PAGE as described above.

Electroelution

Twelve identical samples of 125I labeled virion proteins immunoprecipitated with

either E1D1 or F.2.1 were electrophoresed by SDS-PAGE as described previously using a

preparative comb. The protein was visualized by autoradiography, sliced out of the dried







26

gel and cut into 15 mm length pieces. The gel pieces were swollen in Tris buffer 0.65 M,

pH 6.8 for 10 to 15 min and then placed in an Elutrap (Schleicher and Schuell, Inc.,

Keene, NH){ 100). The chamber was filled with reservoir buffer and applied to a 200 V

current for 12 hours. The protein was collected by electrophoretically passing it through a

special membrane into a 250 Wl1 trap. The amount of radiolabeled protein was determined

at various time-points during the electroelution by taking 10 pl samples and counting by

gamma counter. At the end of the elution the protein was collected and stored at -20'C to

await proteolytic digestion.

V8 Protease Digestion

Samples containing equal amounts of 1251 were partially digested {30) with

Staphylococcus aureus V8 protease (Sigma) in 62.5 mM Tris-HCl, pH 6.8 for 30 min at

37C. The reaction was stopped by addition of an equal volume of sample buffer

containing 2-mercaptoethanol and boiling for 3 min. Boiled samples were then analyzed by

SDS-PAGE in 12.5% acrylamide cross-linked with 0.28% N,N'-methylene-bis-acrylamide

(BIS).

Results

Biosynthesis

The monoclonal antibody, E1D1, was selected for EBV specificity by indirect

immunofluorescence on EBV-producing Cl-13 cells and non-producing Raji cells. The

protein recognized by the antibody was determined by immunoprecipitation of C1-13 virus

and BJAB cells labeled with 125I (Figure 2.1). The antibody precipitated a protein of

approximately 80 kDa in the DATD cross-linked gels which corresponds to the elec-

trophoretic mobility of gp85 in this type of gel. However, the same antibody did not react

with any proteins from the virus-negative BJAB cell line. This was the first indication that

E1D1 may be reacting with the major membrane antigen gp85. In order to determine if the

protein immunoprecipitated by E1D1 had the same biochemical characteristics as gp85,


















AN -f -


205-






116


97


66


1 2




Figure 2.1. SDS-PAGE analysis in 9% acrylamide of proteins immuno-
precipitated by E1D1 from 125I labeled TPA induced C1-13 cells (lane 1)
and BJAB cells (lane 2). Molecular weight markers in kDa.


















205-


116-


97-


66-


- .p:::.,


45- W


1 2


Figure 2.2. SDS-PAGE analysis in 9% acrylamide of proteins immuno-
precipitated by E1D1 from [35S]methionine labeled TPA induced Cl-13 cells
grown in the presence of tunicamycin (lane 1), PAA (lane 3) or without
drug (lane 2). Molecular weight markers in kDa.


q


Mom 4000 ONIM







29

virus-producing cells were biosynthetically labeled with [35S]methionine in the presence of

PAA and tunicamycin and immunoprecipitated with E1D1 (Figure 2.2). The failure of

EID1 to immunoprecipitate the protein from cells grown in the presence of tunicamycin

indicated that its glycosylation had been altered sufficiently to eliminate reactivity with the

antibodies. The disappearance of the 80 kDa protein from cells grown in the presence of

PAA would suggest that it is sensitive to the inhibition of viral DNA synthesis and there-

fore classified as a late protein.

Glycoslation

Glycosylation of the protein reactive with E1D1 was shown by labeling with [3H]

labeled sugars. [3H]glucosamine was used to biosynthetically label the TPA-induced Cl-13

cells which were then immunoprecipitated by E1D 1 (Figure 2.3). The glycoprotein of 80

kDa precipitated by E1D1 was labeled quite efficiently as compared with p105. A rabbit

anti-EBV serum, R2, reacted with the major membrane antigens gp350/220 in a densely

broad band greater than 150 kDa. Also within that region was a band of approximately 170

kDa which is either part of the gp350/220 band or a separate entity. A faint amount of

gp350/220 coprecipitated with the 80 kDa glycoprotein specific for E1D1; similar copre-

cipitations of major membrane antigens are occasionally seen with extremely efficiently

labeled samples. This may indicate an association of the major membrane antigens in the

envelope spikes or incomplete solubilization by detergent. A major glycoprotein of 110

kDa was precipitated by the anti-EBV serum as well as the rabbit anti-gB serum (R77).

This rabbit serum, made against the B-galactosidase fusion protein of the BALF4 open

reading frame, reacted with a protein in the same region as the 80 kDa protein brought

down by E1D1. Whether this is a coprecipitation as mentioned above or a specific reaction

between E1D1 and gB cannot be determined from this gel.

The question of the similarity between the two proteins of approximately 80-85 kDa

seen in the [3H]glucosamine gel was partly resolved by labeling with [3H]galactose (Figure

2.4). These results indicate that a protein other than gp85 is probably specifically reacting



















-205K




S.- 116K


-97K


3 66K





-45K








Figure 2.3. SDS-PAGE analysis in 9% acrylamide of proteins immuno-
precipitated by (from left to right): R77 (rabbit anti-gB), E8D2, R2 (rabbit
anti-EBV), and E1D1 from [3H]glucosamine labeled TPA-induced Cl-13
cells.

















N___.. ---


* -205K


44*


116K


97K


S- 66K






* 45K


Figure 2.4. SDS-PAGE analysis in 9% acrylamide of proteins immuno-
precipitated by (from left to right): E1D1, E8D2, and R77 from
[3H]galactose labeled TPA-induced Cl-13 cells.








32
with the R77 serum. The mature gB of 110 kDa is labeled more efficiently than either the

80 kDa protein brought down by E 1D1 or p105. However, the glycoprotein of approxi-

mately 70-75 kDa which reacted with the anti-gB serum had the greatest radioactive inten-

sity, indicating a definite difference in the molecular weights of the potential partially pro-

cessed gB precursor (pgB) and the putative gp85.

Cross-immunoprecipitation

Although the 80 kDa protein immunoprecipitated by E1D1 had all the same charac-

teristics as the molecule recognized by F.2.1, we still could not be certain that they were the

same protein. At this point we were fortunate to receive the hybridoma line producing

F.2.1 enabling us to obtain larger amounts of the antibody which could be used in compar-
ative studies. Both antibodies were used in a cross-adsorption experiment to see if the

precipitation of the protein by one antibody could abolish the reactivity of the other. E1D1

was used to repetitively immunoprecipitate 1251 labeled Cl-13 virus. The precipitate from

each adsorption as well as the cross-adsorption with the opposite antibody was analyzed by
SDS-PAGE. The consecutive immunoprecipitations using E1D1 (Figure 2.5) resulted in a

reduction in amount of protein with each respective sample. Adsorption with protein A-

agarose alone preceded the precipitation with F.2.1. In comparison to a control precipi-

tation using F.2.1, the amount of protein was drastically reduced, although still present.

This gel displays the prominent 60-66 kDa protein efficiently labeled with 1251 but not
[35S]methionine, [3H]galactose, [3H]glucosamine which reacts with E1D1 but not F2.1

and appeared to be either more abundant than the 80 kDa protein or cleared less efficiently

by E1D1. The 60-66 kDa protein was of sufficient concern to require further examination
(see Chapter 3).

A parallel experiment using F.2.1 for repetitive immunoprecipitations followed by
E1D1 showed similar results (Figure 2.6). The residual protein which remained following

clearance with the first antibody perhaps represented a population of molecules in which the

specific epitope recognized by the antibody was not accessible. Therefore, when reacted





















1 2 3 4 5
- om


6 7
4* ..o 4. .


205-



116-


do


45-


Figure 2.5. SDS-PAGE analysis in 9% acrylamide of proteins immuno-
precipitated by E1D1 (lanes 1-5), protein A-agarose alone (lane 6) and
F.2.1 (lanes 7 and 8) from 1251 labeled CI-13 virus.




















1 2 3 4 5
am e..-e -Me- ews-', .- a-


6 7 8
Mo.p


e
; .:


Figure 2.6. SDS-PAGE analysis in 9% acrylamide of proteins immuno-
precipitated by F.2.1 (lanes 1-5), protein A-agarose alone (lane 6) and
E1D1 (lanes 7 and 8) from 1251 labeled Cl-13 virus.


205-



116-


97-

66-




45-









with the second antibody, some of the protein remained to be precipitated. This would

account for reduction without total clearance of the protein. The epitopes recognized by

E1DI and F.2.1 are distinct since they differ in their abilities to neutralize and bind

denatured protein.

Proteolvtic Digestion Patterns

In order to obtain sufficient quantity of the glycoproteins recognized by E1D1 and

F.2.1 for comparison by V8 protease digestion, they were immunoprecipitated by each

antibody and electroeluted from preparative gels. The electroelution successfully removed

the majority of the protein from the acrylamide gel over a period of 10 hours (Figure 2.7)

as well as separating the non-TCA precipitable counts from the labeled protein. The


18390

16551
1 \* Non-TCA
14712
TCA
12873

S 11034

9195

7356

5517

3678

1839

0

1 2 3 4 5 6 7 8 9
Time (hours)


Figure 2.7. Radioactivity electroeluted from 1251 labeled gp85 immunoprecipitated
with F.2.1, separated by SDS-PAGE, cut out of the gel and eluted in an Elutrap.
indicate total counts eluted into trap, n indicate TCA precipitable counts.




















(0
co 'N 'No
I I^


bo
00

4-I4 C.
bo


40-so-

Isl
v 0











0a^ >0
v 8 5^




t)l

'-4



-c




C ag 4s
2 0^


J~*. H.
4,









electroeluted proteins were then digested with S. aureus V8 protease and analyzed by

SDS-PAGE (Figure 2.8). The molecules had very little natural degradation during the

elution process but the addition of 10 gpg, 50 gpg, and 100 gg of the protease resulted in

significant digestion. At each of the different concentrations of protease, the digestion

patterns of the two proteins were identical.

Discussion

Although gp85 had been previously described and characterized as one of the major

EBV membrane glycoproteins, the ability to study the glycoprotein further was restricted

by the few monoclonal antibodies which reacted with it. The only two monoclonal
antibodies made, F.2.1 and G.3.1, were not originally available to us in an unlimited

supply. Although several groups produced monoclonal antibodies to other membrane anti-

gens, none reported additional antibodies to gp85; this was probably due to the relatively

small amounts of the protein which are actually found in the infected cell membrane as well

as the virion envelope.
A new monoclonal antibody (E1D1) produced in our laboratory which immunopre-

cipitated a molecule of the same electrophoretic mobility as gp85 potentially provided us

with a new tool to study this protein fully. However, it was first necessary to show that

E1D1 was actually reacting with the same molecule. By biochemical characterization using

biosynthesis inhibitors and alterations, as well as sugar labeling, we were able to show that

the 85 kDa protein which reacted with E1D1 had the same characteristics as that described

for gp85. E1D1 reacted with a protein which had the same molecular mass as gp85, con-
tained approximately the same amount of N-linked glycosylation, and was affected by

tunicamycin, PAA, and monensin (data not shown) as had been originally described of

gp85. E1D1 was originally selected for by screening virus producing cells with surface
immunofluorescence, and the small speckled fluorescence which was obtained using E1D1

was characteristic of gp85 compared to fluorescent staining patterns obtained when other







38

membrane antigens are seen such as p 105 which has a stringy appearance and gp350/220

which demonstrates a very strong patchy fluorescence. This task of comparing the two

proteins was further facilitated by the acquisition of the hybridoma producing the mono-

clonal antibody F.2.1. With both antibodies readily available we were then able to make a

comparison of the two proteins using competitive immunoprecipitations and Cleveland

digestions and show that they were the same.

Although the two antibodies did not reciprocally clear all the protein from the lysate

in the cross-adsorption immunoprecipitations, the majority of the protein which reacted

with either monoclonal antibody was removed. The residual protein which either antibody
was unable to immunoprecipitate was felt to be a population of gp85 which did not bear the

epitope for that specific antibody but did present the epitope necessary for binding with the

other antibody. It seemed reasonable to assume that the two monoclonal antibodies were

reacting with different epitopes on the protein since their characteristics in neutralization

experiments were quite distinct; F.2.1 is a very efficient neutralizing antibody (see Chapter

5) whereas E1D1 neutralizes only weakly if at all. Partial digestions of each of the proteins
revealed a remarkably identical pattern which demonstrated that the two proteins were

indeed the same. The confirmation that E1D1 reacted with gp85 gave additional importance

to the concurrent experiments being performed to purify gp85 by affinity chromatography

and also provided another monoclonal antibody which could be used to study the nature of

the protein.












CHAPTER III
PURIFICATION OF gp85

Introduction
Various methods exist for purification of proteins such as ion exchange chromatog-

raphy {91}, molecular sieving (160), adsorption chromatography {31), affinity chro-

matography { 124, 139) and electrophoresis { 116). The purity and the condition of the

isolated protein depends greatly on the method used. We had two objectives in purifying

gp85. One was to obtain purified protein in sufficient amounts for N-terminal sequencing;

this would allow us to identify the gene encoding gp85. Our intention was thus to develop

a protocol for isolation of about 100 g.g material with greater than 85% purity in a relatively

undenatured state. The other was to obtain material for the production of monospecific

antibodies that would recognize the native molecule and, preferably, inhibit virus infec-

tivity. Affinity chromatography with an antibody which specifically reacted with gp85

appeared to be a logical choice. We began by using the only gp85 specific antibody,

E1D1, which we had in sufficient quantities for coupling to Sepharose to purify gp85 from

virus producing cell lysates.


Materials and Methods
Virus Preparation

Epstein-Barr virus concentrated 250-fold from culture supernatant of 7 day TPA-
induced Cl-13 cells or MCUV5 cells was used for the purification of gp85. In order to

trace the purification of the protein, 200 ml of virus producing cells were labeled with
[35S]methionine as described in the previous chapter. The labeled cells were then fraction-

ated on a Ficoll-Hypaque gradient to separate the virus producing cells from non-producing

cells by difference in density of the two populations. Cells were washed two times in PBS

39









and resuspended at 1 x 107 cells/ml. Lymphocyte Separation Medium (LSM)(Organon
Teknika Corp., Durham, N.C.) was diluted 50% and 25% in PBS and carefully layered so

that a sharp interface existed between the upper 25% and lower 50% fractions. One
milliliter of the cells were placed on top of the LSM gradient and centrifuged at 400 g for

15 min without brakes. The interface was collected and washed in PBS three times to

remove LSM.
Both the labeled cells and unlabeled virus were pelleted and resuspended together in

8 ml lysing buffer (1 M Tris-HCI pH 7.2, 0.15 M NaCi, 1% Triton X-100, 1% deoxy-
cholate, 100 U/ml Aprotinin, and 1 mM PMSF). Lysis of membranes was facilitated by
probe sonication for 30 sec and sonicated material was centrifuged at 100,000 g for 1 hour

to remove the non-soluble nucleocapsid and cytoskeleton. TCA precipitation of 10 Wl of
the lysate was counted by B-scintillation counter to determine the amount of labeled protein
present. This lysate was then used for affinity chromatography.
Affinity Column
An affinity column was made by coupling E1D1 to cyanogen bromide activated

Sepharose (CNBr-Sepharose)(Sigma, St. Louis, MO) { 161). Seven milliliters of the
purified antibody at 7 mg/ml was added to CNBr-Sepharose which had been swollen and

washed in 0.2 M carbonate/bicarbonate buffer, pH 8.6. The coupling material was rocked
for 4 hours at room temperature, after which unbound antibody was removed by

centrifugation at 400 g for 10 min. Protein concentration of unbound antibody was

determined by the Lowry method ({ 125} to estimate coupling efficiency. Ethanolamine

(1M, pH 8.0) was used to block activated Sepharose by reacting overnight at 4'C. The
coupled beads were washed in low pH, high pH succession 10 times using 0.1M acetate
buffer and 0.1 M carbonate buffer. A column of approximately 5 ml was poured by
allowing the beads to settle followed by washing with 10 column volumes of carbonate

buffer.









Before use the column was washed and equilibrated with 100 ml of lysing buffer.

The lysate was then passed over the column 5 times and the column washed with washing

buffer (0.05 M Tris-HCl, pH 7.2, 0.5 M NaCi, 0.5% Triton X-100, 100 U/ml Aprotinin,

and 1mM PMSF) until all unbound radioactivity was removed. An additional 50-100 mls

of washing buffer was passed over the column after background levels of radioactivity

were reached to further clear the column of any unbound proteins. Three molar potassium

thiocyanate (KSCN) in 30 ml PBS followed by PBS was used to elute the bound protein.
Eluate was collected in 1 ml fractions. All fractions containing [35S] counts were pooled

and dialyzed against PBS with 0.01% sodium azide for 12 hours with four changes of
buffer until the presence of KSCN was undetectable by reacting with supersaturated

barium chloride. The protein was then concentrated by dialysis against PEG to

approximately 3 ml.
Silver Staining and Autoradiography

The purity of the protein was determined by silver staining and autoradiography of

the sample analyzed by SDS-PAGE. One hundred micoliters of the purified, concentrated

protein was boiled in 100 0l sample buffer and applied to gel electrophoresis on a 9%

acrylamide gel cross-linked with DATD as described previously. The gel was then silver

stained by a modified method of Merril {79, 133). Basically, fixation of the gel was done

in 30% ethanol, 10% acetic acid for 1 hour followed by two washes in 10% ethanol. The

gel was then reduced with 0.5% Farmer's reducer (2.27 mM sodium thiosulfate, and 60.7
mM potassium ferricyanide) for approximately 30 sec and washed with running dH20 for

5 min followed by five 10-min washes in dH20. The gel was soaked for 30 min in 0.1%
M silver nitrate and then washed three times in dH20. The silver impregnated gel was

developed with 2.5% sodium bicarbonate, with 0.02% formaldehyde until the background

of the gel began to change color and was washed in dH20. Because the silver stain has a
quenching effect on the B-rays emitted from the [35S]methionine labeled proteins, some of

the gels were stained directly with coomassie blue as described previously. Enhancement









of the radioactive signal by infusing with PPO as described previously was done for both

silver stained and coomassie blue stained gels after which the gels were dried and exposed

to XAR film for fluorography.

Antibody Production

The purified protein was used as antigen for the production of antibodies in rabbits.

One rabbit, R18, was immunized with protein purified from MCUV5 cells and virus. The

first injection included labeled Cl-13 cells. A second rabbit, R24, was immunized with

protein purified from Cl-13 cells and virus. New Zealand White rabbits (North Florida

Rabbitry) were injected subcutaneously with 1.5 ml of the purified antigen emulsified with

1.5 ml complete Freund's adjuvant. Subsequent injections used the same amount emulsi-

fied with incomplete Freund's adjuvant. After 3 approximately biweekly injections the

rabbits were bled from the ear vein 30 min after anesthetizing with 0.1 ml/ kg of Innovar-

vet (fentanyl 0.04% and droperidol 2%). The rabbit was given two additional injections

and bled again which was repeated one more time. R24 developed a serious abscess from
an injection site Pasteurella infection and therefore only received four injections before it

had to be sacrificed. The rabbit serum was purified by affinity chromatography on protein

A-agarose. One other rabbit died of Pasteurella pneumonia after five injections of purified

antigen but before any immune serum was obtained. The major drawback besides the rela-

tive lack of protein available to us was in keeping the rabbits healthy long enough, despite

their bacterial infested surroundings, to give enough injections of the purified gp85 for

high-titered antisera.

Results


By monitoring the amount of [35S] eluted from the column after the addition of 3M

KSCN, we were able to trace the elution of the protein. All of the bound protein eluted in a







43

single peak and was removed rapidly in the presence of the eluting agent (Fig. 3.1).

Approximately 1% of the CPM loaded was obtained in the eluate (Table 3.1).

Purified gp85

The purity of the protein eluted from the affinity column was shown by silver stain-

ing and autoradiography of the eluted sample after SDS-PAGE (Figure 3.2). The lysate

which was applied to the column as well as the flow-through obtained after adsorption of

the protein to the column contained a substantial amount of protein. The eluted protein

came off in a relatively pure state with one major [35S] labeled contaminant around 45 kDa

which is probably the major cytoskeletal component, actin. Protein contaminants which

were not metabolically labeled and undetected by silver staining could not be discounted.

Densitometric tracing of the autoradiograph (Figure 3.3) of this gel indicated that over 85%

of the CPM separated by SDS-PAGE was contained in the band corresponding to gp85.




530,
477_ _
424_ _
371
318 \
265
212_ _
159_ _
106
53
0 i I I I I I I I
1 5 10 15 20 25
Fraction
Figure 3.1. Elution profile of [ 3Smethionine contained in 3 ml samples from an
E1D1 antibody affinity column after eluting adsorbed proteins with 3 M KSCN
followed with extensive PBS washing.















Table 3.1. Summary of gp85 purification experiments with analysis of radiolabeled
protein purified from total lysate of labeled and unlabeled virus producing cells.
Experi- Composition Lysate Concentratel %

ment Unlabel Label2 Total CPM Volume CPM3 of total
1 MCUV5 Cl-13 48000 5.0 ml 362 0.7
2 MCUV5 MCUV 41040 2.8 ml 344 0.8
2a 2nd pass of Exp. 2 38173 1.0 ml 105 0.2
3 2nd pass of Exp. 1 38328 5.0 ml 56 0.1
4 MCUV5 MCUV5 108797 3.8 ml 1414 1.2
5 2nd pass of Exp. 4 89215 4.5 ml 405 0.4
6 MCUV5 MCUV5 115680 5.3 ml 890 0.7
7 MCUV5 MCUV5 94780 1.0 ml 1059 1.1
8 MCUV5 MCUV5 147465 6.0 ml 1356 0.9
9 C1-13 CI-13 171792 3.7 ml 2864 1.6
10 C0-13 CI-13 160000 4.0 ml 2112 1.3
11 Cl-13 Cl-13 18754 4.1 ml 369 1.9
12 C0-13 0-13 188370 6.0 ml 1392 0.7
13 Cl-13 CI-13 169640 4.0 ml 1732 1.0
14 Cl-13 C0-13 116200 4.0 ml 1210 1.1
15 C0-13 Cl-13 60760 4.0 ml 840 1.3

1- Concentrated sample after eluting from affinity column and dialysed against PEG 20,000
2- Cells labeled with [35S]methionine
3- Number of counts estimated in total volume determined by counting 10 tl sample and
multiplying by total volume.
























116K-


97.4K-



66K- saw"











1 2 3



Figure 3.2. SDS-PAGE analysis in 9% acrylamide of [35S]methionine
labeled purified gp85 eluted from the E1D1 affinity column (lane 3) com-
pared with the whole lysate passed over the column (lane 1) and flow-
through (lane 2) which did not adsorb to the column.





























Figure 3.3. Densitometric tracing of autoradiograph intensities determined by laser
scanning. Affinity purified gp85 was analysed by SDS-PAGE and detected by auto-
radiograph of radiolabeled proteins. Lane containing purified gp85 was scanned by
laser probe with increased intensity of autoradiograph exposure indicated by density.







Although the molecular weight of the eluted protein was substantially less than 85

kDa, it was within the range normally found for this protein which was confirmed by

immunoprecipitating the protein from the original lysate using the normal method of protein

A-agarose precipitation with E1D1 (Figure 3.4). Purified gp85 was shown to migrate with

similar mobility as the protein immunoprecipitated from the lysate with E1D 1 and protein

A-agarose prior to purification. However, no detectable gp85 was immunoprecipitated

from the flow through of the affinity column by E1D1 even though another antibody D4.16

faintly precipitates gp350/220 from the flow through.


45 K 66 K 97 K 116K


MOLECULAR WEIGHT













II i i l Mii.m


.IOh







48

Protein assays, including the Lowry method { 125), the Bradford method (18) and

BCA assay ({ 189) were never able to detect protein in the purified, concentrated sample.

Even when several batches of purified protein were combined and concentrated further

using a Centricon-10 (Amicon Corp., Danvers, MA), no protein was detectable by any of

these methods.
Polyclonal Rabbit Serum

Following a series of subcutaneous injections of the purified protein into rabbits, an

antibody response to the antigen was achieved. SDS-PAGE analysis of the immunopre-

cipitation using immune serum as well as control antibodies showed the ability of the poly-

clonal rabbit seras to react with gp85 (Figure 3.5) from Cl-13 virus (panel A) and MCUV5

virus (panel B). Rabbit serum from R18 which was immunized with gp85 purified from

MCUV5 virus reacted better with gp85 from MCUV5 than from Cl-13 while R24 reacted

better with the gp85 from the Cl-13 virus from which its antigen was purified. F.2.1 and
E1D1 also immunoprecipitated gp85. The monoclonal antibody E8D2 reacted with p105 as

well as multiple bands of lesser molecular weight in the Cl-13 lysate; p105 has been shown

to produce a number of breakdown products by Dr. Soman in our laboratory possibly
explaining the large number of bands present in materials immunoprecipitated with anti-

body E8D2. The major glycoprotein gp220 and a small amount of gp350 was precipitated

from Cl-13 and gp350 with a minor amount of gp220 from MCUV5 virus with antibody

72A. 1. The species specificity of the high molecular weight gp350/220 correlates with the

amount of each of these two glycoproteins produced in the two cell lines {168). The

amounts of radiolabeled proteins were substantially greater in in the Cl-13 virus than the

MCUV5 which reflects the quantity of virus produced by the two cell lines. A number of

additional polypeptides were also apparent in the precipitations from C1-13 virus. Many of

these are probably breakdown or cleavage products which appear with an efficiently labeled

sample. The glycoprotein gp85 appears to have a major cleavage product (see Chapter 5)




















LOO
~~~ **^^ '"" 0



a I .I


11 9







00
I C
I: U,
IgW
C.,
I .~tr~f
I r 0
Cd N





















con
toI:

(*0

t3 t a m



-4
C14










II
cn8e,



Vt
M]I 2i
Nl 00l
m e o



SMid









which may correspond to a major band in gp85 immunoprecipitated samples in the lower

molecular weight range of 35-40 kDa. One other major band at 66 kDa appeared in the

samples precipitated by the two polyclonal rabbit antisera as well as E1D1 was the subject

of further investigation.

Presence of BSA in Immunoprecipitated Samples

The 66 kDa polypeptide precipitated with E1D1 from 125I labeled virus was electro-
eluted as described previously and compared with 1251-labeled BSA by V8 protease diges-

tion (Figure 3.6). The amount of protein eluted from the 66 kDa protein which E1D1 pre-

cipitated was substantially less than the radiolabeled BSA. Therefore different exposures

of the same gel were used to obtain equivalent photographic intensities. The pattern of

digested peptides resulting from treatment with 10 g.g, 50 gg, and 100 jig of S. aureus V8

protease was similar for both proteins digested.


Discussion
With the monoclonal antibody E1D 1 available in sufficient quantities, we began
purification of gp85 using affinity chromatography. Initially a small column of approxi-

mately 5 mls of coupled sepharose beads were used for extracting the protein from lysates

of virus-producing cells. The method of column chromatography was compared with

batch preparation but we found that better purity of protein was achieved if the protein were

bound and eluted with the affinity matrix packed in a column. Subsequent purifications

were performed using two to four affinity columns in order to increase the amount of pro-

tein purified. Flow-through from a purification was rechromatographed to determine

whether all available gp85 had been bound. No radioactivity was retrieved in this second

run. This was either due to efficient clearance in the first purification or breakdown of the

protein during the storage of the sample between purifications.

The cells which were used for the initial purifications were those of the marmoset
lineage which produced virus. Since we hoped to be able to use the purified protein for









N-terminal sequencing to map the gene coding for gp85, we used this MCUV5 virus

because of its similarity to the B95-8 virus used for genomic sequencing (5). However,

this cell line produces less virus than the human lymphoblastoid cell line Cl-13, so later

purifications were from this virus with presumably greater amount of protein produced.

The failure of protein assays to detect any amounts of protein, even after several

purifications had been pooled and concentrated, indicated that the amounts of this

glycoprotein relative to other viral glycoproteins such as gp350/220 was extremely small.

Although much protein was probably lost at many of the purification steps, the likelihood

that gp85 is only a fraction of the total viral proteins was confirmed by the difficulty in

raising monoclonal antibodies against it. At the time this work was done, although

methods were being introduced for micro-sequencing extremely small amounts of protein

{ 1), we were unable to produce enough purified protein for N-terminal sequencing of the

whole protein. As a membrane protein, the N-terminus would probably have been

blocked requiring even greater amounts of protein to make peptide fragments.

However, undaunted by the failure to measure the protein by standard protein
assays, either because of the limited quantity or interference by detergents, and by silver

staining, we knew that some protein was being eluted in a fairly pure state as determined by

the autoradiographs of eluted [35S]methionine labeled samples which had been elec-

trophoresed. The eluate was therefore injected into rabbits with eventual antibody pro-

duction to gp85. The rabbit which received gp85 purified primarily from MCUV5 cells

had a much greater capacity to immunoprecipitate gp85 from MCUV5 lysates while the

opposite was true of rabbit serum made by inoculation with gp85 purified from Cl-13.
Although some genomic differences exist between the viruses obtained from these two

sources, the membrane antigens produced are essentially the same. The slight specificity

which these rabbit antibodies demonstrated could reflect recognition of minor variations in

glycosylation between the species { 114) or very minor genomic mutations.









The purified gp85 was also injected in BALB/c mice for the production of immune

spleen cells for hybridoma production. The response of the mice to the protein reflected

that either they were not capable of recognizing the protein which was doubtful since they

produced antibodies against it in the native viral form, or the amount of protein which could

be emulsified and injected into a mouse was insufficient for antibody production. Although

multiple injections were given into many mice, the immune sera never reflected positive

production of antibodies, thereby warranting the fusion of immune spleen cells for hybri-

doma production.

The ability of the rabbit antibodies to react with a protein of 66 kDa increased with
continued immunization; therefore it became apparent that the immunizations of purified

gp85 contained this contaminant. Because E1D1 also reacted with this polypeptide it was

not inconceivable that it was being purified by the antibody as well as gp85 even though its

presence was not readily visible in the autoradiographs. The 66 kDa protein was present in

samples precipitated from 1251 extrinsically labeled virus and cells but not [35S]methionine

labeled cells therefore accounting for its absence in the affinity chromatography autoradio-

graphs. The inability to label the 66 kDa protein biosynthetically suggested that its source

was other than viral or cellular, but that it was sufficiently available on the surface of the
viral envelope or cell membrane to be surface labeled with 125I. Considering these charac-

teristics as well as its molecular weight, BSA appeared to be a likely candidate for the con-

taminant. The hypothesis that the 66 kDa protein was indeed BSA was confirmed by

partial proteolytic digestions.

The ability of E1D1 to immunoprecipitate BSA could be explained by the ability of
BSA to non-specifically stick to gp85 in such a manner so as to be coprecipitated by the

monoclonal antibody. However, if the two molecules tend to stick together, it would seem

likely that other monoclonal antibodies to gp85, such as F.2.1 would likewise precipitate

both molecules. E1D1 appears to be unique it its ability to precipitate both gp85 and BSA,

which may indicate that it actually recognizes an epitope which is present on the BSA









molecule. Sequence analysis which was done after the gp85 gene had been determined

suggested that there are actually several sequences in BSA with weak homology to gp85.

Although the virus used for radiolabeling and lysing was removed from serum-containing

media, sufficient quantity of residual BSA could have been present. We have also

experienced BSA as a contaminant in various inoculation preparations using Freund's

adjuvant. Rabbits immunized with a synthetic peptide which had never been in the

presence of BSA produced antibodies which reacted with BSA; therefore, the BSA was

probably inherit in the adjuvant.

Although the purification of gp85 did not result in sufficient amounts of protein for

N-terminal sequencing, the production of polyclonal, monospecific rabbit serum to gp85

provided useful antibodies for additional studies of the glycoprotein.













CHAPTER IV

MAPPING OF gp85 TO THE GENOME

Introduction
Although the entire EBV genome had been sequenced, the gene encoding gp85 had

not been identified. Several groups attempted expressing the two major open reading

frames, BILF2 and BDLF3, which were originally speculated as potential genes for this

protein {5), but those efforts did not produce a protein with the appropriate characteristics.

Since gp85 is not easily purified in a quantity sufficient for sequencing {Chapter 2 and ref.

80) even though microsequencing techniques have improved to allow for sequencing of

extremely small amounts of protein { 1), we looked into alternative methods for identifying

the gene which encodes gp85.

As early as 1965 antibodies to synthetic peptides were shown to react with native
proteins (2); since that time synthetic peptides have been used quite extensively to raise

antibodies to various virus proteins (28, 147, 199). Synthetic peptides are useful, there-

fore, for identifying genes which encode proteins, determining similarity of proteins,

localization of proteins, proteolytic processing, protein purification, and protein function

{46, 216). Although many of the antibodies raised against a protein react with confor-

mational rather than sequential epitopes { 14), it is possible to make anti-peptide antibodies

that react with native protein. Some of the factors which improve the chances for a

sequence to be an available antigenic site include the flexibility of the region surrounding

the chain, hydrophilicity of the amino acids in that chain, the secondary structure main-

tained by the protein, and location of the region in the overall structure of the protein

(216). Although the structure of a protein is rarely known prior to selecting antigenic

determinants, computer programs which analyze the potential structure and chain

55









characteristics facilitate the likelihood of selecting regions which may be of antigenic

importance { 14, 49, 62, 94, 115, 221}.

Another method of determining proteins encoded by selected open reading frames

involves the expression of those genes in prokaryotic and eukaryotic expression systems.

The expressed protein can then be used to produce antibodies which react with a native

protein specified by the virus. Our experience had already involved antibody production

using synthetic peptides; therefore, we used the technology of synthetic peptide immunol-

ogy to facilitate the identification of the gene of the EBV DNA which encodes gp85.


Materials and Methods
Peptide Synthesis

Peptide V-1, cys-ser-leu-glu-arg-glu-asp-arg-asp-ala-trp-his-leu-pro-ala-tyr-lys.,

corresponding to amino acids 518-533 of the BXLF2 open reading frame of EBV DNA,

with an additional cys residue attached to the N-terminus to facilitate coupling and eliminate

coupling of shorter incomplete peptides was synthesized { 132) by the Protein Chemistry

Core Facility at the University of Florida. Two other peptides were used for controls,

peptide E, cys-tyr-ser-arg-pro-leu-val-ser-phe, corresponding to amino acids 551-559 of

the EBV BALF4 open reading frame, and peptide B, cys-leu-ile-asp-asp-val-ser-leu-ile-

arg-gly-gly-thr-arg-lys, corresponding to residues 399-413 of human C9 with an additional

N-terminal cys. Both were synthesized by Biosearch, San Rafael, CA. The purity of the

peptide was checked by Dr. Soman using reverse phase high performance liquid chro-

matography on a C18 column and an acetonitrile gradient

Coupling Procedures

The coupling reagent m-maleimidobenzoyl sulfosuccinimide ester (sulfo-MBS,

Pierce Chemical Co., Rockford, IL){ 19) was used to attach the synthetic peptides, via the

N-terminal cysteine residue, to keyhole limpet hemocyanin (KLH) (69, 123). Briefly,

KLH and sulfo-MBS were dissolved in 10 mM sodium phosphate buffer pH 7.2. Two







57
hundred microliters of KLH at a concentration of 20 mg per ml were reacted for 30 min at

room temperature with 85 1l sulfo-MBS at a concentration of 6 mg per ml. The KLH-MB

was sieved through Sephadex G-25 to remove unreacted sulfo-MBS and collected in a

volume of approximately 5 ml. Five mg peptide in 1 ml PBS containing 2 mM dithio-

threitol were added to the sulfo-MBS, allowed to react for 3 hours at room temperature and

then stored in small volumes at -20C for use.

Antibody Production

Polyclonal antibodies to peptide V-1 (R27) were made in an adult New Zealand

White rabbit (specific pathogen free). The animal was first inoculated subcutaneously with

400 gig of peptide conjugated to KLH and emulsified in Complete Freund's Adjuvant. It

was reinoculated three times at weekly intervals with the same amount of KLH coupled

peptide which was emulsified in Incomplete Freund's Adjuvant. Immunoglobulin G (IgG)

was purified by Protein A-Sepharose affinity chromatography from preimmune and

immune sera as described previously (Chapter 2) and then further purified by affinity

chromatography on peptide V-1 coupled to CNBr-activated Sepharose. Bound IgG was

eluted from the peptide coupled sepharose with 0.1 M acetic acid in 0.15 M NaCl.

Approximately 10% of the immunogobulin reactive with the peptide in ELISA did not bind

to the peptide column.

Enzyme-Linked Immunosorbant Assay

Reactivity of antisera with synthetic peptides was measured in ELISA assays done in
96-well microtiter plates {212). Sixty micrograms of peptide in carbonate-bicarbonate

buffer, pH 9.6, were added to each well and dried overnight at 37'C. Plates were then

incubated sequentially with 5% skimmed milk in PBS-Tween 20, dilutions of rabbit or

mouse antibody, goat anti-rabbit or anti-mouse antibody conjugated to horse radish peroxi-

dase (Organon Teknika Corp., Durham, N.C.) and the substrate orthophenylenediamine

(Sigma). Plates were washed between each incubation with 0.05% Tween 20 in 0.85%









NaCI; the final enzyme reaction was stopped with 2 N H2SO4 and analyzed for color

change at 492 nm.

Western Blotting

Radiolabeling, immunoprecipitations, and proteolytic digestions were carried out as
described previously (Chapter 2). For Western Blots the immunoprecipitated proteins were

electrophoresed in 9% acrylamide cross-linked with DATD and then electrotransferred onto

nitrocellulose membranes (0.45 gm pore size; Schleicher and Schuell, Inc., Keene,

NH){211). The transferred sheets were treated for 3 hours with blocking buffer (10 mM

Tris hydrochloride, pH 7.2, 0.15 M NaCI, 5% skimmed milk, and 0.05% sodium azide)

and reacted for 3 hours with blocking buffer containing rabbit antibody at 400 gg/ml. They

were then washed five to six times with wash buffer (10 mM Tris hydrochloride, pH 7.2,

0.15 M NaCI, 0.3% Tween 20) for 10 min each, followed by an overnight wash. The

washed sheets were reacted with alkaline phosphatase-conjugated goat anti-rabbit anti-

bodies (Hyclone, Logan, UT) for 3 hours and the bound anti-rabbit antibodies detected by

reacting with substrate (5-bromo-4-chloro-3-indophosphate and nitroblue tetrazolium;
Sigma).


Results
Selection of V-1 Peptide Sequence

The original analysis of the hypothetical proteins encoded by the EBV DNA sug-
gested that the BDLF3 open reading frame was a potential candidate for the gp85 gene (5).

This open reading frame is predicted to encode a protein of approximately 24 kDa with nine

potential N-linked glycosylation sites. Although extensive glycosylation of a smaller back-
bone polypeptide is seen with gp350/220, the previous work characterizing gp85 had sug-

gested that the protein stripped of its N-linked sugars was approximately 70 kDa. There-

fore, additional computer analysis (112) done by Dr. Hutt-Fletcher of the potential










1 20 40
MOLLCVFCLVLLWEVGAASLSEVKLHLDIEGHASHYTIPWTELMAKVPGL

60 80 100
SPEALWREANVTEDLASMLNRYKLIYKTSGTLGIALAEPVDIPAVSEGSM

120 140
QVDASKVHPGVISGLNSPACMLSAPLEKQLFYYIGTMLPNTRPHSYVFYQ

160 180 200
LRCHLSYVALLSINGDKFQYTGAMTSKFLMGTYKRVTEKGDEHVLSLVFGK

220 240
TKDLPDLRGPFSYPSLTSAQSGDYSLVIVTTFVHYANFHNYFVPNLKDMP

260 280 300
SRAVTMTAASYARYVLQKLVLLEMKGGCREPELDTETLTTMFEVSVAFFK

320 340
VGHAVGETGNGCVDLRWLAKSFFELTVLKDIIGICYGATVKGMQSYGLER

360 380 400
LAAMLMATVKMEELGHLTTEKQEYALRLATVGYPKAGVYSGLIGGATSVL

420 440
LSAYNRHPLFQPLHTVMRETLFIGSHVVLRELRLNVTTQGPNLALYQLLS
I
460 480 500
TALCSALEIGEVLRGLALGTESGLFSPCYLSLRFDLTRDKLLSMAPQEAT

520 V-1 540
LDQAAVSNAVDGFLGRI SLEREDRDAWHLPAY VDRLDKVLMTIPLINV

560 580 600
TIISSDREVRGSALYEASTTYLSSSLFLSPVIMNKCSQGAVAGEPRQIP

620 640
KIQNFTRTQKSCIFCGFALLSYDEKEGLETITYITSQEVQNSILSSNYFD

660 680 700
FDNLHVHYLLLTTNGTVMEIAGLYEERAHVVLATILYFIAFALGIFLVHK


IVMFFL
Figure 4.1. Predicted amino acid sequence of protein potentially coded by
BXLF2 open reading frame. Solid underlines (- ) indicate hydrophobic
regions predicted by the program of Klein and colleagues to have a high
probability of being membrane spanning regions; I indicate potential
N-glycosylation sites; boxed area ( ) indicate residues synthesized.









membrane associated proteins indicated that an open reading frame, BXLF2 which poten-

tially encodes a polypeptide of 78 kDa with five potential N-linked glycosylation sites {5)

appeared to be a more likely candidate for the gene of gp85 (Figure 4.1). Computer analy-

sis for hydrophobic regions of the protein (112) indicated that three stretches had a high

probability of being membrane spanning segments. One of these regions is located at the

amino terminus from amino acids three to nineteen and may represent a signal sequence. A

second region is at the carboxy terminus, stretching from amino acids 680-696 and may

anchor the protein in the membrane; the third spans a 16 amino acid region from amino

acids 538-554 and may either be a second anchor sequence or perhaps a hydrophobic

region similar to that found in fusion proteins of ortho and paramyxoviruses (223). Since

other information concurrently being gathered in the laboratory indicated that gp85 may

play a role in penetration of the virus post-binding, the possibility that this region could be

involved in fusion made this gene additionally attractive as a candidate for gp85.

Selection of an amino acid sequence within this reading frame for synthesis was

made after computer analysis (151) to predict a composite surface profile indicating

regions with good potential for inducing antibodies that recognize linear epitopes on the

native protein (Figure 4.2). Residues with a composite profile values of greater than 50%

are those considered to be most likely to include surface accessible residues. Of the seven

stretches of the BXLF2 sequence that had surface profile values of greater than 90%, the

sequence corresponding to amino acids 518-528, which is adjacent to the internal hydro-

phobic sequence, was first chosen for synthesis. Since the likely secondary structure of

the sequence 530-533 was predicted by the algorithm of Gamier and colleagues {62) to be

a turn, peptide V-1 was extended to include amino acids 518-533. A cysteine residue was

added to the N-terminus of the peptide to facilitate coupling to carrier protein. The V-1

sequence is not duplicated in EBV DNA. The peptide was provided in pure form as indi-

cated from reverse phase HPLC (Figure 4.3).








































3mVA 33VjflS


. .= = .
- aI
______________________
8
T m 3jn


Fn
10,





-4W
o .
















* 4)

lii
itEi










18.75


[ b -bA 20 Time (min) 25 0 37 1

Figure 4.3. HPLC profile of 20 gg V-1 peptide loaded on a C18 column and run at
0.5 ml/min with a 10 to 100% acetonitrile gradient. Peptide was removed with 80%
acetonitrile.


Production of Anti-V-1 Antibodies

Successful immunization of the rabbit with peptide V-1 was first determined by the

ability of the immunoglobulin purified from preimmune and immune sera to react with

peptide V-1 and not with unrelated EBV specified peptide D, derived from the BALF4 (gB)

open reading frame, and peptide B, derived from the human C9 sequence (Figure 4.4).

Antibody against V-1 was purified by affinity chromatography on protein-A agarose fol-

lowed by purification on immobilized peptide. The reactivity of the purified antibody was

compared to that of the two gp85 monoclonal antibodies, F.2.1 and E1D1 and with the

rabbit anti-gp85, R24, and rabbit anti-peptide E, R16. Neither of the two monoclonal anti-

bodies reacted with any peptides and the rabbit anti-peptide E failed to react with V-1 or B

but reacted strongly with peptide E. The monospecific rabbit antibody, R24, reacted

weakly, but specifically with peptide V-1 but not with either of the other two peptides. The

preimmune sera from these rabbits failed to react at similar concentrations of immuno-

globulin with any of the peptides.













- repdie v-


0.3
0.0. -

3.0..
2.7. Peptide D
2.4..
2.1.
1.8.
1.5.
1.2
0.9 _
0.6
0.3_
0.0 |-4- I I "-
.0.0 li


Peptide B


U.Un1


5000


2500


1250


625
Antibody (ng)


1
Figure 4.4. Reactivity in ELISA of monoclonal antibodies and purified rabbit antibodies
with peptides V-l, D,and B. Symbols: R24 (rabbit anti-gp85), R27 (rabbit
anti-V-1), *R16(rabbitanti-B), a F.2.1 E1D1.


'lr o A I -* F


L









Reactivity in Western Blots with gp85

The weak reactivity of R24 with peptide V-1 suggested that this sequence might be

present in gp85. We have not found the gp85 protein to be detectable in Western blots of

electrophoresed lysates of virus, presumably because of its lack of abundance. However,

reactivity with rabbit antibody has been observed if the protein is enriched by immunopre-

cipitation with F.2.1 before transfer. We therefore immunoprecipitated gp85 from CI-13

virus with F.2.1, electrophoresed and transferred it to nitrocellulose filters and examined

whether purified R27 immune sera or preimmune sera reacted with the transferred protein

(Figure 4.5). The immune sera but not the preimmune sera reacted strongly with a

molecule of approximately 82 kDa bound to the nitrocellulose. The mobility of this protein

was consistent with that of gp85 in DATD cross-linked polyacrylamide gels.

Immunoprecipitation by Anti-V-1

Immunoprecipitation of 1251 labeled virus or TPA-induced Cl-13 cells labeled with
[3H]glucosamine (Figure 4.6) also indicated that anti-V-1 recognized a glycoprotein with a

mobility in gels similar to that of gp85 immunoprecipitated with either F.2.1 or the

monospecific rabbit antibody R24. Although a number of bands were present in the glu-

cosamine labeled samples, the reaction with a protein with electrophoretic mobility similar

to gp85 was quite specific.
Proteolytic Digestions

The similarity of the proteins immunoprecipitated by F.2.1 and R27 was further

examined by comparing V8 protease digestion patterns of the two. The proteins were

immunoprecipitated and eluted as described previously and digested with varying amounts

of V8 protease. The digestion products were analyzed by SDS-PAGE (Figure 4.7).

Comparison of the digestion patterns indicated that the two proteins were identical.















1 2


116k-


84k-

58k-


48k-



36k-

26k-


Figure 4.5. Reactivity of preimmune antibody (lane 1) and anti V-1 (lane
2) with gp85 immunoprecipitated by monoclonal antibody F.2.1 from Cl-
13 cell-grown virus, separated by SDS-PAGE and blotted onto nitro-
cellulose. The arrow indicates the band recognized by anti-V-1.















cI -0 N (D I 0



I

' I I I


I


.-. o ,' '






67








= L


LOO





.co
igl
(0 to cc' t 0 >
(0 t C)N N N -S
S10 I I^ ^ ^^ we6 .1 Oi *I
In ^^SB^fj ^7

*4 -V.:earn 0
^^^^T^ ^0
.M 4

--,^
m -a~










Discussion

The data presented in this chapter indicate that antibodies made to a synthetic peptide

derived from the BXLF2 open reading frame recognize the same glycoprotein as mono-

clonal antibody F.2.1 and are thus consistent with the hypothesis that BXLF2 encodes the

gene for gp85 { 148). The use of synthetic peptides derived from known gene sequences

is an indirect method for protein mapping, but is nevertheless a useful alternative approach

to gene expression that has been extremely useful for isolation and characterization of gene

products (222). These data confirm that synthetic peptides can be used for the production
of antibodies which react with native protein. A careful analysis of the primary amino acid

sequence probably increases the likelihood of selecting a region which will do this. The
sequence represented by the V-1 peptide encompassed less than 3% of the BXLF2 open

reading frame, so the fact that the monospecific rabbit antibody, made against the entire

native gp85 glycoprotein, reacted only weakly with the peptide was not surprising.

The results of this work have recently been independently confirmed by Heineman

and colleagues {80) through the expression of a fusion protein encoded by the BXLF2

gene. Antibodies to the protein express in bacteria reacted with gp85. Also, F.2.1 was

capable of immunoprecipitating the fusion protein which had been radiolabeled and ana-

lyzed by SDS-PAGE.

Several of the EBV genes have been shown to have counterparts in other her-

pesviruses (39, 68, 159) and the BXLF2 sequence has been reported to include stretches

that have weak homology with genes coding for HSV-1 glycoprotein gH {131) and
glycoprotein II of varicella-zoster virus (70, 106). Preliminary data (Balachandran, Oba

and Hutt-Fletcher) reveal that the monospecific rabbit anti-gp85 used in this study reacts
with gH of HSV-1, indicating that the deduced homology is mirrored by shared anti-

genicity. Both gpIll and gH are important to spread of virus in tissue culture (20, 106),

and HSV gH is essential for efficient egress of virus from infected cells (67). The







69
importance of gp85 to virus infectivity is demonstrated by the fact that antibody to it can

neutralize infectivity in a complement-independent manner. Antibody F.2.1 does not

inhibit virus binding, but does inhibit penetration of virus through the cell membrane

{136). Therefore it is likely that this glycoprotein performs an essential function in entry

of the virus and perhaps, like its human herpesvirus counterparts, is involved in egress of

the virus.













CHAPTER V

ANTIBODIES TO gp85

Introduction
Because of the specific nature with which antibodies bind to antigens, their useful-

ness in studying proteins has been invaluable. Both polyclonal antibodies and monoclonal

antibodies have significant use for protein studies. The production of monoclonal anti-

bodies, as originally described by Kohler and Milstein {113), has greatly improved the

utility of antibodies by providing a means of obtaining an unlimited quantity of antibody

with specificity to a single epitope or at least a very limited number of cross reactive epi-

topes. These monoclonal antibodies allow entire proteins to be studied by small segments,

thereby identifying the essential regions and those which are of lesser importance. If one

can provide an antigen, then it is usually feasible to produce an immune response in mice,

hamsters, or rats which can then be used for fusing with myeloma cells to make hybrido-

mas producing monoclonal antibodies. Very often the limiting factor in this procedure is

having sufficient quantities of antigen to be used for immunization. Other difficulties

include the ability of testing for the presence of antibodies produced by the hybridomas,

which very often parallels the availability of antigen used for immunization. Recent

advances in understanding the immunology of presentation of antigen associated with

major histocompatibility (MHC) class II antigens and the use of synthetic peptides to
enhance antigenic selection are improving the ability to make monoclonal antibodies which

recognize desired proteins. Although producing monoclonal antibodies can be a time-con-

suming, tedious process their continued supply provides the tools necessary for a wide

variety of studies. However, since a monoclonal antibody is usually restricted to a single

binding site on a protein, very often its use is limited when conditions alter the form or
70







71
nature of its ligand. Therefore, it is useful to produce polyclonal antibodies, particularly if

monospecificity can be ensured.

Polyclonal antibodies, usually produced in rabbits, provide large, but not unlimited

supply of antibodies useful for multiple purposes. Production of polyclonal antibodies in

rabbits can be a relatively quick procedure, producing high titers of antibody in less than a

month, which can then render valuable information quite rapidly. When immunizing with

antigens of small quantity the time required for good response is increased but ultimately a

response is usually achieved. Polyclonal antibodies also recognize a wide spectrum of

antigenic determinants, decreasing the unique specificity which they recognize, but

increasing the potential that they may recognize altered or non-native epitopes (224).

The major reason for the production of antibodies in this work was to provide the

tools necessary for studying the biologically important functional epitopes on gp85. Stan-

dard procedures originally employed eventually evolved into more unique methods in order

to increase the chances of making antibodies which could recognize the protein. A variety

of antibodies, both monoclonal and polyclonal have been produced which react with the

native protein, with synthetic peptides, have neutralizing capability and demonstrate

involved immunologic processes of antigen recognition. Although these antibodies are

actually only the beginning of further studies, their production itself gave rise to new

insights about gp85 and its relation to the immunologic process.

Materials and Methods

Mouse Immunizations

Immunizations of all antigens were done in BALB/c mice. The antigen to be inocu-

lated was emulsifed in an equal volume of Complete Freund's adjuvant so that final volume

injected subcutaneously was 0.5 ml/mice. Subsequent injections were emulsified in

Incomplete Freund's adjuvant. Emulsification was done by drawing the total volume of

adjuvant into a syringe. Approximately 100-200 pl of the antigen in PBS was drawn up









into the syringe and vortexed for several minutes and repeated until all antigen was drawn

up. The mixture was vortexed in the syringe until complete emulsification as determined

by non-separation of an emulsification drop on water. Mice were inoculated at 1-2 week

intervals by dorsal subcutaneous injection. After three injections the serum from a tail vein

bleed was tested for reactivity with antigen or whole virus by ELISA or surface fluores-

cence. After high titers of antibodies were achieved, or after multiple injections if all

response was continually negative, the mice were boosted intraperitoneally (IP) with 0.5 ml

of antigen in PBS.

The immunization protocols varied for different types of response being sought. For

the production of cross-reactive antibodies, the mice were injected subcutaneously three

times with 5 x 106 virus-producing Cl-13 cells. Following a three month rest the mice

were boosted IP with 0.5 ml HSV-1 (KOS strain) which had been grown to a maximum

cytopathic effect and UV-irradiated after which the mice were used for fusion three days

later. Immunization of mice with virus for the production of a variety of anti-EBV anti-

bodies was done as described above using MCUV5 virus concentrated 250-fold resus-

pended in PBS and emulsified with adjuvant. A final challenge of virus alone was given IP

three days before fusing. Peptides were immunized with two methods in order to get anti-

bodies against the peptide as well as antibodies against whole protein. Peptide immu-

nization consisted of a series of injections of peptide coupled to KLH as described in

Chapter 4, followed by a final boost IP with 100 pgg peptide alone three days prior to

fusion. Peptide priming was achieved by immunizing the mice with 100 tg of peptide dis-

solved in PBS and emulsified with adjuvant. After four or five injections of the peptide,

the mice were innoculated with 0.25 ml of MCUV5 virus emulsified in adjuvant Five

days later the mice were challenged IP with 0.5 ml of the virus and used for fusion after

three days.









Fusion

Hybridomas were made by fusing mice spleen cells with the myeloma cell line SP2/0

Ag 14. After the appropriate immunization protocol, and preferably when antibody titers

were high, a mouse was anesthetized with Metofane and bled by cardiac puncture using a

23-guage needle. Exsanguination euthanized the mouse after which the spleen was care-

fully removed and placed in sterile Dulbecco's modified essential medium supplemented

with non-essential amino acids, sodium pyruvate and 0.1M Hepes (SDMEM) and 20%

fetal calf serum (FCS) (Gibco). The spleen cells were released from the capsule by

grinding the cut up spleen between two frosted-end slides after which the red blood cells

were lysed with ammonium chloride followed by two washes of the spleen cells in

SDMEM. Hyperimmune spleens yielded more than 108 cells. When fewer cells were

present, the fusions were not very successful. The spleen cells were combined with SP2/0

myeloma cells at a ratio of five spleen cells to one SP2/0 cell, washed in serum-free

SDMEM and fused together with 50% polyethylene glycol 1500 for one minute. Serum

free media was slowly added and the fused cells were then washed to remove PEG and

plated into 24-well tissue culture plates at 1 x 106 cells in 300 pl SDMEM with 20% FCS

per well. After 24 hours, 100 pl of SDMEM containing four times the normal con-

centration of hypoxanthine, thymidine and aminopterine (HAT) was added to each well.

Cells were fed when media was depleted and outgrowths were screened for the production

of antibody and transferred into petri dishes for further growth.
Hybridoma Screening

Culture supernatant from hybridomas were screened by indirect immunofluorescence

of EBV-producing cells as described previously. Various patterns of fluorescence were

identified and helpful in determining the nature of antibody present; gp85 antibody very

often emitted a very weak speckled fluorescence which could easily be overlooked in com-

parison to the more pronounced gp350/220 staining. Antibodies were selected which

reacted with a similar number of cells as control anti-viral antibodies. After initial screening









with Cl-13 cells, the positive cultures were screened again using both the TPA-induced

Cl-13 cells and EBV membrane protein-negative Raji cells.

Hybridomas which were making antibodies against peptides were screened by

ELISA. The ELISA was done as described previously using 50 Wl of the supernatant from

the hybridoma culture. Reactions to the specific peptide were compared with response to a
non-related peptide and those testing positive were grown up for further testing by surface

fluorescence and immunoprecipitation. Antibodies responding to the peptide priming
protocol were initially screened using surface fluorescence as well as ELISA. Antibodies

from the cross-reactive fusion were screened by ELISA against HSV-1 viral proteins (7)

and Bovine hamster kidney (BHK) virus-negative cell proteins diluted 1/20 from a 2 mg/ml

antigen solution and plated at 50 pl/well as well as by surface fluorescence to EBV mem-

brane antigens. Those cultures which were positive for HSV-1 or EBV were selected for

additional growth.
Additional screening of all positive cultures was done following the freezing of the

cells. Supernatant removed when the cells were frozen was concentrated 5-10 times by

dialysis against PEG 20,000 and then used for immunoprecipitating virus lysed in

non-SDS containing lysing buffer and virus lysed in RIPA. The hybridomas were mono-
cloned by limiting dilution and high titered ascitic fluid obtained by injecting 5 x 106 cells

into Pristane primed mice. The monoclonal antibodies were then purified by protein-A

chromatography as described previously.
Synthetic Peptides V-2 and V-3

Two additional peptides from the BXLF2 open reading frame were selected for the
production of monoclonal and polyclonal antibodies (Figure 5.3). Both peptides were
synthesized by the Protein Chemistry Facility at the University of Florida as was peptide

V-1. Peptide V-2, cys-thr-met-leu-pro-asn-thr-arg-pro-his-ser-tyr, is an 11 amino acid
sequence corresponding to residues 138-147 of the open reading frame, and peptide V-3,
cys-gly-thr-try-lys-arg-val-thr-glu-lys-gly-asp-glu, is a 12 amino acid sequence









corresponding to residues 180-191 of the same gene. Homogeneity of peptide V-2 was
done by Dr. Soman using HPLC with a C18 column and 10-100% acetonitrile gradient; and

HPLC of the V-3 peptide was done by Mike Jarpe using a 0-40% acetonitrile gradient.

Both peptides were coupled to KLH as described previously for immunization. Peptide V-

3 required dissolving in 10% DMSO in one-tenth the final volume of PBS. The final con-

centration of DMSO was 1%. Three weekly injections of the peptides conjugated to KLH
were given to R33 (peptide V-2) and R32 (peptide V-3), after which the immune serum

was obtained and purified as described previously using respective homologous

immobilized peptide columns.

Neutralization

Neutralization of virus infectivity by antibody was expressed as the ability to inhibit

virus induced immunoglobulin synthesis by fresh human B cells. Leukocytes were

obtained from heparinized human peripheral blood by flotation on Ficoll-Hypaque and

depleted of T cells by a double cycle of resetting with 2-aminoethylthioisouronium treated
sheep erythrocytes (155) and centrifugation over 60% Percoll (Pharmacia Fine Chemicals,

Piscataway, NJ){98). A total of 200,000 cells were incubated with or without virus that

had been preincubated with an equal volume of preimmune rabbit antibody and test anti-
body adjusted so that the total amount of immunoglobulin remained constant at 100 gg per

ml; all antibodies were heated at 56'C for 30 min to inactivate complement. Cells and virus

were plated in a total volume of 100 gl in 96-well round-bottomed tissue culture plates.

After 6 days in culture, 100 p1 of medium was added to each well. On day 12, the

immunoglobulin concentrations in the media were measured by a double-sandwich micro-
ELISA (212) with appropriate concentrations of rabbit anti-human immunoglobulin,
peroxidase-conjugated rabbit anti-human immunoglobulin, and the substrate hydrogen
peroxide with 5-aminosalicylic acid.









Results
Monoclonal Antibodies

A variety of immunization protocols were followed in order to obtain monoclonal

antibodies to gp85. Early attempts at raising antibodies in mice included injection of

homogenized gel slices containing the 85 kDa region and immunization with gp85 purified

by affinity chromatography (see Chapter 3). However, we were unable to achieve a

hyperimmune response with either of these methods, most likely because of the lack of

abundance of this glycoprotein. The spleens from these mice contained only the normal

number of cells and fusing these cells resulted in a poor percentage of outgrowths with no

positive hybridomas. Therefore we endeavored to use novel methods for immunization in

order to enhance the ability of the immune system to recognize the non-abundant gp85.

Since several of the glycoproteins seemed to have cross-reactive epitopes among the human

herpesviruses, one of the earlier methods employed was that of cross-reactive immu-

nization. This method of challenging mice with HSV- 1 to enhance a secondary immune

response in mice previously immunized with EBV and rested for several months, gave us a

number of monoclonal antibodies which responded to both EBV and HSV. Antibodies of

particular interest were those characteristic of a secondary immune response rather than the

IgM antibodies produced in primary immunity. Therefore, the production of antibodies

(Table 5.1) which were of the IgG class indicated that cross-immunization with different

viruses was capable of activating memory cells which recognized antigenic epitopes from

both viruses. Although none of the monoclonal antibodies produced by this method

reacted with gp85, a very antigenically cross-reactive protein, gB {8,22,176) was the tar-

get of several of these antibodies {Figure 5.1).

Earlier monoclonal antibodies produced in our laboratory were made against the Cl-

13 strain of the virus. After multiple attempts to raise monoclonal antibodies to purified

protein, we decided to try whole virus immunization again with the idea of producing

monoclonal antibodies to a variety of EBV proteins. Immunization with virus concentrated







77

Table 5.1. Summary of hybridomas with positive reaction to EBV or HSV-1 proteins.

Surface Fluorescence

Name ELISAI C1-13 Raji Kos BHK Protein2 Isotype3


X1B6 0.166 ++++ gB IgG3
X2C3 0.406 + IgGza
X4A5 0.164 ++++ IgG2b
X4B5 0.369 ++ IgM
X5A5 0.222 IgM
X7B3 0.430 ++ IgG2b
X8C4 1.344 ++++ gB IgG2a
X9D5 0.379 +++ *
SP2/0 0.050 -
1- Absorbance values at 492 nm detecting antibodies bound to HSV-1 (KOS strain) antigen
bound to plates
2- Protein which was detected by immunoprecipitation from HSV and EBV lysates.
3- Class of antibody produced by clone as detected by double immunodiffusion in Ouchterlony
plates
No detectable protein


from MCUV5 cell cultures resulted in hyperimmune mice which produced a very high per-

centage outgrowth fusion. Screening by surface fluorescence was difficult because the

large number of positive wells, but a number of hybridomas (C clones) were grown which

reacted with EBV (Table 5.2). Immunoprecipitation with concentrated culture supernatant

from these hybridomas showed that a variety of proteins was being recognized by these

antibodies, including gp350/220, p105, gp85, and additional polypeptides of 170 kDa, 120

kDa and a number of lower molecular weight proteins which were either breakdown

products, precursors, cleavage products, previously undescribed proteins or a combination

of these (Figure 5.2). Four hybridomas appeared to react with a protein of 85 kDa which

migrated similarly as gp85, but two of these cell lines were lost due to a bad batch of

serum.













1 2 3


4 5


T


-205k


-116 k

-97.4k


-66k


-45k


Figure 5.1. SDS-PAGE in 9% acrylamide of proteins immunoprecipi-
tated from [35S]methionine labeled BHK cells (lanes 1) and HSV-1
infected BHK cells (lanes 2 and 4) and HSV-2 infected BHK cells (lanes
3 and 5) with the monoclonal antibodies X1B6 (lanes 1-3) and B3, anti-
gB monoclonal antibody (lanes 4 and 5).


4ra
4r
Now







79

Table 5.2. Summary of hybridomas produced against EBV proteins by immunization
with MCUV5 virus.

Immunofluorescence Protein Molecular weight (kDa)3
Name CI-13 Rajil Cl-13 Raji2 C1-13 MCUV5
C1A2 ++ -
C1B1 + + 66 66
C1D1 + ++ 85 170
C2B3 + + + 85
C2B4 + + 140
C2D4 + + ++ -
C3A3 ++ + +++ 220 95-140
C3B1 + + + 85,110 -
C3B3 + + + + 85,120,200
C3B6 + + + 116,125
C3C4 + + + + 85
C3C5 + + 140 85,140
C3D4 ++ + +++ 220 350
C3D6 + -
C4A1 + + + -
C7A1 + + + 28, 105
C7A3 + + 350
C7B5 + + 29
C7C2 +++ -
C10D4 + + 66-85

1- Cell surface attachment of early culture supernatant antibodies detected by fluorescein conju-
gated anti-mouse
2- Similar surface fluorescence using 10-fold concentrated culture supernatant
3- Molecular weight of 1251 labeled proteins




















ik


CO



I1
c D

Ib

CO


V 1 M
I CM


9


-ja.






in
S 30







ec




M,
2OG






Hil


a


b


I


-I


qD


I (
o> <











1 20 40
MOLLCVFCLVLLWEVGAASLSEVKLHLDIEGHASHYTIPWTELMAKVPGL

60 80 100
SPEALWREANVTEDLASMLNRYKLIYKTSGTLGIALAEPVDIPAVSEGSM
I
120 140 V-2
QVDASKVHPGVISGLNSPACMLSAPLEKQLFYYTMLPNTRPHSYVFYQ

160 180 V-3 200
LRCHLSYVALLSINGDKFQYTGAMTSKFLM3TYKR VTEKGDtVLSLVFGK

220 240
TKDLPDLRGPFSYPSLTSAQSGDYSLVIVTTFVHYANFHNYFVPNLKDMP

260 280 300
SRAVTMTAASYARYVLQKLVLLEMKGGCREPELDTETLTrMFEVSVAFFK

320 340
VGHAVGETGNGCVDLRWLAKSFFELTVLKDIIGICYGATVKGMQSYGLER

360 380 400
LAAMLMATVKMEELGHLTTEKQEYALRLATVGYPKAGVYSGLIGGATSVL

420 440
LSAYNRHPLFQPLHTVMRETLFIGSHVVLRELRLNVTTQGPNLALYQLLS

460 480 500
TALCSALEIGEVLRGLALGTESGLFSPCYLSLRFDLTRDKLLSMAPQEAT

520 V-1 540
LDQAAVSNAVDGFLGRSLEREDRDAWHLPAYCVDRLDKVLMIIPLINV
-Ig-
560 580 600
TFIISSDREVRGSALYEASTTYLSSSLFLSPVIMNKCSQGAVAGEPRQIP

620 640
KIQNFTRTQKSCIFCGFALLSYDEKEGLETTTYITSQEVQNSILSSNYFD

660 680 700
FDNLHVHYLLLITNGTVMEIAGLYEERAHVVLAIILYFIAFALGIFLVHK


IVMFFL
Figure 5.3. Predicted amino acid sequence of protein potentially coded by
BXLF2 open reading frame. Solid underlines (- ) indicate hydrophobic
regions predicted by the program of Klein and colleagues to have a high
probability of being membrane spanning regions; 0 indicate potential
N-glycosylation sites; boxed area (r- ) indicate residues synthesized.











9.47
















0 5 10 15 20 25 30 35 40
Time (min)

Figure 5.4. HPLC profile of 20 gg V-2 peptide loaded on a Czg column and run at
0.5 ml/min with a 10 to 100% acetonitrile gradient. Peptide was removed with 40%
acetonitrile.


Peptide Antibodies

Monoclonal antibodies were also produced using peptide V-1 as either the sole

immunogen or to prime the immune response for later immunization with whole virus.

Two monoclonal antibodies were made against peptide V-1 which reacted specifically with

the peptide in ELISA, only one of which was capable of immunoprecipitating the native

protein from lysed virus. The same peptide was used for priming mice which were then

boosted with whole virus. Several hybridomas were produced by this method which

reacted with the virus producing cells in surface fluorescence, but only one, VM4C3, could

be shown to react with the native gp85 molecule (Figure 5.7). This antibody had a

fluorescent staining pattern typical of gp85 and immunoprecipitated gp85: however, the

antibody did not react with thbodies werptide V- in ELISAs. Therefore, although this anybody
antibody did not react with the peptide V-i in ELISAs. Therefore, although this antibody










was made by successive immunization with uncoupled peptide, a method which failed

repeatedly to produce an immune response, boosting with the virus allowed the primed

immune system to recognize a different epitope on the gp85 molecule. The fact that one out

of five hybridomas produced by this method (most of the fused hybridoma cells failed to

grow due to the same batch of bad serum) reacted with gp85 (a high percentage based on

previous experiments) suggesting that the cells had been primed for recognizing gp85.

Having mapped the gene for gp85, it became immediately obvious that the pro-

duction of monoclonal antibodies to gp85 should be much easier in the future. While

waiting to obtain the cloned fusion protein expressing gp85 in order to get large quantities

of the protein for immunization, we had two additional peptides synthesized. The HPLC


I 1 3.08


0

0
.0


I I I I
5 10 15 20
Time (min)


Figure 5.5. HPLC profile of peptide V-3 run on a Cis column with
a 0-40% acetonitrile gradient.












,! 1~n6' I


3.0-
2.7-
2.4-
2.1-
1.8-
1.5-
1.2-
0.9-
0.6-
0.3-
0.0-


V-3 Peptide


IT ------


V-1 Peptide


0.01

500 50 5 0.5
Antibody (ng/well)
Figure 5.6. Reactivity in ELISA of purified rabbit antibodies with peptides V-1, V-2,
and V-3. Symbols: A R24 (rabbit anti-gp85), 0 R27 (rabbit anti-V-i), R32
(rabbit anti-V-3), R33 (rabbit anti-V-2).


I- -j


















3 4 5 6 7


205k-





116k-


p

: .


97.4k-


66k-


45k-


Figure 5.7. SDS-PAGE in 9% acrylamide of proteins immunoprecipi-
tated from 125I-labeled CI-13 virus with the following antibodies:
VM4C3 (lane 1), F.2.1 (lane 2), EID1 (lane 3), R33 (lane 4), R32 (lane
5), R27 (lane 6) and preimmune rabbit serum (lane 7).


401*









profile of these two peptides indicates their relative purity (Figure 5.4 and Figure 5.5).

These two peptides were coupled to KLH and injected into rabbits. After three weekly

injections the rabbits were bled and anti-peptide antibodies purified by chromatography

with protein-A agarose and immobilized peptide. The specificity of each antiserum for its

appropriate peptide was demonstrated by ELISA (Figure 5.6). The polyclonal rabbit anti-

gp85 which reacted weakly with peptide V-1 did not react with peptides V-2 and V-3.

Somewhat of a surprising result was the proteins which these two antisera immuno-

precipitated (Figure 5.7). The anti-V-2 antibodies reacted with a polypeptide of approxi-

mately 35 kDa which was labeled quite strongly and another protein of approximately 40

kDa. It is feasible that this anti-peptide antibody reacts only with a cleavage product of

gp85 rather than with the mature molecule. The reactivity of the anti-V-3 antibodies with a

polypeptide of approximately 170 kDa without any specific reaction with the native gp85

was equally perplexing. Both of these two rabbit sera react with a faint positive speckled

fluorescence on TPA induced Cl-13 cells which contrasts with the negative results using

anti-V- antisera.

All antibodies to gp85 which have been produced thus far, both in these efforts and

by others are compared to show their specificity with virus strains, membrane antigens in

surface fluorescence and the three peptide regions, and their neutralizing ability if tested

(Table 5.3).

Neutralization

Of primary interest in studying gp85 with monoclonal and polyclonal antibodies are

the ability of these antibodies to block infectivity. In the many assays performed by Susan
Turk, F.2.1 was shown to be an efficient neutralizing antibody. However, ElD1 neu-

tralizes only weakly, if at all as measured by the ability of infective virus to promote
immunoglobulin synthesis in fresh T-depleted B cells. The rabbit anti-V-1 immunoglobu-

lin failed to neutralize virus, whereas the antibodies to peptide V-3 and peptide V-2

appeared to have a slightly higher degree of immunoglobulin synthesis inhibition. The









peptide primed monoclonal antibody VM4C3 also showed a weak inhibition of

immunoglobulin synthesis consistent in several assays. The ability of these antibodies to

partially neutralize at higher concentrations with a non-linear decrease in inhibition at lower

concentrations may indicate that the bound antibodies react with a site adjacent to, but not

identical with, epitopes crucial to the infectivity.


Table 5.3. Summary of antibodies produced which react with gp85 or potentially
related protein with specificity of reaction by immunoprecipitation, surface fluores-
cence, ELISA, and neutralization.


Antibody Immunogen MCUV51 Cl-131 SF2 Peptide3 Neutralize

(kDa) (kDa)
Monoclonal
F.2.1 B95-8 cells 85 85 + +
E1D1 Cl-13 cells 85 85 + -
C3C5 MCUV5 virus 85 85 + ND
VM4C3 V-1, Virus 85 85 + +?
VK1B3 V-1, V-1KLH + V-1 ND
VK1B6 V-1, V-1KLH 85 85 V-1 ND
C2B3 MCUV5 virus 85 85 + ND ND
C1D1 MCUV5 virus 170? 170? + ND ND

Polyclonal
R18 gp85 (MCUV5) 85 (30) 85 (30) + -
R24 gp85 (CI-13) 85 85 + V-1 -
R27 V-1 KLH 85 85 V-1 -
R32 V-3 KLH 170 170 + V-3 +?
R33 V-2 KLH 35,40 35,40 + V-2 -


1- Molecular weight of protein(s) immunoprecipitated by the antibody from respective lysate
2- Surface fluorescence reaction seen when antibody is reacted with virus producing cells and
visualized with fluorescein conjugated antibody
3- Reactivity of antibody with peptide in ELISA









Table 5.4. Comparison of antibody inhibition of the ability of MCUV5 cell-grown virus to
induce immnuoglobulin synthesis by fresh T-depleted human leukocytes.

Antibody:preimmune Ig concn (ng/ml) with: %
Antibody rabbit antibody ---------------------------- Inhibition
(ng/culture) Antibody Antibody
alone + virus


F.2.1


E1D1


R27
(anti-V-1)


None

F.2.1


R32
(anti V-3)



R33
(anti V-2)



VM4C3




None


1000:0
300:700
100:900
30:970
10:990
1000:0
300:700
100:900
30:970
10:990
1000:0
300:700
100:900
30:970
10:990
0:1000

1000:0
300:700
100:900
30:970
10:990
1000:0
300:700
100:900
30:970
10:990
1000:0
300:700
100:900
30:970
10:990
1000:0
300:700
100:900
30:970
10:990
0


70
73
86
112
112
146

573
808
466
520
556
332
333
381
635
347
441
234
263
403
354
396
429
446
427
576
441


0
207
541
1409
3256
3040
4863
4863
4729
2137
3562
3393
2560
4077
3993
5241

508
1130
1517
4700
5908
5856
6881
9440
11465
9533
6841
10452
9387
13647
13753
4764
10235
10762
6755
10587
14897


100
96
90
73
39









Discussion

Monoclonal and polyclonal antibodies have been used extensively to describe, char-

acterize, and study proteins expressed by many different cells and viruses. In order to

clarify the role that gp85 has in the infectious process, monoclonal antibodies, particularly

those with neutralizing ability are very important. Throughout the scope of this work,

many different methods have been employed in order to raise antibodies against a protein

which through time has proven to be relatively difficult to study. Human sera from patients

infected with EBV demonstrate that antibodies to gp85 are fairly abundant during acute

infection. This production of antibodies to gp85 is not easily duplicated in murine systems

using virus produced from lymphoblastoid cell lines. Therefore novel methods were used

to improve the likelihood of antibody production against gp85.

The failure of the cross-reactive immunization to result in antibodies reactive with

EBV gp85 or its HSV homolog, gH was most likely a result of the relative small amount of

either of these proteins in their respective virion. Although both proteins perform a func-

tion which is vital for infectivity, they are not as abundant as other viral glycoproteins. The

role gp85 plays in getting the virus inside a cell probably occurs after the virion has

adsorbed to the cell membrane; therefore, the location or position of the glycoprotein may

be more crucial than the actual quantity of the protein. The cross-reactive immunization

resulted in the majority of antibodies recognizing the gB molecule of HSV. This protein

has been shown to be extremely abundant or at least very immunogenic in many of the her-

pesviruses {8, 22) with a high degree of cross-reaction between viruses. The fact that we

were able to produce immunoglobulin of the IgG class, characteristic of a secondary

immune response, would indicate that this method is useful as an immunization protocol

for cross-reactive proteins.

The other novel method used for immunization was that of peptide priming. It had

been our observation that peptide alone was only weakly immunogenic, if at all, and that

the peptide must be coupled to KLH for efficient antigenicity. Recently, the role of the







90
carrier protein in improving the immune response to a small hapten was described as that of

stimulating a helper T-cell response rather than simply providing a large molecule which

facilitated presentation of the antigen {71). The molecular mass of the carrier is irrelevant

to the system beyond the fact that larger proteins are more likely to contain determinants for

helper T-cell clones of a greater number of histocompatibility types. Several groups have

been able to overcome a non-responsiveness to synthetic peptides by coupling these pep-

tides to T-cell helper determinants, thereby enabling the peptide to be recognized by that

particular genetic strain of mouse {61). Other recent work has indicated that antibodies to

a hepatitis viral envelope protein could be obtained by priming with a single T-cell determi-

nant from the nucleocapsid followed by boosting with intact virus (134). This is thought

to be a function of the recognition of foreign antigens by both T cells and B cells and their

respective role in the immune response; a response which requires the immune system to

recognize the virus as a single whole molecule rather than seeing it by its separate compo-

nents { 117). Internalization of the aggregated structure by B cells results in expression of

T-helper determinants from all of the viral proteins in association with MHC class II anti-

gens. The helper T cells, which had been previously primed with a peptide from the

nucleocapsid, then bind to the B cells expressing not only the nucleocapsid protein but the

envelope protein as well resulting in the production of antibodies against both proteins. If

the same hypothesis can be applied to the expression of multiple epitopes on the gp85

molecule, then it would be possible that a B cell which was expressing T-helper determi-

nants from different gp85 regions would be recognized by the helper T cells which had

been primed by the peptide V-1 immunizations. The resulting affect on the B cell would be

the production of antibodies to regions of gp85 other than the sequence from which V-1

was derived. Further manipulation of these immunologic principles involved in this hypo-

thesis may greatly facilitate the production of antibodies to specific proteins which in the

past have proven only weakly immunogenic.







91
By producing antibodies to several known regions on the gp85 molecule using syn-

thetic peptides, the mapping of crucial epitopes of this molecule may be facilitated. This

assumes, however, that some of the epitopes recognized by these anti-peptide antibodies

are crucial for infectivity. The rabbit antibodies which were produced against the three

synthetic peptides have raised more questions regarding the structure of the protein as well

as providing clues to at least one partially neutralizing epitope. Since the rabbit anti-V-1

antibodies we originally produced recognized a molecule of 85 kDa, which we were able to

show to be gp85, we did not expected to find that antibodies to two other peptides from the

same gene would precipitate proteins with such unique characteristics as a 35-40 kDa

molecular mass and 170 kDa molecular mass. Clearly, additional work will be required to

determine the nature of these polypeptides. It is possible that the smaller proteins are

cleavage products of the mature gp85 and that the larger protein are a SDS-resistant dimer

of gp85 or a complex of gp85 with anther protein or proteins. Alternatively, they may be

proteins other than gp85 which share sequence homology with these two unique peptides.

The failure of either of the anti-V-2 or anti-V-3 antibodies to immunoprecipitate the

mature gp85 was equally surprising since it is assumed that an antibody to a sequential

epitope which could recognize either a cleaved protein or a dimerized protein should also be

able to recognize the protein which has not been altered. Closer examinations of the hun-

dreds of immunoprecipitations performed during this work using a variety of techniques

and conditions revealed occasional appearance of the 170 kDa protein under conditions

when the amount or radioactive intensity of the protein was high. The monoclonal anti-

bodies E1DI and C1D1 immunoprecipitate the 170 kDa protein as well as a rabbit serum

raised against the alcelaphine herpesvirus-1. Similarly, F.2.1 appears to precipitate a low

molecular weight protein of approximately 35 kDa besides the mature gp85. The relation

of these proteins, once established, will assist in mapping the antibody binding sites on the

protein and may provide essential insight about the structure/function of this protein.









Monoclonal antibodies to these peptides will also provide interesting tools which may

clarify the results obtained with the polyclonal rabbit sera.

Studies of the neutralizing capacities of antibodies is the link between the characteri-

zation of the protein and antibodies which react with it and the understanding of how the

virus functions. Since EBV does not produce a lytic infection, neutralization must be mea-

sured by indirect means. The assay developed by Dr. Hutt-Fletcher measures the ability of

the virus to infect and thereby activate B cells to produce immunoglobulin. The sensitivity

of this cumbersome assay allows for measurement of the degree of interference which an

antibody may have on the ability of the virus to infect B cells. Although none of the anti-

bodies produced in this work neutralize EBV as well as F.2.1, several show the potential

of partial neutralization. Many of the antibodies, such as E1D1 bind to the virus as deter-

mined by surface fluorescence, but fail to neutralize infectivity with any notable efficiency.

Other antibodies, such as R27 not only fail to neutralize but are also not seen binding the

protein by surface fluorescence. This may indicate that the site at which this antibody binds

is not available on the molecule in its native structural configuration. We can also not rule

out the possibility that the portion of the molecule which contains the V-1 sequence is actu-

ally found internal to the membrane. This is fairly unlikely, however, since the structure of

the molecule would indicate that it can only pass through the membrane a maximum of two

times, and sequences from all three peptides would need to be located on the same side of

the membrane. Both sequences for peptides V-2 and V-3 appear to be external to the

membrane since antibodies can be seen bound to the protein by surface immunofluores-

cence.

The information garnered from neutralization, surface fluorescence and structure of

the protein is essential in epitope mapping of the important domains of the glycoprotein.

The process of epitope mapping the sites on gp85 which bear functional significance

requires the use of antibodies to the protein of sufficient diversity so as to allow com-

parisons in function and binding. The monoclonal antibodies and polyclonal antibodies







93
which have been produced in this work should facilitate the process by which the func-

tional domains of the protein are established, thus resulting in a better understanding of the

process by which EBV infects cells.














CHAPTER VI

SUMMARY AND CONCLUSIONS

Recapitulation

A careful consideration of the characteristics which EBV has developed in its sym-

biotic relationship with its host will reveal that this virus has evolved into a niche which

allows it to persist, even as an infectious agent, in an almost harmless balance with man.

The fact that the virus is rarely implicated in various neoplasms as well as numerous

lymphoproliferative abnormalities is an indication of a disruption in the careful immuno-

logic balance within which the virus developed. EBV has carefully evolved a mechanism

for infecting two cell populations, epithelial cells of the nasopharynx and B lymphocytes.

This mechanism allows entry of the virus by attachment to a molecule, CR2, expressed on

the surface for functional purpose of those cells. Although the virus causes a rapid prolif-

eration and transformation in infected cells, it does so in the B-cell lineage which possesses

natural immunologic checks. Once the immune system has righted itself of the imbalance,

the virus persists in an episomal state without malaise until once again the balance is dis-

turbed. The source of this disruption, whether striking at the immunologic balance, or

disturbing the chemical nature of the transformed cell by carcinogenesis, is probably

always required for further damaging affects associated with the virus.

Critical in our understanding of the biologic activity of EBV is an understanding of

how it successfully infects cells and what components, such as surface membrane proteins

are important for this process. Among the many proteins encoded by Epstein-Barr virus,

the membrane antigens have special functions which have specifically selected for their

expression on the outer surface of the virion envelope. Of these membrane antigens, gp85




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EWBN46S29_D1GMP7 INGEST_TIME 2012-02-20T21:43:10Z PACKAGE AA00009086_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES