IDENTIFYING THE GENETIC BASIS OF ATTENUATION IN MAREK’S DISEASE VIRUS
VIA EXPERIMENTAL EVOLUTION
By
Evin Hildebrandt

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Genetics - Doctor of Philosophy
2014

I

ABSTRACT
IDENTIFYING THE GENETIC BASIS OF ATTENUATION IN MAREK’S DISEASE VIRUS
VIA EXPERIMENTAL EVOLUTION
By
Evin Hildebrandt

Marek’s disease virus (MDV), an oncogenic alphaherpesvirus of chickens, causes up to
$2 billion in loses a year due to Marek’s disease (MD). Therefore control of this economically
important disease is critical. The primary method to control MD is vaccination. Attenuated, or
weakened, strains of MDV have been generated via repeated in vitro serial passage to generate
avirulent MDV strains that have been used as successful MD vaccines. Despite introduction of
several vaccines since the 1970’s, more virulent strains of MDV have evolved to break vaccinal
protection. Therefore, development of new MD vaccines is necessary. To address this concern,
we sought to better understand the molecular basis of attenuation in MDV to provide information
that may assist in the rationale design of MD vaccines.
Three attenuated replicates of a virulent MDV were serially passed in vitro for over 100
passages. DNA and RNA from attenuated viruses were deep sequenced using Illumina nextgeneration sequencers to identify changes in DNA sequence or expression following attenuation.
Top candidate mutations identified via sequencing were used to generate seven recombinant
viruses using red-mediated recombineering for mutations within UL42, UL46, UL5, two
involving LORF2 and two mutations within ICP4. These recombinant viruses were tested in vivo
to determine the impact of these mutations on MD incidence, in vivo replication and horizontal
transmission. Point mutations within UL42, UL46, LORF2-Promoter and ICP4 did not cause
II

observable phenotypic changes compared to the parental virus. A single point mutation within
LORF2-Intron and a double mutant involving ICP4 both resulting in 100% MD in challenged
birds but failed to transmit horizontally to uninfected contact birds. Finally, a point mutation
within UL5 reduced MD incidence by over 90%, significantly reduced in vivo replication, and
eliminated horizontal transmission.
Further characterization of this UL5 point mutation determined that it increased in vitro
replication in growth curves, yet head-to-head competition of the Mut UL5 virus versus parental
virus showed the parental virus outcompeted the mutant virus. Furthermore, serial passage of
Mut UL5 in vivo did not result in increased in MD incidence, in vivo replication or result in
reversion or compensatory mutations to UL5 after passage through birds. Trials testing vaccinal
protection of the Mut UL5 virus showed the virus provided partial protection against challenge
with virulent MDV, yet did not exceed protection achieved through use of traditional vaccines.
Therefore, use of this point mutation in combination with other candidate mutations was tested.
Addition of the UL5 mutation with ΔMeq, a candidate vaccine with high protection and
replication but also induces bursal-thymic atrophy (BTA), resulted in a recombinant virus that
replicated at low levels and did not cause BTA, yet reduced levels of vaccinal protection,
indicating an intricate relationship between replication levels, BTA and vaccinal protection.
This study shows that a variety of genes are mutated during attenuation, and particularly
mutations within DNA replication genes, such as UL5, appear to play an important role in
attenuation. We also determined that experimental evolution is a process that not only can
identify mutations involved in attenuation, but also offer protection as a vaccine to provide
information for further development of MD vaccines.

III

ACKNOWLEDGEMENTS

I would like to thank everyone who made it possible not only to complete my project, but
to make my time as a Ph.D student memorable and enjoyable. First off, I must thank my advisor
Hans Cheng. As a scientist and advisor, I cannot think of anyone who is a better person to fill
those roles. His knowledge, guidance, patience and enthusiasm provided the perfect and
supportive work environment to not only be productive in regards labwork, but also intellectually
develop as a scientist. I wish all graduate students could have the pleasure of working with
advisors that were at least half the caliber of Hans, since they would be hard pressed to find ones
that would come close to equaling his skills as a mentor.
As in my papers, I must acknowledge Laurie Molitor, Spencer Jackson and Lonnie
Milam for their technical assistance in completion of my experiments. Without their help I might
still be sitting out in the bird house bleeding chicks, or extracting DNA from countless samples
with an obscene numbers of boxes of samples still sitting in the freezer waiting for me to work
through.
Laurie showed me the ropes on how to conduct all the vital bird work, including how to
challenge birds, work in necropsy and bleed birds like a pro, even those teeny tiny chicks less
than a week old that at first seemed impossible to not immediately cause a hematoma and ruin all
chances of obtaining any sample…all while somehow manage to have fun doing it too. Anyone
who has ever touched a pipette while Lonnie Milam has worked at ADOL surely must thank him
for his absolutely priceless knowledge and skills regarding nearly any assay you might even
consider doing in a lab. From basic tissue culture, transfection, recombineering, growth curves,
qPCR, DNA extraction, PCR, and nearly any other molecular biology technique his extensive
iv

expertise and willingness to freely and happily offer advice, reagents, assistance or guidance has
helped me complete a myriad of assays and overcome those numerous tricky complications I
have run across over the course of my work at ADOL. Without his knowledge I know I would
not have been able to successfully complete as much as I have.
Finally, but definitely not least, I also must thank all of my excellent colleagues, coworkers and friends I have worked with including Laurie Molitor, Sudeep Perumbakkam, Cari
Hearn, Suga Sugaleshini and countless others who have spent time at ADOL and in the Cheng
Lab. They have made the long hours spent working in the lab not only productive thanks to their
help, but also enjoyable thanks to the oftentimes strange ways we passed the time during the
myriad of time spent waiting for one assay or another to finish.
I also must embarrassingly thank the staff at the East Lansing Taco Bell for whom the
members of the Cheng lab, myself included, have become regular customers. Who knew Taco
Bell was such a productive hotbed for in-depth scientific discussion, brain-storming and
troubleshooting? To my friends and co-conspirators in the Cheng Lab, I’d like to horribly misappropriate and butcher a classic quote because no matter what, “We’ll always have Paris Taco
Bell Happy Hour”.

v

TABLE OF CONTENTS
LIST OF TABLES ......................................................................................................................... ix
LIST OF FIGURES ........................................................................................................................ x
KEY TO ABBREVIATIONS ...................................................................................................... xiii
CHAPTER 1. INTRODUCTION ................................................................................................... 1
VIRUSES, ATTENUATION AND VACCINATION ............................................................... 1
MAREK’S DISEASE AND VACCINATION ........................................................................... 4
MAREK’S DISEASE VIRUS AND ATTENUATION ............................................................. 7
IDENTIFYING THE GENETIC BASIS OF ATTENUATION IN MDV ............................... 12
Objective 1: Generate parallel, attenuated replicates of MDV, all initially derived from an
identical, virulent MDV. ........................................................................................................ 12
1a: Produce attenuated viruses through serial passage in vitro. ............................................ 12
1b: Determine at which passage each replicate became completely attenuated. ................... 13
Objective 2: Identify sequence variation and expression differences in attenuated strains that
appeared likely candidates responsible for creation of attenuated strains. ............................ 13
Objective 3: Experimentally validate candidate mutations which appear to correspond with
attenuation. ............................................................................................................................ 13
Objective 4: Further characterization of mutations shown to alter virulence ........................ 14
4a: Comparison of in vitro replication of recombinant UL5 mutant virus to growth rates of
the virulent parental virus. ..................................................................................................... 14
4b: Serial passage of the attenuated UL5 recombinant virus through birds to identify
restoration of virulence or increased in vivo replication due to adaptation during selection for
improved in vivo replication. ................................................................................................. 14
4c: Determine vaccinal potential of UL5 recombinant viruses ............................................. 15
REFERENCES .......................................................................................................................... 16
CHAPTER 2. CHARACTERIZING THE MOLECULAR BASIS OF ATTENUATION OF
MAREKS DISEASE VIRUS VIA IN VITRO SERIAL PASSAGE IDENTIFIES DE NOVO
MUTATION IN THE HELICASE-PRIMASE SUBUNIT GENE UL5 AND OTHER
CANDIDATES ASSOCIATED WITH REDUCED VIRULENCE ............................................ 20
ABSTRACT .............................................................................................................................. 20
IMPORTANCE ..................................................................................................................... 21
INTRODUCTION ..................................................................................................................... 21
MATERIALS AND METHODS .............................................................................................. 24
Tissue Culture ........................................................................................................................ 24
Viruses ................................................................................................................................... 24
In Vivo Virulence Trials ........................................................................................................ 25
Virulence Trials of Mixed Virulent and Avirulent MDV...................................................... 25
Sequencing of Viral Stocks ................................................................................................... 26
Identification of Mutations .................................................................................................... 26
Tracking SNV Kinetics ......................................................................................................... 26
Pathway Analysis of Mutated Genes ..................................................................................... 27
vi

Generation of Defined Recombinant Viruses ........................................................................ 27
In Vivo Characterization of Mutant Viruses .......................................................................... 30
In Vivo Replication of MDV ................................................................................................. 30
RESULTS.................................................................................................................................. 31
Attenuation of Serially Passed Replicates ............................................................................. 31
MD Incidence in Defined Mixtures ....................................................................................... 31
Next-Generation Sequencing Analysis of Attenuated Replicates ......................................... 32
Candidate Mutations and Characterization of Recombinant Viruses .................................... 47
DISCUSSION ........................................................................................................................... 54
ACKNOWLEDGEMENTS ...................................................................................................... 61
REFERENCES .......................................................................................................................... 62
CHAPTER 3. MUTATIONS WITHIN ICP4 ACQUIRED DURING IN VITRO
ATTENUATION DO NOT ALTER VIRULENCE OF RECOMBINANT MAREK’S DISEASE
VIRUSES IN VIVO ....................................................................................................................... 66
ABSTRACT .............................................................................................................................. 66
INTRODUCTION ..................................................................................................................... 67
MATERIALS AND METHODS .............................................................................................. 69
Recombinant Viruses and Tissue Culture ............................................................................. 69
In Vivo Characterization of Mutant Viruses .......................................................................... 71
Sequence Comparison of ICP4 among Alphaherpesviruses ................................................. 71
RESULTS.................................................................................................................................. 72
In Vivo Trials ......................................................................................................................... 72
Homology of Alphaherpesviruses ICP4 Protein Sequence ................................................... 77
DISCUSSION ........................................................................................................................... 85
Conclusions ........................................................................................................................... 90
ACKNOWLEDGEMENTS ...................................................................................................... 90
REFERENCES .......................................................................................................................... 91
CHAPTER 4. A SINGLE NUCLEOTIDE MUTATION WITHIN THE UL5 HELICASEPRIMASE SUBUNIT GENE NOT ONLY REDUCES VIRULENCE, BUT ALSO PROVIDES
VACCINAL PROTECTION AGAINST MAREK'S DISEASE.................................................. 94
ABSTRACT .............................................................................................................................. 94
INTRODUCTION ..................................................................................................................... 95
MATERIALS AND METHODS .............................................................................................. 97
Viruses and Cell Culture........................................................................................................ 97
In Vivo Challenge .................................................................................................................. 97
Vaccinal Protection of Viruses .............................................................................................. 98
RESULTS.................................................................................................................................. 99
PI of Viruses ........................................................................................................................ 103
DISCUSSION ......................................................................................................................... 104
ACKNOWLEDGEMENTS .................................................................................................... 107
REFERENCES ........................................................................................................................ 108

vii

CHAPTER 5. ADDITION OF A UL5 HELICASE-PRIMASE SUBUNIT POINT MUTATION
ELIMINATES BURSAL-THYMIC ATROPHY OF MAREK’S DISEASE VIRUS ∆MEQ
RECOMBINANT VIRUS, BUT REDUCES VACCCINAL PROTECTION ........................... 111
ABSTRACT ............................................................................................................................ 111
INTRODUCTION ................................................................................................................... 112
MATERIALS AND METHODS ............................................................................................ 115
Viruses and Tissue Culture .................................................................................................. 115
In Vivo Replication and Bursal Thymic Atrophy ................................................................ 115
Vaccinal Protection.............................................................................................................. 116
RESULTS................................................................................................................................ 117
In Vivo Replication .............................................................................................................. 117
BTA of Meq-Based Viruses ................................................................................................ 117
Protective Efficacy of Vaccines .......................................................................................... 117
DISCUSSION ......................................................................................................................... 123
ACKNOWLEDGEMENTS .................................................................................................... 127
REFERENCES ........................................................................................................................ 128
CHAPTER 6. CHARACTERIZATION OF MAREK’S DISEASE VIRUS POINT MUTATION
WITHIN UL5 HELICASE-PRIMASE SUBUNIT IDENTIFIED WITHIN ATTENUATED
VIRUS SHOWN TO REDUCE VIRULENSE AND IN VIVO REPLICATION ...................... 131
ABSTRACT ............................................................................................................................ 131
INTRODUCTION ................................................................................................................... 132
MATERIALS AND METHODS ............................................................................................ 133
Tissue Culture: Cell and Virus Propagation ........................................................................ 133
UL5+UL46 Double Mutant In Vivo Replication and MD Incidence .................................. 133
Viral Growth Curves ........................................................................................................... 136
Viral Competition ................................................................................................................ 136
Serial Passage of Mut UL5-I682R In Vivo .......................................................................... 137
RESULTS................................................................................................................................ 138
UL5+UL46 Double Mutant In Vivo Replication and MD Incidence .................................. 138
Mut UL5-I682R In Vitro Replication; Viral Growth Curves and Competitions ................. 138
Serial Passage of Mut UL5-I682R In Vivo .......................................................................... 139
DISCUSSION ......................................................................................................................... 147
ACKNOWLEDGEMENTS .................................................................................................... 150
REFERENCES ........................................................................................................................ 151
FURTHER WORK ..................................................................................................................... 154
HORIZONTAL TRANSMISSION OF RECOMBINANT VIRUSES ............................... 154
IN VITRO REPLICATION OF UL46 RECOMBINANT VIRUSES ................................ 155
FURTHER CHARACTERIZATION OF IN VITRO REPLICATION OF MUT UL5 ..... 156

viii

LIST OF TABLES

TABLE 2-1. Genomic Regions and PCR Primers Used for Targeted Resequencing of SNVs ... 28
TABLE 2-2. Mutational Primers for Generation of Recombinant Viruses .................................. 29
TABLE 2-3. Mutations in Sequenced Viral Stocks ...................................................................... 37
TABLE 2-4. SNVs Located in Serially Passed Md5B40BAC Replicates ................................... 38
TABLE 2-5. Depth of Coverage for Sequenced Viruses .............................................................. 42
TABLE 2-6. ICP4 Mutation in Attenuated Viruses...................................................................... 44
TABLE 2-7. Top Candidate Mutation Frequencies in DNA vs RNA .......................................... 45
TABLE 2-8. Significantly Differentially Expressed Genes in Attenuated MDV Replicates vs
Virulent MDV ............................................................................................................................... 46
TABLE 2-9. MD Incidence Due to Candidate Recombinant Viruses .......................................... 52
TABLE 3-1. Mutational Primers for Generation of Recombinant ICP4 Viruses ......................... 70
TABLE 3-2. MD Incidence of Recombinant ICP4 Viruses ......................................................... 74
TABLE 3-3. Alphaherpesvirus ICP4 Sequences .......................................................................... 79
TABLE 4-1. MD Incidenence and Vaccinal Protection ............................................................. 102
TABLE 5-1. Vaccinal Protection Against Challenge with 648A MDV ..................................... 122
TABLE 6-1. Primers ................................................................................................................... 135
TABLE 6-2. MD Incidence of Single and Double UL5 Recombinant Viruses ......................... 142
TABLE 6-3. Disease Incidence Pre and Post Serial Passage In Vivo ......................................... 145

ix

LIST OF FIGURES

FIGURE 1-1. Introduction of MD Vaccines and the Increased Virulence of MDV. Adapted
Witter 1996 ..................................................................................................................................... 6
FIGURE 2-1. Virulence of MDV replicates derived from a serially passed, BAC-cloned virus.
Two bird trials (A and B) were conducted for Reps 1, 2, and 3 each at p55, p65, and p75, e.g.,
R1-A is Replicate 1, trial A at the indicated passages. In addition, the control strain Md5 was
tested at p41, 51, and p61.............................................................................................................. 35
FIGURE 2-2. Disease incidence in chickens challenged with defined mixtures of virulent and
avirulent MDV. Bird trials conducted in duplicate are indicated by “A” and “B,” e.g., 5:495-A
was trial A where the inoculation consisted of 5 PFU B40-derived MDV and 495 rMd5∆Meq. 36
FIGURE 2-3. Genes with nonsynonymous mutations within attenuated MDV replicates. Serially
passed MDV replicates were compared to identify commonly mutated genes shared among
attenuated replicates. ..................................................................................................................... 43
FIGURE 2-4. Kinetics of candidate SNP frequencies over serial passage. Targeted resequencing
of candidate SNPs allowed for the frequency of each mutation to be determined from the earliest
passage in which the mutation occurred until attenuation. ........................................................... 50
FIGURE 2-5. Survival of birds challenged with recombinant MDVs. Mortality due to MD over
the course of the 10 weeks was determined for the five recombinant viruses and the virulent B40derived MDV positive control. ..................................................................................................... 51
FIGURE 2-6. Comparison of in vivo replication for attenuated replicates and recombinant
viruses. Due to housing and chick availability, recombinant viruses were tested incrementally,
therefore positive (Md5B40-c1) and negative (uninfected) controls are matched to the respective
recombinant viruses tested collectively by matched symbols (i.e. *, ^, #, etc.). Trials for Mut
UL5 were conducted in duplicate and each trial indicated is by “A” or “B.” .............................. 53
FIGURE 3-1. Survival curves of birds challenged with ICP4 recombinant MDV. Mortality of
birds during the 12 week experiment until termination was compared for the two ICP4 mutant
viruses relative to the parental Md5B40BAC virus. ..................................................................... 75
FIGURE 3-2. Viral replication rates of birds challenged with ICP4 recombinant MDV. In vivo
replication of recombinant ICP4 MDV versus parental Md5B40BAC virus. .............................. 76

x

FIGURE 3-3. Amino acid alignment of N-terminal end of ICP4. Individual sequences for the
three mardiviruses are shown while all other 17 herpesvirus ICP4 sequences are combined and
shown as a non-mardivirus consensus sequence. The region containing amino acids 52-69 unique
to MDV in Gallid herpesvirus 2 is indicated by the boxed region, while amino acids 60, 62 and
63 are designated with an *. .......................................................................................................... 80
FIGURE 3-4. Amino acid alignment of 20 ICP4 species and HSV-1. The S1630P mutated
position in Gallid herpesvirus 2 is indicated by *. Identical amino acids at the same position are
indicated in black boxes, while amino acids with comparable properties but different identities
are indicated by shaded boxes....................................................................................................... 81
FIGURE 3-5. Maximum likelihood phylogenetic tree based on the complete amino acid
sequence of ICP4 from 20 herpesvirus species. Artificially rooted with the outgroup of UL36
amino acid sequence from Gallid herpesvirus 2. .......................................................................... 84
FIGURE 4-1. Survival of birds challenged with JM/102W virus. A minimum of 17, 15I5 x 71
maternal antibody negative chicks were vaccinated with either HVT or Mut UL5-I682R virus at
day of hatch, while unvaccinated birds were used as controls, before challenge with 500 PFU of
JM/102W virus at 5 days of age and mortality shown over the course of the experiment until
termination. ................................................................................................................................. 100
FIGURE 4-2. Survival of birds challenged with Md5B40BAC-c1 virus. A minimum of 17, 15I5
x 71 maternal antibody negative chicks were vaccinated with either bivalent (HVT+SB-1)
vaccine or Mut UL5-I682R of virus at day of hatch, while unvaccinated birds were used as
controls, before challenge with 500 PFU of Md5B40BAC-c1 virus at 5 days of age and mortality
shown over the course of the experiment until termination. ....................................................... 101
FIGURE 5-1. In vivo replication of MDV within PBLs in vaccinated, unvaccinated or 648A
challenged birds at 6, 13 and 20 days post vaccination. Unvaccinated birds challenged with 648A
experienced severe mortality around 2 weeks post infection, resulting in no data for the final
timepoint indicated with an “x”. Panel A designated trial one and panel B indicates trial 2. .... 119
FIGURE 5-2. Weight of lymphoid organs in vaccinated and unvaccinated birds relative to total
body weight in Trial 1 (Panel A) and Trial 2 (B). ...................................................................... 120
FIGURE 5-3. Survival of vaccinated birds following challenge with vv+ 648A MDV. Panel A
shows data from the first trial, with panel B depicting data from second replicate trial. Data for
birds vaccinated with bivalent vaccines was unavailable for trial two due to flood causing
complete loss of the isolator housing that lot.............................................................................. 121

xi

FIGURE 6-1. In Vivo Replication of UL5 Single and Double Recombinant Viruses. Splenic
DNA was extracted for qPCR analysis to quantify viral loads of the single UL5 mutant virus or
double Mut UL5+UL46 recombinant viruses relative to uninfected birds (Neg) or wildtype
Md5B40 challenged birds. Results from two separate trials of Mut UL5 are indicated by “a” and
“b”. .............................................................................................................................................. 141
FIGURE 6-2. Growth curves of Wild-type MdB40BAC-c1 and Mut UL5. 50 PFU of parental
Md5B40BAC or Mut UL5 viruses were used to infect p60 plates containing a CEF. Triplicate
plates of each virus were then trypsinized to harvested and freeze plates at 0, 1, 2, 3, 4 and 5 days
post infection. Dilutions of these harvested plates were used to titer and calculate the number of
plaques for all timepoints for the two viruses to compare plaque yield of Mut UL5 versus
Md5B40BAC viruses. ................................................................................................................. 143
FIGURE 6-3. Head-to-Head competition of wild-type Md5B40BAC-c1 vs Mut UL5. 2,000 PFU
of either Mut UL5 (UL5) or Md5B40BAC (B40) viruses were used to infect p100 plates
containing a CEF monolayer. Three plates were then infected with 2,000 PFU of both viruses
and designated plates A,B or C (B40:UL5 A, B, C). These plates were serially passed and at each
passage stocks saved for DNA extraction for pyrosequencing analysis to quantify the proportion
of DNA containing the wild-type allele or the mutant allele following serial passage. ............. 144
FIGURE 6-4. In Vivo Replication of Md5B40BAC-c1 and Mut UL5 within PBL and spleen (s)
tissue Pre and Post serial passage in vivo. Birds challenged with the Mut UL5 (UL5) or
Md4B40BAC (B40) viruses were bled for collection of PBLs or splenic DNA and extracted for
qPCR to quantify viral load of the respective viruses prior to in vivo passage (p0) or after three
passages through birds while selecting for birds with the highest viral loads (p3) to determine
MDV levels in viruses following selection for greater in vivo replication. ................................ 146

xii

KEY TO ABBREVIATIONS

ACUC

Animal Care and Use Committee

ADOL

Avian Disease and Oncology Lab

BAC

Bacterial artificial chromosome

BTA

Bursal-thymic atrophy

CEF

Chicken embryo fibroblast

CVI988

A.K.A Rispens vaccine

DEF

Duck embryo fibroblast

DPI

Days post infection

DPV

Days post vaccination

FBS

Fetal bovine serum

GAPDH

Glyceraldehyde 3-phosphate dehydrogenase

gB

Glycoprotein B

HB

Horsfall-Bauer isolation unit

HSV-1

Herpes simplex virus 1 or Human herpesvirus 1

HVT

Turkey herpesvirus

ICP4

Immediate early transcriptional regulator gene RS1

xiii

IRL

Internal repeat long region of MDV genome

IRS

Internal repeat short region of MDV genome

m

Pathotype of MDV designated “mild”

MD

Marek’s disease

MDV

Marek’s disease virus

MEQ

“MDV EcoQ” the MDV oncogene

PBL

Peripheral blood lymphocytes

PBS

Phosphate buffered saline

PFU

Plaque forming unit

PI

Protective index

TRL

Terminal repeat long region of MDV genome

TRS

Terminal repeat short region of MDV genome

UL

Unique long region of MDV genome

US

Unique short region of MDV genome

v

Pathotype of MDV designated “virulent”

vv

Pathotype of MDV designated “very virulent”

vv+

Pathotype of MDV designated “very virulent plus”

xiv

CHAPTER 1. INTRODUCTION

VIRUSES, ATTENUATION AND VACCINATION

Viral attenuation is a process describing the loss of virulence from a virulent, disease
causing virus to an avirulent virus that is no longer able to cause disease. These attenuated or
‘weakened’ viruses often provide a useful and functional purpose as vaccines to protect against
virulent, disease causing strains. Live viruses have been used to protect against disease causing
strains since the 18th century when first employed by Edward Jenner’s use of the Cowpox virus
as a vaccine providing protection against smallpox (1).
Prior to the 18th century, a variety of methods had been described as routes to preventing
smallpox, such as a multitude of herbal remedies, dietary restrictions, temperature regulation and
other treatments were documented, all with variable degrees of success (2). The most successful
of these of methods were based on observations that survivors of smallpox were immune to
future outbreaks. One method based on this observation that likely had been in use since the
seventh century was known as variolation, in which material from the pustules of smallpoxaffected individuals was used to inoculate naïve individuals with the intention to protect against
smallpox (3). While this method did provide utility and protection against smallpox, it also
conferred a high risk of disease transmission, not only for development of smallpox itself in 23% of individual variolated, but also transmission of other diseases, such as syphilis (2). To
avoid dangerous complications from variolation, observations that milkmaids tending cows often
developed cowpox marks, but did not contract smallpox or suffer any symptoms more severe
than mild poxmarks from the cows. This provided an alternative to using smallpox directly in
variolation. Instead, transmission of material from cowpox pustules led to protection against
1

smallpox without the risk of developing the deadly disease itself, as well as helping to name the
process of vaccination thanks to the source material derived from the cow, or ‘vaca’ in latin
(4).This widespread application of a live virus as a vaccine for control of a deadly virus
established the groundwork for proliferation of vaccine development, eventually culminating in
the eradication of smallpox and the control of a multitude of diseases such as polio, measles,
rabies, tuberculosis and many more diseases that are successfully prevented and controlled
through the use of vaccination (1).
A variety of techniques are currently used to create vaccines. This includes heat or
chemicals to produce inactivated vaccines from disease-causing viruses to generate a killed
vaccine, or use of recombinant DNA to express only a specific region or epitope of the microbe
to stimulate immunity against the pathogen, which are known as subunit vaccines. However, of
the different vaccine types, live-attenuated vaccines are often considered the version that
provides the greatest level of protection. Live-attenuated vaccines can provide a stronger and
more robust immunity against pathogens and often can provide life-long immunity due to low
levels of replication of live-attenuated viruses, leading to greater persistence of protection (5).
Live-attenuated vaccines have been generated by several methods, all of which are
capable of producing an avirulent virus that is also protective against challenge with a virulent
virus. One of the first methods implemented utilizes viral species that are naturally avirulent in
certain species and able to protect against other closely related viruses that are pathogenic to that
host, as illustrated by Jenner’s use of the closely related, apathogenic Cowpox virus against the
deadly Smallpox in humans. A second method employed to generate attenuated viruses is based
on prolonged exposure of the virulent virus to conditions detrimental or suboptimal for
replication of the virulent virus, leading to a weakened virus poorly adapted for replication and
2

unable to cause disease when reintroduced into the host species as a vaccine. The first labattenuated vaccine against rabies was created using this method in which neuronal samples from
affected rabbits were removed and dried for several days to weaken the rabies virus before
administering the neuronal tissue containing the virus as a vaccine (6). These initial rabies
vaccines provided an alternative following exposure to a rabid animal and were able to prevent a
lethal outcome that previously had no medical recourse. Vaccines based upon this technique
using infected tissues from species such as rabbits or sheep is still employed in some countries
such as Ethiopia and India today due to their inexpensive cost (7, 8), but further refinement and
development of rabies vaccines has resulted in even safer lines of rabies vaccines generated via
cultivation in cell lines, which the World Health Organization (WHO) recommends for
replacement of earlier tissue-derived vaccines (9).
In cell culture based methods for generation of attenuated viruses, virulent viruses are
grown and serially passed in tissue culture conditions, which are significantly different from the
in vivo host environment that the virulent viruses had evolved to replicate, transmit and cause
disease. It has long been known that passage of virulent viruses through non-target species or
sub-optimal conditions leads to a selection for avirulent viruses during this process (10). For
example, over the course of repeated passage in tissue culture, viruses can evolve and adapt to in
vitro conditions, leading to a loss of virulence when returned to the host species where they are
no longer are optimized for replication within in vivo conditions. One of the first, well-known
vaccines generated via this method of serial passage was the Sabin vaccine in which the polio
virus was cultivated in monkey kidney cells, resulting in attenuation of the polio virus for use as
a live attenuated vaccine (1). In vitro serial passage to attenuate virulent viruses is a process

3

used to generate many of today’s vaccines against diseases in addition to polio and rabies, such
as measles, mumps, chicken pox, yellow fever, and influenza (5).
Widespread use of live attenuated vaccines is not just limited to controlling human
diseases, but many diseases affecting livestock and animals as well. Marek’s disease (MD) is one
such example of a major economic disease affecting livestock that is controlled by vaccination
with live-attenuated viruses. MD is caused by an oncogenic alphaherpevirus of chickens known
as Gallid herpesvirus 2, also commonly known as Marek’s disease virus (MDV). MD is a T-cell
lymphoma resulting in paralysis, nerve lesions, tumors and eventually death of affected chickens.
MD costs the poultry industry $1-2 billion annually as a result of losses due to MD, therefore
control of MD is critical for modern poultry production (11) . The primary method to control
MD has been the use of vaccines, with the first anti-tumor vaccine in any species first developed
for prevention of MD (12).
MAREK’S DISEASE AND VACCINATION
The first attenuated MDV developed in 1969, known as HPRS-16, was an attenuated
virus generated via repeated serial passage of a virulent MDV strain through chicken kidney cells
until after approximately 30 in vitro passages the resulting virus became apathogenic. This
demonstrated that MDV could be attenuated through repeated passage in tissue culture, with the
practical application of this technique for vaccine development. Shortly after demonstrating the
attenuation of MDV via in vitro serial passage, a closely related herpesvirus of turkeys (HVT),
known as Melaegrid herpesvirus 1, was determined to be apathogenic in chickens and also
conferred vaccinal protection against challenge with virulent MDV. HVT was licensed and
widely used as a vaccine against MDV throughout the 1970’s until more virulent strains of MDV
classified as ‘virulent’ (v) arose that were able to break vaccinal protection of HVT (Figure 1-1).
4

This resulted in the introduction of Bivalent vaccines in the 1980’s containing a combination of
two MDV vaccines, such as HVT plus an avirulent strain such as SB-1, which interacted
synergistically to provide greater protection when combined than either virus provided alone.
Bivalent vaccines provided the best protection against outbreaks of MD until the early 1990’s
when even more virulent field strains classified as ‘very virulent’ (vv) evolved able to cause MD
even in birds vaccinated with Bivalent vaccines. This resulted in the introduction of the current
‘gold-standard’ vaccine known as CVI988/Rispens. Rispens is currently the most protective
vaccine commercially available able to protect against even strains classified as ‘vv.’ Despite the
success of Rispens to control MD since the 1990’s, it appears that new strains classified as ‘very
virulent plus’ (vv+) are now circulating within flocks, raising concerns that vaccines exceeding
the protection currently provided by Rispens may soon be required (13).
Despite the success of MDV vaccines at reducing MD losses, vaccination itself may be
driving the evolution of more virulent MD strains. In a vivid example of the Red-Queen
hypothesis, MDV vaccines and virulent MDV field strains appear to be in an evolutionary arms
race in which following introduction of a new, more protective vaccine, field strains of MDV
respond in kind with new, more virulent strains able to overcome the new vaccine, leading to this
cycle of continued vaccine development, introduction and viral evolution (14). While MD
vaccines prevent the symptoms of MD such as tumors and nerve lesions, they are non-sterilizing
vaccines, in which vaccinated birds are still able to be infected with virulent field strains of
MDV (13). This results in both attenuated vaccine strains and virulent field strain of MDV coinfecting and replicating within the same bird, providing a selection force for the evolution of
more virulent MDV strains able to break vaccinal protection.

5

FIGURE 1-1. Introduction of MD Vaccines and the Increased Virulence of MDV. Adapted
Witter 1996

6

Due to the intrinsic limitations of non-sterilizing vaccines, it appears likely that
successive lines of MD vaccines must be periodically developed and introduced to combat the
continued evolution of more virulent MDV field strains. While vaccines derived from in vitro
serial passage, such as HPRS-16 and Rispens, have been successfully created, there is concern
that there may be a limit to the maximal vaccinal protection against MD as seen with current
vaccine development. Protective indices of ten candidate vaccines were compared against
Rispens, currently the most-protective licensed vaccine available. Two of the ten viruses had
high levels of protection comparable to Rispens, indicating that new candidate vaccines can be
developed to equal current protection levels (15). Unfortunately, none of the candidate had
protection levels exceeding Rispens, indicating that development of a more protective vaccine to
address the evolution of vv+ strains of MDV may be a difficult task.
MAREK’S DISEASE VIRUS AND ATTENUATION
In order to address this challenge, a better understanding of the genetic basis of
attenuation may assist in the development of more protective candidate vaccines. As shown by
Witter et al. (15), blind serial passage can reliably produce attenuated vaccines, of which some
may provide appreciable levels of vaccinal protection, but with the current opportunities
available with recombinant DNA technology and next-generation sequencing, it could be
possible to design and engineer an attenuated virus able to provide even greater levels of
protection against vv+ strains of MDV. In order to precisely design attenuated viruses, it is
necessary to better understand the process of attenuation. Despite the utilization of attenuation of
MDV since 1969 in generating candidate vaccines, little is known regarding what drives the
process of attenuation at a genetic level.

7

Early observations of attenuated MDV strains showed a signification loss in the
quantities of A-antigen, later identified as glycoprotein C (gC), which is encoded by UL44.
Expression of gC was shown to be significantly reduced and lost after passage, correlating with a
loss of virulence (16). Despite this correlation, expression of gC is variable, in which oncogenic
viruses have been identified that do not produce gC (17) and attenuated viruses that produce high
levels of gC expression (18). This indicates that loss of gC expression during serial passage is
only a general trend and not the causative force for attenuation. This conclusion is further
supported by sequencing of gC in attenuated and virulent strains that found no differences in the
sequence or promoter of gC itself to account for changes in virulence of MDV strains, suggesting
the role of other factors for alteration of gC expression besides direct mutation of gC (19).
Another proposed mechanism for the attenuation of MDV during in vitro serial passage
involved the genomic region known as the 132 bp repeats. As the name suggests, this is a
repetitive, 132 base pair long region normally present as two to three tandemly repeated copies
(20, 21). During serial passage, a proliferation in the number of tandem copies of the 132 bp
repeats occurs while viruses are becoming attenuated. Despite the correlation of this
phenomenon with attenuation, recombinant viruses testing the impact of mutating the number of
copies of the 132 bp repeats show expansion of these repeats is not sufficient for attenuation.
Incorporation of 9 tandem copies of the 132 bp repeat region into a virulent MDV and following
only two passages in vitro resulted in viruses with up to 20 copies of the 132 bp repeats, yet the
resulting viruses were still virulent despite having the 132 bp repeat expansion pattern common
in attenuated viruses (22). Furthermore, complete deletion of the 132 bp region in a virulent virus
prior to repeated in vitro serial passage showed that despite lacking this region, the 132 bp

8

deletion virus was still able to become attenuated during prolonged passage in tissue culture,
proving that the 132 bp repeats are not necessary for attenuation (23).
Understanding the basis of attenuation has been a topic of research for many years, yet
despite the clear correlation of several traits common to attenuated viruses, such as the loss of gC
expression and expansion of 132 bp repeats, it is clear a more systematic approach to
determining candidate genes involved in attenuation is necessary. With the advent of nextgeneration sequencing (NGS) technology, it is now possible to sequence the complete genome of
a multitude of viruses to compare variation among various virulent strains, attenuated viruses and
vaccines. The complete genome of several MD vaccines, such as SB-1 and Rispens, have been
sequenced and compared to oncogenic strains to find differences between vaccines and virulent
MDV (24, 25). The serially passed and attenuated 648A strain of MDV has also been sequenced
to identify differences in the attenuated virus relative to the parental 648A virus prior to passage
to identify numerous mutations occurring during serial passage (26). Therefore, due to NGS
technology, it is now possible to quickly and inexpensively sequence a multitude of viruses and
identify mutations found in avirulent viruses compared to virulent progenitors. Despite the
ability to catalogue these mutations, further steps are necessary to determine the effect of these
mutations on virulence because, as seen with gC expression and the 132 bp repeats, correlation
does not equate causation, in which mutations that appear correlated with attenuation may not be
the causative factor for loss of virulence.
While NGS may provide a list of mutations for consideration, it is the role of
recombineering for generation of recombinant viruses that can provide the necessary bridge
between genotype and phenotype to determine the impact of mutations identified via NGS in
biological context of the MDV genome as a whole. With the generation of BAC-cloned MDV
9

viruses containing the complete MDV genome incorporated into a BAC-backbone, a multitude
of techniques are now available to allow the manipulation and mutation of MDV. Using Redmediated recombineering, it is now possible to introduce defined point mutations, whole gene
deletions and insertions to routinely be incorporated into BAC-cloned viruses to create
recombinant viruses testing the impact of various mutations of interest and yield near-isogenic
recombinant viruses containing only the desired alternations (27, 28).
A second benefit of BAC-cloned viruses is the ability to generate more homogenous viral
populations. Cloning MDV genomes to create BAC-derived viruses allow genetically
homogenous viral stocks comprised of one genotypically and phenotypically uniform virus of a
designated virulence level. Compared to BAC-cloned viruses, which are highly uniform viral
populations as all members are derived from the same transfected DNA molecule, MDV strains
are often described as quasi-species composed of a variety of diverse viral genotypes (26, 29,
30). While these unique genotypes are all present within the same strain and each MDV strain is
classified into categories based on their severity and virulence, not all members of the viral
population are necessarily phenotypically, much less genotypically, identical and share the same
virulence level (31), therefore, genetic homogeneity within strains have been shown to be an
important component of pathogenicity in herpesviruses (32, 33). Traditionally, MDV strains
have been classified as mild (m), virulent (v), very virulent (vv) or very virulent plus (vv+) based
on their phenotypes and ability to be controlled by the successive lines of vaccines developed,
yet it is not assured that any strain designated as one of the preceding virulence levels
exclusively contains viral subpopulations that all share the same virulence level. For example,
there is no limitation preventing a strain denoted overall as vv+ from also containing a mixture of
subpopulations with unique genotypes and phenotypes that differ from the vv+ designation, such
10

as vv or v, or even avirulent viruses (31). During cloning of the Md11 strain of MDV, several
unique genotypes were isolated as BAC clones, of which one had the entire terminal repeat short
region of the genome replaced with part of a duck chromosome, supporting the viewpoint that
MDV strains are often genotypically diverse collections of viruses only broadly described
collectively as one strain (29). This genetically diverse quasi-species aspect of MDV complicates
comparisons between parental and derived strains because, while there may be one consensus
genotype sequenced for the strain, there are subpopulations within the progenitor virus which
differ from the consensus sequence. This is a particular complication when trying to determine
the source for genetic variation observed in serially passed attenuated strain which have been
evolved from a virulent strain. During serial passage in vitro, an obvious phenotypic change
occurs from the virulent parental virus to the resulting avirulent attenuated viruses, but there are
two major forces that could account for that change; selection or de novo mutation. Due to the
quasi-species nature of MDV strains, avirulent MDV genotypes already present at low
frequencies within the parental virus could simply be selected for during repeated serial passage
and drive an increase in frequency within the population, eventually resulting in replacing the
majority of the virulent strain with avirulent sequences. As seen in previous studies, avirulent
BAC-cloned viruses have been isolated from a very virulent strain, this phenomenon has already
been observed and described (31).
Alternatively, attenuation could be driven by de novo mutation occurring within virulent
genotypes during serial passage as a result of errors during replication to generate new, avirulent
genotypes within the virulent strain. Use of a virulent MDV-BAC cloned virus assists by
ensuring a genotypically homogeneous viral population of only virulent viruses all sharing the
same BAC-derived genome, therefore, complications arising from passing viral quasi-species is
11

eliminated, as well as guaranteeing that any genetic variation observed during passage must have
resulted de novo during replication of the in vitro culture.
IDENTIFYING THE GENETIC BASIS OF ATTENUATION IN MDV
Due to the advantages offered by NGS technologies to inexpensively and quickly
sequence the complete genome of several species, the availability of BAC-cloned MDV viruses
allowing homogenous virulent viral stocks to be created, as well as recombineering technologies
for generation of recombinant viruses to validate candidate mutations, a variety of techniques are
now available to assist in exploring and elucidating the mechanisms and routes driving the
attenuation of MDV during in vitro serial passage. This will provide necessary information
regarding attenuation and provide a starting point for the rational design of future vaccines. With
a better understanding of the genetic mechanism driving attenuation it may be possible to
precisely design and engineer an avirulent virus for use as a candidate vaccine without relying on
blind serial passage in vitro to generate a more protective vaccine, especially considering that
current attempts to create more protective vaccines have failed to produce a vaccine significantly
more protective than Rispens (15).
In order to address this question and elucidate the genetic basis of attenuation of MDV at
the genetic level we had the following objectives:
Objective 1: Generate parallel, attenuated replicates of MDV, all initially derived from an
identical, virulent MDV.
1a: Produce attenuated viruses through serial passage in vitro.
A virulent MDV BAC clone known as Md5B40BAC-c1 (B40BAC) was used as the
initial homogeneous viral stock to infect chicken embryo fibroblast (CEF) cells. Three replicates
12

were grown in parallel and passed continuously for over 100 passages. Viral stocks of the three
replicates were saved after every 10 passages for testing in bird trials to determine at which
passage attenuation occurred.
1b: Determine at which passage each replicate became completely attenuated.
In vivo trials using the serially passed viral stocks to infect birds were conducted to
determine when attenuation first occurred in each of the three parallel lines. Based on resulted
generated in earlier studies it was predicted that the virulent MDV replicates would attenuated
within approximately 30 to 100 passages (16, 24, 34, 35).
Objective 2: Identify sequence variation and expression differences in attenuated strains
that appeared likely candidates responsible for creation of attenuated strains.
Sequencing both the DNA and mRNA from the lowest attenuated passage of the three
MDV replicates identified differences in both the sequence and expression of genes in attenuated
strains compared to the parental B40BAC. The frequency of these candidate mutations were then
tracked over the course of serial passage using viral stocks saved every 10 passages. Mutations
believed to be causative for attenuation were expected to be high frequency, nonsynonymous
mutations that were found within genes commonly mutated within the replicate attenuated
populations and whose frequencies increased during in vitro passage at similar timepoints that
corresponded to a decrease in virulence in vivo.
Objective 3: Experimentally validate candidate mutations which appear to correspond with
attenuation.
Mutations identified from the attenuated strains were used to replace the original virulent
forms in the parental Md5B40BAC clone. The resulting recombinant viruses generated were
13

tested in birds to determine their ability to cause MD, or if incorporation of candidate mutations
caused the engineered viruses to become avirulent. Recombinant viruses with an attenuated
phenotype would confirm that mutations integrated in the recombinant viruses were responsible
for changing the virus from a virulent to an attenuated form.
Objective 4: Further characterization of mutations shown to alter virulence

Mutations identified in Objective 3 shown to alter in vivo phenotypes of recombinant
MDVs were further characterized to describe in greater detail their in vitro and in vivo
characteristics. In particular, a point mutation within the UL5 helicase-primase subunit was
identified which reduced virulence by nearly 90% and was selected for further characterization.
4a: Comparison of in vitro replication of recombinant UL5 mutant virus to growth rates of the
virulent parental virus.

Mutations found at high frequencies in the attenuated viral populations were predicted to
have had a selective advantage for in vitro growth, driving their increase in frequency during the
course of serial passage. To determine if recombinant viruses resulting in reduced virulence in
vivo also had an effect on in vitro replication one step growth curves were conducted comparing
replication of the attenuated UL5 recombinant virus to the parental Md4B40BAC in vitro.
4b: Serial passage of the attenuated UL5 recombinant virus through birds to identify
restoration of virulence or increased in vivo replication due to adaptation during selection for
improved in vivo replication.
To determine how selective pressures within the bird may impact the point mutation
within the recombinant virus, leading to either reversion or compensatory mutations occurring to
14

restore virulence and negate the effects of the point mutation, the UL5 recombinant virus was
serially passed through birds while selecting for individuals with the highest replication levels in
vivo for re-isolation and passage of the virus in vitro.
4c: Determine vaccinal potential of UL5 recombinant viruses
The UL5 recombinant virus was tested as candidate vaccines to determine if mutations identified
using the method outlined in Objective 1 was able to identify SNVs which not only reduce
virulence, but also provided vaccinal protection as proof of concept for this method in assisting
in the future rational design of vaccines. The UL5 point mutation was tested for vaccinal
protection as a single point mutation in a recombinant virus as well as a recombinant virus
containing both the UL5 point mutation and deletion of the Meq oncogene, which is another
mutation proposed for use as a candidate vaccine.

15

REFERENCES

16

REFERENCES

1.
Plotkin SA, Orenstein WA, Offit PA. 2008. A short history of vaccination, p. 1–12. In
Vaccines. Elsevier Health Sciences.
2.
Riedel S. 2005. Edward Jenner and the history of smallpox and vaccination. Proc. Bayl.
Univ. Med. Cent. 18:21–25.
3.
Gross CP, Sepkowitz KA. 1998. The myth of the medical breakthrough: smallpox,
vaccination, and Jenner reconsidered. Int. J. Infect. Dis. IJID Off. Publ. Int. Soc. Infect. Dis.
3:54–60.
4.
121.

Willis NJ. 1997. Edward Jenner and the eradication of smallpox. Scott. Med. J. 42:118–

5.
Edwards KM, Halasa NB. 2007. Vaccines and Antisera for the Prevention and
Treatment of Infectious Diseases, p. 445–450. In Schaechter’s Mechanisms of Microbial
Disease, 4th ed.
6.
Wu X, Smith TG, Rupprecht CE. 2011. From brain passage to cell adaptation: the road
of human rabies vaccine development. Expert Rev. Vaccines 10:1597–1608.
7.
Hurisa B, Mengesha A, Newayesilassie B, Kerga S, Kebede G, Bankovisky D, Metlin
A, Urga K. 2013. Production of Cell Culture Based Anti- rabies Vaccine in Ethiopia. Procedia
Vaccinol. 7:2–7.
8.
K SA, V S, K P, M B. 2004. Diagnostic dilemma in flaccid paralysis following antirabies vaccine. Neurol. India 52:132.
9.

Rabies vaccines WHO position paper. Wkly. Epidemiol. Rec. WER 82:425–436.

10.
Plotkin SA, Plotkin SL. 2011. The development of vaccines: how the past led to the
future. Nat Rev Micro 9:889–893.
11.
Morrow C, Fehler F. 2004. Marek’s disease: a world-wide problem, p. 49–61. In
Davison, F, Nair, V (eds.), Marek’s disease. Academic Press, San Diego, CA.
12.
Perdue ML, Seal BS. 2000. Impact of Avian Viruses, p. 550–586. In Hurst, CJ (ed.),
Viral Ecology. Academic Press.
13.
Gimeno IM. 2008. Marek’s disease vaccines: A solution for today but a worry for
tomorrow? Vaccine 26, Supplement 3:C31–C41.
14.
Witter RL. 1997. Increased virulence of Marek’s disease virus field isolates. Avian Dis.
41:149–163.
15.
Witter RL, Kreager KS. 2004. Serotype 1 viruses modified by backpassage or
insertional mutagenesis: approaching the threshold of vaccine efficacy in Marek’s disease. Avian
Dis. 48:768–782.
17

16.
Churchill AE, Chubb RC, Baxendale W. 1969. The attenuation, with loss of
oncogenicity, of the herpes-type virus of Marek’s disease (strain HPRS-16) on passage in cell
culture. J. Gen. Virol. 4:557–564.
17.
Purchase H, Burmester B, Cunningham C. 1971. Responses of cell cultures from
various avian species to Marek’s disease virus and herpesvirus of turkeys. Am. J. Vet. Res.
32:1811–23.
18.
Biggs P, Payne L, Milne B, Churchill A, Chubb R, Powell D, Harris A. 1970. Field
trials with an attenuated cell associated vaccine for Marek’s disease. Vet. Rec. 87:704–709.
19.
Wilson MR, Southwick RA, Pulaski JT, Tieber VL, Hong Y, Coussens PM. 1994.
Molecular analysis of the glycoprotein C-negative phenotype of attenuated Marek’s disease
virus. Virology 199.
20.
Maotani K, Kanamori A, Ikuta K, Ueda S, Kato S, Hirai K. 1986. Amplification of a
tandem direct repeat within inverted repeats of Marek’s disease virus DNA during serial in vitro
passage. J. Virol. 58:657–660.
21.
Silva RF, Witter RL. 1985. Genomic expansion of Marek’s disease virus DNA is
associated with serial in vitro passage. J. Virol. 54:690–696.
22.
Silva RF, Reddy SM, Lupiani B. 2004. Expansion of a unique region in the Marek’s
disease virus genome occurs concomitantly with attenuation but is not sufficient to cause
attenuation. J. Virol. 78:733–740.
23.
Silva R, Gimeno I. 2007. Oncogenic Marek’s disease viruses lacking the 132 base pair
repeats can still be attenuated by serial in vitro cell culture passages - Springer. Virus Genes
34:87–90.
24.
Spatz SJ, Schat KA. 2011. Comparative genomic sequence analysis of the Marek’s
disease vaccine strain SB-1. Virus Genes 42:331–338.
25.
Spatz SJ, Petherbridge L, Zhao Y, Nair V. 2007. Comparative full-length sequence
analysis of oncogenic and vaccine (Rispens) strains of Marek’s disease virus. J. Gen. Virol.
88:1080–1096.
26.
Spatz SJ. 2010. Accumulation of attenuating mutations in varying proportions within a
high passage very virulent plus strain of Gallid herpesvirus type 2. Virus Res. 149:135–142.
27.
Tischer B, von Einem J, Kaufer B, Osterrrieder N. Two-step Red-mediated
recombination for versatile high-efficiency markerless DNA manipulation in Escherichia coli.
Biotechniques 40:191–197.
28.
Van Pijkeren J-P, Britton RA. 2012. High efficiency recombineering in lactic acid
bacteria. Nucleic Acids Res. 40:e76.
29.
Niikura M, Dodgson J, Cheng H. 2006. Direct evidence of host genome acquisition by
the alphaherpesvirus Marek’s disease virus. Arch. Virol. 151:537–549.

18

30.
Spatz SJ, Zhao Y, Petherbridge L, Smith LP, Baigent SJ, Nair V. 2007. Comparative
sequence analysis of a highly oncogenic but horizontal spread-defective clone of Marek’s disease
virus. Virus Genes 35:753–766.
31.
Spatz SJ, Smith LP, Baigent SJ, Petherbridge L, Nair V. 2011. Genotypic
characterization of two bacterial artificial chromosome clones derived from a single DNA source
of the very virulent gallid herpesvirus-2 strain C12/130. J. Gen. Virol. 92:1500–1507.
32.
Bower JR, Mao H, Durishin C, Rozenbom E, Detwiler M, Rempinski D, Karban
TL, Rosenthal KS. 1999. Intrastrain Variants of Herpes Simplex Virus Type 1 Isolated from a
Neonate with Fatal Disseminated Infection Differ in the ICP34.5 Gene, Glycoprotein Processing,
and Neuroinvasiveness. J. Virol. 73:3843–3853.
33.
Montgomery A, Centifanto Y. 1989. Heterogeneity within an HSV-1 wild-type strain
and its importance in pathogenesis. Proc. Soc. Exp. Biol. Med. Soc. Exp. Biol. Med. N. Y. N
191:362–369.
34.
Schat KA, Calnek BW, Fabricant J, Graham DL. 1985. Pathogenesis of infection with
attenuated Marek’s disease virus strains 1,2. Avian Pathol. 14:127–146.
35.
Dudnikova E, Vlasov A, Norkina S, Kireev D, Witter RL. 2009. Factors Influencing
The Attenuation of Serotype 1 Marek’s Disease Virus by Serial Cell Culture Passage and
Evaluation of Attenuated Strains for Protection and Replication. Avian Dis. 53:63–72.

19

CHAPTER 2. CHARACTERIZING THE MOLECULAR BASIS OF ATTENUATION OF
MAREKS DISEASE VIRUS VIA IN VITRO SERIAL PASSAGE IDENTIFIES DE NOVO
MUTATION IN THE HELICASE-PRIMASE SUBUNIT GENE UL5 AND OTHER
CANDIDATES ASSOCIATED WITH REDUCED VIRULENCE
Published Journal of Virology. Hildebrandt E, Dunn JR, Perumbakkam S,
Niikura M, Cheng HH. 2014 2014 Jun; 88(11): 6232-42.
ABSTRACT
Marek’s disease (MD) is a lymphoproliferative disease of chickens caused by the
oncogenic Gallid herpesvirus 2, commonly known as Marek’s disease virus (MDV). MD
vaccines, the primary control method, are often generated by repeated in vitro serial passage of
this highly cell-associated virus to attenuate virulent MDV strains. To understand the genetic
basis of attenuation, we used experimental evolution by serially passing three virulent MDV
replicates generated from an infectious bacterial artificial chromosome (BAC) clone. All
replicates became completely or highly attenuated indicating de novo mutation, and not selection
amongst quasi-species existing in a strain, is the primary driving force for the reduction in
virulence. Sequence analysis of the attenuated replicates revealed 41-95 single nucleotide
variants (SNVs) at 2% or higher frequency in each population, and several candidate genes
containing high frequency, nonsynonymous mutations. Five candidate mutations were
incorporated into recombinant viruses to determine their in vivo effect. SNVs within UL42
(DNA polymerase auxiliary subunit) and UL46 (tegument) had no measurable influence, while
two independent mutations in LORF2 (a gene of unknown function) improved survival time of
birds but did not alter disease incidence. A fifth SNV located within UL5 (helicase-primase
subunit) greatly reduced in vivo viral replication, increased survival time of birds, and resulted in
20

only 0-11% disease incidence. This study shows that multiple genes, often within pathways
involving DNA replication and transcriptional regulation, are involved in de novo attenuation of
MDV and provides targets for the rational design of future MD vaccines.
IMPORTANCE
Marek’s disease virus (MDV) is a very important pathogen in chickens that costs the
worldwide poultry industry $1-2 billion annually. Marek’s disease (MD) vaccines, the primary
control method, are often produced by passing virulent strains in cell culture until attenuated. To
understand this process, we identified all the changes in the viral genome that occurred during
repeated cell passage. We find that a single mutation in the UL5 gene, which encodes a viral
protein necessary for DNA replication, reduces disease incidence by 90% or more. In addition,
other candidate genes were identified. This information should lead to the development of more
effective and rationally-designed MD vaccines leading to improved animal health and welfare,
and lower costs to consumers.
INTRODUCTION
Marek’s disease virus (MDV, aka Gallid herpesvirus 2) is an oncogenic alphaherpesvirus
that causes Marek’s disease (MD) in chickens, which is characterized by T-cell lymphomas,
nerve lesions, and death in affected birds. MD costs the worldwide commercial poultry industry
$1-2 billion in losses annually (1), therefore, control of MD is vital for economic viability.
Vaccines have been successfully used to control MD since 1970 with the introduction of HVT, a
related turkey herpesvirus (2). Since then, additional vaccines with better efficacy have been
introduced to combat field strains that evolved to overcome existing vaccinal protection (3). This
continuous need to improve MD vaccines is likely driven by the fact that MD vaccines are non-

21

sterilizing and, thus, do not prevent vaccinated birds from becoming infected with virulent field
strains to collectively replicate and evolve within the same bird (4). Therefore, new vaccines
must be periodically developed and introduced to combat the evolution of more virulent strains
of MDV. Currently the most protective vaccine, known as Rispens or CVI988, is an attenuated
MDV strain developed from in vitro serial passage of a virulent virus until the resulting isolate
became avirulent (5).
The process of attenuation via in vitro serial passage has been used to successfully
generate candidate vaccines against MD since 1969 (6). Despite this widespread and wellestablished history of use, the underlying mechanism(s) behind attenuation remains poorly
understood. Expansion of 132 bp repeats within the inverted repeat regions flanking the unique
long region of the MDV genome has been consistently observed during serial passage of MDV
strains and, therefore, was once postulated to be a driving cause for attenuation. Despite the
ubiquitous nature of this 132 repeat region expansion, this phenomenon is correlated with, but
not causal for, attenuation because recombinant viruses completely lacking the 132 bp repeat
region are still virulent and able to become attenuated via serial passage in vitro (7). Recently,
next generation sequencing (NGS) has led to a massive expansion in sequencing power, allowing
characterization at the molecular genetic level of the complete genomes for many viruses and
strains, including serially passed MDV strains such as 648A (8) and several classical MD
vaccines including Rispens and SB-1 (9, 10). This increase in sequencing power allows
comparisons for identification of a range of polymorphisms among strains that might account for
differences in virulence, yet connecting genetic variation with phenotypic differences still
remains an additional and necessary step to understand attenuation at the genetic level.

22

To determine the genetic basis of attenuation, viral populations generated from the same
infectious Md5 BAC clone of MDV were serially passed in vitro to generate three attenuated
replicates. For each replicate, the lowest attenuated passage was sequenced to ~250-1,000x depth
of coverage via NGS. The resulting sequence information allowed us to accurately identify even
low frequency mutations and track their occurrence within the viral populations as they evolved
from the virulent parental virus and increased in frequency throughout the population.
Furthermore, an additional benefit of using a BAC-derived MDV, besides ensuring initially
uniform replicates before passage, is that candidate mutations identified in the serially passed
attenuated replicates could be incorporated via Red-mediated recombination into the cloned viral
genome allowing us to generate near-isogenic viruses. These recombinant viruses were then
tested in bird trials in order to measure the phenotypic effects of candidate mutations. This
approach led to the identification of several mutations that increased survival time of infected
birds, and, most notably, a mutation within UL5, the helicase-primase subunit, that reduced
MDV replication levels in vivo and lessened virulence, resulting in a reduction in disease
incidence by over 90%.

23

MATERIALS AND METHODS
Tissue Culture. Chicken embryo fibroblasts (CEFs) and duck embryo fibroblasts (DEFs) were
used to culture viruses. Cultures were maintained in 1:1 mixture of Leibovitz’s L-15 and
McCoy’s 5A (LM) media supplemented with fetal bovine serum (FBS), 200 U/ ml penicillin, 20
2

incubator. Cells were

plated with 4% FBS LM media and maintained in 1% FBS LM media. For storage as viral
stocks, infected cells were suspended in freezing media composed of 10% DMSO, 45% FBS,
and 45% LM media and kept in liquid nitrogen.

Viruses. A virulent Md5 BAC-derived virus stock was generated by transfecting Md5B40BAC,
an infectious pBeloBAC11 clone that contains the complete MDV Md5 genome (11), into CEF
cells. Md5B40BAC-derived virus clone 1 (Md5B40BAC-c1) at passage 4 (p4) was used as the
parental virus to infect three separate CEF plates with 200 plaque forming units (PFU) of virus.
These three replicates were designated Rep 1 p5, Rep 2 p5, and Rep 3 p5. The original, uncloned
Md5 strain p11 was also used to infect an additional plate of CEFs with 200 PFU of the virus and
passed as a positive control. Plates containing a confluent monolayer of CEFs in 4% LM media
were infected with the appropriate viral stock and after 24 hours, the media was changed and
maintained in 1% media. When cytopathic effects developed on the entire monolayer, the viruses
were passed by trypsinizing the monolayer and collecting infected cells in 1% LM media before
co-culturing infected cells with fresh CEFs. Ten percent of the total cells harvested were
transferred to a new plate containing a confluent monolayer of CEFs in 4% LM media to
complete one passage. This process was conducted repeatedly on the three Md5B40BAC-c1
replicates and the Md5 strain for 100 passages. At every 10 passages (i.e., p15, p25, p35, etc., for

24

BAC-derived reps and p21, p31, p41 etc., for Md5 strain), viral stocks were saved for in vivo
bird trials to determine virulence levels and identify the earliest passage at which attenuation
occurred in each replicate.

In Vivo Virulence Trials. In order to determine virulence, viral stocks were used to challenge
ADOL 15I5 x 71 maternal antibody negative chicks. A minimum of 17 day-old chicks were
challenged intra-abdominally with 500 PFU of the designated virus and housed for eight weeks
in Horsfall-Bauer (HB) units. Moribund birds, or those that survived up to eight weeks post
challenge, were terminated and examined via necropsy for signs of MD, including tumors and
nerve enlargement. All experiments were approved by the USDA, Avian Disease and Oncology
Laboratory Animal Care and Use Committee (ACUC). The ACUC guidelines established and
approved by the ADOL ACUC (April 2005) and the Guide for the Care and Use of Laboratory
Animals by the Institute for Laboratory Animal Research (2011) were followed throughout the
experiments.

Virulence Trials of Mixed Virulent and Avirulent MDV. Viral mixtures containing various
known quantities of virulent Md5B40BAC-c1 p5 and avirulent ∆MeqBAC, a recombinant virus
lacking both copies of the Meq (R-LORF7) oncogene (12), were used to infect birds. Mixtures
contained 5, 10, 25, or 50 PFUs of the virulent Md5B40BAC-c1 p5 were mixed with ∆MeqBAC
to yield a total of 500 PFU of virus to challenge each bird. An additional series of birds were
infected with the same quantity of Md5B40BAC-c1 p5 as before (5, 10, 25, 50 PFUs) but
without the addition of the ∆MeqBAC. Birds were challenged and disease incidence measured as
described previously.

25

Sequencing of Viral Stocks. To identify changes in both the genome and gene expression of the
attenuated replicates relative to the virulent progenitor virus, DNA and RNA of the lowest
attenuated passage for each replicate, as well as the Md5 strain p11 and Md5B40BAC-c1 p5
parental viruses, were sequenced using the Illumina GAIIx platform (San Diego, CA). DNA
from heavily infected plates was extracted and enriched for viral DNA using a micrococcal
nuclease protocol (13) while RNA was extracted using the Stratagene Absolutely RNA Kit
(Agilent Technologies; Santa Clara, CA). Library preparation and sequencing was conducted by
the Michigan State University Research Technology and Support Facility (www.rtsf.msu.edu).

Identification of Mutations. After processing, filtering, and trimming reads based on quality
scores, Illumina reads were mapped to the sequenced Md5B40BAC reference genome
(HQ149526.1) using BWA (14). Single nucleotide variants (SNVs) present in attenuated
replicates were called using VarScan (15) with a p<0.005. SNVs were identified from RNA
sequence data following the same procedure described above, with analysis for differential
expression utilizing BWA for mapping, followed by processing with Cufflinks (16) .

Tracking SNV Kinetics. Primers were designed to amplify eight regions of the MDV genome
containing 16 candidate SNVs identified via sequencing present at >20% in the attenuated
replicates (Table 2-1). Template DNA was obtained from viral stocks of the serially passed
replicates saved every 10 passages (from p15 to final attenuated passage of p65, p75 or p85
depending on replicate). SNV-containing regions were amplified using Phusion High-Fidelity
DNA Polymerase (New England Biolabs; Ipswich, MA) and the resulting amplicons barcoded

26

and pooled for Illumina GAIIx sequencing. Sequenced reads were processed as previously
described to determine SNV frequency at each passage level in the replicates.

Pathway Analysis of Mutated Genes. The Database for Annotation, Visualization and
Integrated Discovery (DAVID) (17, 18) was used to identify gene pathways that were enriched
for mutations in the attenuated viruses. A list of genes that contained mutations at 20% or greater
frequency in any of the four serially passed replicates (Rep 1-3 and Md5 strain) was submitted to
DAVID for analysis using the default parameters.

Generation of Defined Recombinant Viruses. Single point mutations in UL42, UL46, UL5,
and two different mutations within LORF2 (aka MDV012) were incorporated individually into
the parental BAC using two-step Red-mediated recombineering (19). Phusion High-Fidelity
DNA Polymerase (New England Biolabs) and primer sets (Table 2-2) were used to mutate
Md5B40BAC, yielding five independent recombinants: Mut UL42-D207G, Mut UL46-Q117R,
Mut UL5-I682R and two involving LORF2, referred to as LORF2-Promoter and LORF2-Intron.
Mutations in the BAC were verified in the bacterial stocks via Sanger sequencing before
transfecting purified BAC DNA into DEF cells using the calcium phosphate method (20) to
create viral stocks, which were again sequenced to confirm the desired mutation.

27

28

29

In Vivo Characterization of Mutant Viruses. Day-old, 15I5 x 71 maternal antibody negative
chicks were infected with 500 PFU of the recombinant MDV. Eighteen infected chicks were
housed in one HB isolator to measure MD incidence, while an additional 21 infected chicks were
housed separately in another HB isolator with six uninfected contact chicks to test for horizontal
transmission of recombinant viruses between birds. Additionally, one HB isolator containing
uninfected negative control birds and another containing positive controls challenged with 500
PFU of the parental Md5B40BAC-c1 virus were included for each trial. For each recombinant
virus, five birds were sacrificed at 7, 14, and 21 days post infection (dpi) to collect spleen
samples for qPCR. All surviving birds after 10 weeks post-infection were euthanized and
examined via necropsy. Recombinant viruses that showed altered disease incidence in
preliminary trials were used to challenge two more isolators each with a minimum of 17 day-old,
1515 x 71 maternal antibody negative chicks with 500 PFU of the virus, as previously described,
in order to replicate measures of disease incidence.

In Vivo Replication of MDV. DNA extracted from spleens of infected birds was used to assay
the in vivo replication levels of serially passed replicates and recombinant viruses relative to
control virulent Md5B40BAC-c1 p5. Extracted DNA was used to quantify the relative copy
number of MDV genomes present via qPCR using primers for chicken GAPDH and MDV gB
with the Taqman Fast Universal PCR kit (Applied Biosciences; Foster City, CA) as described by
Gimeno et al. (21).

30

RESULTS

Attenuation of Serially Passed Replicates
To assess the ability to identify de novo mutations in the MDV genome, three replicates
derived from the virulent Md5B40BAC-c1 virus and one from the Md5 parental strain (p11), the
strain originally used to clone the MDV genome, were serially passed in vitro. To determine the
earliest passage at which viral attenuation occurred, birds were challenged with the passed MDV
replicates and MD incidence measured. As expected, the genetically heterogeneous Md5 p11
strain became attenuated at p61, with no virulence observed after 50 serial passages (Fig. 2-1). In
addition, all passed replicates became attenuated, with complete attenuation occurring at passage
65 and 75 for Reps 2 and 3, respectively, while Rep 1 retained low levels of virulence (6% MD)
even at passage 85, indicating the rate of attenuation is a variable process even among initially
identical replicates (Fig. 2-1).

MD Incidence in Defined Mixtures
Defined mixtures containing known quantities of virulent Md5B40BAC-c1 p5 and
avirulent ∆MeqBAC were used to simulate intermediate time points of the serially passed
replicates, which were mixed populations composed of both attenuated and virulent viruses at
various frequencies. Additionally, equal PFUs of virulent Md5B40BAC-c1 p5 as used in the
mixed populations, but without addition of ∆MeqBAC, were used as a comparison to determine
if virulence was simply a matter of the raw quantity of virulent MDV used to infect a bird, or if
infectivity would be influenced by the presence of additional avirulent ∆MeqBAC virions.
Increasing the percentage of virulent virus within a total dose of 500 PFU increased the MD

31

induced by mixtures of Md5B40BAC-c1: ∆MeqBAC, as expected, but the same pattern did not
hold for birds infected with identical levels of Md5B40BAC-c1 alone (Fig. 2-2). For example,
birds infected with 5 PFU of the virulent Md5B40BAC-c1 along with 495 PFU of the avirulent
∆MeqBAC resulted in ~15% of the birds developing MD, yet birds challenged with 5 PFU of
Md5B40BAC-c1 by itself resulted in over 90% MD (Fig. 2-2). Thus, our test for virulence is
very sensitive, being virtually saturated at 5 PFU of Md5B40BAC-c1. While the quantity of
additional avirulent virus reduces disease incidence following challenge with a mixed viral
population, we still detect MD, even after infection with a population in which 1% of the PFU
derive from virulent MDV. Therefore, mutations that lead to attenuation in our samples should
exist in the population at substantial frequencies in order to explain the loss of virulence for the
population as a whole.

Next-Generation Sequencing Analysis of Attenuated Replicates
Based on the results of tests above, the completely attenuated Rep 2 p65, Rep 3 p75, and
Md5 strain p61, as well as the >90% attenuated Rep 1 at p75 and p85, were chosen for NGS to
identify and quantify mutations in the attenuated viral genomes. The parental Md5B40BAC-c1
p5 virus used to generate the three replicate strains was also sequenced to determine any
differences compared to the previously published Md5B40BAC reference sequence (GenBank
accession no. HQ149525.1). This screening also allowed us to identify mutations already present
in the progenitor virus prior to passage to eliminate pre-existing SNVs from consideration as
causative mutations for attenuation. Analysis of the data identified six SNVs fixed at 100% in the
sequenced Md5B40BAC-c1 p5 viral stocks that differed from the reference. Due to the presence
of these variations in the progenitor stock and, consequently, all resulting BAC-derived
32

replicates, these SNVs were excluded from further analysis. While the number of SNVs
identified in the sequenced viruses varied among replicates, 19-68% of called SNVs were present
at frequencies of less than 2% in the viral populations (Table 2-3). To screen for the optimal
candidate mutations involved in attenuation and eliminate false SNV calls due to sequencing
error, mutations present at less than 2% were excluded from further analysis; the standard error
of the frequency estimate was 0.9% and 0.4% for 250x and 1,000x depths of coverage,
respectively. The total number remaining ranged from 41-95 SNVs, depending on the replicate
(Table 2-4). These SNVs occurred within both coding and non-coding regions of the MDV
genome, and of those within coding regions, over 60% were nonsynonymous mutations.
Among attenuated replicates, eight identical nucleotide changes were present at
frequencies greater than 2% in at least two of the four attenuated viruses (data not shown). Five
of the eight identical nucleotide changes were instances in which one viral replicate had the
mutation at a high frequency while other replicates containing the same mutation were at low
frequencies <10%. Only three mutations were present at moderate or high frequencies in two or
more of the attenuated viral replicates. One mutation at moderate frequencies in two replicates
was present at nucleotide 171,661, which is located downstream of SORF2a. This mutation
occurred at frequencies of 22% and 27% in Rep 1 and Rep 3, respectively. Of the two high
frequency mutations, one was identified within the three regions of the a-like sequence at
positions 897, 140,705, and 183,749 occurring between 50-89% in the four attenuated replicates,
while the final point mutation at nucleotide 100,014 in UL42 was found at 84% and 85% in Rep
1 and Md5 p61, respectively. Due to the logistical complexity of incorporating mutations within
the three a-like regions present within herpesvirus genomes, that mutation was not considered for

33

creation as a recombinant virus, while the mutation within UL42 was deemed an excellent
candidate.

34

FIGURE 2-1. Virulence of MDV replicates derived from a serially passed, BAC-cloned virus.
Two bird trials (A and B) were conducted for Reps 1, 2, and 3 each at p55, p65, and p75, e.g.,
R1-A is Replicate 1, trial A at the indicated passages. In addition, the control strain Md5 was
tested at p41, 51, and p61.

35

FIGURE 2-2. Disease incidence in chickens challenged with defined mixtures of virulent and
avirulent MDV. Bird trials conducted in duplicate are indicated by “A” and “B,” e.g., 5:495-A
was trial A where the inoculation consisted of 5 PFU B40-derived MDV and 495 rMd5∆Meq.

36

37

38

Besides SNV mutations, we looked for changes in gene copy number variation reflected
by differences in the depth of sequenced coverage over the MDV genome of attenuated
replicates versus virulent virus. Previous studies of the MDV 132 bp repeats has shown this
region to greatly expand in copy number during serial passage, although this increase in copy
number is not causative for the loss of virulence (7). In the virulent Md5B40BAC-c1 p5 virus
before passage, the depth of coverage across the entire MDV genome was fairly consistent, with
the only exception being a larger number of reads mapping to the inverted repeat region
containing the a-like sequences and telomeric repeat regions-mTMR (Table 2-5). This expansion
involving the repetitive a-like sequence and mTMR was also present in the three attenuated
replicates. After serial passage, a 7-58x fold increase in depth of coverage corresponding to the
132 bp repeat regions was observed in the attenuated viruses, as would be expected (Table 2-5).
Candidate attenuation-causal genes were identified by screening for genes mutated more
than once among the serially passed populations. Seventy-four percent of genes containing
nonsynonymous mutations were mutated exclusively within only one of the three Md5B40BACc1 reps (Fig. 2-3). Five genes had nonsynonymous mutations in two of the three replicates: UL5
(helicase-primase helicase subunit), US8 (gE), US1 (ICP22), R-LORF4 (unknown function), and
UL46 (VP11/VP12). There were only two genes with missense mutations in all three replicates:
UL26 (VP24), which had an identical, but low frequency (<5%) mutation in all three replicates
at nucleotide 59,317, and RS1 (ICP4). A third gene, LORF2, also contained unique mutations
within all three BAC-derived replicates, although these mutations were not all nonsynonymous
changes affecting protein sequence. Generally, mutated genes contained only 1-2 unique
polymorphisms within a gene per replicate. Only ICP4 contained multiple, nonsynonymous
mutations in all replicates (Table 2-6). In the completely attenuated Reps 2 and 3, both replicates

39

had multiple, high-frequency, nonsynonymous ICP4 SNVs such as G62V (Rep 2, 83.8%),
T195A (Rep 2, 39.8%), Q63H (Rep 3, 100%), L256S (Rep 3, 65.9%), and S1630P (Rep 3,
85.7%). These mutations suggest ICP4 as a top candidate gene. The three Md5B40BAC-c1
replicates were not the only attenuated viruses containing high-frequency mutations within ICP4.
The serially passed Md5 strain also possessed the SNV Y60C at 93%. All fully attenuated virus
stocks except Rep 1 contained a high-frequency SNV in ICP4 within bases that encode amino
acids 60-63, including the completely fixed mutation in Rep 3 encoding amino acid 62,
suggesting that mutations within this region in particular may play an important role in
attenuation.
Following sequencing of DNA to identify mutations in the attenuated replicates, SNVs
were also called in the sequenced RNA isolated from infected plaques grown on CEF. While
SNVs were first called in genomic DNA data, those mutations may not necessarily be expressed
in mRNA, so the frequency of SNVs in DNA was compared to their frequency in RNA.
Frequencies of mutations first identified in DNA were extremely consistent and comparable to
their frequencies in RNA (Table 2-7). On the whole, SNV frequencies in RNA did not differ by
more than 5% from their measured frequency called in DNA. Of the top candidate mutations
identified, the sole exception to this trend was LORF2 in Rep 2 p65, which had the greatest
difference in SNV frequency in RNA vs. DNA of 8.8%. This provides confirmation of the
accuracy of the sequencing and variant calling, as well as indicating that there was no apparent
preferential expression of wild-type or mutated alleles within a viral population.
RNA data were then analyzed to determine if there was differential gene expression
between attenuated and virulent viruses. Comparing each attenuated replicate to the virulent
parental virus identified 5-14 genes that had significantly differential expression (Table 2-8).

40

Analyzing genes that are differentially expressed among the attenuated replicates showed one
gene, UL45 (envelopment protein), was differentially expressed in all attenuated replicates with
at least a 1.5-2.9 fold increase in all attenuated viruses. In addition to UL45, R-LORF2 (vIL-8)
was also differentially expressed in all replicates, although for Rep 1 there was only differential
expression at p85. Despite the differential expression of these genes, there were no mutations
within coding regions or nearby promoter regions of the genes themselves that occurred at
frequencies exceeding 20%.
Analysis for gene pathways that were enriched for mutations among the attenuated
replicates using DAVID only found one annotated cluster of functional genes that was enriched
(DAVID Enrichment Score 0.56) for the attenuated replicates. This cluster contained genes
involved in transcriptional regulation and regulation of RNA metabolic processes (GO:0006355
and GO:0051252), including the genes UL46, and ICP4 previously identified as candidate genes
for attenuation.

41

42

FIGURE 2-3. Genes with nonsynonymous mutations within attenuated MDV replicates. Serially
passed MDV replicates were compared to identify commonly mutated genes shared among
attenuated replicates.

43

44

45

46

Candidate Mutations and Characterization of Recombinant Viruses
Given that sequence analysis identified only a modest number of non-synonymous
mutations in genes and the fact that the isolates were either completely or very highly attenuated,
we further narrowed our focus to those mutations whose frequency exceeded 20% in the
attenuated populations. We conducted targeted resequencing of 16 candidate SNVs identified in
the attenuated replicates at consecutive 10 passage intervals from p15 until attenuation. This
allowed us to estimate when these mutations first occurred and correlate increases in candidate
SNV frequencies with decreasing virulence in vivo. Mutations that were at high frequencies in
the final, attenuated population and whose increase in frequency occurred roughly around
passages 55-65, when a drop in virulence of the serially passed populations was most
predominant when tested in vivo (Fig. 2-1), were identified as top candidates, particularly if they
were nonsynonymous mutations (Fig. 2-4). Of the candidate SNVs included in targeted
resequencing, five mutations were picked as the top candidates to first test as recombinant
viruses due to their location within either the unique long or short region of the MDV genome
(Fig. 2-4). Based on mutation frequencies and a concentration of high frequency nonsynonymous
mutations clustered around amino acids 60-63 in ICP4, ICP4 appears to be a top candidate for
involvement in attenuation. However, ICP4 was not included among the five mutations initially
tested using recombinant viruses due to its location within the repeat regions of the herpesvirus
genome, as additional recombineering steps would have been required for mutating both copies
of ICP4, instead of a single mutation necessary for genes located in unique regions. Despite this
complication, characterizing ICP4 mutations via recombinant viruses is currently underway.
These five mutations within four genes were incorporated into Md5B40BAC to generate
the following recombinant viruses: Mut UL42-D207G, Mut UL46-Q117R, Mut UL5-I682R, Mut

47

LORF2-Promoter, and Mut LORF2-Intron. The first three recombinant virus mutations altered
amino acids within the dsDNA-binding protein/DNA polymerase processivity subunit, the
VP11/VP12 tegument protein, and the helicase-primase subunit, respectively, while the two final
mutations both are non-coding mutations that may affect expression of a fourth gene of unknown
function, known as LORF2.
In vivo bird trials of the five recombinant viruses indicated that two of the mutations,
UL42-D207G and UL46-Q117R, had no effect on survival or MD incidence (Fig. 2-5 and Table
2-9). Three recombinant viruses, Mut LORF2-Promoter, Mut LORF2-Intron, and Mut UL5I682R all increased length of survival in birds challenged with the recombinant viruses relative
to the virulent Md5B40BAC-c1, but upon termination and examination via necropsy both
LORF2 mutants exhibited 100% MD incidence; though interestingly, Mut LORF2-Intron did not
appear to transmit horizontally. Only Mut UL5-I682R not only influenced survival time, but also
resulted in an almost complete reduction in virulence. Birds infected with Mut UL5-I682R
survived to the end of the 10 week experiment and upon termination, only 11% of infected birds
developed any symptoms of MD, in comparison to 100% of control Md5B40BAC-c1-infected
birds (Table 2-9). Further independent trials measuring disease incidence of Mut UL5-I682R
resulted in 0% (0/17 birds) and 6% (1/18 birds) of birds developing MD.
To determine if attenuated replicates and recombinant viruses retained the ability to
replicate in vivo or if loss of oncogenicity resulted from loss of viral replication in vivo, splenic
DNA from infected birds was extracted to measure MDV replication in birds using qPCR ratios
comparing copy numbers of viral gB DNA versus copies of chicken GAPDH genes averaged
over the three collection time points (Fig. 2-6). The average relative number of gB/GAPDH
copies in Md5B40BAC-c1 ranged from 0.14-0.33, based on four separate bird trials. The three

48

serially passed Md5B40BAC-c1 replicates exhibited about a 10-fold reduction in viral
replication, while the recombinant Mut UL5 I682R also had a ratio comparable to these
attenuated viruses of 0.016 (Fig. 2-6). The two virulent recombinant viruses, Mut UL42-D207G
and Mut UL46-Q117R, had ratios comparable to the virulent virus (Fig. 6). The LORF2
mutations, which resulted in greater survival for challenged birds yet still exhibited 100% MD
incidence, exhibited values of 0.13-0.14, which were towards the lower range of values observed
with Md5B40BAC-c1. Virulent viruses, such as Md5B40BAC-c1, Mut UL42-D207G, Mut
UL46-Q117R, and both Mut LORF2 recombinants, all exhibited significantly higher levels of
MDV than the three serially passed attenuated viruses or Mut UL5-I682R. While attenuated
viruses had low levels of replication compared to virulent viruses, their levels of replication were
higher than the background of uninfected controls, indicating they still replicated in vivo. Not
only was a low level of MDV detectable via qPCR in attenuated viruses, but viable viruses were
able to be reisolated from peripheral blood lymphocytes of infected, but MDV negative, birds
challenged with the attenuated viruses when plated on DEF cells to produce plaques (data not
shown), further supporting that these attenuated viruses replicated in vivo.
Based on the dramatic decrease in virulence due to the UL5 mutation, a revertant of Mut
UL5-I682R was created called Revt UL5-R682I. This revertant virus not only replicated at high
levels in vivo, but also restored levels of virulence comparable to that of the wild-type,
Md5B40BAC-c1 virus (Fig. 2-6 and Table 2-9).

49

FIGURE 2-4. Kinetics of candidate SNP frequencies over serial passage. Targeted resequencing
of candidate SNPs allowed for the frequency of each mutation to be determined from the earliest
passage in which the mutation occurred until attenuation.

50

FIGURE 2-5. Survival of birds challenged with recombinant MDVs. Mortality due to MD over
the course of the 10 weeks was determined for the five recombinant viruses and the virulent B40derived MDV positive control.

51

52

FIGURE 2-6. Comparison of in vivo replication for attenuated replicates and recombinant
viruses. Due to housing and chick availability, recombinant viruses were tested incrementally,
therefore positive (Md5B40-c1) and negative (uninfected) controls are matched to the respective
recombinant viruses tested collectively by matched symbols (i.e. *, ^, #, etc.). Trials for Mut
UL5 were conducted in duplicate and each trial indicated is by “A” or “B.”

53

DISCUSSION

This work shows attenuation of a genetically homogeneous MDV reproducibly occurs
after repeated serial passage in vitro, indicating that de novo mutations arising during passage are
sufficient to generate new, avirulent viral populations. A variety of mutations were identified
within each attenuated MDV replicate relative to the parental virulent virus, of which several
candidate mutations were found within genes of interest including ICP4 (transcriptional
transactivator), UL42 (DNA polymerase subunit), UL46 (VP11/VP12 tegument protein), UL5
(helicase-primase subunit), and LORF2 (unknown function). In particular, no single mutation
rose to high frequency in every attenuated replicate, nor was any single gene mutated at high
frequencies in every replicate. Two genes, ICP4 and UL26, were mutated in all attenuated
replicates but, at least in Rep 1, the frequencies of the respective mutations suggest that they are
not solely responsible for attenuation. Furthermore, we experimentally identified individual
mutations in UL5 and, to a lesser extent, in LORF2, that resulted in a phenotypic change that
contributed to attenuation, while mutations within UL5 have been shown to result in a reduction
in virulence in other herpesviruses, suggesting UL5 is a factor in virulence of herpesviruses (22,
23).
While attenuation of a virulent virus via serial passage in vitro is not a new or unusual
phenomenon unique to MDV, the use of virus generated from a BAC-cloned MDV in this study
allowed for a more detailed insight into the mechanism(s) of attenuation than previous studies (6,
8). When trying to understand the driving mechanism that causes a virulent viral population to
become avirulent, the two most probable explanations are selection or mutation. Like many
viruses, MDV strains, even those that are plaque purified, are often described as quasi-species
composed of genetically diverse genotypes existing within a strain population (8, 11). Thus,
54

while the viral strain may be described as a virulent, there can be unique subpopulations that do
not share the same degree of virulence compared to the whole population. Therefore, viral
attenuation could be driven by selection for pre-existing variation of low-frequency, avirulent
genotypes within a strain. A second explanation proposes de novo mutations that occur during
viral replication over serial passage generate new avirulent genotypes to attenuate the virus.
Therefore, by passing both a traditional virulent MDV strain and viruses from a BAC-derived
clone, we could distinguish whether attenuation requires pre-existing variation within a strain, or
if normal mutation rates occurring during serial passage are sufficient to generate avirulent
viruses de novo from the genotypically homogenous virus. Not only did the Md5 strain become
attenuated, as expected, but the three serially passed Md5B40BAC-c1 replicates attenuated as
well. This result shows, that despite beginning from a single virulent genotype, de novo mutation
is sufficient to generate avirulent viruses, and attenuation is not simply selection upon preexisting variation within quasi-species. However, as the Md5 strain became attenuated at earlier
passages, it is possible that selection of pre-existing variants in the population may also be a
contributing mechanism.
To better understand attenuation, we then compared duration of time before attenuation
between the serially passed replicates, and the first noticeable difference was that time until
attenuation varied between replicates. While a reduction in virulence occurred in all replicates,
attenuation did not occur simultaneously in all replicates despite being derived from identical
BAC-derived virus stocks before passage. Rep 2 was completely attenuated at p65, Rep 3 at p75,
and the virulence of Rep 1 was greatly reduced, but not completely attenuated, at p85. Despite
these differences, all replicates showed dramatic decreases in virulence with a 50-80% decrease
in MD incidence within a span of ~10 passages. This rapid change suggests a small number of

55

mutations of large affect are likely responsible for attenuation. In contrast, if attenuation was the
result of an accumulation of many small, additive mutations that progressively accrued over
time, we would have expected a slower, more gradual decline in virulence (24, 25).
Simply looking at the total number of mutations that occurred at >2% in the viral
population, no replicate had greater than 100 mutations with the number of SNVs in the three
replicates ranging from 41 to 95. While this is a relatively small number of mutations within a
genome of ~184 Kb, the actual number of candidate mutations is reduced even further to yield a
combined list of only 16 candidate SNVs from the three Md5B40BAC-c1 replicates when
considering only mutations occurring at high frequencies exceeding 20% in the population,
which is still a conservative cut-off value. As a result, we conclude that there are only a small
handful of probable candidate mutations occurring at high frequency to cause attenuation of the
population. Further filtering of these SNVs based on frequency kinetics in relation to decreasing
virulence identified five candidate mutations and genes underlying attenuation.
Further comparisons among the three attenuated replicates to find commonly mutated
genes identified ICP4 as a prime candidate gene. ICP4 contained several nonsynonymous
candidate SNVs at frequencies greater than 60% in the completely attenuated MDV replicates,
including one mutation completely fixed at 100% in a replicate. As ICP4 is an immediate early
transcriptional regulator that controls expression of other immediate early, early, and late genes,
it is reasonable that mutations altering normal ICP4 function could have a cascade affect, altering
regulation of many downstream genes. RNA-seq analysis of attenuated viruses identified 5-14
genes that were differentially expressed but when compared to the list of SNVs identified there
were no mutations within the genes themselves to explain their differential expression. Instead,
mutations within upstream regulators, such as ICP4, could explain this observation in which

56

mutations within ICP4 could lead to larger, widespread effects within the population resulting
from relatively few mutations occurring within ICP4.
The results of in vivo trials of recombinant viruses involving a single SNV within UL5
encoding the helicase-primase subunit showed that a single mutation can have a substantial
impact on the phenotype of recombinant viruses. Unlike the other four recombinant viruses that
retained 100% disease incidence in vivo, the UL5 I682R mutation reduced MD incidence by at
least 89% or more, depending on the trial. This mutant virus also had much lower levels of viral
replication in vivo compared to the virulent Md5B40BAC-c1. All serially passed Reps had
comparable levels of MDV replication in vivo, as did the Mut UL5-I682R. As a result of this
dramatic phenotypic change due to the mutation within UL5, we created a revertant virus to
verify that the results of in vivo trials could be attributed to the desired single point mutation and
not extraneous changes that may have occurred unbeknownst to us during recombineering and
generation of viral stocks. Challenging birds with this virus clearly showed the reversion restored
a wild-type phenotype for survival, levels of in vivo replication, and disease incidence,
confirming our conclusion that the single point mutation within Mut UL5-I682R was responsible
for the observed phenotypic changes. While this single mutation did not cause complete
attenuation of the virus, it does show that even a single nucleotide change is capable of causing
dramatic changes in virulence. This result clearly warrants further study to determine if
introduction of additional candidate mutations, such as the next-highest frequency candidate
mutation within UL46 (VP11/VP12 tegument protein) identified within the same replicate, is
able to have an additive or epistatic influence to cause complete attenuation in a recombinant
virus.

57

This dramatic decrease in disease incidence exhibited by the mutation within UL5helicase-primase within this study indicates the importance of this gene as a factor in virulence.
Sequence comparison of Gallid herpesvirus-1, an alphaherpesvirus also known as infectious
laryngotracheitis virus (ILTV), identified a nonsynonymous mutation within UL5 that was
unique to the vaccine strain and enabled the attenuated virus to be distinguished from four other
virulent ILTV strains (26). Mutations within UL5 have been shown to cause a reduction in
virulence in herpesviruses other than Gallid herpesviruses as well. Studies involving HSV-1 have
identified several mutations within UL5 that have been determined to cause a reduction in
virulence (27). Biochemical analysis of the UL5 Gly 815 mutant shows a decrease in binding
affinity of single-stranded DNA and reduced turnover rate (28). This mutation in HSV-1 is
unable to replicate in vivo alone, but when additional purified components of the heterotrimeric
complex are provided, the mutant was able to perform normal helicase-primase functions. While
the three components of the functional complex, UL5, UL8, and UL52, are all necessary for
normal function in vivo, during in vitro replication UL8 is not. Growth curves measuring the in
vitro replication for our Mut I682R relative to the Md5B40BAC-c1 are yet to be conducted, but
it is clear the recombinant Mut UL5 replicates in vitro to form plaques and generates high viral
titers. Despite this ability to replicate in vitro, its ability to replicate in vivo is severely reduced
compared to the virulent wild-type, which has been observed in other UL5 mutants. Since serial
passage of MDV in tissue culture drives selection for improved in vitro replication, it would be
anticipated that the mutation within UL5 would not detrimentally affect in vitro replication or
even, as a result of selection for improved in vitro growth, the observed Mut I682R could
potentially improve in vitro growth. Whether in vitro replication is altered by the UL5 mutation
compared to the wild-type BAC, growth curves quantifying replication of the Mut UL5-I682R

58

virus may offer an explanation for how this mutation rose to a high frequency in the attenuated
populations.
While this high frequency SNV in UL5 resulting in greatly reduced virulence was only
present in Rep 2, three additional mutations within UL5 were also found within Rep 1, although
these SNVs were present at much lower frequencies than the mutation found within Rep 2. As
seen in ICP4, it appears that identical mutations shared among replicates are rare, but there were
several commonly mutated genes between attenuated replicates, such as with ICP4, UL5, and
LORF2. Experimental evolution studies involving over 100 replicates of E. coli have shown high
degrees of convergent evolution between replicates at the level of genes and pathways, while
identical nucleotide mutations were rarely shared among the replicates (29). Based on these
conclusions, we looked for commonly mutated pathways among attenuated replicates to
determine if there were key pathways involved in attenuation. After screening all mutations for
the most probable candidates we identified five top candidate genes, of which two of the five
candidates were genes involved in DNA replication; a mutation within UL5 and a second one
within UL42. Considering that mutations within genes involving DNA replication are probable
targets for selection to cause adaptation of viruses during in vitro growth, wider inspection of all
mutations within attenuated replicates of the seven genes required for DNA replication in
herpesviruses (UL5, UL8, UL9, UL29, UL30, UL42 and UL52) revealed multiple mutations
within genes involved in DNA replication, especially within Rep 1. Of these genes, we found
eight mutations at frequencies >2% within all the seven genes required for DNA replication
within Rep 1, p85. Clustering of mutations within DNA replication genes, such as the UL5
(helicase-primase) and UL42 (DNA polymerase subunit), as well as the significant phenotypic
effect of the mutation tested within UL5, suggests that mutations in the genes associated with the

59

DNA replication pathway are targeted for selection during serial passage leading to attenuation
of the virus. To further identify commonly mutated pathways, we used DAVID to determine
pathways that were enriched for mutants in the attenuated replicates. Considering only high
frequency, nonsynonymous mutations at >20% in the four attenuated viral populations, we found
pathways containing genes involving transcriptional regulation were enriched for mutations. It is
possible that pathways involving transcriptional regulation, at both the DNA-dependent and
RNA metabolism levels, are mechanisms which, when disturbed, can lead to attenuation. By
finding enrichment for pathways involving transcriptional regulation due to multiple mutations
within genes involved in DNA replication, such as the Mut I682R in UL5, supports our
hypothesis that serial passage in vitro selects for increased replication and adaptation for growth
in tissue culture. Mutations altering normal interactions of these gene products would impact
viral replication and lead to the changes in phenotype observed in attenuated viruses.
In conclusion, our results show that it is possible to attenuate viral BAC-derived clones as
a result of de novo mutation, and that it is apparent there is variation among viruses in the
manner of how they become attenuated. All serially passed replicates contained unique candidate
mutations and, while there may be commonly mutated genes among attenuated viruses such as
ICP4, it appears that there is no single route that all viruses must follow to become attenuated.
Multiple candidate mutations have been identified and tested to determine their effect within
genes of known and unknown function such as LORF2, DNA polymerase subunit, VP11/VP12
tegument protein and helicase-primase subunits. Singly, no one mutation was able to cause
complete attenuation, yet one SNV within the helicase-primase subunit had a dramatic decrease
in virulence indicating an important role for this gene in MDV attenuation. Additional work to
further characterize candidate mutations identified in this study may identify a minimum number

60

of mutations required to create a fully avirulent virus through changing only a small handful,
perhaps 2-3 key mutations. This could allow the precise engineering of avirulent viruses for
production of candidate vaccines via mutation of critical genes involved in attenuation, without
relying on blind serial passage and random chance to generate the next superior MD vaccine
candidate.
ACKNOWLEDGEMENTS
We thank Laurie Molitor and Lonnie Milam for technical assistance and Jerry Dodgson for
critical input in this manuscript. This project was supported in part by National Research
Initiative Competitive Grant no. 2010-65119-20505 from the USDA National Institute of Food
and Agriculture.

61

REFERENCES

62

REFERENCES

1.

Morrow C, Fehler F. 2004. Marek’s disease: a world-wide problem, p 49-61. In Davison
F, Nair V (ed), Marek’s disease. Academic Press, San Diego, CA.

2.

Purchase H, Okazaki W. 1971. Effect of vaccination with herpesvirus of turkeys (HVT)
on horizontal spread of Marek’s disease herpesvirus. Avian Dis. 15:391–397.

3.

Witter R. 1997. Increased virulence of Marek’s disease virus field isolates. Avian Dis.
41:149–163.

4.

Atkins KE, Read AF, Savill NJ, Renz KG, Islam AF, Walkden-Brown SW,
Woolhouse ME. 2013. Vaccination and reduced cohort duration can drive virulence
evolution: Marek’s disease virus and industrialized agriculture. Evolution 67:851–60.

5.

Rispens BH, Vloten H, Mastenbroek N, Maas HJ, Schat KA. 1972. Control of Marek's
disease in the Netherlands. 1. Isolation of an avirulent Marek's disease virus (strain
CVI988) and its use in laboratory vaccination trials. Avian Dis. 16:108–125.

6.

Churchill AE, Chubb RC, Baxendale W. 1969. The attenuation, with loss of
oncogenicity, of the herpes-type virus of Marek’s disease (strain HPRS-16) on passage in
cell culture. J. Gen. Virol. 4:557–564.

7.

Silva RF, Reddy SM, Lupiani B. 2004. Expansion of a unique region in the Marek’s
disease virus genome occurs concomitantly with attenuation but is not sufficient to cause
attenuation. J. Virol. 78:733–740.

8.

Spatz SJ. 2010. Accumulation of attenuating mutations in varying proportions within a
high passage very virulent plus strain of Gallid herpesvirus type 2. Virus Res. 149:135–42.

9.

Spatz SJ, Petherbridge L, Zhao Y, Nair V. 2007. Comparative full-length sequence
analysis of oncogenic and vaccine (Rispens) strains of Marek’s disease virus. J. Gen.
Virol. 88:1080–96.

10.

Spatz SJ, Schat KA. 2011. Comparative genomic sequence analysis of the Marek’s
disease vaccine strain SB-1. Virus Genes 42:331–338.

11.

Niikura M, Kim T, Silva RF, Dodgson J, Cheng HH. 2011. Virulent Marek’s disease
virus generated from infectious bacterial artificial chromosome clones with complete
DNA sequence and the implication of viral genetic homogeneity in pathogenesis. J. Gen.
Virol. 92:598–607.

63

12.
13.

Silva RF, Dunn JR, Cheng HH, Niikura M. 2010. A MEQ-deleted Marek’s disease
virus cloned as a bacterial artificial chromosome is a highly efficacious vaccine. Avian
Dis. 54:862–869.
Volkening JD, Spatz SJ. 2009. Purification of DNA from the cell-associated herpesvirus
Marek’s disease virus for 454 pyrosequencing using micrococcal nuclease digestion and
polyethylene glycol precipitation. J. Virol. Methods 157:55–61.

14.

Li H, Durbin R. 2009. Fast and accurate short read alignment with Burrows-Wheeler
transform. Bioinformatics 25:1754–60.

15.

Koboldt DC, Chen K, Wylie T, Larson DE, McLellan MD, Mardis ER, Weinstock
GM, Wilson RK, Ding L. 2009. VarScan: variant detection in massively parallel
sequencing of individual and pooled samples. Bioinformatics 25:2283–5.

16.

Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg
SL, Wold BJ, Pachter L. 2010. Transcript assembly and quantification by RNA-Seq
reveals unannotated transcripts and isoform switching during cell differentiation. Nat.
Biotech. 28:511–515.

17.

Huang BDW, Sherman BT, Lempicki RA. 2009. Systematic and integrative analysis of
large gene lists using DAVID bioinformatics resources. Nature Protoc. 4:44–57.

18.

Huang DW, Sherman BT, Lempicki RA. 2009. Bioinformatics enrichment tools: paths
toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res.
37:1–13.

19.

Tischer BK, von Einem JV, Kaufer B, Osterrieder N. Two-step Red-mediated
recombination for versatile high-efficiency markerless DNA manipulation in Escherichia
coli. Biotechiques 40:191-197.

20.

Moriuchi H, Moriuchi M, Smith HA, Straus SE, Cohen JI. 1995. Varicella-zoster virus
open reading frame 61 protein is functionally homologous to herpes simplex virus type 1
ICP0. J. Virol. 66:7303-7308.

21.

Gimeno IM, Cortes AL, Silva RF. 2008. Load of challenge Marek’s disease virus DNA
in blood as a criterion for early diagnosis of Marek's disease tumors. Avian Dis. 52:203–
208.

22.

Biswas N, Weller SK. 2001. The UL5 and UL52 subunits of the herpes simplex virus
type 1 helicase-primase subcomplex exhibit a complex interdependence for DNA binding.
J. Biol. Chem. 276:17610–17619.

23.

Jarosinski KW, Osterrieder N, Nair VK, Schat KA. 2005. Attenuation of Marek’s
disease virus by deletion of open reading frame RLORF4 but not RLORF5a. J. Virol.
79:11647–11659.

64

24.

Barrick JE, Lenski RE. 2013. Genome dynamics during experimental evolution. Nature
Rev. Genet. 14:827–839.

25.

Elena SF, Lenski RE. 2003. Evolution experiments with microorganisms: the dynamics
and genetic bases of adaptation. Nature Rev. Genet. 4:457–469.

26.

Spatz SJ, Volkening JD, Keeler CL, Kutish GF, Riblet SM, Boettger CM, Clark KF,
Zsak L, Afonso CL, Mundt ES, Rock DL, Garcia M. 2012. Comparative full genome
analysis of four infectious laryngotracheitis virus (Gallid herpesvirus-1) virulent isolates
from the United States. Virus Genes 44:273-285.

27.

Biswas S, Miguel RN, Sukla S, Field HJ. 2009. A mutation in helicase motif IV of
herpes simplex virus type 1 UL5 that results in reduced growth in vitro and lower
virulence in a murine infection model is related to the predicted helicase structure. J. Gen.
Virol. 90:1937–1942.

28.

Graves-Woodward KL, Weller SK. 1996. Replacement of gly815 in helicase motif V
alters the single-stranded DNA-dependent ATPase activity of the herpes simplex virus
type 1 helicase-primase. J. Biol. Chem. 271:13629–13635.

29.

Tenaillon O, Rodríguez-Verdugo A, Gaut RL, McDonald P, Bennett AF, Long AD,
Gaut BS. 2013. The molecular diversity of adaptive convergence. Science 335:457–460.

65

CHAPTER 3. MUTATIONS WITHIN ICP4 ACQUIRED DURING IN VITRO
ATTENUATION DO NOT ALTER VIRULENCE OF RECOMBINANT MAREK’S DISEASE
VIRUSES IN VIVO
ABSTRACT
Marek’s disease (MD) is a T-cell lymphoma of chickens caused by the oncogenic
Marek’s disease virus (MDV). MD is primarily controlled by live-attenuated vaccines generated
by repeated in vitro serial passage. Previous efforts to characterize attenuated MDVs identified
numerous mutations, particularly a convergence of high-frequency mutations around amino acids
60-63 within ICP4 (RS1). Therefore, ICP4 was considered a candidate gene deserving further
characterization. Recombinant MDVs were generated containing a single Q63H mutation or
double Q63H+S1630P mutations. Despite the repetitive nature of mutations within ICP4 in
attenuated viruses, neither recombinant virus decreased virulence, although one mutant reduced
in vivo replication and failed to transmit horizontally. Our results indicate these mutations are
insufficient to reduce disease incidence in infected birds and suggest variants in ICP4 do not
directly alter virulence, but rather may enhance MDV replication rates in vitro, offering an
explanation for the widespread occurrence of ICP4 mutations in a variety of attenuated
herpesviruses.

66

INTRODUCTION
Gallid herpesvirus 2, also commonly known as Marek’s disease virus (MDV), is an
oncogenic alphaherpesvirus of chickens. Afflicted birds display symptoms of Marek’s disease
(MD) including depression, cachexia, and paralysis due to viral-induced T cell lymphomas that
ultimately result in death. The primary mode of MD control has been via the use of vaccines,
which has significantly reduced losses due to MD since the 1970’s (1). Despite the widespread
use of vaccinal control of MD, virulent strains of MDV able to overcome vaccines have
emerged, leading to the need for periodic introduction of new and more effective MD vaccines
approximately every 10 years (2). Currently the most protective vaccine commercially available,
CVI988/Rispens, has been in use in the United States since the 1990’s. Concern regarding the
potential for MD outbreaks in Rispens-vaccinated flocks has highlighted the importance for
vaccine development (3). One common method for vaccine production has been the use of live
attenuated viruses generated via repeated in vitro serial passage. Several MD vaccines, such as
HPRSS-16 and Rispens, were generated by in vitro serial passage and provide testament to the
utility of this process in vaccine development (4, 5).
To better understand this process and identify candidate genes involved in attenuation,
previously we serially passed and sequenced the complete genome of four attenuated MDV
replicates (three Md5BAC-derived viruses and one Md5 strain) to identify candidate mutations
and genes involved in attenuation (6). Several candidate mutations present within genes in the
unique long (UL) or unique short (US) regions of the viral genome were characterized via
recombinant viruses. Among the five point mutations tested, one recombinant virus revealed a
single nucleotide mutation in UL5 (helicase-primase subunit) able to reduce disease incidence by
90% or more. Due to the additional complexity involved in mutating genes present within the

67

two long and short repeat regions (TRL/IRL and TRS/IRS) in MDV, these candidate mutations
were not initially tested.
Therefore, the purpose of this study was the characterization of a top candidate gene,
ICP4, identified within the repeat regions of the MDV genome. ICP4, encoded by RS1, is an
immediate early transcriptional regulator in herpesviruses located within the TRS/IRS region of
MDV and a gene commonly mutated in all of the attenuated replicates. All four passed viruses
had mutations within ICP4, with three viruses containing high frequency (80-100%)
nonsynonymous mutations within amino acids 60-63 in ICP4. Three other high frequency (40%
or higher) nonsynonymous mutations in ICP4 were also observed, including one at amino acid
position 1630 (85% frequency). Due to these numerous high frequency, parallel mutations within
the attenuated replicates, ICP4 was considered a candidate gene for attenuation. Using Redmediated recombineering, we generated two recombinant viruses to examine how these
mutations in ICP4 impacted virulence of MDV.

68

MATERIALS AND METHODS
Recombinant Viruses and Tissue Culture. ICP4 mutations were incorporated into
Md5B40BAC, the pBeloBAC11 clone containing the entire MDV (Md5 strain) genome, using
Red-mediated recombineering (7–9). To generate recombinant viruses containing the desired
single nucleotide variant (SNV) within both copies of ICP4, one copy of ICP4 was first deleted
using the ∆ICP4 primer set (Table 3-1) based on the process of manipulating genes within repeat
regions described by Engel et al.(10). Point mutations were then incorporated within the
remaining copy of ICP4 via additional rounds of recombineering using the primer sets Mut ICP4
Q63H and Mut ICP4 S1630P. The resulting mutants, designated as Mut ICP4-1 and Mut ICP4-2,
both contained the mutation Q63H, while the latter also had the additional SNV S1630P. The
Q63H mutation was fixed 100% in one of the serially passed attenuated replicates during our
previous study (6), while the S1630P mutation was the second highest frequency ICP4 mutation
within the same replicate present at 85%. All introduced mutations were confirmed via Sanger
sequencing of mutated BACs before transfecting BAC DNA into duck embryo fibroblasts (DEF)
via the calcium phosphate method (11). Additionally, mutations in the generated viral stocks
were confirmed by Sanger sequencing. Viral stocks were then titrated on DEF cells to quantify
the number of plaque forming units (PFU) in frozen viral stocks.

69

70

In Vivo Characterization of Mutant Viruses. Maternal antibody negative, ADOL 15I5 x 71 day
old chicks were challenged intraabdominally with 500 PFU of each recombinant MDV or the
original Md5B40BAC-c1 as a positive control, and housed in Horsfall-Bauer (HB) isolators for
12 weeks. Uninfected birds were housed as negative controls in a separate isolator. All surviving
birds, or those that were moribund during the course of the experiment, were terminated and
examined via necropsy to determine disease incidence. To determine horizontal transmission,
uninfected contact birds were housed in the same HB unit with challenged birds. Five infected
birds were removed for sampling at 7, 14 and 21 days post infection (dpi) to collect spleen
tissues to quantify in vivo replication of viruses via qPCR comparing the relative quantity of the
MDV gB gene to chicken GAPDH gene using a Taqman Fast Universal PCR kit (Applied
Biosciences; Foster City, CA) and primers described by Gimeno et al. (12). All in vivo bird trials
were approved by the USDA, Avian Disease and Oncology Laboratory Animal Care and Use
Committee (ACUC). Based on ACUC guidelines established and approved by the ADOL ACUC
(April 2005) and the Guide for the Care and Use of Laboratory Animals by the Institute for
Laboratory Animal Research (2011), procedures for bird housing and care were followed
throughout the duration of the experiment.

Sequence Comparison of ICP4 among Alphaherpesviruses
The complete amino acid sequence of the immediate early transcriptional regulator gene ICP4
from 20 different species of alphaherpesviruses was obtained from the NCBI database. All 20
sequences were aligned using ClustalW (13) and a maximum likelihood phylogenic tree
constructed in Mega6 (14), while pairwise comparisons for each of the 20 ICP4 sequences
relative to herpes simplex virus 1 (HSV-1) ICP4 (GenBank accession no. AGZ01922.1) were
71

aligned via the UniProt protein resources database (15) to calculate the percentage of shared
sequence identity among ICP4 sequences.
RESULTS
In Vivo Trials
Disease incidence resulting from challenge with either Mut ICP4-1 (single ICP4 Q63H
SNV) or Mut ICP4-2 (double Q63H and S1630P SNVs) were 100% and 95%, respectively.
Compared to the parental virus that showed 84% MD incidence, these mutations did not reduce
virulence of the viruses in infected birds (Table 3-2). While most challenged birds developed
MD, the disease incidence in contact birds varied significantly between the two recombinant
ICP4 viruses. Contact birds housed with the Mut ICP4-1 virus all developed MD, yet no contact
birds mixed with birds challenged with the Mut ICP4-2 recombinant virus developed MD (Table
3-2).
While both recombinant viruses resulted in over 90% disease incidence, there was
delayed mortality for birds challenged with the double Mut ICP4-2 recombinant virus compared
to Mut ICP4-1 and Md5B40BAC-c1. Birds challenged with the single Mut ICP4-1 recombinant
virus experienced survival comparable to the original Md5B40BAC-c1 (Fig. 3-1). On the other
hand, the double Mut ICP4-2 virus exhibited lower mortality from week 4 onwards (Fig. 3-1),
with approximately 20% of birds surviving the full length of the 12 week experiment compared
to birds challenged with either Mut ICP4-1 or Md5B40BAC-c1, which all exhibited severe
morbidity and euthanasia of all birds by week 10.
Replication of MDV within the spleen of birds sacrificed at 7, 14 and 21 dpi as
determined by qPCR showed a range among the three viruses (Fig. 3-2). Compared to Mut ICP41 and the wild-type Md5B40BAC virus, the double mutant Mut ICP4-2 recombinant virus
72

showed lower levels of in vivo replication at 7 and 14 dpi, with significant differences in
replication at day 14 (ANOVA p-value=0.0136), though this difference was not observed at 21
dpi. Further pairwise comparisons of in vivo replication at day 14 showed significant differences
in replication between Mut ICP4-1 and Md5B40BAC-c1, with differences between Mut ICP4-1
and Md5B40BAC-c1 falling short the threshold for significance (T-test p= 0.0101 and p=0.0557,
respectively)

73

74

FIGURE 3-1. Survival curves of birds challenged with ICP4 recombinant MDV. Mortality of
birds during the 12 week experiment until termination was compared for the two ICP4 mutant
viruses relative to the parental Md5B40BAC virus.

75

FIGURE 3-2. Viral replication rates of birds challenged with ICP4 recombinant MDV. In vivo
replication of recombinant ICP4 MDV versus parental Md5B40BAC virus.

76

Homology of Alphaherpesviruses ICP4 Protein Sequence
Comparing the prototypical alphaherpesvirus (HSV-1) to the 20 alphaherpesviruses
sequences via pairwise alignment of HSV-1 to each respective species sequence showed a high
degree of variation among ICP4 between species. Amino acid sequence identity for ICP4 of the
20 species surveyed relative to HSV-1 ranged from approximately 15-71%, with an average
shared sequence identity of 36% of amino acids identical among the query species and HSV-1
(Table 3-3). Alignment of HSV-1 to Mardivirus sequences show an average sequence identity of
17% among the three Mardivirus species, versus an average shared sequence identity of 39% for
HSV-1 alignments with non-Mardivirus species. Alignment of HSV-1 ICP4 and the virulent
MDV ICP4 sequences revealed the lowest degree of shared sequence identify of approximately
15%.
Alignment of all 20 ICP4 sequences showed a high degree of variation among
alphaherpesviruses surveyed, not only in coding sequence but also in total length. The average
length of ICP4 in the 20 alphaherpesviruses surveyed was approximately 1,500 amino acids.
However, the average length of ICP4 in Mardivirus species infecting chicken (Gallid herpesvirus
2, Gallid herpesvirus 3 and Meleagrid herpesvirus 1) revealed significantly larger (2,173 amino
acids) ICP4 proteins, with Gallid herpesvirus 2 having the longest sequence of any herpesvirus
species surveyed at 2,323 amino acids (Table 3-3). The majority of these nearly 700 additional
amino acids are found at the N terminus of ICP4 in Mardiviruses. Clustal alignment of the 20
herpesvirus species show that for the first 500 amino acids of MDV ICP4, the only species with
homologous amino acids to this region are Gallid herpesvirus 3 and Meleagrid herpesvirus 1,
both of which also belong to the Maridvirus genus (Fig. 3-3). Furthermore, the cluster of high
frequency mutations within ICP4 identified in the attenuated MDV replicates around amino acids

77

60-63 occurs within this additional region that is unique relative to other alphaherpesviruses.
Even among Mardiviruses that possess a longer ICP4, the region around amino acids 52-69 is
exclusively found only in the virulent Gallid herpesvirus 2 species, preventing functional
predictions for this mutated region based on homology to other species (Fig. 3-3).
Alignment of ICP4 sequences showed the regions flanking the second mutation at amino
acid 1630 are present in the majority of herpesvirus species, yet among those amino acids
present, there is considerable variation in amino acids present at and around this position (Fig. 34). The MDV wild type serine is an amino acid seen in other herpesvirus species, such as Gallid
herpesvirus 1 and both HSV-1 and HSV-2. The mutation of a serine to proline residue at position
1630 is predicted to be a tolerated change by programs such as SIFT (16), as well an
evolutionarily acceptable amino acid considering that proline is commonly seen in several
herpesvirus species, such as Equid herpesvirus 1, Suid herpesvirus 1, Psittacid herpesvirus 1, and
the closely related Melaeagrid herpesvirus 1 (Fig. 3-4). Based on the alignment to HSV-1, the
MDV S1630P ICP4 mutation corresponds to roughly the same serine residue at position 683 in
HSV-1. This position is expected to be a globular region within the third defined region of ICP4
in HSV-1. This position is also characterized as occurring within a functional domain implicated
for nuclear localization (17), yet the corresponding amino acid regions within MDV ICP4 do not
appear to contain a known nuclear localization signal surrounding the mutated region in MDV
ICP4 based on various prediction programs (NucPred, NLS Mapper and NLStradamus; data not
shown).

78

79

FIGURE 3-3. Amino acid alignment of N-terminal end of ICP4. Individual sequences for the three mardiviruses are
shown while all other 17 herpesvirus ICP4 sequences are combined and shown as a non-mardivirus consensus sequence.
The region containing amino acids 52-69 unique to MDV in Gallid herpesvirus 2 is indicated by the boxed region, while
amino acids 60, 62 and 63 are designated with an *.

80

FIGURE 3-4. Amino acid alignment of 20 ICP4 species and HSV-1. The S1630P mutated
position in Gallid herpesvirus 2 is indicated by *. Identical amino acids at the same position are
indicated in black boxes, while amino acids with comparable properties but different identities
are indicated by shaded boxes.

81

Phylogenetic trees of all 20 alphaherpesvirus differed when considering either a
combination of multiple genes or only ICP4. Our phylogenetic tree constructed based on the
complete sequence of ICP4 from the 20 herpesvirus species had an overall topology and
grouping comparable to trees previously constructed using a compilation of several genes (18,
19). However, despite the overall similarity for most genera between the two trees, there was a
change in the relationship of the virulent Iltovirus, known as ILTV, and the Mardiviruses
between trees based only on ICP4 or using multiple genes. Combining several different genes to
construct one overall phylogenetic tree groups the closely related Gallid herpesviruses 1 and 2
together as sister taxa, with the avirulent Melaeagrid herpesvirus 1 branching from the Gallid
herpesvirus node to collectively classify the Mardiviruses into one group known as α3 viruses,
while the virulent Iltovirus, Gallid herpesvirus 1, that causes avian infectious laryngotracheitis in
chickens, branches separately from the Mardivirus lineage and is classified singly into a group
known as α4 (18, 19). However, when considering only the ICP4 amino acid sequence, there was
a change in the relationship of the virulent Iltovirus and the Mardiviruses. The resulting ICP4
tree has the most closely related sister taxa classified as the virulent Gallid herpesvirus-2 and
ILTV, which are shown to split from the avirulent Gallid herpesvirus-3 and Meleagrid
herpesvirus 1 lineage (Fig. 3-5). This yields classification of the three Mardiviruses and single
Iltovirus species along lines of virulence, unlike previous studies that classifies these species
relationships based on genera lines, in which the split between groups is among Mardi and
Iltoviruses instead of virulent and avirulent viruses. This relationship is dependent on unique
ICP4 sequences, as trimming ICP4 sequences to only use regions known to be highly conserved
among alphaherpesviruses, such as regions 2 and 4 (17, 20), results in a phylogenetic tree with

82

the expected grouping of the Mardivirus and Iltovirus genera as separate branches, restoring the
groupings of α3 and α4 (data not shown).

83

FIGURE 3-5. Maximum likelihood phylogenetic tree based on the complete amino acid
sequence of ICP4 from 20 herpesvirus species. Artificially rooted with the outgroup of UL36
amino acid sequence from Gallid herpesvirus 2.

84

DISCUSSION
During previous in vitro serial passage of MDV to find candidate genes involved in
attenuation, ICP4 was found to be mutated repeatedly in all attenuated MDV replicates
sequenced (6). These attenuated viruses contained nonsynonymous mutations at amino acids 60,
62 and 63 at frequencies of 96%, 83% and 100%, respectively, in three of the four attenuated
populations. Besides these clustered mutations, three more nonsynonymous mutations exceeding
frequencies of 40% or more were also seen at other positions within ICP4 in the attenuated
replicates. In addition to these serially passed Md5 replicates, mutations within ICP4 in
attenuated MDV have been identified by other groups as well. During serial passage and
sequencing of an attenuated 648A strain of MDV, three high frequency mutations within ICP4,
two nonsynonymous and one silent mutation, were found within the attenuated strain (21). Even
during complete genome sequence analysis of the current “gold-standard” MD vaccine known as
Rispens, mutations were found within ICP4 (22). Studies including alphaherpesviruses besides
MDV also show a similar trend of mutations within ICP4 in attenuated viruses. Complete
genome sequencing of an attenuated and highly passed vaccine strain of the pseudorabies virus,
Suid herpesvirus 1, identified 11 mutations within ICP4 that were unique to a vaccine strain
compared to two sequenced virulent strains (23). Furthermore, in the Oka vaccine strain of the
varicella-zoster virus (VZV), also known as human herpesvirus 3 (HHV-3), a disproportionate
number of mutations (15 mutations, of which 8 were nonsynonymous) were found within ICP4
in the vaccine strain relative to the virulent parental virus (24).
Despite evidence from a variety of attenuated alphaherpesviruses suggesting mutations
within ICP4 may be a causative factor driving attenuation of the virus, the two recombinant
viruses tested containing high frequency candidate SNVs within ICP4 did not cause any

85

reduction in virulence of MDV, although the Mut ICP4-2 recombinant virus did fail to transmit
horizontally to contact birds. Recombinant MDV viruses that are virulent and induced MD in
challenged birds but failed to transmit horizontally to cause MD in contact birds have been
previously reported for mutations in UL44 (gC) and UL13 (protein kinase) (7, 25), as well as
with a recombinant virus with a point mutation within LORF2 (6). One possible explanation for
the failure of Mut ICP4-2 to infect contact birds is due to lower levels of in vivo replication at
days 7 and 14, during the times traditionally classified as early cytolytic and latent phases of
infection (26). The lowest virus load among the three time points was seen during the transition
between the early cytolytic phase to latency (7 dpi) for Mut ICP4-2. This could affect
transmission of the recombinant virus because as chickens mature they develop an age-related
resistance towards MD. Therefore, young chicks are most susceptible to MDV infection, and due
to a decrease or delay in shedding of virulent MDV by Mut ICP4-2, the potential for
transmission and infection to older contact birds would decrease (7). Alternatively, these
mutations within ICP4 may impair replication within the feather follicles, leading to an inability
to spread through shed feather dander. Despite the indication that mutations within ICP4 appear
to be correlated with attenuation in a multitude of alphaherpesviruses, recombinant viruses
containing point mutations within ICP4 show that mutations within ICP4 alone are not sufficient
for attenuation in MDV. This conclusion is supported by experiments using the Oka vaccine
strain of VZV in which transactivation of downstream promoters regulated by ICP4 was
compared between mutated versions of ICP4 from the vaccine strain versus the parental virus.
Cohrs et al. (27) established that regulation and transcription of downstream promoters between
the mutated and wild-type ICP4 was comparable, leading to their conclusion that mutations
within ICP4 alone are not sufficient to cause attenuation.

86

Mutations within the VZV Oka vaccine strain did not affect transactivation of viral
promoters, but it is known that ICP4 contains different regions with specific functions in addition
to transactivation, such as DNA binding, nuclear localization, and regulation of late genes (17).
Therefore, we sought to computationally determine what function observed mutations within
MDV ICP4 may impact, particularly those within amino acids 60-63, which contained mutations
in three completely attenuated MDV replicates via sequence comparisons to other
alphaherpesviruses. Alignment of the complete amino acid sequence of ICP4 revealed significant
variation among alphaherpesviruses, which has previously been described. It is known that
certain regions of ICP4, such as regions 2 and 4, are highly conserved among herpesviruses,
while other regions are not particularly well conserved among species, such as regions 1, 3 and 5
(17, 20). Specifically, comparisons among genera show that Mardiviruses have significantly
longer ICP4 of approximately 2,173 amino acids in length compared to the average length of
1,500 amino acids for non-Mardiviruses. Furthermore, in our attenuated MDV replicates
containing several high frequency mutations around amino acids 60-63, these mutations occur
within a region found uniquely in MDV yet absent in other closely related Mardiviruses.
Considering that MDV is not only virulent, but is also an oncogenic virus, unlike all other
alphaherpesviruses which do not cause tumors, it is difficult to predict what role these unique
regions in MDV ICP4 provide during the life cycle of this oncogenic herpesvirus, particularly in
light of the fact that closely related apathogenic Mardiviruses lack this commonly mutated region
of ICP4.
The second mutation at amino acid 1630 occurs at a position present within the 20
herpesvirus species surveyed unlike the first mutation at amino acid 63. The mutation of a serine
residue to a proline at position 1630 appears to be an evolutionarily accepted change, with

87

proline commonly found in several other species of herpesviruses. Despite this apparently
tolerated amino acid change, it is only with addition of this second mutation at position 1630 in
conjunction with the mutation at amino acid 63 that changes in horizontal transmission were
observed. The role for this second mutation at position 1630 is difficult to predict due to the high
degree of variation among herpesvirus species for amino acids within this designated region 3 of
ICP4, which has been shown to be less highly conserved than regions 2 and 4 (17, 20), leading to
a poorer understanding of functional roles for diverse regions such as region 3. Analysis of
MDV’s ICP4 primary structure via software designed for modeling and prediction of protein
structure, such as for MDV ICP4 sequences predict that the regions of MDV ICP4 containing the
two mutations engineered are found within disordered, coiled regions of the protein but due to
lack of MDV ICP4 structural predictions, little data is available for hypothesizing roles for the
observed mutated regions.
Due to limitations of comparisons of ICP4 for specific amino acid positions among the
herpesviruses surveyed, further comparison of the entire ICP4 sequence via phylogenetic
analysis of the 20 alphaherpesvirus species based on the complete ICP4 amino acid sequence
showed a discrepancy between the relationship of Mardiviruses and Iltoviruses of chicken
compared to traditional trees based on compilation of data involving multiple genes. Previous
studies have classified the three Mardiviruses (Melaeagrid herpesvirus 1, Gallid herpesvirus 2
and Gallid herpesvirus 3) collectedly in one group known as α3 viruses, while the Iltovirus,
Gallid herpesvirus 1 (ILTV), was classified in a separate group designated α4, as would be
expected (19). Instead, based on ICP4 sequences, these relationships are no longer in line with
genera classification, but show closer homology between the virulent viruses of Gallid
herpesvirus 2 and ILTV now grouped as sister taxa, despite belonging to different genera. Thus,

88

it appears there may be a greater commonality between evolutionary pressure affecting the
sequence of ICP4 within virulent viruses compared to avirulent viruses for Mardi and Iltoviruses.
Collectively, these data suggests that mutations within ICP4 may play a supporting role
in attenuation, cooperatively with the addition of other mutations that reduce virulence. Due to a
selective advantage for viruses able to replicate faster during serial passage, mutations that
increase in vitro replication would be expected in attenuated viruses despite the fact that those
mutations themselves are not causative for the loss of virulence resulting from serial passage. It
is possible that mutations within ICP4 seen in numerous attenuated viruses may not reduce
virulence but would appear correlated with attenuation due to positive selective forces during in
vitro passage. This would result in mutations that impact in vitro replication to occur
concurrently with mutations that reduce virulence in vivo and are the causative mutation for
attenuation and the loss of virulence. One such potential scenario involves a point mutation
within the UL5 helicase-primase gene identified at 65% in an attenuated MDV viral population
that likely could have occurred concurrently within the same viral genome containing the ICP4
mutation G62V present at 83% in the same population. This mutation in UL5 reduced virulence
by nearly 90% (6), therefore, it may be postulated that mutations within ICP4, which increase in
vitro replication and lead to a selective advantage in vitro (24), could result in attenuated viruses
generated after repeated serial passage due to combination with additional mutations that alter
virulence, such as the UL5 mutation.
Despite the large body of examples correlating mutations within ICP4 with attenuation in
alphaherpesviruses, it remains unknown what function these mutations may impart to attenuated
viruses. Without further characterization it would be difficult to further postulate what role these
additional 700 amino acids confer in MDV relative to other alphaherpesviruses, or understand

89

the impact of the numerous mutations within ICP4 seen not only within attenuated MDV, but
among other alphaherpesviruses as well.

Conclusions
Despite showing multiple nonsynoymous mutations at high frequency in attenuated
MDVs, neither Q63H nor S1630P mutations were sufficient to reduce viral virulence and are not
causative factors for the loss of virulence during in vitro serial passage. Therefore, recombinant
MDVs with defined ICP4 mutations are not likely to be viable candidates as MD vaccines.

ACKNOWLEDGEMENTS
We thank Laurie Molitor, Lonnie Milam, and Spencer Jackson for technical support. This project
was supported in part by National Research Initiative Competitive Grant no. 2010-65119-20505
from the USDA National Institute of Food and Agriculture.

90

REFERENCES

91

REFERENCES

1.
Purchase HG, Okazaki W. 1971. Effect of vaccination with herpesvirus of turkeys (HVT) on
horizontal spread of Marek’s disease herpesvirus. Avian Dis. 15:391–397.
2.
163.

Witter RL. 1997. Increased virulence of Marek’s disease virus field isolates. Avian Dis. 41:149–

3.
Gimeno IM. 2008. Marek’s disease vaccines: A solution for today but a worry for tomorrow?
Vaccine 26, Supplement 3:C31–C41.
4.
Churchill AE, Chubb RC, Baxendale W. 1969. The attenuation, with loss of oncogenicity, of
the herpes-type virus of Marek’s disease (strain HPRS-16) on passage in cell culture. J. Gen. Virol.
4:557–564.
5.
Rispens BH, Vloten H van, Mastenbroek N, Maas HJL, Schat KA. 1972. Control of Marek’s
disease in the Netherlands. I. Isolation of an avirulent Marek’s disease virus (strain CVI988) and its use in
laboratory vaccination trials. Avian Dis. 16:108–125.
6.
Hildebrandt E, Dunn JR, Perumbakkam S, Niikura M, Cheng HH. 2014. Characterizing the
molecular basis of attenuation of Marek’s disease virus via in vitro serial passage identifies de novo
mutations in the helicase-primase subunit gene UL5 and other candidates associated with reduced
virulence. J. Virol.
7.
Jarosinski KW, Margulis NG, Kamil JP, Spatz SJ, Nair VK, Osterrieder N. 2007. Horizontal
Transmission of Marek’s Disease Virus Requires US2, the UL13 Protein Kinase, and gC. J. Virol.
81:10575–10587.
8.
Tischer B, von Einem J, Kaufer B, Osterrrieder N. Two-step Red-mediated recombination for
versatile high-efficiency markerless DNA manipulation in Escherichia coli. Biotechniques 40:191–197.
9.
Niikura M, Kim T, Silva RF, Dodgson J, Cheng HH. 2011. Virulent Marek’s disease virus
generated from infectious bacterial artificial chromosome clones with complete DNA sequence and the
implication of viral genetic homogeneity in pathogenesis. J. Gen. Virol. 92:598–607.
10.
Engel AT, Selvaraj RK, Kamil JP, Osterrieder N, Kaufer BB. 2012. Marek’s disease viral
interleukin-8 promotes lymphoma formation through targeted recruitment of B cells and CD4+ CD25+ T
cells. J. Virol. 86:8536–8545.
11.
Moriuchi H, Moriuchi M, Smith HA, Straus SE, Cohen JI. 1992. Varicella-zoster virus open
reading frame 61 protein is functionally homologous to herpes simplex virus type 1 ICP0. J. Virol.
66:7303–7308.
12.
Gimeno IM, Cortes AL, Silva RF. 2008. Load of challenge Marek’s disease virus DNA in
blood as a criterion for early diagnosis of Marek’s disease tumors. Avian Dis. 52:203–208.
13.
Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H,
Valentin F, Wallace IM, Wilm A, Lopez R, Thompson JD, Gibson TJ, Higgins DG. 2007. Clustal W
and Clustal X version 2.0. Bioinformatics 23:2947–2948.

92

14.
Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. 2013. MEGA6: molecular
evolutionary genetics analysis version 6.0. Mol. Biol. Evol.
15.
The UniProt Consortium. 2014. Activities at the Universal Protein Resource (UniProt). Nucleic
Acids Res. 42:D191–D198.
16.
Kumar P, Henikoff S, Ng PC. 2009. Predicting the effects of coding non-synonymous variants
on protein function using the SIFT algorithm. Nat. Protoc. 4:1073–1081.
17.
Wyrwicz LS, Rychlewski L. 2007. Fold recognition insights into function of herpes ICP4
protein. Acta Biochim. Pol. 54:551–559.
18.
McGeoch DJ, Cook S. 1994. Molecular phylogeny of the Alphaherpesvirinae subfamily and a
proposed evolutionary timescale. J. Mol. Biol. 238:9–22.
19.
McGeoch DJ, Dolan A, Ralph AC. 2000. Toward a comprehensive phylogeny for mammalian
and avian herpesviruses. J. Virol. 74:10401–10406.
20.
McGeoch DJ, Dolan A, Donald S, Brauer DHK. 1986. Complete DNA sequence of the short
repeat region in the genome of herpes simplex virus type 1. Nucleic Acids Res. 14:1727–1745.
21.
Spatz SJ. 2010. Accumulation of attenuating mutations in varying proportions within a high
passage very virulent plus strain of Gallid herpesvirus type 2. Virus Res. 149:135–142.
22.
Spatz SJ, Petherbridge L, Zhao Y, Nair V. 2007. Comparative full-length sequence analysis of
oncogenic and vaccine (Rispens) strains of Marek’s disease virus. J. Gen. Virol. 88:1080–1096.
23.
Szpara ML, Tafuri YR, Parsons L, Shamim SR, Verstrepen KJ, Legendre M, Enquist LW.
2011. A wide extent of inter-strain diversity in virulent and vaccine strains of alphaherpesviruses. PLoS
Pathog. 7.
24.
Gomi Y, Sunamachi H, Mori Y, Nagaike K, Takahashi M, Yamanishi K. 2002. Comparison
of the complete DNA sequences of the Oka varicella vaccine and its parental virus. J. Virol. 76:11447–
11459.
25.
Jarosinski KW, Osterrieder N. 2010. Further Analysis of Marek’s Disease Virus Horizontal
Transmission Confirms That UL44 (gC) and UL13 Protein Kinase Activity Are Essential, while US2 Is
Nonessential. J. Virol. 84:7911–7916.
26.
Baigent SJ, Davison F. 2004. Marek’s disease virus: biology and lifecycle, p. 62–77. In Davison,
F, Nair, V (eds.), Marek’s disease. Academic Press, San Diego, CA.
27.
Cohrs RJ, Gilden DH, Gomi Y, Yamanishi K, Cohen JI. 2006. Comparison of virus
transcription during lytic infection of the Oka parental and vaccine strains of Varicella-Zoster virus. J.
Virol. 80:2076–2082.

93

CHAPTER 4. A SINGLE NUCLEOTIDE MUTATION WITHIN THE UL5 HELICASEPRIMASE SUBUNIT GENE NOT ONLY REDUCES VIRULENCE, BUT ALSO PROVIDES
VACCINAL PROTECTION AGAINST MAREK'S DISEASE
ABSTRACT
Marek’s disease virus (MDV) is an oncogenic herpesvirus that afflicts chickens with the disease
known as Marek’s disease (MD). This virus causes tumors, nerve lesions, immunosuppression,
and death of affected birds. Vaccines are the primary method for control of MD but due to the
periodic evolution of field strains, it is necessary to explore the development of new MD
vaccines. MD vaccines are often attenuated MDV strains generated through serial passage in
vitro. We previously used experimental evolution of MDV to provide a better understanding of
the genetic basis of attenuation. During complete genome sequencing of evolved MDV
populations, we identified a point mutation within the UL5 helicase-primase gene that reduced
disease incidence by 89-100%. To determine if experimental evolution also identifies mutations
that provide protective qualities as potential vaccine candidates, we tested the UL5 recombinant
virus as a vaccine and compared its protection to commercial HVT and bivalent (HVT + SB-1)
vaccines. Both commercial vaccines resulted in higher protection against MD than the UL5
recombinant virus, although the UL5 virus did provide protection against developing MD in 4670% of birds challenged. This indicates that a mutation within the UL5 helicase-primase gene
not only reduces virulence, but also confers protection against challenge with virulent MDV,
providing support that not only can experimental evolution identify candidate mutations involved
in attenuation, but also identify potential candidates for use in vaccine development.

94

INTRODUCTION
Marek’s disease virus (MDV; aka Gallid herpesvirus 2) is an oncogenic alphaherpesvirus that
cause Marek’s disease (MD) in chickens. Annually, MD causes nearly $1-2 billion losses
worldwide, therefore, control and prevention of this disease is important for the poultry industry
(1). Vaccines are the primary method for controlling MD, which were first employed in 1969
with the use of the live attenuated strain HPRS-16 (2). A related turkey herpesvirus, known as
HVT which is nononcogenic in chickens, was introduced shortly afterwards and has been widely
used to protect against MD since 1971 (3). Despite the introduction of early MD vaccines, field
strains of MDV evolved to greater virulence that were no longer controlled by vaccines such as
HVT (4). This led to the introduction of bivalent vaccines based on HVT with the addition of a
second vaccine, such as the nononcogenic SB-1 strain, in which the two strains interacted
synergistically to protect better than single vaccine preparations alone (5). While this
combination provided protection for several years, more virulent strains once again emerged
requiring vaccines with greater protection, leading to the U.S. introduction in the 1990s of the
Rispens, or CVI998 strain, currently considered the gold standard of MD vaccines (6). Despite,
or likely as a result of, the availability and use of these protective vaccines, virulent strains are
now appearing able to cause outbreaks even within flocks vaccinated with Rispens (4).
This evolution of more virulent strains requiring the successional introduction of more
efficacious vaccines illustrates the importance of continuous vaccine development. Several
vaccines, including both HPRS-16 and Rispens, are live attenuated viruses resulting from serial
passage in vitro (2, 7). Considering the history of in vitro serial passage for generation of
attenuated vaccines, this process could further yield new vaccine candidates. Despite the utility
of this method for vaccine development, the underlying genetic mechanism behind the process of

95

attenuation remains unclear. With the advent of next-generation sequencing, many MDV strains
have now been sequenced, including vaccines such as SB-1 and Rispens, as well as attenuated
viruses such as a highly passed strain of 648A (8–10). Sequencing complete MDV genomes
provides valuable insight into genetic differences between virulent and nononcogenic viruses, yet
still requires validation of candidate polymorphisms identified to determine their impact on
virulence.
Previously, we serially passed three replicate strains of a virulent, BAC-cloned MDV
virus and sequenced the resulting attenuated viruses. We identified candidate mutations believed
to play a role in attenuation and created recombinant viruses testing 5 unique mutations via Redmediated recombineering. One nonsynonymous point mutation within UL5, which encodes the
helicase-primase subunit, significantly reduced both in vivo replication and virulence when used
to challenge birds. With the availability of BAC-derived MDVs and ability to create recombinant
viruses by Red-mediated recombineering, there is now the opportunity to design and engineer
candidate vaccines incorporating mutations into the MDV genome to quickly and precisely
create avirulent viruses to yield a more protective attenuated virus. Therefore, we wanted to
determine if this single nucleotide mutation within UL5 conferred any protection against
challenge with virulent MDV and could be considered for future development of vaccine
candidates, or if the UL5 mutation decreased virulence with no bearing on protection against
MD.

96

MATERIALS AND METHODS
Viruses and Cell Culture. Viral stocks were propagated in secondary duck embryo fibroblast
cells in a 1:1 mixture of Leibovitz’s L-15 and McCoy’s 5A (LM) media supplemented with 14% fetal bovine serum (FBS), penicillin (10,000 U per 100 ml), 200 g/ml streptomycin, and 0.2
g/ml amphotericin B in a 37 °C, 5% CO2 incubator. The recombinant virus, Mut UL5-I682R,
was generated from the Md5B40BAC using Red-mediated recombineering, as previously
described (11). Md5B40BAC-c1 (12) and JM/102W challenge viruses, as well as the bivalent
vaccine, were derived from ADOL stocks, while the HVT vaccine originated from a commercial
source (13).

In Vivo Challenge. A minimum of seventeen maternal antibody negative, 15I5 x 71 chicks were
vaccinated intraabdominally upon day-of-hatch with 2,000 plaque forming units (PFU) of HVT,
bivalent HVT+SB-1 (1,000 PFU of each virus to yield a collective dose of 2,000 PFU in total),
or the recombinant Mut UL5-I682R. Vaccinated birds were then infected five days later with 500
PFU of the virulent challenge virus, either JM/102W or Md5B40BAC-c1, and housed in
Horsfield-Bauer isolators for 10 weeks. Isolators of unvaccinated birds were also challenged
with virulent viruses as unvaccinated controls. To determine how increasing the quantity of
PFUs administered to match commercial doses affected disease incidence due solely to Mut
UL5-I682R, an additional set of birds challenged with 2,000 PFU Mut UL5-I682R only was
included. Moribund birds, and all birds that survived past 10 weeks, were terminated and
examined via necropsy for signs of MD to determine disease incidence. All in vivo bird trials
were approved by the USDA, Avian Disease and Oncology Laboratory Animal Care and Use
Committee (ACUC). Based on ACUC guidelines established and approved by the ADOL ACUC

97

(April 2005) and the Guide for the Care and Use of Laboratory Animals by the Institute for
Laboratory Animal Research (2011), procedures for bird housing and care were followed
throughout the duration of the experiment.

Vaccinal Protection of Viruses. To compare the protection of Mut UL5-I682R versus
commercial HVT or bivalent vaccines against challenge with virulent MDV, the protective index
(PI) of vaccines was calculated based on MD incidence. The percent MD incidence in birds
challenged only with virulent MDV strains was compared against levels of MD incidence
resulting from vaccinated birds infected with the same virulent virus, and the protective index
calculated as follows;
PI= ((% MD incidence in unvaccinated) – (% MD incidence in vaccinated))/ (% MD incidence
in unvaccinated).
Fischer’s exact test was used to determine if there were statistically significant
differences in levels of protection between Mut UL5-I682R and commercial vaccines when
challenged with the same virulent MDV strain.

98

RESULTS
Unvaccinated birds challenged with virulent virus, either JM/102W or Md5B40BAC-c1,
developed MD and resulted in 89-100% mortality of infected birds before the end of the
experiment (Fig. 4-1 and 4-2). On the contrary, between 89-100% of birds vaccinated with either
commercial vaccine, either HVT or bivalent, survived the full length of the experiment and upon
termination had disease incidences <11% (Fig. 4-1 and 4-2, Table 4-1). Vaccination with the
recombinant virus Mut UL5-I682R resulted in approximately 80% of birds surviving the full
length of the experiment regardless of which challenge virus, either JM/102W or Md5B40BACc1, was used to infect birds (Fig. 4-1 and 4-2). Examination upon termination revealed birds
vaccinated with Mut UL5-I682R resulted in 30% and 48% MD following challenge with virulent
JM/102W or Md5B40BAC-c1 viruses, respectively (Table 4-1). Of the birds administered 2,000
PFU of recombinant UL5-I682R virus only, 90% survived the full 10 week experiment and upon
termination 23% of birds were MD positive, indicating increasing the quantity of the Mut UL5I682R virus delivered by four-fold also increased MD incidence relative to earlier trials, in which
500 PFU of Mut Ul5-I682R caused 0 to 11% MD.

99

FIGURE 4-1. Survival of birds challenged with JM/102W virus. A minimum of 17, 15I5 x 71
maternal antibody negative chicks were vaccinated with either HVT or Mut UL5-I682R virus at
day of hatch, while unvaccinated birds were used as controls, before challenge with 500 PFU of
JM/102W virus at 5 days of age and mortality shown over the course of the experiment until
termination.

100

FIGURE 4-2. Survival of birds challenged with Md5B40BAC-c1 virus. A minimum of 17, 15I5
x 71 maternal antibody negative chicks were vaccinated with either bivalent (HVT+SB-1)
vaccine or Mut UL5-I682R of virus at day of hatch, while unvaccinated birds were used as
controls, before challenge with 500 PFU of Md5B40BAC-c1 virus at 5 days of age and mortality
shown over the course of the experiment until termination.

101

102

PI of Viruses. To compare protective efficacy of the vaccines, the PI of the three vaccinating
strains were calculated as described previously. Without vaccination, the JM/102W virus caused
100% MD incidence in the unvaccinated birds. Vaccination with HVT before challenge with
JM/102W protected all birds against MD, resulting in a PI of 100 and MD incidence of 0%,
while the recombinant Mut UL5-I682R resulted in 30% of birds developing MD when
challenged by JM/102W (PI=70). Comparing protection of Mut UL5-I682R versus the bivalent
vaccine when challenged with Md5B40BAC-c1, a more virulent vv strain of MDV, showed a
similar trend. Without vaccination, the virulent challenge virus alone caused high MD incidence
of 89%. After vaccination with the bivalent vaccine, challenge by Md5B40BAC-c1 resulted in
only 11% MD (PI=87) compared to vaccination with Mut UL5-I682R that had higher levels of
MD, 48% (PI=46). Birds administered 2,000 PFU of only the Mut UL5-I682R virus developed
MD at higher levels than previous trials involving only 500 PFU of the virus (11). The higher
dose of Mut UL5-I682R used in all the vaccine trials caused 23% MD without addition of a
following challenge virus. Statistical analysis of disease incidence comparing unvaccinated
versus vaccination with Mut UL5-I682R or either commercial vaccines (HVT or bivalent
vaccines) using Fisher’s exact test showed significant differences among all three groups
(P<0.05). Comparing unvaccinated birds challenged with either virulent virus (JM/102W or
Md5B40BAC-c1) versus birds vaccinated with Mut UL5-I682R prior to challenge showed
significant differences between vaccinated and unvaccinated birds. Furthermore, comparing MD
in birds vaccinated with Mut UL5-I682R and birds vaccinated with a commercial vaccine there
also was a statistically significant difference between protection conferred by Mut UL5-I682R
compared to HVT or bivalent vaccines.

103

In all trials, the commercial vaccine strains (HVT or bivalent) provided higher protection,
although not always complete protection, compared to Mut UL5-I682R against challenge by
virulent virus. While the Mut Ul5-I682R recombinant virus did not protect as well as commercial
vaccines, it did improve the survival of birds and protected birds against virulent MDV strains,
although not to as high degrees as standard vaccine strains.
DISCUSSION
Due to the increasing evolution of more virulent strains of MDV able to overcome
vaccinal protection of even the current, most efficacious vaccines, new candidate vaccines must
be developed to address the disease threat of these new strains. Previously, we used experimental
evolution of three MDV replicates to identify candidate mutations involved in attenuation during
serial passage in vitro. Recombinant viruses testing top candidate mutations revealed the
recombinant virus, Mut UL5-I682R, dramatically altered in vitro phenotype of the virus by
causing a reduction in disease incidence by nearly 90% or more depending on trial. We tested the
Mut UL5-I682R recombinant virus in vaccine trials to determine any protective qualities due to
this SNP, as well as to provide a proof-of-concept that experimental evolution is a viable route
for identification of candidate SNPs and genes for potential use as vaccine candidates.
Comparing protection of the recombinant Mut UL5-I682R versus commercial vaccines
indicated that the commercial vaccines all provided higher degrees of protection against
challenge by virulent viruses. Despite the superior protection conferred by commercial vaccine
strains, Mut UL5-I682R still was able to provide a fair level of protection and improved survival
of birds when challenged with either virulent (v) or very virulent (vv) MDV strains. Considering
that the Mut UL5-I682R recombinant virus results in 0-11% MD in birds infected with low, 500
PFU doses, and up to 40% MD when 2,000 PFUs are administered in vaccine-level doses,
104

indicates that the recombinant mutant containing a single SNP clearly is not a vaccine candidate
currently. Considering that this recombinant virus significantly reduces disease incidence and in
vivo replication (11), while also providing some level of protection against virulent MDV strains,
instead suggests this mutation could potentially be used collectively with other mutations to
generate a completely apathogenic virus, which in turn could be considered for use as a
candidate vaccine.
Additional candidate mutations previously identified within attenuated viruses could be
considered for incorporation into the UL5 recombinant virus for potential vaccine research and
development, although a recombinant virus involving mutations within both within UL5 and the
oncogene Meq may be a superior alternative for initial consideration. The Meq gene has
previously been identified as the oncogene of MDV and a virus in which the Meq gene has been
deleted, known as rMd5∆Meq, resulted in an virus which no longer causes MD (14, 15).
Deletion of the Meq gene in rMd5∆Meq has been shown to not only result in an avirulent virus
that does not caused MD, but also serves as candidate vaccine that provides higher protection
than even Rispens, the current most efficacious vaccine (16). Despite the improved protective
qualities of rMd5∆Meq, this virus still induces bursal-thymus atrophy (BTA) in maternal
antibody negative birds and, as a result, has not been approved as a commercial vaccine (17).
Attempts to eliminate bursal-thymus atrophy by serial passage of rMd5∆Meq in vivo resulted in
a virus that no longer caused BTA in maternal antibody negative birds, but also lost the superior
protective efficacy of the parental rMd5∆Meq relative to Rispens when used in commercial birds
(18). The authors noted that with the loss of BTA after 40 passages in vitro, they also no longer
were able to detect viral DNA, indicating that the attenuated rMd5∆Meq p40 virus was unable to
replicate in vivo. They concluded that BTA is likely the result of vigorous replication of the low

105

passage rMd5∆Meq within the lymphoid organs, and that in vitro passage eliminating in vivo
replication of the high passage p40 rMd5∆Meq virus led to loss of BTA as well. Based on the
hypothesis that high levels of viral replication within lymphoid organs cause BTA and loss of in
vivo viral replication eliminates BTA, suggests one possible solution would be creation of a Meq
deletion virus able to replicate in vivo, albeit at low levels as to not induce BTA. The candidate
mutation we have identified within UL5 is one candidate SNP to propose introducing into a Meq
deletion virus. By itself, the UL5 mutation causes greatly reduced levels of in vivo replication
and, as shown in this study, confers protection against challenge with virulent MDV.
Incorporating both mutations involving deletion of Meq and addition of the UL5 point mutation
into a single recombinant virus would be predicted to generate a virus lacking oncogenicity and
that replicates at low levels in vivo, although the resulting impact on BTA and PI compared to
Rispens and rMd5∆Meq remain unknown.
This experiment shows that a previously identified mutation within UL5 not only greatly
reduces virulence, but also imparts protection against challenge with virulent MDV, although the
PI of the Mut UL5-I682R recombinant virus is lower than current vaccines. This single mutation
alone is not appropriate for use as a vaccine, yet it is possible that the use of this SNP, in
conjunction with other identified mutations such as deletion of Meq, may be considered for
development of new vaccines addressing the evolution of emerging vv+ strains of MDV. This
experimental also shows that experimental evolution of MDV is a process not only able to
identify candidate mutations involved in attenuation and but identifies mutations that provide
protection against challenge by virulent MDV and that can be considered for use in vaccine
development.

106

ACKNOWLEDGEMENTS
We thank Laurie Molitor, Spencer Jackson and Lonnie Milam for, as always, excellent technical
support. This project was supported in part by National Research Initiative Competitive Grant
no. 2010-65119-20505 from the USDA National Institute of Food and Agriculture.

107

REFERENCES

108

REFERENCES

1.
Morrow C, Fehler F. 2004. Marek’s disease: a world-wide problem, p. 49–61. In
Davison, F, Nair, V (eds.), Marek’s disease. Academic Press, San Diego, CA.
2.
Churchill AE, Chubb RC, Baxendale W. 1969. The attenuation, with loss of
oncogenicity, of the herpes-type virus of Marek’s disease (strain HPRS-16) on passage in cell
culture. J. Gen. Virol. 4:557–564.
3.
Purchase HG, Okazaki W. 1971. Effect of vaccination with herpesvirus of turkeys
(HVT) on horizontal spread of Marek’s disease herpesvirus. Avian Dis. 15:391–397.
4.
Davison F, Nair V. 2005. Use of Marek’s disease vaccines: could they be driving the
virus to increasing virulence? Expert Rev. Vaccines 4:77–88.
5.
Witter RL, Lee LF. 1984. Polyvalent Marek’s disease vaccines: Safety, efficacy and
protective synergism in chickens with maternal antibodies. Avian Pathol. 13:75–92.
6.
Witter RL. 2001. Protective efficacy of Marek’s disease vaccines, p. 57–90. In Hirai,
PDK (ed.), Marek’s Disease. Springer Berlin Heidelberg.
7.
Rispens BH, Vloten H van, Mastenbroek N, Maas HJL, Schat KA. 1972. Control of
Marek’s disease in the Netherlands. I. Isolation of an avirulent Marek’s disease virus (strain
CVI988) and its use in laboratory vaccination trials. Avian Dis. 16:108–125.
8.
Spatz SJ, Schat KA. 2011. Comparative genomic sequence analysis of the Marek’s
disease vaccine strain SB-1. Virus Genes 42:331–338.
9.
Spatz SJ, Petherbridge L, Zhao Y, Nair V. 2007. Comparative full-length sequence
analysis of oncogenic and vaccine (Rispens) strains of Marek’s disease virus. J. Gen. Virol.
88:1080–1096.
10.
Spatz SJ. 2010. Accumulation of attenuating mutations in varying proportions within a
high passage very virulent plus strain of Gallid herpesvirus type 2. Virus Res. 149:135–142.
11.
Hildebrandt E, Dunn JR, Perumbakkam S, Niikura M, Cheng HH. 2014.
Characterizing the molecular basis of attenuation of Marek’s disease virus via in vitro serial
passage identifies de novo mutations in the helicase-primase subunit gene UL5 and other
candidates associated with reduced virulence. J. Virol.
12.
Niikura M, Kim T, Silva RF, Dodgson J, Cheng HH. 2011. Virulent Marek’s disease
virus generated from infectious bacterial artificial chromosome clones with complete DNA

109

sequence and the implication of viral genetic homogeneity in pathogenesis. J. Gen. Virol.
92:598–607.
13.
Witter RL, Calnek BW, Buscaglia C, Gimeno IM, Schat KA. 2005. Classification of
Marek’s disease viruses according to pathotype: philosophy and methodology. Avian Pathol.
34:75–90.
14.
Jones D, Lee L, Liu JL, Kung HJ, Tillotson JK. 1992. Marek disease virus encodes a
basic-leucine zipper gene resembling the fos/jun oncogenes that is highly expressed in
lymphoblastoid tumors. Proc. Natl. Acad. Sci. U. S. A. 89:4042–4046.
15.
Lupiani B, Lee LF, Cui X, Gimeno I, Anderson A, Morgan RW, Silva RF, Witter
RL, Kung H-J, Reddy SM. 2004. Marek’s disease virus-encoded Meq gene is involved in
transformation of lymphocytes but is dispensable for replication. Proc. Natl. Acad. Sci. U. S. A.
101:11815–11820.
16.
Lee LF, Kreager KS, Arango J, Paraguassu A, Beckman B, Zhang H, Fadly A,
Lupiani B, Reddy SM. 2010. Comparative evaluation of vaccine efficacy of recombinant
Marek’s disease virus vaccine lacking Meq oncogene in commercial chickens. Vaccine 28:1294–
1299.
17.
Lee LF, Heidari M, Zhang H, Lupiani B, Reddy SM, Fadly A. 2012. Cell culture
attenuation eliminates rMd5ΔMeq-induced bursal and thymic atrophy and renders the mutant
virus as an effective and safe vaccine against Marek’s disease. Vaccine 30:5151–5158.
18.
Lee LF, Kreager K, Heidari M, Zhang H, Lupiani B, Reddy SM, Fadly A. 2013.
Properties of a meq-deleted rMd5 Marek’s disease vaccine: protection against virulent MDV
challenge and induction of lymphoid organ atrophy are simultaneously attenuated by serial
passage in vitro. Avian Dis. 57:491–497.

110

CHAPTER 5. ADDITION OF A UL5 HELICASE-PRIMASE SUBUNIT POINT MUTATION
ELIMINATES BURSAL-THYMIC ATROPHY OF MAREK’S DISEASE VIRUS ∆MEQ
RECOMBINANT VIRUS, BUT REDUCES VACCCINAL PROTECTION
ABSTRACT
Marek’s disease virus (MDV) is an oncogenic alphaherpesvirus and the causative agent
of Marek’s disease (MD), a T-cell lymphoma of chickens. Despite widespread usage of vaccines
since the 1970’s to control MD, more virulent strains of MDV have emerged that break vaccinal
protection, necessitating the development of new and more protective MD vaccines. The
rMd5∆Meq (∆Meq) virus, a recombinant Md5 strain MDV lacking the viral oncogene Meq, is
one candidate MD vaccine with great potential, but unfortunately it also causes bursal-thymic
atrophy (BTA) in maternal antibody negative chickens, prohibiting approval as a vaccine.
Previously, we identified a point mutation within UL5 that reduced in vivo replication in
attenuated viruses. We proposed that introduction of the UL5 point mutation into the ∆Meq virus
would reduce in vivo replication and eliminate BTA yet retain high protective abilities. In birds,
the ∆Meq+UL5 recombinant MDV had reduced replication compared to the original ∆Meq
virus, while weights of lymphoid organs indicated that the ∆Meq+UL5 did not induce BTA,
supporting the hypothesis that reduction of in vivo replication would also abolish BTA. Vaccine
trials of the ∆Meq+UL5 virus compared to other ∆Meq based viruses and commercial vaccines
show that, while the ∆Meq+UL5 does provide vaccinal protection, this vaccinal protection was
also reduced compared to the original ∆Meq virus. Therefore, it appears that a very delicate
balance is required between levels of replication able to induce high vaccinal protection, yet not
as high as to induce BTA.

111

Key words: Marek’s disease virus (MDV), Marek’s disease (MD), Meq, bursal thymic atrophy
(BTA), vaccine, bacterial artificial chromosome (BAC), recombinant
INTRODUCTION
Marek’s disease (MD) is an economically significant disease of chickens that costs the
worldwide poultry industry $1-2 billion a year (1). The causative agent of MD is the oncogenic
herpesvirus known as Gallid herpesvirus 2, commonly referred to as Marek’s disease virus
(MDV). Characteristic signs of MD include depression, transient paralysis, nerve lesions,
tumors, and death of affected birds. The primary method used to successfully control MD since
the 1970’s is vaccination. The first vaccine against MD was the attenuated strain of MDV known
as HPRS-16, which was shortly followed by the introduction of another vaccine utilizing the
closely related turkey herpesvirus known as HVT (2, 3).
While vaccines have been used to successfully prevent MD and reduce losses in the
poultry industry for over 40 years ago, more virulent strains have evolved requiring the
introduction of a successive line of new vaccines (4). Disease breaks within HVT-vaccinated
flocks resulted in the need to introduce bivalent vaccines combining two vaccines that interacted
synergistically, such as HVT and another attenuated virus such as SB-1, to yield a vaccine that
was more protective than either vaccine alone (5). Bivalent vaccines controlled losses due to MD
until the evolution of more virulent field strains led to the introduction of the CVI988-Rispens
(Rispens) vaccine in the U.S. (6, 7). Presently, the Rispens vaccine is considered the goldstandard vaccine against MD and provides the highest level of protection commercially
available.
Despite the level of success in controlling MD, there is concern that more virulent strains
of MDV are evolving able to overcome even Rispens vaccinal protection (8). Currently a
112

recombinant cosmid-based Md5 strain MDV in which the Meq oncogene has been deleted,
known as rMd5∆Meq, has shown great potential in several trials. Unfortunately, rMd5∆Meq
causes bursal-thymus atrophy (BTA) in maternal antibody negative birds and, as a result, has not
been licensed for commercial use as a vaccine. Attempts to eliminate BTA of rMd5∆Meq via
repeated serial passage in vitro to attenuate the virus resulted in viruses that did not cause BTA
in antibody negative birds (9), but failed to replicate and lost the superior protective qualities of
the initial rMd5∆Meq virus in commercial antibody positive birds (10). It is postulated that BTA
seen in the otherwise avirulent rMd5∆Meq virus is the result of vigorous in vivo replication of
the virus within the bursa and thymus, leading to atrophy of those lymphoid organs (10, 11).
During previous efforts to identify the driving factors for MDV attenuation at a genetic
level, we identified a point mutation within UL5, the helicase primase subunit, that resulted in
significant reduction of in vivo replication of the virus, while significantly reducing MD
incidence to 0-11% as well as providing protection against challenge with virulent MDV (12).
Considering the hypothesis that loss of BTA seen in the high-passage rMd5∆Meq resulted from
considerable reduction in replication after passage, we predicted that addition of the UL5 point
would reduce replication of ∆Meq to low levels seen in the UL5 point mutant, while
simultaneously eliminating BTA characteristic of low passage ∆Meq, but that it would still
confer vaccinal protection.
A second Meq deletion mutant generated using a BAC construct, known as
Md5B40BAC∆Meq, has previously been created and shown to be as protective as rMd5∆Meq,
and equaling or even surpassing protection of Rispens (13). Therefore, use of the BAC cloned
MDV allowed us to use Red-mediated recombineering to create a double mutant of
Md5B40BAC∆Meq+UL5 (12). We compared in vivo replication and ratios of lymphoid organ

113

weight for low passage ∆Meq, high passage ∆Meq (p41), and the double mutant,
Md5B40BAC∆Meq+UL5, against uninfected birds. Furthermore, we tested the three preceding
∆Meq-based viruses as well as two commercially approved vaccines, CVI988 and bivalent
(HVT+ SB-1), for vaccinal protection against challenge with 648A, a very virulent plus (vv+)
strain of MDV.

114

MATERIALS AND METHODS
Viruses and Tissue Culture. Md5B40∆Meq BAC generated and characterized by Silva et al.
(13) was used for incorporation of the UL5 point mutation using Red-mediated recombineering
protocols and mutational primers as previously described (12). The resulting
Md5B40BAC∆Meq+UL5 (∆Meq+UL5) recombinant mutant contained deletion of both copies
of Meq and the nonsynonymous point mutation of amino acid I682R within UL5. Viral stocks of
Md5B40∆Meq BAC (∆Meq BAC) at p5 and ∆Meq+UL5 p4 were used for birds trials, while
rMd5∆Meq p39 (9, 10) was plated and passed through chicken embryo fibroblast (CEF) for one
additional passage to amplify viral stocks before use of the resulting rMd5∆Meq p40 (∆Meq
p40) for bird trials. The very virulent plus (vv+) MDV 648A (p7) (6)strain was used as the
challenge virus. The two MD vaccines compared were (1) bivalent vaccine consisting of HVT
(p10) and SB-1 (p14) strains and (2) Rispens (p43); of which all viruses used originated from
ADOL stocks. Cells and viruses were cultured in a 1:1 mixture of Leibovitz’s L-15 and 100
McCoy’s 5A (LM) media containing 1-4% of fetal bovine serum (FBS) with antibiotics as
previously described (12).

In Vivo Replication and Bursal Thymic Atrophy. ADOL 15I5 x 71 maternal antibody negative
chicks were vaccinated with 2,000 PFU of ∆Meq BAC p5, ∆Meq p40, or ∆Meq+UL5 p4 at day
of hatch and housed in Horsfall-Bauer (HB) unit isolators. In addition to vaccinated birds, an
isolator of unvaccinated birds was maintained for negative controls. Birds were bled at days 6,
13, and 20 days post vaccination (dpv) and peripheral blood lymphocytes (PBL) isolated using
Histopaque-1077 density gradient separation (Sigma-Aldrich). DNA extracted from PBLs was
used for qPCR analysis to compare in vivo replication among vaccines. The Taqman Fast

115

Universal PCR kit (Applied Biosciences; Foster City, CA) and primers for amplification of
chicken GAPDH and MDV gB were used to calculate the relative ratio of the number of MDV
genomes versus chicken genome numbers as previously described (12, 15).
To monitor BTA, five birds from each of the four lots (∆Meq BAC p5, ∆Meq p40,
∆Meq+UL5 p4, or unvaccinated) were sacrificed at 15 dpv, the time when atrophy is typically
most severe (11), and their total body weight (g) and weights of bursa and thymus (mg)
measured. The ratio of lymphoid organs over total body weight was calculated to compare
against unvaccinated controls to assay atrophy of lymphoid organs in vaccinated birds. Ratios of
lymphoid organs over total body weight were analyzed using ANOVA to determine differences
between vaccinated lots and unvaccinated controls.

Vaccinal Protection. To assay vaccinal protection of Meq-deleted viruses, we compared the
three Meq-deleted viruses previously described against two MD vaccines: Rispens/CVI988 and
bivalent (HVT+SB-1). Maternal antibody negative, ADOL line 15I5 x 71 chicks were vaccinated
with 2,000 PFU of ∆Meq BAC p5, ∆Meq p40, ∆Meq+UL5 p4, bivalent (HVT+SB-1) or Rispens
at day of hatch and housed in HB units. After 5 dpv, vaccinated birds were challenged with 500
pfu vv+ 648A strain of MDV. As positive controls, one lot of unvaccinated birds were
challenged with 500 PFU of 648A at 5 days of age. Unvaccinated and uninfected negative
control birds were also housed and maintained with the experimental lots in HB units for 8
weeks. Birds that died before 8 weeks were examined via necropsy for cause of death, while
after 8 weeks all surviving birds were terminated and examined via necropsy for signs of MD.
To determine vaccinal protection, we calculated protective index (PI) of vaccines as follows; PI=
((% MD incidence in unvaccinated) – (% MD incidence in vaccinated))/ (% MD incidence in

116

unvaccinated)). To compare protection of vaccines, analysis using Fisher’s exact test was
performed to identify statistically significant differences in MD among vaccinated groups.
RESULTS
In Vivo Replication. Unvaccinated birds and birds vaccinated with either ∆Meq+UL5 p4 or
∆Meq p40 all had very low background levels of MDV detected at all three time points (Fig. 51). ∆Meq and virulent 648A challenged birds had high levels of at 6 dpv and 13 days post
challenge, respectively. No data is available at 20 days post challenge for 648A infected birds
due to the severity of this vv+ virus causing mortality of nearly all birds within 20 days after
challenge with 648A. There were statistically significant differences in replication at 6 dpv
(p<0.01), in which the low passage ∆Meq BAC p5 had very high levels of replication at 6 dpv
that dropped dramatically at 13 and 20 dpv to levels comparable to unvaccinated birds and the
two other Meq-based viruses.

BTA of Meq-Based Viruses. Ratios comparing the weights of lymphoid organs, either
individual bursa and thymus weights or total lymphoid organs, to the total body weight of birds
at 15 dpv showed a significant difference in the weights of lymphoid organs in birds vaccinated
with low passage ∆Meq BAC p5 (Fig. 5-2) compared to unvaccinated birds (p<0.001 in all trial
comparisons). Individual bursa and thymus weights, or total lymphoid organ to body weight
ratios of birds vaccinated with either ∆Meq p40 or ∆Meq+UL5 p4 were not significantly
different from unvaccinated (ANOVA p>0.15 in all trial comparisons).

Protective Efficacy of Vaccines. Both trials showed severe mortality in unvaccinated birds
challenged with 648A (Fig. 5-3). Due to infection with this vv+ strain of MDV, over 90% of
117

unvaccinated birds challenged with 648A were deceased within 2-3 weeks post challenge. In trial
one, all birds vaccinated with ∆Meq p40, bivalent, or Rispens prior to challenge survived the full
8 week experiment, while ∆Meq BAC p5 had one death out of fifteen birds (7% mortality) and
four out of 14 birds died when vaccinated with ∆Meq+UL5 p4 (29% mortality). In trial two
similar trends were also observed in which 80-90% of birds vaccinated with either ∆Meq+UL5
p4, ∆Meq BAC p5, ∆Meq p40, or Rispens survived the full 8 week experiment compared to
unvaccinated birds challenged with 648A, which were all deceased by 3 weeks post challenge
(Figure 5-3b).
Examination via necropsy to determine MD incidence revealed very high protection
following vaccination with ∆Meq BAC p5, ∆Meq p40, and Rispens. Of these highly protective
vaccines, ∆Meq BAC p5 had the lowest protective index of 93, while ∆Meq p40 and Rispens had
no MD positive birds, resulting in PIs of 100 for both viruses in Trial 1 (Table 5-1). The
∆Meq+UL5 p4 and bivalent viruses had higher degrees of MD incidence in birds vaccinated with
these viruses, resulting in PIs of 57-43 or 87-73 for each respective virus, showing they did not
confer as strong of protection as the previous three viruses, but they still provided significant
protection against challenge with vv+ MDV. Results of trial two also supported these
conclusions in which ∆Meq BAC p5, ∆Meq p40, and Rispens all provided high protection
against MD (PIs of 94, 94 and 88, respectively), while lower levels of protection were provided
by vaccination with ∆Meq+UL5 p4 (PI 44).

118

FIGURE 5-1. In vivo replication of MDV within PBLs in vaccinated, unvaccinated or 648A
challenged birds at 6, 13 and 20 days post vaccination. Unvaccinated birds challenged with 648A
experienced severe mortality around 2 weeks post infection, resulting in no data for the final
timepoint indicated with an “x”. Panel A designated trial one and panel B indicates trial 2.

119

FIGURE 5-2. Weight of lymphoid organs in vaccinated and unvaccinated birds relative to total
body weight in Trial 1 (Panel A) and Trial 2 (B).

120

FIGURE 5-3. Survival of vaccinated birds following challenge with vv+ 648A MDV. Panel A
shows data from the first trial, with panel B depicting data from second replicate trial. Data for
birds vaccinated with bivalent vaccines was unavailable for trial two due to flood causing
complete loss of the isolator housing that lot.

121

122

DISCUSSION
While low passage ∆Meq is a highly effective vaccine able to protect against vv+ MDV
strains, it also induces BTA in maternal antibody negative chickens, prohibiting commercial
licensing of the virus as a vaccine (11). In addition to BTA, another trait characteristic of low
passage ∆Meq is the vigorous in vivo replication of ∆Meq during early cytolytic replication. This
characteristic was previously described, and is apparent in this study as well (11, 16). Due to loss
of vigorous in vivo replication following serial passage of ∆Meq, it was postulated this high
replication within lymphoid organs was the cause of BTA. We hypothesized that addition of a
point mutation within UL5 shown to cause greatly reduced in vivo replication would abrogate
BTA in low passage ∆Meq, resulting in an avirulent virus that does not induce tumors and
replicates at very low levels as to not cause BTA. Comparing replication via qPCR of PBLs and
the weight of lymphoid organs of the three ∆Meq based viruses compared to unvaccinated birds
supported the hypothesis that reducing in vivo replication in the ∆Meq+UL5 p4 virus also
resulted in a low passage ∆Meq-based virus that failed to cause BTA. In addition to the
anticipated phenotypic changes in regards to in vivo replication and lymphoid organ weight,
addition of the UL5 point mutation also affected vaccinal protection of the ∆Meq+UL5 p4
compared to the original ∆Meq BAC p5 virus. This illustrates the fine interplay between even a
single point mutation intended to reduce in vivo replication and broader implications on vaccinal
protection, suggesting a delicate balance between vaccinal protection, in vivo replication and
BTA.
Previous studies have explored the interaction between MD vaccines and replication.
Studies comparing vaccinal protection of serially passed and fully attenuated MDV strains
relative to earlier passages still containing residual levels of virulence, designated as partially

123

attenuated viruses, showed that partially attenuated viruses had higher viral loads in PBLs and
conferred significantly higher levels of protection than the fully attenuated passages of the
viruses (17). Further studies exploring the interaction between protection and replication of
candidate vaccines compared a set of several vaccine candidates categorized as high protective
(HP) or low protective (LP) vaccines in order to identify characteristics that appeared common
among high or low protective vaccines (18). One trait the authors noted that distinguished high
from low protective vaccines was in vivo replication. Looking at viral load of HP viruses
revealed higher viral loads in all tissues quantified, with significantly higher levels in lymphoid
organs such as the bursa, thymus and spleen.
These studies collectively suggest a correlation of greater vaccinal protection with high
in vivo replication. Unfortunately it is not uncommon to find that highly protective, high
replicating MD vaccine candidates also induce BTA. One such example is the virus known as
RM1, which was generated by cocultivation of MDV and reticuloendotheliosis virus (REV),
leading to retroviral insertions into the MDV genome. RM1 was determined to be attenuated and
replicated efficiently in vivo, yet failed to cause MD and was typified by severe BTA (19).
Further characterization revealed that not only was the attenuated RM1 virus very protective as a
vaccine, but also exhibited high tissue tropism for early dissemination and replication to the
thymus (20). Among several vaccine viruses examined, RM1 showed the earliest and highest
replication within lymphoid organs, specifically the thymus, compared to vaccine viruses that did
not cause BTA.
Additional studies looking at in vivo replication within the thymus comparing both RM1
and low passage ∆Meq against in vivo replication of viruses that do not induce BTA show both
RM1 and low pass ∆Meq have significantly higher levels of replication within the thymus (11).

124

This supports the hypothesis that increased replication, specifically within lymphoid organs such
as the bursa and thymus, leads to atrophy of these sensitive organs. A confounding factor for
vaccine development is that despite the negative effects causing BTA due to high levels of in
vivo replication, increased replication appears to be a factor impacting higher degrees of vaccinal
protection conferred by viruses. Therefore, in vivo replication and the protection qualities of
candidate vaccines appear to be highly interconnected, although not the only factor believed to
play a role in protection, leading to complications when trying to eliminate BTA by altering
excessive replication while still producing a protective vaccine.
Modulation of in vivo replication has been a method proposed for generation of
attenuated vaccines by increasing replication fidelity of the viral polymerase in a polio mouse
model. High-fidelity polio vaccines exhibiting altered tissue tropism resulted in attenuated
viruses with diminished transmission and shed of virus (21). These high-fidelity replication
variant viruses were also able to generate excellent immune responses and induce antibody levels
comparable to, or even exceeding, current Sabin polio vaccines. Considering the trend that
moderate-to-high levels of replication tends to result in more protective MD vaccines than their
low replicating counterparts, it may be advantageous to try to direct replication and limit tissue
tropism for these vaccine candidates. Considering that viruses such as RM1 and low passage
∆Meq replicate at high levels in lymphoid organs, and those tissues suffer atrophy as a result of
this vigorous replication, it may be beneficial to restrict the tissue tropism of prolific replication
to avoid BTA. By diverting the highest replication, this may allow low passage ∆Meq viruses to
maintain their high levels of replication in tissues other than the bursa and thymus and prevent
BTA, while also provoking a strong immune response due to vigorous early cytolytic replication.
Considering that MDV is a highly cell-associated virus whose lifecycle involves preferential

125

infection and replication within B and T cells, it may be difficult to directly apply this approach
with MDV, but it is clear that alternative approaches besides simply reducing in vivo replication
should be considered for development of protective candidate vaccines.
Ongoing development of more protective MD vaccines clearly illustrates the difficulty in
generating vaccines that are not only more protective, but also do not cause undesirable
secondary reactions, as seen with low pass ∆Meq or RM1 viruses which induce BTA. Attempts
to generate new vaccine candidates have yielded many viruses which confer vaccinal protection
equaling that currently provided by Rispens, yet it has been difficult to surpass the protection of
Rispens (22). Two candidate vaccines, RM1 and low passage ∆Meq, both met or exceeded
vaccinal protection commonly seen due to Rispens vaccination, but both viruses also induce
BTA in maternal antibody negative birds, prohibiting their approval as commercial vaccines
(18). Attempts to eliminate BTA in low passage ∆Meq in previous studies by extensive serial
passage resulted in a virus with reduced in vivo replication and which did not cause BTA, but
also reduced protection compared to the original low passage ∆Meq (10). This study aiming to
eliminate BTA in low passage ∆Meq by reducing in vivo replication via addition of the UL5
point mutation showed similar results as previous trials, in which BTA was lost at the expense of
vaccinal protection.
Clearly there is a fine balancing act between generating a highly protective vaccine that
elicits a strong immune response with mild residual pathogenicity, compared to completely
viruses that are completely avirulent yet may not induce as vigorous immune response as the
virus’ partially virulent counterpart. As shown in this study and supported by previous work, it
appears easier to eliminate undesirable BTA characteristics of a virus such as ∆Meq, but a far
more complicated task to do so without also simultaneously reducing vaccinal protection.

126

Perhaps new routes must be considered to address this problem of BTA in highly protective
viruses, such as trying to restrict the tissue tropism of replication, instead of simply reducing
replication to eliminate BTA, which has been shown to be an unsatisfactory method to address
BTA without also undermining high vaccinal protection.

ACKNOWLEDGEMENTS
We would like to thank Laurie Molitor, Spencer Jackson and Lonnie Milam for excellent
technical support.

127

REFERENCES

128

REFERENCES

1.
Morrow C, Fehler F. 2004. Marek’s disease: a world-wide problem, p. 49–61. In
Davison, F, Nair, V (eds.), Marek’s disease. Academic Press, San Diego, CA.
2.
Churchill AE, Chubb RC, Baxendale W. 1969. The attenuation, with loss of
oncogenicity, of the herpes-type virus of Marek’s disease (strain HPRS-16) on passage in cell
culture. J. Gen. Virol. 4:557–564.
3.
Purchase HG, Okazaki W. 1971. Effect of vaccination with herpesvirus of turkeys
(HVT) on horizontal spread of Marek’s disease herpesvirus. Avian Dis. 15:391–397.
4.
Witter RL. 2001. Protective efficacy of Marek’s disease vaccines, p. 57–90. In Hirai,
PDK (ed.), Marek’s Disease. Springer Berlin Heidelberg.
5.
Witter RL, Lee LF. 1984. Polyvalent Marek’s disease vaccines: Safety, efficacy and
protective synergism in chickens with maternal antibodies. Avian Pathol. 13:75–92.
6.
Witter RL. 1997. Increased virulence of Marek’s disease virus field isolates. Avian Dis.
41:149–163.
7.
Rispens BH, Vloten H van, Mastenbroek N, Maas HJL, Schat KA. 1972. Control of
Marek’s disease in the Netherlands. I. Isolation of an avirulent Marek’s disease virus (strain
CVI988) and its use in laboratory vaccination trials. Avian Dis. 16:108–125.
8.
Gimeno IM. 2008. Marek’s disease vaccines: A solution for today but a worry for
tomorrow? Vaccine 26, Supplement 3:C31–C41.
9.
Lee LF, Heidari M, Zhang H, Lupiani B, Reddy SM, Fadly A. 2012. Cell culture
attenuation eliminates rMd5ΔMeq-induced bursal and thymic atrophy and renders the mutant
virus as an effective and safe vaccine against Marek’s disease. Vaccine 30:5151–5158.
10.
Lee LF, Kreager K, Heidari M, Zhang H, Lupiani B, Reddy SM, Fadly A. 2013.
Properties of a meq-deleted rMd5 Marek’s disease vaccine: protection against virulent MDV
challenge and induction of lymphoid organ atrophy are simultaneously attenuated by serial
passage in vitro. Avian Dis. 57:491–497.
11.
Dunn JR, Silva RF. 2012. Ability of MEQ-deleted MDV vaccine candidates to
adversely affect lymphoid organs and chicken weight gain. Avian Dis. 56:494–500.
12.
Hildebrandt E, Dunn JR, Perumbakkam S, Niikura M, Cheng HH. 2014.
Characterizing the molecular basis of attenuation of Marek’s disease virus via in vitro serial
passage identifies de novo mutations in the helicase-primase subunit gene UL5 and other
candidates associated with reduced virulence. J. Virol.
13.
Silva RF, Dunn JR, Cheng HH, Niikura M. 2010. A MEQ-deleted Marek’s disease
virus cloned as a bacterial artificial chromosome is a highly efficacious vaccine. Avian Dis.
54:862–869.
129

14.
Witter RL. 1997. Increased virulence of Marek’s disease virus field isolates. Avian Dis.
41:149–163.
15.
Gimeno IM, Cortes AL, Silva RF. 2008. Load of challenge Marek’s disease virus DNA
in blood as a criterion for early diagnosis of Marek’s disease tumors. Avian Dis. 52:203–208.
16.
Lupiani B, Lee LF, Cui X, Gimeno I, Anderson A, Morgan RW, Silva RF, Witter
RL, Kung H-J, Reddy SM. 2004. Marek’s disease virus-encoded Meq gene is involved in
transformation of lymphocytes but is dispensable for replication. Proc. Natl. Acad. Sci. U. S. A.
101:11815–11820.
17.
Witter RL. 2002. Induction of Strong Protection by Vaccination with Partially
Attenuated Serotype 1 Marek’s Disease Viruses. Avian Dis. 46:925–937.
18.
Gimeno IM, Witter RL, Hunt HD, Reddy SM, Reed WM. 2004. Biocharacteristics
shared by highly protective vaccines against Marek’s disease. Avian Pathol. 33:57–66.
19.
Jones D, Brunovskis P, Witter R, Kung HJ. 1996. Retroviral insertional activation in a
herpesvirus: transcriptional activation of US genes by an integrated long terminal repeat in a
Marek’s disease virus clone. J. Virol. 70:2460–2467.
20.
Gimeno IM, Witter RL, Cortes AL, Reed WM. 2011. Replication ability of three
highly protective Marek’s disease vaccines: implications in lymphoid organ atrophy and
protection. Avian Pathol. 40:573–579.
21.
Vignuzzi M, Wendt E, Andino R. 2008. Engineering attenuated virus vaccines by
controlling replication fidelity. Nat Med 14:154–161.
22.
Witter RL, Kreager KS. 2004. Serotype 1 viruses modified by backpassage or
insertional mutagenesis: approaching the threshold of vaccine efficacy in Marek’s disease. Avian
Dis. 48:768–782.

130

CHAPTER 6. CHARACTERIZATION OF MAREK’S DISEASE VIRUS POINT MUTATION
WITHIN UL5 HELICASE-PRIMASE SUBUNIT IDENTIFIED WITHIN ATTENUATED
VIRUS SHOWN TO REDUCE VIRULENSE AND IN VIVO REPLICATION

ABSTRACT
A nonsynonymous point mutation identified within UL5, which encodes the helicaseprimase protein, subunit in a serially passed, attenuated MDV was shown to reduce in vivo
replication and MD incidence when incorporated into a recombinant virus, known as Mut UL5.
This mutation was further characterized to determine if addition of a second top candidate
mutation would result in complete loss of virulence. Growth curves quantifying in vitro
replication were conducted comparing the mutant UL5 virus to parental wild-type virus, and the
UL5 mutant was serially passed in vivo while selecting for increased in vivo replication and
assaying for reversion of the point mutation during replication within birds. Addition of a second
point mutation did not significantly affect in vivo replication or MD incidence compared to the
single UL5 virus, and serial passage of the single UL5 virus through birds did not yield any
revertant viruses or result in increased replication of disease incidence. Growth curves showed
significantly higher numbers of Mut UL5 plaques compared to the wildtype virus, yet head-tohead competitions of the two viruses resulted in the parental virus outcompeting the Mut UL5,
suggesting traditional growth curves may not provide complete pictures of the true biological
ramifications of in vitro replication for viral populations and quasi-species collectively
replicating within a shared culture.

131

INTRODUCTION
Marek’s disease virus (MDV) is an oncogenic alphaherpesvirus that causes Marek’s
disease (MD) in chickens, which is characterized by depression, paralysis, nerve lesions and tcell lymphomas that ultimately culminates in the death of affected chickens (1). Control of MD
has primarily been through the use of vaccines, of which many highly effective MD vaccines are
attenuated viruses created through in vitro serial passage (2). Due to the continued evolution of
more virulent field strains of MDV, the development of more protective MD vaccines has been a
continued topic of research since the first introduction of MD vaccines in the 1970’s (3–7). In
efforts to better understand the process of attenuation and provide information that may assist in
future vaccine development, we previously serially passed three BAC cloned MDV replicates
and conducted complete genome sequencing of the attenuated replicates via Illumina sequencing
(8). We identified several candidate mutations and created recombinant viruses testing the
phenotypic impact of those mutations and identified a point mutation within the UL5 helicaseprimase subunit that reduced disease incidence by >90% while significantly reducing in vivo
replication. Due to its dramatic impact, we wished to further characterize this point mutation.
Considering that the UL5 recombinant virus still resulted in low levels of MD, we wanted to
determine if adding a second, high-frequency mutation identified in the attenuated replicate
would interact in a synergistic manner to result in complete loss of virulence and a fully
attenuated recombinant virus. Furthermore, we sought to determine if the point mutation affected
in vitro replication, which may provide a selective advantage during serial passage and explain
its rise in frequency within the attenuated viral population. Finally, we also wanted to determine
if, despite the low in vivo replication of the Mut UL5 virus, it was possible to select for improved
in vivo replication, and, if so, whether increased in vivo replication would result in increased

132

virulence through reversion or compensatory mutations that may arise counteracting the original
UL5 point mutation.
MATERIALS AND METHODS
Tissue Culture: Cell and Virus Propagation
Chicken embryo fibroblast (CEF) and duck embryo fibroblasts (DEF) were used to
propagate viruses. All cells were cultured in Leibovitz’s L-15 and McCoy’s 5A (LM) media (1:1
ratio) and supplemented with fetal bovine serum (1-4% FBS) and penicillin, streptomycin, and
amphotericin B (10,000 U per 100 ml, 0.2mg/ml streptomycin, and 0.02µg/ml respectively).
Recombinant viruses were generated via Red-mediated recombineering using techniques
previously described (8–10) and mutational primers indicated in Table 6-1.
UL5+UL46 Double Mutant In Vivo Replication and MD Incidence
We generated and tested a double recombinant virus of Md5B40BAC UL5I682R+UL46-Q117R (Mut UL5+UL46) to see if addition of the UL46 point mutation had a
synergistic or additive effect on virulence compared to the single UL5 mutation. The double Mut
UL5+UL46 virus was created by addition of the UL46-Q117R mutation into the UL5 mutant
BAC using primers listed in Table 6-1 and Red-mediated recombineering as described earlier.
The single Mut UL5-I682R recombinant virus and double Mut UL5+UL46 recombinant viruses
were used to challenge groups of 18 maternal antibody negative ADOL line 151 5x71 chicks,
injected intraabdominally with 500 PFU of virus at day of hatch. Five uninfected contact birds
were also housed with infected birds to determine horizontal transmission of viruses. Birds were
sacrificed for tissue collection at days 7, 14 and 21 post infection and spleen extracted DNA used

133

for qPCR to determine viral load. Birds were housed for (10) weeks and examined via necropsy
upon termination to identify MD positive birds and determine disease incidence.

134

135

Viral Growth Curves
Growth curves comparing replication of the wild-type Md5B40BAC-c1 versus Mut UL5I682R were conducted as follows. For each respective virus, 21 p60 plates containing a confluent
monolayer of CEF cells were infected with a calculated quantity of approximately 50 PFU of the
appropriate virus. Triplicate plates at 1, 2, 3, 4, and 5 days post infection for each virus were
trypsinized with .025% trypsin solution to be individually frozen in vials for later titration of
virus at each timepoint. The total number of plaques were counted for three plates at 6 days post
infection to calculate the precise quantity of PFU actually delivered during inoculation of CEF
plates. After collecting timepoints, the frozen vials were titered. Each vial was titered in triplicate
in a series of dilutions ranging from ¼ to 1/1000 of the original infected p60 plates. Plaques were
counted at 6 days post infection to calculate the number of plaques per p60 plate frozen at each
timepoint.
Viral Competition
2,000 PFU of each Md5B40BAC-c1 and Mut UL5-I682R were mixed (1:1) to infect
three p100 plates containing a CEF monolayer and 10% of harvested cells were passed onto a
new, uninfected CEF monolayer once mature plaques appeared, on average 4-5 days post
infection. At each passage, the remaining harvested cells to quantify the ratio of wild-type to
mutant allele using pyrosequencing to track changes in allele frequencies. DNA was extracted
from aliquots of harvested cells during trypsinization for passage using QIAamp DNA Blood
Extraction Kit (Qiagen Cat# 51106) for PCR amplification and prepared using the Pyromark
Gold Q96 pyrosequencing kit (Qiagen Cat# 972804) with primers described in Table 6-1 and run
on a PSQ 96MA pyrosequencing machine.

136

Serial Passage of Mut UL5-I682R In Vivo
17 maternal antibody negative ADOL line 1515x71 chicks were challenged at one week
of age with 500 PFU of either the parental Md5B40BAC-c1 or recombinant Mut UL5-I682R and
designated as in vivo p0. One week post infection, all birds were bled and PBLs isolated, DNA
extracted and Taqman qPCR conducted to identify birds with the highest levels of MDV for each
of the viral lots. Approximately 3-5 of the birds with the highest levels of MDV were then rebled at two weeks post infection and PBLs from each individual bird isolated and plated onto a
p100 containing a CEF monolayer to reisolate MDV. Infected CEF plates were passed to amplify
viral quantities by 2-3 passages until large quantities of plaques were produced, then harvested
and frozen in vials to be titered. Extracted DNA from the amplified viral stocks was assayed
using pyrosequencing to verify whether the mutant UL5 SNP was still present, or if it had
reverted to the wild-type allele in the highest replicating viruses.
Titered viral stocks from the top 1-2 birds with the highest load of MDV for each viral lot
were used to infect another set of isolators with 500 PFU of the derived Md5B40BAC-c1 and
Mut UL5-I682R viruses described as “in vivo passage 1”. This process of infecting, bleeding,
screening, selecting and amplifying viruses from birds with the highest levels of MDV was
repeated for a total of three passages through birds to select for viruses with the highest levels of
in vivo replication. During the final challenges of the in vivo p3 viruses, two additional groups of
seventeen maternal antibody negative 1515x71 chicks were challenged with the original viruses
prior to passage in vivo to directly compare replication and disease incidence prior and post
passage. These birds were housed and examined via necropsy for MD upon termination 8 weeks
post infection to compare disease incidence of the in vivo p0 and in vivo p3 viruses

137

RESULTS

UL5+UL46 Double Mutant In Vivo Replication and MD Incidence
Recombinant viruses containing a single UL5 point mutation or a double mutant
recombinant virus containing both the UL5 and UL46 point mutation were both shown to
replicate at very low levels in vivo (Figure 1). Viral load of MDV from spleen tissue of infected
birds revealed no significant differences in replication between the single or double recombinant
mutants (ANOVA p=0.4559)
Birds challenged with recombinant UL5 viruses all experienced low levels of MD with
disease incidences of 17% or less (Table 6-2). Although the double mutant had slightly higher
levels of MD compared to the single mutant, the difference in the number of MD positive birds
between the single and double mutant (1 and 3, respectively) was not statistically significant
(Fischer’s exact test p=0.6026).
Mut UL5-I682R In Vitro Replication; Viral Growth Curves and Competitions
Growth curves comparing the parental Md5B40BAC-c1 and Mut UL5-I682R show the
two viruses were initially plated with an average of 54 or 52 PFU, respectively, yet the Mut
UL5-I682R virus produced higher number of plaques for all days 1-5 compared to the parental
Md5B40BAC-c1 (Fig 6-2). There were statistically significant differences in the number of
plaques produced between the two viruses at 1, 2, 3, and 5 days (Unpaired t-test p= 0.0042,
0.0004, 0.0017, and <0.0001, respectively). Differences in plaque numbers at day 4 were not
statistically significant at p= 0.0537.
Head-to-head competition of Md5B40BAC-c1 vs Mut UL5-I682R showed that despite
starting mixed cultures with approximately 50% of each virus, there was a steady progression

138

towards greater quantities of the wild-type Md5B40BAC-c1 allele with each passage in all three
replicates of the mixed UL5:B40 cultures (Figure. 6-3). After 20 passages the frequency of the
wild type allele composed approximately 83-90% of the viral DNA, indicating a greater yield of
the wild-type viral DNA in mixed cultures than Mut Ul5-I682R.
Serial Passage of Mut UL5-I682R In Vivo
Screening viral stocks of the highest replicating viruses generated from in vivo serial
passage via pyrosequencing to determine presence of either the mutant or wild-type allele within
UL5 indicated that there was no reversion of the Mutant UL5 allele to the parental form in any of
the viral stocks isolated from birds with the highest viral load (n=7).
Comparing disease incidence resulting from infection with pre and post in vivo passage
viruses of both Md5B40BAC-c1 and Mut UL5-I682R did not result in an increase in virulence
for either virus as a result of passage and selection for high viral load (Table 6-3). The virulent
parental Md5B40BAC-c1 typically has 90-100% virulence and birds trials comparing the
Md5B40BAC prior and post in vivo passage resulted in all developing MD, therefore this virus
already induces the highest levels of MD, making increases in virulence of Md5B40BAC
impossible to detect. There also were no significant increases in virulence of Mut UL5-I682R
following passage in vivo. Several bird trials of Mut UL5-I682R show that MD incidence
routinely falls within 0-11% MD. In vivo passage p3 of Mut UL5-I682R resulted in <10% MD
incidence, a level of MD which falls within the range of MD typically seen from the pre-passage
Mut UL5-I682R. Comparing viral load via qPCR of PBL and spleen samples from the two
viruses pre and post passage shows that in both tissues, the Mut UL5 viruses replicated at very
low levels which were statistically indistinguishable (T-test p>0.2989). For the higher replicating
MD5B40BAC viruses, there were also no significant differences in viral load in PBLs before and
139

after passage (T-test p=0.400). Within spleen tissues there were higher average levels of MDV
after in vivo passage, although this degree of increase was nonsignificant (T-test p=0.1024) (Fig
6-4).

140

FIGURE 6-1. In Vivo Replication of UL5 Single and Double Recombinant Viruses. Splenic
DNA was extracted for qPCR analysis to quantify viral loads of the single UL5 mutant virus or
double Mut UL5+UL46 recombinant viruses relative to uninfected birds (Neg) or wildtype
Md5B40 challenged birds. Results from two separate trials of Mut UL5 are indicated by “a” and
“b”.

141

142

FIGURE 6-2. Growth curves of Wild-type MdB40BAC-c1 and Mut UL5. 50 PFU of parental
Md5B40BAC or Mut UL5 viruses were used to infect p60 plates containing a CEF. Triplicate
plates of each virus were then trypsinized to harvested and freeze plates at 0, 1, 2, 3, 4 and 5 days
post infection. Dilutions of these harvested plates were used to titer and calculate the number of
plaques for all timepoints for the two viruses to compare plaque yield of Mut UL5 versus
Md5B40BAC viruses.

143

FIGURE 6-3. Head-to-Head competition of wild-type Md5B40BAC-c1 vs Mut UL5. 2,000 PFU
of either Mut UL5 (UL5) or Md5B40BAC (B40) viruses were used to infect p100 plates
containing a CEF monolayer. Three plates were then infected with 2,000 PFU of both viruses
and designated plates A,B or C (B40:UL5 A, B, C). These plates were serially passed and at each
passage stocks saved for DNA extraction for pyrosequencing analysis to quantify the proportion
of DNA containing the wild-type allele or the mutant allele following serial passage.

144

145

FIGURE 6-4. In Vivo Replication of Md5B40BAC-c1 and Mut UL5 within PBL and spleen (s)
tissue Pre and Post serial passage in vivo. Birds challenged with the Mut UL5 (UL5) or
Md4B40BAC (B40) viruses were bled for collection of PBLs or splenic DNA and extracted for
qPCR to quantify viral load of the respective viruses prior to in vivo passage (p0) or after three
passages through birds while selecting for birds with the highest viral loads (p3) to determine
MDV levels in viruses following selection for greater in vivo replication.

146

DISCUSSION
Initially, the UL5 point mutation resulted in a >90% reduction in MD incidence, yet
addition of a second, high-frequency UL46 candidate SNP identified within the same attenuated
replicate did not have an additive or synergistic effect on either MD incidence or in vivo
replication. While the single UL46 point recombinant virus was comparable to the virulent
parental virus for the common traits assayed in MDV infection, such as MD incidence,
replication, survival duration or horizontal transmission (8), the double UL5+UL46 recombinant
also showed that addition of the UL46 mutation did not confer any observable differences in
those key in vivo phenotypes compared to the single UL5 recombinant. The nonsynonymous
UL46 point mutation occurred at a high frequency (99%) within the attenuated replicate and was
expected to confer some phenotypic effect due to the hypothesis that this mutation may provide a
selective advantage in order to explain the near fixation of this mutation within the attenuated
population.
In vitro characterization of replication has been performed by comparing the UL5
recombinant virus growth in vitro to the parental Md5B40BAC-c1. Growth curves showed the
Mut Ul5-I682R virus produced significantly plaques than the wild-type virus. However, in headto-head competitions the wild-type virus outcompeted the UL5 virus, eventually resulting in the
wild-type allele composing 80-90% of the population following serial passage of the viral
popuations. One factor that potentially explains this discrepancy is that counting plaque numbers
in growth curves does not take into account the total number of infectious cells produced within
each plaque or plaque size. Comparing plaques, the number of infected cells, as well as the
quantity of virus produced within each infected cell can vary, allowing for viruses to produce
drastically different quantities of infected cells from one plaque to another. Therefore the UL5

147

mutant virus may have a greater number of overall plaques, yet the number of infectious cells,
and in turn virus, produced by each plaque may be lower than the yield for each plaque of the
Md5B40BAC-c1 virus, allowing for the outcompetition of the wild-type virus when
approximately equal quantities of plaques are initially mixed.
Further work would be necessary to better understand the mechanism between replication
of these two viruses to have a better understanding of if the observed differences in growth curve
vs competition assays was the result of a biological difference in the viruses’ response to
replication within a denser, genetically diverse population due to a difference in copies of MDV
genomes produced or cells infected within a plaque, overall plaque size, soluble products
released within conditioned media or other factors.
Nevertheless, these results show that under traditional in vitro growth curve assays, the
Mut UL5 recombinant virus showed higher replication by producing more plaques than the wildtype virus. Yet under more biologically relevant conditions reflecting the quasi-species nature of
MD viruses, in which mixed populations of the Mut UL5 and wild-type Md5B40BAC viruses
were serially passed and allowed to replicate in head-to-head competition, the result of the headto-head competition was unexpected based upon the result of the growth curve. This suggests
that while in vitro growth curves may provide important information regarding replication
dynamics of viruses, this assay may fail to encompass all aspects which may be pertinent for
replication and competition within the biologically relevant context of populations.
Besides characterizing in vitro replication of the Mut UL5 virus, we also attempted to
increase replication via selection and repeated serial passage in vivo, particularly with an
emphasis on determining the likelihood for the mutation virus to revert or undergo a secondary
compensatory mutation to restore wild-type levels of replication and virulence. Despite serially

148

passing and selecting viruses with the highest viral loads in birds, there was no significant
increase of in vivo replication for either virus. Prior work involving in vivo serial passage of
Md5B40BAC, the same progenitor virus used in this study, showed that the virus initially caused
MD in 18% of the MHC-B13 Line 0 of maternal antibody positive chickens prior to passage,
where after 5 passages virulence of passaged viruses increased to 94% (19). This previous work
shows selection for increased in vivo replication and MD incidence can be observed significantly
within a very small number of passages. However, our use of the Md5B40BAC virus within the
highly susceptible line of maternal antibody negative 1515x71 birds leads to maximum MD
incidence, making it difficult to select for significant increases and adaptation in an equally short
time frame. Conversely, while the Md5B40BAC virus caused 100% MD incidence in birds both
before and after serial passage, the Mut UL5 virus was at the opposite spectrum, resulting in
<6% MD incidence both before and after passage. Despite selecting for the birds with the highest
levels of MDV, even the highest loads of Mut UL5 virus were significantly lower than those of
the virulent Md5B40BAC virus. Due to a lack of increases in either in vivo replication or MD
incidence after selection and serial passage, it seemed likely that compensatory mutations
occurred to counteract the Mut UL5 point mutation. Pyrosequencing of serially passaged viral
stocks confirmed that the UL5 point mutation did not revert to the wild-type Md5B40BAC
allele.
In summary, the UL5 point mutant was shown to result in significantly higher levels of in
vitro replication via traditional growth curves, yet when grown in head-to-head competition with
the parental Md5B40BAC virus, it did not outcompete the parental virus. This suggests
traditional growth curves may fail to account for key biological aspects of replication within
populations. Considering the impact of this single point mutation on both in vitro and in vivo

149

replication, experiments to determine stability of this mutation under selection during in vivo
passage determined that mutations to counteract the detrimental reduction of in vivo replication
do not readily occur even when purposefully selecting for increased replication, suggesting that
despite concern over reversions of point mutations in vaccine candidates, point mutations may be
more stable than initially anticipated.

ACKNOWLEDGEMENTS
We would like to thank Laurie Molitor, Lonnie Milam and Spencer Jackson for
exemplary technical support, as well the USDA National Institute of Food and Agriculture for
partial financial support via the National Research Initiative Competitive Grant number 201065119-20505.

150

REFERENCES

151

REFERENCES

1.
Baigent SJ, Davison F. 2004. Marek’s disease virus: biology and lifecycle, p. 62–77. In
Davison, F, Nair, V (eds.), Marek’s disease. Academic Press, San Diego, CA.
2.
Witter RL. 2001. Protective efficacy of Marek’s disease vaccines, p. 57–90. In Hirai,
PDK (ed.), Marek’s Disease. Springer Berlin Heidelberg.
3.
Churchill AE, Chubb RC, Baxendale W. 1969. The attenuation, with loss of
oncogenicity, of the herpes-type virus of Marek’s disease (strain HPRS-16) on passage in cell
culture. J. Gen. Virol. 4:557–564.
4.
Purchase HG, Okazaki W. 1971. Effect of vaccination with herpesvirus of turkeys
(HVT) on horizontal spread of Marek’s disease herpesvirus. Avian Dis. 15:391–397.
5.
Witter RL. 1982. Protection by attenuated and polyvalent vaccines against highly
virulent strains of Marek’s disease virus. Avian Pathol. 11:49–62.
6.
Witter RL, Lee LF. 1984. Polyvalent Marek’s disease vaccines: Safety, efficacy and
protective synergism in chickens with maternal antibodies. Avian Pathol. 13:75–92.
7.
Rispens BH, Vloten H van, Mastenbroek N, Maas HJL, Schat KA. 1972. Control of
Marek’s disease in the Netherlands. I. Isolation of an avirulent Marek’s disease virus (strain
CVI988) and its use in laboratory vaccination trials. Avian Dis. 16:108–125.
8.
Hildebrandt E, Dunn JR, Perumbakkam S, Niikura M, Cheng HH. 2014.
Characterizing the molecular basis of attenuation of Marek’s disease virus via in vitro serial
passage identifies de novo mutations in the helicase-primase subunit gene UL5 and other
candidates associated with reduced virulence. J. Virol.
9.
Tischer B, von Einem J, Kaufer B, Osterrrieder N. Two-step Red-mediated
recombination for versatile high-efficiency markerless DNA manipulation in Escherichia coli.
Biotechniques 40:191–197.
10.
Niikura M, Kim T, Silva RF, Dodgson J, Cheng HH. 2011. Virulent Marek’s disease
virus generated from infectious bacterial artificial chromosome clones with complete DNA
sequence and the implication of viral genetic homogeneity in pathogenesis. J. Gen. Virol.
92:598–607.
11.
Crute JJ, Tsurumi T, Zhu LA, Weller SK, Olivo PD, Challberg MD, Mocarski ES,
Lehman IR. 1989. Herpes simplex virus 1 helicase-primase: a complex of three herpes-encoded
gene products. Proc. Natl. Acad. Sci. U. S. A. 86:2186–2189.
152

12.
Crute JJ, Lehman IR. 1991. Herpes simplex virus-1 helicase-primase. Physical and
catalytic properties. J. Biol. Chem. 266:4484–4488.
13.
Weller SK, Kuchta RD. 2013. The DNA helicase-primase complex as a target for
herpes viral infection. Expert Opin. Ther. Targets 17:1119–1132.
14.
Calder JM, Stow ND. 1990. Herpes simplex virus helicase-primase: the UL8 protein is
not required for DNA-dependent ATPase and DNA helicase activities. Nucleic Acids Res.
18:3573–3578.
15.
Cavanaugh NA, Ramirez-Aguilar KA, Urban M, Kuchta RD. 2009. Herpes Simplex
Virus-1 Helicase-Primase: Roles of Each Subunit in DNA Binding and Phosphodiester Bond
Formation. Biochemistry (Mosc.) 48:10199–10207.
16.
Graves-Woodward KL, Weller SK. 1996. Replacement of gly815 in helicase motif V
alters the single-stranded DNA-dependent ATPase activity of the herpes simplex virus type 1
helicase-primase. J. Biol. Chem. 271:13629–13635.
17.
Graves-Woodward KL, Gottlieb J, Challberg MD, Weller SK. 1997. Biochemical
analyses of mutations in the HSV-1 helicase-primase that alter ATP hydrolysis, DNA unwinding,
and coupling between hydrolysis and unwinding. J. Biol. Chem. 272:4623–4630.
18.
Saponaro M, Kantidakis T, Mitter R, Kelly GP, Heron M, Williams H, Söding J,
Stewart A, Svejstrup JQ. 2014. RECQL5 Controls Transcript Elongation and Suppresses
Genome Instability Associated with Transcription Stress. Cell 157:1037–1049.
19.
Hunt HD, Dunn JR. 2013. The Influence of Host Genetics on Marek’s Disease Virus
Evolution. Avian Dis. 57:474–482.

153

FURTHER WORK

HORIZONTAL TRANSMISSION OF RECOMBINANT VIRUSES
Based on the result of birds trials involving two recombinant viruses (LORF2-Intron and
ICP4-2, Chapters 2 and 3 respectively) that remained virulent when used to challenge birds, yet
failed to transmit horizontally to uninfected contact birds, further trials of these viruses in greater
depth to better understand the mechanism behind this loss of horizontal transmission would be
warranted. Further studies focusing on horizontal transmission would involve greater numbers of
contact birds, and trials evaluating the impact that the age of shedding birds post challenge has
on transmission. Preliminary trials, of which the results were described in Chapters 2 and 3,
involved a small number of approximately 5 contact birds housed in isolators with an average of
17 infected contacts birds. Challenged birds were infected with the recombinant viruses upon day
of hatch and immediately housed with uninfected contact birds of the same age from the same
hatch.
Future trials would contain a greater number of contact birds. It would be
preferable to increase the number of contact birds to a minimum of 10 birds per isolator to be
housed with at least 10 infected contact birds. Two different approaches could be followed in
order to compare the impact the age of shedder birds has on MD incidence in contact birds. As
previously conducted in the prior trials, challenged shedder birds would be infected and housed
with contact birds from the same hatch immediately upon day of hatch. Alternatively, a second
route would involve challenge and housing of shedder birds from separate hatches approximately
1-2 weeks prior to hatch and housing of the uninfected contact birds. This would allow shedder
birds to establish viral replication and to shed MDV into isolators prior to addition of contact

154

birds, resulting in immediate exposure to infected dander, compared to earlier scenarios in which
there was a delay in possible exposure in contact birds due to time required for infection to
become established.
For further trials, sampling of dander would be required to quantify MDV shed from
feathers. This could be accomplished through collection of whole isolator dander collected
through an air filter, as well as through collection of dander obtained through vacuuming of
individual birds. Invasive sampling of challenged birds sacrificed for collection of feather
follicles would also allow immunohistochemistry straining to identify MDV replication within
the feather follicles prior to release as feather dander. Quantification of MDV within dander and
feather follicle samples will help determine if the recombinant viruses are completely deficient in
horizontal transmission and no MDV is present, or if they are shed in dander and if so, at what
levels compared to wild-type Md5B40BAC virus.
Besides collection and quantification of MDV in feather follicles and dander, viral load
could be quantified through isolation of PBLs from both contact and challenged birds, as well as
within the spleen collected from sacrificed, moribund, or terminated birds during examination
via necropsy to determine MD.
IN VITRO REPLICATION OF UL46 RECOMBINANT VIRUSES
The Mut UL46 recombinant virus was a candidate mutation found at a high frequency of
99% within Rep 46, but this mutation had no in vivo phenotypic affect. Therefore we wish to
determine if this mutation conferred a selective advantage in vitro during replication to offer an
explanation regarding the spread of this mutation with an attenuated rep during serial passage.
Recombinant viruses containing the single point mutations of Mut UL5 and Mut UL46, as well

155

as a recombinant virus containing both mutations have already been constructed and tested in
vivo. Growth curves and in vitro competition comparing the single Mut UL5 to the wild-type
Md5B40BAC-c1 have also already been conducted and described in Chapter 6. We propose
addition of growth curves for the single Mut UL46 and Double Mut UL5+UL46 to compare the
impact of the UL46 point mutation on replication compared to the single MutUL5 and the
Md5B40BAC-c1. We also propose a head-to-head competition of viruses versus the parental
Md5B40BAC-c1. Previously the single Mut UL5 point mutation was shown to produce more
plaques in one-step growth curves compared to the parental Md5B40BAC-c1, yet in head-tohead competitions the parental Md5B40BAC-c1 outcompeted the Mut UL5 in mixed cultures,
suggesting within a mixed population it is more complicated to predict outcome than based
simply upon higher yield as determined in single viral cultures. Therefore in vitro
characterization of recombinant viruses containing the UL46 mutation will help determine if
addition of the UL46 point mutation increased in vitro replication compared to the parental virus,
providing an explanation for the presence of this high frequency mutation within the attenuated
replicate.
FURTHER CHARACTERIZATION OF IN VITRO REPLICATION OF MUT UL5
As described in Chapter 6, the Mut UL5 recombinant virus was shown to produce
significantly higher numbers of plaques compared to the parental Md5B40BAC virus, yet when
the two viruses were grown in head-to-head competitions at roughly initially equal quantities, the
Md5B40BAC virus expanded to dominate nearly 90% of the population over the course of in
vitro serial passage. To better understand this phenomenon, comparisons of plaque size and
average numbers of infected cells produced per plaque would be compared between the Mut
UL5 virus and Md5B40BAC virus to determine if the Mut UL5 virus yields smaller plaques with
156

fewer copies of MDV per plaque compared to the Md5B40BAC. Additionally, to explore the
possibility that the two viruses may produce or cause infected CEF cultures to produce, soluble
and diffusible products impacting growth of neighboring viruses, it would be proposed to explore
the impact of conditioned media on growth of the two viruses. Conditioned media from plates
infected with either Mut UL5 or Md5B40BAC viruses would be concurrently added to plates
being infected with a defined PFU of viruses and assayed via qPCR to determine MDV
quantities in infected cultures and plaque size measured via fluorescent microscopy with ImageJ
software, as would also be conducted and compared to plates infected with a known quantity of
MDV PFUs but cultured normally as described above.

157