IMPORTANT NUCLEOTIDE SEQUENCES INVOLVED IN
LATENCY OF DNA VIRUSES OF ANIMALS
Mark A. Rayfield
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
Mark A. Rayfield
This work is dedicated to my family: to my parents Martin and
Fran whose sacrifices made it all possible, to Glenn who shared the
lean years and to my wife Dot for her loving support of a very tired
and grouchy graduate student.....
I would like to acknowledge the efforts of Tess Korhnak and Dr.
Vishwanath Suryanarayanan with whom I've had the good fortune to be
associated during these studies. I would also like to take this oppor-
tunity to express my appreciation to the members of my advisory committee,
especially Dr. Kenneth I. Berns. A warm thanks is extended to Drs. Bill
Hauswirth, Bill Hollowman and Phil Laipis who have continually supplied
good natured advice and whose laboratories have been a warehouse of
materials and equipment. Finally, I would like to thank my wife Dot and
Gail Crummer, a masochist but dear friend, for their selfless contribu-
tions to the organization and drafting of this manuscript.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS . . . . . . . . . . . . iv
ABSTRACT . . . . . . . . . . . . . . vii
INTRODUCTION . . . . . . . . . . . . . 1
Molecular Biology of HSV . . . . . . . . 7
Genome Size and Composition . . . . . . . . 7
Nucleotide Sequence Arrangement. . . . . . .. 10
Endonuclease Cleavage Map . . . . . . . .. 13
Transcription . . . . . . . . . . . 14
MATERIALS AND METHODS . . . . . . . . . .. 20
Purification of Ganglion Cell DNA and RNA . . . .. 20
Viral DNA Purification . . . . . . . . . 20
Preparation of Cloned DNA . .. . . . . . . 22
DNA Restriction....... .. . . . . . . . 26
Gel Electrophoresis . . . . . . . . . 26
Isolation of Endonuclease Cleavage Products . . . .. 27
Mapping Approach . . . . . . . . . .. 28
Nick Translation . . . . . . . . . . 30
lodination of Cytoplasmic RNA . . . . . . .. 31
5' End Labeling Reactions . . . . . . . .. 31
Southern Blots . . . . . . . . . . 32
Hybridization Reaction . . . . . . . . .. 33
RESULTS . . . . . . . . . . . . . . 35
Experimental Approach .. . . . . . . ....... ... 35
Controls ............. ..... ....... . 37
Hybridization of Ganglia Nucleic Acids to HSV DNA ..... 41
CsCl Enrichment Studies ........ ............ .... 44
Mapping the L/S junction of HSV-1 F strain ......... .. 49
HYBRIDIZATIONS . . . . . . . . . . . . 59
Hybridization of ganglia RNA to the L/S junction ... 59
Cross Hybridization Experiments . . . . . . .. 60
DISCUSSION . . . . . . . . . . . .. .75
LIST OF REFERENCES . . . . . . . . . . . 88
BIOGRAPHICAL SKETCH . . . . . . . . . .. 98
Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
I.IPOPTArIlT NUCLEOTIDE SEQUENCES INVOLVED IN
LATENCY OF DNA VIRUSES OF ANIMALS
Mark A. Rayfield
Chairman: Kenneth I. Berns, Ph.D., M.D.
Major Department: Immunology and Medical Microbiology
We have investigated the molecular status of the Herpes simplex
virus (HSV) genome in latently infected human trigeminal ganglia. The
general aim of this research is to illuminate the genetics of viral
latency during recurrent infections in man. The trigeminal ganglia of
22 patients have been assayed for HSV sequences by DNA/DNA or RNA/DNA
hybridization. None of the patients manifested clinical evidence of
recurrence at the time of death nor was a history of recurrent HSV
lesions noted. Fourteen of the patients were sero-positive for HSV.
Positive serology was assessed as a complement fixation titer of 1:8 or
greater. In 18 of the 22 assays in which trigeminal ganglia were evalu-
ated for the presence of HSV, total ganglionic DNA was cleaved by endo-
nuclease digestion, transferred to nitrocellulose and probed with rich
nick-translated HSV DNA. The nucleic acid from two of the sero-positive
patients failed to demonstrate HSV specific sequences in either DNA or
RNA. Only one of the eight sero-negative patients gave evidence of the
presence of HSV DNA or RNA, the remainder were negative. Thus, there is
a positive correlation between the presence of complement fixing anti-
bodies and the presence of sequestered HSV specific sequences as
measured by these techniques. Further, from the mapping of those RNA
transcripts observed, it is clear that the regions of the genome most
frequently encountered to date are those containing sequences in the
immediate early and early mRNA species of HSV-1.
The involvement of herpes simplex virus (HSV) in recurrent infection
of man is well documented. These recurrences are most frequently mani-
fested as periodic ulcerations of the oral or genital mucosa and infections
of adjacent cutaneous tissues. Recurrent herpetic keratitis is, today,
the most common severe ocular infection in the United States and is the
principle factor involved in blinding corneal disease. Ocular infections
in 50% of the patients experiencing a primary episode will recur within
2 years. Latent virus is the probable source of infection producing
acute episodes of HSV encephalitis and meningitis in adults. It also
serves as a reservoir for virus involved in generalized herpetic infec-
tions in newborns and is potentially involved in carcinoma of the cervix.
Several lines of evidence suggest the involvement of neural tissue
with viral latency in man. Repetitive infections generally occur in
localized tissues which define together the limits of a region inner-
vated by a single sensory ganglia. Trauma to the sensory nerve will
consistently induce a viral recurrence in peripheral tissue. Reactiva-
tion occurs in greater than 90% of the patients undergoing root sections
of the trigeminal nerve; none is observed after trauma to tissue pre-
viously denervated (Carton and Kilborne, 1952). A majority of the patients
who reactivate following neurosurgical treatment for trigeminal neuralgia
have a history of recurrent herpes labialis (Pazin et al., 1978). While
such observations are only indirect evidence that neural tissue is the
site of latency, it is noteworthy that the significant efforts by
investigators to demonstrate virus in peripheral tissues during quiescent
periods have failed (Rustigan et al., 1966; Baringer, 1975)..
HSV can be liberated from cultures of latently infected trigeminal
ganglia from unselected humans (Bastian et al., 1972"; Baringer, 1974).
Using tissue cocultivation with HSV susceptible cell lines or organotypic
cultures for periods of 10-45 days in vitro, the vagus, sacral and
superior cervical ganglia may also be shown to harbor virus in some
individuals (Baringer 1974, 1975; Wlarren et al., 1978; Lonsdale et al.,
1979). EM studies show virions and cytopathology similar to that of HSV
in such cultures. The virus isolated may be identified as HSV on the
basis of both serologic (Bastian et al., 1972; Baringer, 1975) and genetic
criteria (Lonsdale et al., 1979). Similar culture techniques fail to
yield virus from the choroid plexus, trigeminal nerve root or cutaneous
tissues from known positive individuals. This argues that the ganglion
may be the sole site of latency.
Several parameters of latency can be envisioned from studies of
the virus yielded from explants. Serological evaluations of isolates
suggest that the virus remains phenotypically unaltered during latency;
reactivated virus may at some point spread to adjacent ganglia and
become latent in these; ganglia innervating remote peripheral sites may
infrequently become latent with seriologically distinct strains of the
virus; finally, both HSV type I and II may be latent and recur within the
same individual (Stevens and Cook, 1971; Baringer, 1975). Analysis of the
restriction endonuclease digestion patterns of their DNIA and gel electro-
phoresis patterns of isolates proteins confirms these observations
(Lonsdale et al., 1979; Buchman et al., 1978). Further, such studies
have shown that the isolates were unique to an individual unless they
were epidemiologically related and that within the individual multiple
isolates from different ganglia were usually genetically identical.
Infection of peripheral ganglia appears to be subsequent to viral spread
from the site of initial infection in a majority of individuals. These
observations suggest that once a sensory ganglion cell becomes latently
infected it is to some degree refractory to superinfection by a future
viral isolate. Complementation studies involving superinfection of
ganglia with temperature sensitive mutants of HSV indicate that many
harbor defective or parital viral genomes (Brown et al., 1979).
Currently two animal models exist for neural latency of HSV. Virus
inoculation in the foot pad of mice gives rise to a subsequent infection
of the spinal ganglia (Stevens and Cook, 1971)o This infection is
biphasic. A primary acute infection occurs during which high titers of
virus are found in spinal fluid and tissues. Virus titers then subside
and a chronic phase of infection follows. During this period virus may
only be demonstrated by maintenance of ganglia explants in organotypic
cultures. The chronic period probably most closely mirrors the quiescent
periods in many, but the mouse is not clearly known to undergo recurrence.
Intraocular inoculation of rabbits leads to viral spread and involvement
of the trigeminal ganglia (Stevens and Cook, 1972; Goodpasture 1929).
The infected globe will thereafter undergo periods of quiescence and
Several parallels can be drawn between chronically infected murine
neural tissue and latently infected human ganglia. Only the spinal
ganglia associated with sciatic nerves, innervating the site of inocula-
tion, harbor HSV. The virus can be demonstrated by immunofluorescence
and electron microscopy after explanation and in vitro maintenance in
organ cultures. Cultures of the sciatic nerve trunk, thoracic spinal
cord and medulla oblongata do not yield virus. Persistently infected
mice also possess neutralizing antibodies to HSV (Stevens and Cook, 1971).
The lack of virus-specific products during the chronic periods suggests
that the HSV genome may exist as a subviral unit in ganglionic cells.
Searches for products of viral replication in latent tissue confirm
the hypothesis that few ganglionic cells are directly involved with
persistence. In chronically infected mice less than 0.1% of the
ganglionic cells will act as infectious centers for the virus and those
cells positive in fluorescent antibody assays are in the great minority
(Walz et al., 1976; Stevens and Cook, 1972). The presence of HSV-
specific thymidine kinase activity in the absence of late viral protein
markers may indicate an early termination of viral replication in
chronically infected murine ganglia (Yamamoto et al., 1977). It is also
possible that some species of mRNA are very stable or that transcription
continues from "remnants" of the viral genome in some cells.
HSV initially inoculated into the mouse via the ear can be stimu-
lated to reactivate through localized minor trauma to the ear (Hill et
al., 1975). During the quiescent periods in this recurrence model, virus
is absent from the peripheral tissue, but can be found in the 2nd, 3rd
and 4th cervical ganglia. Several forms of stimuli are effective: UV
light, cellophane tape, and stripping or swabbing the ear with xylene.
All such agents produce a localized inflammatory response and provoke
increased levels of prostaglandins within the skin. It is uncertain
what role these factors may play during recurrence. It is clear that
the effective agents do induce changes within the ganglia as evidenced
by increases in infectious virus and tritiated thymidine uptake within
the ganglia 2 to 4 days following trauma to the ear (Hill et al., 1975).
Similar observations have been made in HSV latently infected rabbits.
In this model, viral spread may be clearly followed from the site of
inoculation to the peripheral nervous system and then to the central
nervous system in individuals where meningitis and encephalitis ensue.
At each level of spread the infection is first manifest in neurons and
then adjacent supportive cells, such as Schwann cells or satellite cells
(Goodpasture 1929; Baringer et al., 1974). The rate of viral spread
within the nerve trunk approaches that of retrograde transport of pro-
teins within the ganglia (Stevens 1975). The kinetics of viral infection
and the presence of virus in intra-axonal vesicles has prompted investi-
gators to propose that intra-axonal transport is the principal mode of
viral spread (Cook and Stevens, 1973; Kristensson et al., 1974; Baringer
1975). Seventy percent of the trigeminal ganglia in latent rabbits yield
virus upon explant while cultures of the conjunctiva, lacrimal gland,
cornea, iris, fifth cervical nerve and brain tissue are negative. Immuno-
flourescent antibody studies on serial sections of rabbit trigeminal
ganglia show 1/500 neurons infected with HSV during the first week
following corneal inoculation. This ratio of immunopositive to immuno-
negative neurons falls to approximately 1/2000 to 1/5000 by 20 weeks
following inoculation. At this later date the yield of infectious virus
from explanted organ cultures is still on the order of 90% (Racjani 1978).
These findings closely parallel the observations of Baringer and
Sworeland (1973) in man. Lastly, single abnormal ganglion cells
surrounded by a cuff of mononuclear cells may be seen infrequently as
late as 11 months post infection. The nuclei of such cells contain HSV
particles and their cytoplasm is highly vacolated with virions contained
within intra-axonal vesicles (Baringer 1974). It is unclear whether
such cells are the product of the original infection or remnants of a
recent recurrence. The mass of this evidence argues that the ganglionic
neuron serves as the site of viral latency. The findings also suggest
that latency is not a totally statess" event but is "dynamic" to the
extent that, with even long periods of clinical quiescence, a few neurons
continue to act as infectious centers within the ganglia.
Although recurrence appears to be a natural phenomenon in rabbits,
it can also be artificially induced. Mechanical stimulation is usually
achieved through neurosurgical manipulation of the trigeminal nerve
root. Following such treatment, virus can be isolated from the tears of
80% of the animals within 48 hours (Nesburn et al., 1976). These
findings closely parallel the observations made by Carton and Kilborne
(1959) and Pazin et al (1978) in man.
In situ hybridization of HSV specific complementary RNA probes to
latent murine ganglia genetic material show viral nucleic acid sequences
in a small minority of the neurons (Cook et al., 1974; Zur Hausen and
Schulte-Holthausen, 1975). Analysis of hybridization experiments by
Puga et al (1978) demonstrates approximately 0.1 genome equivalents of
HSV DNA per cell in such ganglia. The quantity of viral RNA fell below
the limits of detection during the chronic period. The techniques
employed in this work were sensitive enough to detect 1 genome equiva-
lent of viral RNA per 2000 cells but at this sensitivity as many as 400
RNA copies per cell would go undetected.
Finally, using a nick translated 3H-labeled HSV DNA probe, viral
specific RNA sequences can be demonstrated in human ganglia sections
with in situ hybridization (Galloway et al,, 1979). Such sequences may
be found in the sacral, thoracic and lumbar ganglia, confirming previous
observations that an individual could be latent at multiple ganglionic
sites. In each case only 0.4-8.0% of the ganglionic neurons possessed
HSV specific nucleic acid sequences..
Molecular Biology of HSV
The infectious HSV particle has evolved a rather complex architec-
ture. A 1000 A nucleocapsid of acosahedral symmetry encloses the double
stranded linear DNA chromosome and various phosphoproteins of the viral
chromatin (Russell et al., 1962; Becker et al., 1968; Spear and Roizman,
1972). This core is itself encompassed by a loose matrix of glycoproteins
and carbohydrates of viral origin. Ultimately, the entire structure is
surrounded by a lipid envelope, giving the infectious virion a diameter
of roughly 1800 A. Quantitatively the mature particle is comprised of
approximately 70% protein, 22% lipid, 7% nucleic acid and 2% carbohy-
drate (Scott and Tokumaru, 1964; Callaghan-et al., 1976). Although the
viral genome has the potential to encode and modify through methylation
or glycosylation all known virion associated proteins, the lipid moiety
of the envelope appears to be principally derived from the phospholipids
of the infected cell's nuclear membrane (Watson et al., 1973; Roizman
et al., 1974). All known protein species associated with the envelope
arise from either virus specified de novo synthesis or modification of
cellular proteins (Kaplan 1973; Spear 1972).
Genome Size and Composition
The DNAs extracted from sucrose gradient purified virions of HSV-1
and HSV-2 differ in bouyant density by 0.002 g/cm3 (Goodheart et al.,
1968; Kieff et al., 1971). Goodheart et al (1968) and Plummer et al
(1970) determined these densities to be 1.725 g/cm3 for HSV-1 and 1.727
g/cm3 for HSV-2. Similar estimates were obtained by Graham (1972) using
E. coli DNA as a density reference. These values predict a G + C ratio
of 66.3% for HSV-l and 68.4% for HSV-2 (Goodheart et al., 1975). Kieff
et al (1971) established 1.726 and 1.728 g/cm for the respective bouyant
densities in CsCl isopycnic gradients using SPOl DNA as a marker. The
resulting 67% and 68% G + C ratios were confirmed by the finding of a
1C difference in the Tm values for HSV-l and HSV-2. While Kieff
reported the Tm of type 1 to be 83C and type 2 to be 84C, Graham et al
(1972) found only a 0.1C difference in Tm values of 82.6 and 82.7C.
Both serotypes of HSV have a sedimentation rate in neutral sucrose
density gradients of 55 S with a co-responding molecular weight of
995x106 daltons (Kieff et al., 1971; Sheldrick and Berthelot, 1974).
The contour length for the linear duplex molecule, based on EM studies,
is approximately 100x106 daltons (Sheldrick and Berthelot, 1974; Wilkie,
1973). The estimated molecular weight of HSV-1 circles is 943x106
daltons; if one considers a 3% redundancy in their terminal overlap this
would suggest a value for the linear molecule of 973x106 daltons
(Grafstrom et al., 1974). In a comparative analysis of restriction endo-
nuclease fragments from 80 isolates of HSV-1 Buchman et al (1978) found
a variation of only 2% in the determination of overall molecular weight
The similarities between HSV-1 and HSV-2 are apparently not limited
to a cursory examination of molecular weight and composition. Liquid
and filter hybridization analysis of their DNAs suggests an 85% homology
over 46% of their genomes; the remaining 56% of their respective genomes
is highly variable with little homology (Kieff et al., 1971). Nearest -
neighbor analysis for the two gives virtually identical frequency patterns.
However, a 2.2% difference in G + C ratios would predict a mismatch of
at least 3850 base pairs between serotypes (Subak-Sharpe et al., 1973).
Polypeptide mapping studies of intratypic recombinants give strong evi-
dence of a collinearity in gene ordering within HSV-1 and HSV-2 chromo-
somes. The reported viability of intratypic recombinants and the ability
of ts mutants of HSV-1 to complement ts mutant of HSV-2 also suggest
little divergence within functionally critical regions of the genomes
(Subak-Sharpe et al., 1972; Morse et al., 1977). Such evidence indicates
extensive similarity in the gross architecture of the two genomes and
argues for a distant but common lineage.
An interesting characteristic of the structure of the HSV chromo-
some is the presence of single-strand interruptions in the DNA. When
alkali denatured DNA from purified virus is sedimented on alkaline
sucrose density gradients numerous bands of single-stranded DNA may be
observed. These bands correspond to fragments of 7x106 daltons to
intact strands 48x106 daltons in weight (Kieff et al., 1971; Wilkie 1973).
Frenkel and Roizman (1972) have distinguished 6 classes of fragments,
ranging from 10x106 to 39x106 daltons, in denatured HSV-1 DNA. The
kinetics of reassociation within the intact size, 39x106 daltons, class
indicated this class was composed of elements representing a single unique
strand of DNA. In conflict with this interpretation are the findings of
Wilkie (1973) which suggested a randomly gapped genome and nonunique
intact strands. To date no biological function has been ascribed to the
single strand gaps and the controversy over random vs. defined sizes has
not been reconciled. Newly synthesized viral DNA has been reported to
contain short stretches of RNA (Biswal et al., 1974). It is unlikely
that such RNA stretches alone would account for these gaps since the
sedimentation profiles in neutral sucrose gradients are indistinguishable
for DNA denatured with formamide versus alkali (Roizman, 1979). The
consequences of a randomly gapped genome in the absence of a mechanism
for repair are self evident. It can be speculated that abortive trans-
cription due to an early truncation of the template could play a role in
viral latency or oncogenic transformation. It is noteworthy that no
repair mechanism has yet been clearly demonstrated.
Nucleotide Sequence Arrangement
If following denaturation, single strands of HSV DNA are permitted
to reanneal a high degree of self annealing can be observed by electron
microscopy. Among those strands that have folded back on themselves
three recognizable structures can be identified: a linear form with
extensive folding at one end, a single stranded circle with a heavily
folded region and a structure consisting of two single stranded circles
joined by a duplex region (Sheldrick and Berthelot, 1974). The forming
of two single stranded circles connected by a double stranded region is
indicative of the presence of two inverted repetitions of the termini
within the molecule. In a similar manner duplex circles may be formed
after treatment of hative HSV DNA with exonuclease III (Sheldrick and
Berthelot, 1974; Grafstrom et al., 1975). Such behavior is consistent
with the presence of natural repetitions at the single strand termini.
Thus the HSV chromosome is bracketed by direct repeats and consists of
two unique gene regions of differential length separated by two tandemly
arranged inverted repetitions of the termini. One implication of such a
model is that following concatameric replication of the molecule the
relative orientation of the two unique gene regions may be inverted
through recombination between the external and internal repetitions
(Thomas and MacHattie, 1967; Sheldrick and Berthelot 1974).
Partial denaturation mapping of the chromosome confirms this archi-
tecture and permits the sizing of the various regions. The long unique
gene region comprises 70.2% of the genome while the shorter unique region
involves 9.4%. The direct and inverted repeats bracketing the long region
each represent 6% of the genome length and can be further divided into
two subregions: the a, a' reiteration and the b,b' reiteration, which
is found only in the long segment. In like manner repeats associated
with the short unique region can be designated a'c' and ca, respectively;
comprising 4.3% of the genome. The a,a' sequences are held in common
between the long and short gene regions while the c, c' and b, b' inver-
sions are limited to the short and long regions respectively (Wadworth
et al., 1975). Partial denaturation studies and subsequent endonuclease
digestion mapping further confirm that the long and short components are
able to invert in their relative orientations. Thus, as originally pre-
dicted by Sheldrick and Berthelot (1974), within a population of HSV DNA
four conformations of the molecule can be determined which differ in the
relative orientations of the long and short unique regions (Hayward 1975;
Delins and Clements, 1976; Wilkie, 1976). The diagram below illustrates
One consequence of such an arrangement is that endonuclease
cleavage fragments spanning the L/S junction will be found in a 0.25
molar ratio relative to unique fragments and the terminal fragments will
be present in a 0.5 molar ratio. This prediction was confirmed by
Hayward (1975) in an analysis of the HindIII and Eco R1 digestion frag-
ments of HSV-1. Morse et al (1977) has suggested that another conse-
quence of this genome architecture is that not all conformations may be
functionally equivalent. If one aligns the amps of intertypic
|TR, IR L^ I I R
oa b b'o'a'c' c a
E3 ---- ---=4-- -1---E
0 10 20 30 40 50 60 70 80 90 100
1 I I I -f-----1----t-- - I I -
L & I
L&S -_ "
Figure 1. Inversion of HSV DNA
recombinants to give the least number of crossover points, the resulting
progeny could only have originated from recombination between parents
which were both in the prototype configuration. Such an approach does
not take into account the contribution of double cross over events
(Morse et al., 1977; Roizman, 1979). To date analysis of recombination
frequencies between genetic markers in the long and short regions has
shed little light on the validity of this assumption.
Endonuclease Cleavage Map
In addition to the size variation in fragments spanning the L/S
joint, arising from inversion of the unique components, two other
classes of heterogenity are evident in cleavage maps of both HSV-1 and
HSV-2. One type of heterogenity maps solely to the unique regions and
appears to define viral strains. The differences within this class
arise from the presence or absence of specific endonuclease cleavage
sites and are found throughout the long and short components (Skare,1975;
Hayward, 1975; Locker and Frenkel, 1979). In a study of eighty HSV-1
isolates, Buchman (1978) determined 19 out of 60 cleavage sites which
could be either present or absent; suggesting the possibility of at
least 219 viral strains within the population. At this level of analysis
the cleavage maps of HSV-1 and HSV-2 are quite distinct. This first
class of heterogenity is less frequent in HSV-2 (Roizman, 1979). The
loss or gain of a single cleavage site need not give rise to a new
clinically or immunologically recognizable viral strain. Thus the number
of immunologically identifiable strains need not approach 219 value.
The ability of HSV to remain latent within an individual with inter-
mittent recurrences would provide a natural reservoir for the accumula-
tion of nonlethal mutants within the population. The lower frequency of
heterogenity in HSV-2 may simply mirror a smaller reservoir within the
populace. The second class of heterogenity maps to the termini of the
long unique region (Wagner and Summers, 1978; Skare and Summer, 1977)
of HSV. These differences derive from the insertion of a 28Glbp segment
into the terminal reiteration of the L component in single or multiple
copies giving rise to a series of minor fragments with a distinct 2xlO5
dalton increment in size (Wagner and Summers, 1978). Here again a
variance between viral strains may occur since a 3.3x105 dalton
increment in size is observed between minor fragments in HSV-1 (F) DNA.
In addition, while the minor fragments from the termini of HSV-1 differ
by 255 base pairs, those arising from the L/S junction differ by only
125 base pairs (Locker and Frenkel, 1979). Subsequent work with HSV-l
(F) has directly linked the tandem addition through recombination of the
501 base pair 'a' segment to the L termini with the generation of those
minor bands (Mocarski et al., 1980). Size variability has also been
reported for endonuclease fragments containing the c, c' repeat regions
(Lonsdale et al., 1979). This heterogenity can again be traced to the
copy number of a 23 base pair repeat present within the c regions. The
reiteration can occur in a variable copy number between viral strains,
between clones of same viral strain and lastly between the c and c'
reiterations of the same strain (Watson and Van de Woude, 1982; Rixon
and Clements, 1982).
Once a permissive host cell is infected, viral replication rapidly
follows in a well-ordered seuqence of events. With uncoating, transcrip-
tion of defined regions of the viral genome proceeds in a highly regu-
lated temporal manner (Clements et al., 1977; Jones and Roizman, 1979).
Transcription of host cell specific RNA (C-RNA) continues at a reduced
rate, but synthesis and processing of cellular mRNA is rapidly inhibited,
as is the appearance of mature cellular rRNA. This inhibition appears
to be coded for by the earliest viral transcripts (Wagner and Roizman,
1969; Kaplan, 1973). For the most part, the reduction in the rate at
which mature rRNA appears is brought about by the interference of
infection with its maturation; synthesis of the 45 s RNA precursor con-
tinues at roughly the same level as that of C-RNA (Wagner and Roizman,
0 10 20
30 I4 I5 6 0
II I I
I I I I l i
\-^- ----- I---- ^+4H-1--I-
30 40 50 60 70 80 90 1C
i- | I -- |I | -- | -- --I --- |!
Figure 2. Restriction Map of HSV-1 and HSV-2 DNAs (Roizman, 1979).
1969)o However, while there is a stable accumulation of C-RNA within
the infected cell, inhibition of host mRNA synthesis is completed to the
extent that no de novo synthesis of cellular mRNA can be observed late
in infection (Kaplan, 1973). As the rate of host cell DNA synthesis
gradually falls off, it is rapidly surpassed by that of nacent viral DNA.
By 7 to 8 hours post infection, host DNA is no longer replicated. This
inhibition is concomitant with that of host cell protein synthesis and
thus may mirror the loss of proteins specifically required by the cell's
replicative machinery (Roizman et al,, 1974). The peak rate of synthesis
for the earliest viral polypeptides is reached between 2 and 4 hours post
infection. The chief function of these proteins is to govern the change
over from cellular to viral macromolecular synthesis. There is a strin-
gent requirement for these immediate early polypeptides (a-proteins)
prior to the production of a second and larger group of proteins, the
early polypeptides (s-proteins), which appear 5 to 7 hours post infection.
These g-polypeptides augment the replication of the HSV genome. They
also govern a' gradual reduction in a-protein synthesis and pave the way
for production of the late structural proteins (y-polypeptides) of the
HSV virion (Honess and Watson, 1974; Pereira et al., 1977).
Viral mRNA is transcribed within the nucleus (Wagner and Roizman,
1969), is polyadenylated at its 3' end (Bachenheimer and Roizman, 1972;
Silverstein et al., 1976) and carries a 5' cap (Moss et al., 1977).
Prior to viral DNA synthesis two classes of viral mRNA, the immediate
early and early transcripts, are abundant and are transcribed at least
in part by host cell RNA polymerase II (Jones and Roizman, 1979; Clements
et al., 1977). Translation of these two classes of transcripts gives
rise to the aforementioned a and polypeptides, respectively. The
immediate early mRNA represents approximately 12% of genome and is made
in the absence of previous viral polypeptide synthesis. It can be iso-
lated in abundance from the cytoplasm of infected cells treated with
cyclohexamide (Jones and Roizman, 1979). These transcripts range from
1,500-5,500 nucleotides in size and map preferentially to the terminally
reiterated portions of the genome (Clements et al., 1977, Holland et al.,
1979). The second class of mRNA which is abundant before viral DNA
synthesis have a stringent requirement for a polypeptide synthesis prior
to its production. The viral DNA polymerase falls within this group.
Early transcripts map throughout the HSV genome and the -size of their
nuclear precursors differs little from that of the polysomal bound popu-
lation suggesting that their promoters map close to the structural genes
(Holland et al., 1979). Together the sequences transcribed prior to DNA
synthesis represent approximately 25% of the genome.
Comparative inhibition studies indicate that more than one polypep-
tide in the a and groups is involved in regulation of transcription
and the transition from a to p to y protein synthesis. For example
canavanine selectively inhibits subsets of these proteins and permits
partial transition from immediate early to early and then to late trans-
cription. As with early transcripts, late transcription occurs through-
out the genome with approximately 50% of the viral RNA at 6 hours post
infection having a length of 5,000 to 10,000 nucleotides. At this time
transcription of most early viral RNA continues with y mRNA species
representing approximately 20% of the accumulated viral mRNA.
The complex overlapping hybridization patterns of viral mRNA to the
HSV genome suggest that an elaborate series of interlocking controls
regulate transcription and translation with multiple RNA species derived
from limited regions of the genome and nucleotide sequences common to
several messengers. As suggested above the immediate early promoters
are compatible with the existing host cell transcriptional machinery yet
early promoters appear to have been modified by one or more a viral
proteins. The insertion of the 5' promoter region of immediate early
(IE) mRNA 3 into sequences upstream from the early thymidine kinase
structural gene permits its transcription as an immediate early species
(Post et al., 1981). As with other mammalian DNA viruses both mRNA
splicing and overlapping coding regions has been documented. With
respect to overlapping coding regions, the 5' termini of two early mRNA's
have been mapped to the region of the HindIII cleavage site at 0.586.
The regulatory signals for transcription of the smaller 1.2 kb mRNA are
found within the larger 5.0 kb transcript. Both mRNAs are unspliced.
The 5.0 kb species encodes a 136,000 dalton protein while the smaller
1.2 kb messenger codes for a 38,000 dalton protein. These early mRNAs
have a 3' co-terminus (Anderson et al., 1981; McLauchlan and Clements,
1982). Analysis of the co-terminus revealed the sequence AAUAAA with an
A + T rich region 3' to the sequence. This is the general format for the
3' termini of all early HISV mRNAs studied to date.
The IE mRNA-4 and IE mRNA-5 share a common 5' sequence. In this
instance the promoter and capped 5' termini map to the reiterated
sequences flanking the short unique region. Both messages share a
common 247 bp leader with a single splice located within the reiterated
sequences. A 'TATA' box is located approximately 25 bases 5' to the
leader and the common splice appears to function only to remove the
introns from a 5' untranslated region (Rixon and Clements, 1982). Since
these are the only messages located in this portion of the genome,
splicing does not increase the coding capacity of this region. It should
be noted that between IE mRNA-3 and IE mRNA-5 is a stretch of 800 nucleo-
tides which is not transcribed. This region may include the origin of
DNA replication (Vlanzy and Frenkel, 1981; Watson and Van de Woude, 1982).
Lastly, comparison of the 5' terminus of the IE mRNA-5 gene to that of
early thymidine kinase gene revealed that the regulatory sequence GGCGATTC
was absent in the immediate early gene while present 80 base 5' to the
early transcribed region (Watson, 1982). Eco R1 endonuclease cleavage
at this sequence causes reduced levels of thymidine kinase expression
(Wigler et al., 1977; McKnight and Gavis, 1980) and the sequence is
similar to the RNA polymerase II regulatory signal GGPyCAATCT described
by Benoist et al (1980). Thus transcriptional control at the promoter
level must distinguish between immediate early, early and late signals
and between origins of transcription within the same class; as in the
case of the two early promoters described above with a 3' co-terminus.
The rationale for studying the molecular status of the latent herpes
genome is in reality twofold. In a general sense, if the latent genome
is present as a subviral unit within the cell it provides a well defined
gene region of which specific questions concerning gene content and
transcription may be asked. In theory, such a genetic system would be
subjected to many of the restraints governing the activities of the genes
of its eukaryotic host, regardless of whether the viral DNA was inte-
grated or funcitonally an episome. The more direct concern of this study
is to approach the question of whether the HSV genome is functionally
dormant during its latency or does itself provide some necessary
activity. An assay for specific regions of the HSV genome should deter-
mine which portions of the viral DNA are most frequently retained in
latency while an evaluation of viral RNA will determine which regions of
the genome are most frequently expressed. The retention and expression
of certain genomic regions common to all HSV latently infected ganglia
argue that the virus may play an active rather than passive role during
MATERIALS AND METHODS
Purification of Ganglion Cell DNA and RNA
Teased ganglia preparations were washed 2 times in 1 ml of lXSSC
(150 mM NaCI and 15 mM Na2C6H507, pH 7.6), Dounce homogenized and trans-
ferred to siliconized 1.5 ml Eppendorf centrifuge tubes. The homogenate
was spun at 10,000 rpm for 2 minutes to pellet nuclei from which the DNA
was extracted. The resulting supernatant was extracted 2 times with
phenol-cresol-chloroform (5:1:1 mixture) and ether extracted 3 times.
From the extracted supernatant, total cytoplasmic RNA was precipitated
with diethylpyrocarbinol (DEPC) treated ethanol and stored at -20C.
For isolation of nuclear DNA, the pellet was first resuspended in 1 ml
lXSSC, 1% sodium dodecyl sulfate (SDS) and then treated with pronase
(1 mg/ml) at 37C overnight. The preparation was then phenol extracted
2 times, ether extracted 3 times and RNase A treated (20 mg/ml) for 30
minutes at 37C to degrade nuclear RNA. This was followed by 2 addi-
tional phenol extractions. The DNA was then precipitated with ethanol
and stored at -20C. This procedure yielded approximately 26 Pg of DNA
pre ganglia preparation.
Viral DNA Purification
HSV-1 DNA was isolated as previously described by Holland (1979).
Briefly, confluent monolayers of human epidermoid carcinoma cells (HEP-2)
were infected with HSV-1 at a multiplicity of 10 PFU/cell and incubated
in Eagle's minimum essential medium with 10% calf serum at 37C (Roizman
1968). Thirty hours post infection the monolayers were scraped into
medium, which was clarified by spinning at 1,000 xg for 15 minutes. The
clarified medium was stored on ice and the cell pellet resuspended in
cold reticulocyte saline buffer (0.01 m NaCl, 0.0015 m MgCl2, 0.01 iM Tris,
pH 7.4) and 0.5% Triton. The cells were lysed by 10 strokes in a Dounce
homogenizer. The cell homogenate and clarified media were then each
spun at 10,000 xg for 15 minutes to remove nuclei and cell debris. The
resulting supernatants were pooled and from these HSV-1 virions were
pelleted by centrifugation at 33,000 xg for 30 minutes. Virions were
resuspended and lysed in 10 mM Tris, pH 7.5, containing 50 mM EDTA, 0.5%
SDS and 1% sarkosyl. The lysate was digested overnight at 37C with
pronase at a final concentration of 2 mg/ml (Walboomers and Ter Schegget,
1976). The digest was phenol extracted twice and ether extracted 3 times.
HSV-1 DNA was purified by banding twice in an isopycnic ethidium bromide -
CsCl gradient. The starting density of CsCI was 1.566 g/cm with 100
tg/ml ethidium bromide and centrifugation was for 48 hours in a Ti50
rotor at 45,000 rpm (Pater et al., 1976).
Partially purified HSV-l F strain DNA was also received from Dr.
Saul Silverstein, Columbia University, New York, NY. These preparations
had been treated in the following manner: Nucleocapsids were purified
from cytoplasmic extracts of infected Vero cells by velocity sedimentation
in sucrose gradients. The virions were lysed in the presence of 0.5% SDS,
0.5% sarcosyl and 10 mM EDTA. The lysate was then extracted with phenol,
chloroform and 2% isoamyl alcohol to remove proteins and lipids. The DNA
was then precipitated with ethanol and stored (Spear and Roizman, 1972).
Viral DNA preparations received in this form were resuspended in
2XSSC and further purified on CsCl density gradients. A single viral
band was obtained at a density of approximately 1.728 g/cm which was well
separated from contaminating, cellular DNA and RNA in the pellet (see
Figure 3). The HSV-l DNA was not highly fragmented as shown by ultra-
centrifugation on 5-20% alkaline sucrose gradients using 32P-labeled
form II SV40 as a size marker (see Figure 4) (Kieff et al., 1971;
Pignatti et al., 1979).
Preparation of Cloned DNA
Two clones of the reiterated junction between the long and short
unique regions of HSV-l DNA designated pRB104 and pRB115 were obtained
from B. Roizman, University of Chicago, Chicago, IL. pRB104 was
generated by the insertion of the Bam HI fragment SP2 of F strain HSV-1
into the Bam HI site in the tetracycline resistance gene of the Escherichia
coli plasmid pBR322; pRB115 was derived in like manner but contained the
Bam HI fragment SP1 of F strain HSV-1 (see Figure 5). The plasmids were
provided in E. coli strain C600SF8 (Post et al., 1980). Stocks were
maintained in 50% glycerol at -70Co Prior to growth, single colony
isolates were selected for ampicillin resistance, and grown overnight at
37C in Luria broth with 100 mg/l ampicillin. Generally 20 mis of the
overnight culture were used to inoculate 2 liters of broth for the pro-
duction of DNA stocks. The cultures were induced with the addition of
200 pg/ml chloramphenicol 15 hours before harvest.
Upon harvest the cultures were centrifuged at 10,000 xg for 15
minutes to pellet the bacteria. The bacterial pellet was washed once in
a solution of 10 mM EDTA and 50 mM Tris, pH 7.5 and then resuspended in
a solution of 25% sucrose, 10 mM EDTA, and 50 mM Tris HC1, Ph 7.5. All
subsequent extractions were conducted at 4C. The bacterial cell wall
was digested with egg white lysozyme (500 pg/m1). After 20 minutes in
the presence of lysozyme, (1 pg/ml) DEPC was added to inhibit breakdown
of nucleic acids and the cells were gently lysed in 0.5% Triton, Low
molecular weight DNA was extracted in the manner described by Hirt (1967).
C. 150 1.720
120 1.700 "
..... ,,, /
30 1'.// / ,640
5 10 15 20 25
Figure 3. CsCI isopycnic gradient purification of HSV DNA
5- 20%Alkuline Sjcrose
S./ 8 1
4 8 1. -
(3H) HSV DNA
\ i/ 32P) SV40 L.NA
Figure 4. Alkaline Sucrose Gradient analysis of HSV DNA
0 10 20 30 40 50 60 70 80P 90 100
1 S It II i I P 5
-- 1E-15 mRNA
SP2 ---------- 104
Origin of pRB104 and pRB115 DNAs
The Hirt supernatant was brought to 0.5 M NaCl and polyethylenegycol
(PEG) 6000 was added to a final concentration of 5%. The plasmid DNA
was permitted to precipitate overnight at 4C. -The resulting pellet was
isolated by centrifugation and resuspended in 50 mn1 Tris, pH 7.6 and 10
mM EDTA; excess PEG was removed by chloroform-phenol extraction and the
nucleic acids were ethanol precipitated in 250 mM sodium acetate, pH 4.7,
at -20C. Residual RNA in the resultant pellet was degraded by RNase A
treatment and form I plasmid DNA was subsequently isolated by ultracen-
trifugation on CsCl-ethidium bromide isopycnic gradients (LePecq, 1970).
Viral DNA (3 ig) was digested with either EcoRl, HindIll or Hpal
endonuclease in a 50 pI reaction mixture at 37C for 2 hours; 1 unit of
enzyme per pg DNA was added per hour. The enzymes were obtained from
Bethesda Research Laboratories, Inc., Rockville, MD. Reaction buffers
used were in accordance with the instructions provided by the supplier.
Ganglionic DNA was digested with EcoRI or HindIII in a 100 il reaction
under the above conditions.
To develop a fine structure restriction of the BamHl fragment
spanning the joint region of HSV-l DNA, cloned viral DNA was digested by
BamHl, AluI, BstEll, TaqI, HincI, Aval and Smal endonucleases. All
cleavage reactions were conducted in accordance with the instructions
provided by Bethesda Research Laboratories, Inc. BstEll and TaqI endo-
nuclease digestions were carried out at 60C: all other digestion reac-
tions were done at 37C.
Restriction enzyme digestions were phenol extracted and then electro-
phoresed on 0.5% or 1% agarose vertical slab gels at 2 V/cm. Viral DNA
was run on 40x15x3 cm 1% slabs with 1 cm wells, each receiving 500 ng
HSV-1 DNA in 10 pl of loading solution. Stock loading solution contained
0.05% bromphenol blue, 0.05% xylene cyanole and 50% glycerol in 0.01M
Tris, pH 7.4. This was diluted 1 to 3 with the digests prior to loading.
Digests of cloned viral DNA were electrophoresed in the same manner.
Cellular DNA was run on 14x12x3 cm 0.5% agarose slabs with 1 cm wells;
10 pg of DNA was loaded as above per well. The running buffer was con-
tinuously circulated and contained 5 mM sodium acetate and 1 mM EDTA in
40 mM Tris, pH 7.8.
For size~determination and mapping, the digests of cloned viral DNA
were also electrophoresed on 6% acrylamide 0.15% bisacrylamide or 12%
acrylamide 0.30% bisacrylamide gels. DNA (3 pg) was loaded as above
per 5 mm well on 40x15x3 cm gels and 1 pg per 5 mm well on 40x15x1.5 cm
vertical slabs. The running current was 3 volts/cm0 The continuously
circulated buffer was 40 mM sodium acetate and 2 mM EDTA in 50 nfmM Tris,
Isolation of Endonuclease Cleavage Products
Endonuclease cleavage products were isolated and purified following
agarose gel electrophoresis in the manner described by Finkelstein and
Rownd (1978). Ethidium bromide stained gels were illuminated with a UV
lamp and the fragment band sliced out. Gel slices were finely ground
and suspended in 200 pl of 100 mM Tris, pH 5.95. Agarose (Calbiochem-
Behring LaJolla, California) was added to a final concentration 0.5 pg/pl
and the suspension was incubated for 2 hours at 37C. The degraded
agarose was removed by centrifugation at 15,000 xg for 10 minutes.
Following two successive phenol and ether extractions the DNA was con-
centrated by ethanol precipitation.
Smaller cleavage fragments were extracted from 6% acrylamide gels,
following the method devised by Maxam and Gilbert (1980). Gel slices
were minced and suspended in a solution of 500 mM ammonium acetate,
10 nmM magnesium acetate, 1 mM EDTA, 0.1% (wt./vol.) sodium dodecyl
sulfate and 10 pg/ml yeast tRNA. This slurry was incubated at 37C for
10 hours and then the residual acrylamide was removed by filtration
through glass wool. The filtrate was phenol extracted once, ether
extracted once and then the extracted DNA was concentrated by ethanol
Bulk purification of cleavage fragments was done by affinity chroma-
tography (BiUnemann and MUller, 1978). Generally 200 jg of digested DNA
were loaded onto a malachite gree substituted bisacrylamide gel
(Beohringer Mannheim Biochemicals, Indianapolis, IN) column in the pre-
sence of 10 mM sodium phosphate and 1 mM EDTA, pH 6.0. The fragments were
eluted through a 0-2 molar gradient of sodium perchlorate in the same
phosphate buffer (see Figure 6). The sodium perchlorate was removed
from the eluate by dialysis against 10 mM Tris, 1 mM EDTA; pH 7.0. The
DNA was the concentrate from the dialyzed eluate by ethanol precipitation.
Endonuclease digestion maps for pRB104 and RB115 were determined
by two methods: double digestion mapping and partial digestion mapping.
Where applicable, new sites were mapped relative to known sites in
reactions where two or more digestions were done to completion before
gel electrophoresis of the fragments. All reactions and subsequent gel
fractionations were carried out as previously discussed. In this manner
the digestion sites for BamHl, BstEll, Hincli, Alul, and Taql were con-
firmed and ma-ped relative to one another. The size of the endonuclease
1.3373 7 7
I 3!37 .2
35 30 25 20 15 10 5
Figure 6. Affinity Chromotography of BamHl digest of pRB104
cleavage fragments was determined by comparison to known digestion
standards of adeno-associated virus (AAV) DNA, lamda phage DNA, or
pBR322 DNA, which were run in parallel wells on the .'same gels (Berns and
Hauswirth, 1978; Cheung et al., 1980). The numerous cleavage products
arising from restriction by either Aval or Smal were oriented relative to
the Alul and Taql sites through partial digestion reactions in the manner
described by Smith and Birnstiel (1976). End labeled Alul or Taql frag-
ments were isolated and partially digested in the presence of 1 unit/Pg
DNA of Aval or Smal endonuclease under standard conditions. Aliquotes
were withdrawn from cleavage mixture and prepared for gel electrophoresis
after 5 minutes, 10 minutes, 20 minutes, 40 minutes and 80 minutes of
1 pg HSV-1 F strain DNA was incubated. for 45 minutes at 15C in a
50 pIl reaction mixture containing 30 rmMdGTP, 30 mMdTTP, 30 mMdATP (sigma),
100 PCi[a 32P]dCTP (400 Ci/m mole; New England Nuclear), 5 mM MgCl2, 4
units DNA-polymerase I, 10 mM -mercaptoethanol, 0.05 mg/ml bovine serum
albumin, 50 mM Tris, pH 7.8, and 4.5 units of DNA-polymerase I (Boehringer
Mannheim Biochemicals, Indianapolis, IN). The reaction was stopped by
the addition of 100 ji of a solution containing 100 mM EDTA, 100 ug
sonicated calf thymus DNA and 10 mM Tris, pH 7.8. The mixture was
extracted with phenol and ether. Finally free [a 32P]dCTP was separated
from labeled DNA by passage through a Sephadex G75 column (IcmxlOcm).
This procedure usually yielded a viral probe with a specific activity of
approximately 3x108 cpm/ug. Cellular DNA was labeled in the same manner
except that the reaction volume was increased to 100 wl and 10 pg DNA
was labeled per reaction in the presence of 0.5-2 ng activated DNase I.
DNase I was activated by incubating the enzyme for 2 hours at 0C in a
solution containing 10 mM Tris, pH 7.6, 5 mM MgCl2 and 1 mg/ml bovine
serum albumin. DNase I was obtained from Worthington Biochemical Cor-
poration, Freehold, NJ and stock solutions of 1 mg/ml in 0.01 n HC1 were
stored at -20C (Rigby et al., 1977).
iodination of Cytoplasmic RNA
Ethanol precipitated cytoplasmic RNA was resuspended in 50 pl of
DEPC treated distilled water and transferred to a 1.5 ml Eppendorf
centrifugation tube containing 4 mCi 125I1 sodium iodide. To this, 20 Il
of 0.2M sodium acetate (pH 4.7) and 5 mM thallic chloride was added.
The tube was sealed and incubated at 70C for 20 minutes. 1251 sodium
iodide with a specific activity of 350-600 mCi/m mole was purchased from
Amersham/Searle Corporation, Arlington Heights, IL. Thallium chloride
was purchased from ICN Pharmaceuticals, Incorporated, Plainview, NY.
After incubation the reaction was chilled and stopped by the addition of
20 il of 0.02 M1-mecaptoethanol, 1 M sodium phosphate, pH 6.8. The
solution was heated to 70C for 15 minutes and 50 vg of yeast tRNA was
added as carrier. Labeled RNA was separated from free 1251 by passage
through a Sephadex G75 column (IcmxlOcm) (Tereba and McCarthy, 1973).
5' End Labeling Reactions
The 5' ends of endonuclease digestion fragments were labeled with
[y 32PdATP in the presence of T4 polynucleotide kinase (Richardson,
1966). The 5' terminal phosphate of the polynucleotide was first removed
by treatment bacterial alkaline phosphatase in 10 mM Tris, pH 7.4 and
100 mM NaCl at 60C for 30 minutes. To halt the reaction 5 mM EDTA was
added and the solution was phenol extracted twice then ether extracted
3 times. The DNA was then precipitated with ethanol and dried under
vacuum. The DNA was redissolved in a solution containing 50 mM Tris,
pH 9.5, 10 mM MgCl2, 5 mM dithiolthreatol, and 50% glycerol. To this
was added 100 mM spermidine and 20 units T4 polynucleotide kinase. The
resulting solution was mixed with 200 pCi of dried [y -32P]dATP and
allowed to react at 37C for 30 minutes. To stop the reaction 5 mM EDTA
was added and the solution was dialyzed against 10 mM1 Tris, pH 7.4,
5 mM EDTA, and 1 M NaCI to remove unreacted [y -32P]dATP. The bacterial
alkaline phosphatase was purchased from Worthington Biochemical Corpora-
tion, Freehold, NJ; the Kinase was obtained from Bethesda Research Labora-
tories, Inc., Rockville, MD; and the [y -32P]dATP was obtained from
Amersham, Arlington Heights, IL. Generally 50-100 pg of DNA were
labeled in a 50 pl reaction volume to a specific activity of 105 cpm/pg
The large "A" fragment of E.coli polymerase I was used to 3' end-
label various digestion fragments. For this purpose the 50 pl cleavage
reaction mixture was blown down to approximately 20 pl and 20 pCi of
[a -32P]dCTP or [a -32P]dATP with 0.9 units of the Klenow polymerase
fragment (New England BioLabs, Beverly, MA) were added. Labeling was
done at 37C for 30 minutes. To half the reaction the sample was placed
at 60C for 10 minutes, then 20 pg of sonicated carrier DNA was added
and the sample was ethanol precipitated.
DNA restriction fragments were transferred from agarose or polyacry-
lamide gels to nitrocellulose filters in the manner described by Southern
(1975). Gels were first stained in a 1 pg/ml ethidium bromide bath,
visualized under UV light and photographed. The right corner of the gel
and the furtherest points of migration were notched to facilitate future
orientation of the gel. The DNA fragments were then denatured in situ
by soaking the gel for 20 minutes in a 1 M KOH solution. The KOH wash
and gel were then tritrated to neutrality with IM HCI-IM Tris. The gel
was then aligned on a prewashed nitrocellulose filter for transfer. The
nitrocellulose filter was prewet by washing for 20 minutes in a 5XSSC
bath. The DNA was eluted from the gel with 1OXSSC. For transfer from
a 1% or greater agarose gel, the gel was placed on a paper wick soaked
in 1OXSSC, overlaid with the nitrocellulose and absorbant material.
This arrangement permitted a higher volume of 10XSSC to pass through the
gel and enhanced transfer of higher molecular weight fragments. For
transfer from agarose gels of less than 1% and polyacrylamide gels, a
slightly different procedure was used in order to decrease diffusion of
the DNA bands during transfer. The filter was again sandwiched between
the gel and absorbant material but arranged in the reverse order so that
transfer was in a downward direction. The gel was overlaid with a paper
wick which was periodically wet with 1OXSSC. Thus a decreased volume of
1OXSSC was passed through the gel and diffusion was diminished. After
transfer the nitrocellulose filter was dried at 80C for 2 hours under
Hybridizations were carried out in sealed plastic bags containing
the filter and a hybridization mixture composed of the following: 50%
formamide, 5XSSC, 0.08% polyvinylpyriodate, 0.08% bovine serum albumin,
0.08% ficoll, 0.5% SDS, 0.02 M Tris (pH 7.4), 50 pg/ml sonicated calf
thymus DNA as a carrier and either 32P or 125I-labeled probe at a con-
centration of at least 106 cpm/ml. All blots were prewashed with the
hybridization mixture minus the probe at 37C for 2 hours prior to the
hybridization. Hybridization was carried out in the presence of the
complete mixture for 72 hours at 37C. The blots were then removed and
washed twice to remove nonspecifically bound label, with the hybridiza-
tion mixture minus carrier and probe for 30 minutes at 37C. The blots
were then rinsed twice with 2XSSC and 0.5% SDS in 50% formamide. The
blots were washed twice in 1XSSC and 0.5% SDS in 50% formamide for 30
minutes at 37C (McConaughy et al., 1969; Tereba and McCarthy, 1973).
Finally, the blots were washed overnight at 65C in 1XSSC and 0.5% SDS,
with 4 changes of the wash.
On occasion the hybridization was carried out at 65C. In these
instances the reaction and washes were done as described above at 37C
but in the absence of formamide. Attempts were made to increase the
sensitivity of the hybridization by adding 10% dextran sulfate to the
hybridization reaction (Wetmur, 1975). The findings were inconsistent
due to variable nonspecific background on the autoradiographs of such
Following the hybridization and washings the blots were dried and
then autoradiographed with Kodak X-Omatic regular intensifying screens
and Dupont Cronex 4 X-ray film at -70C for 3-21 days.
As stated previously, the general aim of this research is to illu-
minate the molecular status of the HSV genome in latently infected human
trigeminal ganglia. Specifically the studies are designed to
1. Detect HSV DNA and RNA sequences in human ganglia.
2. Determine the extent of the viral genome present through
hybridization of HSV DNA to nitrocellulose paper bound ganglia
3. Determine the extent of viral DNA expressed as cytoplasmic
RNA through mapping of ganglia RNA hybridized to nitrocellu-
lose paper bound viral DNA.
4. Correlate genome presence with the expression of a humoral
immune response to HSV antigens in immuno-competent patients
as measured by complement fixation.
5. Evaluate the manner of viral genome latency.
To achieve this end the following experimental approach was taken.
Total cellular DNA and cytoplasmic RNA were extracted from trigeminal
ganglia. Three different experimental routes could then be carried out.
The DNA was either radiolabeled with [a 32P]dCTP by nick translation
(Rigby et al., 1977) or it was digested with a restriction endonuclease,
gel electrophoresed and then transferred and immobilized on nitrocellu-
lose paper by the blotting method of Southern (Southern, 1975; lWahl et
al., 1979) (see Figure 7). Radiolabeled ganglia DNA was subsequently
hybridized to EcoRl digests of HSV DNA immobilized on nitrocellulose
paper under conditions favorable for formation of DNA-DNA hybrids. The
reverse experiment was also done using nick translated [a 32P]dCTP
labeled HSV DNA as a probe and hybridizing with endonuclease digested
Figure 7. DNA/DNA Hybridization Procedure
The third procedure entailed hybridizing 125I-labeled total cytoplasmic
RNA from ganglia to immobilized restriction endonuclease digests of HSV
DNA (Tereba and McCarthy, 1973) (see Figure 8). In all these approaches
viral DNA or RNA sequences complementary to immobilized sequences were
detected by autoradiography at -70C for 1-21 days. The pattern of
radioactive bands was compared to control HSV DNA run in parallel
RNA HSV DNA
125 chemically label
125I RNA I
hybridize ___ Endo R Digest
HSV DNA Blots" '
Figure 8. RNA/DNA Hybridization Procedure
The binding of radiolabeled ganglion DNA or RNA to nitrocellulose
bound viral DNA permits the detection of specific genome regions of HSV
DNA sequences in neural tissue. This is a considerable refinement over
standard filter hybridization methods or in situ methods where only
hybridization to the whole genome is monitored and where a consideration
must be given to high backgrounds due to nonspecific binding of the
The selectivity of the assay is still largely a product of the
purity of the known viral DNA preparation. To assay this purity HSV-l
F strain DNA was nick translated and hybridized to EcoRl endonuclease
digests of DNA from calf thymus tissue, salmon sperm, human foreskin,
Hela M monolayers and Hep 2 cells (see Figure 9). The HSV DNA was pre-
pared by high salt precipitation in the manner described by Pater et al
(1976). This approach is a modification of the high salt extraction
method first devised by Hirt (1967). Nonspecific binding of label
occurred with each digestion assayed. This suggests the presence of
cellular DNA copurifying within the viral DNA preparation. For this
" -. r
Figure 9. Cellular DNA Controls
reason, only HSV-l F strain DNA extracted from sucrose gradient purified
nucleocapsids and banded twice in CsCI density gradients to remove con-
taminating cellular DNA and RNA was used throughout (Spear and Roizman,
1972). To further eliminate the possibility of contamination by copuri-
fying cellular nucleic acids, HSV-1 F strain DNA sequences cloned in
PBR322 were also used in some of the hybridization studies. DNA isolated
from purified nucleocapsids failed to hybridize to cellular DNA in a
repetition of the above described assays. In control experiments such
a probe also failed to bind to DNA from adenovirus (Ad2), adenoassociated
(AAV) virus and Simian Virus 40 (SV40) (see Figure 10). DNAs extracted
from thoracic ganglia and connective tissue of patients have been run in
parallel reactions with trigeminal gnaglia DNA and failed to hybridize
with this probe.
To determine the sensitivity of the hybridization assay reconstruc-
tion experiments were run using HSV and AAV DNA. Samples of AAV DNA
were digested with either BamHl, EcoRl or Alul and then mixed. Starting
with a total of 10 ug of digested AAV DNA serial dilutions of the
preparation were then made. To maintain a total of 10 ug DNA in each
sample sonicated calf thymus DNA was added as needed. The samples were
electrophoresed on a 6% polyacrylamide gel and then transferred to
nitrocellulose by the Southern blotting method. Using a nick translated
AAV probe with a specific activity of 108cpm/ug DNA the sample containing
100 pg of AAV DNA could be visualized by autoradiography after 2 weeks.
A 200bp fragment within this sample would represent 5pg of DNA. Because
fragments within this size range are visualized the assay is sensitive
enough to distinguish 5pg of AAV in 10 ug of calf thymus DNA, a dilution
factor of 2x106. Similar experiments using HSV DNA established that
Figure 10. Viral DNA Controls
endonuclease fragments transferred from 0.5% agarose gels to nitrocellu-
lose could be distinguished at a dilution factor of 106 of Ipg of HSV
DNA in 1 ug of carrier DNA.
Hybridization of Ganglia Nucleic Acids to HSV DNA
The trigeminal ganglia of 18 patients have been assayed for HSV
sequences by DNA/DNA or DNA/RNA hybridization. None of the patients
manifested clinical evidence of recurrence at the time of death nor was
a history of recurrence noted (Table 1).
Table 1. Hybridization of trigeminal nucleic acids to HSV DNA.
Patient CT titer
(N.D.) not determined
Briefly, the findings from the experimental approaches have been
listed in Table 1. The first approach entails the hybridization of a
nick translated 32P labeled viral DNA to either EcoRl or HindIII endo-
nuclease digests of ganglia DNA bound to nitrocellulose filters. The
second is the hybridization of 1251-labeled ganglia cytoplasmic RNA to
restriction endonuclease digests of HSV-1 DNA bound to nitrocellulose.
Eleven of these patients were sero-positive for HSV. A positive serology
here is assessed as a compliment fixation titer of 1:8 or greater. The
trigeminal gnaglia from 9 such individuals were shown to contain HSV
DNA sequences. Viral DNA was not detected in the remaining two members
of this group (949 and 1033) but viral sequences were demonstrated in a
preparation of total cytoplasmic RNA from the ganglia of patient #949.
Only one of the sero-negative individuals (941) yielded ganglia con-
taining latent HSV DNA and viral RNA was also detected in this case.
Figure 11 is an example of the autoradiographs obtained from
hybridization assays involving viral and ganglia DNA. The first lane
(A) is a photograph of the ethidium bromide stain of an EcoRl digest of
HSV-1 run as a marker. The second lane (B) is the autoradiograph of
labeled viral DNA hybridized to EcoRl digests of ganglia DNA. Lane C
is the photograph of the ethidium bromide stain of the HindIll digested
control as it appeared after agarose gel electrophoresis. Lanes D and
E are the autoradiographs of labeled viral DNA bound to HindIII endo-
nuclease digests of ganglia DNA. In the experimental reaction with
ganglia #228 DNA, extensive binding of the probe is evident in the
regions where the larger EcoRl fragments (MW 8.4-17.5x106 dal.) of HSV
would migrate as determined by comparison to the controls. Binding is
also apparent where the EcoRl M (MW 2.7x106 dal.) fragments are expected
to migrate. The viral probe did not hybridize in the region where the
EcoRl K(MIJ 3.5x106 dal.) fragment should be found. EcoRl K maps at the
terminus of the short unique portion of the genome. The fragment may
display an altered migration here, or this portion of the viral genome
may not be represented in the ganglia. It is of interest to note that
AB C D
E' Pl i* ;'. H ,i '' i ir 1 .:-
Figure 11. Viral DNA hybridized to Ganglia DNA
ganglia of patient #227 also revealed an altered migration of the EcoRl
K fragment. Unfortunately the HindiI1 digestion pattern does not permit
such an analysis. Patient #228 is representative of the type of findings
obtained when a major portion of the HSV genome was demonstrated in the
ganglia. In all such cases to date the endonuclease digestion pattern
revealed was consistent with that of HSV-1. In several of the ganglia
only part of the HSV genome could be detected; whether the remainder of
the genome is absent or those portions are present in too few copies for
the assay to reveal is as yet uncertain.
Figure 12 is an autoradiograph of 1251-labeled ganglia cytoplasmic
RNA hybridized to an EcoRl endonuclease digest of HSV-1 F strain DNA
bound to nitrocellulose paper. To the left of the autoradiograph is a
photograph of the cleaved HSV stained with ethidium bromide after agarose
gel electrophoresis and prior to transfer to nitrocellulose. Clearly
not all regions of the viral genome are represented in the ganglia cyto-
plasmic RNA. This would not be the case if the ganglia were actively
supporting a lytic infection as shown in the HSV control. Further,
those fragments present are not uniformly represented in the RNA. For
instance, transcripts of EcoRl K appear most abundant in the RNA of
patient #951. In this manner the viral RNA transcripts found in several
ganglia have been mapped on the HSV-1 genome. Figure 13 summarizes this
To determine whether HSV DNA could be isolated in an episomal form
free from host chromatin total ganglia nucleic acid was fractionated on
CsCl density gradients. As a control human thoracic ganglia DNA was
kinased and run in a parallel gradient. 3H labeled lambda DNA was run
as a marker in all gradients and 50 ug of sonicated calf thymus DNA was
added to each as a carrier. The gradient profiles are shown in Figure
HSV 951 HSV 941
Ganglia RNA Hybridized to Viral DNA
I i i IIH
J 0 G N F MO
LA I IE KI
LA I E IK
9 5 1 1. --
Hpal I i I l I II I I II
LN K ORSUJ M IV
941 ! !- i F---
, I I
HI E II II I I
H F E TOXPL2
Figure 13. Mapping HSV specific RNA found in the Ganglia
3H 32 P
10 it I
9 \ 9
o_ i \'*
X6 I 0\
0.5 I 5x
I I \ ,0.
I 2 3 4 5 6 7 8 9 0 I 12 13 i4 5 6 17
Figure 14. Fractionation of Total Ganglia Nucleic Acid
A B C
2 3 .4 5 7 14 115 Is I7
F PtACT ION
Figure 15. Sectioning of CsCl gradient
Using the lambda marker as a reference the gradients were divided
into five segments. Within gradients A, B and C fractions were pooled
and the total DNA per section isolated and nick translated. The findings
are summarized in Table 2. Approximately 25 ug of DNA from HEP2 mono-
layers infected with HSV-1 F at an m.o.l. of 1 and harvested at 20 hours
post infection was used as a positive control for this assay.
Table 2. CsCl enrichment study
Patient CF-titer I II III
1168 256 -
1090 256 -
1144 128 -
1143 8 -
thoracic ganglia 16 -
HepII ND -
HSV infected Hep2 ND + + -
Mapping the L/S junction of HSV-1 F strain
The immediate early mRNA species of HSV-1 are known to map to the
reiterated portions of the genome. This fact coupled with the observa-
tion that transcripts mapping to this region were present in all posi-
tive ganglia studied to date suggested that this area warranted further
analysis. Hybridizations of ganglia RNA to pBR322 clones of the L/S
junction were undertaken in order to develop a clearer picture of the
degree of transcription. To gain such an understanding it was necessary
to first derive detailed restriction digestion maps of the DNA in
Two plasmids were used for this purpose, pRB104 and pRB115 (Post
et al., 1981; Material & Methods) (see Figure 16). The SP2 insert in
pRB104 is approximately 500bp greater in length than the SP1 species in-
serted into pRB115 (Post, 1981). First the DNA was evaluated by single
and double or triple endonuclease digestions using BamHl, Pstl, HindII
and BstEll. This confirmed the identity of the two plasmids and also
allowed us to map a 520bp insertion to the largest BamHl/HindII fragment
of pRB115 (see Figure 17). The size markers in these assays were known
digests of either pBR322 or AAV DNA. The clones were then screened with
various additional endonucleases to determine their digestion patterns.
A summary of the restriction enzyme fragment sizes is given in Table 3
o 10 20 30 40 50 60 70 80 90 100
ab 4 | 4 a, O12 C
2 51 5
- "-_ iE-mRNA
1 ~ ~-- -- - -\ ^ j --- -
* I 1 'ii II I i 111 |'I !1 I t U B am .
i I l!il I il l ;: I1 ill l i l i .1 i i ; i l ii I s
S P S IP \P
Derivation of Clones pRB104 and pRB115
1 2 3 4 5 6 7 8
Figure 17. Restriction Digests of pRB115 and pRB104
BamHl/Pstl/BstEll digest of pRB115
BamHl/Pstl/BstEll digest of pBR322
BamHl/Pstl digest of pRB115
BamHl/Pstl digest of pBR322
Pstl digest of pRB115
Pstl digest of pRB322
BamHl/HincII digest of pRB115
BamHl/HincII digest of pRB104
Table 3. Endonucleases
BamHl/HindII BamHl/BstEll BamHl/BstEll/TaqI Smal Aval
Clone: 115 104 115 104 115 104 115 115
size (bp) 4600 5200 750 4200 2400 2350 1075 890
1230 1240 2300 2300 1950 2200 677 566
1600 1975 577 483
Enzymes that did not cut:
BglII, Kpnl, Hpal,
Pstl and Sail
Xbal, EcoRl, HindIII,
and a comparison of the Smal and Aval digests of pRB104 and pRB115 are
shown in Figure 18. The mean length of the insert in pRB104 was calcu-
lated to be approximately 6382 base pairs and that of pRB115 was 5830 base
pairs. A comparison of the Smal and Aval digestion patterns of the two
clones indicates that more than a single insertion event has occurred
in the derivation of the SPI, and SP2 fragments. The fragment corres-
ponding to the 1075 bp Smal fragment of pRB115 is only 880bp in pRB104
even though pRB104 contains the larger insert. The 360bp Smal fragment
of pRB115 is overrepresented in pRB104. The 450bp Aval fragment of
pRB115 is overrepresented in pRB104 and the 483bp Aval fragment of
pRB115 is missing in pRB104. The pRB104 Aval digest also has a cluster
of fragments between 330bp and 380bp that are not found in pRB115.
Because of these additional rearrangements the Sinai and Aval maps of
only pRB115 were determined.
1 2 3
Figure 18. Smal and Aval digests of pRB104 and pRB115
Lanes: 1. Smal digest of pRB115
2. Smal digest of pRB104
3. Aval digest of pRB115
4. Aval digest of pRB104
The SP1 insert in pRB115 was excised by BamHl cleavage and the
insert separated in bulk from the vector by affinity chromatography over
a malachite green substituted bisacrylamide gel column. The Smal and
Aval cleavage maps were determined by partial digestion mapping (Smith
and Birnstiel, 1976). The approach was as follows. The DNA was cut with
Alul and then 5' end labeled by the T4 polynucleotide kinase reaction.
The DNA was then recut with Taql endonuclease and the digestion fragments
separated by agarose gel electrophoresis. The three largest fragments
spanning the insert were designated 1, 2 and 3, respectively. These
fragments were aligned on the general map using double digestion assays
with Hincil or BstEll as shown in figure 19a. Partial cleavage mapping
was then done using Smal and Aval on each of the Alul/Taql fragments.
The map was confirmed by first cutting with Taql, labeling and then
recutting with Alul; thus generating an overlapping map going in the
Figure 19b shows the generation of the Smal map for Alul/Taql frag-
ment 2 and the Hincil digestion of fragment 2 and fragment 3. The
accompanying figure 19b shows the mapping strategy and the log plot of
migration distance versus molecular weight from which the size of the
partial Smal fragments was determined. The resulting cleavage map of
fragment 2 is also shown. Figure 20 shows the partial digestion frag-
ments resulting from Aval cleavage of the Alu /Taql fragments 2 and 3.
Sall, EcoRl, BamHl and Alul digests of AAV were used as size markers.
The maps resulting from these gels and similar experiments are shown in
T__o_______ 3I A I
ITo-} Atu I
& i AiI I
Sal I A
SSl I B
-i \ Al, I A
Dislance Migroted (cm)
Figure 19a. Strategy for Partial Digestion Mapping of pRBII5
2 Sma I
HinclI 0 5 10 20 40 80
EcoRI A -
Alul A -
Figure 19b. Smal partial digestion of pRBll5
AAV 0 5 10 20 40 80 5 10 20
Figure 20. Aval Partial Digestion Mapping of pRB115
- 0 ">
._ > E
Hybridization of ganglia RNA to the L/S junction
As stated previously hybridization of total cytoplasmic RNA from
ganglia to Southern blots of clone pRB115 of the L/S junction of HSV-1
F was undertaken in hopes of better delineating the level of transcrip-
tion through this portion of the genome. The 125I-labeled RNA from
five sero-positive patients was evaluated in this manner. The nucleic
acids from thoracic ganglia and a normal lytic infection were also
evaluated as controls. The findings are given in Table 4. As in the
Table 4. Analysis of L/S Junction
Patient CT titer Viral DNA* Viral RNA*
1033 64 -
1034 32 +
1082 128 +
969 16 -
1089 128 +
Hep2 N.D. + +
(N.D.) not done
CsCl enrichment studies DNA or RNA from Hep2 monolayers infected with
HSV-1 F at an m.o.i. of 1 and harvested at 20 hours post infection were
used in the control assays.
Cross Hybridization Experiments
The close biological relationship between AAV, Ad and HSV suggests
the possibility of partial genome homology; however previous attempts
have failed to demonstrate detectable homology. Using the more sensi-
tive technique of cross hybridization of 32P-labeled AAV, HSV or Ad
probes to restriction endonuclease digests of the reciprocal genome
immobilized on nitrocellulose filters, we were able to determine limited
regions of homology. Figure 22 outlines a typical cross hybridization
assay and our findings to date from the hybridization assays are sum-
marized in Table 5.
Table 5. Cross Hybridization Analysis
AAV-2 Ad2 Ad5 SV40 HSV-l Hep2 HelaM Hela
AAV-2 + + + -
Ad2 + + + -
Ad5 + + + -
SV40 + -
HSV-1 + -
(-) no homology
To determine the level of nonspecific hybridization that could occur in
our assays, such as that which might arise from long stretches of high
G+C content, hybridizations were done with the DNA from the Hep2, HelaM
and Hela suspension cell lines. The Ad5, Ad2, and AAV2 virus stocks
were grown in Hela suspension cultures. The HSV-1 virus was grown in
either Hep2 or HelaM monolayers. Thus these controls would also allow
us to detect any cellular DNA which might be copurifying with viral DNA
during its preparation. Our findings were consistently negative although
the controls were hybridized under several degrees of stringency as were
0 -1-' 4-
4-1 *r- (n
4r3 in I E
Q- .r I--
the experimental group. Figure 23 shows the findings of 2 sets of
experiments in which an Ad2 nick translated probe was hybridized to AAV
blots and washed under varying salt concentrations. The hybridizations
were done overnight (as described in the Materials and Methods portion
of this text) at 68C and all washes were done at 68C. On the left of
Figure 23 are the autoradiographs from the hybridizations to a HpaII
blot to AAV and on the right are those from a HaeII blot of AAV. As
one progresses from a 2XSSC wash with 300 mM NaCI to a O.5XSSC wash with
only 75 mM NaCl various bands are seen to drop from the autoradiographs.
In the case of the HaeII digest, clearly all bands are present in the
autoradiograph of the blot washed in 2XSSC but the blot washed in 1XSSC
has lost the HaeII D fragment and the intensity of the HaeII B band is
greatly diminished. The O.5XSSC level of stringency reveals only strong
hybridization to the HaeII A and C fragments with significant binding
remaining to the HaeII E fragment.
When assays are done in the manner described above whole viral DNA
is nick translated as a probe. This approach permits the mapping of
homologous regions on the nitrocellulose bound DNA but fails to correlate
that portion found in the blot with the portion of probe hybridizing in
reciprocal crosses. To obtain this information fragments which failed
to show homology when bound to nitrocellulose were extracted from gels
after endonuclease digestion and gel electrophoresis. These fragments
were nick translated and then used as independent probes in cross hybridi-
zation assays. Figure 24 shows the results of such an assay using the
EcoRl A and C fragments of Ad2 as probes against a HaeII digestion blot
of AAV. On the left is the autoradiograph of total AAV hybridized to
an EcoRl digest of Ad2 under stringent conditions (hybridization over-
night at 68C in 5XSSC and final overnight wash at 68C in O.5XSSC).
SSC \ cs-
Tec 4 s4
, ,.' -
X IX .5X
A' r^ ;.
Figure 23. Autoradiograph of Ad2 Probe Hybridized to AAV
Eco Rl ,:
AAV Ad RI
AAV '" "
Autoradiograph of AAV/Ad2 Cross Hybridizations
The AAV probe binds only to the EcoRl D and F fragments. A self
hybridization with an Ad2 probe was done to confirm that all the diges-
tion fragments were transferred efficiently to the blot. A specificity
control is also shown using a nick translated SV40 probe. The EcoRl
A and C fragments were extracted from another portion of the gel and
nick translated. The findings are shown on the left. A total Ad2 probe
bound to the HaeII A and C fragments of AAV. The EcoRl A and C probes
of Ad2 showed no homology with AAV under these conditions of hybridiza-
tion. The HaeII A and C fragments of AAV contain the terminal inverted
repeats, which have been shown to have a high G+C content. For this
reason HSV-1 DNA, which also has a high G+C content, was used as a
control for nonspecific hybridization and none was found.
Because of its defective nature AAV must be grown in the presence
of Ad. Unfortunately this situation permits the potential cross contam-
ination of AAV viral stocks with Ad DNA. To limit this occurrence both
AAV and Ad DNA stocks were extracted from CsCl gradient purified virions.
The possibility still remains that a small portion of the contaminating
DNA may have been packaged in the reciprocal viral capsid, thus copuri-
fying throughout the preparative procedure. It may be argued that such
an event could in part give rise to the finding demonstrated in the
previous experiments. In dealing with this argument two approaches have
been taken. First, Ad5 stocks which were maintained continually in the
absence of AAV were cut with BamHl giving rise to two fragments of
approximately 14.3x10 dalton and 9.7x10 dalton in size. After electro-
phoresis the digest was transferred to nitrocellulose and hybridized
with a nick translated probe of total AAV which had been grown in cocul-
ture with Ad2. Hybridization with the AAV probe involved only the high
molecular weight Ad5 BamHl fragments and showed no smaller contaminating
AAV fragments. The second approach was to hybridize Ad DNA from a viral
stock shown to be free of AAV by EM studies, to AAV DNA generated from
a clone of the entire genome inserted into pBR322. The Ad DNA for this
assay was donated by Peter McGuire, Department of Biochemistry, Univer-
sity of Florida, and the 620 pBR322 clone of AAV was a gift from Jude
Samulski, Department of Immunology and Medical Microbiology, University
of Florida. In both cases, the results confirm the previous findings
in the AAV and Ad2 cross hybridizations. The hybridization studies on
Ad5 were extended to map the regions of the genome to which binding was
most stringent. For this purpose EcoRl and Hindill digests were used
and again the salt conditions for the hybridizations or washes varied.
The findings are compatible with those previously found for Ad2. A
summary of the cross hybridization mapping studies is shown in Figure
To evaluate in better detail the cross hybridization studies a
computer comparison was done of the entire nucleotide sequence of AAV-2
against the published sequences of the first 11,600 bases of the left
half of Ad2 and the final 8,561 nucleotides of the right half of Ad2.
Three programs were run: the first conducted a random search for
sequences with a minimum perfect match of six bases, the second searched
for a 3 out of 5 base match and the third looked for a 7 out of 10 match.
In general the data showed extensive random homology throughout both
genomes. This is supportive of the less stringent hybridization studies.
If one analyzed the data for sets of adjacent homologies that mapped
co-linearly between the two genomes these sequences were invariably
found within the terminal 20% of the AAV genome, again corroborating the
hybridization data. The following are comparisons of some of the
Ecc R I
A C B
73 .C 97.3
G E C H A B F
Eco R I Acenovirus 2
___________1 t=^ -
AB F D E C
E C F A B 0 G
0 10 20 30 40 50 60 70 80 90 100
Adenc asscci ted Virus
C E F B D A
T'- V P N F B FR K US A L E
S C 2C 3C
40 5C C: 70 80 90 J'0
Figure 25. Mapping of Homologous Regions in AAV and Ad
U 1IL "
^ ~111 -o r.! 0? .
I1 . i,, 1
C uJ -I E )T
^ g .- t_ -o-"T
0l 0 E
sequences found in this manner. "These-regions of homology .have been
numbered and are summarized in Figure 26.
Upon comparison the portion of the Ad2 genome mapping 7930 to 7999
base pairs in from its right terminus shows a 71% sequence homology with
the 60 base pair sequence beginning 40 nucleotidesin on the inverted
terminal repeat of AAV (see Figure 27). This maps within the EcoRl D
fragment of Ad2 78% of the B and C' domains of AAV are conserved as are
the functions between B and B' domains and the C' to A' junction.
Eighty-eight percent of the C reiteration of AAV is retained. Approxi-
mately 20 nucleotides upstream from the homologous stretch shown the
sequence 5' TGCGC 3' is conserved in both genomes; similarly 25 nucleo-
tides downstream in both genomes the sequence 5' CTCCA 3' is retained.
The sequence 5' CTCACCGG 3' shown in figure 27 in Ad2 complements the
3' end of the first leader sequence in the E3 transcription complex
(Herisse et al., 1980).
B B' C C' A'
IIIII I lllll I IIII III1IIII II 11 1 1 HII 1I II I
Ad2 TCGGGCAATCAAAG ACTCCCGGAGCCCGGGCAAAGCACTGGCGGCGGCAGTGGTCGAG 7
I i 7999
T / A G TGT
Figure 27. Ad2 Early Region 3A (Herisse et al., 1980)
Figure 28 details the sequence homology surrounding the beginning
of the second coding region in AAV and a homologous stretch in Ad2
starting 5252 nucleotides in from the right terminus. This 53 base
pair sequence maps to the EcoRl E fragment of Ad2 and is part of
sequence involved in the E3 transcription complex. There is about 69%
total sequence homology between the two genomes over this area and it
should be noted that the ATG start signal found in AAV is absent in the
Ad2 sequence. Figure 29 compares a 26 nucleotide sequence found just
outside the inverted terminal repeat of AAV starting at base pair 144
to a similar region in the terminal repeat of Ad2. 'There is an 87%
homology between the two regions.
AAV TCAGCCAGGTACATGGAGCTGGTCGGGTGGGTCGTGGACAAGGGGATTACCTCGGAG 1306
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
Ad2 GCA CCAG TTTTTGGCGCTGGGTGGGTAGCTTGTAGCTGAGGCGGTTGCC GGAG 5305
Figure 28. AAV Coding Region 3 (Heriss6 et al., 1980)
AAV CTGGAGGGGTGGAGTCGTGACGTGA 168
I I I I 1 1 1 1 I I 1 1
Figure 29. Ad2 Terminal Repetition (Herisse et al., 1980)
Although cross hybridization assays failed to reveal any homology
between HSV and AAV three lines of evidence suggest at least a general
similarity in the architecture of their genomes. The termini of both
genomes contain a tandem array of natural and inverted repeats. Portions
of these repeat regions may be flipped or inverted in their relation to
unique regions of the genomes and this flip-flopping appears to be a
natural consequence of viral DNA replication. And lastly there is an
inherent ability within the replication cycle for maintaining identity
between the termini through some mechanism of gene conversion. The
reiterated sequences in both viruses probably contain the origin of
replication and at least some early regulatory function. A comparison
was made between the terminal reiterated sequences of AAV and the pub-
lished sequence for the L/S junction of HSV-1, which contains the HSV
0D-- O u
cD -c u
0j < -^
0 -0 0 -
terminal repeats also. Figure 30 presents the "flip" and "flop" orien-
tations of the AAV genome to which the HSV sequences were compared.
The direct repeat which brackets the ends of the a reiteration in
the termini of HSV-l has a similar sequence to the B', C, and C' domains
of the AAV hairpin. The short palindrome GCCCGGGC forms the core of
this repeat in HSV-1 DNA. This palindrome is found in the core of the
inverted repeat of AAV also. One finds that 80% of the base sequence
is conserved between the HSV-1 direct repeat and the B', C and C'
domains of AAV DNA. The nucleotide sequences bordering the palindrome
conserve 66% of the B' and C' repeated domains (see Figure 31).
B C C'
AAV GGTCGCCCGACGCCCGGGCTTTGCCCGGGCG 86
11 111 I I I IIIIII 11 1 1 11
HSV GGCCGCGGGGGGCCCGGGCT GCGCCGCCG 213
Figure 31. Homology between short palindromes in HSV-l and AAV
(Mocarski et al., 1980)
By expanding the comparison of this region in HSV with AAV one finds
approximately 72% of the flip orientation of the AAV terminal palindrome
is conserved in the HSV sequences (see Figure 32). Eighty-two percent
of the nucleotides reading from roughly the middle of the A domain of
AAV through the C and into the C' domains are the same. Approximately
110 base pairs upstream in the HSV-l map from the sequence described
above one finds a stretch of 43 nucleotides which can be aligned with
the flop orientation of the AAV termini. Having done this there appears
to be roughly 84% homology between the two genomes in this region. As
one continues to read in a 5' to 3' direction the HSV sequence is highly
compatible with the terminal domains of AAV arranged in the order A, C,
C', B, B', C, C' then A. There is an 84% homology between genomes in
C D C) C(
C) < CC
G (3 CD
F-- D- :
co F- F-
(_> CD (S
CDS -U C3
D- D (S
- CD C)
C) (3S- I-
(S 0: I-
U3--F-^I (S c
(S F- \(S
*a I- \<
C-) (S3 \(
(S C3) t
U3 0 (
- C <) (
- (3 C)
*=3 C3) C
this first half of this array. The second half of the array is con-
served in 68 to 70% of the sequences.
The ends of the HSV sequences displayed in Figure 32 are separated
by 20 nucleotides. This boundary is formed in part by a twelve base
pair inverted repeat.
As one reads downstream From the end of the proceeding sequence
through the A reiteration of HSV, one comes to the direct repeat
defining its far terminus. Again this region can be aligned with the
terminal AAV sequences. About 63% of the sequence of the A domain is
conserved and the short palindrome GCCCGGGC is retained in full within
both genomes (see Figure 33). The sequence ACTCC is found 51 base pairs
downstream from this short palindrome in both AAV and HSV, followed by
the sequence GGTTCCT. In HSV these two short sequences are contiguous
while in AAV they are separated by 10 base pairs.
I 1 I1 1 1 11111 i 111 1 1 1 1 1II 11111111
HSV CCCGCCTTTTTTGCGCGC CGC CCCGCCCGCGGGGGGCCCGGGCTGCC 685
----- DR1 -----685
Figure 33. Homology between the terminus of the short unique region in
HSV-l and the terminus of AAV (Mocarski et al., 1980)
The trigeminal ganglia of 22 patients have been assayed for HSV
sequences by DNA/DNA hybridization: None of the patients manifested
clinical evidence of recurrence at the time of death nor was a history
of recurrence noted. Fourteen of the patients were sero-positive for
HSV. A positive serology was assessed as a complement fixation titer of
1:8 or greater. The RNA from the trigeminal ganglia of 16 patients was
also surveyed for HSV transcripts by RNA/DNA hybridization.
Hybridizations were conducted in 50% formamide and 5XSSC solution
at 37C. Given a Tm of 97C in 1XSSC and a G + C value of 68% the calcu-
lated value of the Tm for HSV in 5XSSC would be approximately 104C
(Sheldrick and Berthelot, 1974). Allowing for a 0.7C decrease in the Tm
value for each 1% of formamide present and a 1C lowering for each 1.5%
mismatching of base pairs (McConaughy et al., 1969), the initial condi-
tions of hybridization would permit a 21% mismatch in base pairing be-
tween the probe and nitrocellulose bound DNA. Subsequent washes in 1XSSC
and 50% formamide at 37C would permit no more than a 16% mismatch to be
retained. In like manner it can be calculated that washes in O.1XSSC
solution alone at 68C would permit no greater than 10% mismatch in base
pairing. Consequently, the signal retained in hybridizations between the
HSV-1 probe and ganglia containing HSV-2 DNA or transcripts would be
greatly subdued. In such instances the strongest binding would occue in
the regions of greatest homology between the two types, namely, the
termini and L/S junction.
Under any of the above conditions control experiments demonstrated
that the HSV-l nick translated probe failed to bind to DNA from unin-
fected human ganglia or connective tissue as well as to the DNAs of AAV,
SV40 and Ad2. All three of these viruses have a demonstrated potential
for latency in human cell lines in vitro. In reconstruction experiments
using blots of 6% acrylamide gels, we were able to detect 5 pg of viral
DNA in the presence of 10 pg of carrier DNA, a dilution factor of
2xlO6. Similarly 10 pg of viral DNA could be detected in the presence
of 10 jg of carrier when hybridization was to Southern blots of 0.5%
agarose gels. Any real differences in the sensitivities of these two
approaches is largely indicative of the increased diffusion witnessed
when transfer was from low percentage agarose gels. If the DNA content
of the normal diploid human cell is 5 pg (Luria and Darnell, 1978), the
10 pg of ganglia DNA normally assayed would represent the product of about
2x106 cells. Thus, given the molecular weight of the HSV genome is 108
daltons, the sensitivity of our hybridization assay is such that we can
detect roughly 1 copy of the HSV genome per 200 cells.
In 18 of the 22 assays in which trigeminal ganglia were evaluated
for the presence of HSV, total ganglionic DNA was cleaved by endonuclease
digestion, transferred to nitrocellulose and probed with nick translated
HSV DNA. Eleven of these patients were sero-positive for HSV. Nine of
the sero-positive were shown to carry HSV DNA and two of these were also
shown to be positive for RNA transcripts. The nucleic acid from two of
the sero-positive patients failed to demonstrate HSV specific sequences
in either DNA or RNA. Only one of the seven sero-negative patients gave
evidence of the presence of HSV DNA or RNA; the remainder were negative.
Thus, there is a positive correlation between the presence of compliment
fixing antibodies and the presence of HSV specific sequences as measured
by this approach.
Two patients warrant some further discussion. Patient #941 was a
1 year old male who died from congenital lobar emphysema. Although both
viral DNA and RNA were detected repeatedly in the trigeminal ganglia no
humoral immune response to HSV could be determined by either complement
fixation or indirect immunofluoresence assays. The patient had not
received sufficient immunosuppressive therapy to negate a humoral re-
sponse nor was there evidence of congenital failure of the humoral immune
system. It is possible the primary viral infection of the infant occurred
before the humoral immune system was competent enough to mount a response
and no subsequent recurrences occurred. No viral DNA was detected in the
ganglia from patient #949 although RNA transcripts from the short repeat
regions of the HSV genome were demonstrated. It may be that only this
portion of the genome was latent in the ganglia; these findings would
then mirror the varied sensitivities of the two experimental approaches.
The detection of only part of the HSV genome in several of the
ganglia raises the question of whether the remainder of the genome is
absent. Brown et al (1979) has reported the detection of latent defec-
tive HSV by complementation of the endogenous genome iwth ts mutants of
HSV-1 in 8 out of 14 individuals previously negative for spontaneous
release of virus. The altered migration or absence of the terminal EcoRl
K fragments revealed in patients #228 and #227 could be examples of such
an event. The frequency with which large deletions could occur would be
greatest in the terminal regions of the linear genome. Conversely, it
may be argued that these fragments have undergone an apparent increase
in molecular weight due to covalent linkage with host DNA or some other
form of rearrangement. The approach taken here would not resolve this
issue. Frequently the regions which we failed to detect gave rise to
smaller endonuclease digestion fragments whose copy number would fall
below our level of detection. The sensitivity of our assay would allow
us to detect a fragment representing 1% of the genome only if it was
present in at least 1 copy per cell. The findings in patients #1082 and
#1089 probably fall within this category.
In a recent repetition of this approach using human brain tissue
Fraser et al (1981) reported the detection of HSV DNA sequences in six of
eleven individuals assayed. Hybridization against total viral DNA
revealed all or portions of the HSV genome were present in the six posi-
tive individuals. Hybridization against the clone pRB115 showed a
marked alteration of the electrophoretic migration pattern of the ter-
minal fragment in oen individual and no change in the two other patients.
The findings of the CsCl enrichment studies lead us to consider the
possibility that the bands we detected by blot hybridizations are the
product of a lytic infection that ensued at the time of death. The
positive controls using DNA from a known HSV infection insures that the
mechanics of this approach were efficient enough to retain sufficient
levels of viral DNA if this were the case. Given the sensitivity of the
hybridization assay itself and the negative findings in known sero-posi-
tive patients a reasonable argument can be made against an ongoing lytic
infection within the ganglia.
RNA from the trigeminal ganglia of 16 patients was screened for HSV
specific transcripts by DNA/RMA hybridization. In 8 of these patients
1251-labeled ganglia cytoplasmic RNA was hybridized to either EcoRl or
Hpal endonuclease digests of HSV-1 F strain DNA bound to nitrocellulose
paper. We were able to demonstrate HSV specific RNA transcripts in two
of three sero-positive patients assayed and one of the five sero-negative
patients. The medical history of the single sero-negative patient, ':941,
has been discussed along with our findings from DNA/DNA hybridizations.
From the mapping of those RNA transcripts observed it is clear that the
regions of the genome most frequently encountered to date are those con-
taining sequences in the immediate early and early mRNA species of HSV-l
(Clements et al., 1979; Roizman, 1979).
Because of the frequency with which transcripts mapping to the region
of the L/S junction were encountered, the cytoplasmic RNA from 9 indivi-
duals was used as a probe in hybridizations against endonuclease digests
of the clone pRB11I5. Eight of these patients were sero-positive. None
of the patients contained RNA which would hybridize with pRB115. This
result was curious in view of the fact that RNA was present which would
hybridize to the larger EcoRl K fragment present in EcoRl digests of the
total HSV genome. It can be argued that the two probes allowed the
recognition of a subset of transcripts associated with latency, namely,
a set of transcripts mapping to a region of the EcoRl K fragment not
common to the BamHl P fragment. The region running downstream from the
right end of BamHl P to the right terminus of EcoRl K is approxiamtely
1.8 kb in size (see Figure 13).
Using the tsB2 mutant Holland et al (1979) has reported finding no
mRNA species of greater than 3 kb in this region. This temperature sen-
sitive mutant is one of a group of ts mutants mapping within the sequences
of the short reiterated region which code for ICP4 (Schaffer et al., 1973).
The viral gene function coding for transition from immediate early to
early polypeptide synthesis also maps to this region and is thought to be
a function of ICP4 (Preston et al., 1978). Studies of protein synthesis
during temperature shift experiments on this group of mutants suggest
that ICP4 is autoregulated (Schaffer et al., 1973). The 5' terminus of
IEmRNA-3 which is believed to code for ICP4 lies within the 1.8 kb region
described above (Watson and Van de Woude, 1982; Clements et al,, 1979;
Anderson et al., 1980). Transcription is normally in a leftward direction
proceeding into the region of the BamHl P fragment at the joint (iaxam
and Gilbert, 1980).
Finally, two ts mutants of ICP4, tsD and tsK, have been assayed for
their ability to establish latency in mice; tsK fails to establish
latency in mice while tsD will (Stevens and Cook, 1971). Both mutants
block viral DNA synthesis. Thus two ts mutants within a single gene
within this region can distinguish between viral functions permitting
subsequent DNA replication and the establishment of latency. In light
of the above findings further evaluation of transcription through this
region during latency is warranted, especially in animal models.
Another portion of the genome highlighted by the transcription
studies is the region mapping from about 0.08 to 0.65, especially the
area from 0.1 to 0.3 (see Figure 13). The bulk of the early viral pro-
teins genes are within these domains (Schaffer et al., 1973). The genes
for the viral thymidine kinase and the viral DNA polymerase of both HSV-1
and HSV-2 map within the region 0.27-0.43 (Clements et al., 1977; Schaffer
et al., 1973). Spear and Roizman (1972) have reported a correlation be-
tween viral induced oncogenic transformation and ICP8 within the region
from 0.30 0.45 map units. Stevens and Cook (1971) have demonstrated
a latency negative phenotype for the mutants tsA, tsS, tsT, and tsl; all
of which map within the region 0.1 to 0.6. It is noteworthy that in
subsequent mappings of RNA transcripts by in situ hybridization,
Galloway and McDougall (1981) have consistently found the region 0.1 to
0.3 represented in ganglia; other regions including the L/S junction were
variably demonstrated. In similar studies of HSV RNA in cervical car-
cinoma cells sequences spanning map positions 0.07 to 0.4, 0.58 to 0.63,
0.82 to 0.85 and 0.94 to 0.96 were represented (McDougall et al., 1982).
These findings are noteworthy because of the collinearity of the HSV-l
and HSV-2 transcription maps. Taken in total, the studies are supportive
of our findings and suggest that a subset of the immediate early and
early genes are involved in the establishment and maintenance of latency.
The binding of 32P-labeled HSV DNA to restriction endonuclease
digests of ganglia DNA, or conversely 1251-labeled ganglia RNA to HSV
DNA has permitted several questions to be considered. Those regions of
the viral genome which are represented have been determined by simply
noting the presence or absence of given viral DNA bands on the autoradio-
graph. Although it is reasonable to expect the entire genome must be
present as a single unit to give rise to recurrence, some cells within
the population may carry only portions of the viral genome. It is there-
fore not surprising to find an apparent unequal molar representation of
viral fragments. This is demonstrated by our findings and suggested by
previous work on viral thymidine kinase in ganglia (Yamamoto et al., 1977).
Finally, the question of whether latent viral DNA is covalently
intefrated into cellular chromosomes was approached. Any digestion frag-
ments which are covalently linked to cellular DNA will migrate in the
agarose gels with an apparent increase in molecular weight as a result
of the added cellular component. Digests of viral DNA present as a free
circular episome would also have an altered migration for the terminal
fragments. DNA sequestered in immature nucleocapsids would give rise to
banding patterns similar to that of wild type viral DNA. In interpreting
such data it should be noted that most virus isolates from ganglia
explants, to date, have yielded unique endonuclease restriction patterns.
Thus for alteration in electrophoretic migration to be reasonably com-
ponent, the shift should be significant and repeatable when the ganglia
DNA is restricted with an enzyme such as Xbal which does not reveal
strain differences in HSV. When the latent virus is HSV-2 the fidelity
of hybridization with the HSV-1 probe could also be significantly reduced.
Thus although our data is compatible with integration of HSV into cellu-
lar genomes, it is not sufficient to prove it.
Hybridization of 1251-labeled cytoplasmic RNA to digests of HSV DNA
demonstrates the presence of viral RNA transcripts in the ganglia. The
design of our experiments using several different restriction digests of
HSV enabled us to corroborate our findings and more finely map these
sequence homologies. Autoradiographic density of hybridized bands per-
mits estimates of the frequency of given RNA species in our sample.
For this reason control reactions were run with the 1251-labeled cyto-
plasmic RNA from both HSV lyticly infected Hep2 cells and uninfected
tissue cultures for comparison. This experimental approach answers the
question of which regions of the latent HSV genome are undergoing active
transcription. The fact that only a subset of the RNA transcripts were
found in latently infected ganglia suggests that some HSV genes may be
necessary for maintenance of the latent state.
A number of general observations have prompted the inclusion of a
comparative analysis of the HSV, Ad, and AAV genomes within this study
of HSV latency. The genomes studies are representative of the three
linear DNA virus families whose members are known to replicate within
the nuceli of human cells. Architecturally each of these linear genomes
contains inverted repeats within their terminal regions. Each of the
viruses have a demonstrated potential in vitro for latency in human cells.
And lastly, the defective parvovirus (AAV) is absolutely dependent upon
coinfection with either an edenovirus or herpesvirus for its multiplica-
In the absence of a coinfecting helper, AAV enters the cell where
its virion is uncoated within the nucleus. No detectable macromolecular
synthesis occurs at this time. The naked AAV DNA may then establish a
stable relationship with the host cell (Hoggan et al., 1966; Cheung et al.,
1980). Subsequent infection with a helper virus promotes the multiplica-
tion of infectious AAV within these latent cells. Although the nature of
this stable relationship between AAV and the host cell is complex, it is
clear that at least a portion of the genome is covalently linked to
cellular DNA and in part copies of the entire genome are tandeomly
arrayed within this linkage with cellular DNA (Handa and Shimogo, 1977;
Cheung et al., 1980). Our findings demonstrate a substantial sequence
homology within apparent regulatory regions between the defective genome
and those of its helpers. This is of particular interest since no bio-
chemical relatedness has been previously demonstrated between these
viruses and the principal role the helper plays in this partnership is
the promoting of AAV macromolecular synthesis.
Although 6 to 12 base pair stretches of homology with AAV were
found randomly throughout the Ad genome these homologies were concen-
trated largely within the terminal sequences of the AAV genome. The
strongest co-linear homologies between regions of the two genomes which
at present appear to have no functional relatedness, involve the EcoRl
D and F fragments of Ad and the terminal repitition of AAV. Two examples
of sequence homology within these regions have been described. The
first of these analyzes the area surrounding the 3' end of the first
leader sequence in the E3 transcription complex of Ad and the termini
of AAV. The second compares the beginning of the second coding region
in AAV and a portion of the E3 transcription complex. The second exam-
ple actually lies within the potential coding sequence of the 14K E3
polypeptide of Ad (Herisse et al., 1980). Because the function of E3
region of Ad is obscure, it is difficult to evaluate these homologies.
There is no evidence to date which directly ascribes a specific regula-
tory role during transcription to the terminal repetition in AAV nor has
a role been assigned to the E3 transcription products of Ad. Transcrip-
tion of the E3 complex is nonessential for growth of Ad or AAV in vitro
(Richardson and 'Jestphal, 1981).
Two other short stretches of homology hying within the left half
of the Ad genome were also found. The first of these is the 26 bp
stretch in AAV (142-168) which is complimentary to a portion of the
inverted terminal repeat of Ad. The second is the AAV sequence sur-
rounding the TATA box 31 bases upstream from the leftward most site of
initiation of transcription. It has a 17/19 base homology with a com-
parable sequence upstream from the initiation of the Ad5 early region
1A transcripts. Lusby et al (1980) had previously reported these are
example regions with a potential role in transcriptional regulation.
Taken in total the data suggest a conservation through evolution of
regions of these genomes involved in transcriptional regulation.
As noted previously several lines of evidence suggest at least a
general similarity in the architecture of the AAV and HSV genomes. The
termini of both genomes contain a tandem array of natural and inverted
repeats. Also portions of these repeated regions may be inverted in
their relationship to unique regions of the genome as a natural conse-
quence of viral DNA replication (Lusby et al., 1980; Mocarski et al.,
1980). And there is apparently an inherent ability within the replica-
tive cycle for maintaining identity between the termini through some
mechanism of gene conversion. Probable origins of replication and at
least some early regulatory functions lie within the reiterated sequences
in both viruses.
The nucleotide sequence of the natural terminal repeat in HSV-1 F
was used for the comparison with the terminal 200 nucleotides of AAV
described in this study. The terminal 200 nucleotides of AAV contain
the 143 bp inverted terminal repeat. This region can further be broken
into two smaller internal palindromic sequences which are bracketed by
a larger repeat and its inverse (Lusby et al., 1980). The 5001 bp
natural repeat in HSV-l F is inverted to form the junction of the long
and short unique regions of the genome; thus each unique arm of the
genome is in turn roughly similar in architecture to the AAV genome, as
is the whole. The 501 bp sequence is composed of 20 direct repeats of a
12 nucleotide sequence, 3 repeats of a 37 base sequence and some inter-
nally unique sequences; the entire domain is bracketed by a 20 base pair
direct repeat (Mocarski et al., 1980).
Our findings demonstrate an extensive homology between portions of
the HSV inverted repeats found at the L/S junction and termini with the
termini of AAV. The regions of HSV homology are 1) those regions of
the "b" reiteration which bracket the long unique region and are imme-
diately adjacent to the "a" repeat, 2) those portions of the "c" reitera-
tion which bracket the short unique region and are contiguous with the
"a" repeat, and 3) the 20 base direct repeat (DRI) which brackets the
"a" repeat. The high G + C content and repetiveness of these regions
would explain our earlier failure to find these homologies by Southern
The extensive homology between DRI and the small internal palin-
dromes of the AAV sequence is of particular interest. The short palin-
drome GCCCGGGC has been totally conserved within both. A role for DRI
has been proposed in both the site specific cleavage of concatomeric
replicative intermediates of HSV and the inversion of the long and short
domains (Mocarski et al., 1980). Similar, deletion mutants within this
sequence in AAV are unable to resolve high molecular weight intermediates
in DNA replication. Fifty-one base pairs downstream from GCCCGGGC in
AAV one finds the sequence ACTCC followed nine base pairs further down-
stream by the sequence GGTTCCT; ACTCC is the probable cleavage site for
the inversion of the AAV terminal palindrome and resolution of the hair-
pin during DNA replication. The sequence ACTCC followed by GGTTCCT is
also found within the "c" reiteration of HSV-1 F fifty-one bases down-
stream from the sequence GCCCGGGC within DRI. The question of whether
this arrangement is coincidental or the sequence actually serves a
similar function in the replication of both genomes will require genetic
analysis. It should be noted that HSV-l F strain DNA contains a single
copy of the "a" reiteration at its right terminus and the putative
cleavage site for resolution of concatomers is thought to lie within or
just outside the "a" reiteration adjacent to the "c" repeat (Mocarski
et al., 1980).
The terminal sequences of AAV DNA appear to be intimately involved
in the integration of latent AAV DNA into cellular chromatin. Endo-
nuclease analysis of AAV DNA from early (10) and late (118) passages of
latently infected cells demonstrated that the only sequences altered
during clonal passage were those within the palindromic region of the
terminal repeat, specifically, the sequence GCCCGGGCT which contains the
Smal recognition site (Cheung et al., 1980). Thus these sequences could
potentially serve a similar role in latency of HSV.
The conservation of similar GC content and architecture between
the termini of several parvoviruses supports the genetic evidence
assigning identical biological roles to these regions. These observa-
tions appear to signify the direct involvement of the termini in viral
multiplication. Further the potential to form nearly identical curci-
form structures through secondary folding of their 3' termini appears
to be stringently conserved even within those genomes which have signif-
icant sequence divergence from that of AAV. Thus the evolution of
this family of small DNA viruses has conserved both primary sequence and
secondary structure within a region of the genome in which both must
interact in the preservation of fidelity in viral replication.
The question of whether sequence homology between AAV and HSV is a
consequence of similar function warrants further investigation in light
of the findings within the parvovirus family. Our observations to date
do not permit the assumption that homologous sequences serve identical
functions within AAV and HSV. A cursory analysis of the homologous
sequences in HSV reveals that they are not self complimentary and cannot
easily be ascribed a secondary structure similar to AAV. Determination
of whether the primary sequence within HSV would lend itself to the
formation of a stable curciform secondary structure is central to this
line of investigation. In conclusion the sequence homology between AAV
and HSV demonstrates the potential for similar mechanics of early regu-
lation and macromolecular synthesis, events which appear to be concomit-
tant with the establishment and maintenance of latency.
LIST OF REFERENCES
Anderson, K., J. Stringer, L. Holland and E. Wagner. 19791 Isolation
and localization of herpes simplex virus type I mRNA. Journal of
Virology. 30: 805-820.
Anderson, K., L. Holland, B. Gaylord and E. Wagner. 1980. Isolation and
transcription of mRNA encoded by a specific region of the Herpes
Simplex virus type I genome. Journal of Virology. 33: 749-759.
Atkinson, M.A., S. Barr and M.C. Timbury. 1978. The fine structure of
cells infected with temperature sensitive mutants of herpes simplex
virus type 2. Journal of General Virology. 2: 517-536.
Bachenheimer, S. and B. Roizmano 1972. Ribonucleic acid synthesis in
cells infected with herpes simplex virus. Journal of Virology.
Baringer, J.R. 1974. Recovery of herpes simplex virus from human
sacral ganglions. New England Journal of Medicine. 291: 828-830.
Baringer, J.R. 1975. Herpes simplex virus infection of nervous tissue
in animals and man. Progress in Medical Virology. 20: 1-26.
Baringer, J.R. and P. Sworeland. 1973. Recovery of herpes simplex
virus from human trigeminal ganglions. New England Journal of
Medicine. 288: 546-650.
Bastian, R.O., A.S. Rabson, C.L. Yee and T.S. Tralka. 1972. Herpes
hominis isolation from human trigeminal ganglion. Science. 178:
Becker, Y., H. Dym and I. Sarov. 1968. Herpes Simplex virus DNA.
Virology. 36: 184-192.
Berns, K.I. and W. Hauswirth. 1978. Parvovirus DNA structure and
replication. Replication of Mammalian Parvoviruses. Cold Spring
Harbor Laboratory Press. pp. 13-32.
Biswal, N., B. Murray and M. Benyesh-Melnick. 1974. Ribonucleotides in
newly synthesized DNA of herpes simplex virus. Virology. 61:
Brown, S.M., J.H. Subak-Sharpe, K.G. Warren, Z. Wroblewska, and H.
Koprowski. 1979. Detection of uninducible herpes simplex virus
genomes latnet in human ganglia explants. Proceedings of the
National Academy of Science, U.S.A. 76: 2364-2368.
Buchman, T.G., B. Roizman, G. Adams and B.H. Storer. 1978. Restriction
endonuclease fingerprinting of herpes simplex virus DNA: a novel
epidemiological tool applied to a nosocomial outbreak. Journal
of Infectious Diseases. 138: 488-498.
BUnemann, H. and W. Muller.
stranded DNA: affinity
Nucleic Acids Research.
1978. Base specific fractionation of double
chromatography on a novel type of absorbant.
Carton, C.A. and E.D. Kilborne. 1952. Activation of latent herpes sim-
plex by trigeminal sensory-root section. New England Journal of
Medicine. 246: 172-176.
Centifanto-Fitzgerald, Y., E. Varnell and H. Kaufman. 1982. Initial
herpes simplex virus type 1 infection prevents ganglionic super-
infection by other strains. Infection and Immunity. 35: 1125-
Centifanto-Fitzgerald, Y., T. Yamaguchi, H. Kaufman, M11. Tognon and B.
Roizman. 1982. Ocular disease pattern induced by herpes simplex
virus is genetically determined by a specific region of viral DNA.
Journal of Experimental Medicine. 155: 475-489.
Cheung, A., D. Hoggan, U. Hauswvirth and K.I. Berns. 1980.
of the adeno-associated virus genome into cellular DNA
infected human Detroit 6 cells. Journal of Virology.
Clements, J., R. Watson, and N. Wilkie. 1977.
herpes simplex virus type I transcription:
on the viral genome. Cell. 12: 275-285.
Temporal regulation of
location of transcripts
Clements, J.D., T. McLaughlan and D.S. McGeoch. 1979. Orientation of
herpes simplex virus type 1 immediate early mRNAs. Nuclear Acid
Research. 7: 77-93.
Comerford, S. 1971. lodination of Nucleic Acids in vitro.
Cook, M.L., V.B. Bastone and J.G. Stevens. 1974. Evidence that neurons
harbor latent herpes simplex virus. Infection and Immunology. 9:
Cook, M.L. and J.G. Stevens. 1973. Pathogenesis of herpetic neuritis
and ganglionitis in mice: evidence for intra-axonal transport of
infection. Infection and Immunity. 7: 272-288.
Cook, M.L. and J.G. Stevens. 1976. Latent Herpetic infections following
experimental viremia. Journal of General Virology. 31: 75-80.
Daniels, C.A., S.G. LeGoff, and A.L. Notkins.
infectious virus-antibody complexes from
patients with recurrent herpes labialis.
1975. Shedding of
vesticular lesions of
Lancet. 2: 524-528.
Delius, Ho and J. Clements. 1976. A parital denaturation map of herpes
simplex virus type 1 DNA: evidence for inversion of the unique
DNA regions. Journal of General Virology. 33: 125-133.
Ecob-Johnston, M.S. and W.O. Whetsell. 1979. Host-cell response to
herpes virus infection in central and peripheral nervous tissue
in vitro. Journal of General Virology. 44: 747-757.
Feldman, L.A., R.D. Sheppard and M.B. Bronstein. 1968.
virus-host cell relationships in organized cultures
nerve tissues. Journal of Virology. 2: 621-628.
Finkelstein, M. and R. Rownd. 1978. A rapid method for extracting DNA
from agarose gels. Plasmid. 1: 557-562.
Fraser, N., W. Lawrence, Z. Wroblewska, D. Gilden, and H. Koprowski.
1981. Herpes Simplex type 1 DNA in human brain tissue. Proceedings
National Academy of Science, U.S.A. 78: 6461-6465.
Frenkel, N. and B. Roizman. 1972. Separation of herpes virus DNA duplex
into unique fragments and intact strand on sedimentation in alkaline
gradients. Journal of Virology. 10: 565-572.
Frenkel, N., H.
from the S
Locker, W. Batterson, G.S. Hayward and B. Roizman. 1976.
herpes simplex virus DNA VI. Defective DNA originates
component. Journal of Virology. 20: 527-531.
Galloway, D.A., C. Genoglio, M. Shevchuk and J.K. McDougall. 1979.
Detection of herpes simplex RNlA in human sensory ganglia. Virology.
Galloway, D. and J. McDougall. 1981. Transformation of rodent cells
by a cloned DNA fragment of Herpes Simplex type 2. Journal of
Virology. 38: 749-760.
Goodpasture, E.W. 1929. Herpetic infection with especial reference to
involvement of the nervous system. Medicine. 8: 223-243.
Grafstrom, R.H., J.C. Alwine, W.L. Steinhart and C.S. Hill. 1974.
Terminal repetitions in herpes simplex virus type I DNA. Cold
Springs Harbor Symposium Quart. Biol. 39: 679-681.
Graham, B., H. Ludwig, D. Bronson, M.
1972. Physiochemical properties
Biochemical and Biophysical Acta.
Benyesh-Melnick and N. Biswal.
of the DNA of herpes viruses.
Handa, H. and H. Shimojo. 1977. Isolation of the viral DNA replication
complex from adeno-associated virus type 1 infected cells. Journal
of Virology. 24: 444-450.
Hayward, G.S., N. Frenkel and B. Roizman. 1975. Anatomy of herpes
simplex virus DNA: strain differences and heterogeneity in the
location of restriction endonuclease cleavage sites. Proceedings
of the National Academy of Science. 72: 1768-1772.
Hayward, G.S., R.J. Jacob, S.C. Wadsworth, and B. Roizman. 1975.
Anatomy of herpes simplex virus: Evidence for four populations of
molecules that differ in the relative orientations of their long
and short components. Biochemistry. 72: 4243-4247.
Herisse, J., G. Courtois and F. Galibert. 1980. Nucleotide sequence of
the EcoRl D fragment of Adenovirus z genome. Nucleic Acid Research.
Heriss6, J. and F. Galibert. 1981. Nucleotide sequence of the EcoRl E
fragment of adenovirus 2 genome. Nucleic Acid Research. 9: 1229-
Hill, T.S. and H.S. Field. 1973. The interaction of herpes simplex
virus with cultures of peripheral nervous tissue: an electron
microscope study. Journal of General Virology. 21: 123-133o
Hill, T.S., H.S. Field and W.A. Blyth. 1975. Acute recurrent infection
with herpes simplex virus in the mouse: a model for studying and
recurrent disease. Journal of General Virology. 28: 341-353.
Hirt, B. 1967. Selective extraction of polyoma DNA from infected
mouse cell cultures. Journal of Molecular Biology. 26: 365-369.
Hoggan, M., N. Blacklow and W. Rowe. 1966. Studies of small DNA viruses
found in various adenovirus preparations! Physical, biological
and immunological characteristics. Proceedings of the National
Academy of Science, U.S.A. 55: 1467-1575.
Holland, L.E., K.P. Anderson, J.R. Stringer and E.K. Wagner. 1979.
Isolation and localization of herpes simplex virus type 1 mRNA
abundant before viral DNA synthesis. Journal of Virology. 3_1:
Honess, M. and D. Watson. 1974. Herpes simplex virus-specific polypep-
tides studied by polyacrylamide gel electrophoresis of immune preci-
pitates. Journal of General Virology. 22: 171-183.
Jones, P. and B. Roizman. 1979. Regulation of Herpesvirus macromolecular
synthesis VIII. The transcription program consists of three phases
during which both extent of transcription and accumulation of RNA
in the cytoplasm are regulated. Journal of Virology. 31: 299-
Kaplan, A.S. (ed.) 1973. The Herpesviruses Academic Press, New York.
Kieff, E.D., S.L. Bachenheimer and B. Roizman. 1971. Size, composition
and structure of the deoxyribonucleic acid of herpes simplex virus
subtypes 1 and 2. Journal of Virology. 8: 125-132.
Knipe, D., W. Ruyechan, R. Honess and B. Roizman. 1979. Molecular
genetics of herpes simplex virus: The terminal a sequences of the
L and S components are obligatorily identical and constitute a part
of a structural gene mapping predominantly in the S component.
Proceedings of the National Academyof Science, U.S.A. 76: 4534-4538.
Kristensson, K., B. Ghetti and H.M. Wisniewski. 1974. Study on the
propagation of herpes simplex virus (type 2) into the brain after
intraocular injection. Brain Research. 69: 189-201.
LePecq, J. 1970. Use of ethidium bromide for separation and determina-
tion of nucleic acids of various conformational forms and measure-
ment of their associated enzymes. ;ethods of Biochemical Analysis.
Lis, J.T. 1980. Fractionation of DNA fragments by polyethylene glycol
induced precipitation. Methods in Enzymology. 65: 347-353.
Locker, H. and N. Frenkel. 1979. Baml, Kpnl and SalI restriction enzyme
maps of the DNAs of herpes simplex virus strains Justin and F:
occurrence of heterogeneities in defined regions of the viral DNA.
Journal of Virology. 32: 429-441.
Lonsdale, D.M., S.M. Brown and J.H. Subak-Sharpe. 1979. The polypep-
tide and the DNA restriction enzyme profiles of spontaneous iso-
lates of herpes simplex virus type 1 from explants of human
trigeminal, superior cervical and vagus ganglia. Journal of
General Virology. 43: 151-171.
Luria, S. and J. Darnell.
New York, New York.
1978. General Virology.
3rd ed. Wiley.
Lusby, E., K. Fife and K. Berns. 1980. Nucleotide sequence of the
inverted terminal repetition in adeno-associated virus DNA.
Journal of Virology. 34: 402-409.
Lusby,E. and K.
Berns. 1982. Mapping the 5' termini of two adeno-
virus RNAs in the left-half of the genome. Journal of
McConaughy, B.L., C.D. Laird and B.S. McCarthy.
reassociation in formamide. Biochemistry.
McDougall, J., C. Crum, C. Fenoglio,
1982. Herpesvirus-specific RNA
uterine cervix. Proceedings of
McLauchlan, J. and J.B. Clements. 1I
herpes simplex virus type 1 mRN/
1969. Nucleic acid
L. Goldstein, and D. Galloway.
and protein in carcinoma of the
National Academy of Science, U.S.A.
)82. A 3' co-terminus of two early
ks. Nucleic Acid Research. 10:
Maxam, A.M. and W. Gilbert. 1980. Sequencing end-labeled DNA with base-
specific chemical cleavages. Methods of Enzymology. 65: 499-460.
Mocarski, E.S., L.E. Post and B. Roizman. 1980. Molecular Engineering
of the herpes simplex virus genome: insertion of a second L-S
junction into the genome causes additional genome inversions. Cell.