 |
INTRODUCTION |
Kaposi's sarcoma-associated herpesvirus
(KSHV),1 also called human
herpesvirus 8, is a 
2-herpesvirus strongly associated
with Kaposi's sarcoma, primary effusion lymphoma, and a plasmablastic variety of multicentric Castleman's disease (1-3). In these
malignancies, the vast majority of tumor cells are latently, as opposed
to lytically, infected. During latent infection, the latency-associated
nuclear antigen (LANA), as well as a small subset of additional viral genes, is expressed, and the genome is replicated and segregated to
infected daughter cells (4-6). LANA is the highly immunogenic gene
product of ORF73 (7, 8). It has been shown to interact with numerous
transcription cofactors, to localize to chromosomes, to specifically
bind DNA within the terminal repeat (TR), to form dimers in solution,
and to maintain a plasmid containing a single copy of the TR as
episomes (9-19).
The ability to commandeer the cellular replication machinery to
replicate extrachromosomal viral genomes has been studied in several
viruses, particularly SV40, human papilloma virus, and Epstein-Barr
virus (EBV), and their respective origin-binding proteins: large T
antigen, E1/E2, and EBNA-1. The recent finding that LANA binds
to a region within the KSHV TR and is capable of maintaining a plasmid
containing a single copy of the KSHV TR indicates that it also belongs
to this functional class of viral origin-binding proteins (9, 17, 18).
Viral origin-binding proteins directly bind DNA in a sequence-specific
manner, but the roles these proteins have in the initiation of
replication is still unclear and may vary between the different
proteins. For example, large T antigen and E1 both have helicase
activity and may themselves serve to unwind DNA during replication,
whereas EBNA-1 seems to be dependent on cellular factors for this
function (20-27). However, it is clear that all these viral proteins
must facilitate the formation of an initiation-of-replication complex, composed primarily of cellular factors, at or near their binding sites.
Each of these viral proteins binds to target DNA sites with distinct
arrangements in the origins of replication (23, 28-31).
Of the known human herpesviruses, EBV has the greatest homology to
KSHV. EBV oriP, which requires the trans-acting
viral protein EBNA-1, has been studied extensively. It is composed of
two elements, the dyad symmetry (DS) and the family of repeats (FR),
separated by ~1 kb of unique sequence. The DS element is composed of
two pairs of EBNA-1-binding sites, with the sites in each pair
separated by 22 bp, center to center. It is the DS element that, in the presence of EBNA-1, is the functional replicator of oriP
(29, 32). The FR element is composed of 20 copies of a 30-bp repeat and
is responsible for long-term maintenance of oriP-containing DNA (33, 34). Recently, it was shown that sequences that lie between
the EBNA-1-binding sites are also important, and it has been
hypothesized that they may be targets of factors involved in the origin
recognition complex (35).
In addition to facilitating DNA replication, viral origin-binding
proteins such as EBNA-1 and E2 also have important transcriptional regulatory effects (36-38). Although many of these activities are crucial for regulating viral gene expression (39-42), it has also been
suggested that transcriptional activation is necessary for efficient
initiation of DNA replication (43). In addition to the origin-binding
proteins themselves acting as transcriptional activators, the sequences
surrounding the origin of replication often contain transcription
factor-binding sites, many of which have been shown to directly
contribute to DNA replication (43, 44). KSHV is similar in that the TR
DNA, which contains the LANA-binding site and presumably the origin,
contains a transcriptional enhancer element (9). Despite the fact that
LANA has been shown to trans-activate many promoters through
interaction with upstream factors, when tethered to DNA by a
Gal4-binding domain, LANA has repeatedly been shown to suppress
transcription (15, 19, 45, 46). In addition, we have previously shown
that LANA suppresses transcription from a reporter construct containing
TR sequences (9).
To further our understanding of how LANA functions as a viral
origin-binding protein, we analyzed the LANA-binding region of the TR
and the contributions of the binding sites in this region to the
ability of LANA to suppress transcription and to promote DNA
replication. Using electrophoretic mobility shift assay (EMSA) and
DNase I footprinting, we show that LANA binds to two sites, LANA-binding site 1 (LBS-1) and LANA-binding site 2 (LBS-2), which are
separated by 22 bp, center to center. Analysis of LBS-1 indicated that
LANA binds with an affinity similar to that of EBNA-1 to the DS element
(47). A detailed quantitative analysis of nucleotides at the termini of
the LANA-binding site identified a core binding site consisting of 16 bp, 13 of which are conserved between LBS-1 and LBS-2. Analysis of
LBS-1 and LBS-2 alone and in tandem indicated that, like the
EBNA-1-binding sites in the DS element, the second site is bound
cooperatively. Analyses using a series of deletion and transversion
mutations showed that LBS-1 and LBS-2 both contribute to the ability of
LANA to suppress transcription and to facilitate replication. Changes
to the binding sites resulted in proportional effects in both
phenotypes, indicating that LANA may carry out these two functions
through related mechanisms.
| |
MATERIALS AND METHODS |
Plasmids--
Plasmid pAG42, used to express the LANA C-terminal
domain, was constructed by cloning the coding sequence from
pcDNA3.1v5HISA/orf73C (9) with NcoI and PmeI.
This sequence was ligated into the pTM1 vector (48) at NcoI
and SmaI. pTR5 is described elsewhere (9).
The plasmids used to produce fragments for EMSA were made by annealing
the oligonucleotides listed in Table I,
digesting them with BamHI/EcoRI, and ligating
them into pBluescript II SK. Constructs were sequenced to ensure
accuracy. Fragments for EMSA were produced by
XhoI/XbaI digestion.
The series of TR mutants was constructed using pAG28, a plasmid
containing one copy of the TR cloned with
KpnI/XhoI from pCRII1TR (9) into the pGL3
promoter (Promega, Madison, WI), and the oligonucleotides listed in
Table I. These oligonucleotides were annealed, digested with
EcoRI and BamHI, and ligated into pBluescript II
SK. After these plasmids were grown and sequenced, the fragments were
excised using Bsu36I and AvaII. These fragments
and the XhoI/AvaII fragment from pAG28 were
ligated into pAG28 at Bsu36I/XhoI in a three-way
ligation. Multiple diagnostic restriction digests and partial
sequencing confirmed all final constructs.
Cell Lines--
CV-1 cell, African green monkey fibroblasts, and
293 human embryonic kidney cells were obtained from American Type
Culture Collection. Cell monolayers were maintained in Dulbecco's
modified Eagle's medium supplemented with 10% fetal bovine serum,
penicillin, and streptomycin at 37 °C under a 5% CO2 atmosphere.
Expression and Partial Purification of the C-terminal Domain of
LANA Using the MVA/T7 Vaccinia Virus Expression System--
CV-1 cells
were infected with MVA/T7 virus as previously described (48-50).
Briefly, a confluent 10-cm plate of CV-1 cells was split 1:2 12 h
prior to infection with MVA/T7 virus at an approximate multiplicity of
infection of 10. Cells were transfected 1 h post-infection with 1 µg of pAG42 plasmid using Effectene (QIAGEN Inc., Valencia,
CA) following the manufacturer's instructions. Cells were harvested
and purified using Ni2+-NTA beads (QIAGEN Inc.) following
the manufacturer's instructions. Briefly, cells were sonicated in
lysis buffer (50 mM Na2HPO4 (pH 8),
500 mM NaCl, 10 mM imidazole, 10 mM
-mercaptoethanol, 1% Triton X-100, and 0.5% Nonidet P-40). Debris
were spun down. The supernatant was incubated with
Ni2+-NTA-agarose beads for 2 h at 4 °C, washed
twice with 2 × 10 ml of wash buffer (50 mM
Na2HPO4 (pH 8), 500 mM NaCl, 20 mM imidazole, and 10 mM -mercaptoethanol),
and eluted with elution buffer (20 mM HEPES, 100 mM KCl, 250 mM imidazole, 10 mM
-mercaptoethanol, and 20% glycerol).
EMSA--
Fragments were created by digesting the respective
plasmids with XbaI/XhoI. Probes were created by
labeling fragments with Klenow (Promega) using
[
-32P]dCTP (3000 Ci/mmol; Amersham Biosciences)
following the manufacturer's instructions. To purify the probe from
non-incorporated nucleotides, we used Sephadex 50 spin columns (Roche
Molecular Biochemicals). r-LANA was incubated with probe at room
temperature for 20 min in a total volume of 20 µl of buffer
containing 10 mM HEPES (pH 7.9), 50 mM KCl, 1 mM EDTA, 10 mM MgCl2, 0.05 µg/µl poly(dI-dC), 0.5 µg/µl bovine serum albumin, 10 mM dithiothreitol, and 10% glycerol. The samples were then
separated by electrophoresis on a native 4% polyacrylamide gel (55 mA
at 4 °C). Gels were dried and either exposed to Kodak film or
analyzed using a Molecular Dynamics PhosphorImager.
DNase I Footprint of EMSA Complexes--
Probes were prepared by
cutting pTR5 with KpnI/XbaI to label the top
strand and with Acc65I/PstI to label the bottom
strand. Labeling was done as described above. Binding reactions were
prepared as described above for EMSA; 1 min prior to loading on the
native gel, 1 µl of DNase I (0.7 units/µl) was added to the
reaction. EMSA was performed as described above, and then the wet gel
was exposed to Kodak film for 2 h to visualize the bands. The
bands were excised and eluted using QIAEX II (QIAGEN Inc.) following the manufacturer's instructions. 8000 counts of each complex and unprotected probe were run on a 6% sequencing gel containing 8 M urea. A G/A sequencing lane was created as a marker for
each probe using formic acid and piperidine (51). The gel was dried and
analyzed using the PhosphorImager.
Transient Transfection Assays--
293 cells were plated at a
density of 4 × 105 cells/well in six-well plates
8-12 h prior to transfection. Plasmids were transfected using FuGENE
6 (Roche Molecular Biochemicals) according to the manufacturer's instructions. To monitor transfection efficiency, we
transfected pcDNA3/LacZ into parallel wells and stained fixed cells
for -galactosidase activity. Transfection efficiencies were
generally between 25 and 35%. Cells were harvested and scored in
luciferase assays following product instructions (Promega).
Short-term Replication Assays--
293 cells were plated at a
density of 3 × 106 cells/100-mm dish. Plasmids were
introduced into cells using the calcium phosphate transfection system
(Invitrogen) following the manufacturer's instructions. Transfection
efficiency was monitored by a parallel transfection of pcDNA3/LacZ.
To exclude non-transfected DNA from the analysis, cells were
trypsinized, washed three times 16 h post-transfection, and seeded
into two plates. Transfected cells were harvested 72 h
post-infection. Cells were lysed in 700 µl of lysis buffer (10 mM Tris, 10 mM EDTA, and 0.6% SDS).
Chromosomal DNA was precipitated at 4 °C overnight by adding 5 M NaCl to a final concentration of 0.85 M. Cell
lysates were centrifuged at 14,000 rpm for 30 min at 4 °C. The
supernatant containing extrachromosomal DNA was subjected to
phenol/chloroform extraction. The extrachromosomal DNA was precipitated
by ethanol precipitation and dissolved in 20 µl of H2O
containing RNase A. 10% of the DNA was digested with HindIII or KpnI to linearize plasmid DNA to
measure input DNA by Southern blot analysis. 90% of the DNA was
subjected to HindIII (or KpnI) and
DpnI digestion in a final volume of 100 µl. 180 units of
DpnI were used for digestion at 37 °C for 48 h.
After digestion, DNA was ethanol-precipitated and redissolved in 20 µl of Tris/EDTA buffer. The single-digested and
double-digested plasmids were electrophoretically separated on 0.8%
agarose gels, transferred to nylon membranes, and assayed by Southern
blot analysis. 20 ng of a 800-bp fragment of TR or 45 ng of the entire
plasmid were labeled using the random prime labeling system
(Amersham Biosciences) and purified with quick spin columns (Roche
Molecular Biochemicals). The blots were hybridized in Church's buffer
at 65 °C, washed, and exposed to a Molecular Dynamics phosphor
screen overnight.
| |
RESULTS |
LANA Binds Two Adjacent Sites in the TR--
We (9) and
others (17, 18) have previously shown that LANA binds to a
region within the TR. Using deletion mutants, it was observed that, in
EMSA analysis, LANA forms multiple complexes, all of which can be
supershifted with antibodies directed against epitope tags of
recombinant proteins. Possible explanations for this phenomenon
included differential post-translational modification, functional
protein degradation products, or higher order complex formation (9, 17,
18). Mapping of a minimal LANA DNA-binding domain in EMSA resulted in
much higher resolution and clearly showed two distinct complexes, both
of which contain LANA as confirmed by supershift analysis (9).
To delineate the composition of these complexes, we expressed a
recombinant LANA protein containing the C-terminal 233 amino acids of
LANA and a His6 tag (r-LANA) using the MVA/T7 vaccinia virus expression system as previously described (9). As shown in Fig.
1A, LANA was highly enriched
after Ni2+-NTA affinity chromatography. This protein was
then used in an EMSA examining the effects of increasing amounts of
r-LANA in binding a constant amount of radiolabeled probe, TR5, which
contains nucleotides 551-675 of the TR (nucleotide numbers are those
defined in Ref. 52). This titration showed that, at low concentrations of r-LANA, a single high mobility complex was formed, but as the protein concentration was increased, a complex of lower mobility began
to form and eventually became the predominate complex (Fig. 1B). This indicated that the protein preparation itself was
a homogeneous species, both in size and in terms of
post-translational modification; thus, neither degradation nor
modification was responsible for the formation of these two complexes.
View larger version (60K):
[in this window]
[in a new window]
|
Fig. 1.
LANA binds to two adjacent sites in the
TR. A, r-LANA expression using the MVA/T7
expression system and enrichment by Ni2+-NTA batch
chromatography shown on a Coomassie blue-stained gel. Lane 1 shows the input protein extract, whereas lane 2 shows the
eluted protein fraction. The arrowhead indicates r-LANA.
Ptn, protein. B, titration of r-LANA on
TR5 (nucleotides 551-675). The end-labeled TR5 probe (0.8 nM) was incubated for 20 min with varying amounts of r-LANA
(0, 0.67, 1.7, 3.4, 6.8, and 10 nM). At low concentrations
of protein, a single high mobility complex (arrowhead b) was
seen by EMSA. At higher protein concentrations, a larger complex
(arrowhead a) was detected, eventually becoming the
predominant complex. C and D, DNase I
footprinting of complexes a and b on one DNA strand and the
complementary strand, respectively. Binding of TR5 in the presence of
10 nM LANA was carried out in the same way as described for
EMSAs, except that 1 min before loading the reaction onto the gel, 0.7 units of DNase I were added. After running the EMSA, the complexes were
visualized by exposure of the wet gel to film. Complexes a and b and
the unbound probe (p) were excised, eluted, and run on a 6%
sequencing gel. A marker (M) that cleaves Gs and
As was run in the first lane as a reference to the location
within TR5. The sequence of the relevant region of TR5 is shown. The
pink bar indicates the protected region within both
complexes. The blue bar indicates the extended footprint
shown in complex a. Arrowheads indicate residues
hypersensitive to cleavage in the presence of r-LANA. Pink
arrowheads indicate residues affected in both complexes;
blue arrowheads indicate residues hypersensitive only in
complex a. E, summation of DNase I footprinting data showing
both strands together. Data are compiled from seven independent
experiments representing both strands. Nucleotides boxed in
pink are areas protected in both complexes. Nucleotides
boxed in blue are areas protected only in complex
a. Nucleotides highlighted in pink are sensitized
to cleavage in both complexes. Nucleotides highlighted in
blue are sensitized only in DNA present in complex a. The
17-bp direct repeat is shown in boldface; the three
transversions within the repeat are denoted with an asterisk
above the nucleotide in LBS-2.
|
|
The fact that the low mobility complex was formed only at higher
concentrations of r-LANA suggested that it was composed of higher order
r-LANA complexes. Based on the observation that LANA and the C-terminal
domain of LANA exist predominantly as dimers in solution (19), we can
assume that the higher mobility complex was composed of one r-LANA
dimer bound to the probe; the low mobility complex would presumably be
composed of another r-LANA dimer interacting with this first complex.
We investigated whether the binding of this second dimer of r-LANA
occurs through protein/protein interaction, a second independent
DNA/protein interaction, or a cooperative DNA/protein interaction. To
discern between these possible scenarios and to better define the
nucleotides of TR5 involved in these interaction(s), we used a modified
DNase I footprinting protocol that allowed us to examine the
DNA/protein interactions in each complex independently (Fig. 1,
C-E). First, we repeated the same kinetic EMSA analysis as
shown in Fig. 1B with the major difference that DNase I was
added to the reaction shortly before running the native gel. After
determining the positions of the high and low mobility complexes by
autoradiography on the wet gel, both complexes and free probe were
extracted from the gel matrix and run on a sequencing gel next to a G/A
sequence ladder originating from the same DNA (51). By examining the
nuclease digestion pattern of each complex compared with that of
unbound probe (arrowhead p), it is clear that the high
mobility complex (arrowhead b) has a protected region that
corresponds to the sequence identified previously as the LANA-binding
site (9, 17, 18), referred to here as LBS-1. The low mobility complex
(arrowhead a) showed an identical protection pattern at
LBS-1, but, in addition, a second protected region of approximately
equal size; we refer to this new site as LBS-2. Examination of the
sequences protected in these regions revealed that the center of each
region, with the exception of three transversions, is composed of the
same 17 bp (Fig. 1E). The fact that these direct repeats are
separated by 22 bp indicates that the binding sites are located on the
same face of the DNA strand, potentially allowing bound proteins to interact (53).
This experiment suggested that, at low LANA concentrations, the high
mobility complex is composed of one dimer bound to the probe; the fact
that no significant protection was seen at LBS-2 indicates that the
vast majority of this complex is composed of r-LANA bound at LBS-1.
However, it is possible that a small percentage of this complex, too
low to be detected by this method, is composed of r-LANA bound only at
LBS-2 and not at LBS-1. It is also clear that the low
mobility complex is composed of two dimers of r-LANA bound to the
probe, one at LBS-1 and one at LBS-2. Two scenarios could explain these
results. First, LBS-2 could bind LANA independent of LBS-1 with a much
lower affinity. Second, LANA could bind cooperatively, such that it
first binds the higher affinity LBS-1 site, which then facilitates
binding to the lower affinity LBS-2 site. The latter possibility is
more likely because we saw virtually no occupation of the second site
at 1.7 nM r-LANA, but nearly complete occupation of the
second site at 10 nM r-LANA (Fig. 1B). The 22-bp spacing between the two binding sites is also a common feature of many
DNA-binding proteins that bind cooperatively (53).
It is worth noting that, when sequencing the highly GC-rich TR, a
compression exists in LBS-2 nucleotides 601-604, such that it
sometimes appears that there are only three Cs instead of four. This
was the case for the first TR sequence reported
(GenBankTM/EBI accession number U75699) (54). However,
subsequent sequencing has shown all four Cs (accession numbers U86666
and AF148805) (52, 55). We also repeatedly confirmed this sequence.
EMSAs using probes containing LBS-1 and LBS-2 with a C deleted at
position 601 have normal affinity for LANA at LBS-1, but greatly
decreased affinity at LBS-2, compared with the wild-type sequence (data not shown).
LANA Binds LBS-1 with a Kd of 1.51 ± 0.16 nM--
When comparing the mechanism of action of LANA in
transcriptional regulation and facilitating replication with that of
other viral origin-binding proteins, it would be informative to know whether it binds with a similar affinity. However, until this point,
all binding assays with LANA have been qualitative in nature. To
determine the Kd of LANA for LBS-1, we constructed a
fragment (TR31) containing the entire LBS-1 sequence protected from
DNase I digestion, but none of LBS-2 (Fig. 1). Annealing of
oligonucleotides representing this highly GC-rich region resulted in
multiple minor annealing products in addition to the desired double-stranded fragment. It was believed that these minor forms may
cause an increased error in this quantitative analysis. To eliminate
this problem, the annealed oligonucleotide was ligated into the
pBluescript II SK polylinker, grown in bacteria, harvested, sequenced,
and finally cut from the plasmid using restriction enzymes. This
resulted in a homogeneous population of fragments containing the TR
sequence embedded in polylinker DNA. This labeled fragment (0.08 nM) was used as probe in each of 13 binding reactions containing equal concentrations of r-LANA and unlabeled fragment at
0-9.9 nM (Fig.
2A). Radiographic densitometry
was used to determine the percentage of probe in the bound and free
forms; this percentage was then used to calculate the total free
fragment and the total bound fragment in each reaction. The results
were plotted, and the best-fit hyperbolic equation showed that the
Kd was equal to 1.51 ± 0.16 nM.
Interestingly, this value is very close to the Kd
reported for the EBNA-1 sites in the DS element, 2.0 nM
(47).
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 2.
Titration of an oligonucleotide containing a
single binding site against a constant protein concentration shows that
LANA binds with Kd = 1. 51 ± 0.16 nM. A, 13 DNA binding reactions were
prepared, each containing equal r-LANA concentrations, 0.08 nM labeled probe containing LBS-1, and varying amounts of
unlabeled probe DNA (0-9.9 nM). After EMSA, the
concentrations of free and bound oligonucleotides were calculated using
radiographic densitometry to determine the fraction of probe in each
form and then multiplying that fraction by the total amount of fragment
present in the reaction. B, these points are shown plotted
on a graph: bound oligonucleotide versus free
oligonucleotide. The data were fit to the hyperbolic equation
Y = Bmax(X)/(Kd + X) and used to calculate the Kd of the
reaction. PTN, protein.
|
|
The Core LANA-binding Site Is 16 bp of the 17-bp Direct
Repeat--
As mentioned, previous experiments to define the binding
site of LANA have always been qualitative in nature; a sequence either formed a complex in EMSA or did not (9, 17, 18). This approach is
insufficient to define a true core binding site because it does not
consider that partial binding sites can often result in partial binding
affinity or that nucleotides flanking a binding site often contribute
to a minor degree to binding affinity. To define the core LANA-binding
site, we examined the boundaries and contributions of peripheral
nucleotides of an individual binding site quantitatively.
Because DNase I is a large enzyme and shows
sequence-dependent preferential cutting, the region of DNA
protected is often much larger than the sequence in physical contact
with the DNA-binding protein. Therefore, we began the analysis with the
full sequence protected from DNase I in the high mobility EMSA complex
(Fig. 1), TR31, and made additional probes lacking consecutive base pairs from each end. To eliminate error introduced by imperfect annealing of oligonucleotides or effects due to differences in length
and flanking sequences, we imbedded each of the TR sequences in
pBluescript SK II at the BamHI and EcoRI sites.
After the sequences of the resulting constructs were confirmed,
fragments for EMSA were created using restriction sites in the
polylinker. This process resulted in each of the probes being a
homologous population, from 89 to 97 bp in length and with virtually
identical flanking sequences.
For each probe, we used EMSA on six samples containing 0.8 nM probe and between 0 and 10 nM r-LANA. By
plotting the free protein concentration versus the
percentage of bound probe and fitting it to a line representing a
hyperbolic equation, we were able to determine the
Kd of each sequence. As expected using this method,
we obtained nearly the same value for the Kd of TR31
(1.36 ± 0.06 nM) as we did in the experiment shown in Fig. 2. After determining the Kd of each sequence,
we calculated the relative Kd in comparison with
that of TR31 (Fig. 3, A and
B). Examination of the relative Kd values
of consecutive probes in these experiments showed that, although
peripheral nucleotides contributed somewhat to binding, the core
binding site is GCCCCATGCCCGGGCG, containing 16 bp of the 17-bp direct
repeat identified in the LANA-binding region. From this experiment, we
can conclude that the nucleotides in this direct repeat are primarily
responsible for binding LANA in both LBS-1 and presumably LBS-2. In
addition, the observed difference in affinity between LBS-1 and LBS-2
suggests that residues 4, 7, and 12 are important for LANA binding to
DNA.
View larger version (37K):
[in this window]
[in a new window]
|
Fig. 3.
The core binding site consists of 16 bp.
Quantitative mapping experiments of the 5'-end of LBS-1 are
shown in A. Experiments mapping the 3'-end are shown in
B. All sequences were imbedded in a polylinker, making the
final probes 89-97 bp in length. The nucleotides that match TR
sequence are shown, and nucleotides within the 17-bp direct repeat are
in boldface. Probe at 0.8 nM was used in binding
reactions with varying amounts of r-LANA (0, 0.67, 1.7, 3.4, 6.8, and
10 nM). After EMSA, the ratio of free probe (p)
to probe·r-LANA complexes (c) in each reaction was
determined by radiographic densitometry. These values were then used to
calculate the affinity of r-LANA for the truncated sequences in
comparison with TR31, which contains the entire region of LBS-1
protected in footprinting experiments (see Fig. 1). Ptn.,
protein.
|
|
LANA Binds to DNA in a Cooperative Fashion--
As stated earlier,
LANA seems to bind LBS-2 only after LBS-1 is occupied, indicating that
it may bind the two cooperatively. This possibility is further
strengthened by the spacing of the two sites, 22 bp from center to
center, which would place the proteins on the same face of the DNA
helix. To address this question, we performed EMSA under quantitative
conditions to compare the affinity of two probes containing the core
binding sites for LBS-1 and LBS-2 (Fig.
4A). LBS-1 alone showed
affinity similar to that observed for probes containing both sites. In
contrast, the affinity of r-LANA for LBS-2 alone was many orders of
magnitude lower than that observed for binding to the second site in
probes containing both sites; to illustrate the weak binding to LBS-2,
a longer exposure was necessary (Fig. 4, A, lanes
7-12; and B, lanes 1-6).
View larger version (45K):
[in this window]
[in a new window]
|
Fig. 4.
LANA binds cooperatively to LBS-2.
A, relative binding of fragments containing the core binding
site of LBS-1 (lanes 1-6) or LBS-2 (lanes
7-12). Each set of reactions was done in the presence of
increasing concentrations of r-LANA (0, 0.67, 1.7, 3.4, 6.8, and 10 nM). The probe·LANA complex is labeled (b). An
overexposure of lanes 7-12 is shown in the right
panel to visualize low level binding. B, binding of a
wild-type DNA fragment containing both LBS-1 and LBS-2 (lanes
1-6) compared with a DNA fragment with LBS-1 deleted (lanes
7-12). Each set of reactions was done in the presence of
increasing concentrations of r-LANA (0, 0.67, 1.7, 3.4, 6.8, and 10 nM). The low (a) and high (b)
mobility complexes are labeled. An overexposure of lanes
7-12 is shown in the right panel to show low level
binding. Ptn., protein.
|
|
This provides further evidence that LANA binds first to the high
affinity site (LBS-1) and then has an increased affinity for LBS-2. To
show conclusively that binding LBS-1 increases the affinity of LANA for
LBS-2, we used EMSA on two additional fragments. LBS-1/LBS-2 contains
TR sequence from nucleotides 555 to 644, encompassing both LBS-1 and
LBS-2. 
/LBS-2 has the identical sequence, except that 20 bp centered
on the core binding site of LBS-1 have been deleted. If LBS-2 binds
LANA independent of LBS-1, the complex formed by /LBS-2 should be
present in approximately the same quantity as the low mobility complex
from LBS-1/LBS-2. However, we observed that r-LANA clearly had much
lower affinity for the LBS-2 site in the absence of an occupied LBS-1
site (Fig. 4B).
These two experiments, combined with the DNase I footprinting data,
show that LANA first binds to the high affinity site (LBS-1) and that
this binding then greatly increases the affinity of LANA for the second
site. Interestingly, EBNA-1 of EBV binds cooperatively to sites in the
DS element in a similar fashion: a high affinity site capable of
binding protein alone paired with a low affinity site dependent on the
first site for significant binding (29). These similarities in
organization and affinity may reflect some necessity in the mechanism
of replication complex formation or initiation.
Both LBS-1 and LBS-2 Contribute to the Ability of LANA to Suppress
Transcription--
We have previously shown that TR sequences harbor
enhancer activity when inserted in a promoter-containing luciferase
reporter; multimerizing the TR leads to increased enhancer
activity (9). This enhancer effect of the TR can be partially
suppressed if LANA or the C-terminal domain of LANA is cotransfected
with the reporter plasmid (9). Hence, we wanted to investigate whether both LBS-1 and LBS-2 contribute to the ability of LANA to suppress transcription from a TR-containing reporter. Therefore, we prepared a
series of mutations in a single unit of the TR to examine the contributions of LBS-1 and LBS-2. In these mutants, we deleted the high
affinity binding site (LBS-1), the low affinity binding site (LBS-2),
or both sites. In addition, two transversion mutants were created: one
in which the strong binding site was converted to a weak site and one
in which the weak site was converted to a strong site. To conserve the
spatial architecture of the DNA, each deletion consisted of 20 bp,
centered around the core binding site to be removed.
Full-length LANA has been shown to act through a plethora of mechanisms
to both activate and suppress transcription. Most of these effects are
independent of DNA binding and facilitated by protein/protein
interactions with transcriptional cofactors (10-14, 45, 46).
However, the N-terminal domain of LANA has been shown to interact with
mSin3 and the related histone deacetylase complex. A LANA N-terminal
domain-Gal4 fusion protein seems to be capable of suppressing
transcription by directing this complex to a GAL4 promoter;
the LANA C-terminal-Gal4 fusion protein was not able to suppress
transcription in this system (15). In contrast, the C-terminal domain
of LANA suppresses transcription from constructs containing the TR,
which contains its native binding sites (9). In this experiment, we
explored the contributions of LBS-1 and LBS-2 to the ability of the
C-terminal domain to suppress transcription from a construct containing
the TR.
The pGL3 promoter plasmid was used as the reporter construct in these
assays. Although full-length LANA has been shown to trans-activate the SV40 promoter (9, 56), the C-terminal domain alone does not have a significant effect on the SV40 promoter, but does retain suppression activity (Fig.
5B) (9). In the absence of any
effector, the wild-type TR (TR1/2) showed 6-7-fold enhancer
activity in comparison with the SV40 promoter alone. A comparable level
of activation was also observed for all of the mutant constructs,
indicating that regions containing LBS-1 and LBS-2 are not responsible
for the transcriptional enhancer activity of the TR. When a construct
expressing the C-terminal domain of LANA was cotransfected with each
reporter plasmid, it became clear that the ability of LANA to bind the
region is essential for transcriptional repression. Transcription from
TR1/2 was suppressed by 42% (Fig. 5B). As expected, the
mutant lacking LBS-1 and LBS-2 (TR/) was not affected by the
presence of effector. The construct containing LBS-2 with the high
affinity site deleted (TR/2) also showed no significant effect; this
was expected considering that binding to LBS-2 was minimal in the
absence of LBS-1 (Fig. 4B). LBS-1 alone with the low
affinity site deleted (TR1/) suppressed transcription about half as
well as the wild-type TR (Fig. 5B), demonstrating that
binding to LBS-2 contributes significantly to the suppression
phenotype. Fig. 5A illustrates the contribution of each
binding site by showing the level of suppression for each construct
relative to the wild-type TR (wild-type TR equals 100%). When the low
affinity LBS-2 site was converted to a high affinity site (TR1/1), the
suppression affect was at or above the level of the wild-type TR,
indicating that it is the affinity of two complexes of LANA for this
region, not the distinct affinities of the sites, that produces full
suppression. When LBS-2 was duplicated (TR2/2), suppression was around
half that seen with the wild-type sequence. This is again consistent
with suppression being proportional to the affinity of LANA for the
region because duplication of the site in combination with the
cooperative binding between the two low affinity sites would facilitate
significantly more binding than LBS-2 alone (TR/2), but still
considerably less binding than a TR containing a high affinity site
(TR1/2). This experiment shows that transcriptional suppression by the
C-terminal domain of LANA is directly proportional to its binding
affinity to the TR.
View larger version (34K):
[in this window]
[in a new window]
|
Fig. 5.
Both LBS-1 and LBS-2 contribute to the
ability of the C-terminal domain of LANA to suppress
transcription. A, table depicting a series of five TR
mutants. The sequences of these mutants were manipulated such that the
LBS-1 or LBS-2 site was either converted to the other site by the three
nucleotide transversions or contains a 20-bp deletion centered on the
relevant core binding site. Also shown is the percentage of suppression
(Sup.) observed for each mutant TR relative to the
suppression of the wild-type (WT) TR. B,
graphical representation of the primary data from the luciferase
reporter assays. The mutant TR reporters (100 ng each) were
cotransfected with 0, 500, or 1000 ng of an effector plasmid expressing
the C-terminal domain of LANA (C-LANA)
(pcDNA3.1v5HIS/LANAC). DNA concentrations of all transfections were
normalized with empty vector. The percentage of suppression
(sup) achieved at 1000 ng of effector as compared with 0 ng
of effector is shown for each reporter construct. - -, no significant
suppression was observed. The results shown were determined from six
independent transfections.
|
|
Both LBS-1 and LBS-2 Contribute to the Ability of LANA to
Facilitate DNA Replication--
LANA has been shown to be sufficient
for long-term episomal maintenance of a plasmid containing the TR,
presumably by binding to TR sequences and tethering the viral genome to
chromosomal structures during mitosis (16, 17, 57). By performing
short-term replication assays, we have recently demonstrated that not
only is LANA involved in genome segregation, but it is also required for de novo DNA synthesis of TR-containing
plasmids.2 Accordingly, we
used the series of TR mutants discussed above to determine whether
LBS-1 and LBS-2 also contribute to the DNA replication activity of
LANA. To make these determinations, we used a short-term replication
assay that measures the presence of newly synthesized DNA 72 h
post-transfection. This assay relies on the ability of DpnI
to recognize and to cleave DNA produced in dam+
Escherichia coli, but not DNA replicated in eukaryotes. The
series of plasmids containing the modified TRs was transfected into 293 cells in the presence or absence of a construct expressing LANA (pcDNA3/orf73) (56). After 72 h, the cells were harvested and subjected to Hirt extraction. 10% of the harvested DNA was then linearized with HindIII to determine the total amount of
intracellular plasmid DNA. The other 90% was digested with an excess
of DpnI. The resulting fragments were then visualized by
Southern blot analysis. Radiographic densitometry was used to quantify
both the input DNA and the DNA resistant to DpnI digestion.
The amount of eukaryotic DNA was adjusted for the input DNA, and the
efficiency of replication for the mutant TRs was compared with that for
the wild-type TR and expressed as relative activity. The relative activities shown in Fig. 6A
represent the average of two independent experiments.
View larger version (50K):
[in this window]
[in a new window]
|
Fig. 6.
Both LBS-1 and LBS-2 contribute to the
ability of LANA to facilitate DNA replication. A, table
depicting a series of five TR mutants. The sequences of these mutants
were manipulated as described in the legend to Fig. 5. B,
results from a DpnI resistance replication assay in which
293 cells were transfected with 10 µg of DNA derived from
dam+ bacteria, causing it to be sensitive to
DpnI cleavage. 72 h later, plasmid DNA was harvested
from these cells using a modified Hirt extraction. 90% of this DNA was
then digested with 180 units of DpnI for 24 h,
completely digesting the DNA replicated in dam+
bacteria, but not the DNA newly synthesized in the eukaryotic cells,
which was then detected by Southern blotting. Two bands (linearized and
supercoiled plasmid) are seen in each positive lane. The
first and second lanes contain vector control DNA
lacking any TR DNA. The third through fourteenth
lanes contain the same vector containing the TR DNA with the
indicated binding site mutations. Even-numbered lanes were
cotransfected with a vector expressing LANA (pcDNA3/orf73) (56).
10% of the DNA harvested in each Hirt extraction was linearized using
HindIII and Southern-blotted to show relative amounts of DNA
input in each DpnI digestion. The DpnI-resistant
DNA and input DNA were quantitated using radiographic densitometry.
After adjusting for the relative input DNA, the replication activity of
each plasmid was calculated in comparison with that of the wild type
(WT); these values for two independent experiments were
averaged and are shown in A. Neg. Con., negative
control.
|
|
The results of this experiment clearly indicate that the binding sites
LBS-1 and LBS-2 are essential for LANA to facilitate DNA replication.
The TR/ mutation showed no replication. The mutant lacking the
low affinity binding site (TR1/) replicated at 51% efficiency. The
mutant lacking the high affinity site (TR/2) replicated with only
14% the efficiency of the wild-type TR. If the second site was
converted to a high affinity site (TR1/1), replication was nearly
identical to that of the wild-type TR. If the first site was converted
to a low affinity site (TR2/2), replication continued at ~46% of
that of the wild-type TR. These results mirror those of transcriptional
suppression, showing that both LBS-1 and LBS-2 are critical for
efficient DNA replication of TR-containing plasmids. Supporting our
data on transcriptional suppression, both binding sites are essential
for maximum replication, and the efficiency of replication is directly
proportional to the combined affinity of the LANA-binding sites within
the direct repeat.
| |
DISCUSSION |
During latent infection, KSHV maintains its genome as an
extrachromosomal episome. Similar to other DNA viruses, SV40, human papilloma virus, and EBV, the replication of the KSHV genome requires a
trans-acting viral protein (in this case, LANA) to bind at
the origin of replication (ori). These viral ori
elements, as well as cellular ori elements in
general, are associated with numerous transcription factors (44,
58-63). The SV40 and human papilloma virus ori elements
each contain numerous cellular transcription factor-binding sites,
whereas EBV oriP is composed of two elements: the DS
element, which contains the replicator element, and the FR element,
which is a potent enhancer in the presence of EBNA-1 (29, 32).
Here, we show that LANA binds to two sites similar to one-half of the
EBV DS element. LBS-1 is a high affinity site capable of facilitating
the cooperative binding of LANA to LBS-2, much like sites 1 and 4 of
the DS element facilitate binding to sites 2 and 3, respectively (29,
64). Either of the DS pair of sites has been shown to be sufficient for
at least partial replication (29, 35, 65). Although the molecular
details have not been established to the same degree, two other
-herpesviruses, herpesvirus saimiri and herpesvirus papio, also
contain similar DS-like elements in their respective ori
elements (66-68). These conserved similarities suggest that the
disparate affinity and cooperative binding of the two sites are
important to extrachromosomal DNA replication. However, manipulation of
this arrangement by converting LBS-2 into a high affinity site did not
cause a significant reduction in replication. It is possible that this
conserved arrangement of sites is not critical to replication of
relatively small plasmids, but is important in the context of
replicating the entire 140-kb genome. It is also possible that the
distinctive arrangement of this element plays a role in the regulation
of origin firing, which may not be critical in rapidly dividing cell lines.
As we better resolve the KSHV latent ori element, one
notable difference from EBV is the lack of an FR element. The FR
element of EBV is essential for long-term maintenance of episomal DNA, presumably by tethering the plasmid to the chromosomes for proper segregation to daughter cells (34, 69-71). Three copies of the DS
element can be used to replace the FR element and to reconstitute its
maintenance function (32). The strong transcriptional enhancer activity
of the FR element indicates that, when bound by EBNA-1, it also has an
effect on DNA structure, perhaps serving a function analogous to the
cellular transcription factors associated with other ori
elements. Although the KSHV putative latent ori element has
no FR-type structure, the TR, which contains LBS-1 and LBS-2, is
repeated 30-40 times in the KSHV genome; this repetition may be
sufficient to serve the maintenance function in a manner similar to the
multimerized DS element. When multimerized, as is the case in the
genome, the KSHV TR is a potent enhancer (9); transcription factors
employed in this enhancer could serve chromatin-remodeling functions
similar to those of the FR element in EBV.
We have shown that the ability of the LANA C-terminal domain to
suppress TR enhancer-dependent transcription is directly
proportional to its ability to bind at LBS-1 and LBS-2. No other viral
origin-binding protein has been reported to negatively regulate
transcription in this way. Large T antigen and E2 of human papilloma
virus both suppress transcription by binding DNA and inhibiting
transcription complex formation at promoters, but not from enhancers
(72, 73). In most cases, E2 and EBNA-1 act as strong transcriptional activators when bound to DNA (36-42). Some qualities of transcription factor activation must be required for efficient replication function, as transcription factor associations are found in most well studied origins of replication (44, 58-63). In many cases, the efficiency of
replication has been linked to the presence of these transcriptional activators (43, 74, 75). With this in mind, it seems strange that LANA
would be acting as a transcriptional suppressor. However, it is the
presence of transcriptional activators that activate replication, not
transcription itself. In fact, transcriptional activity has been shown
to be inversely proportional to replication in autonomously replicating
chromosomes (76). This finding supports the idea that some contribution
of transcription factors, probably chromatin remodeling, is necessary
for the formation of pre-replication complexes and the subsequent
initiation of DNA replication, whereas transcription itself is inhibitory.
Manipulations of the binding sites of LANA cause similar changes in
both its ability to suppress transcription and to facilitate DNA
replication (Figs. 5 and 6). This indicates that the mechanisms of
these processes may be related or interdependent such that replication
may inhibit local transcription, or transcription may need to be
inhibited to facilitate efficient replication. This hypothesis is
further supported by the fact that the LANA C-terminal domain alone is
sufficient for transcriptional suppression and DNA replication
(9).2 Our finding that the TR element contains enhancer
activity is in congruence with the presence of transcriptional activity
within the vicinity of all cellular and viral origins of replication. The mechanism by which LANA suppresses this transcriptional activation and at the same time facilitates DNA replication needs to be further elucidated. It has recently been shown that origin recognition complex/EBNA-1 interaction is critical for oriP function
(77-79). The Origin recognition complex plays an important role in the assembly of heterochromatin and chromosome condensation (80-82), and
LANA has been shown to co-localize with heterochromatic regions (83).
Our results are consistent with a model in which the binding affinity
of LANA for LBS-1 and LBS-2 within the TR determines its ability to
mediate the interaction between the viral episomal genome, the origin
recognition complex, and a manifest heterochromatic environment,
conducive to DNA replication and inhibitory to transcription.
,
TOP
ABSTRACT
INTRODUCTION
Recipient of Research Oncology Training Grant T32 HL 07147 from
the National Institutes of Health.
