Molecular Cell, Vol. 18, 25–36, April 1, 2005, Copyright ©2005 by Elsevier Inc. DOI 10.1016/j.molcel.2005.02.029
Structure of the p53 Binding Domain of HAUSP/USP7
Bound to Epstein-Barr Nuclear Antigen 1:
Implications for EBV-Mediated Immortalization
Vivian Saridakis,1,6Yi Sheng,2,6Feroz Sarkari,1
Melissa N. Holowaty,1Kathy Shire,1Tin Nguyen,1
Rongguang G. Zhang,5Jack Liao,2Weontae Lee,2
Aled M. Edwards,1,3,4Cheryl H. Arrowsmith,2,3,4
and Lori Frappier1,*
1Department of Medical Genetics and Microbiology
2Ontario Cancer Institute and
Department of Medical Biophysics
3Banting and Best Department of Medical Research
4Structural Genomics Consortium
University of Toronto
Toronto, Ontario M5S 1A8
Structural Biology Center
Argonne National Laboratory
9700 South Cass Avenue
Argonne, Illinois 60439
mor suppressor protein, surprisingly none of the EBV
proteins required for immortalization have been shown
to act through p53.
EBNA1 is the only EBV protein consistently ex-
pressed in all proliferating infected cells and plays sev-
eral important roles in EBV latent infection, including
the initiation of EBV DNA replication, the mitotic segre-
gation of the EBV genomes, and transcriptional activa-
tion of other EBV latency proteins (Kieff and Rickinson,
2001). In addition, several pieces of evidence suggest
that EBNA1 plays a direct role in cellular transformation
by EBV. First, EBNA1 is expressed in all EBV-associ-
ated tumors and is the only viral protein expressed in
some of these tumors. Second, transgenic mice ex-
pressing EBNA1 develop malignant B cell lymphomas
(Wilson et al., 1996). Third, the expression of EBNA1 in
Hodgkin cells enhances their ability to form tumors in
nonobese diabetic-SCID mice (Kube et al., 1999).
Fourth, EBV genomes lacking the EBNA1 gene are sev-
eral thousand-fold less efficient at B cell immortaliza-
tion than EBV genomes expressing EBNA1 (Hume et
al., 2003). Fifth, interference with EBNA1 function in
Burkitt’s lymphoma cells by overexpression of the
EBNA1 DNA binding domain increased cell death, sug-
gesting that EBNA1 normally provides a survival func-
tion for these cells (Kennedy et al., 2003).
EBNA1 was recently shown to stably interact with the
ubiquitin-specific protease called USP7 or HAUSP (Her-
pes virus Associated USP; [Holowaty et al., 2003b]),
which was originally identified as a binding target of the
ICP0 protein of herpes simplex virus (Everett et al.,
1997). EBNA1 sequences mediating this interaction
were mapped to within amino acids 395–450, just
N-terminal to the DNA binding domain. An EBNA1 mu-
tant lacking this sequence (?395–450) failed to bind
USP7 but continued to bind other known cellular pro-
tein targets of EBNA1. Functional studies with ?395–
450 showed that USP7 binding was not required for the
replication, segregation, or transcriptional activation
functions of EBNA1 but may inhibit the ability of EBNA1
to activate replication (Holowaty et al., 2003b). The
deletion of the USP7 binding sequence also had no de-
tectable effect on EBNA1 turnover or cell surface pre-
sentation. The lack of requirement of the USP7 interac-
tion for the known EBNA1 functions suggested that the
significance of this interaction may lie in EBNA1-
induced changes to the cell.
A link to the p53 pathway was revealed by Li et al.
(2002b), who showed that USP7 bound and deubiquiti-
nated p53. Overexpression of USP7 stabilized p53, re-
sulting in p53-mediated growth repression and apopto-
sis, whereas decreased USP7 levels destabilized p53.
However, the role of USP7 in p53 regulation was re-
cently shown to be more complicated than originally
thought, as ablation of USP7 expression resulted in p53
accumulation, as opposed to the expected destabiliza-
tion of p53 (Cummins et al., 2004; Li et al., 2004). This
effect has been shown to be the result of the ability of
USP7 to stabilize Mdm2, a ubiquitin ligase that pro-
motes the degradation of p53. Therefore USP7 appears
USP7/HAUSP is a key regulator of p53 and Mdm2 and
is targeted by the Epstein-Barr nuclear antigen 1
(EBNA1) protein of Epstein-Barr virus (EBV). We have
determined the crystal structure of the p53 binding
domain of USP7 alone and bound to an EBNA1 pep-
tide. This domain is an eight-stranded ? sandwich
similar to the TRAF-C domains of TNF-receptor asso-
ciated factors, although the mode of peptide binding
differs significantly from previously observed TRAF-
peptide interactions in the sequence (DPGEGPS) and
the conformation of the bound peptide. NMR chemi-
cal shift analyses of USP7 bound by EBNA1 and p53
indicated that p53 binds the same pocket as EBNA1
but makes less extensive contacts with USP7. Func-
tional studies indicated that EBNA1 binding to USP7
can protect cells from apoptotic challenge by low-
ering p53 levels. The data provide a structural and
conceptual framework for understanding how EBNA1
might contribute to the survival of Epstein-Barr virus-
EBV infects more than 90% of people worldwide and
efficiently immortalizes infected cells, predisposing the
host to a variety of cancers. Cellular immortalization by
EBV occurs as part of its latent infectious cycle and
involves a few EBV proteins including LMP1, which
mimics an activated tumor necrosis factor receptor,
and EBNA2, which activates the transcription of several
cellular and viral genes (Dolcetti and Masucci, 2003).
Whereas cellular transformation by other DNA tumor
viruses (e.g., adenovirus, SV40, and papillomavirus) has
clearly been shown to involve targeting of the p53 tu-
6These authors contributed equally to this work.
Table 1. X-Ray Data Collection, Structure Solution, and Refinement Parameters
X-Ray DataNative Peak EBNA1 Complex
Unit cell axes (Å3)
Se sites (No.)
Total observations (No.)
Unique reflections (No.)
bFigure of merit (%)
Protein atoms (No.)
Water molecules (No.)
Sodium atoms (No.)
Rmsd bonds (Å)
Rmsd angles (°)
Rmsd dihedrals (°)
Rmsd improper (°)
Thermal factor (Å2)
102.6 × 102.6 × 45.2
102.5 × 102.5 × 45.0
70.0 × 70.0 × 45.9
Numbers in brackets refer to the highest resolution shell, 1.97–1.90 Å for the native data, 2.07–2.00 Å for the MAD data, and 1.76–1.70 Å for
the EBNA1 complex data. Data were integrated and scaled by using HKL2000.
aRsym= S|I − <I>|/SI where I is the observed intensity and <I> is the average intensity from multiple observations of symmetry-related
bFigure of Merit of Phasing = |SP(α)eια|/SP(α) where P(α) is the phase probability distribution and α is the phase angle.
cR = S|Fobs− Fcalc|/|Fobs|.
to play multiple roles in regulating the p53-Mdm2
The USP7-p53 interaction occurs between the USP7
N-terminal domain (NTD), amino acids 53–208, and res-
idues 357–382 of the C-terminal regulatory region of
p53 (Hu et al., 2002). This USP7 domain, which is sim-
ilar in sequence to a TRAF domain (Zapata et al., 2001),
is also responsible for the interaction with EBNA1 (Ho-
lowaty et al., 2003a). The fact that EBNA1 and p53 bind
the same domain of USP7 raises the possibility that
EBNA1 affects the regulation of p53 by disrupting the
interaction of USP7 with p53. Indeed, the 395–450 frag-
ment of EBNA1 binds USP7 with 1 ?M affinity, whereas
the p53 regulatory fragment (with or without the tetra-
merization domain) binds with 10-fold lower affinity
(Holowaty et al., 2003a). The EBNA1 peptide 395–450
also displaces the p53 peptide from the p53-USP7
complex. These results indicate that EBNA1 and p53
bind the same or overlapping sites in the USP7 NTD
and suggest that EBNA1 could sequester USP7 from
p53 in vivo, thereby destabilizing p53. In this paper, we
provide structural and functional evidence supporting
a connection between EBNA1 and p53.
structure solution and refinement statistics). The struc-
ture of USP7 was determined by using single anoma-
lous dispersion, and the model was refined to 2.0 Å
resolution. USP7 NTD is composed of a single domain
with approximate dimensions 51 × 30 × 30 Å3, with
structural similarity and identical topology to the
TRAF-C domain of tumor necrosis factor-receptor as-
sociated factors (TRAFs) 2, 3, and 6 (Li et al., 2002a; Ni
et al., 2000; Park et al., 1999; Ye et al., 2002; Ye et al.,
1999). Like all TRAF domains, USP7 forms an eight-
stranded, antiparallel β sandwich (Figure 1A). Strand β7
contains a β bulge, which is important in peptide bind-
ing and found in all TRAF domain structures. Structure
comparison using DALI identified the TRAF domain of
TRAF2 (PDB accession number 1D0A) as the closest
structural neighbor of the USP7 NTD with a Z score of
9.4 and an rmsd of 2.8 Å over 93 Cαatoms (see Fig-
ure 1C for superposition of these domains). A struc-
ture-based sequence alignment for the TRAF domains
of USP7, TRAF2, TRAF3, and TRAF6 is shown in Fig-
Mapping of USP7 Binding Site on EBNA1
An EBNA1 fragment containing amino acids 395–450
stably binds the USP7 NTD (Holowaty et al., 2003a).
Because TRAF domains typically bind peptides of ten
amino acids or less, we imagined that the EBNA1 se-
quence involved in USP7 binding could be further local-
ized to a shorter amino acid sequence. Consensus
binding motifs have been identified that are important
for binding TRAFs 2, 3, and 6 (see below), prompting
us to search for these motifs in the EBNA1 395–450
Structure of the USP7 NTD
To gain insight into the molecular basis for the EBNA1
and p53 interactions with USP7, the structure of the
USP7 NTD (residues 54–204), previously shown to bind
p53 and EBNA1, was determined; first alone and then
in complex with a peptide from EBNA1 (see Table 1 for
Structural Basis for EBNA1 and p53 Binding to USP7
Figure 1. Crystal Structure of the USP7 NTD
with EBNA1 Peptide
(A) Ribbon representation of the crystal
structure of the USP7 NTD bound by EBNA1
peptide (stick form).
(B) Electrostatic surface representation of (A).
(C) Superposition of the TRAF domains of
USP7 (blue) with TRAF2 (silver).
(D) Structure based sequence alignment be-
tween the TRAF domains of USP7, TRAF2,
TRAF3, and TRAF6. Residues that are iden-
tical (asterisks) or conserved (dots) in all four
sequences are indicated. Residues involved
in EBNA1 binding are in bold.
fragments. Because obvious matches to any of these
motifs were not evident, the USP7 binding sequence in
EBNA1 was localized experimentally. A series of EBNA1
peptides fused to GST (420–450, 421–435, 426–440,
431–445, and 436–450) were used to assess interac-
tions with the USP7 NTD by GST pull-down experi-
ments. The results showed a stoichiometric interaction
between USP7 and fusion proteins containing EBNA1
residues 420–450 and 436–450 but little or no interac-
tion with fusion proteins containing EBNA1 residues
421–435, 426–440, and 431–445 (Figure 2A). The USP7
binding sequence of EBNA1 is therefore found within
Figure 2. Interactions of EBNA1 Peptides
with the USP7 NTD
(A) An equimolar mixture of the USP7 NTD
and a GST fusion protein containing the indi-
cated EBNA1 peptide (L) was mixed with
glutathione-sepharose. After washing, pro-
tein was eluted with glutathione (E).
(B) Increasing amounts of EBNA1 peptides
395–450 with wild-type sequence (circles) or
E444A (squares), S447A (triangles) or E444A/
S447A (diamonds) mutations were incubated
with the USP7 NTD and binding was quanti-
fied by change in tryptophan fluorescence.
amino acids 436–450, and residues between 445 and
450 are essential for the USP7 interaction. This informa-
tion was used to guide EBNA1 peptide design for
cocrystallization trials with USP7.
peptide. There were also differences in the lattice con-
tacts formed in crystals of the USP7 NTD in the pres-
ence and absence of EBNA1 peptide. In the absence
of the peptide, the USP7 NTDs interacted through the
peptide binding surface and, as a result, these crystals
were disrupted by soaking in the EBNA1 peptide. In
crystals generated by the USP7-EBNA1 complex, lat-
tice contacts were not observed in the peptide binding
region but were confined to the loops connecting the
β strands (amino acids 77, 80, and 82 interacted with
residues 178, 180, and 181).
Structure Overview of the USP7-EBNA1 Complex
A peptide corresponding to EBNA1 amino acids 441–
450 was cocrystallized with the USP7 NTD. The struc-
ture of the complex was determined by using molecular
replacement, and the model was refined to 1.7 Å resolu-
tion. The EBNA1 peptide (acetyl-DPGEGPSTGP-amide)
was located in the Fo− Fcdifference electron density
map after the initial round of refinement with the protein
model alone (Figure 3E). Excellent density allowed resi-
dues 441#–448# of the EBNA1 peptide to be built unam-
biguously. There was no visible density for the amino-
terminal acetyl moiety and residues 449# and 450#. The
EBNA1 peptide was in an extended conformation and
bound to the edge of the β sandwich. The peptide
formed main chain H bonds solely with strand β7 in an
antiparallel manner, increasing the number of strands in
one of the sheets from four to five (Figures 1A and 3F).
Like the unbound USP7 NTD, the EBNA1 bound do-
main was very similar in structure to other TRAF do-
mains (Figure 1C). No significant conformational changes
were observed between the peptide-free and EBNA1
bound USP7 NTD. The rmsd was 0.4 Å over 102 Cα
residues. Within the region of peptide binding, there
were some minor changes. In the peptide complex, the
side chain of Asp164 rotated to interact with the side
chain hydroxyl of Ser447#, and the side chain of Trp165
rotated to interact with OD1 of Glu444#. Specifically,
atom OD1 of Asp164 moved 3.3 Å toward the peptide,
and the NE1 atom of Trp165 moved 1.4 Å toward the
The EBNA1 peptide bound to a shallow depression on
the surface of USP7 (Figure 1B). Upon binding to USP7,
the EBNA1 peptide buried a total of 742.3 Å2of surface
area. A combination of hydrophobic and hydrophilic in-
teractions was formed between EBNA1 and USP7 (Fig-
ure 3F). These interactions occur predominantly with
strand β7; however, strand β6 and loop β3-β4 are also
involved. H-bonds are formed between OE1 of EBNA1-
Glu444# and NE1 of Trp165 as well as between OG of
EBNA1-Ser447# and OD1 of Asp164. These are the two
interactions that are predicted to confer specificity to
EBNA1 for USP7 as they are the only H-bonding in-
teractions that occur between side chains of USP7 and
EBNA1. The carbonyl group of EBNA1-Ser447# in-
teracts with NH1 of Arg104 from loop β3-β4. OE2 of
EBNA1-Glu444# interacts with the amide of Ser155, and
OE1 of EBNA1-Glu444# interacts with the carbonyl of
Arg153 on strand β6 through bridging water molecules.
The carbonyl of EBNA1-Pro 442# interacts with the am-
ide side chain of Asn169. There are also four main chain
H bonds formed between EBNA1 and strand β7. Hy-
Structural Basis for EBNA1 and p53 Binding to USP7
Figure 3. The Path and Contacts of the EBNA1 Peptide Bound to USP7
(A–C) Transparent surface representation of USP7 bound to EBNA1 peptide (A), TRAF3 bound to TANK peptide (177#-CSVPIQCTDKT-187#;
PDB accession number 1KZZ) (B), and TRAF6 bound to CD40 peptide (230#-KQEPQEODF-238#; PDB accession number 1LB6) (C).
(D) Comparison of the bound conformations of EBNA1 (red), TANK (green), and CD40 (blue) peptides in the identical orientation as above
(generated by superimposing the TRAF domains of the three TRAF-peptide complexes).
(E) Electron density of the EBNA1 peptide. The final EBNA1 model is shown in the Fo− Fcdifference density obtained after the initial rounds
of refinement in the absence of peptide.
(F) Detailed interactions between USP7 (silver) and EBNA1 (charcoal) shown in stereo. The H bonds are indicated by dashed lines.
drophobic interactions occur between the aliphatic
side chains of EBNA1-Glu444# and Phe167, EBNA1-
Ser447# and Phe118, as well as EBNA1-Gly445# and
bound to other TRAF domains. However, the conforma-
tion of the EBNA1 peptide differs from that of all other
TRAF bound peptides whose structures have been de-
termined, as shown in comparison to TRAF3 and
TRAF6 bound peptides (Figures 3A–3D). Peptides
bound to TRAF2 and TRAF3 run from top to bottom of
the TRAF domain in an extended conformation (Figure
3B) cutting across and largely perpendicular to strands
β3, β4, β6, and β7 (Ni et al., 2000; Park et al., 1999; Ye
Comparison of the Peptide Interactions of USP7
and Other TRAF Domains
The EBNA1 peptide binds to a groove within the USP7
TRAF domain at a position similar to that of peptides
et al., 1999). The path of TRAF6 bound peptides is sim-
ilar to this, although it deviates from the TRAF2 peptide
path by 40° and makes more extensive contacts with
strand β7 (Li et al., 2002a; Ye et al., 2002) (Figure 3C).
The most extensive β7 contacts have been seen with
the CD40 peptide, which makes five main-chain H
bonds with strand β7. EBNA1 peptide follows strand β7
even more extensively than CD40, making seven dif-
ferent contacts with six different β7 residues, and, as a
result, bends relative to the other peptides at the posi-
tion of Glu444# (Figures 3A and 3D).
Another difference that distinguishes the EBNA1-
USP7 interaction from other peptide-TRAF interactions
is the sequence of the bound EBNA1 peptide, which
does not conform to the known TRAF binding se-
quence motifs. TRAF-peptide interactions can be clas-
sified into two groups based on their peptide specific-
ity. TRAF6 binds the consensus sequence, PxExx
(f/acidic) (Ye et al., 2002), whereas TRAFs 1, 2, 3, and
5 all bind the consensus sequence (P/S/A/T) × (Q/E)E
and its variants (fSxEE, QEE, and PxQxxD) (Park et al.,
1999; Ye et al., 1999). The TRAF1, TRAF2, TRAF3, and
TRAF5 residues involved in peptide binding are abso-
lutely conserved (including Arg393, Tyr395, Phe447,
Ser453, Ser454, Ser455, and Ser467 according to
TRAF2 numbering) despite the fact that these TRAF do-
mains share only 52%–64% sequence identity (Wajant
et al., 2001).
The EBNA1 peptide bound by USP7 (PGEGPS) does
not match any of the consensus sequences identified
for binding other TRAF domains. Although the EBNA1
sequence contains a PxE motif, which is a found within
some of the TRAF binding consensus sites, the struc-
tural contacts made by these EBNA1 residues are dif-
ferent than those of other TRAF bound peptides. In all
previously solved TRAF domains, the Pro and Gln/Glu
residues in the bound peptides are largely super impo-
sable. However, the Pro and Glu residues of the EBNA1
peptide are out of register in comparison to the TRAF
bound TANK and CD40 peptides (Figure 3D). The Cαof
EBNA1-Glu444# is 7.1 Å away from the Cαof Gln182#
from the TANK peptide and 7.4 Å away from the Cαof
Glu235# from the CD40 peptide. The Cαof Pro442# from
EBNA1 is 9.4 Å away from the Cαof Pro180# from the
TANK peptide and 6.8 Å away from the Cαof Pro233#
from the CD40 peptide. The functions of the PxE motifs
are also different between EBNA1 and the other TRAF
binding peptides. For example, in all TRAF2, TRAF3,
and TRAF6 interactions, the Pro and Gln/Glu residues
of the bound peptide make specific interactions with
the TRAF domain and within the peptide itself (Li et al.,
2002a; Li et al., 2003; McWhirter et al., 1999; Ni et al.,
2000; Park et al., 1999; Ye et al., 2002). In EBNA1-USP7
interaction, the specific interactions are mediated by
the Glu and Ser residues. Thus the PxE sequence in
EBNA1 appears to be fortuitous.
Consistent with the different peptide sequence
bound by USP7, the ten USP7 residues that contact
the EBNA1 peptide are not conserved in other TRAF
domains (Figure 1D), with the exception of Phe118 and
Gly166 (USP7 numbering), which make similar peptide
interactions in all TRAF-peptide structures. Phe118 and
Gly166 make aliphatic interactions with the appropri-
ately positioned peptide residues and two main chain
H bonds with the peptide, respectively. USP7 does not
contain the residues that mediate peptide interactions
in TRAF2 and TRAF3 and are highly conserved in
TRAFs 1, 2, 3, and 5 (Arg393, Tyr395, Phe447, Ser453,
Ser454, Ser455, and Ser467) nor are residues at most
of these positions used for EBNA1 peptide binding.
Similarly, the residues in TRAF6 that are important for
peptide binding (Arg392, Phe471, and Tyr473) are not
conserved in USP7. Therefore, USP7 forms a new class
of peptide binding TRAF domain.
To further investigate the specificity of the USP7
TRAF domain, we tested binding to five different pep-
tides known to bind TRAF2 or TRAF6; namely, the
TRAF2 binding peptides QVPFSKEEC from TNF-R2
(Park et al., 1999), PQQATDDSS from LMP1 (Ye et al.,
1999), and PVQETLH from hCD40 (Ye et al., 1999), and
the TRAF6 binding peptides QMPTEDEY from TRANCE-R
(Ye et al., 2002) and KQEPQEIDF from hCD40 (Ye et al.,
2002). None of these peptides gave detectable binding
to USP7 at any of the concentrations tested (up to 100
?M), whereas the cocrystallized EBNA1 10 mer peptide
bound USP7 with a Kdof 0.86 ?M in the same assay.
Because EBNA1-USP7 sequence-specific contacts in
our structure were mediated by Glu444# and Ser447# of
EBNA1, we also generated the EBNA1 395–450 amino
acid fragment with these two residues mutated to ala-
nines in order to verify their importance for USP7 bind-
ing. In keeping with the structural information, mutation
of these residues severely decreased binding to USP7
such that no binding was detected up to 100 ?M,
whereas the wild-type (wt) 395–450 fragment bound
USP7 with a Kdof 0.97 ?M (Figure 2B). Point mutation
of the Glu444# residue alone only slightly decreased
binding, giving a Kdof 1.36 ?M, whereas no binding
was detected when Ser447# was mutated to alanine
(Figure 2B). Thus, Ser447 plays a major role in USP7
Comparison of the Interaction of p53 and EBNA1
Peptides with USP7 by NMR
EBNA1 and p53 bind to the TRAF domain of USP7 and
compete for binding. To elucidate the molecular basis
of p53 binding, we used NMR chemical shift mapping
to compare the binding of EBNA1 and p53 peptides to
USP7. 85% of the backbone resonances were assigned
for the NTD of USP7 (amino acids 62–205), and NMR
titration experiments were performed on uniformly15N-
labeled USP7 in the presence of unlabeled p53 (amino
acids 355–393) and unlabeled EBNA1 (amino acids
410–450). Upon binding EBNA1 peptide, the reso-
nances in β7 of USP7 showed strong perturbations (Fig-
ure 4A), especially residues Trp165, Gly166, Phe167,
Ser168, Asn169, Phe170, and Met171. The resonances
for some residues in β3, β4, and β6 were also affected,
including Met100, Met102, Phe 117, Phe118, and
Ser155. The results from NMR mapping data agree well
with the crystal structure of the USP7 NTD EBNA1 com-
plex (Figure 4C). Phe117 and Phe118 of β4 form con-
tacts with the backbone amide groups of the peptide.
Though the backbone amides of Met100 and Met102
(β3) seem far away from the peptide, their side chains
are in close proximity of the C terminus of the peptide.
Shorter EBNA1 peptides (amino acids 420–450 and
Structural Basis for EBNA1 and p53 Binding to USP7
Figure 4. Comparison of the Changes in NMR Resonance Frequencies of USP7 upon Binding EBNA1 and p53 Peptides
Composite chemical shift changes versus residue number for the USP7 residues 62–205 in the presence of EBNA1 410–450 (A) and p53 355–
393 (B). The values shown were calculated by using the equation ?δcomp= [?δ2HN+ (?δN/5)2]1/2. The approximate locations of the USP7 NTD
secondary structure elements are shown on top with an arrow for β strands and a rod for α helices. Chemical shifts of 0.15 ?δppm or greater
induced upon binding EBNA1 (C) or p53 (D) peptides are indicated by the colored residues on the surface representation of the USP7 TRAF
domain from the cocrystal structure. Shifted amino acids from strands β3, β4, β6, and β7 are colored in cyan, yellow, purple, and green,
respectively. In (C) the position of the EBNA1 peptide from the cocrystal structure is also shown.
441–450) were also used in these experiments and gave
the same chemical shift perturbation pattern as the
longer one, indicating that EBNA1 residues outside
441–450 do not contact USP7 (data not shown).
The addition of the p53 peptide also induced chemi-
cal shift perturbations in the1H-15N HSQC spectra of
USP7. All chemical shift changes observed in the p53
complex were also observed in the EBNA1 complex.
However, compared with EBNA1 peptides, p53-induced
changes were smaller both in magnitude and in the
number of affected residues (Figures 4B and 4D). In the
p53 complex, strong chemical shift changes were ob-
served in β7 but only for residues Trp165, Gly166, and
Phe167, indicating a smaller binding surface and a
lower binding affinity for p53 peptide. The chemical
shift patterns in the EBNA1 complex suggest that
EBNA1 extends its binding site further along β7 and
forms more specific interaction with the USP7 NTD.
Residues Phe117, Phe118, and Leu119 from β4 were
affected in the p53 complex, suggesting that interac-
tions of the USP7 NTD with the C terminus of the pep-
tides are similar for both p53 and EBNA1 peptides.
Chemical shift changes for the residues in β3 and β6
(Met100, Met102, and Ser155) that were affected in the
EBNA1 complex were not observed in the p53 complex.
Functional Studies on the EBNA1-USP7 Interaction
Our biochemical (Holowaty et al., 2003a) and structural
studies on the EBNA1-USP7 interaction both indicate
that EBNA1 binding to USP7 could disrupt the interac-
tion of USP7 with p53, which would be predicted to
destabilize p53. Therefore EBNA1 expression in human
cells might prevent p53-induced cell cycle arrest or
apoptosis and contribute to cell immortalization by
EBV. To test this possibility, U2OS cells were cotrans-
fected with a plasmid expressing dsRed (transfection
marker) and either an empty plasmid or a plasmid ex-
pressing EBNA1 or the ?395–450 EBNA1 mutant. This
EBNA1 mutant does not bind USP7 but has the same
stability and other known protein interactions as wt
EBNA1 (Holowaty et al., 2003b). The cells were then UV
irradiated to induce apoptosis, which was detected by
deoxynucleotidyltransferase-mediated dUTP nick end
labeling (TUNEL) assay followed by FACS analysis.
EBNA1 expression decreased the percentage of apo-
ptotic cells after UV irradiation, whereasa the ?395–450
EBNA1 mutant had little effect (Figure 5A) despite being
expressed at similar levels as EBNA1 (data not shown).
Similar results were obtained when apoptosis was in-
duced in H1299 p53 null cells by p53 overexpression
from a transfected plasmid and detected by Annexin V
staining (Figure 5B). The percentage of transfected
cells (as determined by the EGFP marker) that was An-
nexin-positive was compared with and without p53
overexpression in the presence and absence of EBNA1.
EBNA1 reduced the percentage of apoptotic cells gen-
erated by p53 overexpression to just above the back-
ground level of Annexin V staining seen in the absence
of p53, but little effect was seen with the ?395–450
EBNA1 mutant. Results in Figures 5A and 5B are repre-
sentative experiments; EBNA1 effects are similar in re-
peat experiments, but absolute numbers of apoptotic
cells vary between experiments. The results indicate
that, at least under these experimental conditions,
EBNA1 can protect cells from apoptotic challenge, and
this effect is largely due to the USP7 binding region
The above results suggest that EBNA1 may lower
p53 levels by disrupting the p53-USP7 interaction. To
test this possibility, we transfected U2OS cells with an
oriP plasmid expressing EBNA1, ?395–450, or no
EBNA1 and grew the cells under selection for the plas-
mid for 2 weeks. At that point, EBNA1 and ?395–450
were found to be expressed in w60% of the cells at
levels equivalent to those in the EBV-transformed Raji
cells (data not shown). Aliquots of the cells were then
analyzed for p53 expression by Western blotting before
and after induction of p53 by UV irradiation (Figures 5C
and 5D). Although the relative level of p53 in the cell
lines expressing and not expressing EBNA1 varied prior
to UV treatment (i.e., p53 levels in EBNA1-expressing
cells were either equivalent to or lower than those lack-
ing EBNA1 expression), the degree to which p53 levels
increased upon UV induction was consistently less in
EBNA1-expressing cells than in cells expressing ?395–
450 or not expressing EBNA1. Therefore, EBNA1 can
interfere with the stabilization of p53 by USP7 in re-
sponse to UV irradiation.
EBNA1-USP7 Interaction in B cells
The interaction between EBNA1 and USP7 to date has
been studied in epithelial cells. This is biologically rel-
evant because EBV infects epithelial cells and can
promote their transformation as evidenced by the in-
duction of nasopharyngeal carcinoma and other epithe-
lial-based tumors. However, because B lymphocytes
are a major site of EBV infection and persistence, we
wanted to verify that the EBNA1-USP7 interaction
could also occur in this cell background. To this end,
we used the EBNA1 affinity column approach, which
(B) H1299 cells were transfected with an EGFP expression plasmid,
a p53 expression plasmid or empty plasmid, and a plasmid ex-
pressing EBNA1, ?395–450, or no EBNA1. Cells were stained for
Annexin V and FACS sorted. The percentage of transfected (EGFP
positive) cells that stained with Annexin V are shown after subtrac-
tion of the background staining seen in the absence of p53 ex-
(C) U2OS cells expressing EBNA1, ?395–450, or no EBNA1 were
lysed before (0), 4 hr after, or 8 hr after UV irradiation and analyzed
for p53 and actin levels by Western blotting.
(D) Combined data from three experiments showing relative levels
of p53 after UV induction in U2OS cells expressing EBNA1, ?395-
450, or no EBNA1. SD is indicated by the error bars.
Figure 5. EBNA1 Expression Affects Cell Survival and p53 Levels
through the USP7 Binding Region
(A) U2OS cells were transfected with plasmids expressing dsRed
and the EBNA1 proteins indicated, then analyzed by TUNEL assay
and flow cytometry before and after UV irradiation. The percentage
of transfected (dsRed positive) cells that became TUNEL positive
after UV irradiation is shown.
Structural Basis for EBNA1 and p53 Binding to USP7
We also tested whether EBNA1 and USP7 would
coimmunoprecipitate from EBV-infected B cells by
using the Raji Burkitt’s lymphoma cell line. We found
that a proportion of the USP7 from Raji cell lysates
could be immunoprecipitated and that EBNA1 coimmu-
EBNA1 efficiently interacts with USP7 in B cells as it
does in epithelial cells.
USP7 plays an important role in regulating cell prolifera-
tion and apoptosis through p53 and Mdm2 interactions.
We have determined the structure of the p53 binding
domain of USP7 alone and bound to an EBNA1 peptide
and found that it forms a TRAF domain. This is consis-
tent with a previous prediction based on limited se-
quence homology between this region of USP7 and
known TRAF domains (Zapata et al., 2001). The location
and orientation of the EBNA1 peptide on USP7 is sim-
ilar to other known peptide-TRAF interactions; how-
ever, the sequence of the bound EBNA1 peptide
(PGEGPS), the specific USP7 residues contacted by
the peptide, and the conformation of the bound peptide
are all significantly different from other TRAF-peptide
The USP7 TRAF domain is also unique in that it is
a monomer. No oligomeric interactions between USP7
NTDs were observed in the crystal structure, consistent
with our previous analysis of full-length USP7 by analy-
tical centrifugation that indicated that USP7 is a mono-
mer (Holowaty et al., 2003a). TRAF proteins that medi-
ate signaling form homo- and heterotrimers that involve
interactions between the loops in the TRAF domains as
well as through a coiled coil found just N-terminal to
the TRAF domain (Chung et al., 2002; Park et al., 1999).
There is no evidence of such a coiled-coil domain in
USP7. Although some interaction of the purified USP7
NTD with in vitro-translated TRAFs was reported (Za-
pata et al., 2001), it remains to be determined if USP7
can heterotrimerize with TRAF proteins in vivo. The only
other TRAF domain whose structure has been deter-
mined is that of the Siah ubiquitin ligase, and this TRAF
domain forms dimers (Polekhina et al., 2002).
p53 binds the USP7 NTD through residues located
between amino acids 357 and 382 but with a 10-fold
lower affinity than EBNA1 (Holowaty et al., 2003a; Hu
et al., 2002). This p53 region does not contain a match
to the USP7 binding sequence of EBNA1 nor is the
ExxS motif that is responsible for sequence-specific
contacts between EBNA1 and USP7 evident in the
USP7 binding region of p53. This indicates that the
peptide binding pocket in the TRAF domain of USP7
can accommodate sequence variation and the exact
p53 sequence (within the 357–382 fragment) that con-
tacts USP7 will have to be determined experimentally.
We are currently defining the p53 sequence bound by
USP7, which should enable the generation of p53-
USP7 cocrystals suitable for structure determination.
EBNA1 binds USP7 through amino acids 442–447.
Alignments between EBNA1 homologs of EBV-like
viruses that infected other primates (cercopithicine her-
pesvirus 15, cynomolgus EBV, and Herpesvirus papio)
Figure 6. EBNA1-USP7 Interaction in B Cells
(A) Comparison of cellular proteins retained on an EBNA1 affinity
column from B cell (BL41) and HeLa cell lysates. A negative control
of BL41 lysates applied to a column lacking EBNA1 is also shown.
Labeled protein bands were identified by MALDI-ToF mass spec-
(B) Coimmunoprecipitation of EBNA1 with USP7 from Raji cell ly-
sates (L) using the indicated amount of USP7 antibody. Western
blots were probed with antibodies against USP7 (top) and EBNA1
we had originally used to profile epithelial cell protein
interactions of EBNA1 (Holowaty et al., 2003b), to com-
pare the EBNA1 interactions seen in HeLa and B cell
(BL41) lysates. The specific protein interactions of
EBNA1 that were defined with HeLa lysates are also
seen in B cell lysates, although the relative efficiency of
the interactions varies in the two lysates (Figure 6A).
Notably, the interaction with USP7 is more prevalent in
the B cell lysate than in the HeLa cell lysate.
mutations in Glu444, Ser447, or both) were generated by expres-
sion of relevant sequences from pET15b (Novagen) and purification
as described in Holowaty et al. (2003a). Resulting clones were se-
showed that, this USP7 binding sequence (PGEGPS) is
absolutely conserved in these viruses, whereas se-
quences in the 395–430 region are highly divergent.
This suggests that EBNA1 residues 442–447 are func-
tionally important for the virus and that USP7 interac-
tions likely also occur with EBV-related viruses.
Because USP7 binding to p53 results in the stabiliza-
tion of p53 (Li et al., 2002b), our structural data on the
EBNA1 and p53 interactions with USP7 predict that
EBNA1 would interfere with the stabilization of p53 by
blocking the p53-USP7 interaction. In keeping with the
prediction, EBNA1, but not an EBNA1 mutant deficient
in USP7 binding, was found to increase the survival of
cells that were induced to undergo apoptosis either by
DNA damage or p53 overexpression. Similar antiapo-
ptotic effects of EBNA1 have been reported by Ken-
nedy et al. (2003). In keeping with this antiapoptotic ef-
fect, EBNA1 was found to decrease the stabilization of
p53 that occurs in response to UV-induced DNA dam-
age, and this effect required the USP7 binding region
of EBNA1. Not surprisingly, we have not seen reproduc-
ible effects of EBNA1 expression on p53 levels in
rapidly growing cells where p53 is unstable and ex-
pressed at low levels. The apoptotic protection experi-
ments presented here were performed in the presence
of EBNA1 that was expressed at levels approximately
8-fold higher than in Raji cells, and it remains to be
determined whether similar protection is conferred by
EBNA1 in EBV-infected cells. However, this possibility
is supported by the findings that (1) EBNA1 binds USP7
in EBV-infected cells and (2) EBNA1 interferes with the
UV-induced stabilization of p53 when expressed at
levels similar to those in EBV-infected cells. Overall, the
data indicate that EBNA1 can indirectly destabilize p53
by binding USP7, which could be important for initial
cell immortalization by EBV, continued proliferation and
survival of latently infected cells, and/or malignant
Crystallization, Data Collection, and Structure
Determination of Unbound USP7 NTD
Crystals of native or Se-Met-enriched USP7 NTD (30 mg/ml) were
obtained in 35% MPD, 0.2 M MgOAc, and 0.1 M MES, (pH 6.5).
Complete native and MAD data sets from frozen crystals were col-
lected at beamline 19ID at the Advanced Photon Source by using
the SBC-3 CCD detector. Data collection statistics are presented
in Table 1. The native and MAD data were merohedrally twinned
with a twinning factor of 0.36 for the native data and 0.40 for the
MAD data. The structure was determined by using single anoma-
lous dispersion with the peak data. There were three USP7 mole-
cules in the asymmetric unit and each was monomeric. The sele-
nium substructure was located as described in Saridakis et al.
(2004), and 73 out of 151 amino acids of molecule A were built
automatically (residues 70–77, 85–100, 116–121, 128–158, and 191–
202). Models of molecules B and C were built manually. The final
protein model of molecule A consists of 97 residues from 65–78,
85–104, 115–142, 150–173, and 188–203. The amino terminus and
four loops are completely disordered. In molecule B, the following
residues were modeled: 67–78, 85–104, 115–142, 150–175, and
189–204 and in molecule C, 67–74, 85–102, 115–142, 150–178, and
188–204. The rmsd between the different molecules ranges from
1.0 Å for molecules B and C to 1.3 Å for molecules A and C or B
and C over the same number of Cαresidues. The final models were
refined to 2.0 Å with an Rcrystof 23.5 and an Rfreeof 29.5 and con-
tain 62 water molecules. All of the residues are in the best regions
of the Ramachandran plot.
Crystallization, Data Collection, and Structure
Determination of EBNA1 Bound USP7 NTD
USP7 NTD (100 mg/ml) was cocrystallized with a 10 mer EBNA1
peptide corresponding to EBNA1 amino acids 441–450 at 1.5-fold
molar excess of peptide. Large clusters of rods appeared after 4
weeks at 4°C in the dark in three conditions containing 30% PEG
4000, 0.1 M Tris (pH 8.5), and either lithium sulfate, magnesium
chloride, or sodium acetate. The structure was determined by using
molecular replacement and was refined and rebuilt as described in
Saridakis et al. (2004). The final model refined to an Rworkof 0.21
and an Rfreeof 0.25. There are 186 water molecules and 21 sodium
ions. All residues are in the most favored and additionally allowed
regions of the Ramachandran plot. Residues 54–62 and 107–111
are disordered in the final model of the complex. A summary of
data collection and refinement statistics is presented in Table 1.
Expression and Purification of USP7
USP7 fragments coding for amino acids 54–205 and 62–205 were
expressed in E. coli from pET15b plasmids and purified as de-
scribed (Holowaty et al., 2003a). Expression of selenomethionine
(Se-Met)-containing USP7 54–205 was conducted in BL21(DE3)
Gold cells according to Doublie (1997) and purified as for native
protein. For NMR measurements, uniformly labeled15N,13C/15N,
and13C/15N/2H USP7 (amino acids 62–205) were produced in M9
media with15N ammonium chloride (0.8 g/l) and13C glucose (2 g/l)
as the sole nitrogen and/or carbon sources, respectively, and using
deuterated water (90%) for 2H-labeled samples. Labeled USP7 62–
205 for NMR was prepared as in Holowaty et al. (2003a) except that
buffers containing 20 mM sodium phosphate (pH 7.5), 250 mM
NaCl, 2 mM DTT, and 90% H2O/10%D2O were used. The concen-
tration of the purified proteins for NMR experiments ranged be-
tween 0.5 and 0.8 mM.
Generation of GST-EBNA1 Fusion Proteins
and Use in USP7 Binding Assays
Oligonucleotides encoding EBNA1 amino acids 421–435, 426–440,
431–445, and 436–450 were expressed as GST-fusions from GST-
2TK plasmid (Amersham) in BL21-CodonPlus E. coli (Stratagene)
at 37°C for 3 hr. Proteins were purified on glutathione-sepharose
(Amersham) by using standard methods then dialyzed against as-
say buffer (20 mM sodium phosphate [pH 7.5], 250 mM NaCl, 2 mM
DTT, 1 mM Benzamidine, and 0.5 mM PMSF). For USP7 binding
assays, purified USP7 NTD (amino acids 62–205) was incubated
with GST or GST-EBNA1 fusion proteins in the assay buffer in a 1:1
molar ratio at 4°C for 1 hr. The mixture was passed through a 0.2 ml
glutathione-sepharose column. After extensive washing with assay
buffer, bound proteins were eluted with 20 mM reduced glutathione
and detected by SDS-PAGE and Coomassie staining.
The EBNA1 peptide crystallized with USP7 consisted of amino
acids 441–450 (DPGEGPSTGP). This peptide was synthesized by
Dalton Chemicals (Toronto, Canada) with both amino terminal acet-
ylation and carboxy-terminal amidation to mimic the native pep-
tides. Peptides containing known TRAF binding sequences used in
USP7 binding assays were also synthesized by Dalton Chemicals.
Human p53 (355–393) and EBNA1 (410–450) peptides used in NMR
studies and the EBNA1 395–450 fragment (with or without point
Assays of USP7 Binding by Tryptophan Fluorescence
EBNA1 and TRAF binding peptides were titrated with the purified
USP7 NTD (62–205), and change in tryptophan fluorescence was
measured as described in Holowaty et al. (2003a).
NMR spectra were acquired at 30°C on a Varian Inova-500 and
600 MHz spectrometers equipped with pulsed field gradient triple-
Structural Basis for EBNA1 and p53 Binding to USP7
resonance probes or a Bruker 600 MHz spectrometer equipped
with a triple resonance cryo probe. The backbone1H,15N, and13C
resonances were assigned by using TROSY-HNCACB, TROSY-
CBCA(CO)NH, HNCACB, HNCO, and
(Salzmann et al., 1998; Kay, 2001). Greater than 80% of the back-
bone atoms were assigned. The interaction of USP7 with EBNA1
and p53 peptides was monitored by using the1H-15N HSQC experi-
ment at 30°C in a buffer containing 20 mM sodium phosphate (pH
7.5), 250 mM NaCl. Briefly, 0.7 mM15N-labeled USP7 (amino acids
62–205) was titrated with unlabeled p53 (residues 355–393) or
EBNA1 (residues 395–450, 410–450, or 441–450) peptides up to a
10:1 peptide:USP7 molar ratio. The data shown were collected by
using a 3:1 and a 5:1 molar ratio of EBNA1:USP7 and p53:USP7,
respectively, and no further changes in chemical shifts were de-
tected in the1H-15N HSQC spectra with higher peptide ratios.
mM MgCl2, 1.26 M potassium acetate, and 75% glycerol), cells
were further dounce homogenized and extracted on ice for 30 min.
The lysate was clarified by centrifugation, dialysed overnight
against 10 mM HEPES (pH 7.9), 150 mM NaCl then precleared by
incubation with 100 ?l bed volume of Protein A Sepharose for 15
min. USP7 was immunoprecipitated from 500 ?l (2.4 mg) lysate with
0 (negative control), 2, and 10 ?l BL851 USP7 antibody (Bethyl
Laboratories) and 20 ?l bed volume Protein A Sepharose. After 5
hr of mixing at 4°C, beads were washed three times in 10 mM
HEPES (pH 7.9), 150 mM NaCl and eluted with 40 ?l 1% SDS. 20
?l of lysate and eluates were analyzed by Western blot using
EBNA1 OT1X monoclonal antibody. The blots were then stripped
and reprobed with BL851 USP7 antibody.
The effect of EBNA1 expression on cell death was determined in
the U2OS osteosarcoma line by TUNEL assay upon induction of
p53 by UV irradiation. Cells were seeded at 50% confluency on 60
mm2dishes in duplicate 18 hr prior to transfection with 3 ?g of
pc3oriP plasmid expressing either EBNA1, EBNA1?395–450, or no
EBNA1 (Holowaty et al., 2003b) and with 250 ng of pDsRed1-N1
(BD Bioscience) as a transfection marker. 24 hr later, one set of
transfections was subjected to UV irradiation in a Stratagene 1800
ultraviolet crosslinker at 50 × 100 ?J/cm2. Cells were grown another
24 hr, then stained by using the In Situ Cell Death Detection Kit,
Fluorescein (Roche Applied Sciences) according to the manufac-
turer. Cells were filtered through a 0.7 ?m strainer cap (BD Biosci-
ence) prior to analysis on a Beckman-Coulter EPICS Elite (Flow
Cytometry Facility, University of Toronto).
The effect of EBNA1 on p53-mediated apoptosis was determined
in the H1299 p53 null osteosarcoma line by Annexin V staining.
Cells were seeded as above, 18 hr prior to transfection with 2 ?g
of pc3oriP plasmid expressing either EBNA1, EBNA1?395–450, or
no EBNA1, 1 ?g of pcDNA3p53 (Leng et al., 2003), and 500 ng of
pEGFP-C1 (BD Biosciences) as a transfection marker. 24 hr later,
cells were harvested and stained by using Annexin V-APC (BD Bio-
sciences) according to the manufacturer. Cells were fixed in 2%
paraformaldehyde/PBS overnight, washed in PBS, then filtered and
analyzed by flow cytometry as described above.
We thank the Oxford Protein Production Facility, University of Ox-
ford for performing initial crystal screens with USP7 and Dr. Dinesh
Christendat for help with X-ray data collection. We also thank Dr.
J. Lukin for assistance with NMR analysis, Cheryl Smith for FACS
analysis, Dr. Sam Benchimol for cell lines and antibodies, Dr. Jaap
Middeldorp for EBNA1 antibody, and Dr. Alan Davidson for use of
his spectrofluorometer. This work was funded by the Canadian
Cancer Society through grants to L.F. and C.H.A. from the National
Cancer Institute of Canada (NCIC). V.S. was supported by a Natural
Sciences and Engineering Council of Canada postdoctoral fellow-
ship and Y.S. was supported by a fellowship from the NCIC.
Received: August 26, 2004
Revised: November 23, 2004
Accepted: February 23, 2005
Published: March 31, 2005
Banks, L., Matlashewski, G., and Crawford, L. (1986). Isolation of
human-p53-specific monoclonal antibodies and their use in the
studies of human p53 expression. Eur. J. Biochem. 159, 529–534.
Chung, J.Y., Park, Y.C., Ye, H., and Wu, H. (2002). All TRAFs are
not created equal: common and distinct molecular mechanisms of
TRAF-mediated signal transduction. J. Cell Sci. 115, 679–688.
Cummins, J.M., Rago, C., Kohli, M., Kinzler, K.W., Lengauer, C., and
Vogelstein, B. (2004). Tumour suppression: disruption of HAUSP
gene stabilizes p53. Nature 428, 486–487.
Dolcetti, R., and Masucci, M.G. (2003). Epstein-Barr virus: induction
and control of cell transformation. J. Cell. Physiol. 196, 207–218.
Doublie, S. (1997). Preparation of selenomethionyl proteins for
phase determination. Methods Enzymol. 276, 523–530.
Everett, R., Meredith, M., Orr, A., Cross, A., Kathoria, M., and Par-
kinson, J. (1997). A novel ubiquitin-specific protease is dyamically
associted with the PML nuclear domain and binds to a herpesvirus
regulatory protein. EMBO J. 16, 1519–1530.
Holowaty, M.N., Sheng, Y., Nguyen, T., Arrowsmith, C., and Frap-
pier, L. (2003a). Protein interaction domains of the ubiqutin specific
protease, USP7/HAUSP. J. Biol. Chem. 278, 47753–47761.
Holowaty, M.N., Zeghouf, M., Wu, H., Tellam, J., Athanasopoulos,
V., Greenblatt, J., and Frappier, L. (2003b). Protein profiling with
Epstein-Barr nuclear antigen 1 reveals an interaction with the her-
pesvirus-associated ubiquitin-specific protease HAUSP/USP7. J.
Biol. Chem. 278, 29987–29994.
Hu, M., Li, P., Li, M., Li, W., Yao, T., Wu, J.-W., Gu, W., Cohen, R.E.,
and Shi, Y. (2002). Crystal structure of a UBP-family deubiquitinat-
ing enzyme in isolation and in complex with ubiquitin aldehyde.
Cell 111, 1041–1054.
Hume, S., Reisbach, G., Feederle, R., Delecluse, H.-J., Bousset,
K., Hammerschmidt, W., and Schepers, A. (2003). The EBV nuclear
antigen 1 (EBNA1) enhances B cell immortalization several thou-
sanfold. Proc. Natl. Acad. Sci. USA 100, 10989–10994.
Kay, L.E. (2001). Nuclear magnetic resonance methods for high mo-
lecular weight proteins: a study involving a complex of maltose
Effect of EBNA1 on p53 Levels
U2OS cells in 60 mm dishes were transfected with 2 ?g of pc3oriP
plasmid expressing either EBNA1, ?395–450, or no EBNA1 by using
Lipofectamine 2000 (Invitrogen). 5 hr later, cells were replated in 15
cm dishes and grown in the presence of 0.4 mg/ml G418 to select
for cells containing the pc3oriP plasmids. After 2 weeks, cells were
either harvested or UV irradiated (as described above) and har-
vested 4 or 8 hr post-UV treatment. Cells were lysed in 9 M urea, 5
mM Tris-HCl (pH 6.8), and sonicated. 30 ?g of total protein was
analyzed by Western blotting using antibodies PAb1801 for p53
(Banks et al., 1986), Ab-1 for actin (Oncogene research products)
and R4 serum for EBNA1 (Holowaty et al., 2003b). Blots were devel-
oped by using the ECL plus system (Amersham), and enhanced
chemifluorescence was quantified by using a Typhoon 9400 scan-
ner (Amersham) and ImageQuant 5.0 software. p53 levels were de-
termined in relationship to the actin loading control.
Assays of EBNA1-USP7 Interactions in B Cells
EBNA1 affinity column assays were conducted by using whole-cell
extracts from BL41 (at 10 mg/ml) and HeLa (at 14 mg/ml) cells,
which were generated as described in Holowaty et al. (2003b). 400
?l of each lysate was applied under physiological salt conditions
to a 40 ?l column containing EBNA1?61–83 (Wu et al., 2002) at 1
mg/ml and, after washing, bound proteins were eluted in high salt
and identified by MALDI-ToF mass spectrometry as previously de-
scribed (Holowaty et al., 2003b). For coimmunoprecipitation experi-
ments, 2.5 × 108Raji (EBV-positive Burkitt’s lymphoma) cells were
lysed by dounce homogenization in 10 mM HEPES (pH 7.9), 1.5
mM MgCl2, 10 mM KCl, 0.1% Triton X-1001, and Roche protease
inhibitors using 1.33 mls buffer per mg cell pellet. After addition of
an equal volume of extraction buffer (50 mM HEPES [pH 7.5], 1.5
Molecular Cell Download full-text
binding protein and beta-cyclodextrin. Methods Enzymol. 339,
Kennedy, G., Komano, J., and Sugden, B. (2003). Epstein-Barr virus
provides a survival factor to Burkitt’s lymphomas. Proc. Natl. Acad.
Sci. USA 100, 14269–14274.
Kieff, E., and Rickinson, A.B. (2001). Epstein-Barr virus and its repli-
cation. In Fields Virology, D.M. Knipe and P.M. Howley, eds. (Phila-
delphia: Lippincott Williams and Wilkins), pp. 2511–2573.
Kube, D., Vockerodt, M., Weber, O., Hell, K., Wolf, J., Haier, B.,
Grasser, F.A., Muller-Lantzsch, N., Kieff, E., Diehl, V., and Tesch, H.
(1999). Expression of Epstein-Barr virus nuclear antigen 1 is associ-
ated with enhanced expression of CD25 in the Hodgkin cell line
L428. J. Virol. 73, 1630–1636.
Leng, R.P., Lin, Y., Ma, W., Wu, H., Lemmers, B., Chung, S., Parant,
J.M., Lozano, G., Hakem, R., and Benchimol, S. (2003). Pirh2, a
p53-induced ubiquitin-protein ligase, promotes p53 degradation.
Cell 112, 779–791.
Li, C., Ni, C.Z., Havert, M.L., Cabezas, E., He, J., Kaiser, D., Reed,
J.C., Satteerthwait, A.C., Cheng, G., and Ely, K.R. (2002a). Down-
stream regulator TANK binds to the CD40 recognition site on
TRAF3. Structure 10, 403–411.
Li, C., Norris, P.S., Ni, C.Z., Havert, M.L., Chiong, E.M., Tran, B.R.,
Cabezas, E., Reed, J.C., Satterthwait, A.C., Ware, C.F., and Ely, K.R.
(2003). Structurally distinct recognition motifs in lymphotoxin-beta
receptor and CD40 for tumor necrosis factor receptor-associated
factor (TRAF)-mediated signaling. J. Biol. Chem. 278, 50523–50529.
Li, M., Chen, D., Shiloh, A., Luo, J., Nikolaev, A.Y., Qin, J., and Gu,
W. (2002b). Deubiquitination of p53 by HAUSP is an important path-
way for p53 stabilization. Nature 416, 648–653.
Li, M., Brooks, C.L., Kon, N., and Gu, W. (2004). A dynamic role of
HAUSP in the p53-Mdm2 pathway. Mol. Cell 13, 879–886.
McWhirter, S.M., Pullen, S.S., Holton, J.M., Crute, J.J., Kehry, M.R.,
and Alber, T. (1999). Crystallographic analysis of CD40 recognition
and signaling by human TRAF2. Proc. Natl. Acad. Sci. USA 96,
Ni, C.Z., Welsh, K., Leo, E., Chiou, C.K., Wu, H., Reed, J.C., and Ely,
K.R. (2000). Molecular basis for CD40 signaling mediated by
TRAF3. Proc. Natl. Acad. Sci. USA 97, 10395–10399.
Park, Y.C., Burkitt, V., Villa, A.R., Tong, L., and Wu, H. (1999). Struc-
tural basis for self-association and receptor recognition of human
TRAF2. Nature 398, 533–538.
Polekhina, G., House, C.M., Traficante, N., Mackay, J.P., Relaix, F.,
Sassoon, D.A., Parker, M.W., and Bowtell, D.D.L. (2002). Siah ubiqu-
itin ligase is structurally related to TRAF and modulates TNF-α signal-
ling. Nat. Struct. Biol. 9, 68–75.
Salzmann, M., Pervushin, K., Wider, G., Senn, H., and Wuthrich, K.
(1998). TROSY in triple-resonance experiments: new perspectives
for sequential NMR assignment of large proteins. Proc. Natl. Acad.
Sci. USA 95, 13585–13590.
Saridakis, V., Yakunin, A., Xu, X., Anandakumar, P., Pennycooke, M.,
Gu, J., Cheung, F., Lew, J.M., Sanishvili, R., Joachimiak, A., et al.
(2004). The structural basis for methylmalonic aciduria. The crystal
structure of archaeal ATP:cobalamin adenosyltransferase. J. Biol.
Chem. 279, 23646–23653.
Wajant, H., Henkler, F., and Scheurich, P. (2001). The TNF-receptor-
associated factor family: scaffold molecules for cytokine receptors,
kinases and their regulators. Cell. Signal. 13, 389–400.
Wilson, J.B., Bell, J.L., and Levine, A.J. (1996). Expression of Ep-
stein-Barr virus nuclear antigen-1 induces B cell neoplasia in trans-
genic mice. EMBO J. 15, 3117–3126.
Wu, H., Kapoor, P., and Frappier, L. (2002). Separation of the DNA
replication, segregation and transcriptional activation functions of
Epstein-Barr nuclear antigen 1. J. Virol. 76, 2480–2490.
Ye, H., Park, Y.C., Kreishman, M., Kieff, E., and Wu, H. (1999). The
structural basis for the recognition of diverse receptor sequences
by TRAF2. Mol. Cell 4, 321–330.
Ye, H., Arron, J.R., Lamothe, B., Cirilli, M., Kobayashi, T., Shevde,
N.K., Segal, D., Dzivenu, O.K., Vologodskaia, M., Yim, M., et al.
(2002). Distinct molecular mechanism for initiating TRAF6 signal-
ling. Nature 418, 443–446.
Zapata, J.M., Pawlowski, K., Haas, E., Ware, C.F., Godzik, A., and
Reed, J.C. (2001). A diverse family of proteins containing tumor
necrosis factor receptor-associated factor domains. J. Biol. Chem.
The atomic coordinates for USP7 NTD and USP7 NTD bound to
the EBNA1 peptide have been deposited in the Protein Data Bank
with the accession numbers 1YZE and 1YY6, respectively.