Molecular Cell, Vol. 17, 331–339, February 4, 2005, Copyright ©2005 by Elsevier Inc.DOI 10.1016/j.molcel.2005.01.008
The Deubiquitinating Enzyme USP1
Regulates the Fanconi Anemia Pathway
tion. MMC and DEB hypersensitivity is a hallmark of FA
and is used as a diagnostic test in the clinic (Auerbach
et al., 1989).
The genetic basis for FA is diverse, and evidence
exists for at least 11 complementation groups (Levitus
et al., 2004). FA is clinically related to various other
hereditary chromosomal instability syndromes, and re-
cent work has shown that the protein products mutated
in Bloom Syndrome, Nijmegen Breakage Syndrome
(NBS), Ataxia Telangiectasia (ATM), and Seckel Syn-
drome (ATR) functionally intersect with the FA signaling
pathway (Andreassen et al., 2004; Meetei et al., 2003b;
Nakanishi et al., 2002; Taniguchi et al., 2002b). Further-
more, hypomorphic mutations in the BRCA2 gene make
up the Fanconi complementation group D1 (FANCD1)
(Howlett et al., 2002).
Based on clinical, biochemical, and cellular pheno-
types, the FA proteins appear to function in a common
cellular signaling network. At least seven of these pro-
teins, FANCA, FANCB, FANCC, FANCE, FANCF, FANCG,
and the ubiquitin E3 ligase FANCL form a nuclear multi-
subunit complex that is critical for the monoubiquitina-
tion of the FANCD2 protein (de Winter et al., 2000; Gor-
don and Buchwald, 2003; Meetei et al., 2003a; Meetei
et al., 2004; Pace et al., 2002). Indeed, functional loss
of any of these FA proteins abrogates S phase and DNA
damage-induced FANCD2 ubiquitination.
izes to nuclear DNA damage foci, where it binds to
BRCA1 and the RAD51 recombinase and colocalizes
with FANCD1/BRCA2 (Taniguchi et al., 2002a; Wang et
al., 2004). It is thought that these nuclear foci mark the
sites of DNA damage-induced double-strand breaks
recombination. A role for FANCD2 and BRCA1 in ho-
mologous recombination is also suggested by their
presence at sites of meiotic recombination in spermato-
genesis (Garcia-Higuera et al., 2001). Although the
monoubiquitination of FANCD2 appears to be a critical
event in efficient DNA repair, the exact molecular func-
tion of FANCD2 is poorly understood.
As mentioned, FANCD2 is also monoubiquitinated
during S phase, and this event is required for normal
quitinated form of FANCD2 (FANCD2-L) disappears
when cells exit S phase and is transiently present in
cells that have been exposed to DNA damage (Garcia-
Higuera et al., 2001; Taniguchi et al., 2002a). Both forms
of FANCD2 are stable and not subject to proteasomal
degradation, indicating that the monoubiquitination
does not serve to target FANCD2-L for degradation.
Instead, it is more likely that a deubiquitinating enzyme
(DUB) removes the ubiquitin moiety after DNA damage
is repaired, and cells resume cycling. Like protein phos-
phorylation, ubiquitination is dynamic and reversible,
involving numerous ubiquitin-conjugating enzymes and
DUBs (Chung and Baek, 1999; D’Andrea and Pellman,
1998; Kim et al., 2003; Shackelford and Pagano, 2004;
Wilkinson, 2000). Homology searches in human genome
Sebastian M.B. Nijman,1,3Tony T. Huang,2,3
Annette M.G. Dirac,1,3Thijn R. Brummelkamp,1
Ron M. Kerkhoven,1Alan D. D’Andrea,2,*
and Rene ´ Bernards1,*
1Division of Molecular Carcinogenesis and Center
for Biomedical Genetics
The Netherlands Cancer Institute
1066 CX Amsterdam
2Department of Radiation Oncology
Dana-Farber Cancer Institute
Harvard Medical School
Boston, Massachusetts 02115
Protein ubiquitination and deubiquitination are dy-
namic processes implicated in the regulation of nu-
merous cellular pathways. Monoubiquitination of the
Fanconi anemia (FA) protein FANCD2 appears to be
critical in the repair of DNA damage because many of
the proteins that are mutated in FA are required for
FANCD2 ubiquitination. By screening a gene family
RNAi library, we identify the deubiquitinating enzyme
USP1 as a novel component of the Fanconi anemia
pathway. Inhibition of USP1 leads to hyperaccumula-
tion of monoubiquitinated FANCD2. Furthermore,
teins colocalize in chromatin after DNA damage. Fi-
nally, analysis of crosslinker-induced chromosomal
aberrations in USP1 knockdown cells suggests a role
in DNA repair. We propose that USP1 deubiquitinates
FANCD2 when cells exit S phase or recommence cy-
cling after a DNA damage insult and may play a critical
role in the FA pathway by recycling FANCD2.
tion against malignant transformation. Genetic disor-
ders that perturb the repair of DNA damage, induced
by either exogenous agents or endogenous events, of-
ten lead to increased cancer susceptibility. One such
disorder is Fanconi anemia (FA), a rare syndrome with
predisposition to a variety of malignancies (D’Andrea,
2003). Genes mutated in FA have also been implicated
in the carcinogenesis of sporadic tumors, underscoring
the broad relevance of studying rare human genetic
At the cellular level, FA is characterized by chromo-
somal instability and hypersensitivity to DNA-crosslink-
ing agents, such as mitomycin C (MMC), cisplatin, di-
poxybutane (DEB), and to alesser extent, ionizing radia-
3These authors contributed equally to this work.
coding for (putative) DUBs. Although DUBs have been
functionally linked with various pathways and pro-
cesses, surprisingly few mammalian DUB substrates
have been identified (Brummelkamp et al., 2003; Cum-
mings et al., 2004; Graner et al., 2004; Kovalenko et al.,
2003; Li et al., 2002; Trompouki et al., 2003). To study
the role of these enzymes in specific pathways, we have
constructed a library of 220 independent vectors ex-
DUBs. Using this library, we have previously identified
the familial tumor suppressor CYLD as a negative regu-
lator of TRAF2 poly-ubiquitination (Brummelkamp et al.,
2002b; Brummelkamp et al., 2003). Here, we identify the
deubiquitinating enzyme USP1 as a novel component
of the FA pathway and proposethat USP1 is the enzyme
that deubiquitinates FANCD2.
cells with the four shRNA vectors and an expression
vector containing a green fluorescent protein tagged
version of USP1 (GFP-USP1). As expected, all four
shRNA vectors efficiently suppressed GFP-USP1 ex-
pression (Figure 2C).
Tostudy endogenousUSP1 protein,a polyclonalanti-
serum directed against the N terminus of USP1 was
generated (see Experimental Procedures) and tested on
synthetic siRNA transfected or control HEK293 cells.
Cells were also treated with the S phase inhibitor hy-
droxyurea (HU) to induce monoubiquitinated FANCD2.
Two bands present in the control lanes were efficiently
downregulated in lysates derived from the USP1 siRNA-
transfected cells (Figure 2D, upper panel). As expected,
the observed USP1 downregulation correlated with the
upregulation of FANCD2-L (Figure 2D, lower panel). The
predicted molecular weight of endogenous, full-length
USP1 is 88 kDa, corresponding to the slower migrating
USP1 species detected by Western blot and consistent
with the size of ectopically expressed USP1 (see Figure
3D, lower panel). The faster migrating band is likely a
proteolytic fragment of USP1. We conclude that in our
experiments, both ectopically expressed and endoge-
nous USP1 protein are efficiently inhibited by RNA inter-
Because FANCD2 monoubiquitination is activated in
S phase, we investigated whether USP1 inhibition re-
sulted in an altered cell cycle distribution or S phase
delay. We retrovirally transduced U2-OS cells with a
knockdown vector targeting USP1. After selection with
puromycin, the cells were synchronized using a double-
thymidine block. Cells were released and samples for
FACS and protein analysis were taken at the indicated
time points. Propidium-iodide (PI) staining of nuclei and
subsequent FACS analysis indicated that cell cycle dis-
tribution and S phase progression of USP1 knockdown
cells was unaffected, suggesting that USP1 inhibition
does not activate a cell cycle checkpoint (Figure 2E).
Furthermore, although FANCD2-L levels decreased sig-
nificantly in the control cells about 4 hr after release,
this decrease appeared to be strongly delayed in the
USP1 knockdown cells (Figure 2F). USP1 inhibition did
lation was not an indirect effect of DNA damage.
was dependent on a functional FA core complex. We
transfected a FANCA-deficient (FA-A) cell line or a
knockdown vector. Subsequently, HU-stimulated or
-untreated cells were analyzed for FANCD2 ubiquitina-
tion (Figure 2G, upper panel). USP1 knockdown did not
result in FANCD2-L accumulation in the FA-A cell line.
cells, indicating that the ability of USP1 to affect
FANCD2 monoubiquitination is dependent on a func-
tional FA signaling pathway.
To investigate whether USP1 requires its protease
activity to affect FANCD2 ubiquitination, we generated
a catalytically inactive USP1 mutant in which the active
site cysteine is replaced by a serine residue (GFP-USP1
C/S) (Papa and Hochstrasser, 1993). Overexpression of
Identification of USP1 as a Regulator
of FANCD2 Monoubiquitination
Previous experiments have indicated that during normal
cell cycle progression, FANCD2 ubiquitination is dy-
namic (Taniguchi et al., 2002a). Therefore, we reasoned
that inhibition of a DUB that cleaves the ubiquitin moiety
from FANCD2 would lead to an overall increase of
FANCD2-L (the monoubiquitinated isoform of FANCD2)
in asynchronous cycling cells. To identify DUBs that
generated in our laboratory (Brummelkamp et al., 2003).
The library currently consists of 55 pools of 4 indepen-
dent shRNA-encoding plasmids targeting 55 DUBs for
suppression by RNA interference (Figure 1A and see
Supplemental Table S1 at http://www.molecule.org/cgi/
content/full/17/3/331/DC1/). We electroporated each
pool of DUB knockdown vectors separately in U2-OS
cells and selected for shRNA expression. After 72 hr,
we analyzed cell lysates by Western blot with an anti-
FANCD2 antibody. As shown in Figure 1B, the pool tar-
geting the ubiquitin-specific protease 1 (USP1, pool 47)
significantly increased the FANCD2-L fraction (Fujiwara
et al.,1998). Theincrease inFANCD2-L wascomparable
to the levels observed in MMC-treated cells (Figure 1B).
Further validation showed that only the pool targeting
USP1 reproducibly had this effect on FANCD2 (Figure
2A and data not shown).
Next, we tested the four independent USP1 shRNA
vectors (A–D) present in the original pool for their ability
to induce FANCD2-L accumulation (Figure 2A). Both
MMC-treated and -untreated cells displayed enhanced
FANCD2 monoubiquitination upon transfection of all
four vectors (A–D). However, vectors A and C were more
potent in inducing FANCD2-L than vectors B and D.
Retroviral delivery of an shRNA targeting USP1 by
using the pRetroSuper vector (pRS) also enhanced
Compared to control cells, retrovirally transduced USP1
knockdown cells displayed enhanced FANCD2 ubiquiti-
nation when stimulated with MMC or left untreated (Fig-
To verify that USP1 expression was indeed inhibited
by the knockdown vectors, we cotransfected HEK293
USP1 Regulates FANCD2
Figure 1. DUB Gene Family Screen
(A) Genome-wide chromosomal locations of
selected DUBs (adapted from ENSEMBL).
(B) U2-OS cells were electroporated with the
individual pools of the knockdown library and
selected for shRNA expression with puromy-
were prepared and immunoblotted using a
FANCD2 specific antibody. As a control for
FANCD2 ubiquitination, cells were treated
overnight with mitomycin C (50 ng/ml MMC).
The unmodified and monoubiquitinated forms
mulation of FANCD2-L (Figure 2H), likely by a dominant
negative effect. Possibly due to efficient proteasomal
degradation (see below), ectopic expression of wild-
type USP1 had a minimal effect on monoubiquitinated
FANCD2. We conclude that USP1 is a regulator of
nating enzyme activity to exert this function.
had disappeared in late S phase (8–10 hr after G1/S
release, Figure 3B). We conclude that USP1 levels are
cell cycle regulated and that USP1 protein is present
when FANCD2-L is deubiquitinated.
USP1 data mining in the SOURCE microarray data-
base indicated that USP1 mRNA is induced in S phase
and coclusters in terms of its cell cycle regulation with
other DNA repair proteins, suggesting a mechanism for
the observed USP1 cell cycle regulation (Whitfield et al.,
2002). The observation that USP1 is cell cycle regulated
could be confirmed with an independent microarray da-
taset (Figure 3C). As indicated, USP1 mRNA levels fol-
genes Rad51 and PCNA.
Whereas USP1 mRNA levels decline when cells exit
S phase, a marked decrease of USP1 protein levels is
already noticeable in late S phase, just prior to cyclin A
that USP1 protein may be degraded by the ubiquitin-
proteasome pathway. To explore this possibility, HEK293
cells were transfected with a HA-tagged murine USP1
(HA-mUsp1) expression construct and, 24 hr later, were
USP1 Is a Cell Cycle-Regulated and Proteasomally
Degraded Nuclear Protein
The USP1 protein sequence contains a putative nuclear
tion by fluorescence microscopy confirmed that USP1
is a nuclear protein (Figure 3A).
Because FANCD2 monoubiquitination is cell cycle
regulated, we investigated whether USP1 protein levels
might also be regulated. HeLa cells were synchronized
with a double thymidine block, and samples for protein
and FACS analysis were taken at the indicated time
points (Figure 3B). Whole-cell lysates were subsequently
immunoblotted for FANCD2, USP1, cyclin A, and ?-tubu-
lin (loading control). USP1 protein levels were high in
Figure 2. USP1 Is a Regulator of FANCD2 Monoubiquitination
(A) U2-OS cells were transfected with USP1 knockdown (USP1kd) vectors as indicated (A–D), puromycin selected, and treated 72 hr later with
MMC (50 ng/ml) overnight or left untreated. A pSUPER vector containing a hairpin targeting murine E2F3 served as a control. Whole cell
lysates were analyzed by immunoblotting with a FANCD2 specific antibody.
(B) U2-OS cells expressing the ecotropic receptor were transduced with pRetroSuper-USP1 or pRetroSuper-GFP. After selection with puromy-
cin, cells were stimulated with MMC (50 ng/ml) overnight or left untreated. Whole-cell lysates were analyzed by immunoblotting using a
(C) HEK293 cells were cotransfected with an expression plasmid containing GFP-USP1 and knockdown vectors as indicated (A–D). GFP
served as a transfection control (lower panel). Whole-cell extracts were immunoblotted using a GFP antibody.
(D) HEK293 cells were transfected with a synthetic siRNA targeting USP1 or a control siRNA targeting LacZ. 72 hr later, cells were treated
for 12 hr with hydroxy-urea (2 mM) or left untreated. Whole-cell lysates were prepared and analyzed with a FANCD2 specific antibody (lower
panel) or USP1 polyclonal anti-serum (upper panel, the asterisk indicates a nonspecific background band).
(E) U2-OS cells expressing the ecotropic receptor were transduced with pRetroSuper-USP1 or pRetroSuper-GFP. After selection with puromy-
cin, cells were synchronized and released, and PI-stained cells were analyzed by FACS. Indicated are the approximate average cell cycle
phase distributions (G1, S, and G2/M) as percentage.
(F) As in (E), whole-cell lysates were immunoblotted with a FANCD2 antibody.
(G) A FANC-A-deficient patient cell line (GM6914) and a complemented derivative were transfected with a synthetic siRNA targeting USP1 or
a control siRNA targeting LacZ. Cells were stimulated for 12 hr with hydroxyurea (2 mM HU) or left untreated. Whole-cell extracts were
immunoblotted for FANCD2 (upper panel) or FANCA (lower panel).
(H) HEK293 cells were transfected with GFP-tagged wild-type USP1 (GFP-USP1WT), a catalytically inactive USP1 mutant (GFP-USP1C/S), or
empty vector. Whole-cell lysates were prepared 72 hr after transfection and analyzed with a FANCD2 antibody.
After release, protein samples were taken at different
time points and analyzed. Ectopically expressed murine
Usp1 protein levels peaked between 3 and 9 hr after
release, coinciding with FANCD2 deubiquitination. At
9–12 hr after release, mUsp1 levels decreased, similar
to the results observed for endogenous USP1. This indi-
cates that in addition to transcriptional control, USP1
protein levels may also be regulated posttranslationally.
To investigate this further we transfected HEK293 cells
with GFP-USP1and treated the cellsovernight with pro-
teasome inhibitor. GFP-USP1 protein was stabilized
immunoprecipitation of Flag-tagged-USP1 (Flag-USP1)
from HA-ubiquitin-cotransfected cells indicated that
USP1 is poly-ubiquitinated (Figure 3F). Taken together,
these results indicate that USP1 levels are regulated in
the cell cycle by transcriptional and posttranslational
USP1 Regulates FANCD2
Figure 3. USP1 Is a Cell Cycle-Regulated Nuclear Protein
(A) U2-OS cells were transfected with GFP-tagged USP1, and 24 hr later, phase contrast and fluorescence (GFP) pictures were taken.
(B) HeLa cells were synchronized with a double-thymidine block and released. Samples were taken as indicated and analyzed by Western
blot and FACS (PI staining, the asterisk indicates a nonspecific background band).
(C) Serum-starved Rat1 cells were stimulated by adding back 10% fetal calf serum. Samples for FACS and microarray analysis were taken
at different time points. Indicated are the cell cycle phases and the2log relative mRNA levels of the cell cycle-regulated genes proliferating
cell nuclear antigen (PCNA), Rad51, and USP1 and the cell cycle-unregulated MIWI-like gene MILI.
(D) HEK293 cells were transfected with HA-tagged murine Usp1 (HA-mUsp1) and synchronized by double-thymidine block. After release, cell
extracts were prepared at the indicated time points and analyzed with a FANCD2 or HA antibody.
(E) HEK293 cells were transfected with GFP-USP1 and, 24 hr later, treated overnight with CBZ-LLL (10 uM). GFP served as transfection
control. Whole-cell lysates were immunoblotted with a GFP antibody.
(F) HEK293 cells were transfected as indicated and treated overnight with CBZ-LLL (10 uM). Cells were lysed, and Flag-tagged USP1 was
immunoprecipitated. Flag-USP1 present in the whole-cell lysates was detected with an anti-Flag antibody (lower panel). Ubiquitinated USP1
was visualized with an HA-specific antibody (upper panel).
USP1 Localizes in Chromatin and Associates
Monoubiquitination of FANCD2 is critical for its localiza-
tion to chromatin after DNA damage (Taniguchi et al.,
tional modification on FANCD2 functions as a targeting
signal, tethering it to nuclear DNA damage foci. Based
on this model one would expect that the enzyme that
deubiquitinates and thereby releases FANCD2 from
DNA is either constitutively localized to chromatin or
recruited to chromatin upon DNA damage. To study
the subnuclear localization of USP1, cell fractionation
Figure 4. USP1 Localizes to Chromatin and Binds to FANCD2
(A) Schematic overview of nuclear fractionation of cells. Briefly, cytoplasm and nucleoplasm were extracted by permeabilization with detergent,
the resulting nuclei were DNase I digested, and chromatin was extracted with ammonium sulfate.
(B) HeLa cells were treated overnight with MMC (160 ng/ml) or left untreated and subsequently fractionated as indicated. Supernatants (S)
and pellets (P) were subjected to Western blot analysis with the indicated antibodies. Lamin B was used as a nuclear matrix (P4) loading control.
(C) HEK293 cells were transfected as indicated. 24 hr after transfection, the cells were left untreated, treated with CBZ-LLL (10 uM), or a
combinationofCBZ-LLL(10 uM)andMMC(50ng/ml).The nextday,cellswerelysedinELB, andFlag-taggedproteinswereimmunoprecipitated
using a Flag antibody. Coimmunoprecipitated FANCD2 was visualized with a specific antibody (upper panel). The blot was subsequently
reprobed with a Flag antibody to visualize the amount of immunoprecipitated Flag-tagged proteins (middle panel). Whole-cell lysates (input)
were Western blotted for FANCD2 (lower panel).
experiments were performed (see schematic diagram,
Figure 4A) (Wang et al., 2004). As expected, FANCD2-L
was found primarily in chromatin fractions after MMC
treatment (Figure 4B, compare S2 with S4 fraction). In
of FANCD2 deubiquitination, endogenous USP1 protein
was also found predominantly in the chromatin fraction
in both MMC-treated and -untreated cells.
To investigate whether USP1 physically associates
with FANCD2, we performed immunoprecipitation ex-
periments. We transfected Flag-USP1 or control Flag-
tagged DUBs (i.e., USP8 and VDUI) in HEK293 cells and
Because Flag-USP1 is being continuously degraded by
the proteasome, the IP was done in the presence of a
proteasome inhibitor. Under these conditions, endoge-
nous FANCD2 was coimmunoprecipitated with USP1,
suggesting that they interact in vivo (Figure 4C, upper
panel). A FANCD2 interaction with the highly expressed
Flag-USP8 or with the lower-expressed VDUI could not
be detected, indicating that the observed USP1/FANCD2
interaction is specific (Figure 4C, middle and upper pan-
els). Similar results were obtained with GFP-tagged
USP1 and with a second, independent FANCD2 anti-
body (data not shown). Combination of proteasome in-
hibitor and MMC treatment did not significantly modu-
late the USP1/FANCD2 interaction. This suggests that
DNA damage does not enhance or disrupt the inter-
USP1 Inhibition Protects Cells from MMC-Induced
Next, we asked whether USP1 inhibition, which in-
sequences for DNA repair. To address this, we inhibited
USP1 in HEK293 cells by RNAi, induced DNA damage
by treating the cells with MMC (10 ng/ml), and analyzed
metaphase spreads for chromosomal breakage. Under
these conditions, many of the control siRNA (siLacZ)-
treated cells displayed at least one chromosomal aber-
ration (Table 1, column 4). Recent studies have sug-
gested a critical upstream role for ATR in activating
number of aberrations per cell and the number of cells
with at least one triradial chromosome (a FA hallmark)
wasincreased bysiRNA-mediatedinhibitionof thisDNA
damage checkpoint kinase (Table 1, columns 3–5) (An-
dreassen et al., 2004). In contrast, inhibition of USP1
appearedto providerelativeresistanceagainst thistype
of DNA damage-induced aberrations in three indepen-
dent experiments. Compared to control samples, the
number of chromosomal aberrations per cell induced
USP1 Regulates FANCD2
Table 1. USP1 Inhibition Protects Cells from MMC-Induced Chromosomal Aberrations
Cells with at Least
(n ? 50)
Fraction of Cells
with One Triradial
(Control ? 1)
Total Break Events
per Cell TransfectionTreatment
3.54 ? 0.21
19 ? 4
10 ? 2
28 ? 11
0.52 ? 0.02
1.42 ? 0.18
Table shows the total number of break events per cell and the number of cells with at least one triradial (displayed as an absolute number
and as a normalized fraction where the value of the lacZ control is arbitrarily set as 1). Indicated are the mean number values from three
independent experiments in which 50 metaphases were analyzed. Standard error values are shown. Cells were treated with 10 ng/ml MMC
or left untreated (see Experimental Procedures).
a2-sample chi-square test siUSP1 versus siLacZ p ? 0.3.
b2-sample chi-square test siUSP1 versus siLacZ p ? 0.0001.
by MMC treatment was inhibited approximately by 50%
when cells were treated with an siRNA targeting USP1
(Table 1, column 3). Also the number of cells displaying
at least one triradial chromosome was lower when com-
paring USP1 knockdown cells to the control cells (Table
1, columns 4–5). This suggests that under these condi-
tions, reduced levels of USP1 can protect cells from
crosslinker-induced DNA damage.
require massive redistribution of USP1. Furthermore,
low levels or absence of USP1 in DNA damage foci
during DNA repair processes would be in agreement
with a role for USP1 in inhibition and recycling of
FANCD2. The mechanisms regulating USP1 activity are
a current focus of study.
The suggestion that USP1 knockdown has a protec-
tive effect against chromosomal aberrations induced by
the DNA crosslinking agent MMC indicates that USP1
functions to inhibit or switch off FANCD2-L-mediated
DNArepair. Furthermore,itmaysuggest thatFANCD2-L
availability—under these conditions—is a rate-limiting
step in homologous recombination following high levels
of DNA damage. However, in this context, it should be
noted that the experimental conditions of MMC-break-
age assays in general may not necessarily reflect a fre-
quently occurring physiological situation. Therefore,
from an evolutionary perspective, USP1 expression lev-
els may not be optimal for dealing with this high level
of DNA damage. Instead, USP1 levels may be more
tuned to deal with low levels of DNA damage encoun-
tered during normal DNA replication. In addition, it
shouldbe notedthatthe constitutivepresence ofmono-
ubiquitinated FANCD2 may lead to inappropriate DNA
repair events not measured in our assays. For example,
the inability to deubiquitinate FANCD2 may lead to ho-
mologous recombination in the absence of DNA da-
mage, possibly leading to chromosomal instability.
Thus, the inability to turn off or reset the FA pathway
after the repair of specific DNA damage sites may have
overall-deleterious effects on genome integrity or cause
increased mutation frequency. Indeed, besides a defect
in DSB repair by homologous recombination (HR) re-
sulting in large deletions, FA cells display lower levels
of mutational repair compared to normal cells (Papado-
poulo et al., 1990; Sonoda et al., 2003). A functional link
is also suggested by the observation that chicken DT40
cells deficient for Rev3, the TLS-associated DNA poly-
merase, display increased levels of sister-chromatid ex-
change and chromosomal breaks (Sonoda et al., 2003).
Therefore, persistent FANCD2 monoubiquitination may
lead to enhanced error-prone repair, perhaps through
the heightened activity of translesion synthesis (TLS)
polymerases (Niedzwiedz et al., 2004). Therefore, USP1
may also function to limit mutagenesis by its regulated
Using a gene family RNAi library, we have identified the
deubiquitinating enzyme USP1 as a novel component
of the Fanconi anemia pathway. We have found that
FANCD2and protectscellsagainstcertain typesofDNA
damage. Furthermore, coimmunoprecipitation and co-
matin after DNA damage suggest a direct role for USP1
in deubiquitination of FANCD2. However, we cannot ex-
clude the possibility that USP1 has other substrates or
that the observed effects are indirect. In addition, since
DUBs have poor substrate specificity in vitro, it is diffi-
cult to obtain direct evidence for deubiquitination of
FANCD2 by USP1 (Mason et al., 2004).
Inhibition of USP1 has allowed us to uncouple DNA
damageorS phasearrestfromthe inductionofFANCD2
monoubiquitination. Because we did not observe a sig-
nificant defect in cell cycle progression of USP1 knock-
down cells, FANCD2 monoubiquitination does not ap-
pear to be sufficient for activation of a cell cycle
Because USP1 levels are relatively constant through-
out S phase, it is tempting to speculate on additional
mechanisms regulating USP1 activity. USP1 deubiquiti-
nating activity may be activated in late S phase by a
yet-undefined posttranslational event(s). The predicted
increase in USP1 activity in late S phase may account
for (1) the rapid conversion of FANCD2-L to FANCD2-S
and (2) the increased degradation of active USP1 by the
Although we cannot exclude the presence of USP1 in
FANCD2/BRCA DNA damage foci, we have not ob-
servedredistribution ofUSP1upontreatment withDNA-
crosslinking agents (data not shown). Possibly the dis-
solving of FANCD2-L foci by deubiquitination does not
For detailed information on the 15K cDNA microarrays used, visit
http://microarrays.nki.nl. Rat 1A cells were serum starved in low-
serum medium (DMEM, 0.01% FCS) and stimulated to reenter the
cell cycle by adding back 10% serum. Time points were taken for
RNA-extraction and FACS analysis. Total RNA was linear amplified
by using the CMF-T7-RNA amplification method. Amplified anti-
transcriptase (Invitrogen) to incorporate dUTP-Cy3 or dUTP-Cy5
(Amersham). The labeled nucleic acid molecules were dissolved in
a hybridization mixture containing 5?SSC, 25% formamide, and
poly-dA [Pharmacia], and tRNA [Roche]) and hybridized to the array
at 42?C for 18 hr. After subsequent washing and drying, the slides
were scanned in the Agilent DNA microarray scanner (G2505B).
Extended protocols can be found at http://microarrays.nki.nl/
download/protocols.html. Two-dye swap pairs of hybridizations
were performed as well as self-self experiments. The raw data were
normalized and2log transformed using the Central Microarray Facil-
ity database software (http://cmfdb.nki.nl).
turn off of error-prone DNA repair processes. In conclu-
sion, USP1 is a novel player in the DNA repair network
by limiting FANCD2 activity and may play a critical role
in the control of homologous recombination by the FA/
Materials, Antibodies, and Plasmids
The generation of the DUB knockdown library has been described
elsewhere (Brummelkamp et al., 2003). In short, four annealed sets
of oligonucleotides encoding short hairpin transcripts correspond-
ing to one DUB enzyme (see Table S1) were cloned individually into
USP1 (sequence C) was generated by ligating an EcoRI/XhoI-
digested pSUPER fragment in pMSCV as described (Brummelkamp
et al., 2002a). Human Flag- and GFP-tagged USP1 were generated
by PCR amplification of Image clone 6473568 containing full-length
USP1 and cloned into pEGFP (clontech) or a modified pcDNA3.1
plasmid containing a 5? sequence coding for the Flag epitope
(pVLAG). To generate human full-length. Flag-tagged VDUI, PCR-
in pVLAG together with a 3? fragment acquired from a human cDNA
library. Human Flag-tagged USP8 was cloned by ligating a BamHI-
digested fragment in pVLAG. Mouse Usp1 was cloned by RT-PCR
from NIH 3T3 cells and subsequently subcloned into an HA vector.
The synthetic siRNA oligonucleotide sequence targeting human
USP1 is 5?-TCGGCAATACTTGCTATCTTA-3?. Anti-FANCD2 (FI17),
GFP (FL), HA-tag (Y-11), and lamin B (M-20) antibodies were ac-
quired from Santa Cruz. Antibodies recognizing cyclin A (Ab-3) and
?-tubulin were purchased from Oncogene research products and
Calbiochem, respectively. Rabbit polyclonal Anti-USP1 antibody
was generated by using an aminoterminal epitope corresponding
to TDSQENEEKASEYRASEIC (Biosource). CBZ-LLL, mitomycin C,
propidium iodide, thymidine, and hydroxyurea were purchased
We thank Helen Pickersgill and members of our labs for critical
cal assistance and G. Draetta for kindly providing a human USP8
construct. We are grateful to P. Andreassen for chromatin fraction-
ation and L. Moreau for chromosome breakage analysis. This work
was supported by the Centre for Biomedical Genetics (CBG), the
Netherlands Organization for Scientific Researc, and National Insti-
tutes of Health (NIH) grants RO1HL52725, RO1DK43889, and
PO1HL54785 (A.D.D.). T.T.H. was supported by Ruth L. Kirschstein
NRSA and is a Blount fellow for Damon Runyon Cancer Research
Foundation. A.M.G.D. was supported by the Danish Cancer Society
and the Danish Medical Research Council.
Received: August 25, 2004
Revised: November 29, 2004
Accepted: January 6, 2005
Published: February 3, 2005
Cell Cultures, G1/S Cell Synchronization, Transfections,
and Chromatin Isolation
All cells were cultured in Dulbecco’s modified Eagle medium
(DMEM) supplemented with 10% fetal calf serum. High-efficiency
electroporation of U2-OS cells and FACS analysis was done as
described (Agami and Bernards, 2000). Double-thymidine cell cycle
block was done according to (Taniguchi et al., 2002a). SiRNA trans-
fections were done using Lipofectamine 2000 (Invitrogen) (Andreas-
sen et al., 2004). Chromatin fractions were isolated as described
(Wang et al., 2004).
Agami, R., and Bernards, R. (2000). Distinct initiation and mainte-
nance mechanisms cooperate to induce G1 cell cycle arrest in re-
sponse to DNA damage. Cell 102, 55–66.
Andreassen, P.R., D’Andrea, A.D., and Taniguchi, T. (2004). ATR
Genes Dev. 18, 1958–1963.
Auerbach, A.D., Rogatko, A., and Schroeder-Kurth, T.M. (1989). In-
ternational Fanconi Anemia Registry: relation of clinical symptoms
to diepoxybutane sensitivity. Blood 73, 391–396.
Brummelkamp, T.R., Bernards, R., and Agami, R. (2002a). Stable
suppression of tumorigenicity by virus-mediated RNA interference.
Cancer Cell 2, 243–247.
Immunoblotting and Immunoprecipitation
Western blots were performed with whole-cell extracts, separated
12% Bis-Tris gradient gels (Invitrogen) and transferred to polyvi-
nylidine difluoride membranes (Millipore). Western blots were
probed with the indicated antibodies. For immunoprecipitation,
HEK293 cells were transfected by calcium phosphate precipitation
with the indicated plasmids; 48–72 hr after transfection, cells were
lysed in ELB buffer (0.25 M NaCl, 0.1% NP-40, 50 mM HEPES [pH
protein complexes were immunoprecipitated with 2 ?g of the indi-
cated antibody conjugated to protein G sepharose beads.
Brummelkamp, T.R., Bernards, R., and Agami, R. (2002b). A system
for stable expression of short interfering RNAs in mammalian cells.
Science 296, 550–553.
Brummelkamp, T.R., Nijman, S.M., Dirac, A.M., and Bernards, R.
(2003). Loss of the cylindromatosis tumour suppressor inhibits apo-
ptosis by activating NF-kappaB. Nature 424, 797–801.
Chromosome Breakage Analysis
Core Facility of the Dana-Farber Cancer Institute (Andreassen et al.,
2004). For the analysis of HEK293 cells transfected with various
siRNAs, cells were plated into T25 flasks 24 hr after transfection.
Beginning the next day, cells were treated with 10 ng/ml MMC for
2 days. After treatment, cells were exposed to colcemid for 12 hr,
swollen with 75 mM KCl, and fixed with 3:1 methanol/acetic acid.
Slides were then stained with Wrights’ stain, and a minimum of 50
metaphases were scored per treatment sample and averaged from
three independent experiments using a blinded approach.
Chung, C.H., and Baek, S.H. (1999). Deubiquitinating enzymes: their
diversity and emerging roles. Biochem. Biophys. Res. Commun.
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.
D’Andrea, A., and Pellman, D. (1998). Deubiquitinating enzymes: a
new class of biological regulators. Crit. Rev. Biochem. Mol. Biol.
USP1 Regulates FANCD2
D’Andrea, A.D. (2003). The Fanconi road to cancer. Genes Dev.
encodes a deubiquitinating enzyme related to a product of the hu-
man tre-2 oncogene. Nature 366, 313–319.
Papadopoulo, D., Guillouf, C., Mohrenweiser, H., and Moustacchi,
E. (1990). Hypomutability in Fanconi anemia cells is associated with
increased deletion frequency at the HPRT locus. Proc. Natl. Acad.
Sci. USA 87, 8383–8387.
Shackelford, J., and Pagano, J.S. (2004). Tumor viruses and cell
signaling pathways: deubiquitination versus ubiquitination. Mol.
Cell. Biol. 24, 5089–5093.
Sonoda, E., Okada, T., Zhao, G.Y., Tateishi, S., Araki, K., Yamaizumi,
M., Yagi, T., Verkaik, N.S., van Gent, D.C., Takata, M., and Takeda,
S. (2003). Multiple roles of Rev3, the catalytic subunit of polzeta in
Taniguchi, T., Garcia-Higuera, I., Andreassen, P.R., Gregory, R.C.,
Grompe, M., and D’Andrea, A.D. (2002a). S-phase-specific interac-
tion of the Fanconi anemia protein, FANCD2, with BRCA1 and
RAD51. Blood 100, 2414–2420.
Taniguchi, T., Garcia-Higuera, I., Xu, B., Andreassen, P.R., Gregory,
R.C., Kim, S.T., Lane, W.S., Kastan, M.B., and D’Andrea, A.D.
(2002b). Convergence of the fanconi anemia and ataxia telangiecta-
sia signaling pathways. Cell 109, 459–472.
Taniguchi, T., Tischkowitz, M., Ameziane, N., Hodgson, S.V., Ma-
thew, C.G., Joenje, H., Mok, S.C., and D’Andrea, A.D. (2003). Disrup-
tion of the Fanconi anemia-BRCA pathway in cisplatin-sensitive
ovarian tumors. Nat. Med. 9, 568–574.
Tischkowitz, M., Ameziane, N., Waisfisz, Q., De Winter, J.P., Harris,
R., Taniguchi, T., D’Andrea, A., Hodgson, S.V., Mathew, C.G., and
Joenje, H. (2003). Bi-allelic silencing of the Fanconi anaemia gene
FANCF in acute myeloid leukaemia. Br. J. Haematol. 123, 469–471.
Trompouki, E., Hatzivassiliou, E., Tsichritzis, T., Farmer, H., Ash-
worth, A., and Mosialos, G. (2003). CYLD is a deubiquitinating en-
members. Nature 424, 793–796.
Wang, X., Andreassen, P.R., and D’Andrea, A.D. (2004). Functional
interaction of monoubiquitinated FANCD2 and BRCA2/FANCD1 in
chromatin. Mol. Cell. Biol. 24, 5850–5862.
Whitfield, M.L., Sherlock, G., Saldanha, A.J., Murray, J.I., Ball, C.A.,
Alexander, K.E., Matese, J.C., Perou, C.M., Hurt, M.M., Brown, P.O.,
and Botstein, D. (2002). Identification of genes periodically ex-
pressed in the human cell cycle and their expression in tumors. Mol.
Biol. Cell 13, 1977–2000.
of proteins for degradation by the proteasome. Semin. Cell Dev.
Biol. 11, 141–148.
de Winter, J.P., van der Weel, L., de Groot, J., Stone, S., Waisfisz,
(2000). The Fanconianemia protein FANCF formsa nuclear complex
with FANCA, FANCC and FANCG. Hum. Mol. Genet. 9, 2665–2674.
Fujiwara, T., Saito, A., Suzuki, M., Shinomiya, H., Suzuki, T., Taka-
hashi, E., Tanigami, A., Ichiyama, A., Chung, C.H., Nakamura, Y.,
and Tanaka, K. (1998). Identification and chromosomal assignment
of USP1, a novel gene encoding a human ubiquitin-specific prote-
ase. Genomics 54, 155–158.
Garcia-Higuera, I., Taniguchi, T., Ganesan, S., Meyn, M.S., Timmers,
C., Hejna, J., Grompe, M., and D’Andrea, A.D. (2001). Interaction of
the Fanconi anemia proteins and BRCA1 in a common pathway.
Mol. Cell 7, 249–262.
Gordon, S.M., and Buchwald, M. (2003). Fanconi anemia protein
complex: mapping protein interactions in the yeast 2- and 3-hybrid
systems. Blood 102, 136–141.
Graner, E., Tang, D., Rossi, S., Baron, A., Migita, T., Weinstein, L.J.,
Lechpammer, M., Huesken, D., Zimmermann, J., Signoretti, S., and
Loda, M. (2004). The isopeptidase USP2a regulates the stability of
fatty acid synthase in prostate cancer. Cancer Cell 5, 253–261.
Howlett, N.G., Taniguchi, T., Olson, S., Cox, B., Waisfisz, Q., De Die-
Smulders, C., Persky, N., Grompe, M., Joenje, H., Pals, G., et al.
(2002). Biallelic inactivation of BRCA2 in Fanconi anemia. Science
Kim, J.H., Park, K.C., Chung, S.S., Bang, O., and Chung, C.H. (2003).
Kovalenko, A., Chable-Bessia, C., Cantarella, G., Israel, A., Wallach,
D., and Courtois, G. (2003). The tumour suppressor CYLD negatively
regulates NF-kappaB signalling by deubiquitination. Nature 424,
Mathew, C.G., Hoatlin, M.E., Waisfisz, Q., Arwert, F., de Winter, J.P.,
and Joenje, H. (2004). Heterogeneity in Fanconi anemia: evidence
for 2 new genetic subtypes. Blood 103, 2498–2503.
Li, M., Chen, D., Shiloh, A., Luo, J., Nikolaev, A.Y., Qin, J., and
Gu, W. (2002). Deubiquitination of p53 by HAUSP is an important
pathway for p53 stabilization. Nature 416, 648–653.
Mason, D.E., Ek, J., Peters, E.C., and Harris, J.L. (2004). Substrate
profiling of deubiquitin hydrolases with a positional scanning library
and mass spectrometry. Biochemistry 43, 6535–6544.
Meetei, A.R., de Winter, J.P., Medhurst, A.L., Wallisch, M., Waisfisz,
Q., van de Vrugt, H.J., Oostra, A.B., Yan, Z., Ling, C., Bishop, C.E.,
et al. (2003a). A novel ubiquitin ligase is deficient in Fanconi anemia.
Nat. Genet. 35, 165–170.
Meetei, A.R., Levitus, M., Xue, Y., Medhurst, A.L., Zwaan, M., Ling,
C., Rooimans, M.A., Bier, P., Hoatlin, M., Pals, G., et al. (2004).
X-linked inheritance of Fanconi anemia complementation group B.
Nat. Genet. 36, 1219–1224.
Meetei, A.R., Sechi, S., Wallisch, M., Yang, D., Young, M.K., Joenje,
H., Hoatlin, M.E.,and Wang, W. (2003b). Amultiprotein nuclear com-
plex connects Fanconi anemia and Bloom syndrome. Mol. Cell. Biol.
Nakanishi, K., Taniguchi, T., Ranganathan, V., New, H.V., Moreau,
L.A., Stotsky, M., Mathew, C.G., Kastan, M.B., Weaver, D.T., and
D’Andrea, A.D. (2002). Interaction of FANCD2 and NBS1 in the DNA
damage response. Nat. Cell Biol. 4, 913–920.
Niedzwiedz, W., Mosedale, G., Johnson, M., Ong, C.Y., Pace, P.,
and Patel, K.J. (2004). The Fanconi anaemia gene FANCC promotes
homologous recombination and error-prone DNA repair. Mol. Cell
Pace, P., Johnson, M., Tan, W.M., Mosedale, G., Sng, C., Hoatlin,
M., de Winter, J., Joenje, H., Gergely, F., and Patel, K.J. (2002).
FANCE: the link between Fanconi anaemia complex assembly and
activity. EMBO J. 21, 3414–3423.
Papa, F.R., and Hochstrasser, M. (1993). The yeast DOA4 gene