The SWI
⁄
SNF protein BAF60b is ubiquitinated through a
signalling process involving Rac GTPase and the RING
finger protein Unkempt
Patrick Lore
`
s
1,2,
*, Orane Visvikis
1,2,
*, Rosa Luna
1,2
, Emmanuel Lemichez
3,4
and Ge
´
rard Gacon
1,2
1 Institut Cochin, Universite
´
Paris Descartes, CNRS (UMR8104), Paris, France
2 INSERM, U567, Paris, France
3 INSERM, U895, Centre Me
´
diterrane
´
en de Me
´
decine Mole
´

work
(Received 4 September 2009, revised 30
November 2009, accepted 8 January
2010)
doi:10.1111/j.1742-4658.2010.07575.x
The SWI ⁄ SNF chromatin remodelling complexes are important regulators
of transcription; they consist of large multisubunit assemblies containing
either Brm or Brg1 as the catalytic ATPase subunit and a variable subset of
approximately 10 Brg⁄ Brm-associated factors (BAF). Among these factors,
BAF60 proteins (BAF60a, BAF60b or BAF60c), which are found in most
complexes, are thought to bridge interactions between transcription factors
and SWI ⁄ SNF complexes. We report here on a Rac-dependent process
leading to BAF60b ubiquitination. Using two-hybrid cloning procedures,
we identified a mammalian RING finger protein homologous to drosophila
Unkempt as a new partner of the activated form of RacGTPases and dem-
onstrated that mammalian Unkempt specifically binds to BAF60b and pro-
motes its ubiquitination in a Rac1-dependent manner. Immunofluorescence
studies demonstrated that Unkempt is primarily localized in the cytoplasmic
compartment, but has the ability to shuttle between the nucleus and the
cytoplasm, suggesting that the Rac- and Unkempt-dependent process lead-
ing to BAF60b ubiquitination takes place in the nuclear compartment.
Ubiquitinated forms of BAF60b were found to accumulate upon treatment
with the proteasome inhibitor MG132, indicating that BAF60b ubiquitina-
tion is of the degradative type and could regulate the level of BAF60b in
SWI ⁄ SNF complexes. Our observations support the new idea of a direct
connection between Rac signalling and chromatin remodelling.
Structured digital abstract
l
MINT-7543606: Rac1 (uniprotkb:P63000) physically interacts (MI:0915) with Unkempt (uni-
protkb:

The SWI ⁄ SNF chromatin remodelling complexes are
evolutionary conserved multimeric enzymes using ATP
hydrolysis to mobilize nucleosomes and remodel
chromatin structure [1–4]; these complexes are large
multisubunit assemblies containing either Brm or Brg1
as the catalytic ATPase subunit and a variable subset
of approximately 10 Brg ⁄ Brm-associated factors
(BAF). Among the later, the 60 kDa subunit BAF60 is
found in most complexes; it can be represented by
BAF60a, BAF60b or BAF60c, which are encoded by
the smarcd1, smarcd2 and smarcd3 genes, respectively
[1]. BAF60 proteins have been shown to interact with
multiple transcription factors and are thought to
bridge interactions between these transcription factors
and SWI ⁄ SNF complexes, thereby allowing the recruit-
ment of SWI ⁄ SNF to target genes [5–9].
Biochemical purification and analysis of SWI ⁄ SNF
complexes have revealed that few to no free subunits are
present within the cells, suggesting that most, if not all,
subunits are assembled into the complexes [1]. Thus,
cells must co-ordinate the expression and degradation of
SWI ⁄ SNF subunits in order to maintain the stoichiome-
try of the complexes. Recent studies have demonstrated
the role of protein–protein interactions, ubiquitination
and proteasomal degradation in regulating SWI⁄ SNF
subunit levels [10,11]. However, the mechanisms leading
to ubiquitination of specific SWI ⁄ SNF subunits and
their regulation have not been molecularly defined.
Ubiquitination consists of the covalent attachment
to proteins of ubiquitin, a highly conserved 76 amino

In a two-hybrid screen for partners of activated
RacGTPase, we isolated a human cDNA sequence
with a partial ORF showing a strong homology with a
previously described drosophila protein, d-Unkempt
[17], and with Unkempt-like sequences from human
and mouse origin (accession UniProtKB ⁄ TrEMBL
Q9H9P5 and Q6RUT6, respectively). Northern blot
analysis revealed ubiquitous expression of a 4.4 kb
mRNA in mouse tissues (not shown). Iterative 5¢
RACE PCR amplification starting from mouse and
human RNA resulted in ORFs encoding two predicted
proteins of 678 and 680 amino acids, respectively, with
quasi-identical sequences (87% identity; 95% similar-
ity) and significant homology with the full-length dro-
sophila d-Unkempt (40% identity; 64% similarity)
(Fig. 1A). The novel human mRNA sequence that
encodes the 680 amino acid Unkempt-like protein has
been assigned the accession number AM944365 by the
EMBL nucleotide sequence database. Alignment of
drosophila Unkempt protein with mouse and human
Unkempt-like sequences revealed conserved zinc finger
motifs in the N-terminal part of the protein and a
RING finger at the C-terminal end (Fig. 1A). For
further studies, we constructed plasmids encoding
glutathione S-transferase (GST)- and hemagglutinin
(HA)-tagged human Unkempt C-terminal region
(UNK-C-ter) and murine full-length Unkempt (UNK-
fl), as well as mutated versions of these proteins, as
shown in Fig. 1B.
Interestingly, the putative interaction between acti-

Unkempt ubiquitination is stimulated by
activated Rac
As mentioned above (Fig. 1A, B), Unkempt contains
in its C-terminal region a conserved RING finger,
a motif that is typical of a large group of E3 ubiquitin
ligases known as RING E3s [14,18]; this prompted us
to investigate whether self-ubiquitination and ubiquitin
A
B
Fig. 1. Structure of Unkempt proteins. (A) Sequence alignment of human, mouse and drosophila Unkempt proteins. Triple identity is shown
in red, double identity in orange. Zinc finger and RING finger motifs are indicated by blue and green bars, respectively, the conserved C ⁄ H
being indicated by asterisks. The position of the C-terminal RING finger deletion in UNK-fl-DRF is indicated by a red arrowhead. (B) Sche-
matic representation of the Unkempt-derived proteins used in the present study. Zinc finger and RING finger motifs are in dark and light
grey, respectively. The C to A mutations of the RING finger in UNK-C-ter double mutant (UNK-C-ter-DM) are indicated.
P. Lore
`
s et al. BAF60b ubiquitination is controlled by Rac and Unkempt
FEBS Journal 277 (2010) 1453–1464 ª 2010 The Authors Journal compilation ª 2010 FEBS 1455
ligase activity were associated with mammalian
Unkempt.
HA-tagged versions of Unkempt and 6-His-tagged
ubiquitin were coexpressed in cultured cells and the
resulting 6-His-tagged ubiquitinated proteins were col-
lected on cobalt beads and analysed by western blot-
ting. Following this procedure, we were able to
demonstrate Unkempt ubiquitination in various cell
lines, including CHO, HEK 293 (not shown) and HeLa.
As shown in Fig. 3A, UNK-C-ter ubiquitination
was clearly detected in HeLa cells; interestingly, the
ubiquitination pattern was enhanced by activated

ases in total cellular lysates is shown (input). The GTPases were
extracted from lysates with GST-UNK-C-ter (WT or RING finger
double mutant) beads, or with GST beads as a control, and pull-
down proteins were revealed by anti-myc western blotting (pull-
down). The results are representative of three experiments.
A
B
Fig. 3. Ubiquitination of Unkempt is dependent on activated Rac.
Ubiquitination of Unkempt was assessed by transfecting HeLa cells
with a combination of expression plasmids encoding 6His-Ub,
myc-Rac1L61 ⁄ N17 and HA-UNK-C-ter or HA-UNK-fl as indicated.
Ubiquitinated proteins were extracted on cobalt beads and immu-
noblotted with anti-HA IgG. The expression of transfected proteins
was monitored on total protein extracts by immunoblotting using
the indicated antibodies. Where indicated, cells were treated with
the proteasome inhibitor MG132 for 4 h prior to lysis. (A) Ubiquiti-
nation of UNK-C-ter. WT, UNK-C-ter wild-type; DM, UNK-C-ter dou-
ble mutant C639A, C670A. (B) Ubiquitination of UNK-fl. WT, UNK-fl
wild-type; DRF, UNK-fl with RING finger deletion. The results are
representative of at least three experiments.
BAF60b ubiquitination is controlled by Rac and Unkempt P. Lore
`
s et al.
1456 FEBS Journal 277 (2010) 1453–1464 ª 2010 The Authors Journal compilation ª 2010 FEBS
RING finger proteins devoid of intrinsic ubiquitin
ligase activity have been found to form oligomeric
complexes, especially with other RING finger proteins,
resulting in active E3 ligases [19–22].
We therefore reasoned that Unkempt may partici-
pate in a protein complex showing an E3 ligase

A
B
Fig. 4. BAF60 proteins and their interactions with UNK-C-ter. (A) Sequence alignment of human BAF60a, b and c proteins used in the pres-
ent study. Triple identity is indicated in red, double identity in orange. The region of BAF60b involved in UNK-C-ter interaction, as mapped
from two-hybrid clones, is underlined (green bar). (B) UNK-C-ter interacts specifically with BAF60b. HeLa cells were transfected with expres-
sion plasmids encoding FLAG-tagged BAF60a, b or c. Proteins were extracted from lysates with GST-UNK-C-ter (WT or RING finger double
mutant) beads, or by GST beads as a control, and pull-down proteins were revealed by anti-FLAG western blotting. Identical results were
obtained in two independent experiments.
P. Lore
`
s et al. BAF60b ubiquitination is controlled by Rac and Unkempt
FEBS Journal 277 (2010) 1453–1464 ª 2010 The Authors Journal compilation ª 2010 FEBS 1457
RING finger motif of UNK-C-ter (C639A ⁄ C670A
double mutant) did not significantly alter the binding of
BAF60b (Fig. 4B).
Considering that BAF60b may be a partner of
Unkempt, we investigated BAF60b ubiquitination and
its regulation by Rac and Unkempt. Attempts to
detect the ubiquitinated fraction of endogenous
BAF60b were unsuccessful, possibly due to the low
sensitivity of available antibodies (not shown). There-
fore, a FLAG-tagged version of BAF60b was
expressed in HeLa cells, and the resulting ubiquitina-
tion of FLAG-BAF60b could be analysed.
As illustrated in Fig. 5A, ubiquitinated forms of
BAF60b were detected in HeLa cells in the absence of
ectopic expression of Unkempt; they were found to
accumulate upon treatment with the proteasome inhibi-
tor MG132, indicating that BAF60b ubiquitination is at
least partly of the degradative type. Interestingly, simi-

panel); of note, ubiquitination assays run in parallel
with BAF60a, b and c demonstrated that, in agreement
with interaction studies (see Fig. 4), BAF60b is the
preferred substrate of Unkempt-dependent ubiquitina-
tion. Similar to the pattern observed in the absence of
ectopic expression of Unkempt (Fig. 5A), ubiquiti-
nated forms of BAF60b generated in the presence of
UNK-C-ter strongly accumulated upon MG132 treat-
ment (Fig. 5C, right panel).
Next, we analysed the effects of Rac activation on
BAF60b ubiquitination. When coexpressed with the
dominant negative mutant Rac1N17 (i.e. in the
absence of activated Rac), BAF60b was poorly ubiqui-
tinated; by contrast, the ubiquitination profile was
strikingly enhanced by coexpression of Rac1L61
(Fig. 5D, lane 1 versus 4). Similarly, in the presence of
exogenous Unkempt, either UNK-C-ter or UNK-fl,
the amount of BAF60b ubiquitination appeared
strongly dependent on Rac activation (Fig. 5D, lane 2
versus 5 and lane 3 versus 6). Interestingly, the stimu-
lation of BAF60b ubiquitination could be replicated
by treating HeLa cells with CNF1, a toxin from uro-
pathogenic Escherichia coli known to strongly activate
endogenous Rac [23], thus confirming the results of
ectopic expression of activated Rac (not shown).
Although Unkempt seems to play a critical role in
BAF60b ubiquitination, it is noteworthy that muta-
ted ⁄ deleted forms UNK-C-ter and UNK DRF retained
full efficiency in BAF60b ubiquitination and Rac-
dependent regulation (Fig. S1), thus suggesting that

(STAT) transcription factors [25]. Finally, recent
studies have convincingly demonstrated a cell cycle-
dependent modulation of the amount of Rac1 in the
nucleus (i.e. accumulation in late G2 and exclusion in
early G1) [26]. These results prompted us to investigate
whether the binding of activated Rac might influence
the shuttling of Unkempt between cytosol and nucleus:
so far we have not been able to demonstrate any dif-
ferential effect of Rac1L61 or Rac1N17 on nuclear
accumulation of Unkempt (not shown).
However, taken together, these data support the idea
that Rac and Unkempt can translocate in the nuclear
compartment and activate BAF60b ubiquitination;
how these processes are co-ordinated remains to be
analysed.
Discussion
Although the results reported above are consistent
with BAF60b being ubiquitinated through a Rac- and
Unkempt-dependent process, the molecular composi-
tion of the E3 ligase involved and the role of Unkempt
RING finger remain uncertain. On the basis of the
results of a mutational analysis (Figs 3 and S1), it
appears that the RING finger of exogenously
expressed Unkempt is not critically involved in the
ubiquitination reaction. A possible explanation is that
exogenously expressed mutants of Unkempt form
dimers ⁄ oligomers with endogenous Unkempt and ⁄ or
associates with other RING finger protein(s), resulting
in active E3 ligase. As already mentioned, there are
multiple examples of RING E3s, the activity of which

involving an unidentified RING finger protein) and
whether and how RacGTP regulates this putative E3
ligase. To address these issues, in vitro studies aimed at
analysing intrinsic E3 ligase activity of recombinant
Unkempt will be required.
Our results also raise the questions of the physiolog-
ical relevance and significance of BAF60b ubiquitina-
tion. Unfortunately, using available antibodies to
BAF60b, we were not able to detect ubiquitinated
forms of endogenous BAF60b. However, in HeLa cells
expressing exogenous BAF60b, we found that BAF60b
is significantly ubiquitinated, even in the absence of
exogenous Unkempt; in addition, the ubiquitinated
forms of BAF60b strongly accumulated in the presence
of MG132, suggesting that the fate of ubiquitinated
BAF60b is proteasomal degradation. Thus, it may be
that ubiquitination results in degradation of an excess
of BAF60b subunits, thereby allowing the stoichiome-
try of SWI ⁄ SNF complexes to be maintained. Another
interesting possibility would be that BAF60b, alone or
in complex with Unkempt, interacts with other uniden-
A
B
Fig. 6. Subcellular localization of BAF60 and
Unkempt proteins. (A) Nuclear localization of
BAF60 proteins. HeLa cells were trans-
fected with FLAG-BAF60a, b or c, and FLAG
immunofluorescence was carried out on
fixed cells the following day. (B) Nucleocyto-
plasmic shuttling of Unkempt. HeLa cells

Several rounds of 5¢ RACE were performed (5¢RACE
System; Invitrogen, Carlsbad, CA, USA) to complete the
human and mouse cDNA sequences, using as templates
human and mouse polyA
+
-enriched fractions extracted from
peripheral blood leukocytes and kidney, respectively. cDNA
sequences deduced from sequencing overlapping 5¢ RACE
fragments were confirmed by resequencing both strands of
the complete cDNAs amplified by RT-PCR (Access RT-
PCR System; Promega, Madison, WI, USA) using 5¢- and
3¢-specific primers.
In the search for Unkempt-interacting proteins, a human
placental cDNA library was screened with UNK-C-ter as
the bait (Hybrigenics S.A., Paris, France); this resulted in
the isolation of four ‘high confidence’-independent clones
encoding overlapping regions of hSMARCD2 ⁄ BAF60b.
DNA plasmids and mutagenesis
Mouse and human Unkempt cDNA were inserted in N-ter-
HA pcDNA-3 (Invitrogen) and in pGEX-4T2 (Pharmacia,
Pfizer, New York, NY, USA) plasmid vectors.
A RING finger deletion mutant was generated from the
full-length sequence in pCDNA3 by NcoI digestion, Kle-
now extremities fill-in and re-ligation. Human cDNAs of
WTRac1 and WTCdc42, dominant negative mutant
Rac1N17 and constitutively activated forms Rac1L61,
Cdc42L61 and RhoAL63, cloned in pRK5-myc plasmids
were obtained from A. Hall (Memorial Sloan-Kettering
Cancer Center, New York, NY, USA); the pRBG4-6His-
myc-Ub plasmid has been used previously [31,32]. Expres-

CAAGTCCAAA and 5¢AAGATCACCTGTGCCTCCAC,
and normalized against endogenous glutamic acid decar-
boxylase mRNA levels, detected by RT-PCR with specific
primers 5¢GTCAGCCGCATCTTCTTTTG and 5¢GCAGA
GATGATGACCCTTT.
Cell culture, reagents and transfections
HeLa (ATCC reference CCL-2), HEK 293 (ATCC reference
CRL-1573) and CHO-K1 (ATCC reference CCL-61) cells
were grown in Dulbecco’s modified Eagle’s medium (Gibco-
BRL, Rockville, MD, USA) supplemented with 10% fetal
bovine serum (Gibco-BRL), 100 lgÆmL
)1
streptomycin,
100 lgÆmL
)1
penicillin and 250 ngÆmL
)1
fungizone (Gibco-
BRL), in a humidified atmosphere of 5% CO
2
at 37 ° C.
Where indicated, cells were treated with proteasome inhibitor
MG132 20 lm (Sigma) for 4 h. Cells were transiently trans-
fected using FuGENE 6 Transfection Reagent (Roche, Basel,
Switzerland) following the manufacturer’s instructions.
Antibodies
The primary antibodies used were M2 mouse monoclonal
antibody to FLAG
Ò
epitope (Sigma), 9E10 mouse mono-

2
, 150 mm NaCl, 1% Triton X-100, 0.5% NP40 and a
protease inhibitor cocktail (Amersham)]. In total, 500 lg
protein in 150 lL was incubated for 2 h at 4 °C with 15 lg
GST or GST-UNK-C-ter coupled to 20 lL glutathione-
sepharose beads (Amersham Biosciences). Pelleted beads
were washed twice with washing buffer (50 mm Tris ⁄ HCl
pH 7.5, 150 mm NaCl, 10 mm MgCl
2
,1mm dithiothreitol,
0.1% Triton X-100, 0.2 mgÆmL
)1
BSA and a protease
inhibitor cocktail). Bound proteins were recovered by
boiling beads in Laemmli sample buffer 2· (Sigma) and
analysed by western blotting.
Ubiquitination assay
HeLa, HEK 293 and CHO-K1 cells seeded in 100 mm Petri
dishes were transfected with a plasmid mix containing
1–3 lg each plasmid encoding 6His-myc-Ub and the indi-
cated proteins; in silencing experiments, HeLa cells were
transfected twice (with a 24 h interval) with plasmid mix
including pSUPER. Twenty hours after transfection, the
cells were washed in phosphate-buffered saline and lysed in
1 mL denaturating buffer (8 m urea, 20 mm Tris ⁄ HCl pH
7.5, 200 mm NaCl, 10 mm imidazole, 0.1% Triton X-100,
5mm N-ethylmaleimide, 10 mm iodoacetic acid); 50 lL
lysate was resuspended in Laemmli sample buffer 2X to
evaluate the total quantity of proteins. 6His-myc-ubiquiti-
nated proteins were recovered by incubating the remaining

References
1 Wang W, Cote J, Xue Y, Zhou S, Khavari PA, Biggar
SR, Muchardt C, Kalpana GV, Goff SP, Yaniv M
et al. (1996) Purification and biochemical heterogeneity
of the mammalian SWI-SNF complex. EMBO J 15,
5370–5382.
2 Olave IA, Reck-Peterson SL & Crabtree GR (2002)
Nuclear actin and actin-related proteins in chromatin
remodeling. Annu Rev Biochem 71, 755–781.
3 de la Serna IL, Ohkawa Y & Imbalzano AN (2006)
Chromatin remodelling in mammalian differentiation:
lessons from ATP-dependent remodellers. Nat Rev
Genet 7, 461–473.
4 Reyes JC, Muchardt C & Yaniv M (1997) Components
of the human SWI ⁄ SNF complex are enriched in active
chromatin and are associated with the nuclear matrix.
J Cell Biol 137, 263–274.
5 Hsiao PW, Fryer CJ, Trotter KW, Wang W & Archer
TK (2003) BAF60a mediates critical interactions
between nuclear receptors and the BRG1 chromatin-
remodeling complex for transactivation. Mol Cell Biol
23, 6210–6220.
6 Debril MB, Gelman L, Fayard E, Annicotte JS, Rocchi
S & Auwerx J (2004) Transcription factors and nuclear
receptors interact with the SWI ⁄ SNF complex through
the BAF60c subunit. J Biol Chem 279, 16677–16686.
7 Ito T, Yamauchi M, Nishina M, Yamamichi N,
Mizutani T, Ui M, Murakami M & Iba H (2001)
BAF60b ubiquitination is controlled by Rac and Unkempt P. Lore
`

VA, Orth A, Chanda SK, Batalov S & Joazeiro CA
(2008) Genome-wide and functional annotation of
human E3 ubiquitin ligases identifies MULAN, a
mitochondrial E3 that regulates the organelle’s dynam-
ics and signaling. PLoS ONE 3, e1487.
15 Pickart CM (2004) Back to the future with ubiquitin.
Cell 116, 181–190.
16 Kirkin V & Dikic I (2007) Role of ubiquitin- and
Ubl-binding proteins in cell signaling. Curr Opin Cell
Biol 19, 199–205.
17 Mohler J, Weiss N, Murli S, Mohammadi S, Vani K,
Vasilakis G, Song CH, Epstein A, Kuang T, English J
et al. (1992) The embryonically active gene, unkempt,
of Drosophila encodes a Cys3His finger protein. Genet-
ics 131, 377–388.
18 Joazeiro CA & Weissman AM (2000) RING finger
proteins: mediators of ubiquitin ligase activity. Cell 102,
549–552.
19 Buchwald G, van der Stoop P, Weichenrieder O,
Perrakis A, van Lohuizen M & Sixma TK (2006) Struc-
ture and E3-ligase activity of the Ring-Ring complex of
polycomb proteins Bmi1 and Ring1b. EMBO J 25,
2465–2474.
20 Poyurovsky MV, Priest C, Kentsis A, Borden KL, Pan
ZQ, Pavletich N & Prives C (2007) The Mdm2 RING
domain C-terminus is required for supramolecular assem-
bly and ubiquitin ligase activity. EMBO J 26, 90–101.
21 Xia Y, Pao GM, Chen HW, Verma IM & Hunter T
(2003) Enhancement of BRCA1 E3 ubiquitin ligase
activity through direct interaction with the BARD1

human cells. Cell 137, 459–471.
29 Corcoran CA, Montalbano J, Sun H, He Q, Huang Y
& Sheikh MS (2009) Identification and characterization
of two novel isoforms of Pirh2 ubiquitin ligase that
negatively regulate p53 independent of RING finger
domains. J Biol Chem 284, 21955–21970.
30 Toure A, Dorseuil O, Morin L, Timmons P, Jegou B,
Reibel L & Gacon G (1998) MgcRacGAP, a new human
GTPase-activating protein for Rac and Cdc42 similar to
Drosophila rotundRacGAP gene product, is expressed in
male germ cells. J Biol Chem 273, 6019–6023.
31 Doye A, Mettouchi A, Bossis G, Clement R, Buisson-
Touati C, Flatau G, Gagnoux L, Piechaczyk M,
Boquet P & Lemichez E (2002) CNF1 exploits the
ubiquitin-proteasome machinery to restrict Rho GTPase
activation for bacterial host cell invasion. Cell 111,
553–564.
32 Visvikis O, Lores P, Boyer L, Chardin P, Lemichez E &
Gacon G (2008) Activated Rac1, but not the tumori-
genic variant Rac1b, is ubiquitinated on Lys 147
through a JNK-regulated process. FEBS J 275,
386–396.
P. Lore
`
s et al. BAF60b ubiquitination is controlled by Rac and Unkempt
FEBS Journal 277 (2010) 1453–1464 ª 2010 The Authors Journal compilation ª 2010 FEBS 1463
Supporting information
The following supplementary material is available:
Fig. S1. Ubiquitination of BAF60b by RING mutants
of Unkempt.