Journal of Biomolecular Screening
The online version of this article can be found at:
2012 17: 1221 originally published online 1 August 2012J Biomol Screen
Emily C. Dykhuizen, Leigh C. Carmody, Nicola Tolliday, Gerald R. Crabtree and Michelle A. J. Palmer
Monitoring Transcriptional Regulation
Screening for Inhibitors of an Essential Chromatin Remodeler in Mouse Embryonic Stem Cells by
On behalf of:
Journal of Biomolecular Screening
can be found at:
Journal of Biomolecular Screening
Additional services and information for
What is This?
- Aug 1, 2012 OnlineFirst Version of Record
- Sep 27, 2012Version of Record >>
by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013jbx.sagepub.comjbx.sagepub.com jbx.sagepub.comjbx.sagepub.comjbx.sagepub.com jbx.sagepub.comjbx.sagepub.comjbx.sagepub.com jbx.sagepub.comjbx.sagepub.comjbx.sagepub.com Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from
Journal of Biomolecular Screening
17(9) 1221 –1230
© 2012 Society for Laboratory
Automation and Screening
An essential component of the embryonic stem (ES) cell
core pluripotency transcriptional network is the SWI/SNF-
like adenosine triphosphate (ATP)–dependent chromatin-
remodeling complex1; several of the subunits have a
genetically dominant role in ES cell development and for-
mation of the inner cell mass.2,3 In addition, the complex is
necessary and, in certain contexts, sufficient for the induc-
tion of pluripotency.4 Mammalian SWI/SNF complexes are
composed of a central ATPase (BRG1 or BRM) and 10 to
12 subunits (referred to as BAFs for BRG1 or BRM-
associated factors). The 11 members are encoded by 20
genes, which exhibit combinatorial assembly for an
astounding 288 predicted combinations. Indeed, ES cells
express a unique assembly of subunits, called esBAF,
which cannot be functionally rescued by the expression of
alternate, homologous subunits (Fig. 1). The combination
of BAF subunits is cell type specific, and subunit switching
is an important determinant of differentiation.5
In addition to the essential role of BAF in pluripotency
and development, several BAF subunits have been
al of Biomolecular Screening XX(X)Dykhuizen et al.
1Department of Pathology and the Howard Hughes Medical Institute,
Stanford University School of Medicine, Stanford, CA, USA
2Chemical Biology Platform, Broad Institute of Harvard and MIT,
Cambridge, MA, USA
Received Apr 2, 2012, and in revised form May 16, 2012. Accepted for
publication Jun 21, 2012.
Supplementary material for this article is available on the Journal of
Biomolecular Screening Web site at http://jbx.sagepub.com/supplemental.
Emily C. Dykhuizen, Department of Pathology, and the Howard Hughes
Medical Institute, Stanford University School of Medicine, Beckman
Center, 279 Campus Drive, Stanford, CA 94305, USA
Michelle A. J. Palmer, Chemical Biology Platform, Broad Institute of
Harvard and MIT, 7 Cambridge Center, Cambridge, MA 02142, USA
Screening for Inhibitors of an Essential
Chromatin Remodeler in Mouse Embryonic
Stem Cells by Monitoring Transcriptional
Emily C. Dykhuizen1, Leigh C. Carmody2, Nicola Tolliday2,
Gerald R. Crabtree1, and Michelle A. J. Palmer2
The SWI/SNF-like adenosine triphosphate (ATP)–dependent chromatin remodeling complex, esBAF, is both necessary
and, in some contexts, sufficient to induce the pluripotent state. Furthermore, mutations in various BAF subunits are
associated with cancer. Little is known regarding the precise mechanism(s) by which this complex exerts its activities.
Thus, it is unclear which protein interactions would be important to disrupt to isolate a relevant readout of mechanism. To
address this, we developed a gene expression–based assay to identify inhibitors of the native esBAF complex. Specifically,
a quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) assay was developed in mouse embryonic stem
(ES) cells to monitor expression of Bmi1, a developmentally important gene repressed by the esBAF complex. The assay
was miniaturized to a 384-well format and used to screen a diverse collection of compounds, including novel products of
diversity-oriented synthesis (DOS). Confirmed hits were validated using a knock-in ES cell reporter line in which luciferase
is inserted into the Bmi1 locus. Several of the validated hits regulate a panel of target genes in a manner similar to the BAF
chromatin-remodeling complex. Together these data indicate that expression-based screening using qRT-PCR is a successful
approach to identify compounds targeting the regulation of key developmental genes in ES cells.
esBAF, chromatin, qRT-PCR, expression-based screening, stem cells
Journal of Biomolecular Screening 17(9)
confirmed as tumor suppressors in humans. As of now,
malignancy-causing BAF subunit mutations include BAF47
(hSNF5, INI1) in nearly 100% of human malignant rhabdoid
tumors6; BAF250a (Arid1a) in ovarian clear cell carcinoma,
gastric cancer, and colorectal cancer7,8; BAF180 (poly-
bromo) in renal clear cell carcinoma9; BAF200 (Arid2) in
hepatocellular carcinoma10; and Brg1 mutations in medulo-
blastoma.11 Cancer-associated perturbations in expression
have been reported for Brm, BAF155, BAF60, BAF57,
BAF45d, and BAF53a, although, as of yet, they have not
been characterized as tumor suppressors.12 Considering the
evidence linking misregulation of the BAF complex to can-
cer initiation, progression, and therapeutic resistance, fur-
ther understanding of the mechanism of the BAF complex
is crucial for understanding the link between chromatin
remodeling and tumor biology.12
Although SWI/SNF complexes in Saccharomyces cere-
visiae exclusively activate genes, microarray studies in
mouse ES cells reveal a predominantly repressive role for
esBAF (approximately 70% of genes directly regulated by
esBAF are repressed). To further complicate matters,
esBAF predominantly represses targets at sites distal to the
promoters; only 12% of esBAF binding occurs at promot-
ers.1 It is unclear how the esBAF complex can be acting in
such a manner, and previous mechanistic studies have only
revealed small glimpses of the whole picture.13–15 Selective
and specific small-molecule probes of esBAF activity will
be invaluable tools with which to elucidate the mechanism(s)
by which this complex chromatin regulator functions; none
In vitro, the BAF complex displays DNA-stimulated
ATPase activity and can mobilize nucleosomes on a nucleo-
somal template. The ATPase subunit, BRG1 or BRM, is suf-
ficient for remodeling activity, and the addition of core
subunits, BAF47, BAF155, and BAF170, increases remodel-
ing activity to a level observed for the whole complex.16 The
subunits not required for ATP-dependent remodeling in vitro,
however, are essential for all of the activities of the complex
in vivo, indicating activities for individual subunits beyond
nucleosome remodeling. Previous evidence supported the
model that the BAF complex mobilized nucleosomes at pro-
moters to create open regions of chromatin for active tran-
scription.17 However, our evidence supports a more repressive
function for the esBAF complex, from a location distal to the
promoter. These data indicate that BAF complexes play more
complex roles in ES cell gene regulation than previously
thought and act via currently unknown mechanisms. We plan
to use small-molecule inhibitors to identify and order the
series of reactions catalyzed by the esBAF complex. We
hypothesize that the temporal control that small-molecule
inhibitors provide will be critical in deciphering the elusive
mechanism of the BAF complex.
No compound has yet been identified as an inhibitor of an
ATP-dependent chromatin-remodeling complex. A number of
approaches could be developed for high-throughput screening
(HTS) of small-molecule libraries to identify inhibitors of
Figure 1. The esBAF complex associated with chromatin. The specificity imparted by the particular composition of BAF subunits in
embryonic stem (ES) cells is in part through the combinatorial assembly of chromatin targeting motifs, including DNA-binding domains,
bromodomains, chromodomains, and plant homeodomains.
Dykhuizen et al.
esBAF activity, including both cell-based and biochemical
strategies. To maximize the physiological relevance of any
hits, we chose to develop a gene expression–based quantitative
reverse transcriptase polymerase chain reaction (qRT-PCR)
assay as a primary screen for compounds that inhibit esBAF-
mediated repression in ES cells. We used the Ambion Cells-
to-Ct kit (Life Technologies, Carlsbad, CA) in a 384-well
format to multiplex the expression of Bmi1, a developmentally
important BAF-repressed gene, with actin in mouse ES cells.
We validated confirmed hits using a mouse ES cell knock-in
luciferase reporter line for Bmi1 expression. Last, we used
qRT-PCR to test the compounds’ ability to regulate the expres-
sion level of a panel of esBAF target genes. By screening a
library of ~30 000 small molecules, including both novel and
pharmacologically active compounds, we identified 20 com-
pounds that transcriptionally mimic the esBAF knockout.
Studies are under way to determine whether the compounds
act directly on the BAF complex or if they inhibit important
transcriptional regulators that act in concert with the BAF
complex. Either possibility will lead to deeper understanding
of the actions of this chromatin-remodeling complex, which
plays an essential role in pluripotency, human tumor suppres-
sion, and cellular senescence.
Materials and Methods
Culture of Mouse ES Cells for qRT-PCR
The feeder-free mouse ES cell line, E14, was maintained on
gelatin-pretreated tissue culture plates in ES media consist-
ing of high-glucose Dulbecco’s modified Eagle’s medium
(DMEM; Invitrogen, Carlsbad, CA) supplemented with
15% fetal bovine serum (FBS; ES cell qualified; Applied
Stem Cell, Menlo Park, CA), 100 µM 2-mercaptoethanol
(Invitrogen), 1% minimum essential medium (MEM) nones-
sential amino acids (Invitrogen), 1 mM Hepes (Invitrogen),
100 U/mL penicillin/streptomycin (Invitrogen), 2 mM
GlutaMAX (Invitrogen), 1 mM sodium pyruvate (Invitrogen),
and 1 U/mL LIF (Millipore, Billerica, MA).
Development and Culture of Bmi-luc
Reporter ES Cell Line
The Bmi-luc reporter line was developed as published with
modifications for incorporation of the firefly luciferase
gene.18 Briefly, ES cells derived from TC1 mice were elec-
troporated with an 8-kb construct consisting of homology
arms flanking the ATG transcription start site of Bmi1,
where the firefly luciferase gene and neomycin were
inserted. After selection, knock-in lines were confirmed
using Southern blot analysis, and neomycin was removed
by transfection with Cre recombinase. Renilla luciferase
was inserted into the genome using a lentiviral vector and
selection with blastomycin, followed by clonal isolation
and propagation to create the final Bmi-luc ES cell reporter
line. Bmi-luc ES cells were maintained on a layer of irradi-
ated mouse embryonic fibroblasts (MEFs) and were plated
on gelatin-coated 384-well plates for screening.
Compound Library and Screening
White opaque tissue culture–treated 384-well plates
(Corning, Corning, NY) were pretreated for 30 min with
0.1% gelatin in PBS (Millipore). The gelatin was removed
by aspiration, and 5000 ES cells were plated per well in in a
total volume of 50 µL ES media. After plating, the ES cells
were incubated for 24 h at 37 °C and 5% CO2. Test com-
pounds (100 nL) were added to ~10 µM final concentration
by pin transfer; each compound was tested in duplicate. The
compound library comprised ~30 000 compounds from
diverse sources, including bioactive compounds (including
Food and Drug Administration [FDA]–approved drugs),
commercially available drug-like molecules, targeted collec-
tions (biased for kinases, chromatin modifiers, etc.), stereo-
chemically diverse compounds, and purified natural products.
After compound addition, cells were incubated for 18 h at
37 °C and 5% CO2. The media were aspirated and cells were
washed with 100 µL phosphate-buffered saline (PBS). Cells-
to-Ct lysis buffer containing DNaseI (10 µL) was added, and
plates were agitated for 5 min at room temperature. Cells-
to-Ct stop solution was added and plates were incubated for
2 min at room temp (19–25 °C). Meanwhile, Cells-to-Ct
reverse transcriptase master mix (8 µL) was dispensed into
new PCR plates. Cell lysate (2 µL) was transferred to the
PCR plate containing RT master mix, and the plates were
incubated at 37 °C for 60 min and then at 95 °C for 5 min in
a PCR block. qPCR master mix (4 µL), consisting of Roche
Taqman master mix, 40× FAM Bmi1 Taqman probe, and 40×
VIC actin Taqman probe (Applied Biosystems, Life
Technologies, Carlsbad, CA), was dispensed into 384-well
qPCR plates. cDNA from the RT reaction (1 µL) was dis-
pensed into the qPCR plate alongside cDNA isolated from
Brg1-depleted ESE14 cells (to serve as a positive control for
elevation of Bmi1 transcript levels). qPCR was performed on
the Roche Lightcycler (Roche Applied Sciences, Indianapolis,
IN) with a 10-min incubation at 95 °C to activate the
enzyme, followed by 40 cycles of 1 min at 60 °C and 15 s at
95 °C. Calculations are based on the 2–ΔΔCT calculations
described by Livak and Schmittgen,19 where –ΔΔCT =
(CT,Bmi1 – CT, Actin)Treated – (CT,Bmi1 – CT, Actin)DMSO control. The
“–ΔΔCT” values were calculated on a per plate basis using 32
neutral control (DMSO-treated) wells and the derived CP
values obtained from the Roche Lightcycler, which are akin
to CT values. –ΔΔCT was calculated for all compounds, and
the –1.33 value for –ΔΔCT was established as a cutoff for hit
calling and cherry-picking purposes (this value corresponds
to a ~2.5-fold increase in Bmi1 expression). For an in-depth
Journal of Biomolecular Screening 17(9)
protocol for implementation of high-throughput RT-PCR for
small-molecule screening assays, including data analysis, see
Phelan et al.16 and Bittker.20
Analysis of Bmi1 Induction Using Luciferase
Reporter ES Cell Line
Bmi1 luciferase knock-in ES cells were plated in 384-well
white opaque tissue culture plates coated with gelatin at a
density of 5000 cells per well in 30 µL media. Cells were
grown for 24 h at 37 °C and 5% CO2 and pinned with 100
nL test compound. The compound-treated cells were grown
for an additional 24 h at 37 °C and 5% CO2 and then
equilibrated to room temperature for 1 h. Dual-Glo
(Promega, Madison, WI) luciferase reagent (25 µL) was
added and plates were incubated for 45 min at room tem-
perature. Firefly luciferase levels were read on an EnVision
plate reader (PerkinElmer, Waltham, MA), and 25 µL Dual-
Glo renilla luciferase reagent was added. Plates were incu-
bated for 15 min at room temperature, and renilla luciferase
levels were read. Measurements were calculated as a ratio
of firefly and renilla luciferase levels.
Additional Analysis of Brg1 Targets by qRT-PCR
cDNA (1 µL) obtained by RT reactions from lysates of ES cells
treated with hit compounds (used at the concentration that
gave the maximal Bmi1 induction during the concentration-
response confirmation studies) was screened using SYBR
Green for Bmi1, as well as additional BRG1 targets, Phox2b,
Fgf4, Bmp4, Socs3, Cbx7, Ring1a, and Eed. Gapdh was used
as the housekeeping gene. Melting curves and sequencing
were performed for each set of primers to confirm specificity.
List of SYBR Green qPCR Primers
Brg1: Forward: CGGTTGTGAGTGACGATGAC;
Bmi1: Forward: TACCATGAATGGAACCAGCA;
Phox2b: Forward: CAGGGACCAGAGCAGT; reverse:
Actin: Forward: TTGCTGACAGGATGCAGAAG;
Gapdh: Forward: TGCACCACCAACTGCTTAG;
FGF4: Forward: GGGTGTGGTGAGCATCTTCGGA;
Cbx7: Forward: TCTCAGGGCAGTCCTTGTCT;
Ring1a: Forward: CCTGGACATGCTGAAGAACA;
Eed: Forward: CTGGCAAAATGGAGGATGAT;
Bmp4: Forward: ACAATGTGACACGGTGGGA-
AAC; reverse: TGTGGGTGATGCTTGGGACTAC
Socs3: Forward: ATTTCGCTTCGGGACTAGC;
shBrg1 construct: Open Biosystems PLKO.1 vector:
TRCN0000071386; hairpin sequence: CCGGCG-
CAATAATGTGTCGGGCGTTTTTG targeted to
Selection of a Reporter Gene for esBAF Activity
Using ChIP-seq and microarray data from a conditional
Brg1 knockout ES cell line, our lab has determined that
the esBAF complex is enriched both at highly expressed
ES cell-specific genes and at repressed developmental or
lineage-specific genes.1,21 A major group of genes directly
repressed by esBAF are members of the Polycomb repres-
sive complex, including Bmi1, Cbx7, Eed, Ring1a, Phc1,
Phc2, and Suz12. Of these, we chose to further validate
Bmi1 as a reporter of esBAF activity. Bmi1 is a particularly
interesting target due to its essential role in the maintenance
and self-renewal of hematopoietic and neural stem cells, as
well as its role as an oncogene.22 Thus, although our pri-
mary focus is to identify inhibitors of the BAF complex,
compounds that regulate the expression of Bmi1 indepen-
dent of the BAF complex will also be of interest and will be
pursued separately. Using qRT-PCR, we observed a 10-fold
increase of Bmi1 expression upon tamoxifen-induced Cre-
mediated Brg1 deletion from a conditional knockout ES
cells line, Brg1f/f, actin-CreER (Fig. 2A).21 There was no
delay between the decrease in Brg1 and the subsequent
increase in Bmi1, providing additional evidence that Bmi1
is a direct target of the esBAF complex (Fig. 2B). We con-
firmed the regulation of Bmi1 transcription in ES cells by
the BAF complex by testing a short-hairpin RNA (shRNA)
against Brg1 in E14 ES cells, a commonly used feeder-free
ES cell line (Fig. 2C). Knockdown of Brg1 for 72 h dis-
plays a similar robust increase in the transcript of Bmi1 as
the Brg1 knockout (KO) cell line (Fig. 2D). Taken together,
these data validate Bmi1 as a reporter for esBAF activity.
Development and Validation of a Bmi1
We developed a robust qRT-PCR expression-based assay to
measure Bmi1 transcript levels in mouse ES cells as a
reporter of esBAF-mediated repression. Using mouse E14
cells and Ambion’s Cells-to-Ct system with Applied
Dykhuizen et al.
Biosystem’s Taqman probes, we optimized individual assay
parameters (cell number, incubation period prior to adding
compound, incubation period with compound, qPCR mas-
ter mix type, housekeeping gene selection, and Taqman
probe concentration) for successful assay execution in a
384-well format. Since esBAF is necessary for ES cell
viability and pluripotency, esBAF inhibitors may cause a
decrease in cell number that occludes any increase in Bmi1
expression. Therefore, to correct for cell number and com-
pound toxicity, we felt it important to multiplex Bmi1 and
actin expression using compatible Taqman probes. We
screened a number of housekeeping genes with cDNA from
Brg1f/f and actin-CreER ES cells, including GAPDH,
actin, and HSP90. We found that all the housekeeping
genes tested gave very similar results (Suppl. Fig. S1). Due
to the commercial availability of large volumes of actin-
VIC Taqman probes, we chose to perform the screen with
actin as the housekeeping gene. To validate the assay as
“HTS-ready,” we performed a pilot screen of 2200 com-
pounds, comprising compounds with known bioactivity
(including FDA-approved drugs), commercially available
drug-like compounds, purified natural products, and novel
diversity-oriented synthesis (DOS) compounds targeted to
chromatin regulators. Wells were treated in duplicate for 18
h. The neutral control was DMSO alone, and the positive
control was cDNA isolated from E14 ES cells treated with
shRNA against Brg1 (the ATPase subunit of esBAF) for 3
days. The coefficient of variation (CV) values for biologi-
cal replicates of the CT values of Bmi1 expression were
0.6% to 2.4%, and the CV values for actin were 1% to
3.6%. In addition, the 0.9 Z factor was making it an
extremely robust assay. Hits were defined as compounds in
which both replicates produced a 50% induction of Bmi1
levels compared with Brg1 knockdown. The induction of
Bmi1 was calculated as the change in CT value for Bmi1
over the change in CT value for actin. We validated that
these primers have between 97% and 100% amplification
efficiency and are sensitive down to very low transcript
levels, making this simplified calculation accurate. In fact,
we found that ΔΔCT provides an accurate normalization of
Bmi1 transcript levels for cell numbers ranging from 5000
to 150 cells plated per well (Fig. 3). Twelve hits were iden-
tified in the pilot screen, giving a hit rate of 0.75%.
High-Throughput qRT-PCR Screen and
Validation of Hits
Having demonstrated that the qRT-PCR assay was robust
and capable of detecting hits, we initiated a screen of an
additional 28 000 compounds (see Suppl. Fig. S2 for a
Figure 2. Depletion of BRG1 results in increased Bmi1 expression. (A) BRG1 protein expression in embryonic stem (ES) cells
homozygous for a floxed Brg1 allele with an actin-CreER transgene after 72 h of EtOH or tamoxifen treatment. (B) Quantitative reverse
transcriptase polymerase chain reaction (qRT-PCR) of Brg1 and Bmi1 mRNA expression following tamoxifen treatment in the Brg1f/f-
actin-CreER ES cell line. Fold change is calculated using Gapdh as an internal control. (C) BRG1 protein expression in E14 ES cells 72 h
after treatment with an empty lentiviral construct or a lentiviral construct containing shBrg1. (D) Induction of Bmi1 transcript levels in
E14 ES cells 72 h after treatment with shBrg1.
Journal of Biomolecular Screening 17(9)
depiction of the workflow). E14 ES cells were plated and
treated in 384-well plates as described above for the pilot
screen. Excellent screening statistics were observed: CVs
were comparable to those observed in the pilot screen, and
excellent reproducibility was observed between replicates.
From this larger screen, using a hit threshold of a 2.5-fold
increase or greater in Bmi1 expression, 98 compounds were
identified as hits for a final hit rate of 0.33% (Fig. 4A).
Eighty-two hits were available for “cherry picking,” and
these compounds were retested across a concentration-
response curve using the primary qRT-PCR assay. From
these 82 retested hits, 37 compounds produced a dose-
dependent response in the qRT-PCR assay (a representative
concentration-response curve is shown in Fig. 4B). The
remaining compounds were eliminated due to a lack of
dose-dependent response, observed toxicity at high concen-
trations, or nonactive compounds.
As an orthogonal approach to assessing esBAF activity, we
developed a mouse ES cell line with a luciferase reporter
knocked into the Bmi1 locus.18 We found a reproducible induc-
tion of luciferase upon RNAi-based depletion of Brg1, when
corrected for cell number (Fig. 5). Since the esBAF complex is
necessary for both pluripotency and self-renewal in ES cells,
depletion of Brg1 by RNAi causes a decrease in cell prolifera-
tion. Similarly, inhibitors of esBAF would likely also cause a
reduction in cell viability that would confound the ability to
monitor an increase in Bmi1 expression. Therefore, we desired
a way to correct for cell number. We created an internal control
in the knock-in cell line by infecting it with a lentivirus con-
taining the renilla luciferase gene under the EF-1a promoter.
Luciferase from Renilla reniformis uses a different substrate
than firefly luciferase, allowing for multiplexing of readouts
using commercially available kits. We optimized and validated
the ES reporter cell line using Promega’s Dual-Glo luciferase
reagent and tested all 82 “cherry-picked” hits. Forty-three
compounds showed dose-dependent behavior in the secondary
luciferase reporter screen (a representative concentration-
response curve is shown in Fig. 4C). Comparison of the per-
formance of hits in both the primary qRT-PCR assay and the
secondary luciferase reporter assay identified 34 compounds
that showed dose-dependent responses in both assays.
Compounds determined to be hits in the qPCR screen but not
the luciferase screen were not followed up on due to the pos-
sibility that they may only appear to be hits by downregulating
actin transcript levels instead of upregulating Bmi1 levels.
Many of the hits show a decrease in actin levels at high con-
centrations due to a loss of cell viability (as we would expect
for esBAF inhibitors) that would be indistinguishable from a
compound that interferes with actin transcription. Thus, the
luciferase secondary screen was important for eliminating
these potential artifacts. The high overlap between the two
complementary screens (80%–90%) validates our approach
for identifying compounds that upregulate Bmi1 expression in
In addition to confirming a compound’s ability to induce
Bmi1 transcription, we used the cDNA obtained during the
rescreen of hits to perform qRT-PCR for additional devel-
opmentally important esBAF targets identified from micro-
array and ChIP-seq experiments. The ability to resample the
cDNA samples obtained during the rescreen of hits for mul-
tiple relevant targets is an advantage of the qRT-PCR
approach. We initially used SYBR Green to examine the
effects of hits on the transcriptional regulation of Phox2b
and Fgf4, in addition to Bmi1.1 To rule out any potential
artifacts obtained from using actin as a housekeeping gene,
we tested both actin and GAPDH as housekeeping genes in
SYBR Green studies. We found similar results with both
housekeeping genes. For all SYBR Green primers tested,
melting curves were generated and the products were vali-
dated with sequencing. This first screen of Bmi1, Phox2b,
and FGF4 levels served to eliminate compounds that may
have general repressive activities or may be acting on a
separate target or pathway responsible for Bmi1 regulation.
Of the 34 confirmed and validated hits tested, 22 com-
pounds regulated these targets in a manner similar to esBAF
(Fig. 6). Following analysis of mRNA expression levels of
five additional esBAF transcriptional targets (Eed, Cbx7,
Ring1a, Bmp4, and Socs3), another 2 compounds were
eliminated, leaving 20 compounds for follow-up studies
focusing on determining the mechanism of action. These
studies are currently under way.
To identify compounds that inhibit the esBAF complex, we
developed a gene expression–based screening system using
Figure 3. CT value as a function of number of cells plated per
well of a 384-well plate. The direct relationship between Bmi1 and
actin CT values means that the ΔΔCT value provides an accurate
normalization of Bmi1 transcript levels by actin, even at low
densities of cells.
Dykhuizen et al.
mouse ES cells. By monitoring mRNA expression levels of
Bmi1, a developmentally important esBAF-repressed target
gene, we were able to identify a series of compounds that
mimic the transcriptional action of the BAF complex in
mouse ES cells. Measurement of the endogenous Bmi1
mRNA expression levels by qRT-PCR afforded several
advantages as a primary screening strategy. It provided a
straightforward way to screen for transcriptional activity in
ES cells without genetic manipulation, thus reducing the
risk of assay artifacts. In addition, we found there to be
minimal optimization required to produce an assay with
low variability and high reproducibility. Both of these
points are concerns with standard reporter gene assays
(e.g., luciferase, β-galactosidase reporters). Although there
are limits on the number of compounds that can be screened
with this method (primarily due to cost and throughput fac-
tors), we have determined that gene expression–based
screening is a physiologically relevant, robust method for
screening a medium-sized library of compounds in ES
Although unconventional, we found this screening method
to be an advantageous way to screen for inhibitors of novel
transcriptional regulators for which very little is known
mechanistically and for which there are no known inhibitors.
Figure 4. (A) Scatter plot of the quantitative polymerase chain reaction (qPCR) primary screen. Broad ID refers to a unique compound
identifier assigned to each compound screened. The activity score is calculated as –ΔΔCT. An activity score of –1.33 was set as the
cutoff for hits (green line), which correlates to a 2.5-fold increase in Bmi1 transcript levels. Solid gray line represents no change in Bmi1
transcript levels (mean value for DMSO-treated wells); dotted gray lines represent three standard deviations from this mean. Active: both
replicates meet the hit calling threshold; inconclusive: only one replicate meets the hit calling threshold; inactive: neither replicate meets
the hit calling threshold. (B) Dose response of Bmi1 induction by a representative hit (compound 63), as determined by quantitative
reverse transcriptase PCR (qRT-PCR). (C) Dose response of Bmi1 induction by compound 63, as determined by luciferase levels using
the reporter embryonic stem (ES) cell line.
Journal of Biomolecular Screening 17(9)
We were able to use the qRT-PCR screen to validate the target
and to identify a panel of potential esBAF inhibitors for
follow-up studies. The qRT-PCR screen was very robust and
reproducible, making it fairly quick to identify and confirm
hits with a minimal number of false positives. Furthermore,
storage of the cDNA isolated from compound-treated cells
enabled rapid interrogation of additional targets to validate
the biological relevance of confirmed hits. From this work,
we identified a panel of 20 compounds that transcriptionally
mimic the Brg1 knockout phenotype in mouse ES cells.
Several factors, including cost, throughput, and ES cell-
specific considerations, will likely make a qRT-PCR approach
less feasible when screening larger collections (hundreds of
thousands to millions) of compounds. Efforts are currently
under way to miniaturize the assay into a 1536-well format to
expand the feasibility to larger screens.
What still remains to determine is the mechanism of
action for these validated hit compounds. As we still do not
understand the mechanism of BAF-mediated repression,
this remains a significant challenge. Our first step will be
to identify compounds that bind directly to the esBAF
complex using mass spectrometry–based binding assays.23
BAF is known to catalyze the movement and displacement
of nucleosomes. Since the ATPase function of BRG1 is
necessary for this function, our next step will be to charac-
terize any BAF binders in an in vitro ATPase assay and an
in vitro nucleosome remodeling assay.13,14 For compounds
that bind to esBAF but do not inhibit remodeling activity in
vitro, we will determine the subunit targeted by the small
molecules using seven established embryonic stem cell
lines that have been created in our lab with null mutations
in subunits of the complex. Finding inhibitors that target
certain domains, such as plant homeodomain fingers, chro-
modomains, or bromodomains, will help elucidate the
roles of individual subunits in esBAF targeting and may
explain how different combinations of BAF subunits pro-
mote cell-specific transcriptional programs.24,25 We will
investigate how these compounds affect the known in vivo
effects of the BAF complex. As mentioned earlier, only a
few subunits are required for nucleosome remodeling
activity in vitro, whereas many other subunits are required
for the in vivo activities of BAF. We will investigate the
effect of compounds on the activities, which include but
are not limited to nucleosome positioning,26 changes in his-
tone variants and histone modifications, DNase I hypersen-
sitivity sites associated with open chromatin in the promoter
regions of actively transcribed genes,15,27 looping and
higher-order chromatin structure,28,29 transcription factor
occupancy,1 and the cellular effects related to loss of
Figure 5. Increase in Bmi1 levels upon BRG1 depletion as
indicated by the luciferase reporter. Data are shown at 72 h after
viral infection with shRNA against Brg1.
Figure 6. Transcriptional profiles of E14 ES cells treated with shBrg1 (Brg1 KD) and three representative hits. The fold change was
calculated as comparison to DMSO-treated embryonic stem (ES) cells using Gapdh as the housekeeping gene. Twenty compounds out of
~30 000 screened produced a transcriptional profile similar to Brg1-deficient cells.
Dykhuizen et al.
Using a transcriptional reporter of esBAF repression not
only allows us to identify compounds that target the esBAF
complex directly but will also enable us to identify small-
molecule inhibitors of proteins that act on or with the BAF
complex. Future work will focus on determining if the vali-
dated hits inhibit players in putative ES cell-specific signal-
ing pathways that regulate esBAF function or inhibit proteins
that act in concert with the esBAF complex, including tran-
scription factors and chromatin-modifying enzymes required
for esBAF binding or activity. For compounds that do not
bind directly to the complex, we plan to use established
assays for target identification,30 with the goal of identifying
novel protein partners required for BAF activity in ES cells.
Small-molecule inhibitors will be used as tools to pro-
vide the precision and temporal control needed to tease out
the precise functions and mechanisms of the BAF complex.
These compounds will prove useful to researchers studying
stem cells, chromatin, and cancer. These compounds may
even one day be used for a therapeutic function, as an
increase in Brg1 expression in adult cardiomyocytes has
recently been shown to be a major player in hypertrophy.31
We thank Stuart L. Schreiber for his generous support and guid-
ance (GM38627, S.L.S.), Naoki Hosen for his generous gift of the
Bmi1-GFP construct, P. J. Aspesi and Vihren Kolev for technical
advice, and Lili Wang and Josh Bittker for critical reading of the
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect
to the research, authorship, and/or publication of this article.
The authors disclosed receipt of the following financial support for
the research, authorship, and/or publication of this article: This work
was supported by the National Institute of General Medical Sciences
(GM38627), the National Cancer Institute (N01-CO-12400), the
National Institute of Neurological Disorders and Stroke (NS046789),
and the American Cancer Society (121535-PF-11-145-01-DMC).
The content of this publication does not necessarily reflect the views
or policies of the Department of Health and Human Service, nor does
mention of trade names, commercial products, or organizations imply
endorsement by the U.S. government.
1. Ho, L. N.; Jothi, R.; Ronan, J. L.; Cui, K. R.; Zhao, K. J.;
Crabtree, G. R. An Embryonic Stem Cell Chromatin Remod-
eling Complex, esBAF, Is an Essential Component of the Core
Pluripotency Transcriptional Network. Proc. Natl. Acad. Sci.
U. S. A. 2010, 106, 5187–5191.
2. Ho, L.; Ronan, J. L.; Wu, J.; Staahl, B. T.; Chen, L.; Kuo,
A.; Lessard, J.; Nesvizhskii, A. I.; Ranish, J.; Crabtree, G. R.
An Embryonic Stem Cell Chromatin Remodeling Complex,
esBAF, Is Essential for Embryonic Stem Cell Self-Renewal
and Pluripotency. Proc. Natl. Acad. Sci. U. S. A. 2009, 106,
3. Mak, A. B.; Ni, Z.; Hewel, J. A.; Chen, G. I.; Zhong, G.; Karam-
boulas, K.; Blakely, K.; Smiley, S.; Marcon, E.; Roudeva, D.;
et al. A Lentiviral Functional Proteomics Approach Identifies
Chromatin Remodeling Complexes Important for the Induc-
tion of Pluripotency. Mol. Cell Proteomics 2010, 9, 811–8123.
4. Singhal, N.; Graumann, J.; Wu, G. M.; Arauzo-Bravo, M. J.;
Han, D. W.; Greber, B.; Gentile, L.; Mann, M.; Scholer, H.
R. Chromatin-Remodeling Components of the BAF Complex
Facilitate Reprogramming. Cell 2010, 141, 943–955.
5. Wu, J. I.; Lessard, J.; Crabtree, G. R. Understanding the Words
of Chromatin Regulation. Cell 2009, 136, 200–206.
6. Versteege, I.; Sevenet, N.; Lange, J.; Rousseau-Merck, M. F.;
Ambros, P.; Handgretinger, R.; Aurias, A.; Delattre, O. Trun-
cating Mutations of hSNF5/INI1 in Aggressive Paediatric
Cancer. Nature 1998, 394, 203–206.
7. Jones, S.; Wang, T. L.; Shih, I. M.; Mao, T. L.; Nakayama, K.;
Roden, R.; Glas, R.; Slamon, D.; Diaz, L. A.; Vogelstein, B.; et al.
Frequent Mutations of Chromatin Remodeling Gene ARID1A in
Ovarian Clear Cell Carcinoma. Science 2010, 330, 228–231.
8. Wang, K.; Kan, J. S.; Yuen, S. T.; Shi, S. T.; Chu, K. M.; Law,
S.; Chan, T. L.; Kan, Z. Y.; Chan, A. S. Y.; Tsui, W. Y.; et al.
Exome Sequencing Identifies Frequent Mutation of ARID1A
in Molecular Subtypes of Gastric Cancer. Nat. Genet. 2011,
9. Varela, I.; Tarpey, P.; Raine, K.; Huang, D.; Ong, C. K.; Ste-
phens, P.; Davies, H.; Jones, D.; Lin, M. L.; Teague, J.; et al.
Exome Sequencing Identifies Frequent Mutation of the SWI/
SNF Complex Gene PBRM1 in Renal Carcinoma. Nature
2011, 469, 539–542.
10. Li, M.; Zhao, H.; Zhang, X. S.; Wood, L. D.; Anders, R. A.;
Choti, M. A.; Pawlik, T. M.; Daniel, H. D.; Kannangai, R.;
Offerhaus, G. J. A.; et al. Inactivating Mutations of the Chro-
matin Remodeling Gene ARID2 in Hepatocellular Carcinoma.
Nat. Genet. 2011, 43, 828–829.
11. Parsons, D. W.; Li, M.; Zhang, X. S.; Jones, S.; Leary, R.
J.; Lin, J. C. H.; Boca, S. M.; Carter, H.; Samayoa, J.; Bet-
tegowda, C.; et al. The Genetic Landscape of the Childhood
Cancer Medulloblastoma. Science 2011, 331, 435–439.
12. Weissman, B.; Knudsen, K. E. Hijacking the Chromatin
Remodeling Machinery: Impact of SWI/SNF Perturbations in
Cancer. Cancer Res. 2009, 69, 8223–8230.
13. Logie, C.; Peterson, C. L. Purification and biochemical prop-
erties of yeast SWI/SNF complex. In Chromatin; Academic
Press: San Diego, CA, 1999; Vol. 304, pp 726–741.
14. Bultman, S. J.; Gebuhr, T. C.; Magnuson, T. A Brg1 Mutation
That Uncouples ATPase Activity from Chromatin Remodeling
Reveals an Essential Role for SWI/SNF-Related Complexes
in Beta-Globin Expression and Erythroid Development.
Genes Dev. 2005, 19, 2849–2861.
Journal of Biomolecular Screening 17(9)
15. Kim, S. I.; Bresnick, E. H.; Bultman, S. J. BRG1 Directly
Regulates Nucleosome Structure and Chromatin Looping of
the Alpha Globin Locus to Activate Transcription. Nucleic
Acids Res. 2009, 37, 6019–6027.
16. Phelan, M. L.; Sif, S.; Narlikar, G. J.; Kingston, R. E. Recon-
stitution of a Core Chromatin Remodeling Complex from
SWI/SNF Subunits. Mol. Cell 1999, 3, 247–253.
17. Letimier, F. A.; Passini, N.; Gasparian, S.; Bianchi, E.; Rogge,
L. Chromatin Remodeling by the SWI/SNF-like BAF Com-
plex and STAT4 Activation Synergistically Induce IL-12R
Beta 2 Expression during Human Th1 Cell Differentiation.
Embo J. 2007, 26, 1292–1302.
18. Hosen, N.; Yamane, T.; Muijtjens, M.; Pham, K.; Clarke, M.
F.; Weissman, I. L. Bmi-1-Green Fluorescent Protein-Knock-
in Mice Reveal the Dynamic Regulation of Bmi-1 Expression
in Normal and Leukemic Hematopoietic Cells. Stem Cells
2007, 25, 1635–1644.
19. Livak, K. J.; Schmittgen, T. D. Analysis of Relative Gene
Expression Data Using Real-Time Quantitative PCR and
the 2(T)(-Delta Delta C) Method. Methods 2001, 25,
20. Bittker, J. A. High-Throughput RT-PCR for Small-Molecule
Screening Assays. Curr. Protoc. Chem. Biol. 2012, 4, 49–63.
21. Ho, L.; Miller, E. L.; Ronan, J. L.; Ho, W. Q.; Jothi, R.; Crab-
tree, G. R. esBAF Facilitates Pluripotency by Conditioning
the Genome for LIF/STAT3 Signalling and by Regulating
Polycomb Function. Nat. Cell Biol. 2011, 13, 903–913.
22. Schuringa, J. J.; Vellenga, E. Role of the Polycomb Group Gene
BMI1 in Normal and Leukemic Hematopoietic Stem and Pro-
genitor Cells. Curr. Opin. Hematol. 2010, 17, 294–299.
23. Cancilla, M. T.; Leavell, M. D.; Chow, J.; Leary, J. A. Mass
Spectrometry and Immobilized Enzymes for the Screening of
Inhibitor Libraries. Proc. Natl. Acad. Sci. U. S. A. 2000, 97,
24. Lessard, J.; Wu, J. I.; Ranish, J. A.; Wan, M.; Winslow, M. M.;
Staahl, B. T.; Wu, H.; Aebersold, R.; Graef, I. A.; Crabtree, G.
R. An Essential Switch in Subunit Composition of a Chroma-
tin Remodeling Complex during Neural Development. Neu-
ron 2007, 55, 201–215.
25. Yoo, A. S.; Staahl, B. T.; Chen, L.; Crabtree, G. R. MicroRNA-
Mediated Switching of Chromatin-Remodelling Complexes
in Neural Development. Nature 2009, 460, 642–646.
26. Micrococcal Nuclease-Southern blot Assay. Nat. Methods
2005, 2, 719–720.
27. Gilbert, N.; Ramsahoye, B. The Relationship between Chromatin
Structure and Transcriptional Activity in Mammalian Genomes.
Brief. Funct. Genomics Proteomics 2005, 4, 129–142.
28. Hagege, H.; Klous, P.; Braem, C.; Splinter, E.; Dekker, J.;
Cathala, G.; de Laat, W.; Forne, T. Quantitative Analysis of
Chromosome Conformation Capture Assays (3C-qPCR). Nat.
Protoc. 2007, 2, 1722–1733.
29. Dostie, J.; Dekker, J. Mapping Networks of Physical Interac-
tions between Genomic Elements Using 5C Technology. Nat.
Protoc. 2007, 2, 988–1002.
30. Tashiro, E.; Imoto, M. Target Identification of Bioactive Com-
pounds. Bioorg. Med. Chem. 2012, 20, 1910–1921.
31. Hang, C. T.; Yang, J.; Han, P.; Cheng, H. L.; Shang, C.; Ash-
ley, E.; Zhou, B.; Chang, C. P. Chromatin Regulation by Brg1
Underlies Heart Muscle Development and Disease. Nature
2010, 466, 62–67.