Activator-binding domains of the SWI
⁄
SNF chromatin
remodeling complex characterized in vitro are required
for its recruitment to promoters in vivo
Monica E. Ferreira
1,2
, Philippe Prochasson
3,
*, Kurt D. Berndt
1,2
, Jerry L. Workman
3
and Anthony
P. H. Wright
1,2
1 School of Life Sciences, So
¨
derto
¨
rns Ho
¨
gskola, Huddinge, Sweden
2 Department of Biosciences and Medical Nutrition, Karolinska Institutet, Huddinge, Sweden
3 Stowers Institute for Medical Research, Kansas City, MO, USA
ATP-dependent chromatin remodeling complexes are a
group of enzymes that modulate transcriptional activa-
tion, as well as other chromosomal processes, by con-
trolling the accessibility of specific DNA sequences
within chromatin. A large number of remodeling com-
plexes have been identified based on similarities

rns Ho
¨
gskola, S-141 89 Huddinge,
Sweden
Fax: +46 8 6084510
Tel: +46 8 6084708
E-mail:
*Present address
Department of Pathology and Laboratory
Medicine, University of Kansas Medical
Center, KS, USA
(Received 15 December 2008, revised 20
February 2009, accepted 23 February 2009)
doi:10.1111/j.1742-4658.2009.06979.x
Interaction between acidic activation domains and the activator-binding
domains of Swi1 and Snf5 of the yeast SWI ⁄ SNF chromatin remodeling
complex has previously been characterized in vitro. Although deletion of
both activator-binding domains leads to phenotypes that differ from the
wild-type, their relative importance for SWI ⁄ SNF recruitment to target
genes has not been investigated. In the present study, we used chromatin
immunoprecipitation assays to investigate the individual and collective
importance of the activator-binding domains for SWI ⁄ SNF recruitment to
genes within the GAL regulon in vivo. We also investigated the conse-
quences of defective SWI ⁄ SNF recruitment for target gene activation. We
demonstrate that deletion of both activator-binding domains essentially
abolishes galactose-induced SWI ⁄ SNF recruitment and causes a reduction
in transcriptional activation similar in magnitude to that associated with a
complete loss of SWI ⁄ SNF activity. The activator-binding domains in Swi1
and Snf5 make approximately equal contributions to the recruitment of
SWI ⁄ SNF to each of the genes studied. The requirement for SWI ⁄ SNF

provided evidence for a direct interaction between acti-
vators and coactivators in vivo. Two main approaches
have been adopted to identify direct targets. Using fluo-
rescence resonance energy transfer (FRET) to measure
in vivo interactions between proteins fused to derivatives
of the green fluorescent protein, the activation domain
of the transcriptional activator Gal4 was reported to
interact directly with the Tra1 subunit of the histone
acetyl transferase complex SAGA [30]. A photo-cross-
linking strategy has independently identified Tra1 as a
direct target of Gal4, and several other acidic activators
[31]. Thus, the FRET and cross-linking approaches
appear to cross-validate each other. A photo-cross-
linking approach has identified the Swi1, Snf5 and
Swi2 ⁄ Snf2 subunits of the SWI ⁄ SNF complex as direct
targets bound by several acidic activators [32]. Subse-
quently, two partially redundant regions of Swi1 and
Snf5, respectively, were shown to mediate interaction
with transcriptional activators in vitro [33]. Recombi-
nant activator interaction domains interact with
activation domains in vitro using a two-step mechanism,
where rapid ionic interaction with an unfolded activa-
tion domain is followed by a slow entropy-driven step
during which the activation domain folds [34]. A similar
mechanism has been reported in another activator target
interaction context and it has been suggested that the
intrinsic conformational flexibility of the interaction
mechanism may facilitate activator interactions with
different coactivator targets [35]. Deletion of both acti-
vator-binding domains does not disrupt the composi-

study the GAL regulon, which is known to be regu-
lated by the Gal4 activator protein. We first tested a
group of known Gal4 target genes [38] to identify a
group of genes showing robust induction under our
conditions. Table 1 shows that several GAL genes are
activated shortly after addition of galactose to cultures
grown with raffinose as carbon source. Induction of
PGM2, FUR4, MTH1 and PCL10 was not detected
under our conditions. It is likely that these genes are
induced at a later time after galactose addition. Using
the identified set of galactose-induced genes, we next
screened for those genes that require the SWI ⁄ SNF
complex for induction by galactose. For this purpose,
we used strain YPP33, in which the entire ORFs of
SWI1 and SNF5 are both deleted, because deletion of
these genes has been shown to disrupt the integrity
of the SWI ⁄ SNF complex [39,40]. The most appropriate
time for revealing SWI ⁄ SNF dependence was found to
Mechanism of SWI ⁄ SNF recruitment to promoters M. E. Ferreira et al.
2558 FEBS Journal 276 (2009) 2557–2565 ª 2009 The Authors Journal compilation ª 2009 FEBS
be 30 min after galactose addition to the cultures.
Under these conditions GAL1, GAL10, GCY1, GAL2
and GAL7 showed a high degree of SWI ⁄ SNF depen-
dence for induction by galactose (Fig. 1). GAL80 and
GAL3 did not show significant SWI ⁄ SNF dependence,
nor did PGM2, which was included as a control to
represent those genes that were not induced at this
time-point. We conclude that a subset of Gal4-induc-
ible genes were induced under our conditions, and that
a further subset of these genes were SWI ⁄ SNF depen-

are required for promoter recruitment of the
SWI
⁄
SNF complex
We next studied the level of SWI ⁄ SNF recruitment to
the GAL1-10, GAL2 and GAL7 regulatory regions in
strains lacking either or both the Swi1 (residues 329–
655) and Snf5 (residues 2–327) activator-binding
domains, 30 min after galactose induction. Quantifying
recruitment in relation to background binding to the
promoter of the SWI ⁄ SNF-independent ILS1 gene, we
found that SWI ⁄ SNF recruitment to GAL1-10 and
GAL2 is reduced by approximately 50% in strains
lacking either the Swi1 or the Snf5 activator-binding
domains (Fig. 3). In the strain lacking both activator-
binding domains, the level of SWI ⁄ SNF recruitment is
reduced to background levels. We have consistently
observed that the activation domains are required for
the recruitment of SWI ⁄ SNF to the GAL7 gene, but
technical problems have thus far prevented acquisition
of quantitative data. We conclude that the activator-
binding domains identified in vitro are essential for
SWI ⁄ SNF recruitment in vivo and that Swi1 and Snf5
Table 1. Wild-type induction 30 min post galactose addition:
screening by real-time RT-PCR.
Gene Fold induction
GAL2 522
GAL1 > 1000
GAL7 > 1000
GAL10 197

SNF-dependent genes
It was next necessary to determine whether the defect
in SWI ⁄ SNF recruitment resulting from the deletion of
one or both activator-binding domains would lead to
reduced galactose-induced expression of target genes.
Figure 4 shows that induction of GAL1, GAL10,
GAL2 and GAL7 is severely reduced in a strain lacking
both the Swi1 and Snf5 activator-binding domains.
Recruitment of the SWI ⁄ SNF complex via activator
interactions with these activator-binding domains is
thus critical for normal galactose induction of the
tested genes. Interestingly, the strains lacking either the
Fig. 3. The activator-binding domains of the SWI ⁄ SNF complex are
required for its recruitment to promoters. Enrichment of the immu-
noprecipitated GAL1-10 and GAL2 promoters under inducing condi-
tions (galactose, 30 min) in the wild-type strain (black bars, Wt),
strain YPP310 (light grey bars, DDABDswi1 + snf5) lacking both
activator-binding domains (SWI1D329-655, SNF5D2-327), strain
YPP211 (white bars, DABDsnf5) lacking the Snf5 activator-binding
domain (SNF5D2-327) and strain YPP247 (dark grey bars, DAB-
Dswi1) lacking the Swi1 activator-binding domain (SWI1D329-655).
The level of SWI ⁄ SNF enrichment on the promoters is relative to
the negative control promoter, ILS1. Error bars indicate the values
obtained from two independent experiments.
A
B
C
Fig. 2. SWI ⁄ SNF is recruited to the GAL1-10, GAL7 and GAL2 pro-
moters within 30 min of galactose induction. (A) Schematic of the
investigated GAL promoter regions. Solid black lines indicate the

DNA binding sites in target genes but its activity is
repressed by its association with the Gal80 repressor
protein. Upon addition of galactose to such cultures,
the Gal3 protein antagonizes Gal80-mediated repres-
sion and Gal4 is immediately able to recruit necessary
factors to its target genes [41]. As shown in the present
study, the SWI ⁄ SNF complex is rapidly recruited to a
subset of Gal4 target genes within 30 min of galactose
addition, and defects affecting the SWI ⁄ SNF complex
caused reduced activation of these genes within the
same time window. The activator-binding domains in
the Swi1 and Snf5 subunits of the SWI ⁄ SNF complex
are essential for recruitment of the SWI ⁄ SNF complex
to these genes, as well for their subsequent activation.
Our observation thus strongly supports the model pro-
posing that the activator-binding domains, largely
defined and characterized in vitro, are necessary for
SWI ⁄ SNF recruitment in vivo. This is crucial for
understanding the roles played by the different
SWI ⁄ SNF subunits and, from the results obtained in
the present study, we can now add that other regions
of SWI ⁄ SNF cannot replace the recruiting function of
the Swi1 and Snf5 activator-binding domains. The
in vivo validation of the Swi1 and Snf5 activator-bind-
ing domains is also of interest in relation to the two-
step binding mechanism between Gal4 and Swi1 that
has been characterized in vitro [34], and our results
suggest that the coupled binding and folding mecha-
nism is likely to be relevant in vivo. This binding mech-
anism is assumed to be generally important for

tor-binding domain (SWI1D329-655), relative to the wild-type tested
in parallel (black bars, Wt, level arbitrarily set to 1). GAL transcript
levels were quantified by real-time PCR after cDNA synthesis and
normalized against transcript levels of a control (ILS1) to obtain nor-
malized GAL expression. Normalized GAL expression levels were
subsequently scaled to give a wild-type value of 1 for all tested
genes. Error bars indicate the SD of means obtained from three
independent cultures.
M. E. Ferreira et al. Mechanism of SWI ⁄ SNF recruitment to promoters
FEBS Journal 276 (2009) 2557–2565 ª 2009 The Authors Journal compilation ª 2009 FEBS 2561
latter alternative is consistent with measurements
showing rapid dynamics of activator and coactivator
association and disassociation with chromatin in vivo
[42,43]. For the genes investigated in the present study,
we have not observed differences in the relative impor-
tance of the Swi1 and Snf5 activator interaction
domains during SWI ⁄ SNF recruitment to different
genes. It remains possible that the two interaction
domains play differential roles on other SWI ⁄ SNF-
dependent genes or under different physiological
conditions; for example, during mitosis when a larger
number of genes become SWI ⁄ SNF dependent [44].
We have previously shown that such differences can
exist as demonstrated by the observation that different
subunits of the Tup-Ssn6 corepressor complex are
differentially important for the repression of different
target genes in fission yeast [45,46].
Swi1 and Snf5 are known to be required for the
structural integrity of the SWI ⁄ SNF complex [39,40].
The defect in transcriptional activation of GAL genes

(GAL1, GAL10, GAL2 and GAL7), whereas genes
showing lower inducibility by galactose after 30 min
(GAL3, GAL80 and PGM2) were not significantly
dependent on SWI ⁄ SNF. Thus, the extent of
SWI ⁄ SNF dependence may depend on the extent
and ⁄ or rate of induction. GCY1 is an exception
because it was strongly dependent on SWI ⁄ SNF, even
though it is relatively mildly induced by galactose. The
results obtained in the present study suggest that
GCY1 is an indirect target of SWI ⁄ SNF because we
were unable to detect an association of SWI ⁄ SNF with
GCY1. However, other explanations are possible. For
example, SWI ⁄ SNF might interact with regions of
GCY1 that are not detected by the primers used.
Unlike the other SWI ⁄ SNF-dependent genes, basal
levels of GCY1 expression were also dependent on
SWI ⁄ SNF (data not shown). It is therefore likely that
a factor required for GCY1 expression in both unin-
duced and induced conditions is SWI ⁄ SNF dependent.
Although an association of SWI ⁄ SNF with the GAL1-
10 and GAL7 genes has been reported previously [50],
albeit under different conditions than those reported
here, SWI ⁄ SNF association with GAL2 is a novel find-
ing of the present study. Furthermore, we demonstrate
that SWI ⁄ SNF is important for the efficient activation
of a number of GAL genes upon galactose induction
of cultures grown with raffinose as carbon source,
which comprise conditions where SWI ⁄ SNF has previ-
ously been reported not to play a role [51]. This
discrepancy may be explained by differences in assay

Swi1 and 2–327 of Snf5, were made using the Cre-loxP
system [52] and verified by PCR.
Galactose induction experiments
Yeast were pre-cultured in yeast-peptone-dextrose medium
at 30 °C for approximately 24 h, washed and diluted in
sterile-filtered YP 2% raffinose, supplemented with adenine,
and grown for at least 16 h until a D
600
of approximately
0.4–0.7 was reached, at which point each culture was split
in two and induced for 30 min at 30 °C by the addition of
one-tenth of the culture volume of a 20% galactose solu-
tion, or mock induced using an equal volume of sterile
deionized water.
Preparation of RNA
Samples for total RNA extraction were harvested at room
temperature by centrifugation, immediately frozen in liquid
nitrogen and stored at )70 °C until purification. Total
RNA was extracted using hot phenol extraction, followed
by further purification using an RNA Easy mini kit (Qia-
gen, Solna, Sweden). Total RNA was treated with amplifi-
cation grade DNase I (Invitrogen, Lidingo
¨
, Sweden)
according to manufacturer’s instructions, followed by
another round of purification using an RNA Easy mini kit.
The quality of RNA was checked by electrophoresis of
total RNA, and by control RT-PCR, using 50 ng total
RNA per 25 lL of reaction, 200 nm each of ARN1 specific
primers and SuperScript III Platinum One-Step RT-PCR

1, 5 and 10 lg of Snf2 antibody per reaction, and the
routinely-used beads only as a control.
Real-time PCR and RT-PCR
Quantitative PCR was performed in duplicate 25-lL reac-
tions using the MyiQ Single-Color Real-Time PCR Detec-
tion System (Bio-Rad, Sundbyberg, Sweden), 200 nm of
each primer and, unless otherwise stated, iQ SYBR Green
Supermix (Bio-Rad). One-step quantitative RT-PCR was
performed with SuperScript III Platinum SYBR Green One-
Step qRT-PCR kit (Invitrogen) supplemented with 10 n m
fluorescein (Invitrogen), using 50 ng of total RNA per reac-
tion, and expression of ARN1 was used for normalization.
For two-step qRT-PCR, cDNA was prepared using iScript
cDNA Synthesis kit (Bio-Rad), containing oligo-dT and
random primers. cDNA corresponding to 50 ng of input
total RNA was used for each subsequent quantitative PCR
reaction with specific primers, and expression of ILS1 was
used for normalization. Unless otherwise stated, the ILS1
promoter region was used for normalization in chromatin
immunoprecipitation experiments for comparison of wild-
type and mutant strains. Primers used for quantification of
transcripts and promoter regions were for the most part
designed using primer3 software [54]. Primer sequences are
available from the authors upon request.
Acknowledgements
We thank Professor Hans Ronne at Uppsala Univer-
sity for sharing the SNF2 deletion strain. This work
was supported by a grant to A.W. from the Swedish
Research Council, and by a Leukemia and Lymphoma
Society Special Fellowship to P.P.

899–903.
7 Quinn J, Fyrberg AM, Ganster RW, Schmidt MC &
Peterson CL (1996) DNA-binding properties of the
yeast SWI ⁄ SNF complex. Nature 379, 844–847.
8 Bazett-Jones DP, Cote J, Landel CC, Peterson CL &
Workman JL (1999) The SWI ⁄ SNF complex creates
loop domains in DNA and polynucleosome arrays and
can disrupt DNA-histone contacts within these
domains. Mol Cell Biol 19, 1470–1478.
9 Whitehouse I, Flaus A, Cairns BR, White MF, Work-
man JL & Owen-Hughes T (1999) Nucleosome mobili-
zation catalysed by the yeast SWI ⁄ SNF complex.
Nature 400, 784–787.
10 Bruno M, Flaus A, Stockdale C, Rencurel C, Ferreira
H & Owen-Hughes T (2003) Histone H2A ⁄ H2B dimer
exchange by ATP-dependent chromatin remodeling
activities. Mol Cell 12, 1599–1606.
11 Boeger H, Griesenbeck J, Strattan JS & Kornberg RD
(2004) Removal of promoter nucleosomes by disassem-
bly rather than sliding in vivo. Mol Cell 14, 667–673.
12 Gutie
´
rrez JL, Chandy M, Carrozza MJ & Workman JL
(2007) Activation domains drive nucleosome eviction by
SWI ⁄ SNF. EMBO J 26, 730–740.
13 Dechassa ML, Zhang B, Horowitz-Scherer R, Persinger
J, Woodcock CL, Peterson CL & Bartholomew B
(2008) Architecture of the SWI ⁄ SNF-nucleosome com-
plex. Mol Cell Biol 28, 6010–6021.
14 Holstege FC, Jennings EG, Wyrick JJ, Lee TI, Hengart-

promoter. J Biol Chem 278, 11026–11031.
23 Yuan GC, Liu YJ, Dion MF, Slack MD, Wu LF, Alts-
chuler SJ & Rando OJ (2005) Genome-scale identifica-
tion of nucleosome positions in S. cerevisiae. Science
309, 626–630.
24 Dhasarathy A & Kladde MP (2005) Promoter occu-
pancy is a major determinant of chromatin remodeling
enzyme requirements. Mol Cell Biol 25, 2698–2707.
25 Ryan MP, Jones R & Morse RH (1998) SWI-SNF com-
plex participation in transcriptional activation at a step
subsequent to activator binding. Mol Cell Biol 18,
1774–1782.
26 Neely KE, Hassan AH, Wallberg AE, Steger DJ, Cairns
BR, Wright AP & Workman JL (1999) Activation
domain-mediated targeting of the SWI ⁄ SNF complex to
promoters stimulates transcription from nucleosome
arrays. Mol Cell 4, 649–655.
27 Wallberg AE, Neely KE, Hassan AH, Gustafsson JA,
Workman JL & Wright AP (2000) Recruitment of the
SWI-SNF chromatin remodeling complex as a mecha-
nism of gene activation by the glucocorticoid receptor
tau1 activation domain. Mol Cell Biol 20, 2004–2013.
28 Yudkovsky N, Logie C, Hahn S & Peterson CL (1999)
Recruitment of the SWI ⁄ SNF chromatin remodeling
complex by transcriptional activators. Genes Dev 13,
2369–2374.
29 Natarajan K, Jackson BM, Zhou H, Winston F & Hin-
nebusch AG (1999) Transcriptional activation by Gcn4p
involves independent interactions with the SWI ⁄ SNF
complex and the SRB ⁄ mediator. Mol Cell 4, 657–664.

somes. Cell 111, 369–379.
37 Wu JI, Lessard J, Olave IA, Qiu Z, Ghosh A, Graef IA
& Crabtree GR (2007) Regulation of dendritic develop-
ment by neuron-specific chromatin remodeling com-
plexes. Neuron 56, 94–108.
38 Ren B, Robert F, Wyrick JJ, Aparicio O, Jennings EG,
Simon I, Zeitlinger J, Schreiber J, Hannett N, Kanin E
et al. (2000) Genome-wide location and function of
DNA binding proteins. Science 290, 2306–2309.
39 Peterson CL, Dingwall A & Scott MP (1994) Five
SWI ⁄ SNF gene products are components of a large
multisubunit complex required for transcriptional
enhancement. Proc Natl Acad Sci USA 91, 2905–2908.
40 Peterson CL & Herskowitz I (1992) Characterization of
the yeast SWI1, SWI2, and SWI3 genes, which encode
a global activator of transcription. Cell 68, 573–583.
41 Campbell RN, Leverentz MK, Ryan LA & Reece RJ
(2008) Metabolic control of transcription: paradigms
and lessons from Saccharomyces cerevisiae. Biochem J
414, 177–187.
42 Johnson TA, Elbi C, Parekh BS, Hager GL & John S
(2008) Chromatin remodeling complexes interact
dynamically with a glucocorticoid receptor-regulated
promoter. Mol Biol Cell 19, 3308–3322.
43 Nagaich AK, Walker DA, Wolford R & Hager GL
(2004) Rapid periodic binding and displacement of the
glucocorticoid receptor during chromatin remodeling.
Mol Cell 14, 163–174.
44 Krebs JE, Fry CJ, Samuels ML & Peterson CL (2000)
Global role for chromatin remodeling enzymes in mito-

required for transcriptional memory at the yeast GAL
gene cluster. Genes Dev 21, 997–1004.
52 Hegemann JH, Gldener U & Kohler GJ (2006) Gene
disruption in the budding yeast Saccharomyces cerevisi-
ae. Methods Mol Biol 313, 129–144.
53 Robyr D & Grunstein M (2003) Genomewide histone
acetylation microarrays. Methods 31, 83–89.
54 Rozen S & Skaletsky H (2000) Primer3 on the WWW
for general users and for biologist programmers. Meth-
ods Mol Biol 132, 365–386.
55 Thomas BJ & Rothstein R (1989) Elevated recombina-
tion rates in transcriptionally active DNA. Cell 56, 619–
630.
56 Winzeler EA, Shoemaker DD, Astromoff A, Liang H,
Anderson K, Andre B, Bangham R, Benito R, Boeke
JD, Bussey H et al. (1999) Functional characterization
of the S. cerevisiae genome by gene deletion and paral-
lel analysis. Science 285, 901–906.
M. E. Ferreira et al. Mechanism of SWI ⁄ SNF recruitment to promoters
FEBS Journal 276 (2009) 2557–2565 ª 2009 The Authors Journal compilation ª 2009 FEBS 2565