The ins and outs of ATP-dependent chromatin remodeling in budding yeast:
Biophysical and proteomic perspectives
Joke J.F.A. van Vugta,1, Michael Ranesa,1, Coen Campsteijnb, Colin Logiea,⁎
aDepartment of Molecular Biology, NCMLS, Radboud University, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands
bSars International Centre for Marine Molecular Biology, Thormøhlensgt. 55, N-5008 Bergen, Norway
Received 2 November 2006; received in revised form 22 January 2007; accepted 29 January 2007
Available online 9 February 2007
ATP-dependent chromatin remodeling is performed by multi-subunit protein complexes. Over the last years, the identity of these factors has
been unveiled in yeast and many parallels have been drawn with animal and plant systems, indicating that sophisticated chromatin transactions
evolved prior to their divergence. Here we review current knowledge pertaining to the molecular mode of action of ATP-dependent chromatin
remodeling, from single molecule studies to genome-wide genetic and proteomic studies. We focus on the budding yeast versions of SWI/SNF,
RSC, DDM1, ISWI, CHD1, INO80 and SWR1.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Nucleosome; Chromatin; SNF2; STH1; ISW1; ISW2; INO80; SWR1; CHD1; YFR038w
DNA forms the backbone of numerous processes in the cell.
In order to play its role well, DNA has to be available when
needed and otherwise be silent and tightly compacted to fit in
the nucleus. DNA is packed in nucleosomes, where 147 base
pairs of DNA are wrapped about 1.7 left-handed turns around
the histone octamer . Nucleosomes have the capacity to form
higher order structures that are called chromatin fibres .
Chromatin remodeling is mediated by two distinct classes of
molecular processes, which are highly intertwined. The first
class remodels chromatin by post-translational modification of
histones, such as acetylation, methylation, phosphorylation,
ubiquitylation,andsumoylation [3–5].Thesecond classisATP-
By changing the DNA–histone interactions ATPases can slide,
eject, insert or restructure histone octamers [6,7].
Over the past decade it has become clear that remodeling by
ATPases and by covalent modification are closely linked
processes. Furthermore, ATPases appear to fulfil both redundant
and non-redundant roles as well as appearing to function
consecutively and sometimes antagonistically. Histone modi-
fiers and ATP-dependent remodellers share subunits and
interact with both DNA and RNA polymerase subunits, raising
many fascinating mechanistic questions.
Here we review ATP-dependent chromatin remodeling
enzymes in light of the ever more sophisticated approaches
that are being deployed to understand their molecular mode of
action, all the way down to single molecule analyses.
Furthermore, we used a large, well curated, proteomic database
as a backdrop to review recent literature on ATP-dependent
This review focuses on five SNF2 subfamilies, i.e. SWI2/
SNF2, ISWI, CHD1, INO80 and DDM1, because these have
demonstrated nucleosome remodeling activity. Orthologs of
these ATPases are present in fungi, plants and animals
suggesting that their basic cellular functions have been
maintained throughout eukaryotic evolution.
2. Current insight in ATP-dependent chromatin
In this section we review the biophysical knowledge that has
accumulated over the last decade concerning ATP-dependent
Biochimica et Biophysica Acta 1769 (2007) 153–171
E-mail address: email@example.com (C. Logie).
1These two authors contributed equally to this work.
0167-4781/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
chromatin remodeling factors. In the course of chromatin
remodeling, histone octamers can be restructured, partly
released from DNA or repositioned. Restructuring entails
insertion of histone variants, while repositioning can either
involve octamer sliding over the DNA or complete dissociation.
Nucleosomes count fourteen histone–DNA interactions ,
which all have to be broken when the nucleosome is moved,
though this does not have to occur simultaneously. Approxi-
mately 1 kcal/mol is needed to break each DNA–histone
contact, which adds to 12–14 kcal/mol for each nucleosome .
Single-pair fluorescence resonance energy transfer indicated
that at physiological salt concentrations a 164 bp DNA segment
containing a nucleosome positioning sequence is fully wrapped
around a single octamer for 250 ms before spontaneous partial
unwrapping and rewrapping within 10–50 ms . Unwrapping
may be enough for protein complexes to access DNA, but to
reposition and restructure nucleosomes in a defined manner,
ATP-dependent chromatin remodellers have to come into play.
All known ATP-dependent chromatin remodellers belong to
the super family 2 (SF2) of DEAD/H-box nucleic acid-
stimulated ATPases [10,11]. SF2 super family members include
DNA and RNA translocases, some of which also have helicase
activity, as well as type I restriction enzymes. In Saccharomyces
cerevisiae there are five different subfamilies of SNF2-type
ATP-dependent chromatin remodellers for which nucleosome
remodeling has been documented. Four of these can be
classified on the basis of conserved protein motifs in the
ATPases. SWI2/SNF2-types have bromodomains that bind
acetylated lysines [12,13], ISWI-types harbor SANT and
SLIDE domains that involve histone tail and linker DNA
binding, respectively , CHD-types bear chromo domains
that can bind methylated lysines  and INO80-types have
DBINO domains that are predicted to bind DNA . Finally,
YFR038w, known in plants as DDM1, represents a distinct
subfamily for which no specific protein motif has been found,
aside from homology within the ATPase domain [17,18]. The
first four types are known to reside in multi-subunit complexes
that can harbor up to 17 subunits. Whether YFR038w/DDM1
forms a stable complex with other proteins remains to be
ATP-dependent chromatin remodellers display similar rates
of ATP hydrolysis (Vmax∼1000 ATP/min) when provided with
normal chromatin [19,20], but when naked DNA, octamers or
nucleosomes deficient of linker DNA or histone tails are served
as substrate, striking differences in ATPase activities are
observed. Yeast Swi2 and Sth1 ATPase activity is maximally
stimulated by free DNA with no further stimulatory effect by
nucleosomes [20,21]. ISWI and INO80 are stimulated by DNA,
but maximally by nucleosomes [20,22,23]. ISWI binds
nucleosomes that display free DNA best and needs the N-
terminal tail of histone H4 to be fully active . The ATPase
activity of dMi-2 (CHD-type) is highly stimulated by nucleo-
somes lacking free DNA and only slightly by naked DNA
[20,25]. Contrary to the ISWI bearing Drosophila NURF
complex, Xenopus Mi-2 complex appears not to rely on histone
tails but on another moiety that is provided by H3/H4 tetramers
. ATP hydrolysis by DDM1 is comparable to that of the
SWI2/SNF2-types as it is greatly stimulated by naked DNA and
only slightly more by nucleosomes .
Although the ATPase subunits of all chromatin remodellers
have some homology with helicases, no intrinsic helicase
activity has been reported and no single strand DNA regions
appear to be introduced in the course of remodeling [26–28].
The INO80 complex does have helicase activity , but this is
due to the presence of two RuvB-like helicase subunits in the
complex. Despite the lack of helicase activity, the isolated
ATPases do display ATP-dependent triple-strand displacement.
The SWI2/SNF2-type Sth1 ATPase from yeast needs a double
helix to begin triple strand displacement , whereas ISWI
only displaces a third strand when it is located within 50 bp of a
How exactly ATP hydrolysis is converted into a mechanical
force that can break DNA–histone interactions is still unknown.
Models that have been proposed so far involve DNA
translocation, nucleosome sliding, intra- and inter-nucleosomal
DNA looping and DNA twisting. Although variations exist in
stimulation of ATP hydrolysis, SWI2/SNF2 and ISWI have in
common that ATP hydrolysis results in DNA translocation
[30,31]. DDM1, INO80 and CHD-type remodellers have not
been tested for translocation, yet. ATPase subunits alone also
translocate DNA [30,32], though ISWI-like ATPases need
nucleosomes to efficiently hydrolyze ATP, which may reflect
their requirement for histone tail binding [33,34]. ATP is turned
over continuously during translocation until the end of the DNA
molecule is reached, which can be illustrated by the facts that
ensemble ATPase activity is proportional to the length of the
linear DNA molecules provided and that DNA minicircles
maximally stimulate ATPase activity . The effectiveness of
single stranded DNA in stimulating ATPase activity suggests
that translocation occurs along one strand of the DNA duplex
. Introduction of 5 or 10 bp DNA gaps showed that
chromatin remodellers track in a 3′–5′ direction along the DNA
[31,35]. Moreover, experiments on the SWI/SNF complex
revealed that it binds to the minor groove, as DNA binding was
disrupted by the minor groove binding agent distamycin A
Exactly how many DNA binding sites and histone contacts
chromatin remodellers comprise is still unknown. The yeast
RSC complex (SWI/SNF type) has about six of its subunits
bound to free or remodeled DNA and at most four when bound
to nucleosomes before remodeling . SWI/SNF and RSC are
known to bind to the nucleosome core, whereas ISWI not only
binds the nucleosome core but also linker DNA [35,38].
Binding DNA at two different locations gives chromatin
remodellers the opportunity to form DNA loops. DNA looping
has so far been reported for SWI2/SNF2- and ISWI-type
remodellers [21,39–41]. In general SWI/SNF-type remodellers
seem to form larger DNA loops than ISWI-type remodellers,
about 50 and 10 bp respectively [30,35,38,39,42]. The DNA
bulge/loop length is correlated with the nucleosome step size of
remodellers, though with 28 bp the SWI/SNF nucleosome step
size is somewhat shorter than its DNA loop size . Magnetic
tweezer experiments with RSC on naked DNA and SWI/SNF
on nucleosomal DNA reveal that loop length increases
154J.J.F.A. van Vugt et al. / Biochimica et Biophysica Acta 1769 (2007) 153–171
moderately with increasing ATP concentration [21,44].
Whereas SWI2/SNF2-type remodellers induce rapid and large
DNA loops of approximately 400 bp within 2 sec on bare DNA,
on nucleosomal DNA they create average loop sizes of only
100 bp at 13 bp/s [21,44]. Intranucleosomal DNA looping is the
most substantiated and therefore favoured model at the moment
to describe the mechanism of chromatin remodeling [7,21,30,
35,38,39,41,44,45]. In this model nucleosomal DNA is thought
to be pulled in by means of DNA translocation, causing
breakage of part of the octamer–DNA contacts. The octamer
can then move to a new position on the DNA via bulge/loop
propagation. Notably, in this model the complete release of
octamers from DNA is not necessary. This model has also been
proposed for octamer transfer in trans, since the part of the
octamer that is not bound to DNA during loop formation can
bind to another DNA molecule after which the octamer can be
further released from the DNA of origin . A closely related
model invokes internucleosomal DNA looping. This model
predicts octamer sliding when the chromatin remodeller forms a
DNA loop with the right size at the side of the nucleosome .
An additional mechanism of ATP-dependent chromatin
remodeling can be envisaged, namely octamer conformation
remodeling, for example by altering [H3/H4]2 tetramer
handedness . However, cross linking of histone octamers
did not prevent SWI/SNF mediated nucleosome repositioning,
indicating that octamer quaternary conformation remodeling is
unlikely to be an absolute prerequisite step [28,39,41]. A
chromatin remodeling complex that explicitly performs octamer
restructuring is the SWR1 complex (INO80 type), which
exchanges H2Awith the yeast H2Avariant Htz1 or its ortholog
in mammals H2A.Z [49,50]. In fact, the SWI/SNF complex can
also promote exchange of H2A and H2B, as was first suggested
in 1994  and later observed on Mouse Mammary Tumor
Virus (MMTV) promoters reconstituted with nucleosomes
[51,52]. The rates of H2A/H2B dimer loss appeared to be
strongly affected by the underlying DNA sequence since dimer
loss was mainly observed on the second nucleosome of the
MMTV array . Not surprisingly, this reaction can also be
modulated by transcription factors .
There is no evidence for sequence specific binding of
chromatin remodellers to DNA. Indeed, SF2 members are
thought to interact with the phosphodiester backbone of DNA
independent fashion . As discussed above, the observed
the MMTV promoter is likely a consequence of distinctive
contacts. It is known that nucleosome positioning can be
influenced by DNA sequence, independent of chromatin
remodellers . In budding yeast this positioning code occurs
genome-wide and explains approximately 50% of the in vivo
octamer positions . Eukaryotic genomes may even use this
positioning code to define specific chromosome functions.
Chromatin remodeling enzymes can overrule this octamer–
DNA sequence preference and reposition nucleosomes when
necessary.SWI/SNF causes disordered nucleosome positioning,
thereby promoting transcription factor binding and gene
activation . In contrast ISWI type remodellers space
nucleosomes with regular distance from one another to generate
tightly packed, nuclease inaccessible DNA [56–58], although
there are exceptions to this, like the Drosophila NURF complex
. Drosophila ACF, another ISWI complex, can assemble
histone H1 on chromatin, further indicating that ISWI
remodellers are involved in higher order chromatin organization
[7,57,60]. Drosophila CHD1 also catalyzes the periodic spacing
of nucleosomes . The effect of INO80 and Yfr038w on
nucleosome spacing has not yet been reported.
ATP-dependent chromatin remodellers can generate negative
supercoils in DNA, which to date has been observed for SWI/
SNF, ISWI and CHD type complexes [21,61]. Magnetic
tweezers experiments on RSC demonstrated that negative coils
were introduced in the translocated DNA and positive coils
accumulated in front of the translocated DNA, indicating that
DNA moves in a right-handed fashion through the remodeller
. Supercoils can be introduced by DNA twisting. At low
remodeling enzyme concentrations the presence of nicks
reduced remodeling efficiency two to threefold for both RSC
and Sth1, indicating that DNA twisting is needed for chromatin
remodeling by this complex . On the other hand, ISWI still
not mean that ISWI does not twist DNA to remodel chromatin,
since it can establish ATP-dependent superhelical torsion in
DNA . Recently, DNA twisting was argued to be an early
step in nucleosome remodeling by both ISWI and SWI/SNF
It is currently not clear whether all SNF2-type chromatin
remodeling complexes actually employ a common mechanism,
though it seems reasonable to assume that at least part of the
catalytic cycle is conserved [63–66]. Even if the basic
mechanism of nucleosome movement would be the same for
all ATP-dependent chromatin remodellers, their effects on
chromatin are different. Some of these differences are due to the
presence of specific subunits bound to the ATPases . Yeast
ISW1a and ISW1b complexes, for example, harbor the same
ATPase subunit but move octamers in opposite directions in
mononucleosome experiments . The diverse and sometimes
contradictory effects that remodellers have on chromatin are
also exemplified by experiments conducted with the histone
chaperone Nap1. In the presence of ATP and a 1000-fold excess
of this chaperone protein, yeast RSC complex disassembled
nucleosomes, thereby dissociating H2A/H2B dimers first .
However, upon addition of as many octamers as Nap1
chaperones, RSC reassembled all naked DNA into nucleosomal
particles . On the other hand, Drosophila CHD1 has only
been reported to assemble nucleosomes onto DNA in the
presence of Nap1 and ATP . Rather than functional
differences, these different results may well reflect subtly
different experimental set-ups. Moreover, since different
subunits within one complex may also reflect participation in
different processes, it is conceivable that they only influence the
outcome of remodeling in specific chromatin contexts.
Application of approaches such as exchange of protein domains
can be very helpful to formally establish intrinsic molecular
differences among the SNF2 enzymes .
155J.J.F.A. van Vugt et al. / Biochimica et Biophysica Acta 1769 (2007) 153–171
What about the rate of chromatin remodeling? Coupled
restriction enzyme reactions on an 11-mer nucleosome array
using SWI/SNF and RSC complexes provided a lowest estimate
of 4.5 min per nucleosome per complex . Recently, using an
elegant DNA duplex unzipping set-up on single reconstituted
nucleosomes, SWI/SNF was shown to slide nucleosomes
approximately 28 bp in 0.5 to 1 min, with a catalytic efficiency
of 0.4 min−1per nM enzyme . On the other hand, single
molecule experiments show that SWI2/SNF2-type remodellers
pull on nucleosomal DNA at approximately 13 bp/s, and altered
nucleosomal states generated by SWI/SNF are obtained in less
than a second [39,44]. Together these results suggest that
octamer repositioning represents the end-product of a much
faster running ATP hydrolysis cycle.
We expect that application of single molecule techniques like
atomic force microscopy and magnetic and optical tweezers will
soon provide a more detailed view of the kinetic parameters of
ATP-dependent nucleosome remodeling. To date the focus has
mainly been on ‘simple’ mononucleosomes positioned on short
stretches of DNA. The next level of chromatin complexity is the
‘beads on a string’ structure. As single molecule techniques
provide precise force and distance measurements they can also
be applied to study higher order chromatin structural transitions,
even with megabase-sized nucleosomal arrays or native
chromosomes . Single molecule techniques hold the
promise that they will reveal how the structure of nucleosomal
arrays is changed by individual chromatin remodeling com-
plexes and what histone moieties are engaged by ATP-
dependent remodellers. Furthermore, they can also be applied
to address the role of individual subunits and their domains.
Finally, we expect single molecule techniques to shed light on
the structural consequences that histone variants and post
translational modifications have on chromatin higher order
folding and remodeling.
3. Comparison of S. cerevisiae chromatin remodeling
complexes: a proteomic perspective
ATP-dependent chromatin remodeling is performed by
multi-subunit complexes. Subunits are generally defined as
proteins that physically interact with each other in the context
of biochemically stable complexes. To date, application of
mass spectrometry on purified complexes has uncovered the
identity of most subunits of the budding yeast chromatin
remodeling complexes. Besides core subunits, the complexes
transiently associate with other proteins, so-called interactors.
Here we compare the different ATP-dependent remodeling
complexes from a proteomic perspective. We used data from
unbiased yeast two-hybrid and mass spectrometry coupled
affinity screens that is available in public protein interaction
databases and in the literature [73–76]. Furthermore, co-
immunoprecipitation data have also been incorporated in the
BIOGRID database (http://www.thebiogrid.org/, ). To
visualize and analyze protein–protein interaction networks
we employed the user-friendly Osprey program  that is
freely available on the Biogrid website and we encourage
interested readers to use this program to interactively browse
the Osprey networks displayed on the figures (see on-line
The present approach has the drawbacks that in the genome-
wide proteomic screens true positive interactions may have
been missed and that false positive interactions may be present.
We do expect that future proteomic surveys will rely on more
sophisticated bioinformatic approaches that result in more
accurate interactomes than what we present here, where all
physical interactions were considered without applying any
type of filter or weighting matrix.
The interactomes of the budding yeast Sth1, Snf2, Ino80,
Swr1, Isw1, Isw2, Chd1 and Yfr038w ATPases are shown in
Fig. 1. We clustered interacting proteins with overlapping gene
ontology (GO) process annotations to survey the different
processes in which remodellers play a role. We colored edges to
highlight interactors that are subunits of the ATPase complexes
according to functional interdependence and biochemical
criteria such as co-elution of the subunits by ion exchange
and size exclusion chromatography.
3.1. DDM1 subfamily; involvement in DNA methylation and
Orthologs of DDM1 are found throughout the eukaryotes,
like yeast (Yfr038w), mouse (Lsh) and human (HELLS). In
Arabidopsis thaliana DDM1 is required to maintain DNA
methylation patterns . In A. thaliana ddm1 mutants,
transposons and pericentromeric satellite repeats are no longer
silent, which leads to the occurrence of new phenotypes at high
frequencies through transposon induced mutations [18,78,79].
Moreover, small RNAs involved in post-transcriptional silen-
cing are strongly reduced over transposons and satellite repeats
. Finally, the level of pericentric H3K9 methylation
(heterochromatin) is greatly decreased while the level of
H3K4 methylation (euchromatin) is increased. Demonstration
of plant DDM1-mediated chromatin remodeling remains
restricted to in vitro assays, showing ATP-dependent nucleo-
some repositioning on a short DNA fragment . Future
studies should reveal how DDM1 links nucleosome remodeling
to DNA methylation. In S. cerevisiae however, Yfr038w
activity cannot involve small RNA mediated RNA degradation
Fig. 1. SNF2 ATPase interactome. All the yeast factors known to interact with the SNF2 ATPases Sth1, Snf2, Isw1, Isw2, Chd1, Ino80, Swr1 and Yfr038w were
collected from the Biogrid website, as well as from original research publications. Bona fide subunits of the ATPase complexes (see text for details) are connected to
their respective ATPases by bold edges. Proteins known to be subunits of two complexes were put at the interfaces between two wheels. Color scheme; nodes are
colored according to current biogrid (http://www.thebiogrid.org/) gene ontology biological process annotation as indicated in the legend (Nog1 and Noc2 are the two
‘ribosome biogenesis’ nodes). The occurrence of each GO term in this network versus the entire available physical interaction network is indicated between brackets.
Letter types are blue for sequence-specific DNA-binding factors, red for protein kinases and purple for known ubiquitin- and SUMO-metabolism nodes. The PDF
version of this diagram can be queried for text strings using the ‘ctrl’ F search function.
156J.J.F.A. van Vugt et al. / Biochimica et Biophysica Acta 1769 (2007) 153–171
DNA Metabolism (74/221)
DNA Repair (11/164)
DNA recombination (1/97)
DNA Replication (4/89)
DNA Damage Response (7/87)
Ribosome biogenesis (2/16)
RNA processing (19/385)
Protein Biosynthesis (14/329)
Signal Transduction (2/213)
Cell Cycle (13/230)
Cell Organization and Biogenesis (19/610)
Carbohydrate Metabolism (1/124)
Stress Response (8/169)
Protein Amino Acid Phosphorylation (7/137)
Protein Transport (9/367)
Protein Degradation (4/220)
157J.J.F.A. van Vugt et al. / Biochimica et Biophysica Acta 1769 (2007) 153–171
nor DNA methylation as budding yeast has not been reported to
harbor these features. The only interactor of Yfr038w is Srb4
(Fig. 1), a subunit of the mediator complex that functions in
transcription regulation , a fact that may perhaps suggest a
role for Yfr038w in transcription.
Mouse Lsh co-localizes and immuno-precipitates with
heterochromatin protein 1 (Hp1), suggesting that Lsh is
involved in pericentric heterochromatin formation . Never-
theless, Lsh also associates with other chromosomal sequences
. Recently, it was shown that Lsh cooperates functionally
and associates physically with the de novo DNA methylases
Dnmt3a and Dnmt3b . Finally, analysis of explanted Lsh−/−
mouse embryonic ovaries showed that Lsh is essential to
complete homologous chromosome synapsis . Whether this
phenotype is due to loss of DNA methylation at interspersed
retroviral elements and centromeric tandem repeats, or to an
inability to remodel nucleosomes for meiotic factors involved in
homolog synapsis remains to be established . Whether yeast
Yfr038w plays a role in meiosis is as yet unknown.
3.2. The SWI2/SNF2 subfamily: roles in transcription, DNA
repair and sister chromatid cohesion
The first ATP-dependent chromatin remodeling complex to
be described was the yeast SWI/SNF complex [85,86]. This
complex was identified through analysis of mutants that showed
similar pleiotropic phenotypes, in addition to being impaired for
mating type switching (SWI) and for sucrose metabolism under
non-fermenting conditions (SNF). Subsequently, mutations
were isolated that could suppress these phenotypes and most
of these affected the capacity of chromatin to repress
transcription . Despite the broad range of cellular processes
that depend on SWI/SNF, this twelve subunit complex is not
essential for viability and at 100–200 copies per cell it is not
very abundant [88,89]. In rich growth medium, proper
expression levels of 5% of S. cerevisiae genes strongly depend
on SWI/SNF [90,91]. This may be an underestimate of the SWI/
SNF targets as many SWI/SNF-dependent genes are regulated
by changes in environmental/growth conditions and would not
be expressed under the conditions where the microarray
experiments were conducted. The role of SWI/SNF in gene
transcription was further confirmed genetically as it was shown
that SWI/SNF becomes essential to cells that lack the H3/H2B
histone tail acetyltransferase Gcn5 or that carry mutations in
mediator subunits . SWI/SNF also plays a role in nucleotide
excision repair and double strand break repair, although the
exact nature of its involvement in DNA repair is still under
In 1996, a second SWI/SNF-type complex was purified on
the basis of sequence similarity with SWI/SNF subunits and it
was prosaically named ‘Remodels the Structure of Chromatin’
(RSC) . RSC contains 17 subunits of which ten are
essential. Five RSC subunits are paralogs of SWI/SNF subunits
and three are shared with SWI/SNF (Fig. 1). RSC is about ten-
fold more abundant in the cell than SWI/SNF and comprises
two isoforms (Campsteijn et al., submitted, [88,97]). RSC has
been implicated in transcription regulation , sister chroma-
tid cohesion , chromosome stability [100,101] and DNA
Although RSC and SWI/SNF show a high degree of
structural similarity, they differ at the phenotypic level .
SWI/SNF appears to antagonize chromatin assembly since
point mutations in histones that destabilize histone octamer–
DNA interactions suppress swi/snf phenotypes . On the
other hand, multiple mutations in chromatin components that
suppress swi/snf phenotypes actually exacerbate rsc pheno-
types, suggesting that RSC collaborates with chromatin re-
assembly factors, such as Spt6 that reforms chromatin in the
wake of elongating polymerases [102,104,105], likely in a
reaction that is coupled to histone acetyltransferase activity
The Swi2/Snf2 ATPase interaction network can be divided
into several clusters based on associated GO processes (Fig. 1).
The largest cluster comprises the twelve subunits (Fig. 1, bold
pink edges). Taf14 is not only a subunit of SWI/SNF, but also of
INO80, TFIID, TFIIF, the mediator complex and the Sas3
histone H3 acetyltransferase-based NuA3 complex that recog-
nizes trimethylated histone H3 K4 . This impressive list of
complexes indicates that Taf14 is a specialized regulatory factor
that functions in chromatin remodeling in a process that is
tightly coupled to RNA polymerase II mediated transcription
initiation. However, the actual nature of Taf14 activity is as yet
undefined. The Taf14 paralog Sas5 and a third yeast factor,
Yaf9, all bear a so-called YEATS domain that is probably
involved in binding histone H3 . Sas5 is a subunit of the
H4 K16 acetyltransferase SAS-I and Yaf9 is a subunit of both
the NuA4 H4/H2A acetyltransferase and of the SWR1 complex
[109,110]. Since genetic analyses showed that cells lacking two
of the three YEATS domain-containing proteins were very sick
or unviable , Taf14, Sas5 and Yaf9 likely play overlapping
roles in RNA polymerase II transcription at the level of
inscription, erasure and interpretation of epigenetic marks that
discriminate silenced genes from poised genes.
Another cluster of Swi2/Snf2 interactors consists of the
sequence specific DNA binding transcription factors, Hir1,
Hir2, Yap1, Hap4 and Gcn4  (Fig. 1), that are likely to
recruit the SWI/SNF complex to their targets. Hir1 and Hir2 are
involved in restricting histone gene expression to S-phase.Hap4
is part of the yeast equivalent of the NF-Y CCAAT factor that is
a global regulator of respiratory gene expression and Gcn4
activates amino acid biosynthesis genes.
The yeast SUMO homologue, Smt3, and the Rio2 and Mck1
kinases will be discussed later in this review.
The main cluster in the Sth1 interactome contains the 17
RSC subunits (Fig. 1, bold blue edges). Rsc1, Rsc2 and Rsc4
harbor tandem bromodomains, which together with the
bromodomain in Sth1, endow RSC with seven of the fifteen
bromodomains in the yeast genome. We have found that
depletion of Sth1 deregulates about 40% of all yeast open
reading frame mRNAs, suggesting that RSC plays a unique role
in maintenance of transcriptional homeostasis (Özdemir et al,
unpublished data). This finding may relate to a finding by
Soutourina et al. (2006) who showed that the C-terminus of the
Rsc4 subunit interacts with Rpb5, a shared subunit of RNA
158J.J.F.A. van Vugt et al. / Biochimica et Biophysica Acta 1769 (2007) 153–171
polymerases I, II and III . Indeed, RSC stimulates passage
of RNA polymerase II through a nucleosome and this process is
strongly stimulated by SAGA and NuA4, suggesting that
histone acetylation recruits RSC . A role for Sth1 in
transcription is further supported by interaction with a cluster of
general transcription factors including the TATA box binding
factor, Spt15/Tbp1, Taf14 and the TFIIH associated Met18/
Similar to SWI/SNF, DNA sequence-specific transcription
factors are found in the RSC interactome; Rtg1, Gcn4 and Crt1.
Rtg1 is a transcription activator of genes involved in inter-
organelle communication [113,114]. Crt1 also binds to Snf2,
Isw1 and Isw2 (Fig. 1), and represses genes that have to be
induced in the course of the DNA damage response .
A unique set of Sth1 interactors comprises the cohesin
subunits Smc1, Smc3, Mcd1 and Scc3 that are involved in
chromosome cohesion and double strand DNA break repair
[116,117]. RSC is indeed required for cohesion of sister
chromatids along the chromosome arms, but not at centromeres
SWI/SNF and RSC share Arp7 and Arp9, two actin related
proteins (ARPs). Interestingly, ARPs are also found in INO80,
SWR1 as well as in the Esa1 HAT-based NuA4 histone
acetyltransferase complex also known as Tip60 in humans
. The role of ARPs in chromatin remodeling has been
proposed to involve chaperone activity  and binding to
chromatin harboring specific histone post translational mod-
ifications . Notably, neither Isw1, Isw2 nor Chd1 appear to
contact ARPs (Fig. 1).
3.3. The ISWI subfamily: association with transcription
regulation and DNA replication
The ISWI (Imitation of Switch) subfamily comprises two
ATPases, Isw1 and Isw2, both estimated to be present at ∼1500
copies per yeast cell . They are present in a variety of
complexes that have multiple functions, including transcription
factor displacement, promoter nucleosome remodeling, tran-
scription elongation and replication .
Isw1 has been reported in two complexes; in ISW1a with
Ioc3 and in ISW1b with Ioc2 and Ioc4 [122,123]. Recent
findings indicate that the ISW1a complex is involved in
nucleosome repositioning at promoters to block transcription
[68,123,124]. This repositioning can impede binding of the
TATA box binding protein Tbp1/Spt15 . Upon transcrip-
tion initiation, ISW1b predominantly re-positions nucleosomes
in the coding region, which in turn regulates RNA polymerase
to shift into the elongation phase [121,123]. Thus Isw1 can
function in both transcription repression and activation, given
the complex it resides in.
Inspection of the interactome of Isw1 (Fig. 1) reveals a
cluster of proteins that function in transcription regulation,
including the Rpc40 subunit of RNA polymerases I and III.
Another transcription-linked interactor is Sin3, a partner of the
Rpd3 histone deacetylase that functions as repressor of
promoter activity when associated with the large Rpd3
complex, while Sin3 participates in transcription elongation
when part of the small Rpd3 complex . These findings
illustrate the fact that individual nucleosome remodeling
enzymes can have dual outcomes in transcription regulation,
depending on the location of the nucleosome and the
transcriptional phase [125,126].
A notable Isw1 interactor is Esc8, which shows a high degree
of homology with Ioc3. Esc8 has been implicated in telomere
and mating type loci silencing . Indeed, Esc8 interacts with
Sir2, the founding member of the Sirtuin class of NAD-
dependent histone deacetylases that is an established sub-
telomeric domain silencing factor . Isw1 also interacts
with Pri1, a subunit of DNA primase, with topoisomerase I
(Top1) that relaxes supercoiled DNA and with Mini-Chromo-
some Maintenance 5 (Mcm5), a subunit of the MCM helicase
complex that melts DNA in front of replication forks.
Intriguingly, Mcm5 has also been shown to play a central role
in epigenetic gene silencing in S. cerevisiae .
Isw2 binds Itc1, a factor involved in transcription repression
, whose targets include early meiotic genes. This complex
functions in parallel to the Sin3/Rpd3 histone deacetylase
complex [130,131] and both complexes are targeted by the
The second described Isw2 complex is known as the yeast
chromatin accessibility complex (yCHRAC) and comprises
Dls1 and Dpb4 in addition to Isw2 and Itc1 . yCHRAC has
been implicated in the maintenance and inheritance of telomeric
silencing marks during S-phase . Dpb4 functions with
Dpb3 and together they may form a H2A–H2B-like dimer
through their histone-fold motifs , see also . Notably,
Dbp4 and Dpb3 are also subunits of DNA polymerase epsilon,
whose dsDNA binding and transcription silencing activity they
appear to promote [132,133].
The link to replication is further backed by the presence
of three factors in the Isw2 interactome that play roles in
DNA replication as well as in gene expression regulation,
namely Abf1, Mcm1 and Rap1. Abf1 (autonomous replicat-
ing sequence (ARS)-binding factor 1) is a DNA binding
protein that is involved in transcriptional activation, gene
silencing, DNA replication and chromatin remodeling
[135,136]. Abf1 binds to over 360 loci in the genome
 including a large number of ARSs. Mcm1 binds to
another subset of ARSs and also stimulates DNA replication
initiation . Rap1 has been implicated in replication
origin regulation, mating type silencing and telomere length
determination as well as being crucial for transcription
activation of ribosome subunit genes [135,139]. Altogether,
Abf1, Mcm1 and Rap1 target ∼700 yeast loci , and it
is striking that these factors cluster with Isw2 as they share
the uncommon properties of functioning in transcription
activation and repression, and in the regulation of replication
origins. Whether Ume6, the fourth transcription regulator of
the Isw2 interactome, directly impinges on replication origins
or replication factor dependent transcriptional silencing is
The Isw2 interactome further contains the nucleolar
proteins, Noc2, Noc3, Nop4, Nop7 and Nug1 (Fig. 1), that
159J.J.F.A. van Vugt et al. / Biochimica et Biophysica Acta 1769 (2007) 153–171
also link to nucleolar factors in the Snf2 and Chd1 interactomes
(see also the interactive version of Fig. 2, provided as
supplemental data). These proteins are implicated in rRNA
processing and ribosome biogenesis and maturation .
However, Noc3 can also bind to chromatin to promote the
association of DNA replication factors and stimulate replication
initiation . How Isw2 impinges on DNA replication and
epigenetic inheritance of transcriptional silencing remains to be
elucidated. An explanation as to why replication and transcrip-
tion silencing are coupled may perhaps be found in the fact that
ongoing transcription can antagonize replication origin usage
3.4. CHD1; a role in chromatin remodeling during
The CHD subfamily includes many paralogs in higher
eukaryote species . The sole yeast CHD1 subfamily
member,Chd1, is present in yeast at about 1600 copies per yeast
cell . Early microarray and ChIP data revealed that loss of
Chd1 deregulates transcription of 2–4% of all yeast genes and
that it is present on chromatin throughout the genome .
is casein kinase—Cka1, Cka2, Ckb1 and Ckb2 . There are
more than 10,000 casein kinase molecules per cell  and they
have a variety of roles and substrates in the nucleus. Drosophila
Mi-2, a distant relative of Chd1, also associates with casein
kinase and it can be phosphorylated by this kinase .
Furthermore, dephosphorylated dMi-2 displayed more nucleo-
some remodeling activity than the phosphorylated form .
The possibility exists that the interaction between chromodo-
main-bearing SNF2-type ATPases and casein kinases serves not
kinase to specific loci at specific times.
Chd1 physically interacts with 3 distinct RNA elongation
complexes; PAF (Rtf1, ∼6500 per cell), FACT (Pob3, Spt16;
∼20,000 per cell) and Spt4/5 complexes (∼2300 per cell) (Fig.
1, see also ). Interestingly, all of these elongation factors
interact physically with casein kinase . This coherent set of
physical interactions very strongly implicates Chd1 in tran-
scription elongation and mRNA processing [144,147].
Recently, in vitro transcription experiments clarified some of
the relationships between SWI/SNF remodeling, p300 mediated
acetylation, the elongation factors FACT and PAF and histone
H2B monoubiquitylation . The ligand inducible retinoic
acid receptor was shown to rely on recruitment of SWI/SNF and
the acetylase p300 to generate a RNA polymerase II preinitia-
tion complex. FACTassociated with stalled RNA polymerase II
after which PAF was recruited, followed by H2B ubiquitylation
factors. Interestingly, in this reconstituted system, monoubiqui-
tylation of H2B substantially increased transcription and
spreading of PAF, FACT and the ubiquitylation factors into
the transcribed region . In humans, H2B monoubiquityla-
tion also leads to in vivo histone H3 K4 trimethylation
[149,150], which recruits human Chd1 . However, since
Chd1 was not included in the Pavri study , Chd1's
contribution to elongation per se remains to be determined.
Fifteen out of thirty-two Chd1-associated proteins are
subunits of both the yeast SAGA and SLIK complexes,
including five TBP associated factors Taf5, Taf6, Taf9, Taf10
and Taf12 that are also subunits of the TFIID complex (Fig. 1,
bold orange edges, [152,153]). SAGA and SLIK are 1.8 MDa
Gcn5-dependent histone H3/H2B acetyltransferase complexes.
Ubp8 is a deubiquitylase that is common to both complexes.
Ubp8 plays an important role in mRNA biogenesis
[146,152,154] as it removes the H2B monoubiquitylation
mark imposed by Rad6 , presumably in a ubiquitylation/
deubiquitylation cycle that accompanies transcription elonga-
tion in order to prevent inappropriate intragenic transcription
initiation [105,125]. Intriguingly, H2B monoubiquitylation also
lies at the crossroads of transcription and DNA repair .
Finally, besides an as yet undefined role in transcription
elongation, Chd1 might also play a role in DNA replication
dependent processes because two of its chief interactors, the
Spt16 and Pob3 subunits of yeast FACT, have also been
implicated in replication [157,158].
3.5. INO80 subfamily: Histone octamer restructuring in DNA
repair and transcription
Like the SWI2/SNF2 and ISWI subfamilies, the INO80
subfamily has two representative ATPases, Ino80 and Swr1,
estimated at 6850 and 656 copies per cell, respectively .
They are the catalytic subunits of the INO80 and SWR1
complexes [29,50,159]. Over the last few years INO80 and
SWR1 have been shown to play important roles in DNA
repair and transcription regulation. Both complexes share
actin (Act1) and Arp4 and both harbor exclusive ARPs (Fig.
1). Arp4 interacts with histone H2A that is phosphorylated on
serine 129 at double strand DNA breaks (DSBs) .
Consequently Arp4 could mediate association of INO80,
SWR1 and of the NuA4 H4/H2A acetyltransferase complexes
with a 10 kb domain surrounding DSBs . Nhp10, an
INO80 subunit, has also been reported to bind phosphorylated
H2A . Since the recruitment of SWR1 and INO80 is
preceded by NuA4, this complex possibly helps the
recruitment of SWR1 to DSBs by means of Bdf1 binding
to histones acetylated by NuA4 .
Like Act1 and Arp4, Rvb1 and Rvb2 are shared by INO80
and SWR1. They are part of the AAA+ super family of ATP
binding proteins [29,161] and belong to a group of ATP-
dependent DNA helicases that catalyze four-way junction
migration in bacteria. Surprisingly, the Drosophila Rvb1 and
Rvb2 orthologs, Pontin52 and Reptin52, have been shown to
play opposite roles in transcription regulation downstream of β-
catenin . Whether this took place via their role in the fly
equivalents of NuA4, SWR1 or INO80 remains an open
question though. The yeast Rvb factors are necessary for the
catalytic activity and for proper assembly of the INO80 complex
. Microarray data has shown that Rvb1 and Rvb2 regulate
about 500 yeast genes that overlap with the 153 Ino80 targets
. INO80 may therefore function in transcription regulation
as well as in DSB repair, a notion that is strengthened by the
presence of Taf14 in INO80 (Fig. 1).
160 J.J.F.A. van Vugt et al. / Biochimica et Biophysica Acta 1769 (2007) 153–171
The Ino80 ATPase forms a complex with 13 other proteins
(Fig. 1, bold mustard edges), of which Ies4, Ies5 and Ies6 co-
purified at low salt concentration only . Arp5 and Arp8
have also proven to be necessary for the enzymatic activity of
INO80  and they have been observed to bind directly to
histone cores .
Taf1 is located in the Ino80 interactome (Fig. 1). Taf1 is not
part of SAGAandis better known as TAFII250 ,the largest
subunit of TFIID. Besides being a kinase , Taf1 may also
be an acetyl transferase . Despite its prominent position in
the basal transcription factor complex TFIID, the role of Taf1
kinase activity is unknown and whether Taf1 interacts
functionally with INO80 is unknown. Taf1 also connects to
the ISW1a subunit Ioc3 that itself connects back to Rvb1 and
Rvb2. Taf1 might thus modulate ISW1a interaction with
promoters and RNA polymerase II, in parallel to the switch in
phosphorylation of serines 5 and 2 of the heptad repeat-bearing
carboxy-terminal tail of subunit 1 of RNA polymerase II that
accompanies promoter clearance by RNA polymerase II
Finally, Fkh2, a transcription factor of the forkhead family
that regulates the cell cycle and pseudohyphal growth, is found
in the Ino80 interactome, further strengthening the link between
INO80 and transcription regulation.
SWR1 forms a complex with 14 other proteins, including
Htz1 (Fig. 1, bold red edges). Htz1 is the yeast ortholog of the
variant histone H2A.Z. Htz1 is incorporated by SWR1 into
nucleosomes in an ATP-dependent nucleosome remodeling/
restructuring reaction [49,50,110,167,168]. Various ChIP
experiments have shown that Htz1 is highly enriched at
promoter sites. This does not necessarily mean transcription is
‘ON’ however, since promoters of inactive genes are also
enriched with Htz1 [168,169]. Htz1 incorporation may there-
fore be interpreted as an epigenetic mark for ‘poised’ genes. The
mechanism of this process is not yet fully understood but
histone tail acetylation seems to play a role [5,169] since NuA4
complex acetylates Htz1 K14 after its assembly into chromatin
by SWR1 [167,170]. How the SWR1 complex is recruited to
chromatin is not known, though the SWR1 subunit Bdf1 might
be involved. Bdf1 is important for transcription initiation at
TATA-containing promoters and is also part of the TFIID
complex . Bdf1 contains two bromodomains which may
recognize acetylated histone tails and thus recruit SWR1 to
acetylated histone octamers, much like the bromodomains of
SWI/SNF, RSC and SAGA [13,169,172]. As outlined in the last
section of this review, genetic evidence strongly implicates
SWR1 and Htz1 incorporation in epigenetic control of
3.6. Remodeling and double strand break repair
Both INO80 and SWR1 complexes can be recruited to
chromosome breaks by binding Arp4 that is bound to serine 129
phosphorylated H2A. The role of SWR1 in DNA repair may
relate to acetylation of the H4 N-terminal tail that has been
implicated in Mec1 promoted DSB repair , and, as this
complex is known to restructure octamers to incorporate H2A.Z
histone, it is tempting to think that SWR1 also restructures
nucleosomes that harbor phosphorylated H2A. Indeed, Papmi-
chos et al. showed that Ino80 is required to maintain
phosphorylated levels of H2A at DSBs and that Ino80
antagonizes Swr1 mediated Htz1 incorporation . They
proposed that INO80 and SWR1 catalyze H2A variant
exchange cycles that control cell cycle re-entry after DNA
damage. Essentially, in this model, INO80 allows cells to
segregate broken chromosomes by maintaining the specialized
phosphorylated H2A harboring chromatin structures surround-
ing DSBs, thus permitting checkpoint adaptation . INO80
is also known to be required for H3 and H2B loss in ∼40 kb
domains surrounding the DSB . Recently, Murr et al.
(2006) showed in a set of elegant experiments that similar
principles apply to mammalian cells , where the Tip60
complex appears like a mixed NuA4–SWR1 complex . On
the other hand, SWI/SNF and RSC are also recruited to break
sites, whereby SWI/SNF associates to the breaking sites before
RSC . SWI/SNF was therefore proposed to act during the
strand invasion stage and RSC when the invading strand is
extended . Notably, an additional early role in DSB repair
was proposed for RSC in a separate study . Altogether,
SWI/SNF, INO80, RSC and SWR1 appear to work in concert
with histone kinases, acetyltransferases and deacetylases to
promote DSB repair [178,179]. To what extent, if any,
chromatin remodeling for repair differs from remodeling for
transcription is unknown. Furthermore, when homologous
recombination ensues, finding out whether and how remodeling
differs at the broken and at the intact chromosome (the sister
chromatid in S/G2/M or on the homolog in diploids) are
interesting issues that remain largely unexplored.
4. Interrelations between the different ATPase interactors
The currently available yeast protein physical interaction
network we employed contains 5106 protein nodes and 30418
binary interactions . The SNF2-type ATPase interactor
network contains 190 protein nodes (Figs. 1–3). In the
previous section we compared the different S. cerevisiae
ATP-dependent chromatin remodeling complexes from an
ATPase-centered point of view (Fig. 1). This provided insight
in the processes these complexes drive individually and
collectively and indicated extensive interactions with the
enzymes that add post-translational modifications to their
substrate, namely histones and polymerase associated factors.
More insight on proteomic and genetic interrelations of the
different subfamilies of ATP-dependent chromatin remodellers
can be obtained from their subunits' interactions. To this end
the entire Biogrid physical interaction dataset was super-
imposed on the remodeller template of Fig. 1, after which all
newly added proteins were deleted, resulting in 1081
interactions among the 190 ATPase interactors (Fig. 2).
Even a cursory look reveals that many of the actual complex
subunits (Fig. 1, bold colored edges) indeed interact amongst
themselves (Fig. 2, bold colored edges). Below we highlight a
number of inter-complex interactions that were not discussed
161J.J.F.A. van Vugt et al. / Biochimica et Biophysica Acta 1769 (2007) 153–171
4.1. Histones and high mobility group proteins
In the entire physical data set that is currently available,
histones interact with 75 (H3) to 226 (H4) factors. H2A.Z
interacts with 35 factors and the H1 linker histone homolog
(Hho1) interacts with 6 other proteins. The SNF2-ATPase
interactome displayed in Fig. 1 encompasses from 27% (H4) to
39% (H2A) of the core histone interactors, indicating that about
one in three of all documented histone interactors interact
physically with one or more of the SNF2-type ATPase
interactors we consider in this review. In the case of Htz1,
63% of its interactors are represented.
A major class of abundant chromatin components are high
mobility group (HMG) proteins. They bind the minor groove of
DNA with limited or no sequence specificity, often bending
DNA [180,181]. Three HMG subfamilies can be distinguished:
HMGA (contains AT-hooks), HMGB (contains Box domains)
and HMGN (binds to nucleosomes) . They function in
chromatin remodeling, transcription, replication, and recombi-
nation [180,182–184]. In S. cerevisiae only the HMGB proteins
are present. There are 6 HMGB proteins known in yeast of
which five, i.e. Nhp6a, Nhp6b, Nhp10, Hmo1 and Abf2, are
represented in the SNF2-ATPase network. In terms of their own
interactors; 85% are present in the network for Nhp10 (in
Fig. 2. Physical interactions amongst ATPase interactors. Physical interactions involving the 190 protein nodes shown on Fig. 1 were downloaded from the Biogrid
website http://www.thebiogrid.org/). We included the following types of interactions: Affinity capture coupled to mass spectrometry, affinity capture coupled to
western analysis, co-purification, yeast two hybrid and reconstituted complex. Nodes for known bona fide subunits as well as interactions within each individual
ATPase interactome are highlighted in the same colors. Nodes for proteins that are subunits of multiple ATPase complexes are colored black. The interactions between
the Isw1 ATPase and RSC subunits are indicated by green edges, Casein kinase 2 interactions by Bordeaux red edges, TAF14 interactions in brown and TBP
interactions in black. The remaining binary interactions are shown in grey. This network has 190 nodes and 1081 binary interactions, of which 657 are located within
any one of the ATPase's interactomes. An interactive version of this network is available as supplemental material in the form of an Osprey data file.
162 J.J.F.A. van Vugt et al. / Biochimica et Biophysica Acta 1769 (2007) 153–171
INO80, RSC, ISW1 and ISW2) [185,186], 61% for Abf2 (in
RSC, ISW1, SWI/SNF and CHD1) [187,188], 59% for Nhp6a/b
(in RSC, ISW1, ISW2, and FACT and Spt5 in the CHD1)
[183,184,189–191], while 25% of Hmo1's interactors (ISW1,
cohesin subunits, RSC and SWR1) [192,193] are represented.
As a random sample of 190 factors would be expected to yield
3.7% of any factors' interactors, it is clear that yeast HMG
proteins are intimately linked to ATP-dependent chromatin
remodeling, much like the histones, playing diverse roles in
Smt3 is the yeast equivalent of SUMO. It is represented here
because several groups reported a large set of yeast proteins that
are conjugated to Smt3 [194–197]. Altogether, 265 yeast
proteins have been reported to be modified by Smt3, equivalent
to ∼5% of the yeast proteome. The remodeller interactome in
Fig. 2 contains 33 Smt3-linked nodes, indicating a 3-fold
enrichment of sumoylated factors in this network. The Isw1
interactome is most enriched with Smt3 interactors (9/31),
including Isw1 itself, and the HMGfactor Hmo1. Second comes
RSC (9/44), then SWI/SNF (5/30, including Snf2 itself) and
INO80 (5/24), followed by CHD1 (3/32), ISW2 (2/17) and
SWR1 (1/19). Exactly what happens to Smt3-conjugated
proteins remains to be elucidated and this may well differ for
different factors, although a functional link with transcription
repression appears to be a recurrent theme [198,199]. As
sumoylation, ubiquitylation, acetylation and methylation all
target lysines, there are some obvious opportunities for
regulatory interactions when the same lysine is a target for
multiple modifications, as has been proposed for histone
4.3. Synthetic interactions between tubulin and the RSC and
A link between RSC and microtubule mediated chromosome
(re)capture is suggested by the fact that several RSC subunits,
but no other SNF2-type remodellers, display synthetic lethal
interactions with alpha-tubulin  (Fig. 3, purple edges). On
the other hand, Htz1 is recruited to centromeric regions in a
SWR1-dependent manner  and the SWR1-specific sub-
units Htz1, Arp6, Swc1 and Swc6 show synthetic phenotypes
with gamma-tubulin, which nucleates microtubules . RSC
and SWR1 therefore appear to perform distinct microtubule
apparatus-related tasks. Whether these observations reflect
indirect transcriptional responses or direct connections to the
chromosome arms (RSC) or the kinetochore (SWR1) is
currently unknown .
4.4. RSC and ISW1 relate to cohesin and to the TATA box
binding factor TBP
One of the interrelationships already visible in Fig. 1 that was
not discussed previously is the interaction between Isw1 and
eight RSC subunits (Fig. 2, green edges). Furthermore, Ioc3
contacts Rsc8, thereby linking the promoter nucleosome
targeted ISW1a complex to RSC function. Since human ISWI
has been implicated in the establishment of cohesin complexes
, sister chromatid cohesion could be a process that
connects Isw1 to RSC in yeast. Also Isw2 interacts with
cohesin, by means of the cohesin subunit Scc3 (Fig. 1). With
their link to cohesin [99,205], RSC and ISWI complexes may
therefore be involved in the genesis of higher order chromatid
organization. Furthermore, the relation between RSC and Isw1
appears to involve TBP, since a member of the SNF2 ATPase
family called Mot1 in yeast and BTAF1 in higher eukaryotes
, interacts with Ioc3, Isw1, Rsc3 and Sth1. Mot1 is thought
to act as a mobilizing factor for TBP .
TBP interacts with 39 factors in Fig. 2 (32% of all its
interactors, black edges on Fig. 2). Taf14 is another promiscu-
ous factor in the remodeller network. It also has 39 interactors
(50% of all its interactors, brown edges in Fig. 2). Besides
contacting TAFs found in SAGA and TFIID, TBP and Taf14
have different interactors in the remodeller network; TBP being
primarily a RSC interactor (11/17 subunits) while Taf14
primarily interacts with SWI/SNF (10/12 subunits) and
INO80 (all subunits). What these differences between TBP
and Taf14 mean at the functional level remains to be
understood. Searching the 78 physical partners of Taf14 for
‘repair’ yielded one node, Ino80, while 70 had a ‘transcription
process’ annotation. This strongly suggests that Taf14 does not
participate in DNA repair in a specific way.
4.5. Implication of Casein kinase 2 in chromatin remodeling
Casein kinase 2 (CK2) has a variety of functions in the cell,
including the regulation of transcription and cell growth. CK2
interacts with 24 of the 190 factors in the SNF2 interactome
(Fig. 2, Bordeaux edges), versus 127 in the entire yeast data set,
representing a 5-fold ‘enrichment’ of CK2 interactors in this
network. As discussed above, Chd1 andelongation factors Pob3
and Spt16 (FACT), Rtf1 (PAF), and Spt5 all appear to
extensively contact CK2  and are therefore likely
substrates for CK2. The SWR1 complex may also be targeted
by CK2, via Swc5 and Bdf1. Bdf1 plays a role in transcription
initiation by binding to TFIID and is indeed known to be
phosphorylated by the CK2 complex . As previously
noted, Bdf1 interacts with TBP, and TBP is also connected to
the CK2 complex (Fig. 2).
In the remodeller network, CK2 is also connected to
nucleolus-linked processes through interaction with the eight
nucleolar proteins Pwp1, Nop1, Nop4, Nop7, Nug1, Noc3,
Nog1 and Rrp5. In light of the replication connection of Noc3, it
remains to be seen whether the nucleolar CK2 links involve
intranuclear localization of ARSs to the vicinity of the nucleolus
and/or a direct function in ribosome biogenesis [89,140].
4.6. Presence of other enzymes in the SNF2-ATPase network
Besides casein kinase, the interactome displayed in Fig. 1
harbors four other kinases; Rim15, Mck1, Rio2 and Taf1.
Rim15 is a PAS kinase family member that controls yeast cell
163J.J.F.A. van Vugt et al. / Biochimica et Biophysica Acta 1769 (2007) 153–171
Mitochondria Not Determined
Nucleus + Nucleolus
Nucleus + Cytoplasm
Fig. 3. Synthetic genetic interactions within the SNF2 ATPase interactome. All the available synthetic genetic interaction involving the 190 nodes shown in Figs. 1 and
2weredownloadedfromthebiogridwebsite.We included syntheticlethalandsyntheticgrowthdefects.Syntheticlethalinteractionsdisplayedbyalpha-tubulin(Tub1)
are shown in purple, Nhp6 interactions in light brown, Swi2/Snf2 interactions are shown in black, those involving Rsc1 in light blue, those of Isw1 and its interactors
Lge1, Sin3 and Esc8 in several shades of green, those made by the Ino80 interactors Ies3 and Ies5 in shades of red, and those made by the PAF complexsubunit Rtf1 in
orange.Note the large numberof genetic interactions that converge on the SWR1complexand Htz1(see text for details). Colorscheme;nodesare coloredaccordingto
current knowledge as to the subcellular location of each protein , as indicated in the legend. This Osprey network has 190 nodes and 198 edges and available as
supplemental material in the form of an Osprey data file.
164J.J.F.A. van Vugt et al. / Biochimica et Biophysica Acta 1769 (2007) 153–171
entry into the resting phase and early meiotic gene expression
. It functions downstream of PKA and TOR and appears to
control the transcription factors Msn2, Msn4 and Gis1 . As
Rim15 is an interactor of the RSC Sth1 ATPase, phosphoryla-
tion of RSC at Msn2, Msn4 or Gis1 targetgenes may be a means
by which Rim15 controls transcription . Mck1 is a GSK3-
type kinase and it interacts with the Swi2/Snf2 ATPase. It
further interacts with Rsc1, Psa1 and Ume6 and has been
implicated in chromosome segregation and entry in meiosis
[211,212]. Rio2 is essential for processing the 20S pre-rRNA
into 18S . Its presence in the SWI/SNF interactome could
therefore reflect a so far unidentified link between SWI/SNF
and the nucleolus.
Besides kinases there are two phosphatase regulators in the
yeast chromatin remodeller interactome, both of which are
located in the Swr1 interactome, namely Sis2/Hal3 and Pil1.
Sis2 is the negative regulatory subunit of Ppz1 (protein
phosphatase 1) and connects to the alpha subunit of CK2
(Fig. 2). Ppz1 is important for the osmotic stress response .
It regulates intracellular pH and G1 progression, probably
because intracellular pH normally rises at this point of the cell
cycle . Whether Swr1 is a target for Ppz1 is unknown,
although as both Sis2 and Ppz1 are partially nuclear this is a
possibility. Pil1 is a component of eisosomes, which are large
immobile patch structures at the cell cortex associated with
endocytosis, and Pil1 may therefore appear as an inappropriate
node in the remodeller interactome. However, Pil1 interacts
with six SWR1 subunits and with Rtg2, a specific subunit of the
SAGA related SLIK H3/H2B acetyltransferase, as well as with
Rad1 and Rad59 DSB repair factors. Pil1 therefore has a high
probability of being a true SWR1 interactor. We speculate that
Pil1 connects chromatin remodeling to chromatin interactions
with the nuclear membrane and perhaps actin.
5. Genetic interrelations gleaned from large scale synthetic
So far we reviewed physical interactions of chromatin
remodeling complexes. As yeast has been a leading model
system for high throughput genetic analysis whereby ‘synthetic’
phenotypes are scored in double mutants, we superimposed the
available genetic data on the SNF2 ATPase network (Fig. 3).
A major drawback of this approach is that until recently,
essential genes were not included when screening for synthetic
phenotypes on a genome-wide scale, because viable deletions
alleles were used in such screens. The synthetic lethal or
sickness map shown in Fig. 3 therefore largely lacks inter-
actions involving the essential factors of the network.
There are obvious genetic interactions between the SWR1
interactome and the other subfamily interactomes. This suggests
that SWR1 drives a process, in a unique fashion that is
redundant with Rsc1 (Fig. 3, blue edges), Isw1 and its partners
Esc8, Lge1 and Sin3 (Fig. 3, green edges). The same applies to
the Rtf1 subunit of the PAF complex in the Chd1 interactome
(Fig. 3, orange edges) as well as the Ies3 and Ies5 proteins in the
Ino80 interactome (red edges) and the SAGA/SLIK complexes
(not colored in Fig. 3). Except for Spt20, all the SWR1 synthetic
genetic interactions are shared by Htz1, supporting the notion
that the main function of SWR1 is to restructure nucleosomes
by replacing H2A molecules with Htz1. However, absence of
Htz1 is also lethal in the absence of Swi2/Snf2 , while
SWR1 subunits do not show genetic interactions with Swi2/
Snf2 (Fig. 3). This suggests that Htz1 can be incorporated by a
SWR1-, and SWI/SNF-, independent mechanism, perhaps by
forming a complex with NAP1 [69,167].
Swi2/Snf2 in turn is essential when Gcn5 is missing and this
is likely due to redundancy of the bromodomains of both
proteins, that targets them to acetylated nucleosomes [13,111].
Interestingly, the Isw1 interactor Lge1 displays a total of 156
synthetic lethal interactions in a synthetic genetic interaction
network that contains 2629 factors and 13333 edges (Biogrid,
http://www.thebiogrid.org/). Of those, 17 are present in our
remodeller map and nine involve SWR1 subunits (Fig. 3).
Remarkably, 133 of the 156 lge1Δ synthetic lethal interactions
are shared with bre1Δ, and 101 are shared with rad6Δ, both of
which are central factors to H2B ubiquitylation [217,218].
Furthermore, Bre1 and Lge1 associate physically, strongly
suggesting that the redundant function of Lge1 and the SWR1
complex pertains to H2B ubiquitylation that leads to H3 K4 tri-
methylation . This in turn suggests that Htz1 incorporation
and H3 K4 methylation are two redundant epigenetic euchro-
matic marks in yeast.
What exactly happens when an ATP-dependent remodeling
enzyme engages nucleosomes is still not fully elucidated and
single molecule approaches hold much promise in this respect.
In our opinion, it will be important to compare multiple
remodellers side by side in the different single molecule set-ups,
so as to find out what aspects of their reaction cycles are shared
and which ones are unique. The next frontier in single molecule
studies will be to enable ex vivo studies, whereby native
chromatin is isolated and studied at the single molecule level.
subunit complexes that have a great diversity of interacting
proteins. Ahighdegree ofcooperationappears tooccuramongst
these complexes, whereby multiple ATPases and their subunits
are involved in different – though sometimes redundant – steps
of processes such as transcription initiation, elongation or DNA
replication. Furthermore, ATP-dependent remodeling com-
plexes specifically interact physically and functionally with
many enzyme complexes that modify histones at the post trans-
lational level, as well as with modified forms of nucleosomal
histones. Much remains to be discovered in yeast, still. One
outstanding question is whether subunits actually exchange
between complexes, an issue that is mechanistically even more
important when a subunit is shared by different types of
complexes, such as ARPs, Bdf1, Taf14, etc. A second emerging
issue concerns post-translational regulation of chromatin com-
plexes themselves. For example, we have identified 16
phosphorylation, 22 acetylation and 2 methylation sites on 13
RSC subunits (Campsteijn et al. submitted). It will be important
to discriminate modifications that globally affect complexes,
165J.J.F.A. van Vugt et al. / Biochimica et Biophysica Acta 1769 (2007) 153–171
such as those that would accompany complex assembly or cell
cycle progression, from those that are imprinted onto complexes
variables such as cell cycle progression and cell type in pro-
teomic and biochemical approaches will undoubtedly shine new
light on issues regarding functional redundancy, compositional
dynamics and targeting.
It is likely that much more insight will be gained from future
high throughput genetic screens. Quantitative analysisof colony
size , inclusion of hypomorphic alleles of essential factors
, as well as screens involving more than two mutations
should greatly help unravel redundant functions. Additionally,
synthetic genetic interactions can be identified under conditions
that challenge the ability of yeast to survive exposure to specific
drugs, such as those that result from DNA insults or mitotic
spindle defects. Finally, devising computational ways to
integrate high throughput chromatin IP, transcriptomic, proteo-
mic and genetic data should further afford detailed insight into
the function of chromatin components in medically relevant
fields of research .
We thank John van Noort for useful comments and Nevan J
Krogan for making submitted data available. We apologise to
those researchers whose papers we did not refer to due to space
constraints. JvV is supported by the Eurocores initiative
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at doi:10.1016/j.bbaexp.2007.01.013.
 K. Luger, A.W. Mader, R.K. Richmond, D.F. Sargent, T.J. Richmond,
Crystal structure of the nucleosome core particle at 2.8 A resolution,
Nature 389 (1997) 251–260.
 J.C. Hansen, Conformational dynamics of the chromatin fiber in solution:
determinants, mechanisms, and functions, Annu. Rev. Biophys. Biomol.
Struct. 31 (2002) 361–392.
 O.J. Rando, T.H. Chi, G.R. Crabtree, Second messenger control of
chromatin remodeling, Nat. Struct. Biol. 10 (2003) 81–83.
 B.D. Strahl, C.D. Allis, The language of covalent histone modifications,
Nature 403 (2000) 41–45.
 C.B. Millar, M. Grunstein, Genome-wide patterns of histone modifica-
tions in yeast, Nat. Rev., Mol. Cell Biol. 7 (2006) 657–666.
 L. Mohrmann, C.P. Verrijzer, Composition and functional specificity of
SWI2/SNF2 class chromatin remodeling complexes, Biochim. Biophys.
Acta 1681 (2005) 59–73.
 A. Saha, J. Wittmeyer, B.R. Cairns, Chromatin remodeling: the industrial
revolution of DNA around histones, Nat. Rev., Mol. Cell Biol. 7 (2006)
 J.M. Gottesfeld, K. Luger, Energetics and affinity of the histone octamer
for defined DNA sequences, Biochemistry 40 (2001) 10927–10933.
 M. Tomschik, H. Zheng, K. van Holde, J. Zlatanova, S.H. Leuba, Fast,
long-range reversible conformational fluctuations in nucleosomes
revealed by single-pair fluorescence resonance energy transfer, Proc.
Natl. Acad. Sci. 102 (2005) 3278–3283.
 J.A. Eisen, K.S. Sweder, P.C. Hanawalt, Evolution of the SF2 family of
proteins: subfamilies with distinct sequences and functions, Nucleic
Acids Res. 23 (1995) 2715–2723.
 M.R. Singleton, D.B. Wigley, Modularity and specialization in super-
family 1 and 2 helicases, J. Bacteriol. 184 (2002) 1819–1826.
 F. Winston, C.D. Allis, The bromodomain: a chromatin-targeting
module? Nat. Struct. Biol. 6 (1999) 601–604.
 A.H. Hassan, P. Prochasson, K.E. Neely, S.C. Galasinski, M. Chandy,
M.J. Carrozza, J.L. Workman, Function and selectivity of bromodo-
mains in anchoring chromatin-modifying complexes to promoter
nucleosomes, Cell 111 (2002) 369–379.
 T. Grune, J. Brzeski, A. Eberharter, C.R. Clapier, D.F.V. Corona, P.B.
Becker, C.W. Muller, Crystal structure and functional analysis of a
nucleosome recognition module of the remodeling factor ISWI, Mol. Cell
12 (2003) 449–460.
 A.J. Bannister, P. Zegerman, J.F. Partridge, E.A. Miska, J.O. Thomas, R.C.
Allshire, T. Kouzarides, Selective recognition of methylated lysine 9 on
histone H3 by the HP1 chromo domain, Nature 410 (2001) 120–124.
 R. Bakshi, T. Prakash, D. Dash, V. Brahmachari, In silico characterization
of the INO80 subfamily of SWI2/SNF2 chromatin remodeling proteins,
Biochem. Biophys. Res. Commun. 320 (2004) 197–204.
 K. Dennis, T. Fan, T. Geiman, Q. Yan, K. Muegge, Lsh, a member of the
SNF2 family, is required for genome-wide methylation, Genes Dev. 15
 A. Vongs, T. Kakutani, R.A. Martienssen, E.J. Richards, Arabidopsis
thaliana DNA methylation mutants, Science 260 (1993) 1926–1928.
 J. Brzeski, A. Jerzmanowski, Deficient in DNA methylation 1 (DDM1)
defines a novel family of chromatin-remodeling factors, J. Biol. Chem.
278 (2003) 823–826.
 L.A. Boyer, C. Logie, E. Bonte, P.B. Becker, P.A. Wade, A.P. Wolffe,
C. Wu, A.N. Imbalzano, C.L. Peterson, Functional delineation of three
groups of the ATP-dependent family of chromatin remodeling enzymes,
J. Biol. Chem. 275 (2000) 18864–18870.
 G. Lia, E. Praly, H. Ferreira, C. Stockdale, Y. Ching Tse-Dinh, D. Dunlap,
V. Croquette, D.Bensimon,T. Owen-Hughes, Directobservationof DNA
distortion by the RSC complex, Mol. Cell 21 (2006) 417–425.
 J. Jin, Y. Cai, T. Yao, A.J. Gottschalk, L. Florens, S.K. Swanson, J.L.
Gutierrez,M.K. Coleman,J.L. Workman,A. Mushegian,M.P. Washburn,
R.C. Conaway, J.W. Conaway, A mammalian chromatin remodeling
complexwith similarities to the yeastINO80 complex,J. Biol.Chem. 280
 T. Tsukiyama, C. Wu, Purification and properties of an ATP-dependent
nucleosome remodeling factor, Cell 83 (1995) 1011–1020.
 C.R. Clapier, G. Langst, D.F.V. Corona, P.B. Becker, K.P. Nightingale,
Critical role for the histone H4 N terminus in nucleosome remodeling by
ISWI, Mol. Cell. Biol. 21 (2001) 875–883.
 A. Brehm, G. Langst, J. Kehle, C.R. Clapier, A. Imhof, A. Eberharter,
J. Muller, P.B. Becker, dMi-2 and ISWI chromatin remodeling factors
have distinct nucleosome binding and mobilization properties, EMBO
J. 19 (2000) 4332–4341.
 J. Cote, J. Quinn, J.L. Workman, C.L. Peterson, Stimulation of GAL4
derivative binding to nucleosomal DNA by the yeast SWI/SNF complex,
Science 265 (1994) 53–60.
 J. Quinn, A.M. Fyrberg, R.W. Ganster, M.C. Schmidt, C.L. Peterson,
DNA-binding properties of the yeast SWI/SNF complex, Nature 379
 J. Cote, C.L. Peterson, J.L. Workman, Perturbation of nucleosome core
structure by the SWI/SNF complex persists after its detachment,
enchancing subsequent transcription factor binding, Proc. Natl. Acad.
Sci. 95 (1998) 4947–4952.
 X. Shen, G. Mizuguchi, A. Hamiche, C. Wu, A chromatin remodeling
complex involved in transcription and DNA processing, Nature 406
 A. Saha, J. Wittmeyer, B.R. Cairns, Chromatin remodeling by RSC
involves ATP-dependent DNA translocation, Genes Dev. 16 (2002)
 I. Whitehouse, C. Stockdale, A. Flause, M.D. Szcezelkun, T. Owen-
Hughes, Evidence for DNA translocation by the ISWI chromatin-
remodeling enzyme, Mol. Cell. Biol. 23 (2003) 1935–1945.
166J.J.F.A. van Vugt et al. / Biochimica et Biophysica Acta 1769 (2007) 153–171
 M.L. Phelan, S. Sif,G.J. Narlikar, R.E. Kingston,Reconstitutionof a core
chromatin remodeling complex from SWI/SNF subunits, Mol. Cell 3
 P.T. Georgel, T. Tsukiyama, C. Wu, Role of histone tails in nucleosome
remodeling by Drosophila NURF, EMBO J. 16 (1997) 4717–4726.
 D.F. Corona, G. Längst, C.R. Clapier, E.J. Bonte, S. Ferrari, J.W.
Tamkun, P.B. Becker, ISWI is an ATP-dependent nucleosomeremodeling
factor, Mol. Cell 3 (1999) 239–245.
 A. Saha, J. Wittmeyer, B.R. Cairns, Chromatin remodeling through
directional DNA translocation from an internal nucleosomal site, Nat.
Struct. Mol. Biol. 12 (2005) 747–755.
 K.E. Neely, J.L. Workman, The complexity of chromatin remodeling and
its links to cancer, Biochim. Biophys. Acta 1603 (2002) 19–29.
 S.M. Sengupta, M. VanKanegan,J. Persinger, C. Logie,B.R. Cairns, C.L.
Peterson, B. Bartholomew, The interactions of yeast SWI/SNF and RSC
with the nucleosome before and after chromatin remodeling, J. Biol.
Chem. 276 (2001) 12636–12644.
 M. Zofall, J. Persinger, S.R. Kassabov, B. Bartholomew, Chromatin
remodeling by ISW2 and SWI/SNF requires DNA translocation inside
the nucleosome, Nat. Struct. Mol. Biol. 13 (2006) 339–346.
 S.R. Kassabov, B. Zhang, J. Persinger, B. Bartholomew, SWI/SNF unwraps,
slides and rewraps the nucleosome, Mol. Cell 11 (2003) 391–403.
 R. Strohner, M. Wachsmuth, K. Dachauer, J. Mazurkiewicz, J.
Hochstatter, K. Rippe, G. Längst, A ‘loop recapture’ mechanism for
ACF-dependent nucleosome remodeling, Nat. Struct. Mol. Biol. 12
 D.P. Bazett-Jones, J. Cote, C.C. Landel, C.L. Peterson, J.L. Workman,
The SWI/SNF complex creates loop domains in DNA and polynucleo-
somearrays andcandisrupt DNA–histone contactswithinthese domains,
Mol. Cell. Biol. 19 (1999) 1470–1478.
 R. Schwanbeck, H. Xiao, C. Wu, Spatial contacts and nucleosome step
movements induced by the NURF chromatin remodeling complex,
J. Biol. Chem. 279 (2004) 39933–39941.
 A. Shundrovsky, C.L. Smith, J.T. Lis, C.L. Peterson, M.D. Wang,
Probing SWI/SNF remodeling of the nucleosome by unzipping single
DNA molecules, Nat. Struct. Mol. Biol. 13 (2006) 549–554.
 Y. Zhang, C.L. Smith, A. Saha, S.W. Grill, S. Mihardja, S.B. Smith, B.R.
Cairns, C.L. Peterson, C. Bustamante, DNA translocation and loop
formation mechanism of chromatin remodeling by SWI/SNF and RSC,
Mol Cell 24 (2006) 559–568.
 G. Langst, P.B. Becker, Nucleosome remodeling: one mechanism, many
phenomena? Biochim. Biophys. Acta 1677 (2004) 58–63.
 Y. Lorch, M. Zhang, R.D. Kornberg, Histone octamer transfer by a
chromatin-remodeling complex, Cell 96 (1999) 389–392.
 W. Li, S.-X. Dou, P. Xie, P.-Y. Wang, Brownian dynamics simulation of
directional sliding of histone octamers caused by DNA bending, Phys.
Rev., E 73 (2006) 051909.
 A. Hamiche, H. Richard-Foy, The switch in the helical handedness of the
histone (H3–H4)2 tetramer within a nucleoprotein particle requires a
reorientation of the H3–H3 interface, J. Biol. Chem. 273 (1998)
 G. Mizuguchi, X. Shen, J. Landry, W.H. Wu, S. Sen, C. Wu, ATP-driven
exchange of histone H2AZ variant catalyzed by SWR1 chromatin
remodeling complex, Science 303 (2004) 343–348.
 N.J. Krogan, M.-C. Keogh, N. Datta, C. Sawa, O.W. Ryan, H. Ding, R.A.
Haw, J. Pootoolal, A. Tong, V. Canadien, D.P. Richards, X. Wu, A. Emili,
T.R. Hughes, S. Buratowski, J.F. Greenblatt, A Snf2 family ATPase
complex required for recruitment of the histone H2A variant Htz1, Mol.
Cell 12 (2003) 1565–1576.
 G.P. Vicent, A.S. Nacht, C.L. Smith, C.L. Peterson, S. Dimitrov,
M. Beato, DNA instructed displacement of histones H2A and H2B at an
inducible promoter, Mol. Cell 16 (2004) 439–452.
 R. Bash, H. Wang, C. Anderson, J. Yodh, G. Hager, S.M. Lindsay,
during nucleosome remodeling, FEBS Lett. 580 (2006) 4757–4761.
 A.K. Nagaich, D.A. Walker, R. Wolford, G.L. Hager, Rapid periodic
binding anddisplacementof the glucocorticoid receptor duringchromatin
remodeling, Mol Cell 14 (2004) 163–174.
 E. Segal, Y. Fondufe-Mittendorf, L. Chen, A. Thastrom, Y. Field, I.K.
Moore, J.-P.Z. Wang, J. Widom, A genomic code for nucleosome
positioning, Nature 442 (2006) 772–778.
 G.R. Schnitzler, C.L. Cheung, J.H. Hafner, A.J. Saurin, R.E. Kingston,
C.M. Lieber, Direct imaging of human SWI/SNF-remodeled mono- and
polynucleosomes by atomic force microscopy employing carbon
nanotube tips, Mol. Cell. Biol. 21 (2001) 8504–8511.
 T.A. Blank, P.B.Becker, The effect of nucleosome phasingsequencesand
DNA topology on nucleosome spacing, J. Mol. Biol. 260 (1996) 1–8.
 A. Lusser, D.L. Urwin, J.T. Kadonaga, Distinct activities of CHD1 and
ACF in ATP-dependent chromatin assembly, Nat. Struct. Mol. Biol. 12
 I. Whitehouse, T. Tsukiyama, Antagonistic forces that position nucleo-
somes in vivo, Nat. Struct. Mol. Biol. 13 (2006) 633–640.
 P.D. Varga-Weisz, M. Wilm, E. Bonte, K. Dumas, M. Mann, P.B. Becker,
Chromatin-remodeling factor CHRAC contains the ATPases ISWI and
topoisomerase II, Nature 388 (1997) 598–602.
 B.R. Cairns, Chromatin remodeling complexes: strength in diversity,
precision through specialization, Curr. Opin. Genet. Dev. 15 (2005)
 K. Havas, A. Flaus, M.L. Phelan, R. Kinston, P.A. Wade, D.M.J. Lilley,
T. Owen-Hughes, Generation of superhelical torsion by ATP-dependent
chromatin remodeling activities, Cell 103 (2000) 1133–1142.
 G. Langst, P.B. Becker, ISWI induces nucleosome sliding on nicked
DNA, Mol. Cell 8 (2001) 1085–1092.
 H. Durr, C. Korner, M. Muller, V. Hickmann, K.P. Hopfner, X-ray
structures of the Sulfolobus solffataricus SWI2/SNF2 ATPase core and its
complex with DNA, Cell 121 (2005) 363–373.
 S. Korolev, J. Hsieh, G.H. Gauss, T.M. Lohman, G. Waksman, Major
domain swiveling revealed by the crystal structures of complexes of
E. coli Rep helicase bound to single-stranded DNA and ADP, Cell 90
 N.H. Thoma, B.K. Czyzewski, A.A. Alexeev, A.V. Mazin, S.C.
Kowalczykowski, N.P. Pavletich, Structure of the SWI2/SNF2 chroma-
tin-remodeling domain of eukaryotic Rad54, Nat. Struct. Mol. Biol. 12
 S.S. Velankar, P. Soultanas, M.S. Dillingham, H.S. Subramanya, D.B.
Wigley, Crystal structures of complexes of PcrA DNA helicase with a
DNA substrate indicate an inchworm mechanism, Cell 97 (1999) 75–84.
 X. He, H.Y. Fan, G.J. Narlikar, R.E. Kingston, Human ACF1 alters the
remodeling strategy of SNF2h, J. Biol. Chem. 281 (2006) 28636–28647.
 C. Stockdale, A. Flaus, H. Ferreira, T. Owen-Hughes, Analysis of
nucleosome repositioning by yeast ISWI and Chd1 chromatin remodeling
complexes, Journal of Biological Chemistry 281 (2006) 16279–16288.
 Y. Lorch, B. Maier-Davis, R.D. Kornberg, Chromatin remodeling by
nucleosome disassembly in vitro, Proc. Natl. Acad. Sci. 103 (2006)
 H.-Y. Fan, K.W. Trotter, T.K. Archer, R.E. Kingston, Swapping function
of two chromatin remodeling complexes, Mol. Cell 17 (2005) 805–815.
 C. Logie, C.L. Peterson, Catalytic activity of the yeast SWI/SNF complex
on reconstituted nucleosome arrays, EMBO J. 16 (1997) 6772–6782.
 M. Poirier, S. Eroglu, D. Chatenay, J.F. Marko, Reversible and
irreversible unfolding of mitotic newt chromosomes by applied force,
Mol. Biol. Cell 11 (2000) 269–276.
 T. Reguly, A. Breitkreutz, L. Boucher, B.J. Breitkreutz, G.C. Hon, C.L.
Myers, A. Parsons, H. Friesen, R. Oughtred, A. Tong, C. Stark, Y. Ho,
D. Botstein, B. Andrews, C. Boone, O.G. Troyanskya, T. Ideker,
K. Dolinski, N.N. Batada, M. Tyers, Comprehensive curation and
analysis of global interaction networks in Saccharomyces cerevisiae,
J. Biol. 5 (2006) 11.
 S. Peri, J.D. Navarro, R. Amanchy, T.Z. Kristiansen, C.K. Jonnalagadda,
V. Surendranath, V. Niranjan, B. Muthusamy, T.K. Gandhi, M. Gronborg,
N. Ibarrola, N. Deshpande, K. Shanker, H.N. Shivashankar, B.P. Rashmi,
M.A. Ramya, Z. Zhao, K.N. Chandrika, N. Padma, H.C. Harsha, A.J.
Yatish, M.P. Kavitha, M. Menezes, D.R. Choudhury, S. Suresh,
N. Ghosh, R. Saravana, S. Chandran, S. Krishna, M. Joy, S.K. Anand,
V. Madavan, A. Joseph, G.W. Wong, W.P. Schiemann, S.N.
Constantinescu, L. Huang, R. Khosravi-Far, H. Steen, M. Tewari,
167J.J.F.A. van Vugt et al. / Biochimica et Biophysica Acta 1769 (2007) 153–171