Paralogue switching in the Nucleosome Remodelling and Deacetylase complex can drive

functional specificity

Xavier J Reid1, Jason K K Low1, Joel P Mackay1, *

1
School of Life and Environmental Sciences, University of Sydney, NSW 2006 Australia

*
To whom correspondence should be addressed:

JPM: joel.mackay@sydney.edu.au

Abstract

The Nucleosome Remodelling and Deacetylase (NuRD) complex is an essential transcriptional

regulator in all complex animals. All seven core subunits of the complex exist as multiple paralogues,
raising the question of whether the complex might utilize paralogue switching to achieve cell-type
specific functions. In this review, we examine the evidence for this idea, making use of published
quantitative proteomics data to dissect NuRD composition in 20 different tissues, as well as a large-
scale CRISPR knockout screen carried out in more than 1000 human cancer cell lines. These data,
together with recent reports, provide strong support for the idea that distinct permutations of the
NuRD complex with tailored functions might regulate tissue specific gene expression programs.

Introduction

Chromatin is the compact structure of DNA in the nucleus of eukaryotic cells. It is a dynamic structure
that provides a basal layer to transcriptional regulation – densely packed chromatin restricts access
of gene promoters to the transcriptional machinery whereas more open chromatin facilitates access.
A central mechanism by which chromatin density is altered is through the repositioning or eviction
of histone octamers, the core unit of chromatin, by chromatin remodelling enzymes. Chromatin
remodellers are therefore essential to the cell, and their failure to execute cell type specific
transcriptional programs can result in developmental abnormalities and diseases such as cancer [1].

There are four classes of chromatin remodelling enzymes – CHD, SWI/SNF, ISWI and INO80 – and
in humans these classes comprise the proteins CHD1-9, SMARCA2 and SMARCA4, SMARCA1
and SMARCA5, and SRCAP and EP400, respectively. Further functional diversification results from
assembling into a range of chromatin remodelling complexes [2].

One such chromatin remodelling complex is the Nucleosome Remodelling and Deacetylase (NuRD)
complex, which utilises a CHD-family chromatin remodelling enzyme to regulate gene expression in
many different contexts and at different scales. NuRD was initially described as a transcriptional co-
repressor complex [3], although the current view is that NuRD is capable of acting as a co-activator,
a co-repressor, or a balance of both [4-6]. For example, at sites of active transcription in embryonic
stem (ES) cells, NuRD maintains gene expression within appropriate ranges rather than switching
particular genes on or off [6]. In contrast, in erythroid cells, NuRD plays a more specific role of fully
repressing the expression of fetal γ-globin genes [5]. Overall, it is clear that NuRD is essential for
normal development and for homeostasis in multiple tissues.

In addition to the CHD remodelling enzyme, NuRD comprises the subunits GATAD2, MBD, MTA,
HDAC, RBBP and CDK2AP (Figure 1). In mammals, each subunit has at least two paralogues that
can be incorporated into the complex – these are CHD3/4/5, GATAD2A/B, MBD2/3, MTA1/2/3,
HDAC/1/2, RBBP4/7 and CDK2AP1/2. The paralogues share medium to high sequence identity,
ranging from about 40% in the case of GATAD2A/B to about 90% for RBBP4/7. Emerging evidence
suggests that NuRD complexes that feature different paralogue combinations can drive distinct
functional outcomes - a phenomenon termed paralogue switching. Herein, we use this term to
encompass changes in paralogue composition that might occur as a function of time in a single
cell/tissue or simply as a function of location (that is, different paralogue architectures that might be
observed in different cell or tissue types). Because of the number of paralogues involved, the potential
clearly exists for a whole constellation of NuRD complexes with distinct functionalities.

Figure 1. Schematic and domain topology of the NuRD complex. HMG, HMG domain; PHD,
PHD domain; CD, chromodomain; CTD1/2, C-terminal domain; CC, coiled coil; MBD, MBD
domain; BAH, BAH domain, ELM, ELM domain; SANT, SANT domain; ZF, zinc finger; R1/2,
RBBP4/7 binding motif, WD40; WD40 domain; DD, dimerization domain. The grey dashed box
around R2 indicates that MTA3 does not contain this motif. Coloured boxes in the topology cartoons
indicate domains for which there are experimentally determined structures.

The paralogue switching concept raises several important questions. How does paralogue switching
affect nucleosome substrate recognition, and ultimately genomic targeting? Does switching occur in
a cell type dependent manner to create tissue specific functions? What are the mechanisms that
contribute to and facilitate paralogue switching and what are the cellular consequences? Here we
address these questions by tying together recent examples of NuRD paralogue switching with a meta-
analysis of published proteomics data and CRISPR knockout screens. We first discuss features of the
paralogues (the NuRD building blocks) that might dictate complex assembly, then consider the effect
of cellular context (the location and neighbourhood) on NuRD composition and functional output.
Finally, we describe compelling examples of the phenomenon, which foreshadow what we suspect
will be a wide-ranging mechanism of gene regulation.

What do the building blocks tell us?

Insight into the different architectural styles of NuRD first requires appreciation of the bricks and
mortar - the subunits from which the complex is built. The defining subunit of the complex is CHD4.
This ATP-dependent DNA translocase is capable of remodelling chromatin independently of the
NuRD complex [7, 8], immediately raising the question of what the NuRD complex can do that CHD4
alone cannot. The answer to that question lies in what each subunit can bring to the table, whether it
be additional enzymatic activities, diversified or specialised chromatin recognition activity, or
interaction with and recruitment by subunit- or perhaps paralogue-specific transcription factors.

The paralogues CHD3/4/5 each contain two PHD zinc fingers and two chromodomains. PHD
domains often act to recognize the modification state of histone H3K4 (methylation in particular)
whereas chromodomains commonly read histone lysine methylation state [9]. The PHD domains of
CHD3 and CHD4 do not, however, bind methylated H3K4 and have instead been shown to
preferentially bind to H3K9me3 and H3K9ac [10-13]. In contrast, the PHD domains of CHD5 prefer
to bind unmodified histone H3 – including non-methylated H3K4 [14, 15]. The chromodomains of
CHD4 are atypical in that they bind DNA rather than recognising histone lysine methylation [16].
However, a recent structure of CHD4 bound to a nucleosome shows that the double chromodomain
lies close to the H3 N-terminal tail, raising the possibility of an interaction [17]. The chromodomains
of CHD5 on the other hand can bind H3K27me3 [18].

These data hint at differences in the chromatin binding potential of CHD3/4/5, suggesting that free
CHD3 and CHD4 might preferentially interact with nucleosomes bearing the H3K9me2/3 and
H3K9ac modifications and lacking H3K4 methylation, whereas free CHD5 might prefer unmodified
or H3K27me3-bearing nucleosomes. It is worth noting that the histone binding properties of
CHD3/4/5 have only been studied in the context of isolated domains and histone peptides, indicating
the need for experiments with full-length CHD proteins and intact nucleosomes [19]. In this context,
two recent preprints demonstrate that the binding selectivity of the BPTF bromodomain-PHD domain
module differs significantly in the context of a nucleosome rather than histone peptides [20, 21].
Indeed, CHD4 ChIPseq data in RH4 cells, mouse ES cells, mouse neurons, and Drosophila S2 cells
show the remodeller occupying chromatin with marks that include H3K4me1, H3K4me3, H3K27ac,
H3K4me1 + H3K27ac and H3K4me3 + H2A.Z [1, 6, 22-25] - in apparent contrast to peptide binding
data.

The contribution of HDAC to the NuRD complex appears straightforward – the addition of lysine
deacetylase activity. However, not only are the targets of NuRD-bound HDACs poorly defined
(including even the question of whether the targets are histones or other chromatin-related proteins),
it is also not clear at a biochemical level what one HDAC paralogue might offer the complex over the
other. HDAC1/2 appear to act redundantly more often than other NuRD paralogues, having
overlapping functions in cardiac development [26], cell cycle regulation and haematopoiesis [27],
and adipogenesis [28]. Several instances of distinct roles have, however, been reported: HDAC1 but
not HDAC2 is essential in embryonic stem cell differentiation [29], whereas knockdown of HDAC1
but not HDAC2 in U2OS cells induces apoptosis [30]. Although we note that HDAC1/2 are
components of several chromatin regulatory complexes, these examples of cell type dependent,
paralogue specific functions for HDAC1/2 suggest the possibility that paralogue switching in the
NuRD complex might in part be a function of cellular context.

MTA1/2/3 are recognized as the structural scaffold upon which the NuRD complex is assembled [31,
32]. These proteins all bear an N-terminal BAH domain that, in other proteins, has been shown to
recognize a range of histone modifications, including H3K27me3 [33], H4K20me2 [34], and
H3K9me2 [35] – though no similar activities have yet been identified for the MTA BAH domain. It
is notable, however, that the structure of an MTA1-HDAC1 subcomplex shows the BAH domain of
MTA1 poised next to the active site of HDAC1, in an ideal position to bring nucleosomes to the
HDAC1 active site [36]. The MTA subunit also directly binds the RBBP subunits, and whereas
MTA1/2 have two RBBP binding motifs and can each recruit two RBBP subunits to NuRD [37, 38],
MTA3 is likely to recruit only one via its single RBBP binding motif, indicating that NuRD
stoichiometry can be regulated by paralogue choice.

A more extreme example of paralogue-specific determination of complex stoichiometry lies in the

recent report that the transcriptional coregulator PWWP2A binds to an HDAC-MTA-RBBP NuRD
subcomplex – in direct competition with the MBD-GATAD2-CHD4 portion of NuRD [39]. This
interaction occurs preferentially with MTA1 rather than MTA2, suggesting paralogue-specific
decoupling of the two enzymatic modules of NuRD by a co-regulatory protein [39]. PWWP2A
localises to H2A.Z containing nucleosomes, and therefore might target MTA1-specifc NuRD
subcomplexes to these regions.

RBBP4 and RBBP7 are single-domain -propeller proteins that have been shown to directly bind
numerous protein partners, including histones H3 and H4 [37, 38, 40] as well as a wide range of
transcriptional regulatory proteins such as FOG1, SALL1, BCL11A/B, PHF6, AEBP2 and ZNF827
[41-44]. RBBP4/7 bind to histone H4 and MTA1 via the same surface [37, 38], implying that when
RBBP4/7 are part of the NuRD complex they do not contact histone H4. The histone H3 contact
surface does, however, remain available – leaving NuRD-bound RBBPs with the potential to
recognise this histone [45]. Interestingly, a recent preprint shows that, in embryonic stem cells (ESCs),
disruption of RBBP4 results in a genome wide increase in H3K27ac levels, suggesting RBBP4 might
act to target HDAC containing complexes such as NuRD to chromatin [46]. Notably, disruption to
RBBP4, but not to RBBP7, has a profound negative effect on H3K27me3 levels within gene bodies,
pointing to non-overlapping functions for the two paralogues in this context.

MBD2 and MBD3 are DNA-binding proteins that differ in their selectivity for methylated DNA.
MBD2 binds almost exclusively to methylated CpG sequences whereas MBD3 exhibits a similar
affinity for methylated and unmethylated DNA [47-49]. ChIP-seq analysis shows that although the
majority of both MBD2 and MBD3 DNA binding sites are CpG islands, only MBD2 binds methylated
CpG islands [50, 51]. This selectivity provides the NuRD complex with an opportunity to bind distinct
genomic targets via differential incorporation of MBD2 or MBD3, and thereby to play different
functional roles. For example, in HeLa cells, MBD2 recognizes sites that are highly enriched for
methylated DNA and is typically associated with repressed genes. In contrast, MBD3 sites are
depleted in methylation and are associated with active genes [50], though some uncertainty remains
regarding the binding preferences of MBD3 [48, 49, 52, 53].
The GATAD2 paralogues are the least characterised of the NuRD subunits. Biochemical data indicate
that these proteins bridge MBD and CHD subunits (Figure 1), recruiting CHD to the complex as a
peripheral subunit [7, 31, 54-56]. Although these proteins are predicted to contain a GATA-class zinc
finger domain, there is no evidence that this domain has DNA-binding activity. GST-pulldown data
suggest that the interaction of GATAD2A with MBD proteins might be mediated by two separate
domains, whereas GATAD2B might use only its N-terminal coiled coil domain to contact MBDs
[57].

GATAD2A and GATAD2B are also SUMOylated in vivo, which appears to modulate their binding
to other NuRD subunits [58]. GATAD2A is SUMOylated at two sites, (K30 and K487), whereas
GATAD2B is SUMOylated only at one (K33). HDAC1 binding to GATAD2A, but not GATAD2B,
is dependent on GATAD2A K30 SUMOylation; in contrast, RBBP7 binding to GATAD2B but not
GATAD2A is dependent on GATAD2B K33 SUMOylation [58]. Furthermore, HDAC2 binding to
GATAD2A might also be dependent on GATAD2A phosphorylation at S96 [59]. Together these data
suggest that post-translational modification of NuRD paralogues might be one mechanism that
licences paralogue specific NuRD complex formation.

Another factor that might contribute to paralogue specific complex formation is how much of each
building block the cell has at hand. Based on relative concentrations in the cell, it could be expected
that a more abundant paralogue will have a better chance of being incorporated into the complex. Our
analysis in Box 1, which is based on published quantitative mass spectrometry data [60, 61], shows
that CHD4, MBD3, MTA2 and HDAC2 are more abundant than their paralogues across a range of
human cancer cell lines, suggesting that an ‘average’ NuRD complex might be more likely to be built
around these paralogues. An interesting question then is whether specific NuRD assemblies operate
on different scales based on the abundance of their constituents and their availability in the cell. For
example, the more abundant MBD3-NuRD has been shown to modulate the expression of a large
number of genes [6], whereas MBD2-NuRD is known to specifically repress the fetal globin genes
[5]. It might thus be that less abundant NuRD assemblies are employed as ‘precision’ gene regulation
tools.

Overall, it is reasonably clear at the subunit level what is added to the NuRD complex with each
building block: HDACs bring an additional enzymatic activity, whereas the RBBPs interact with both
histones and a host transcription factors – presumably to recruit the complex to its genomic targets.
Based on the properties discussed so far, however, the impact of choosing one paralogue over the
other is less clear. The CHD paralogues show some differences in nucleosome binding potential, the
MBDs have differences in selectivity for methylated and unmethylated DNA, and the MTAs may be
able to regulate both chromatin binding and NuRD stoichiometry - although the functional outcomes
of these differences remain unknown. Knockout of the NuRD paralogues across a library of human
cancer cell lines (see Box 2) [62], however, suggests that there are significant functional differences
between the paralogues. This is particularly the case for CHD4 and RBBP4; these two paralogues are
essential in virtually all cell lines, suggesting that their function is not compensated for by their
counterparts. Other paralogues appear to be essential in a tissue-restricted manner, and overall, these
knockout data point to the existence of paralogue specific functions in certain tissues.
Box 1: How much are we talking about?

The biochemical properties of these building blocks indicate that NuRD has the potential to recognise
more diverse chromatin substrates than CHD4 alone and to display more diverse activities. However,
the impact of each paralogue on NuRD assembly and thus on gene regulation in any given cell at any
given time will depend strongly on the relative concentrations at which they are expressed.

We used two quantitative proteomics datasets [60, 61] to estimate the number of molecules per cell
for each paralogue in 375 human cancer cell lines that represent 20 different tissues. To do so, we
used cell lines and proteins that overlapped between the two datasets (one dataset contains absolute
quantification of the human proteome in four human cancer cell lines, whereas the other dataset
contains relative quantification of the human proteome in 375 human cancer cell lines) to convert the
relative quantification data of the larger study to an absolute quantification dataset (in molecules per
cell). We first noted that, of the chromatin remodellers that were detected in both datasets (CHD1,
CHD3, CHD4, CHD6, CHD8, CHD9, SMARCA5, SRCAP and EP400), CHD4 and SMARCA5 are
at least ten-fold more abundant than any of the others, suggesting that NuRD is one of the more
prolific chromatin remodelling complexes in the cell (Figure IA). Figure IB shows the number of
molecules of each NuRD subunit per cell (aggregated across all paralogues) for all 375 cell lines. It
is notable that the median abundance for NuRD subunits
(220,000:310,000:320,000:500,000:400,000:1,000,000 copies per cell for
CHD:GATAD2:MBD:MTA:HDAC:RBBP) is in good agreement with the stoichiometry of
1:1:1:2:2:4 that was recently deduced from targeted and absolute quantitative mass spectrometry data
[31], despite the fact that several subunits (CHD, HDAC, and RBBP) are members of other protein
complexes.

At the paralogue level, there are substantial differences in abundance for most subunits (Figure IC).
CHD4 is ten-fold more abundant than CHD3 (CHD5 was not detected in one dataset), MBD3 is eight-
fold more abundant than MBD2, MTA2 is four-fold more abundant than MTA1 and 11-fold more
abundant than MTA3, and HDAC2 is three-fold more abundant than HDAC1. These differences are
consistent across all tissue types, with only a few exceptions in individual cell lines. In comparison,
GATAD2A & GATAD2B and RBBP4 & RBBP7 are equally abundant on average across all cell
lines, and this too is consistent across all tissue types. It could therefore be expected, based on
abundance alone, that an ‘average’ NuRD complex might be more likely to be built around CHD4,
MBD3, MTA2 and HDAC2.
Figure I. Concentrations of the NuRD building blocks in the cell. A. Distribution of abundances
of various human chromatin remodellers across 375 human cancer cell lines [60, 61]. B. Histogram
showing the distribution of NuRD subunit abundance in molecules per cell, measured across the same
set of cell lines [60, 61]. C. Distribution of NuRD paralogue abundance in molecules per cell,
measured across the same set of cell lines.

Box 2: Building block demolition

We analysed CRISPR knockout data from the Cancer Dependency Map project [62] to ascertain how
essential each paralogue is to cell viability and to infer the likely level of redundancy between
paralogues. This dataset comprises >17,000 gene knockouts in more than 1000 human cancer cell
lines and is reported as a CERES score; a lower score indicates a higher level of importance (–1 is
the median score for pan-essential genes and zero indicates a non-essential gene [62]). Figure II
shows the mean CERES scores for all NuRD paralogues by tissue type. Strikingly, CHD4 and RBBP4
are essential in all tissue types, whereas their paralogues do not appreciably affect viability.
GATAD2A, MTA2 and HDAC1 are essential in more cell lines than their counterparts across a wide
range of tissues, whereas MBD2 and MBD3 are largely comparable, except in lymphocytes and
plasma cells where MBD2 is essential in more of these cell lines than MBD3. Overall, these data
show that loss of individual NuRD paralogues, in many instances, cannot be completely compensated
by their counterparts, pointing strongly to the existence of paralogue specific functions in different
tissues. This observation is particularly surprising in the case of RBBP4, given the equal abundance
of the two paralogues across many cell types and their ~90% sequence identity (again we note that,
similar to the case for HDACs, RBBP4/7 are found in multiple complexes, meaning that this
difference might not be attributable to NuRD activity).

Figure II. NuRD paralogues are essential to cell viability. Mean effect size (CERES score) of
CRISPR-mediated knockout of each NuRD paralogue in >1000 human cancer cell lines, grouped into
tissue type [62]. A score of –1 is the median score for pan-essential genes and zero indicates a non-
essential gene. The total number of cell lines used for each tissue type is indicated.

Location, location, location

Chromatin modifying complexes are no strangers to the idea of cell type or lineage specific
assemblies. Several such assemblies have been described for the SWI/SNF-family BAF complex, and
these act upon very different nucleosome substrates [63, 64]. For example, an embryonic stem cell-
specific BAF complex, esBAF, is critical for pluripotency [65]. Similarly, the repressive CoREST
complex, which shares the HDAC1/2 subunits with NuRD, has been shown to exhibit lineage-
restricted assemblies [66]. Does NuRD act in the same manner – with specific NuRD assemblies
required to execute appropriate transcriptional programs? If so, then location – tissue identity or
developmental stage – might provide the backdrop for such diversity.

Our correlation analysis described in Box 3, based on the protein abundance data described in Box 1,
suggests that certain paralogues might play important roles in a specific tissue, even if they are not
the most abundant paralogue nor even part of the most abundant NuRD assembly. For example,
Figure 2 suggests the existence of a NuRD complex in bone in which CHD3 and MTA1 are important,
yet CHD3 and MTA1 are far less abundant in bone (and in other tissues) than their counterparts.
Conversely, although GATAD2 and RBBP paralogues are equally abundant in all tissues, we see
clear paralogue preferences in several tissues. For example, RBBP4 is highly preferred in breast, CNS,
lymphocytes, ovary and pancreas, whereas this preference is switched to RBBP7 in gastric and upper
aerodigestive tissues. These apparent differences are particularly surprising for RBBP4/7 due to their
high sequence identity.

Similarly, in most tissues, MBD3 is found in the most highly correlated NuRD complex but there is
a quantitative switch to MBD2 in plasma cells, lymphocytes, prostate and skin. The identification of
MBD2-NuRD in plasma cells and lymphocytes is particularly noteworthy because the importance of
MBD2-NuRD on the other side of the haematopoietic lineage, in red blood cells, is well established.
Consistent with our proposed MBD3 to MBD2 switch in the lymphoid lineage, MBD2 has been
shown to be indispensable for both lymphopoiesis and leukemogenesis [67, 68] and thus MBD2-
NuRD might play important roles throughout haematopoiesis.

Strikingly, many of the tissue specific paralogue preferences we identify have been implicated in the
cancers that these tissues represent. For example, (i) CHD4, MTA3 and HDAC1 are overexpressed
in hepatocellular (liver) carcinoma and correlate with poor patient survival [69], (ii) GATAD2A, but
not GATAD2B, is upregulated in bladder (urinary tract) cancer [70], (iii) MTA2, but not MTA1 or
MTA3, is significantly downregulated after EGFR knockdown in head and neck (upper aerodigestive)
cancer [71], and (iv) MBD3, but not MBD2, is mutated in ovarian and uterine cancers [72].

Although we have identified tissue specific NuRDs apparently built around particular paralogues, we
also observe many cases where paralogues are equally represented in the most correlated complexes.
For example, whereas we propose that MBD3, MTA1 & HDAC2 dominate the NuRD complex in
colorectal cell lines, no clear preferences are observed for the CHD, GATAD & RBBP paralogues.
This difference might indicate that multiple NuRD assemblies exist in some tissues or might simply
represent the loss of resolution in our analysis that arises from combining different cell types under a
single tissue category. For example, the cell lines that make up ‘blood’ in our dataset include T-cells,
B-cells and erythroblasts, among others. A T-cell NuRD might differ from a NuRD in erythroblasts,
but this distinction would be lost by grouping the cell types. Our strategy of combining cell types
from a single tissue does, however, increase confidence in differences that we do observe.

We speculate that the tissue specific NuRD complexes we have identified play a specific/important
role in that tissue. In some cases, this is reflected in the CRISPR knockout screens; for example, we
suggest an MBD2-NuRD as being more important in plasma cells and lymphocytes, the two lineages
that have a higher proportion of cell lines in which MBD2 is essential.
Figure 2. NuRD composition in different tissue types. Plots for each tissue showing the frequency
of each paralogue in the ten paralogue-specific NuRD permutations with the highest total correlation
scores. The total number of cell lines used for each tissue type is indicated. The process for identifying
tissue-specific NuRD assemblies is described in detail in Box 3.

Box 3: Possibilities expanded: identification of tissue-specific NuRD assemblies

It is well established that proteome abundance data can be leveraged to build correlation networks
and to infer both physical and functional associations between proteins [60, 73-76]. Recently, this
approach has been used to identify subunit and paralogue variation within protein complexes across
different cell types [77]. We therefore reasoned that we could take a similar approach to identify
tissue-specific NuRD assemblies. We used the quantitative proteomics dataset [60] to create a
correlation map of abundances for NuRD paralogues for each of the 20 tissues (Figure III). We first
calculated Pearson correlation coefficients for the abundance of each paralogue pair across all cell
lines in a given tissue (e.g., CHD4 abundance vs HDAC1 abundance and vs HDAC2 abundance etc).
We then summed the coefficients for each possible permutation of the NuRD complex (144 in total,
excluding the possibility of ‘mixed complexes’ containing both paralogues of one subunit) to get an
overall correlation score for each paralogue-specific complex. We ranked the complexes by overall
correlation score and, for each tissue, plotted the frequency of each paralogue in the top ten most
highly correlated NuRD complexes. In each tissue we identify at least two paralogues that are
significantly overrepresented in the top ten most correlated complexes (Figure 2), suggesting that a
NuRD complex built from those paralogues exists in that tissue.
Figure III. Flowchart for the identification of tissue-specific NuRD assemblies. In each tissue
and for each paralogue permutation of the NuRD complex, a total correlation score was calculated
by summing all pairwise Pearson correlation coefficients within that complex for cell lines derived
from that tissue, and then all complexes were ranked by total correlation score.

Who’s in the neighbourhood?

The functional output of a NuRD assembly will be determined in part by who else is in the cellular
neighbourhood. One important hub within a neighbourhood comprises the interactions between
NuRD and transcription factors. These interactions are often tissue and paralogue specific and have
the potential to direct the right NuRD complex to the appropriate genomic targets (though the
distinction between recruitment and simple co-occupancy is challenging to demonstrate).

As an example, MTA1 plays a role in T-cell development through its interaction with the T-cell
specific transcriptional repressor BCL11B [78]. Although both MTA1 and MTA2 can interact with
BCL11B, only MTA1 enacts transcriptional repression at BCL11B target promoters [78], indicating
that in T-cell development MTA1-NuRD is responsible for BCL11B mediated repression. In contrast,
during B-cell development, it is MTA2-NuRD that interacts with the B-cell specific transcription
factors AILIOS and BCL11A [79-81]. In pre-B cells, MTA2-NuRD represses genes involved in T-
cell differentiation, showing that MTA2-NuRD is important for commitment to the B-cell lineage
[79]. Despite this MTA2-NuRD specific activity, MTA1 and MTA3 are also expressed throughout
B-cell development, although no MTA1/3 specific functions have been identified [79]. MTA1- and
MTA3-NuRD complexes might therefore play distinct roles in B-cell development, a scenario that is
hinted at by the interaction between MTA3 and the B-cell specific transcriptional repressor BCL-6,
which prevents terminal differentiation of B-cells into plasma cells [82]. Thus, MTA paralogues
therefore differentially interact with tissue restricted transcription factors such as BCL11A/B and
BCL-6 to regulate cell fate.

MBD3, but not MBD2 has also been reported to interact with BCL11A and might therefore
participate in B-cell development [80, 83]. Rather than enforcing B-cell lineage commitment as the
interaction with BCL11A might suggest, MBD3-NuRD appears to be a mediator of both B- and T-
cell lineage commitment [84]. In the context of ES cell differentiation, MBD3-NuRD interacts with
the transcription factor ZBTB2, although rather than interacting directly and specifically through
MBD3, the BTB-link domain of ZBTB2 interacts with GATAD2A/B [85]. ZFP296 is another protein
involved in ES cell differentiation that interacts with MBD3-NuRD.[86]. The ZFP296-MBD3-NuRD
interaction is restricted to ES cells, underlining the importance of tissue restricted transcription factors
in recruiting NuRD to enact cell type specific functions. MBD3-NuRD, but apparently not MBD2, is
also recruited by FOG1 to GATA1 targets [41]. As GATA1 is regarded as the master regulator of
erythropoiesis [87], this suggests that MBD3-NuRD is involved in erythroid development. In
intestinal stem cells, MBD3-NuRD is recruited by c-JUN to repress the intestinal stem cell marker
lgr5 and control cellular proliferation [88]. It is clear that, depending on cell type, MBD3-NuRD
plays an important role in both stem cell proliferation and cell fate determination. In the case of
MBD2, the enigmatic NuRD subunit CDK2AP1 recruits MBD2-NuRD to repress master regulators
of the epithelial-mesenchymal transition (EMT) and induce cell differentiation towards epithelial
cells [89].

The chromatin reader protein ZMYND8 recruits GATAD2A-NuRD to sites of DNA damage through
recognition of histone acetylation [90, 91]. The ZMYND8-GATAD2A interaction is mediated
through the MYND domain of ZYMND8, which engages a PPPL motif found in GATAD2A but not
GATAD2B, although GATAD2B is still recruited to sites of DNA damage independently of
ZMYND8. GATAD2A and GATAD2B knockouts show that recruitment of MBD2 to the same sites
of DNA damage is dependent on the ZMYND8-GATAD2A interaction but not on GATAD2B [90].
Together these data point to two different NuRD complexes – a GATAD2A-MBD2-NuRD and a
GATAD2B-NuRD – being recruited to sites of DNA damage via two distinct mechanisms.

A number of other studies have demonstrated paralogue specific NuRD recruitment by transcription
factors – though with functional outcomes that are less clear. For example, the corepressor TRIM28,
which is involved in the maintenance of ES cell pluripotency and haematopoiesis [92], recruits CHD3
but not CHD4 to effect repression at TRIM28 targets [93]. The transcriptional repressor PHF6
interacts specifically with RBBP4 [44, 94, 95]. Because PHF6 has been implicated in a range of
processes such as cell cycle regulation [96], neural development [97] and tumour suppression [98],
the relevance of the PHF6-RBBP4 interaction is not yet clear. The two regions of PHF6 that interact
with RBBP4 share high sequence identity with an N-terminal region present in other NuRD
interacting transcription factors such as SALL1, BCL11A/B and FOG1 [44]. This region is, however,
conserved between RBBP4 and RBBP7, flagging the possibility that paralogue specificity in this case
does not stem from sequence differences at the direct interaction site.
Case studies for NuRD paralogue specificity

We conclude with several examples that underscore the link between paralogue composition and
cellular function. NuRD represses the fetal γ-globin genes and is therefore a key regulator of the
switch from fetal to adult globin expression [5, 99-104]. This NuRD complex was initially described
as an MBD2-NuRD complex [100-102], and CHD4 [99] and GATAD2A [54, 103] were also shown
to be the paralogues required for γ-globin silencing. Additionally, an intrinsically disordered region
in MBD2 both enhances its specificity for methylated DNA and specifically recruits MTA2, HDAC2
and RBBP4, which are required for γ-globin repression [100, 104]. These findings were elegantly
corroborated in a recent study from the Bauer lab, which demonstrated that a highly specific NuRD
complex containing CHD4, GATAD2A, MBD2, MTA2, HDAC2 and RBBP4 is responsible for the
silencing of the γ-globin genes [5]. They first used a CRISPR screen in an erythroid progenitor cell
line to show that CHD4, GATAD2A, HDAC2, MBD2 and MTA2 are required for fetal haemoglobin
(HbF) repression, whereas their counterparts CHD3, GATAD2B, MBD3, MTA1/3 and RBBP4/7 are
not. In agreement with our analysis of CRISPR knockout screens, it was also found that CHD4 and
RBBP4 – but no other NuRD paralogues – are essential for cell fitness.

To consolidate these findings, they used MTA2 and CHD4 as bait for affinity purifications from the
same cell line, followed by label-free quantification by mass spectrometry (IP-MS). Both MTA2 and
CHD4 pulled down the same paralogues identified as essential for HbF repression in the CRISPR
screen: GATAD2A, MBD2, HDAC2, MTA2 and RBBP4. It was also observed that, in the case of
the MBD, HDAC and RBBP subunits, the preferred paralogues in the complex did not correlate with
paralogue abundance (inferred from mRNA expression data), suggesting the ‘active’ assembly of this
particular NuRD permutant. This example is the first of a fully paralogue specific NuRD complex
with a well-defined function. The data also open up the possibility that therapeutic intervention
(targeting NuRD to reactivate -globin) in haemoglobinopathies such as β-thalassemia and sickle cell
disease could target one NuRD permutation to reduce side effects that might be associated with
wholesale disruption of an essential co-regulator complex.

NuRD paralogue switching is also heavily involved in other developmental processes such as
neurogenesis and brain development. Expression of CHD5 is largely restricted to the brain [105],
hinting at a brain specific function for CHD5-NuRD; indeed a CHD5-GATAD2B-MTA3-HDAC2-
RBBP7 NuRD complex was isolated from rat brain, and depletion of CHD5 altered the expression of
neuronal genes involved in aging, Alzheimer’s disease and neuronal function [105]. Further, a defined
program of CHD paralogue switching occurs throughout brain development in mice [106]. CHD4-
NuRD is required for the proliferation of neural progenitor cells, CHD5-NuRD promotes early neural
migration, and CHD3-NuRD is involved in late neural migration and cortical layer specification.
Importantly, the effects of depletion of each CHD paralogue could not be rescued by its counterparts.
CHD5 is also required for neuronal differentiation, highlighting the importance of CHD5-NuRD in
neurogenesis (18). Work from the Hendrich lab has shown that unlike the CHD5-NuRD complex,
MBD3-NuRD is not required for neuronal lineage commitment [107]. Similar to its role in B-cell
development, MBD3-NuRD acts as a mediator to ensure appropriate lineage decisions [107], rather
than promoting either progenitor cell proliferation or commitment to a certain lineage. The natural
question that arises from these findings (and remains to be answered) is which MBD is present in the
respective CHD3/4/5 NuRD complexes during brain development.
Finally, MBD paralogue choice is critical to stem cell pluripotency and cell fate determination. Of
particular interest in recent years has been the role of MBD3-NuRD in the context of ES cell
pluripotency and the formation of induced pluripotent stem cells (iPSCs). The first indication of the
importance of MBD3 for pluripotency was the embryonic lethality associated with knockout of
MBD3, which contrasted with the viability of MBD2 knockout mice [108]. Subsequent work from
several labs has shown that MBD3-NuRD regulates the expression of pluripotency associated genes
and is therefore essential for correct differentiation of ES cells [86, 109-113]. More controversial has
been the role of MBD3-NuRD in the formation of induced pluripotent stem cells (iPSCs); several
labs have shown that MBD3-NuRD, further identified by the Hanna lab to be CHD4-GATAD2A-
MBD3-NuRD, acts as a roadblock to iPSC reprogramming of somatic cells [114-117]. In contrast,
the Hendrich & Silva labs have shown that MBD3-NuRD facilitates the formation of iPSCs from
neural stem cells [111]. Despite these conflicting findings, it is evident that MBD3-NuRD is
necessary for ES cell fate determination. It is possible that in some cellular contexts MBD3-NuRD
prevents iPSC reprogramming, whereas in others the same complex facilitates the process.

Concluding remarks

Control of paralogue composition provides the NuRD complex with opportunities for nuanced
chromatin recognition, either through direct engagement by its various reader domains or through
paralogue specific interactions with transcription factors or other regulatory proteins. Paralogue
specific interactions with tissue restricted transcription factors, exemplified by the MTA1-BCL11B
interaction in T-cell development and the MTA2-BCL11A interaction in B-cell development,
highlight the importance of paralogue switching in eliciting tissue specific functions.

Identification of the unique NuRD permutant that drives -globin repression demonstrates the high
level of specificity that can be enacted in assembly of the complex. It is very likely, especially in light
of our paralogue correlation analysis, that this discovery is the tip of the iceberg, foreshadowing the
widespread use of paralogue switching to add an additional layer of specificity to genome regulation.
The next steps now will be to confirm these complexes experimentally, and to determine their
genomic targets and functions.

One of the most intriguing questions is that of how these specific NuRD complexes form. Examples
presented in this review suggest that post-translational modifications of paralogues might play a role,
as evidenced by the SUMOylation of GATAD2A/B. Additionally, several NuRD paralogues contain
intrinsically disordered regions. These are typically the most sequence diverse parts of each subunit,
and this diversity might serve to elicit paralogue-level specificity. Establishment of the mechanism(s)
underlying paralogue switching, the extent of its diversity and the functional consequences of such
switching will lead both to a deeper understanding of genome biology and to new opportunities to
modulate gene expression with a higher level of precision.
References

1. Marques, J.G., B.E. Gryder, B. Pavlovic, Y. Chung, Q.A. Ngo, F. Frommelt, M. Gstaiger, Y. Song, K.
Benischke, D. Laubscher, M. Wachtel, J. Khan, and B.W. Schafer, Nurd subunit chd4 regulates super‐
enhancer accessibility in rhabdomyosarcoma and represents a general tumor dependency. Elife,
2020. 9.
2. Reyes, A.A., R.D. Marcum, and Y. He, Structure and function of chromatin remodelers. J Mol Biol,
2021. 433: 166929.
3. Xue, Y., J. Wong, G.T. Moreno, M.K. Young, J. Cote, and W. Wang, Nurd, a novel complex with both
atp‐dependent chromatin‐remodeling and histone deacetylase activities. Mol Cell, 1998. 2: 851‐61.
4. Williams, C.J., T. Naito, P.G. Arco, J.R. Seavitt, S.M. Cashman, B. De Souza, X. Qi, P. Keables, U.H.
Von Andrian, and K. Georgopoulos, The chromatin remodeler mi‐2beta is required for cd4
expression and t cell development. Immunity, 2004. 20: 719‐33.
5. Sher, F., M. Hossain, D. Seruggia, V.A.C. Schoonenberg, Q. Yao, P. Cifani, L.M.K. Dassama, M.A. Cole,
C. Ren, D.S. Vinjamur, C. Macias‐Trevino, K. Luk, C. McGuckin, P.G. Schupp, M.C. Canver, R. Kurita,
Y. Nakamura, Y. Fujiwara, S.A. Wolfe, L. Pinello, T. Maeda, A. Kentsis, S.H. Orkin, and D.E. Bauer,
Rational targeting of a nurd subcomplex guided by comprehensive in situ mutagenesis. Nat Genet,
2019. 51: 1149‐1159.
6. Bornelov, S., N. Reynolds, M. Xenophontos, S. Gharbi, E. Johnstone, R. Floyd, M. Ralser, J. Signolet,
R. Loos, S. Dietmann, P. Bertone, and B. Hendrich, The nucleosome remodeling and deacetylation
complex modulates chromatin structure at sites of active transcription to fine‐tune gene expression.
Mol Cell, 2018. 71: 56‐72 e4.
7. Low, J.K., S.R. Webb, A.P. Silva, H. Saathoff, D.P. Ryan, M. Torrado, M. Brofelth, B.L. Parker, N.E.
Shepherd, and J.P. Mackay, Chd4 is a peripheral component of the nucleosome remodeling and
deacetylase complex. J Biol Chem, 2016. 291: 15853‐66.
8. Zhong, Y., B.P. Paudel, D.P. Ryan, J.K.K. Low, C. Franck, K. Patel, M.J. Bedward, M. Torrado, R.J.
Payne, A.M. van Oijen, and J.P. Mackay, Chd4 slides nucleosomes by decoupling entry‐ and exit‐side
DNA translocation. Nat Commun, 2020. 11: 1519.
9. Musselman, C.A., M.E. Lalonde, J. Cote, and T.G. Kutateladze, Perceiving the epigenetic landscape
through histone readers. Nat Struct Mol Biol, 2012. 19: 1218‐27.
10. Tencer, A.H., K.L. Cox, L. Di, J.B. Bridgers, J. Lyu, X. Wang, J.K. Sims, T.M. Weaver, H.F. Allen, Y.
Zhang, J. Gatchalian, M.A. Darcy, M.D. Gibson, J. Ikebe, W. Li, P.A. Wade, J.J. Hayes, B.D. Strahl, H.
Kono, M.G. Poirier, C.A. Musselman, and T.G. Kutateladze, Covalent modifications of histone h3k9
promote binding of chd3. Cell Rep, 2017. 21: 455‐466.
11. Musselman, C.A., J. Ramirez, J.K. Sims, R.E. Mansfield, S.S. Oliver, J.M. Denu, J.P. Mackay, P.A.
Wade, J. Hagman, and T.G. Kutateladze, Bivalent recognition of nucleosomes by the tandem phd
fingers of the chd4 atpase is required for chd4‐mediated repression. Proc Natl Acad Sci U S A, 2012.
109: 787‐92.
12. Mansfield, R.E., C.A. Musselman, A.H. Kwan, S.S. Oliver, A.L. Garske, F. Davrazou, J.M. Denu, T.G.
Kutateladze, and J.P. Mackay, Plant homeodomain (phd) fingers of chd4 are histone h3‐binding
modules with preference for unmodified h3k4 and methylated h3k9. J Biol Chem, 2011. 286: 11779‐
91.
13. Musselman, C.A., R.E. Mansfield, A.L. Garske, F. Davrazou, A.H. Kwan, S.S. Oliver, H. O'Leary, J.M.
Denu, J.P. Mackay, and T.G. Kutateladze, Binding of the chd4 phd2 finger to histone h3 is modulated
by covalent modifications. Biochem J, 2009. 423: 179‐87.
14. Oliver, S.S., C.A. Musselman, R. Srinivasan, J.P. Svaren, T.G. Kutateladze, and J.M. Denu, Multivalent
recognition of histone tails by the phd fingers of chd5. Biochemistry, 2012. 51: 6534‐44.
15. Paul, S., A. Kuo, T. Schalch, H. Vogel, L. Joshua‐Tor, W.R. McCombie, O. Gozani, M. Hammell, and
A.A. Mills, Chd5 requires phd‐mediated histone 3 binding for tumor suppression. Cell Rep, 2013. 3:
92‐102.
16. Bouazoune, K., A. Mitterweger, G. Langst, A. Imhof, A. Akhtar, P.B. Becker, and A. Brehm, The dmi‐2
chromodomains are DNA binding modules important for atp‐dependent nucleosome mobilization.
EMBO J, 2002. 21: 2430‐40.
17. Farnung, L., M. Ochmann, and P. Cramer, Nucleosome‐chd4 chromatin remodeler structure maps
human disease mutations. Elife, 2020. 9.
18. Egan, C.M., U. Nyman, J. Skotte, G. Streubel, S. Turner, D.J. O'Connell, V. Rraklli, M.J. Dolan, N.
Chadderton, K. Hansen, G.J. Farrar, K. Helin, J. Holmberg, and A.P. Bracken, Chd5 is required for
neurogenesis and has a dual role in facilitating gene expression and polycomb gene repression. Dev
Cell, 2013. 26: 223‐36.
19. Gatchalian, J., X. Wang, J. Ikebe, K.L. Cox, A.H. Tencer, Y. Zhang, N.L. Burge, L. Di, M.D. Gibson, C.A.
Musselman, M.G. Poirier, H. Kono, J.J. Hayes, and T.G. Kutateladze, Accessibility of the histone h3
tail in the nucleosome for binding of paired readers. Nat Commun, 2017. 8: 1489.
20. Marunde, M.R., H.A. Fuchs, J.M. Burg, I.K. Popova, A. Vaidya, N.W. Hall, M.J. Meiners, R. Watson,
S.A. Howard, K. Novitzky, E. McAnarney, M.A. Cheek, Z.‐W. Sun, B.J. Venters, M.‐C. Keogh, and C.A.
Musselman, Nucleosome conformation dictates the histone code. bioRxiv, 2022:
2022.02.21.481373.
21. Jain, K., M.R. Marunde, J.M. Burg, S.L. Gloor, F.M. Joseph, Z.B. Gillespie, K.L. Rodriguez, S.A.
Howard, I.K. Popova, N.W. Hall, A. Vaidya, S.W. Cooke, K.E.W. Namitz, B.C. Taylor, E.N. Weinzapfel,
M.A. Cheek, M.J. Meiners, K. Krajewski, M.S. Cosgrove, N.L. Young, M.‐C. Keogh, and B.D. Strahl, An
acetylation‐mediated chromatin switch governs h3k4 methylation read‐write capability. bioRxiv,
2022: 2022.02.28.482307.
22. Reynolds, N., M. Salmon‐Divon, H. Dvinge, A. Hynes‐Allen, G. Balasooriya, D. Leaford, A. Behrens, P.
Bertone, and B. Hendrich, Nurd‐mediated deacetylation of h3k27 facilitates recruitment of
polycomb repressive complex 2 to direct gene repression. EMBO J, 2012. 31: 593‐605.
23. Yang, Y., T. Yamada, K.K. Hill, M. Hemberg, N.C. Reddy, H.Y. Cho, A.N. Guthrie, A. Oldenborg, S.A.
Heiney, S. Ohmae, J.F. Medina, T.E. Holy, and A. Bonni, Chromatin remodeling inactivates activity
genes and regulates neural coding. Science, 2016. 353: 300‐305.
24. Goodman, J.V., T. Yamada, Y. Yang, L. Kong, D.Y. Wu, G. Zhao, H.W. Gabel, and A. Bonni, The
chromatin remodeling enzyme chd4 regulates genome architecture in the mouse brain. Nat
Commun, 2020. 11: 3419.
25. Lenz, J., R. Liefke, J. Funk, S. Shoup, A. Nist, T. Stiewe, R. Schulz, Y. Tokusumi, L. Albert, H. Raifer, K.
Forstemann, O. Vazquez, T. Tokusumi, N. Fossett, and A. Brehm, Ush regulates hemocyte‐specific
gene expression, fatty acid metabolism and cell cycle progression and cooperates with dnurd to
orchestrate hematopoiesis. PLoS Genet, 2021. 17: e1009318.
26. Montgomery, R.L., C.A. Davis, M.J. Potthoff, M. Haberland, J. Fielitz, X. Qi, J.A. Hill, J.A. Richardson,
and E.N. Olson, Histone deacetylases 1 and 2 redundantly regulate cardiac morphogenesis, growth,
and contractility. Genes Dev, 2007. 21: 1790‐802.
27. Wilting, R.H., E. Yanover, M.R. Heideman, H. Jacobs, J. Horner, J. van der Torre, R.A. DePinho, and
J.H. Dannenberg, Overlapping functions of hdac1 and hdac2 in cell cycle regulation and
haematopoiesis. EMBO J, 2010. 29: 2586‐97.
28. Haberland, M., M. Carrer, M.H. Mokalled, R.L. Montgomery, and E.N. Olson, Redundant control of
adipogenesis by histone deacetylases 1 and 2. J Biol Chem, 2010. 285: 14663‐70.
29. Dovey, O.M., C.T. Foster, and S.M. Cowley, Histone deacetylase 1 (hdac1), but not hdac2, controls
embryonic stem cell differentiation. Proc Natl Acad Sci U S A, 2010. 107: 8242‐7.
30. Senese, S., K. Zaragoza, S. Minardi, I. Muradore, S. Ronzoni, A. Passafaro, L. Bernard, G.F. Draetta,
M. Alcalay, C. Seiser, and S. Chiocca, Role for histone deacetylase 1 in human tumor cell
proliferation. Mol Cell Biol, 2007. 27: 4784‐95.
31. Low, J.K.K., A.P.G. Silva, M. Sharifi Tabar, M. Torrado, S.R. Webb, B.L. Parker, M. Sana, C. Smits, J.W.
Schmidberger, L. Brillault, M.J. Jackman, D.C. Williams, Jr., G.A. Blobel, S.B. Hake, N.E. Shepherd,
M.J. Landsberg, and J.P. Mackay, The nucleosome remodeling and deacetylase complex has an
asymmetric, dynamic, and modular architecture. Cell Rep, 2020. 33: 108450.
32. Millard, C.J., N. Varma, A. Saleh, K. Morris, P.J. Watson, A.R. Bottrill, L. Fairall, C.J. Smith, and J.W.
Schwabe, The structure of the core nurd repression complex provides insights into its interaction
with chromatin. Elife, 2016. 5: e13941.
33. Fan, H., J. Lu, Y. Guo, D. Li, Z.M. Zhang, Y.H. Tsai, W.C. Pi, J.H. Ahn, W. Gong, Y. Xiang, D.F. Allison, H.
Geng, S. He, Y. Diao, W.Y. Chen, B.D. Strahl, L. Cai, J. Song, and G.G. Wang, Bahcc1 binds h3k27me3
 via a conserved bah module to mediate gene silencing and oncogenesis. Nat Genet, 2020. 52: 1384‐
1396.
34. Kuo, A.J., J. Song, P. Cheung, S. Ishibe‐Murakami, S. Yamazoe, J.K. Chen, D.J. Patel, and O. Gozani,
The bah domain of orc1 links h4k20me2 to DNA replication licensing and meier‐gorlin syndrome.
Nature, 2012. 484: 115‐9.
35. Du, J., X. Zhong, Y.V. Bernatavichute, H. Stroud, S. Feng, E. Caro, A.A. Vashisht, J. Terragni, H.G.
Chin, A. Tu, J. Hetzel, J.A. Wohlschlegel, S. Pradhan, D.J. Patel, and S.E. Jacobsen, Dual binding of
chromomethylase domains to h3k9me2‐containing nucleosomes directs DNA methylation in plants.
Cell, 2012. 151: 167‐80.
36. Millard, C.J., P.J. Watson, I. Celardo, Y. Gordiyenko, S.M. Cowley, C.V. Robinson, L. Fairall, and J.W.
Schwabe, Class i hdacs share a common mechanism of regulation by inositol phosphates. Mol Cell,
2013. 51: 57‐67.
37. Schmidberger, J.W., M. Sharifi Tabar, M. Torrado, A.P. Silva, M.J. Landsberg, L. Brillault, S. AlQarni,
Y.C. Zeng, B.L. Parker, J.K. Low, and J.P. Mackay, The mta1 subunit of the nucleosome remodeling
and deacetylase complex can recruit two copies of rbbp4/7. Protein Sci, 2016. 25: 1472‐82.
38. Alqarni, S.S., A. Murthy, W. Zhang, M.R. Przewloka, A.P. Silva, A.A. Watson, S. Lejon, X.Y. Pei, A.H.
Smits, S.L. Kloet, H. Wang, N.E. Shepherd, P.H. Stokes, G.A. Blobel, M. Vermeulen, D.M. Glover, J.P.
Mackay, and E.D. Laue, Insight into the architecture of the nurd complex: Structure of the rbap48‐
mta1 subcomplex. J Biol Chem, 2014. 289: 21844‐55.
39. Link, S., R.M.M. Spitzer, M. Sana, M. Torrado, M.C. Volker‐Albert, E.C. Keilhauer, T. Burgold, S.
Punzeler, J.K.K. Low, I. Lindstrom, A. Nist, C. Regnard, T. Stiewe, B. Hendrich, A. Imhof, M. Mann,
J.P. Mackay, M. Bartkuhn, and S.B. Hake, Pwwp2a binds distinct chromatin moieties and interacts
with an mta1‐specific core nurd complex. Nat Commun, 2018. 9: 4300.
40. Murzina, N.V., X.Y. Pei, W. Zhang, M. Sparkes, J. Vicente‐Garcia, J.V. Pratap, S.H. McLaughlin, T.R.
Ben‐Shahar, A. Verreault, B.F. Luisi, and E.D. Laue, Structural basis for the recognition of histone h4
by the histone‐chaperone rbap46. Structure, 2008. 16: 1077‐85.
41. Hong, W., M. Nakazawa, Y.Y. Chen, R. Kori, C.R. Vakoc, C. Rakowski, and G.A. Blobel, Fog‐1 recruits
the nurd repressor complex to mediate transcriptional repression by gata‐1. EMBO J, 2005. 24:
2367‐78.
42. Lauberth, S.M. and M. Rauchman, A conserved 12‐amino acid motif in sall1 recruits the nucleosome
remodeling and deacetylase corepressor complex. J Biol Chem, 2006. 281: 23922‐31.
43. Lejon, S., S.Y. Thong, A. Murthy, S. AlQarni, N.V. Murzina, G.A. Blobel, E.D. Laue, and J.P. Mackay,
Insights into association of the nurd complex with fog‐1 from the crystal structure of an rbap48.Fog‐
1 complex. J Biol Chem, 2011. 286: 1196‐203.
44. Liu, Z., F. Li, K. Ruan, J. Zhang, Y. Mei, J. Wu, and Y. Shi, Structural and functional insights into the
human borjeson‐forssman‐lehmann syndrome‐associated protein phf6. J Biol Chem, 2014. 289:
10069‐83.
45. Migliori, V., J. Muller, S. Phalke, D. Low, M. Bezzi, W.C. Mok, S.K. Sahu, J. Gunaratne, P. Capasso, C.
Bassi, V. Cecatiello, A. De Marco, W. Blackstock, V. Kuznetsov, B. Amati, M. Mapelli, and E.
Guccione, Symmetric dimethylation of h3r2 is a newly identified histone mark that supports
euchromatin maintenance. Nat Struct Mol Biol, 2012. 19: 136‐44.
46. Mu, W., N.S. Murcia, K.N. Smith, D.U. Menon, D. Yee, and T. Magnuson, Rbbp4 modulates gene
activity through acetylation and methylation of histone h3 lysine 27. bioRxiv, 2021:
2021.09.09.459568.
47. Cramer, J.M., J.N. Scarsdale, N.M. Walavalkar, W.A. Buchwald, G.D. Ginder, and D.C. Williams, Jr.,
Probing the dynamic distribution of bound states for methylcytosine‐binding domains on DNA. J Biol
Chem, 2014. 289: 1294‐302.
48. Liu, K., M. Lei, Z. Wu, B. Gan, H. Cheng, Y. Li, and J. Min, Structural analyses reveal that mbd3 is a
methylated cg binder. FEBS J, 2019. 286: 3240‐3254.
49. Liu, K., C. Xu, M. Lei, A. Yang, P. Loppnau, T.R. Hughes, and J. Min, Structural basis for the ability of
mbd domains to bind methyl‐cg and tg sites in DNA. J Biol Chem, 2018. 293: 7344‐7354.
50. Gunther, K., M. Rust, J. Leers, T. Boettger, M. Scharfe, M. Jarek, M. Bartkuhn, and R. Renkawitz,
Differential roles for mbd2 and mbd3 at methylated cpg islands, active promoters and binding to
exon sequences. Nucleic Acids Res, 2013. 41: 3010‐21.
51. Menafra, R., A.B. Brinkman, F. Matarese, G. Franci, S.J. Bartels, L. Nguyen, T. Shimbo, P.A. Wade,
N.C. Hubner, and H.G. Stunnenberg, Genome‐wide binding of mbd2 reveals strong preference for
highly methylated loci. PLoS One, 2014. 9: e99603.
52. Yildirim, O., R. Li, J.H. Hung, P.B. Chen, X. Dong, L.S. Ee, Z. Weng, O.J. Rando, and T.G. Fazzio,
Mbd3/nurd complex regulates expression of 5‐hydroxymethylcytosine marked genes in embryonic
stem cells. Cell, 2011. 147: 1498‐510.
53. Buchmuller, B.C., B. Kosel, and D. Summerer, Complete profiling of methyl‐cpg‐binding domains for
combinations of cytosine modifications at cpg dinucleotides reveals differential read‐out in normal
and rett‐associated states. Sci Rep, 2020. 10: 4053.
54. Gnanapragasam, M.N., J.N. Scarsdale, M.L. Amaya, H.D. Webb, M.A. Desai, N.M. Walavalkar, S.Z.
Wang, S. Zu Zhu, G.D. Ginder, and D.C. Williams, Jr., P66alpha‐mbd2 coiled‐coil interaction and
recruitment of mi‐2 are critical for globin gene silencing by the mbd2‐nurd complex. Proc Natl Acad
Sci U S A, 2011. 108: 7487‐92.
55. Zhang, W., A. Aubert, J.M. Gomez de Segura, M. Karuppasamy, S. Basu, A.S. Murthy, A. Diamante,
T.A. Drury, J. Balmer, J. Cramard, A.A. Watson, D. Lando, S.F. Lee, M. Palayret, S.L. Kloet, A.H. Smits,
M.J. Deery, M. Vermeulen, B. Hendrich, D. Klenerman, C. Schaffitzel, I. Berger, and E.D. Laue, The
nucleosome remodeling and deacetylase complex nurd is built from preformed catalytically active
sub‐modules. J Mol Biol, 2016. 428: 2931‐42.
56. Torrado, M., J.K.K. Low, A.P.G. Silva, J.W. Schmidberger, M. Sana, M. Sharifi Tabar, M.E. Isilak, C.S.
Winning, C. Kwong, M.J. Bedward, M.J. Sperlazza, D.C. Williams, Jr., N.E. Shepherd, and J.P. Mackay,
Refinement of the subunit interaction network within the nucleosome remodelling and deacetylase
(nurd) complex. FEBS J, 2017. 284: 4216‐4232.
57. Brackertz, M., J. Boeke, R. Zhang, and R. Renkawitz, Two highly related p66 proteins comprise a new
family of potent transcriptional repressors interacting with mbd2 and mbd3. J Biol Chem, 2002. 277:
40958‐66.
58. Gong, Z., M. Brackertz, and R. Renkawitz, Sumo modification enhances p66‐mediated
transcriptional repression of the mi‐2/nurd complex. Mol Cell Biol, 2006. 26: 4519‐28.
59. Ding, X., S. Liu, M. Tian, W. Zhang, T. Zhu, D. Li, J. Wu, H. Deng, Y. Jia, W. Xie, H. Xie, and J.S. Guan,
Activity‐induced histone modifications govern neurexin‐1 mrna splicing and memory preservation.
Nat Neurosci, 2017. 20: 690‐699.
60. Nusinow, D.P., J. Szpyt, M. Ghandi, C.M. Rose, E.R. McDonald, 3rd, M. Kalocsay, J. Jane‐Valbuena, E.
Gelfand, D.K. Schweppe, M. Jedrychowski, J. Golji, D.A. Porter, T. Rejtar, Y.K. Wang, G.V. Kryukov, F.
Stegmeier, B.K. Erickson, L.A. Garraway, W.R. Sellers, and S.P. Gygi, Quantitative proteomics of the
cancer cell line encyclopedia. Cell, 2020. 180: 387‐402 e16.
61. Wisniewski, J.R., M.Y. Hein, J. Cox, and M. Mann, A "proteomic ruler" for protein copy number and
concentration estimation without spike‐in standards. Mol Cell Proteomics, 2014. 13: 3497‐506.
62. Meyers, R.M., J.G. Bryan, J.M. McFarland, B.A. Weir, A.E. Sizemore, H. Xu, N.V. Dharia, P.G.
Montgomery, G.S. Cowley, S. Pantel, A. Goodale, Y. Lee, L.D. Ali, G. Jiang, R. Lubonja, W.F.
Harrington, M. Strickland, T. Wu, D.C. Hawes, V.A. Zhivich, M.R. Wyatt, Z. Kalani, J.J. Chang, M.
Okamoto, K. Stegmaier, T.R. Golub, J.S. Boehm, F. Vazquez, D.E. Root, W.C. Hahn, and A. Tsherniak,
Computational correction of copy number effect improves specificity of crispr‐cas9 essentiality
screens in cancer cells. Nat Genet, 2017. 49: 1779‐1784.
63. Centore, R.C., G.J. Sandoval, L.M.M. Soares, C. Kadoch, and H.M. Chan, Mammalian swi/snf
chromatin remodeling complexes: Emerging mechanisms and therapeutic strategies. Trends Genet,
2020. 36: 936‐950.
64. Mashtalir, N., H.T. Dao, A. Sankar, H. Liu, A.J. Corin, J.D. Bagert, E.J. Ge, A.R. D'Avino, M. Filipovski,
B.C. Michel, G.P. Dann, T.W. Muir, and C. Kadoch, Chromatin landscape signals differentially dictate
the activities of mswi/snf family complexes. Science, 2021. 373: 306‐315.
65. Ho, L., J.L. Ronan, J. Wu, B.T. Staahl, L. Chen, A. Kuo, J. Lessard, A.I. Nesvizhskii, J. Ranish, and G.R.
Crabtree, An embryonic stem cell chromatin remodeling complex, esbaf, is essential for embryonic
stem cell self‐renewal and pluripotency. Proc Natl Acad Sci U S A, 2009. 106: 5181‐6.
66. Macinkovic, I., I. Theofel, T. Hundertmark, K. Kovac, S. Awe, J. Lenz, I. Forne, B. Lamp, A. Nist, A.
Imhof, T. Stiewe, R. Renkawitz‐Pohl, C. Rathke, and A. Brehm, Distinct corest complexes act in a cell‐
type‐specific manner. Nucleic Acids Res, 2019. 47: 11649‐11666.
67. Zhou, M., K. Zhou, L. Cheng, X. Chen, J. Wang, X.M. Wang, Y. Zhang, Q. Yu, S. Zhang, D. Wang, L.
Huang, M. Huang, D. Ma, T. Cheng, C.Y. Wang, W. Yuan, and J. Zhou, Mbd2 ablation impairs
lymphopoiesis and impedes progression and maintenance of t‐all. Cancer Res, 2018. 78: 1632‐1642.
68. Zhou, K., M. Zhou, L. Cheng, X. Chen, X. Wang, Y. Chu, Q. Yu, S. Zhang, N. Wang, L. Zhao, D. Wang, L.
Huang, C. Wang, W. Yuan, and J. Zhou, Loss of mbd2 attenuates mll‐af9‐driven leukemogenesis by
suppressing the leukemic cell cycle via cdkn1c. Oncogenesis, 2021. 10: 79.
69. Shao, S., H. Cao, Z. Wang, D. Zhou, C. Wu, S. Wang, D. Xia, and D. Zhang, Chd4/nurd complex
regulates complement gene expression and correlates with cd8 t cell infiltration in human
hepatocellular carcinoma. Clin Epigenetics, 2020. 12: 31.
70. Zhang, M., H. Li, D. Zou, and J. Gao, Ruguo key genes and tumor driving factors identification of
bladder cancer based on the rna‐seq profile. Onco Targets Ther, 2016. 9: 2717‐23.
71. Yang, W., Q. Cai, V.W. Lui, P.A. Everley, J. Kim, N. Bhola, K.M. Quesnelle, B.R. Zetter, H. Steen, M.R.
Freeman, and J.R. Grandis, Quantitative proteomics analysis reveals molecular networks regulated
by epidermal growth factor receptor level in head and neck cancer. J Proteome Res, 2010. 9: 3073‐
82.
72. Zhao, S., S. Bellone, S. Lopez, D. Thakral, C. Schwab, D.P. English, J. Black, E. Cocco, J. Choi, L.
Zammataro, F. Predolini, E. Bonazzoli, M. Bi, N. Buza, P. Hui, S. Wong, M. Abu‐Khalaf, A. Ravaggi, E.
Bignotti, E. Bandiera, C. Romani, P. Todeschini, R. Tassi, L. Zanotti, F. Odicino, S. Pecorelli, C.
Donzelli, L. Ardighieri, F. Facchetti, M. Falchetti, D.A. Silasi, E. Ratner, M. Azodi, P.E. Schwartz, S.
Mane, R. Angioli, C. Terranova, C.M. Quick, B. Edraki, K. Bilguvar, M. Lee, M. Choi, A.L. Stiegler, T.J.
Boggon, J. Schlessinger, R.P. Lifton, and A.D. Santin, Mutational landscape of uterine and ovarian
carcinosarcomas implicates histone genes in epithelial‐mesenchymal transition. Proc Natl Acad Sci
U S A, 2016. 113: 12238‐12243.
73. Ryan, C.J., S. Kennedy, I. Bajrami, D. Matallanas, and C.J. Lord, A compendium of co‐regulated
protein complexes in breast cancer reveals collateral loss events. Cell Syst, 2017. 5: 399‐409 e5.
74. Lapek, J.D., Jr., P. Greninger, R. Morris, A. Amzallag, I. Pruteanu‐Malinici, C.H. Benes, and W. Haas,
Detection of dysregulated protein‐association networks by high‐throughput proteomics predicts
cancer vulnerabilities. Nat Biotechnol, 2017. 35: 983‐989.
75. Roumeliotis, T.I., S.P. Williams, E. Goncalves, C. Alsinet, M. Del Castillo Velasco‐Herrera, N. Aben,
F.Z. Ghavidel, M. Michaut, M. Schubert, S. Price, J.C. Wright, L. Yu, M. Yang, R. Dienstmann, J.
Guinney, P. Beltrao, A. Brazma, M. Pardo, O. Stegle, D.J. Adams, L. Wessels, J. Saez‐Rodriguez, U.
McDermott, and J.S. Choudhary, Genomic determinants of protein abundance variation in
colorectal cancer cells. Cell Rep, 2017. 20: 2201‐2214.
76. Pozniak, Y., N. Balint‐Lahat, J.D. Rudolph, C. Lindskog, R. Katzir, C. Avivi, F. Ponten, E. Ruppin, I.
Barshack, and T. Geiger, System‐wide clinical proteomics of breast cancer reveals global remodeling
of tissue homeostasis. Cell Syst, 2016. 2: 172‐84.
77. Ori, A., M. Iskar, K. Buczak, P. Kastritis, L. Parca, A. Andres‐Pons, S. Singer, P. Bork, and M. Beck,
Spatiotemporal variation of mammalian protein complex stoichiometries. Genome Biol, 2016. 17:
47.
78. Cismasiu, V.B., K. Adamo, J. Gecewicz, J. Duque, Q. Lin, and D. Avram, Bcl11b functionally associates
with the nurd complex in t lymphocytes to repress targeted promoter. Oncogene, 2005. 24: 6753‐
64.
79. Lu, X., C.S. Chu, T. Fang, V. Rayon‐Estrada, F. Fang, A. Patke, Y. Qian, S.H. Clarke, A.M. Melnick, Y.
Zhang, F.N. Papavasiliou, and R.G. Roeder, Mta2/nurd regulates b cell development and cooperates
with oca‐b in controlling the pre‐b to immature b cell transition. Cell Rep, 2019. 28: 472‐485 e5.
80. Xu, J., D.E. Bauer, M.A. Kerenyi, T.D. Vo, S. Hou, Y.J. Hsu, H. Yao, J.J. Trowbridge, G. Mandel, and
S.H. Orkin, Corepressor‐dependent silencing of fetal hemoglobin expression by bcl11a. Proc Natl
Acad Sci U S A, 2013. 110: 6518‐23.
81. Masuda, T., X. Wang, M. Maeda, M.C. Canver, F. Sher, A.P. Funnell, C. Fisher, M. Suciu, G.E. Martyn,
L.J. Norton, C. Zhu, R. Kurita, Y. Nakamura, J. Xu, D.R. Higgs, M. Crossley, D.E. Bauer, S.H. Orkin, P.V.
Kharchenko, and T. Maeda, Transcription factors lrf and bcl11a independently repress expression of
fetal hemoglobin. Science, 2016. 351: 285‐9.
82. Fujita, N., D.L. Jaye, C. Geigerman, A. Akyildiz, M.R. Mooney, J.M. Boss, and P.A. Wade, Mta3 and
the mi‐2/nurd complex regulate cell fate during b lymphocyte differentiation. Cell, 2004. 119: 75‐86.
83. Xu, J., V.G. Sankaran, M. Ni, T.F. Menne, R.V. Puram, W. Kim, and S.H. Orkin, Transcriptional
silencing of {gamma}‐globin by bcl11a involves long‐range interactions and cooperation with sox6.
Genes Dev, 2010. 24: 783‐98.
84. Loughran, S.J., F. Comoglio, F.K. Hamey, A. Giustacchini, Y. Errami, E. Earp, B. Gottgens, S.E.W.
Jacobsen, A.J. Mead, B. Hendrich, and A.R. Green, Mbd3/nurd controls lymphoid cell fate and
inhibits tumorigenesis by repressing a b cell transcriptional program. J Exp Med, 2017. 214: 3085‐
3104.
85. Olivieri, D., S. Paramanathan, A.F. Bardet, D. Hess, S.A. Smallwood, U. Elling, and J. Betschinger, The
btb‐domain transcription factor zbtb2 recruits chromatin remodelers and a histone chaperone
during the exit from pluripotency. J Biol Chem, 2021. 297: 100947.
86. Kloet, S.L., I.D. Karemaker, L. van Voorthuijsen, R.G.H. Lindeboom, M.P. Baltissen, R.R. Edupuganti,
D.W. Poramba‐Liyanage, P. Jansen, and M. Vermeulen, Nurd‐interacting protein zfp296 regulates
genome‐wide nurd localization and differentiation of mouse embryonic stem cells. Nat Commun,
2018. 9: 4588.
87. Ferreira, R., K. Ohneda, M. Yamamoto, and S. Philipsen, Gata1 function, a paradigm for
transcription factors in hematopoiesis. Mol Cell Biol, 2005. 25: 1215‐27.
88. Aguilera, C., K. Nakagawa, R. Sancho, A. Chakraborty, B. Hendrich, and A. Behrens, C‐jun n‐terminal
phosphorylation antagonises recruitment of the mbd3/nurd repressor complex. Nature, 2011. 469:
231‐5.
89. Mohd‐Sarip, A., M. Teeuwssen, A.G. Bot, M.J. De Herdt, S.M. Willems, R.J. Baatenburg de Jong,
L.H.J. Looijenga, D. Zatreanu, K. Bezstarosti, J. van Riet, E. Oole, W.F.J. van Ijcken, H.J.G. van de
Werken, J.A. Demmers, R. Fodde, and C.P. Verrijzer, Doc1‐dependent recruitment of nurd reveals
antagonism with swi/snf during epithelial‐mesenchymal transition in oral cancer cells. Cell Rep,
2017. 20: 61‐75.
90. Spruijt, C.G., M.S. Luijsterburg, R. Menafra, R.G. Lindeboom, P.W. Jansen, R.R. Edupuganti, M.P.
Baltissen, W.W. Wiegant, M.C. Voelker‐Albert, F. Matarese, A. Mensinga, I. Poser, H.R. Vos, H.G.
Stunnenberg, H. van Attikum, and M. Vermeulen, Zmynd8 co‐localizes with nurd on target genes
and regulates poly(adp‐ribose)‐dependent recruitment of gatad2a/nurd to sites of DNA damage.
Cell Rep, 2016. 17: 783‐798.
91. Gong, F., L.Y. Chiu, B. Cox, F. Aymard, T. Clouaire, J.W. Leung, M. Cammarata, M. Perez, P. Agarwal,
J.S. Brodbelt, G. Legube, and K.M. Miller, Screen identifies bromodomain protein zmynd8 in
chromatin recognition of transcription‐associated DNA damage that promotes homologous
recombination. Genes Dev, 2015. 29: 197‐211.
92. Hosoya, T., M. Clifford, R. Losson, O. Tanabe, and J.D. Engel, Trim28 is essential for erythroblast
differentiation in the mouse. Blood, 2013. 122: 3798‐807.
93. Schultz, D.C., J.R. Friedman, and F.J. Rauscher, 3rd, Targeting histone deacetylase complexes via
krab‐zinc finger proteins: The phd and bromodomains of kap‐1 form a cooperative unit that recruits
a novel isoform of the mi‐2alpha subunit of nurd. Genes Dev, 2001. 15: 428‐43.
94. Liu, Z., F. Li, B. Zhang, S. Li, J. Wu, and Y. Shi, Structural basis of plant homeodomain finger 6 (phf6)
recognition by the retinoblastoma binding protein 4 (rbbp4) component of the nucleosome
remodeling and deacetylase (nurd) complex. J Biol Chem, 2015. 290: 6630‐8.
95. Todd, M.A. and D.J. Picketts, Phf6 interacts with the nucleosome remodeling and deacetylation
(nurd) complex. J Proteome Res, 2012. 11: 4326‐37.
96. Wang, J., J.W. Leung, Z. Gong, L. Feng, X. Shi, and J. Chen, Phf6 regulates cell cycle progression by
suppressing ribosomal rna synthesis. J Biol Chem, 2013. 288: 3174‐83.
97. Lower, K.M., G. Turner, B.A. Kerr, K.D. Mathews, M.A. Shaw, A.K. Gedeon, S. Schelley, H.E. Hoyme,
S.M. White, M.B. Delatycki, A.K. Lampe, J. Clayton‐Smith, H. Stewart, C.M. van Ravenswaay, B.B. de
Vries, B. Cox, M. Grompe, S. Ross, P. Thomas, J.C. Mulley, and J. Gecz, Mutations in phf6 are
associated with borjeson‐forssman‐lehmann syndrome. Nat Genet, 2002. 32: 661‐5.
98. Van Vlierberghe, P., T. Palomero, H. Khiabanian, J. Van der Meulen, M. Castillo, N. Van Roy, B. De
Moerloose, J. Philippe, S. Gonzalez‐Garcia, M.L. Toribio, T. Taghon, L. Zuurbier, B. Cauwelier, C.J.
Harrison, C. Schwab, M. Pisecker, S. Strehl, A.W. Langerak, J. Gecz, E. Sonneveld, R. Pieters, E.
Paietta, J.M. Rowe, P.H. Wiernik, Y. Benoit, J. Soulier, B. Poppe, X. Yao, C. Cordon‐Cardo, J.
 Meijerink, R. Rabadan, F. Speleman, and A. Ferrando, Phf6 mutations in t‐cell acute lymphoblastic
leukemia. Nat Genet, 2010. 42: 338‐42.
99. Amaya, M., M. Desai, M.N. Gnanapragasam, S.Z. Wang, S. Zu Zhu, D.C. Williams, Jr., and G.D.
Ginder, Mi2beta‐mediated silencing of the fetal gamma‐globin gene in adult erythroid cells. Blood,
2013. 121: 3493‐501.
100. Yu, X., A. Azzo, S.M. Bilinovich, X. Li, M. Dozmorov, R. Kurita, Y. Nakamura, D.C. Williams, Jr., and
G.D. Ginder, Disruption of the mbd2‐nurd complex but not mbd3‐nurd induces high level hbf
expression in human adult erythroid cells. Haematologica, 2019. 104: 2361‐2371.
101. Rupon, J.W., S.Z. Wang, K. Gaensler, J. Lloyd, and G.D. Ginder, Methyl binding domain protein 2
mediates gamma‐globin gene silencing in adult human betayac transgenic mice. Proc Natl Acad Sci
U S A, 2006. 103: 6617‐22.
102. Kransdorf, E.P., S.Z. Wang, S.Z. Zhu, T.B. Langston, J.W. Rupon, and G.D. Ginder, Mbd2 is a critical
component of a methyl cytosine‐binding protein complex isolated from primary erythroid cells.
Blood, 2006. 108: 2836‐45.
103. Liang, Y., X. Zhang, Y. Liu, L. Wang, Y. Ye, X. Tan, J. Pu, Q. Zhang, X. Bao, X. Wei, D. Li, R. Kurita, Y.
Nakamura, D. Li, and X. Xu, Gata zinc finger domain‐containing protein 2a (gatad2a) deficiency
reactivates fetal haemoglobin in patients with beta‐thalassaemia through impaired formation of
methyl‐binding domain protein 2 (mbd2)‐containing nucleosome remodelling and deacetylation
(nurd) complex. Br J Haematol, 2021. 193: 1220‐1227.
104. Desai, M.A., H.D. Webb, L.M. Sinanan, J.N. Scarsdale, N.M. Walavalkar, G.D. Ginder, and D.C.
Williams, Jr., An intrinsically disordered region of methyl‐cpg binding domain protein 2 (mbd2)
recruits the histone deacetylase core of the nurd complex. Nucleic Acids Res, 2015. 43: 3100‐13.
105. Potts, R.C., P. Zhang, A.L. Wurster, P. Precht, M.R. Mughal, W.H. Wood, 3rd, Y. Zhang, K.G. Becker,
M.P. Mattson, and M.J. Pazin, Chd5, a brain‐specific paralog of mi2 chromatin remodeling enzymes,
regulates expression of neuronal genes. PLoS One, 2011. 6: e24515.
106. Nitarska, J., J.G. Smith, W.T. Sherlock, M.M. Hillege, A. Nott, W.D. Barshop, A.A. Vashisht, J.A.
Wohlschlegel, R. Mitter, and A. Riccio, A functional switch of nurd chromatin remodeling complex
subunits regulates mouse cortical development. Cell Rep, 2016. 17: 1683‐1698.
107. Knock, E., J. Pereira, P.D. Lombard, A. Dimond, D. Leaford, F.J. Livesey, and B. Hendrich, The methyl
binding domain 3/nucleosome remodelling and deacetylase complex regulates neural cell fate
determination and terminal differentiation in the cerebral cortex. Neural Dev, 2015. 10: 13.
108. Hendrich, B., J. Guy, B. Ramsahoye, V.A. Wilson, and A. Bird, Closely related proteins mbd2 and
mbd3 play distinctive but interacting roles in mouse development. Genes Dev, 2001. 15: 710‐23.
109. Kaji, K., I.M. Caballero, R. MacLeod, J. Nichols, V.A. Wilson, and B. Hendrich, The nurd component
mbd3 is required for pluripotency of embryonic stem cells. Nat Cell Biol, 2006. 8: 285‐92.
110. Kaji, K., J. Nichols, and B. Hendrich, Mbd3, a component of the nurd co‐repressor complex, is
required for development of pluripotent cells. Development, 2007. 134: 1123‐32.
111. Dos Santos, R.L., L. Tosti, A. Radzisheuskaya, I.M. Caballero, K. Kaji, B. Hendrich, and J.C.R. Silva,
Mbd3/nurd facilitates induction of pluripotency in a context‐dependent manner. Cell Stem Cell,
2014. 15: 392.
112. Ragheb, R., S. Gharbi, J. Cramard, O. Ogundele, S.L. Kloet, T. Burgold, M. Vermeulen, N. Reynolds,
and B. Hendrich, Differential regulation of lineage commitment in human and mouse primed
pluripotent stem cells by the nucleosome remodelling and deacetylation complex. Stem Cell Res,
2020. 46: 101867.
113. Reynolds, N., P. Latos, A. Hynes‐Allen, R. Loos, D. Leaford, A. O'Shaughnessy, O. Mosaku, J. Signolet,
P. Brennecke, T. Kalkan, I. Costello, P. Humphreys, W. Mansfield, K. Nakagawa, J. Strouboulis, A.
Behrens, P. Bertone, and B. Hendrich, Nurd suppresses pluripotency gene expression to promote
transcriptional heterogeneity and lineage commitment. Cell Stem Cell, 2012. 10: 583‐94.
114. Rais, Y., A. Zviran, S. Geula, O. Gafni, E. Chomsky, S. Viukov, A.A. Mansour, I. Caspi, V. Krupalnik, M.
Zerbib, I. Maza, N. Mor, D. Baran, L. Weinberger, D.A. Jaitin, D. Lara‐Astiaso, R. Blecher‐Gonen, Z.
Shipony, Z. Mukamel, T. Hagai, S. Gilad, D. Amann‐Zalcenstein, A. Tanay, I. Amit, N. Novershtern,
and J.H. Hanna, Deterministic direct reprogramming of somatic cells to pluripotency. Nature, 2013.
502: 65‐70.
115. Fidalgo, M., F. Faiola, C.F. Pereira, J. Ding, A. Saunders, J. Gingold, C. Schaniel, I.R. Lemischka, J.C.
Silva, and J. Wang, Zfp281 mediates nanog autorepression through recruitment of the nurd complex
and inhibits somatic cell reprogramming. Proc Natl Acad Sci U S A, 2012. 109: 16202‐7.
116. Luo, M., T. Ling, W. Xie, H. Sun, Y. Zhou, Q. Zhu, M. Shen, L. Zong, G. Lyu, Y. Zhao, T. Ye, J. Gu, W.
Tao, Z. Lu, and I. Grummt, Nurd blocks reprogramming of mouse somatic cells into pluripotent stem
cells. Stem Cells, 2013. 31: 1278‐86.
117. Mor, N., Y. Rais, D. Sheban, S. Peles, A. Aguilera‐Castrejon, A. Zviran, D. Elinger, S. Viukov, S. Geula,
V. Krupalnik, M. Zerbib, E. Chomsky, L. Lasman, T. Shani, J. Bayerl, O. Gafni, S. Hanna, J.D.
Buenrostro, T. Hagai, H. Masika, G. Vainorius, Y. Bergman, W.J. Greenleaf, M.A. Esteban, U. Elling,
Y. Levin, R. Massarwa, Y. Merbl, N. Novershtern, and J.H. Hanna, Neutralizing gatad2a‐chd4‐
mbd3/nurd complex facilitates deterministic induction of naive pluripotency. Cell Stem Cell, 2018.
23: 412‐425 e10.