Article

The Structure of the ZMYN

D8/Drebrin ComplexSuggests a Cytoplasmic Sequestering Mechanism ofZMYND8 by Drebrin

Graphical Abstract

Highlights

d ZMYND8 PHD-BROMO-PWWP tandem specifically binds to

non-histone protein Drebrin

d The ZMYND8 PHD-BROMO-PWWP tandem/Drebrin ADF-H

domain complex structure is solved

d Drebrin binding blocks histone marker access to ZMYND8

d Drebrin can sequester ZMYND8 in cytoplasm and thus

regulate its nuclear activity

Yao et al., 2017, Structure 25, 1657–1666November 7, 2017 ª 2017 Elsevier Ltd.http://dx.doi.org/10.1016/j.str.2017.08.014

Authors

NingningYao, JianchaoLi, HaiyangLiu,

Jun Wan, Wei Liu, Mingjie Zhang

Correspondenceliuw@ust.hk (W.L.),mzhang@ust.hk (M.Z.)

In Brief

The crystal structure of ZMYND8 PHD-

BROMO-PWWP tandem in complex

Drebrin ADF-H domain solved by Yao

et al. reveals how ZMYND8 can

specifically bind to non-histone proteins

and how the histone marker reader might

be sequestered in cytoplasm by

specifically binding to an actin binding

protein.

Data Resources

5Y1Z

Structure

Article

The Structure of the ZMYND8/DrebrinComplex Suggests a CytoplasmicSequestering Mechanism of ZMYND8 by DrebrinNingning Yao,1,4 Jianchao Li,2,4 Haiyang Liu,1,2 Jun Wan,1,2 Wei Liu,1,2,* and Mingjie Zhang1,2,3,5,*1Shenzhen Key Laboratory for Neuronal Structural Biology, Biomedical Research Institute, Shenzhen Peking University-The Hong Kong

University of Science and Technology Medical Center, Shenzhen 518036, China2Division of Life Science, State Key Laboratory of Molecular Neuroscience, Hong Kong University of Science and Technology, Kowloon,Hong Kong3Center of Systems Biology and Human Health, School of Science and Institute for Advanced Study, Hong Kong University of Science and

Technology, Clear Water Bay, Kowloon, Hong Kong4These authors contributed equally5Lead Contact

*Correspondence: liuw@ust.hk (W.L.), mzhang@ust.hk (M.Z.)

http://dx.doi.org/10.1016/j.str.2017.08.014

SUMMARY

Malfunctions of the actin binding protein Drebrinhave been implicated in various human diseasessuch as Alzheimer’s disease, cognitive impairments,cancer, and digestive disorders, though with poorlyunderstood mechanisms. The ADF-H domain of Dre-brin does not contain actin binding and depolymeriz-ing activity. Instead, it binds to a histone markerreader, ZMYND8. Here we present the high-resolu-tion crystal structure of Drebrin ADF-H in complexwith the ZMYND8 PHD-BROMO-PWWP tandem,elucidating the mechanistic basis governing thehighly specific interaction of the two proteins. Thestructure reveals that the ZMYND8 PHD-BROMO-PWWP tandem forms a structural supramodule thatis necessary for binding to Drebrin ADF-H. DrebrinADF-H competes with modified histone for bindingto ZMYND8. Binding of Drebrin can shuttle ZMYND8from nucleus to cytoplasm in living cells. Takentogether, our study uncovers a non-actin target bind-ing mode for ADF-H domains, and suggests thatDrebrin may regulate activities of epigenetic readerZMYND8 via its cytoplasmic sequestration.

INTRODUCTION

Shuttling proteins between nuclei and cytoplasm is a general

strategy used by eukaryotic cells to control spatially the activ-

ities of proteins (Gama-Carvalho and Carmo-Fonseca, 2001;

Xu and Massague, 2004). Neurons are no exceptions. In fact

the extremely polarized morphology of neurons often requires

more coordinated protein distributions between different com-

partments of cells including nuclei and cytoplasm. For example,

stimulations of a neuron via its synapses can induce long-lasting

transcriptional event alterations in its nuclei. Such long-range

communication between synapses and nuclei may involve direct

Stru

shuttling of proteins between these two compartments (Ch’ng

et al., 2012; Deisseroth et al., 2003; Ivanova et al., 2015; Panayo-

tis et al., 2015). As with other cell types, neurons frequently shut-

tle proteins between nucleus and cytoplasm in response to

different stimuli. For example, some highly conserved RNA bind-

ing proteins (RBPs), which are specifically expressed in the ner-

vous system, could be shuttled between nucleus and cytoplasm

in the presence of nucleo-cytoplasmic shuttling signals and play

important roles in neurite differentiation (Colombrita et al., 2013;

Kasashima et al., 1999). Similar nuclear/cytoplasm translocation

phenomena are also known in the case of the Suppressor of

fused (Sufu), a negative regulator of the Hedgehog signal trans-

duction pathway, and histone deacetylase 7 (HDAC7), a negative

gene expression regulator and the chromatin-associated high-

mobility group proteins (Li et al., 2004; Paces-Fessy et al.,

2004; Pedersen et al., 2010). Such nuclear/cytoplasm shuttling

of these proteins are often regulated by either post-translational

modification or protein-protein interactions.

Zinc-finger MYND-type-containing 8 (ZMYND8), also known

as PRKCBP1, RACK-7, and Spikar, was first identified as a pro-

tein kinase C binding partner (Fossey et al., 2000) and was later

shown to be a transcriptional regulator (Malovannaya et al.,

2011). It contains an N-terminal nuclear localization sequence

(NLS) and a PHD-BROMO-PWWP (PBP) tandem, and a C-termi-

nal coiled-coil (CC) domain and a myeloid, Nervy, and DEAF-1

(MYND) domain. Its domain organization is similar to that of

ZMYND11 (also known as BS69), which uses the N-terminal

BROMO-PWWP tandem to recognize histone H3.3K36me3

(Guo et al., 2014; Wen et al., 2014). Both ZMYND8 and

ZMYND11 were reported to be tumor suppressors (Ansieau

and Sergeant, 2003).

Studies of ZMYND8 have so far been mainly focused on

its roles in nucleus, where it can act as reader for post-transla-

tionally modified histones, form complex with lysine-specific

demethylase 5C (KDM5C) in regulating active enhancers, co-

localize with nucleosome remodeling and histone deacetylase

(NuRD) complex in DNA damage response, bind to REST core-

pressor 2 (RCOR2) and thus regulate neuronal differentiation,

and interact with lysine-specific demethylase 5D (KDM5D,

also known as JARID1D) to suppress metastasis-linked gene

cture 25, 1657–1666, November 7, 2017 ª 2017 Elsevier Ltd. 1657

Figure 1. Biochemical Characterizations of ZMYND8/Drebrin Interaction

(A) Domain organizations of ZMYND8 and Drebrin.

(B) Analytical gel filtration-based assay showing that ZMYND8 PBP tandem and Drebrin ADF-H domain can interact with each other in vitro.

(C) ITC-based assay showing that Drebrin ADF-H domain can bind to ZMYND8 PBP tandem with a KD of �4.3 mM.

(D) ITC results showing that including the Drebrin CC-Hel domain and ZMYND8 N-terminal NLS did not further enhance the binding.

expression (Adhikary et al., 2016; Gong et al., 2015; Li et al.,

2016; Savitsky et al., 2016; Shen et al., 2016; Spruijt et al.,

2016; Zeng et al., 2010). Recently, ZMYND8 was shown to

interact with a neuronal actin binding protein known as develop-

mentally regulated brain protein (Drebrin) (Yamazaki et al., 2014).

It was shown that the localization of ZMYND8 to the dendritic

spines depends on the binding of Drebrin, suggesting a role of

Drebrin in regulating the distribution of ZMYND8 between nu-

cleus and the synapses.

Drebrin was first discovered as the developmentally regulated

brain protein (Shirao et al., 1988). The expression levels of

different isoforms of Drebrin (Drebrin A and E) were shown to

be developmentally regulated: Drebrin E is prominently ex-

pressed during early neuronal development and Drebrin A is ex-

pressed in more mature neurons (Imamura et al., 1992; Shirao

and Obata, 1985). Drebrin is highly expressed in neuron and re-

ported to regulate actin cytoskeleton in dendritic spines (Lin and

Koleske, 2010; Sekino et al., 2007; Shirao and Gonzalez-Billault,

2013). The expression level of Drebrin is tightly controlled. Over-

expression of Drebrin can cause elongation of the spines without

affecting spine density, while knockout or knockdown of Drebrin

will lead to reduction of the density of dendritic spine and

decrease in the PSD-95 clustering in post-synaptic densities (Ta-

kahashi et al., 2003). Downregulation of the expression level of

Drebrin has also been linked to several human genetic diseases,

such as Alzheimer’s disease, Down syndrome, and mild cogni-

tive impairments (Counts et al., 2012; Harigaya et al., 1996;

Shim and Lubec, 2002). In addition to neurons, Drebrin is also ex-

1658 Structure 25, 1657–1666, November 7, 2017

pressed in other tissues including lung, intestine, and liver. Either

overexpression or underexpression of Drebrin has been impli-

cated in various human diseases including cancer and microvilli

inclusion disease, possibly due to impairments of cell polarity in

tissues caused by altered levels of Drebrin (Bazellieres et al.,

2012; Dart et al., 2017; Dun and Chilton, 2010; Lin et al., 2014;

Peitsch et al., 2005). It is tempting to hypothesize that, in addition

to directly binding to actin filaments, Drebrin may also control

transcriptional activities of ZMYND8 by modulating its nuclear

localization in neurons as well as other cell types, thereby indi-

rectly exerting diverse regulatory roles in various tissues.

Both Drebrin E and Drebrin A contain an N-terminal highly con-

served actin-depolymerizing factor homology (ADF-H) domain

(Figure 1A), which loses both actin-depolymerizing activity and

actin binding ability (Grintsevich et al., 2010; Poukkula et al.,

2011). Instead, they use their central putative CC domain and he-

lical (Hel) domain to interact with and bundle actin filaments

(Grintsevich et al., 2010; Worth et al., 2013). The lacking of actin

binding of the Drebrin ADF-H domain raises a natural question:

what is the function of this evolutionarily conserved ADF-H

domain? The recent discovery that the Drebrin ADF-H domain

can bind to ZMYND8 suggests that it may function as a general

protein binding module instead of an actin binding domain.

In this study, we elucidated the molecular basis governing

the unexpected ZMYND8 PBP/Drebrin ADF-H interaction. The

structure of the ZMYND8 PBP/Drebrin ADF-H complex revealed

that the interaction between ZMYND8 and Drebrin is highly

specific. No binding could be detected between ZMYND11

Figure 2. Overall Structure of the ZMYND8

PBP/Drebrin ADF-H Complex

(A) Ribbon representations of the ZMYND8 PBP/

Drebrin ADF-H complex crystal structure. The

Drebrin ADF-H domain is colored in green. The

PHD, BROMO, Zinc Finger, PWWP, andC-terminal

extension are colored in pink, violet, magenta,

hotpink, and purple, respectively (as shown in

Figure 1A). The zinc ions are shown as gray

spheres. This color coding applies to all figures

except when otherwise indicated.

(B) Comparison of the ZMYND8 PBP structure (in

magenta) in the ZMYND8 PBP/Drebrin ADF-H

complex with the recently reported ZMYND8 PBP

apo-form structures (PDB: 4COS in light blue;

PDB: 5B73 in light yellow). The histone binding

sites for PHD, BROMO, and PWWP domains are

indicated with cyan, marine, and blue ellipses,

respectively.

(C) Superposition of the BROMO-PWWP tandems

of ZMYND8 (magenta) and ZMYND11 (PDB: 4N4H

in yellow) showing that the overlapping binding site

of Drebrin ADF-H (green) to ZMYND8 and

H3.3K36me3 (orange) to ZMYND11.

(D) Structural details of the overlapping binding

of Drebrin and histone on PWWP domain. D1:

ZMYND8/Drebrin. D2: ZMYND11/H3.3K36me3.

D3: superposition of ZMYND8/Drebrin and

ZMYND11/H3.3K36me3. Critical residues are

shown as sticks.

andDrebrin or between ZMYND8 andDrebrin-like. The ZMYND8

PBP/Drebrin ADF-H complex structure further suggests that

Drebrin binding and histone binding are mutually exclusive in

the ZMYND8 PBP tandem. Our structure also explains why the

Drebrin ADF-H domain fails to bind actin filaments. Finally, we

demonstrated using a heterologous cell culture assay that nu-

clear localized ZMYND8 can be effectively shuttled out to the

cytoplasm by co-expression of wild-type Drebrin but not by

ZMYND8 binding-defective mutants designed on the basis of

the complex structure.

RESULTS

ZMYND8 Can Bind to Drebrin ADF-H Domain with ItsPHD-BROMO-PWWP TandemA previous study has shown that the interaction between

ZMYND8 and Drebrin is mediated by the PBP tandem and

ADF-H domain (Yamazaki et al., 2014) (Figure 1A). Here we first

purified the recombinant ZMYND8 PBP tandem (amino acids

Structu

[aa] 103–426) and the Drebrin ADF-H

domain (aa 1–135) to validate and quantify

the interaction. On an analytical gel filtra-

tion column, a 1:1 mixture of ZMYND8

PBP and Drebrin ADF-H was eluted at a

volume clearly smaller than each of the

two individual proteins alone (Figure 1B),

suggesting that the two proteins in the

mixture form a stable complex. An

isothermal titration calorimetry (ITC) assay

was used to quantify the interaction, and

the titration experiment showed that the two proteins bind to

each other with a 1:1 stoichiometry and with a dissociation con-

stant (KD) of�4.3 mM (Figure 1C). Extending ZMYND8 to include

the conserved N-terminal NLS sequence (aa 1–426) and extend-

ing Drebrin ADF-H to include the C-terminal CC-Hel region (aa

1–309) did not alter the binding affinity between the two proteins

(Figures 1C and 1D), indicating that ZMYND8 PBP and Drebrin

ADF-H are the minimal regions responsible for the two proteins

to interact with each other.

Overall Structure of ZMYND8 PHD-BROMO-PWWP/Drebrin ADF-H ComplexTo understand the molecular basis of the interaction, we set out

to determine the crystal structure of the ZMYND8 PBP/Drebrin

ADF-H complex. After numerous trials, we obtained several

well-diffracted (to 2.65-A resolution) crystals by fusing the Dre-

brin ADF-H domain at the C terminus of ZMYND8 PBP with a

22-residue flexible linker (Figure 2A and Table 1). Consistent

with the recently solved crystal structures in its apo form (Li

re 25, 1657–1666, November 7, 2017 1659

Table 1. X-Ray Crystallographic Data Collection and Model

Refinement

Data Collection

Datasets ZMYND8 PBP/Drebrin ADF-H

Space group P212121

Wavelength (A) 0.9780

Unit cell parameters (A) a = 54.18, b = 138.10, c = 161.16

a = b = g = 90�

Resolution range (A) 50–2.65 (2.70–2.65)

No. of unique reflections 34,299 (1,638)

Redundancy 6.5 (6.3)

I/s 25.9 (2.5)

Completeness (%) 98.9 (94.5)

Rmergea (%) 9.2 (95.1)

CC1/2 (last-resolution shell)b 0.758

Structure Refinement

Resolution (A) 15–2.65 (2.77–2.65)

Rcrystc/Rfree

d(%) 17.81/23.37 (27.97/33.62)

RMSD bonds (A)/angles (�) 0.009/0.941

Average B factor (A2)e 58.1

No. of atoms

Protein atoms 6,702

Water 48

Ligands 42

No. of reflections

Working set 32,354 (2,717)

Test set 1,648 (149)

Ramachandran plot regionse

Favored (%) 97.6

Allowed (%) 2.4

Outliers (%) 0

Numbers in parentheses represent the value for the highest-resolu-

tion shell.aRmerge = S jIi � <I>j/SIi, where Ii is the intensity of measured reflection

and <I> is the mean intensity of all symmetry-related reflections.bCC1/2 was defined by Karplus and Diederichs (2012).cRcryst = S jjFcalcj � jFobsjj/SFobs, where Fobs and Fcalc are observed and

calculated structure factors.dRfree = ST jjFcalcj � jFobsjj/SFobs, where T is a test dataset of about 5% of

the total unique reflections randomly chosen and set aside prior to

refinement.eB factors and Ramachandran plot statistics are calculated using Mol-

Probity (Chen et al., 2010).

et al., 2016; Savitsky et al., 2016), the ZMYND8 PBP tandem

consists of an N-terminal PHD domain bound with two Zn2+

ions, a BROMO domain and a PWWP domain separated by a

zinc finger. The four domains of the ZMYND8 PBP tandem

interact with each other sequentially and intimately to form an in-

tegral structural supramodule. The binding of Drebrin ADF-H

domain does not induce obvious conformational changes of

the PBP supramodule (root-mean-square deviation [RMSD]

with the apo structures PDB: 4COS and PDB: 5B73 of 0.55 A

and 0.58 A, respectively; Figure 2B). The Drebrin ADF-H domain

fits snugly into the groove formed by the BROMO, zinc-finger,

and PWWP domains of the ZMYND8 PBP supramodule, and

1660 Structure 25, 1657–1666, November 7, 2017

makes contacts with both the BROMO domain and the PWWP

domain. The structure suggests that formation of the ZMYND8

PBP supramodule is essential for its binding to Drebrin. The

structure of the ZMYND8 PBP supramodule solved here also

showed a high similarity to that of the ZMYND11 BROMO-

PWWP tandem (RMSD of 1.21 A, Figure 2C).

Detailed Structural Analysis Shows that the Interactionbetween ZMYND8 PHD-BROMO-PWWP and DrebrinADF-H Is Highly SpecificADF-H binds to two separate sites of the ZMYND8 PBP tandem,

one on the PWWP domain and the other on the BROMO domain

(Figure 3A). The interface involving the PWWP domain is mainly

mediated by the aA of Drebrin ADF-H. A salt bridge formed by

Arg10D (the superscript ‘‘D’’ represents Drebrin) and Asp332Z

(‘‘Z’’ denotes ZMYND8) is critical for the interaction (Figure 3B),

as charge-reversemutations of either of the two residues individ-

ually totally disrupted the binding (Figure 3D). Hydrophobic inter-

actions by Leu11D, Leu14D, and Tyr17D and Phe308Z, Thr355Z,

and Ile358Z further anchor the aA helix of ADF-H to the

PWWP site (Figure 3B). In addition, a hydrogen bond between

Glu107D and His331Z also contributes to the binding (Figure 3B).

Mutating the His331Z to Ala abolished the binding (Figure 3D).

The binding in the second site is much less extensive and only

involves hydrophobic residues of Cys96D from Drebrin ADF-H

aC and Ile257Z, Val260Z from the ZMYND8 BROMO domain.

Substitutions of Cys96D or Val260Z to polar Gln led to a 10- to

100-fold decrease of the binding (Figure 3D).

Both the Drebrin and Drebrin-like protein (also known as

HIP-55, mAbp1, and SH3P7) are mammalian orthologs of yeast

actin binding protein 1 (Abp1), amember of the ABP family of pro-

teins (Kessels et al., 2000). They share similar domain architec-

tures in their N terminus. Especially their ADF-H domains share

more than 40% sequence identity (Figure S1A), and the crystal

structure reported here can be superimposedwell with the previ-

ously solved nuclear magnetic resonance structure of Drebrin-

like ADF-H domain (RMSD �1.16 A) (Goroncy et al., 2009). We

wondered whether ZMYND8 can also interact with the Drebrin-

like ADF-H domain. We carefully analyzed the sequences and

found that although the ADF-H domain of Drebrin-like shared

high identity with Drebrin and only a few residues involved in

ZMYND8binding are different, Drebrin-like ADF-Hhas nodetect-

able binding to ZMYND8 PBP (Figures 3D, S1B, and S1C). The

Drebrin ADF-H and ZMYND8 PBP complex structure can explain

the lack of binding of Drebrin-like ADF-H to ZMYND8 PBP. The

residue corresponding to Arg10 in Drebrin is a Gly in Drebrin-

like (Figure S2), and Arg10 is absolutely essential for Drebrin

ADF-H to bind to ZMYND8 PBP (Figures 3B and 3D). As ex-

pected, when Arg10 of Drebrin ADF-H was replaced by Gly, the

mutant protein failed to interact with ZMYND8 (Figures 3D and

S1D). In parallel, ZMYND11, a close relative of ZMYND8, cannot

bind toDrebrin either (Figures 3D, S1E, andS1F). Taken together,

these biochemical data further support that the interaction be-

tween ZMYND8 and Drebrin is highly specific.

The Drebrin ADF-H Binding and Histone Peptide Bindingon ZMYND8 PBP Tandem Are Mutually ExclusivePHD, BROMO, and PWWP domains are well known to recog-

nize histone markers (Filippakopoulos and Knapp, 2012; Qin

Figure 3. The Detailed Interactions between ZMYND8 PBP and Drebrin ADF-H

(A) The ZMYND8 PBP/Drebrin ADF-H interface can be divided into two parts as highlighted by solid and dashed boxes.

(B) Stereo view of the detailed interactions in the ADF-H/PWWP interface.

(C) Detailed interactions in the ADF-H/BROMO interface.

(D) Summary of the ITC-derived dissociation constants showing that mutations of the critical residues in either of the interfaces weakened or abolished the

binding. Also, no binding was detected between ZMYND8/Drebrin-like and ZMYND11/Drebrin. WT, wild-type.

and Min, 2014; Sanchez and Zhou, 2011). In the apo form of

the PBP supramodule, all the known histone binding sites

are accessible to their respective binders (Figure 2B). Upon

forming a complex with Drebrin ADF-H, the histone binding

site on the PWWP domain is occupied by ADF-H, indicating

that the Drebrin and histone binding on ZMYND8 PBP supra-

module may be mutually exclusive. Consistent with this anal-

ysis, the histone H3.3K36me3 binding site on the ZMYND11

PWWP domain (thus inferring also the histone peptide binding

site of ZMYND8 PWWP) overlaps with the Drebrin binding site

(Figures 2C and 2D) (Wen et al., 2014). In particular, the gua-

nidinium group of Arg10 in Drebrin and the trimethylated

amide group of K36 in H3.3 are at a near identical position

in the two structures (Figure 2D), suggesting that Drebrin

ADF-H can block histone peptide from binding to ZMYND8

PWWP. This analysis is supported by a recent finding showing

that mutations in the PWWP domain weaken the ZMYND8/his-

tone peptide binding (Savitsky et al., 2016). Considering that

the binding affinity of ZMYND8 PBP to Drebrin ADF-H is com-

parable with those of ZMYND8 PBP to histone peptides (with

KD in the micromolar range) (Li et al., 2016; Savitsky et al.,

2016), it is biochemically possible that Drebrin can compete

with histone peptides for binding to the ZMYND8 PBP

supramodule.

Comparison of ADF-H Domain Complex StructuresSuggests a Non-canonical Function of Drebrin ADF-HDomainThe ZMYND8 PBP/Drebrin ADF-H complex structure solved in

this work reveals a previously unknown function of the ADF-H

domain. Instead of binding to actin, Drebrin ADF-H binds to his-

tonemarker recognition domains. The aC helix of Drebrin ADF-H

interacts with the ZMYND8 BROMO domain (Figure 4A). Previ-

ous biochemical and structural studies have focused on how

other ADF-H domains bind to actin (Lappalainen et al., 1998;

Poukkula et al., 2011). The structures of twinfilin C-terminal

ADF-H or cofilin ADF-H domain in complex with G-actin or

F-actin showed that the long aC, situated at the cleft between

actin subdomains 1 and 3, constitutes the major actin binding

element of ADF-H (Figure 4B) (Galkin et al., 2011; Paavilainen

et al., 2008). A similar binding mode has also been observed in

the structure of GMF ADF-H bound to the actin-related protein

2 (ACTR2) of Arp2/3 complex (Luan and Nolen, 2013). Interest-

ingly, the ADF-H domain of cofilin and twinfilin is structurally

homologous to Gelsolin segment 1, and they both utilize their

respective aC to bind to the same groove between subdomains

1 and 3 of actin, although the detailed interacting residues are

slightly different (Mclaughlin et al., 1993; Paavilainen et al.,

2008). On another note, the crystal structure of cofilin ADF-H

Structure 25, 1657–1666, November 7, 2017 1661

Figure 4. aC in ADF-H Domains Is Commonly Used to Recognize Targets

(A–C) Crystal structures of Drebrin ADF-H/ZMYND8 (A, magenta), twinfilin ADF-H/Actin (B, orange), and cofilin ADF-H/LIMK kinase (C, salmon) complexes

showing that the aC (highlighted in green) is the key element for these ADF-H domains (shown in gray) to interact with their respective targets. Residues critical in

binding are shown as blue sticks.

(D) Sequence alignments of the commonly used aC from all human ADF-H domains. The residues critical for actin binding are highlighted in red boxes and the

residue critical for ZMYND8 binding in Drebrin is highlighted in a magenta box.

domain in complex with its regulator LIMK1 kinase domain has

recently been reported (Hamill et al., 2016). LIMK1 is known to

phosphorylate cofilin and inhibit cofilin’s actin binding and depo-

lymerizing capacity. Interestingly, the aC of cofilin ADF-H is also

chiefly responsible for binding to the LIMK1 kinase domain

(Figure 4C).

Sequence alignments of the aC from all identified human

ADF-H domains showed that several critical residues (two posi-

tively charged residues Arg/Lys at position 2, Arg/Lys at position

11, and the two hydrophobic residues Val/Leu/Ile at position 1

and Met at position 5) responsible for actin binding are con-

served in cofilin or twinfilin but not in Drebrin or Drebrin-like (Fig-

ure 4D). This is consistent with the extremely low actin binding

affinity of Drebrin and Drebrin-like ADF-H domains (Grintsevich

et al., 2010; Kessels et al., 2000; Poukkula et al., 2011). The

Drebrin ADF-H domain instead has gained a new function in

binding to the histone marker reader ZMYND8. The Cys at the

1662 Structure 25, 1657–1666, November 7, 2017

fifth position of aC of Drebrin is used for hydrophobic interaction

with the ZMYND8 BROMO domain (Figure 4D). The correspond-

ing residue in cofilin is Met5, which is responsible for both actin

and LIMK1 binding. The corresponding residue at this position in

Drebrin-like is aGly (Figure 4D), andwe have shown that Drebrin-

like was not able to bind to ZMYND8. The structural and amino

acid sequence analyses suggest that the diverse amino acid se-

quences of the structurally invariant aC helix of ADF-H domains

may allow these ADF-H domains to bind to proteins with very

different functions.

Drebrin Binding-Induced ZMYND8 CytoplasmLocalizationThe Drebrin-dependent synaptic localization of ZMYND8 in

neurons (Yamazaki et al., 2014) suggested that Drebrin can

sequester ZMYND8 at synapses and prevent it from entering

the nucleus to perform its nuclear function as a transcriptional

Figure 5. ZMYND8/Drebrin Interaction Is

Essential for the Cytoplasmic Localization

of ZMYND8

(A) Representative fluorescent images of HEK293T

cells expressing Myc-ZMYND8 alone or co-ex-

pressing Myc-ZMYND8. A1: Myc-ZMYND8 wild-

type (WT) alone; A2: Myc-ZMYND8 WT and HA-

Drebrin WT; A3: Myc-ZMYND8 and HA-Drebrin

R10G; A4: Myc-ZMYND8 H331A and HA-Drebrin

WT. The nuclei were shown by DAPI (blue) staining.

(B) Quantification of percentages of cells with

ZMYND8 in nuclei in different groups of experi-

ments; 300–800 cells were quantified in each

group of experiments. Values are expressed in

means ± SD and analyzed using one-way ANOVA

with Tukey’s multiple comparison test by Graph-

Pad Prism 6. ns, not significant; ****p < 0.0001.

(C) Schematic cartoon showing the proposed

function of ZMYND8/Drebrin complex. C1: the

synaptic localization of ZMYND8 by binding to actin

cytoskeletal associating protein Drebrin. C2: the

epigenetic function of ZMYND8 by binding to his-

tone markers.

regulator. We used HEK293T heterologous cells to test this hy-

pothesis. When overexpressed alone, ZMYND8 was predomi-

nantly localized in the nucleus as shown by its co-localization

with DAPI (Figures 5A1 and 5B). Co-expression of hemagglutinin

(HA)-Drebrin with Myc-ZMYND8 led to effective redistribution of

ZMYND8 from nucleus to cytoplasm (Figure 5A2). Importantly,

specific mutations either on Drebrin (R10G) or ZMYND8

(H331A), both of which disrupt the formation of the Drebrin/

ZMYND8 complex, were defective in Drebrin-induced cyto-

plasm localization of ZMYND8, indicating that the redistribution

of ZMYND8 from nucleus to cytoplasm is Drebrin binding depen-

dent (Figures 5A3, 5A4, and 5B).

DISCUSSION

In the current study, by detailed biochemical and structural char-

acterizations, we demonstrate that ZMYND8 PBP tandem can

specifically interact with the Drebrin ADF-H domain. We also

show that ZMYND8 can relocalize from nucleus to cytoplasm

upon binding to Drebrin in living cells. Drebrin is an important

actin-regulating protein in synapses. Although ZMYND8 is well

known for its epigenetic function in the nucleus, the discovery

Structu

of the ZMYND8/Drebrin interaction sug-

gests that the nuclear level of ZMYND8

may be regulated by Drebrin-mediated

binding in neuronal synapses (Figure 5C).

Considering that the expression levels of

Drebrin vary at different brain develop-

mental stages (Shirao and Obata, 1985),

the nuclear distribution of ZMYND8 might

also be developmentally regulated via

binding to Drebrin. We wish to emphasize

that the current biochemical, structural,

and heterologous living cell studies,

together with an earlier study showing

Drebrin-dependent synaptic localization

of ZMYND8 in cultured neurons (Yamazaki et al., 2014), suggest

a possible Drebrin-mediated nuclear/cytoplasm shuttling mech-

anism of ZMYND8 in neurons. When Drebrin expression level is

high, this cytoskeleton-associated actin binding protein can act

as a cytoplasmic anchor to sequester ZMYND8 in the cytosol by

the interaction described in this work.Without such sequestering

(due to either lowered expression level of Drebrin or regulated

dissociation of the Drebrin/ZMYND8 complex), ZMYND8 can

enter the nucleus via its NLS-mediated transportation to perform

its epigenetic function. Further studies are required to verify

whether such a shuttling mechanism is indeed adopted in neu-

rons. Nevertheless, the structural information obtained in this

study will be valuable for guiding future functional studies of

the ZMYND8/Drebrin interaction in nervous systems.

The ZMYND8 PBP/Drebrin ADF-H complex presented here

has also revealed a previously unrecognized role of the ADF-H

domain. ADF-H domains were originally identified as actin-

severing modules to promote filamentous actin depolymeriza-

tion (Lappalainen et al., 1998). ADF-H domain-containing pro-

teins can be grouped into five classes. However, members of

only two classes (cofilin and twinfilin) have been unequivocally

shown to contain actin binding and severing activities (Poukkula

re 25, 1657–1666, November 7, 2017 1663

et al., 2011). The GMF class does not bind to actin but can asso-

ciate with an actin-related protein (ACTR2) in the Arp2/3 complex

(Luan and Nolen, 2013). The fact that Coactosin or the Drebrin

classes can only very weakly interact with F-actin suggests

that the ADF-H domain may have functions other than binding

to and severing actin filaments. In the Drebrin ADF-H case, res-

idues critical for actin binding in the aChelix aremissing. Instead,

this commonly used actin binding region becomes the ZMYND8

BROMO domain binding site. Together with a second site that

binds to the PWWP domain of ZMYND8, the interaction between

Drebrin ADF-H and ZMYND8 PBP is highly specific and with a

quite high affinity (Figure 3).

PHD, BROMO, and PWWP, the three well-characterized do-

mains involved in epigenetics regulation, are well known as

readers of epigenetically modified histones. In this study, we

elucidated the molecular mechanism of the interaction between

the ZMYND8 PBP tandem and the synaptic cytoskeletal protein

Drebrin. To the best of our knowledge, this is the first structural

study elucidating how histone marker readers might also bind

to non-histone targets. This finding expands our knowledge of

how histone marker readers may be regulated by binding

proteins outside the cell nucleus. Recently, several BROMO

domain-containing proteins have been reported to have both nu-

cleus and cytoplasm localizations (Filippakopoulos and Knapp,

2012). Similar cytoplasm localization of histone recognition do-

mains containing proteins, e.g., bromodomain-containing pro-

tein 7 (BRD7), bromodomain and PHD finger-containing protein

1 (BRPF1), and PHD finger protein 20-like protein 1 (PHF20L1),

are also found in the Human Protein Atlas database (http://

www.proteinatlas.org/). Although it is not clear how the nu-

clear/cytoplasm localizations of these proteins are determined,

our study of the Drebrin/ZMYND8 interaction provides a clue

to how histone maker readers may also play roles in cellular

localizations via binding to proteins outside the nucleus.

STAR+METHODS

Detailed methods are provided in the online version of this paper

and include the following:

d KEY RESOURCES TABLE

d CONTACT FOR REAGENT AND RESOURCE SHARING

d EXPERIMENTAL MODEL AND SUBJECT DETAILS

166

B HEK293T Cells Culture

d METHOD DETAILS

B Cloning and Protein Purification

B The Size Exclusion Chromatography Coupled with

Multiple Angle Light Scattering Assay

B The Isothermal Titration Calorimetry Assay

B Crystallography

B Cell Imaging and Data Analysis

d QUANTIFICATION AND STATISTICAL ANALYSIS

d DATA AND SOFTWARE AVAILABILITY

B Data Resources

SUPPLEMENTAL INFORMATION

Supplemental Information includes two figures and can be found with this

article online at http://dx.doi.org/10.1016/j.str.2017.08.014.

4 Structure 25, 1657–1666, November 7, 2017

AUTHOR CONTRIBUTIONS

J.L., W.L., J.W., and M.Z. designed the experiments. N.Y. and H.L. performed

biochemistry and cell biology experiments; J.L. is responsible for the structural

biology experiments. All authors analyzed the results. J.L., W.L., and M.Z.

wrote the manuscript. M.Z. coordinated the research.

ACKNOWLEDGMENTS

We thank the Shanghai Synchrotron Radiation Facility (SSRF) BL17U1

and BL19U1 for X-ray beam times. This work was supported by grants

from the National Basic Research Program of China (2014CB910204)

and National Key R&D Program (2016YFA0501903) from the Minister of

Science and Technology of China, Natural Science Foundation of

Guangdong Province (2016A030312016), Shenzhen Basic Research grant

(JCYJ20160229153100269), and RGC of Hong Kong (16103614; 16149516;

and AoE-M09-12) to M.Z., and grants from the National Natural Science Foun-

dation of China (No. 31400647 and No. 31670765) and Shenzhen Basic

Research grants (JCYJ20160427185712266 and JCYJ20170411090807530)

to W.L. M.Z. is a Kerry Holdings Professor in Science and a Senior Fellow of

IAS at HKUST.

Received: April 11, 2017

Revised: July 27, 2017

Accepted: August 28, 2017

Published: September 28, 2017

REFERENCES

Adams, P.D., Afonine, P.V., Bunkoczi, G., Chen, V.B., Davis, I.W., Echols, N.,

Headd, J.J., Hung, L.W., Kapral, G.J., Grosse-Kunstleve, R.W., et al. (2010).

PHENIX: a comprehensive Python-based system for macromolecular struc-

ture solution. Acta Crystallogr. D Biol. Crystallogr. 66, 213–221.

Adhikary, S., Sanyal, S., Basu, M., Sengupta, I., Sen, S., Srivastava, D.K., Roy,

S., and Das, C. (2016). Selective recognition of H3.1K36 dimethylation/H4K16

acetylation facilitates the regulation of all-trans-retinoic acid (ATRA)-respon-

sive genes by putative chromatin reader ZMYND8. J. Biol. Chem. 291,

2664–2681.

Ansieau, S., and Sergeant, A. (2003). BS69 and RACK7, a potential novel class

of tumor suppressor genes. Pathol. Biol. (Paris) 51, 397–399.

Bazellieres, E., Massey-Harroche, D., Barthelemy-Requin, M., Richard, F.,

Arsanto, J.P., and Le Bivic, A. (2012). Apico-basal elongation requires a dre-

brin-E-EB3 complex in columnar human epithelial cells. J. Cell Sci. 125,

919–931.

Ch’ng, T.H., Uzgil, B., Lin, P., Avliyakulov, N.K., O’Dell, T.J., and Martin, K.C.

(2012). Activity-dependent transport of the transcriptional coactivator

CRTC1 from synapse to nucleus. Cell 150, 207–221.

Chen, V.B., Arendall, W.B., 3rd, Headd, J.J., Keedy, D.A., Immormino, R.M.,

Kapral, G.J., Murray, L.W., Richardson, J.S., and Richardson, D.C. (2010).

MolProbity: all-atom structure validation for macromolecular crystallography.

Acta Crystallogr. D Biol. Crystallogr. 66, 12–21.

Colombrita, C., Silani, V., and Ratti, A. (2013). ELAV proteins along evolution:

back to the nucleus? Mol. Cell. Neurosci. 56, 447–455.

Counts,S.E.,He,B.,Nadeem,M.,Wuu, J., Scheff, S.W., andMufson,E.J. (2012).

Hippocampal drebrin loss in mild cognitive impairment. Neurodegener. Dis. 10,

216–219.

Dart, A.E., Worth, D.C., Muir, G., Chandra, A., Morris, J.D., McKee, C., Verrill,

C., Bryant, R.J., and Gordon-Weeks, P.R. (2017). The drebrin/EB3 pathway

drives invasive activity in prostate cancer. Oncogene 36, 4111–4123.

Deisseroth, K., Mermelstein, P.G., Xia, H.H., and Tsien, R.W. (2003). Signaling

from synapse to nucleus: the logic behind the mechanisms. Curr. Opin.

Neurobiol. 13, 354–365.

Dun, X.P., and Chilton, J.K. (2010). Control of cell shape and plasticity during

development and disease by the actin-binding protein Drebrin. Histol.

Histopathol. 25, 533–540.

Emsley, P., Lohkamp, B., Scott, W.G., and Cowtan, K. (2010). Features and

development of Coot. Acta Crystallogr. D Biol. Crystallogr. 66, 486–501.

Filippakopoulos, P., and Knapp, S. (2012). The bromodomain interaction

module. FEBS Lett. 586, 2692–2704.

Fossey, S.C., Kuroda, S., Price, J.A., Pendleton, J.K., Freedman, B.I., and

Bowden, D.W. (2000). Identification and characterization of PRKCBP1, a

candidate RACK-like protein. Mamm. Genome 11, 919–925.

Galkin, V.E., Orlova, A., Kudryashov, D.S., Solodukhin, A., Reisler, E.,

Schroder, G.F., and Egelman, E.H. (2011). Remodeling of actin filaments by

ADF/cofilin proteins. Proc. Natl. Acad. Sci. USA 108, 20568–20572.

Gama-Carvalho, M., and Carmo-Fonseca, M. (2001). The rules and roles of

nucleocytoplasmic shuttling proteins. FEBS Lett. 498, 157–163.

Gong, F., Chiu, L.Y., Cox, B., Aymard, F., Clouaire, T., Leung, J.W.,

Cammarata, M., Perez, M., Agarwal, P., Brodbelt, J.S., et al. (2015). Screen

identifies bromodomain protein ZMYND8 in chromatin recognition of tran-

scription-associated DNA damage that promotes homologous recombination.

Genes Dev. 29, 197–211.

Goroncy, A.K., Koshiba, S., Tochio, N., Tomizawa, T., Sato, M., Inoue, M.,

Watanabe, S., Hayashizaki, Y., Tanaka, A., Kigawa, T., et al. (2009). NMR so-

lution structures of actin depolymerizing factor homology domains. Protein

Sci. 18, 2384–2392.

Grintsevich, E.E., Galkin, V.E., Orlova, A., Ytterberg, A.J., Mikati, M.M.,

Kudryashov, D.S., Loo, J.A., Egelman, E.H., and Reisler, E. (2010). Mapping

of Drebrin binding site on F-Actin. J. Mol. Biol. 398, 542–554.

Guo, R., Zheng, L., Park, J.W., Lv, R., Chen, H., Jiao, F., Xu, W., Mu, S., Wen,

H., Qiu, J., et al. (2014). BS69/ZMYND11 reads and connects histone H3.3

lysine 36 trimethylation-decorated chromatin to regulated pre-mRNA process-

ing. Mol. Cell 56, 298–310.

Hamill, S., Lou, H.J., Turk, B.E., and Boggon, T.J. (2016). Structural basis for

noncanonical substrate recognition of cofilin/ADF proteins by LIM kinases.

Mol. Cell 62, 397–408.

Harigaya, Y., Shoji, M., Shirao, T., and Hirai, S. (1996). Disappearance of actin-

binding protein, drebrin, from hippocampal synapses in Alzheimer’s disease.

J. Neurosci. Res. 43, 87–92.

Imamura, K., Shirao, T., Mori, K., and Obata, K. (1992). Changes of drebrin

expression in the visual cortex of the cat during development. Neurosci.

Res. 13, 33–41.

Ivanova, D., Dirks, A., Montenegro-Venegas, C., Schone, C., Altrock, W.D.,

Marini, C., Frischknecht, R., Schanze, D., Zenker, M., Gundelfinger, E.D.,

et al. (2015). Synaptic activity controls localization and function of CtBP1 via

binding to Bassoon and Piccolo. EMBO J. 34, 1056–1077.

Karplus, P.A., and Diederichs, K. (2012). Linking crystallographic model and

data quality. Science 336, 1030–1033.

Kasashima, K., Terashima, K., Yamamoto, K., Sakashita, E., and Sakamoto, H.

(1999). Cytoplasmic localization is required for the mammalian ELAV-like pro-

tein HuD to induce neuronal differentiation. Genes Cells 4, 667–683.

Kessels, M.M., Engqvist-Goldstein, A.E.Y., andDrubin, D.G. (2000). Association

ofmouse actin-binding protein 1 (mAbp1/SH3P7), an Src kinase target, with dy-

namic regions of the cortical actin cytoskeleton in response to Rac1 activation.

Mol. Biol. Cell 11, 393–412.

Lappalainen, P., Kessels, M.M., Cope, M.J., and Drubin, D.G. (1998). The ADF

homology (ADF-H) domain: a highly exploited actin-binding module. Mol. Biol.

Cell 9, 1951–1959.

Li, X., Song, S., Liu, Y., Ko, S.H., and Kao, H.Y. (2004). Phosphorylation of the

histone deacetylase 7 modulates its stability and association with 14-3-3 pro-

teins. J. Biol. Chem. 279, 34201–34208.

Li, N., Li, Y., Lv, J., Zheng, X., Wen, H., Shen, H., Zhu, G., Chen, T.Y., Dhar,

S.S., Kan, P.Y., et al. (2016). ZMYND8 reads the dual histone mark

H3K4me1-H3K14ac to antagonize the expression of metastasis-linked genes.

Mol. Cell 63, 470–484.

Lin, Q.F., Tan, H.T., Lim, T.K., Khoo, A., Lim, K.H., and Chung, M.C.M. (2014).

iTRAQ analysis of colorectal cancer cell lines suggests Drebrin (DBN1) is over-

expressed during liver metastasis. Proteomics 14, 1434–1443.

Lin, Y.C., and Koleske, A.J. (2010). Mechanisms of synapse and dendrite

maintenance and their disruption in psychiatric and neurodegenerative disor-

ders. Annu. Rev. Neurosci. 33, 349–378.

Luan, Q., and Nolen, B.J. (2013). Structural basis for regulation of Arp2/3 com-

plex by GMF. Nat. Struct. Mol. Biol. 20, 1062–1068.

Malovannaya, A., Lanz, R.B., Jung, S.Y., Bulynko, Y., Le, N.T., Chan, D.W.,

Ding, C., Shi, Y., Yucer, N., Krenciute, G., et al. (2011). Analysis of the human

endogenous coregulator complexome. Cell 145, 787–799.

Mccoy, A.J., Grosse-Kunstleve, R.W., Adams, P.D., Winn, M.D., Storoni, L.C.,

and Read, R.J. (2007). Phaser crystallographic software. J. Appl. Crystallogr.

40, 658–674.

Mclaughlin, P.J., Gooch, J.T., Mannherz, H.G., and Weeds, A.G. (1993).

Structure of gelsolin segment-1-actin complex and the mechanism of filament

severing. Nature 364, 685–692.

Otwinowski, Z., and Minor, W. (1997). Processing of X-ray diffraction data

collected in oscillation mode. Methods Enzymol. 276, 307–326.

Paavilainen, V.O., Oksanen, E., Goldman, A., and Lappalainen, P. (2008).

Structure of the actin-depolymerizing factor homology domain in complex

with actin. J. Cell Biol. 182, 51–59.

Paces-Fessy,M.,Boucher,D.,Petit, E., Paute-Briand,S., andBlanchet-Tournier,

M.F. (2004). The negative regulator of Gli, Suppressor of fused (Sufu), interacts

with SAP18, Galectin3 and other nuclear proteins. Biochem. J. 378, 353–362.

Panayotis, N., Karpova, A., Kreutz, M.R., and Fainzilber, M. (2015).

Macromolecular transport in synapse to nucleus communication. Trends

Neurosci. 38, 108–116.

Pedersen, D.S., Merkle, T., Marktl, B., Lildballe, D.L., Antosch, M., Bergmann,

T., Tonsing, K., Anselmetti, D., and Grasser, K.D. (2010). Nucleocytoplasmic

distribution of the Arabidopsis chromatin-associated HMGB2/3 and HMGB4

proteins. Plant Physiol. 154, 1831–1841.

Peitsch, W.K., Hofmann, I., Bulkescher, J., Hergt, M., Spring, H., Bleyl, U.,

Goerdt, S., and Franke, W.W. (2005). Drebrin, an actin-binding, cell-type char-

acteristic protein: induction and localization in epithelial skin tumors and

cultured keratinocytes. J. Invest. Dermatol. 125, 761–774.

Poukkula, M., Kremneva, E., Serlachius, M., and Lappalainen, P. (2011). Actin-

depolymerizing factor homology domain: a conserved fold performing diverse

roles in cytoskeletal dynamics. Cytoskeleton (Hoboken) 68, 471–490.

Qin, S., and Min, J. (2014). Structure and function of the nucleosome-binding

PWWP domain. Trends Biochem. Sci. 39, 536–547.

Sanchez, R., and Zhou, M.M. (2011). The PHD finger: a versatile epigenome

reader. Trends Biochem. Sci. 36, 364–372.

Savitsky, P., Krojer, T., Fujisawa, T., Lambert, J.P., Picaud, S., Wang, C.Y.,

Shanle, E.K., Krajewski, K., Friedrichsen, H., Kanapin, A., et al. (2016).

Multivalent histone and DNA engagement by a PHD/BRD/PWWP triple reader

cassette recruits ZMYND8 to K14ac-rich chromatin. Cell Rep. 17, 2724–2737.

Sekino, Y., Kojima, N., and Shirao, T. (2007). Role of actin cytoskeleton in den-

dritic spine morphogenesis. Neurochem. Int. 51, 92–104.

Shen, H., Xu, W., Guo, R., Rong, B., Gu, L., Wang, Z., He, C., Zheng, L., Hu, X.,

Hu, Z., et al. (2016). Suppression of enhancer overactivation by a RACK7-his-

tone demethylase complex. Cell 165, 331–342.

Shim, K.S., and Lubec, G. (2002). Drebrin, a dendritic spine protein, is manifold

decreased in brains of patients with Alzheimer’s disease and Down syndrome.

Neurosci. Lett. 324, 209–212.

Shirao, T., andGonzalez-Billault, C. (2013). Actin filaments andmicrotubules in

dendritic spines. J. Neurochem. 126, 155–164.

Shirao, T., Kojima,N., Kato, Y., andObata, K. (1988).Molecular cloning of a cDNA

for the developmentally regulated brain protein, drebrin. Brain Res. 464, 71–74.

Shirao, T., and Obata, K. (1985). Two acidic proteins associated with brain

development in chick embryo. J. Neurochem. 44, 1210–1216.

Spruijt, C.G., Luijsterburg, M.S., Menafra, R., Lindeboom, R.G., Jansen, P.W.,

Edupuganti, R.R., Baltissen, M.P., Wiegant, W.W., Voelker-Albert, M.C.,

Matarese, F., et al. (2016). 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. 17, 783–798.

Structure 25, 1657–1666, November 7, 2017 1665

Takahashi, H., Sekino, Y., Tanaka, S., Mizui, T., Kishi, S., and Shirao, T. (2003).

Drebrin-dependent actin clustering in dendritic filopodia governs synaptic

targeting of postsynaptic density-95 and dendritic spine morphogenesis.

J. Neurosci. 23, 6586–6595.

Wen, H., Li, Y., Xi, Y., Jiang, S., Stratton, S., Peng, D., Tanaka, K., Ren, Y., Xia,

Z., Wu, J., et al. (2014). ZMYND11 links histone H3.3K36me3 to transcription

elongation and tumour suppression. Nature 508, 263–268.

Worth, D.C., Daly, C.N., Geraldo, S., Oozeer, F., and Gordon-Weeks, P.R.

(2013). Drebrin contains a cryptic F-actin-bundling activity regulated by

Cdk5 phosphorylation. J. Cell Biol. 202, 793–806.

1666 Structure 25, 1657–1666, November 7, 2017

Xu, L., and Massague, A. (2004). Nucleocytoplasmic shuttling of signal trans-

ducers. Nat. Rev. Mol. Cell Biol. 5, 209–219.

Yamazaki, H., Kojima, N., Kato, K., Hirose, E., Iwasaki, T., Mizui, T., Takahashi,

H., Hanamura, K., Roppongi, R.T., Koibuchi, N., et al. (2014). Spikar, a novel

drebrin-binding protein, regulates the formation and stabilization of dendritic

spines. J. Neurochem. 128, 507–522.

Zeng, W.L., Kong, Q.H., Li, C.C., and Mao, B.Y. (2010). Xenopus RCOR2

(REST corepressor 2) interacts with ZMYND8, which is involved in neural dif-

ferentiation. Biochem. Biophys. Res. Commun. 394, 1024–1029.

STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER

Antibodies

Mouse monoclonal anti-HA Santa Cruz Biotechnology Cat#sc7392; RRID: AB_627809

Rabbit polyclonal anti-Myc Proteintech Cat#16286-1-AP; RRID: AB_11182162

Chemicals, Peptides, and Recombinant Proteins

Dulbecco’s Modified Eagle Medium (DMEM) Thermo Fisher Scientific Cat#11995065

fetal bovine serum (FBS) Thermo Fisher Scientific Cat#16000044

penicillin-streptomycin Thermo Fisher Scientific Cat#15140122

Recombinant protein: Human Drebrin ADF-H (aa:

1-135, ref# NP_004386.2)

This paper N/A

Recombinant protein: Human Drebrin ADF-H-CC-

Hel (aa: 1-309, ref# NP_004386.2)

This paper N/A

Recombinant protein: Human ZMYND8 NLS-PBP

(aa: 1-426, ref# NP_898868.1)

This paper N/A

Recombinant protein: Human ZMYND8 PBP (aa:

103-426, ref# NP_898868.1)

This paper N/A

Recombinant protein: Mouse Drebrin-like ADF-H

(aa: 1-138, ref# NP_001139780.1)

This paper N/A

Recombinant protein: Mouse ZMYND11 Bromo-

PWWP (aa: 155-367, ref# NP_001334401.1)

This paper N/A

Recombinant protein: ZMYND8 PBP-linker-

Drebrin ADF-H (ZMYND8 Q103-S426-

GSGSGENLYFQGGSGGSGGSGS-Drebrin

M1-N135)

This paper N/A

Critical Commercial Assays

ViaFect� Transfection Reagent Promega Cat#E4981

Deposited Data

Crystal structure of ZMYND8 PBP/Drebrin ADF-H This paper PDB: 5Y1Z

Experimental Models: Cell Lines

Human: HEK293T cells ATCC CRL-3216

Experimental Models: Organisms/Strains

Escherichia coli: Rosetta (DE3) Novagen Cat#70954

Escherichia coli: BL21 (DE3) Invitrogen Cat#C600003

Recombinant DNA

Plasmid: HA-Drebrin This paper N/A

Plasmid: Myc-ZMYND8 This paper N/A

Plasmid: S142A Drebrin-YFP Addgene Cat#58335

Plasmid: GFP-ZMYND8 Addgene Cat#65401

Plasmid: pET32M.3C Human Drebrin ADF-H This paper N/A

Plasmid: pET32M.3C Human Drebrin ADF-H-

CC-Hel

This paper N/A

Plasmid: pET32M.3C Human ZMYND8 NLS-PBP This paper N/A

Plasmid: pET32M.3C Human ZMYND8 PBP This paper N/A

Plasmid: pET32M.3C Mouse Drebrin-like ADF-H This paper N/A

Plasmid: pET32M.3C Mouse ZMYND11

Bromo-PWWP

This paper N/A

Plasmid: pET32M.3C ZMYND8 PBP-linker-Drebrin

ADF-H

This paper N/A

(Continued on next page)

Structure 25, 1657–1666.e1–e3, November 7, 2017 e1

Continued

REAGENT or RESOURCE SOURCE IDENTIFIER

Software and Algorithms

Origin7.0 OriginLab http://www.originlab.com/

GraphPad Prism GraphPad Software Inc. http://www.graphpad.com/scientific-software/

prism/

HKL2000 (Otwinowski and Minor, 1997) http://www.hkl-xray.com/

PHASER (Mccoy et al., 2007) http://www.phaser.cimr.cam.ac.uk/

PHENIX (Adams et al., 2010) http://www.phenix-online.org/

Coot (Emsley et al., 2010) http://www2.mrc-lmb.cam.ac.uk/Personal/

pemsley/coot/

PyMOL DeLano Scientific LLC http://www.pymol.org/

ASTRA6 Wyatt http://www.wyatt.com/products/software/

astra.html

ImageJ NIH https://imagej.nih.gov/ij/

CONTACT FOR REAGENT AND RESOURCE SHARING

Further information and requests for reagents may be directed to, and will be fulfilled by the Lead Contact Mingjie Zhang (mzhang@

ust.hk).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

HEK293T Cells CultureHEK293T cells (fromATCC) were cultured in Dulbecco’sModified EagleMedium (DMEM) supplementedwith 10% fetal bovine serum

(FBS), and 1% of penicillin-streptomycin at 37�C with 5% CO2.

METHOD DETAILS

Cloning and Protein PurificationThe coding sequences of Drebrin fragments (residues 1-135 & 1-309) were PCR amplified using the full-length human 70 kDa Drebrin

cDNA template (NCBI accession code: NM_004395.3) . The coding sequences of ZMYND8 (residues 103-426&1-426) fragments

were PCR amplified using the full-length human 130 kDa ZMYND8 cDNA template (NCBI accession code: NM_183047.3). The

coding sequences of Drebrin-like (residues 1-138, NCBI accession code: NM_001146308.1) and ZMYND11 (residues 155-367,

NCBI accession code: NM_001347472.1) were PCR amplified from a mouse cDNA library. For the fusion construct used for crystal-

lization and the complex structure determination, Drebrin 1-135 was fused to ZMYND8 103-426 C-terminus by standard two-step

PCR with a 22-residue flexible linker ‘GSGSGENLYFQGGSGGSGGSGS’ consisted of a TEV cleavage site ‘ENLYFQ’ flanking by

Glycine and Serine residues. All point mutations were created by the standard PCR-based mutagenesis method and confirmed

by DNA sequencing. All of these coding sequences were cloned into a modified pET32a vector for protein expression. All Drebrin

or Drebrin-like proteins were expressed in Escherichia coli BL21 (DE3), while all ZMYND8 and ZMYND11 proteins were expressed

in Escherichia coli Rosetta (DE3) with additional ZnCl2 (final concentration in medium: 200 mM). The N-terminal thioredoxin-His6-

tagged proteins were purified with a Ni2+ Sepharose� 6 Fast Flow column and followed by a Superdex-200 prep grade size-exclu-

sion chromatography. The thioredoxin tag was cut by 3C protease and the proteins were further purified by another Superdex-200

prep grade size-exclusion chromatography.

The Size Exclusion Chromatography Coupled with Multiple Angle Light Scattering AssayProtein samples (typically 100 mL at a concentration of 50 mM, pre-equilibrated with column buffer) were injected into an AKTA FPLC

system with a Superose 12 10/300 GL column (GE Healthcare) using the column buffer of 50 mM Tris-HCl (pH 7.5), 1 mM DTT, and

100 mM NaCl. The chromatography system was coupled to a multi-angle light scattering system equipped with a 18 angles static

light scattering detector (Dawn, Wyatt) and a differential refractive index detector (Optilab, Wyatt). The elution profiles were analyzed

using the ASTRA 6 software (Wyatt).

The Isothermal Titration Calorimetry AssayIsothermal titration calorimetry (ITC) measurements were carried out on the MicroCal ITC200 calorimeter (Malvern) at 16�C. All pro-teins were dissolved in 50mMTris buffer containing 100mMNaCl, and 1mMDTT at pH 7.5. The protein concentrations in the syringe

e2 Structure 25, 1657–1666.e1–e3, November 7, 2017

vary from 500 mM-600 mM,while the concentrations of the binding partners in the cell vary from 50 mM-60 mM. Each titration point was

performed by injecting a 2 mL aliquot of the protein sample from the syringe into the protein sample in the cell at a time interval of 120

seconds to ensure that the titration peak returned to the baseline. The titration data were analyzed by Origin7.0 (Microcal) and fitted

by the one-site binding model.

CrystallographyCrystals of the Drebrin and ZMYND8 complex (in 50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM DTT buffer) were obtained by

the sitting drop vapor diffusion method at 16�C. The crystals were grown in buffer containing 0.2 M ammonium sulfate, 0.1 M Bis-

Tris, pH 6.0 and 23%w/v polyethylene glycol 3,350 and then soaked in crystallization solution containing additional 10% v/v glycerol

for cryo-protection. Diffraction data were collected at the Shanghai Synchrotron Radiation Facility BL17U1 at 100 K. Data were pro-

cessed and scaled using HKL2000 (Otwinowski and Minor, 1997).

Structure of the ZMYND8 PBP/Drebrin ADF-H complex was solved by molecular replacement with the models of the ZMYND8

PBP apo form (PDB code: 4COS) and the Drebrin-like ADF-H domain (PDB code: 1X67) using PHASER (Mccoy et al., 2007). Further

manual model building and refinement were completed iteratively using Coot (Emsley et al., 2010) and PHENIX (Adams et al., 2010).

The final model was validated byMolProbity (Chen et al., 2010). The final refinement statistics are summarized in Table 1. All structure

figures were prepared by PyMOL (http://www.pymol.org).

Cell Imaging and Data AnalysisHEK293T cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), and

1% of penicillin-streptomycin at 37�C with 5% CO2. Cells were transfected using ViaFect� Transfection Reagent (Promega) as per

manufacturer’s protocol.

All the images were acquired using a Zeiss LSM 710 laser-scanning confocal microscope, with a 633 1.4 oil objective and pinhole

setting to 1 Airy unit. Data were collected from 3 independent batches of cultures. In each batch, at least 300 fluorescence-positive

cells were counted for each group of experiments. Experimentswere conducted in a double-blinded fashion. The distribution ratios in

nucleus and cytoplasm were quantified and analyzed using GraphPad Prism 6. Data were expressed as mean ± SD. All other groups

were compared with Myc-ZMYND8 only group by one-way ANOVA with Tukey’s multiple comparison test.

QUANTIFICATION AND STATISTICAL ANALYSIS

Data of nuclear localization assays were collected from three independent batches of cultures (>300 fluorescence positive cells per

group per batch). Data were expressed as mean ± SD and analyzed using one-way ANOVAwith Tukey’s multiple comparison test by

GraphPad Prism 6; ns: not significant; ****: p<0.0001. All experiments were performed in a double-blinded fashion.

DATA AND SOFTWARE AVAILABILITY

Data ResourcesThe atomic coordinates of ZMYND8 PBP/Drebrin ADF-H complex have been deposited to the Protein Data Bank under accession

code PDB: 5Y1Z.

Structure 25, 1657–1666.e1–e3, November 7, 2017 e3

Structure, Volume 25

Supplemental Information

The Structure of the ZMYND8/Drebrin

Complex Suggests a Cytoplasmic

Sequestering Mechanism of ZMYND8 by Drebrin

Ningning Yao, Jianchao Li, Haiyang Liu, Jun Wan, Wei Liu, and Mingjie Zhang

 

 

 

Figure S1: Biochemical characterizations of ZMYND8/Drebrin-like and ZMYND11/Drebrin

interaction. Related to Figure 3.

(A) Domain organizations of Drebrin/Drebrin-like and ZMYND8/ZMYND11. (B) Analytical gel

filtration based assay and ITC-based assay showing no interaction between ZMYND8 PBP

tandem and Drebrin-like ADF-H domain. (D) ITC-based assay showing that Drebrin ADF-H

domain carrying R10G mutation failed to bind to ZMYND8 PBP tandem. (E&F) Analytical gel

filtration based assay and ITC-based assay showing that ZMYND11 BP tandem cannot bind to

Drebrin ADF-H domain in vitro.

   

 

 

 

Figure S2: Sequence alignment of Drebrin and Drebrin-like ADF-H domains from different

species. Related to Figure 3.

The secondary structure elements are labelled according to the Drebrin ADF-H structure. Residues

that are identical and highly similar are indicated in dark green and light green boxes, respectively.

The residues involved in ZMYND8 binding are highlighted with magenta boxes.