Autoinhibitory Structure of the WW Domain of HYPB/SETD2 Regulates Its Interaction with the Proline-Rich Region of Huntingtin

Structure Article

Yong-Guang Gao,1,* Hui Yang,1 Jian Zhao,1 Ya-Jun Jiang,1 and Hong-Yu Hu1,* 1State Key Laboratory of Molecular Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai 200031, China *Correspondence: [email protected] (Y.-G.G.), [email protected] (H.-Y.H.) http://dx.doi.org/10.1016/j.str.2013.12.005

SUMMARY fragment, a monoclonal antibody recognizing Htt PRR, significantly inhibits the aggregation, as well as the neurotoxicity, Huntington’s disease (HD) is an autosomally domi- induced by mutant Htt protein (Khoshnan et al., 2002; Southwell nant neurodegenerative disorder caused by expan- et al., 2008). Many tryptophan (WW) or Src homology 3 (SH3) sion of polyglutamine (polyQ) in the huntingtin (Htt) domain-containing proteins have been identified that interact protein. Htt yeast two-hybrid protein B (HYPB/ with the PRR of Htt, such as Htt yeast two-hybrid protein SETD2), a histone methyltransferase, directly inter- HYPA (also named FBP11) and HYPB (known as SETD2) (Faber acts with Htt and is involved in HD pathology. et al., 1998), SH3GL3/endophilin-A3 (Sittler et al., 1998), protein kinase C and casein kinase substrate in neuron 1 (PACSIN1/ Using NMR techniques, we characterized a poly- syndapin) (Modregger et al., 2002), and CA150 (Holbert et al., proline (polyP) stretch at the C terminus of HYPB, 2001). Many investigations have revealed that mutant Htt with which directly interacts with the following WW expanded polyQ impaired its normal interactions with these domain and leads this domain predominantly to partner proteins. So, it is speculative that HD neuropathology be in a closed conformational state. The solution is related to the interference of the normal function of these pro- structure shows that the polyP stretch extends teins by polyQ expansion (Goehler et al., 2004; Harjes and from the back and binds to the WW core domain Wanker, 2003; Landles and Bates, 2004; Li et al., 2003; Li and in a typical binding mode. This autoinhibitory struc- Li, 2004). ture regulates interaction between the WW domain Human HYPB/SETD2 was previously found by yeast two- of HYPB and the proline-rich region (PRR) of Htt, hybrid assay to interact with the N-terminal PRR region of as evidenced by NMR and immunofluorescence Htt protein through its WW domain (Faber et al., 1998; Passani et al., 2000). WW domains are small protein interaction techniques. This work provides structural and modules that are found in many eukaryotic proteins (Bork mechanistic insights into the intramolecular regula- and Sudol, 1994). The domain folds into a stable, triple- tion of the WW domain in Htt-interacting partners stranded b sheet structure and interacts with the proteins and will be helpful for understanding the pathology with proline-rich or proline-containing motifs (Chen and Sudol, of HD. 1995; Kay et al., 2000; Macias et al., 1996). Further studies have revealed that HYPB, as a histone H3 lysine 36 (H3K36)- specific methyltransferase, is responsible for virtually all global INTRODUCTION and transcription-dependent H3K36 trimethylation throughout the nucleus (Edmunds et al., 2008; Sun et al., 2005). HYPB Huntington’s disease (HD) is an autosomally dominant neuro- contains a polyproline (polyP) region, which is a potential degenerative disorder caused by the expansion of polyglut- WW domain-binding segment. This polyP region is followed amine (polyQ) in huntingtin (Htt), a ubiquitously expressed exactly by the WW domain, suggesting a possibility that the protein with an as yet unknown function (DiFiglia et al., 1995; polyP peptide segment interacts with the following WW MacDonald et al., 1993). The underlying mechanism leading domain intramolecularly. to neurodegeneration in HD has not been fully elucidated. By using nuclear magnetic resonance (NMR) and other tech- Human Htt is a large protein containing 3,144 residues with niques in this study, we revealed that the preceding polyP stretch an N-terminal polyQ domain (MacDonald et al., 1993). The interacts with the following WW core domain of HYPB to form a polyQ domain usually contains 35 or fewer residues in unaf- structural entity. This WW-polyP interaction leads to the WW fected individuals, whereas the number in HD patients extends domain predominantly in a closed conformational state. More- to 37 or more glutamine residues (Bates et al., 2002). A proline- over, the interaction between the WW domain of HYPB and rich region (PRR) containing 40 residues directly follows the PRR of Htt is finely modulated intramolecularly by the polyP polyQ domain in sequence. stretch. We propose that this autoinhibitory structure exquisitely The involvement of PRR in the pathological process of HD can regulates the interaction of HYPB with Htt, which will help us be observed from the fact that MW7 single-chain variable better understand the pathology of HD.

WW domain intramolecularly. We further characterized the intramolecular interaction between PP2 and WW in the PP2WW fragment by conducting a paramagnetic relaxation enhancement (PRE) experiment (Figure S1B). The peaks of N20, V36, I37, and T38 in the WW domain were signiﬁcantly broadened when S-(2,2,5,5-tetramethyl-2,5-dihydro-1H-pyrrol-3-yl)methyl meth- anesulfonothioate was labeled to the N terminus of PP2WW (Fig- ure S1C). This reconﬁrms that the ligand-binding surface of the WW domain in PP2WW is covered by its N-terminal polyP segment.

The PP2WW Domain Has Two Conformers in Solution PP2WW is a well-structured domain with intramolecular interactions between the WW core and its preceding polyP stretch. Interestingly, PP2WW exhibited two sets of resonance peaks, the major signals and the minor signals, from residues such as D26, Y34, and Q42 in the 1H-15N HSQC spectrum (Figure 2A), suggesting that PP2WW has two conformational states in equilibrium in solution. The pattern of the minor signals in PP2WW was very similar to that of the WW domain (Figure 2B). These signals correspond to the open state, in which the polyP stretch does not contact the core domain. The major signals are consid- ered to belong to the closed state, in which the polyP stretch Figure 1. Relationship between the WW Domain and Its Preceding contacts the core domain to form a structural entity. On the basis polyP Stretches of HYPB of the peak intensities (heights) in the HSQC spectra, we esti- (A) Domain architecture of HYPB highlighting the amino acid sequences of the mated that approximately an 86% population of PP2WW is in WW domain and its preceding polyP stretches. The sequence of human HYPB/SETD2 is based on the database update (RefSeq:NP_054878 and the closed conformation and a 14% population is in the UniProtKB:Q9BYW2). SET, SET domain; PRR, proline-rich region; PP, poly- open conformation. We prepared two mutants of PP2WW by 5 7 9 11 proline; PP1 and PP2, the two polyP stretches; PP12WW, WW domain with substituting PSP and KPK to AAA, respectively, in the PP2 PP12 (PP1 and PP2). region—namely, SPP2WW and KPP2WW—and examined (B) Amino acid sequences of WW and PP2WW from HYPB. The N-terminal which residues in the polyP region were responsible for contact- GlySer residues are from the cloning sties. PP2WW, WW domain with PP2. ing the WW core. As a result, SPP2WW was totally in the (C) Overlay of the 1H-15N HSQC spectra of WW (red) and PP12WW (green). (D) Overlay of the spectra of WW (red) and PP2WW (green). The concentration open fraction (Figure 2C), whereas KPP2WW was absolutely 5 7 for each sample is 500 mM. The chemical shift assignments of the WW in the closed conformation (Figure 2D). Thus, the PSP tripep- domain are also labeled in the spectra. tide is important for PP2WW to maintain the closed state, See also Table 1 and Figure S1. whereas 9KPK11 is necessary for the open conformation. These open and closed states were also visualized by fluorescence RESULTS emission spectra of these WW domain proteins (Figure S2). As previously reported in the literature (Jiang et al., 2001; Koepf The polyP Stretch of HYPB Interacts with the WW et al., 1999), WW and SPP2WW have a maximum emission Domain Intramolecularly wavelength of approximately 352 nm, whereas PP2WW and HYPB/SETD2 is a SET domain-containing protein (Wagner and KPP2WW have a maximum emission wavelength at 347 nm. Carpenter, 2012). It functions as a methyltransferase as its yeast The blue shifts of the fluorescence maximums reflect that the ortholog Set2p (Sun et al., 2005). Sequence analysis indicates WW moieties in PP2WW and KPP2WW are covered by the pre- that HYPB contains a polyP region followed directly by a WW ceding polyP stretch; that is, both of them are predominantly in domain in the C terminus (Figure 1A). The polyP region can be the polyP-bound closed states. divided into two peptide segments: PP1 and PP2. To test whether these peptide segments interact with the following Solution Structure of the PP2WW Domain from HYPB WW core domain intramolecularly, we constructed three protein To understand the mechanism underlying the conformational ex- fragments of HYPB: PP12WW, PP2WW, and WW (Figure 1B). change of PP2WW in detail, we determined the solution struc- Next, we acquired their 1H-15N heteronuclear single quantum ture of PP2WW in solution by using NMR techniques. PP2WW correlation (HSQC) spectra and performed chemical shift assign- has two conformers in equilibrium, which may give rise to diffi- ments of WW and PP2WW (Figure S1A available online). Com- culty in structural analysis by NMR techniques. We first per- parison of the spectra of PP12WW and PP2WW with the WW formed chemical shift and nuclear Overhauser effect (NOE) spectrum showed that many resonance peaks corresponding assignments of WW and KPP2WW and determined their solution to the residues resided on the typical ligand-binding surface of structures (Table 1). In general, the WW structure mimics the the WW domain were significantly perturbed by the polyP region open-state structure of PP2WW (Figure S3A), whereas (Figure 1C), especially by the PP2 segment (Figure 1D). This in- KPP2WW, similarly to PP2WW, is totally in a closed state (Fig- dicates that the PP2 segment interacts directly with the following ure S3B). On the basis of these two structures, we determined

Figure 2. PP2WW Has Two Conformational States in Solution (A) A representative region of the 1H-15N HSQC spectrum of PP2WW. In the spectrum, there are two sets of signals for some residues: one major peak represents the closed state (c), and the minor one indicates the open state (o). (B) As in (A), the spectrum of WW shows only single peaks. (C) As in (A), the spectrum of SPP2WW shows only single peaks corresponding to the open state. SPP2WW, a mutant of PP2WW replacing 5PSP7 with AAA. (D) As in (A), the spectrum of KPP2WW shows only single peaks corresponding to the closed state. KPP2WW, a mutant of PP2WW replacing 9KPK11 with AAA. See also Figure S2. the solution structure of the closed conformer of PP2WW. The Binding Specificities of the HYPB WW Domain with the experimental restraints, structural statistics, and some typical Htt-PRR Portions NOE correlations of PP2WW and other forms are displayed in HYPB is one of the interacting partners of Htt protein (Faber Table 1 and Table S1. In the structure, the WW domain moiety et al., 1998), in which the N-terminal PRR motif is proposed to retains a canonical WW fold composed mainly of an antiparallel specifically interact with the WW domain of HYPB. However, b sheet of three strands (Figures 3A and 3B). The N-terminal HYPB stays in an autoinhibitory state in most events, whereby polyP segment extends from the back of the ligand-binding sur- the WW core domain is covered by the preceding polyP stretch. face of the WW domain and contacts the core domain in a typical To elucidate the mechanism by which HYPB interacts with binding mode (Figure 3B). The WW core domain contacts the Htt PRR, we synthesized three peptides corresponding to the polyP peptide chain through the b3 strand and the loop between N-terminal (Pept-1), central (Pept-2), and C-terminal (Pept-3) b1 and b2 strands. NOE signals were detected between Leu4 portions of Htt PRR (Figure 4A) (Gao et al., 2006) and compared and Val36 and between Pro7 and Thr41 (Figure 3C). Abundant their binding specificities with the WW domain using an NMR contacts were also observed between the residues in the polyP titration method (Figure S4). The data showed that both Pept-1 segment (Pro7, Pro8, and Ser9) and those in the WW domain and Pept-3 interacted with the WW domain of HYPB (Figures (Trp21, Arg39, Gln40, and hr41). 4B, 4C, S4A, and S4C), whereas Pept-2 did not (Figure S4B). The structure of the WW domain is very similar to that of The peptide-binding properties of the WW domain of HYPB PP2WW in the open state (Figure S3A), whereas the KPP2WW with the Htt PRR peptides are very similar to those of the WW mutant is absolutely in the closed state without an open-state domain pair of HYPA (Gao et al., 2006; Jiang et al., 2011). For fraction (Figure S3B). Structural comparison of PP2WW with WW titration with Pept-123, a similar perturbation pattern was WW suggests that the local structure of the WW core domain observed in the HSQC spectra of the WW domain (Figure S4D), in PP2WW is not influenced considerably by contact with the except that many peaks became weak or disappeared to some polyP segment (Figure S3C). Compared with PP2WW, however, extent during titration. It suggests that binding of Pept-123 to KPP2WW has an additional short helix in the mutated site (Fig- WW may undergo conformational changes or intermediate ex- ure S3D) due to introduction of three Ala residues that may change at the NMR time scale. The saturation curves for the enhance a helical formation. chemical shift changes of the five representative residues in

Table 1. Experimental Restraints and Structural Statistics for the WW Domain Proteins Number of PP2WW KPP2WW experimental restraints WW (21–43) (3–43) (3–43) Total unambiguous 439 360 486 distance restraints Intraresidual 208 186 296 Sequential (j I–jj = 1) 122 103 121 Medium (2 % j i–jj % 4) 40 17 18 Long range (j I–jj R 5) 69 54 51 Dihedral angle restraints 52 89 94 Structural statistics Coordinate precision (A˚ )a Backbone (N, Ca, C’) 0.77 ± 0.21 1.16 ± 0.21 0.76 ± 0.12 Heavy atoms 1.78 ± 0.33 2.39 ± 0.38 1.80 ± 0.22 Backbone, 2nd structure 0.43 ± 0.11 0.55 ± 0.14 0.46 ± 0.13 Heavy atoms, 1.49 ± 0.41 1.62 ± 0.41 1.72 ± 0.29 2nd structure RMSD from experimental restraints NOE distances (A˚ ) 0.09 ± 0.011 0.15 ± 0.02 0.08 ± 0.01 Dihedral angles () 1.96 ± 0.3 3.15 ± 0.30 2.73 ± 0.17 RMSD from idealized geometry Bonds (A˚ ) 0.009 ± 0.000 0.007 ± 0.000 0.006 ± 0.000 Angles () 0.92 ± 0.03 0.87 ± 0.04 0.71 ± 0.04 Impropers () 2.44 ± 0.17 2.53 ± 0.1 2.20 ± 0.09 Ramachandran analysis Figure 3. NMR Solution Structure of PP2WW from HYPB Residues in most 74.2 82.3 86.6 (A) Superposition of the backbone traces of the 15 lowest-energy structures. favored regions The backbone trace of residues 3–43 is shown. Residues in additionally 20.7 16.3 12.3 (B) Ribbon diagram of a representative structure of PP2WW. allowed regions (C) Detailed view of the ribbon diagram of PP2WW. PP2WW is shown with both Residues in generously 4.32 0.16 0.97 surface and ribbon presentation. The side chains of residues that contact each allowed regions other are shown with sticks in the polyP stretch (red) and in the WW domain (blue). The structures shown were generated with MOLMOL software or the Residues in disallowed 0.81 1.17 0.12 UCSF Chimera tool. regions See also Figure S3. aThe coordinate precision is defined as the average RMS deviation among the 15 final structures. state of PP2WW (Figure 5B), whereas for KPP2WW, the peak shift pattern was like that of the closed state (Figure 5C). The the WW domain upon titration with Pept-1 and Pept-3 were well- plots for the chemical shift changes of the peaks from K30 fitted with 1:1 stoichiometry (Figures 4D and 4E), giving rise to against the concentration of Pept-1 are shown in Figure 5D. the dissociation constants (Kd) of about 300 mM for both pep- Along the titration curves, we observed that KPP2WW as well tides. This suggests that there is no distinction for the WW as the closed fraction of PP2WW bound very weakly to Pept-1, domain of HYPB interacting with the two peptide portions of whereas SPP2WW and the open fraction of PP2WW bound Htt PRR. with Pept-1 directly with considerable affinity. By fitting a 1:1

stoichiometry, the Kd value for SPP2WW binding with Pept-1 PP2WW Binds with Htt Pept-1 through Its Open State was calculated to be 250 mM, which is consistent with that for PP2WW contains two conformers in solution, and the molar ratio free WW binding with Pept-1 or Pept-3 (Figure 4). The half-satu- of the closed to open state is 6:1. We performed NMR titrations ration concentration (IC50) of the open fraction of PP2WW was of PP2WW, SPP2WW, and KPP2WW with Htt Pept-1 (Figure 5). significantly lower than that of SPP2WW, probably due to the Intriguingly, with the addition of Pept-1, both peaks for a residue kinetic limit of the exchange between the closed and open states of PP2WW, such as K30, corresponding to the closed and open in PP2WW. We also observed that the closed state of PP2WW states traced to a single peak corresponding to the bound state had a more significant peak shift than KPP2WW, indicating with Pept-1 (Figure 5A). This indicates that Pept-1 directly binds that an equilibrium shift from the closed to the open fraction to the open-state fraction of PP2WW and consequently causes occurred during ligand binding. Collectively, these observations shifting of the equilibrium from the closed to the open state. strongly suggest that the interaction of HYPB with Htt PRR is For SPP2WW titrated with Pept-1, the single peak of K30 shifted finely regulated by intramolecular interaction of the WW domain to the bound state in a pattern very similar to that of the open with its preceding polyP stretch.

Figure 4. Three Portions of Htt PRR Have Different Binding Afﬁnities with the WW Domain (A) Sequence of the PRR region from huntingtin (Htt-PP35). The PRR is divided into three portions: Pept-1, Pept-2, and Pept-3. (B and C) Chemical shift perturbations of the WW domain (100 mM) upon titration with Pept-1 (B) or Pept-3 (C) to a molar ratio of 5:1 (Pept to WW). (D and E) Titration curves for Pept-1 (D) or Pept-3 (E) binding to WW. By ﬁtting the data to a bimolecular binding algorithm, the dissociation constant values for Pept-1 and Pept-3 binding to WW are about 300 and 250 mM, respectively. See also Figure S4.

of the 15N-labeled GB1-PRR peptide with WW and PP2WW (Figure 6). The HSQC spectra showed that some resonance peaks of the residues in Htt PRR were significantly perturbed by WW domain binding, but PP2WW caused less effect on the peak intensities (heights) than WW (Figures 5A and 5B). Compared with WW, PP2WW exhibited reduced binding affinity with GB1-PRR Competitive Interaction of the WW Domain with Htt PRR (Figure 6C), indicating that the polyP stretch significantly in- Peptides hibited the binding of the WW core domain with Htt PRR. This To test whether the intramolecular polyP stretch in PP2WW suggests that competitive interaction with the WW domain influences its interaction with Htt PRR, we prepared a GB1 occurs between intramolecular and intermolecular proline-rich protein fusion of Htt PRR (GB1-PRR) and performed titration peptides.

Figure 5. Interaction of PP2WW with Htt Pept-1 in Equilibrium (A) Traces of the chemical shift changes of a representative residue (K30) upon titration with Pept-1. With the addition of Pept-1, both the major and minor peaks corresponding to the closed (c) and open (o) states move toward a merged one corresponding to the bound state (b) of PP2WW with Pept-1. The concentration of PP2WW is 100 mM. (B and C) As in (A), traces of the chemical shift changes (K30) of SPP2WW (B) or KPP2WW (C) upon titration with Pept-1 are shown. (D) Titration curves for PP2WW, SPP2WW, and KPP2WW binding with Pept-1. By ﬁtting the saturation curve of the chemical shift changes of SPP2WW against ligand concentration, the dissociation constant for SPP2WW binding with Pept-1 is 250 mM.

When cotransfected with Htt100Q, WW-NLS was colocalized with Htt100Q mostly in the cytoplasm, suggesting that it was sequestered to the cytoplasm from the nucleus by interacting with

Htt100Q (Figures 7A and 7C). Under similar conditions, however, PP2WW-NLS was not sequestered to the cytoplasm by Htt100Q (Figures 7B and 7C), owing to the inhibitory effect against its

binding with Htt100Q. When removal of PRR from Htt100Q (Htt100Q-DPRR), the mutant could not sequester WW-NLS or PP2WW-NLS to the cytoplasm (Figure 7D), indicating that this

sequestration by Htt100Q was dependent on the specific interaction between the WW domain and Htt PRR. Collectively, these results demonstrate that the intramolecular interaction in PP2WW significantly inhibits its binding with Htt, by which the polyP stretch finely regulates the sequestration of HYPB by

Htt100Q in cells.

DISCUSSION

Although both WW domain and proline-rich motif are very small fractions in a normal functional protein, their diverse interactions are critically important to transcription and signal transduction pathways (Neduva and Russell, 2007). These interactions should be deliberately orchestrated by different cellular factors or regulatory modes. Studying the interactions between Htt and its partners based on the small WW domains and PRR motifs are extremely important to understanding the functional regulation of Htt and the pathology of HD.

Intramolecular Interactions in Htt-Interacting Partners HYPB is a single WW domain-containing protein that harbors several polyP stretches in its entire sequence, suggesting a possibility that they have intramolecular interactions with the core WW domain for structurally allosteric regulation as previously reported in the literature (Cheetham, 2004; Laughlin et al., 2012; Trudeau et al., 2013). It is likely that the N-terminal polyP region also plays a role in modulating the interaction of HYPB with Htt by interacting with the C-terminal WW domain, but it may be diffi- cult for structural and mechanistic analysis. So, we investigated only the C-terminal polyP stretches preceding the central WW Figure 6. The polyP Stretch Impedes PP2WW Binding with Htt PRR domain. Our studies demonstrate that the WW domain of (A and B) NMR titration of 15N-labeled GB1-PP35 with WW (A) or PP2WW (B). HYPB specifically interacts with the preceding polyP stretch to The resonance peaks for PP35 are indicated by arrows. The concentration of form a structural entity and that this autoinhibitory structure pro- GB1-PRR is 100 mM. vides an arena for allosterically regulating its interaction with Htt (C) Titration curves for two representative peaks, peak 1 (red) and peak 2 PRR. This study reports that a polyP stretch acts on autoinhibi- (blue), from GB1-PP35 caused by addition of WW (triangle) or PP2WW (circle). tion of the nearby core WW domain and allosteric regulation of its binding to other proline-rich motifs. Analogously, Htt yeast Intramolecular Interaction in PP2WW Inhibits Its Binding two-hybrid protein C (HYPC) contains a pair of WW domains with Htt in a Cell Model and several polyP regions flanking the domain pair (Faber We next examined whether the interaction between the WW et al., 1998; Passani et al., 2000). The novel sequence and struc- domain of HYPB and Htt PRR was affected by the intramolecular ture are reminiscent of an intramolecular regulation of the WW interaction using a cell model (Jiang et al., 2011). We prepared domain pair by the polyP stretches for interacting with Htt two constructs of WW and PP2WW from HYPB fused with a PRR. These Htt-interacting partners that putatively have intra- nuclear localization sequence (NLS) (WW-NLS and PP2WW- molecular allosteric regulation deserve further exploration. NLS) and performed immunofluorescence experiments (Fig- ure 7). The confocal microscopic imaging showed that, in most Intramolecular Regulation of the WW Domain of HYPB HEK293T cells, the overexpressed WW-NLS and PP2WW-NLS for Interacting with Htt PRR were localized in the nucleus, whereas the overexpressed In our NMR studies, PP2WW maintained equilibrium between polyQ-expanded Htt (Htt100Q, residues 1–171) was mostly local- the closed and open states in solution. The closed state has a ized in the cytoplasm as reported previously (Jiang et al., 2011). major population (86%), whereas the open state is the minor

Figure 7. Confocal Fluorescence Microscopy Images Showing Colocalization of Htt100Q with WW or PP2WW in Cells

(A) Colocalization of Htt100Q with WW of HYPB in cytoplasm. Normally, HYPB WW fused with a C-terminal NLS sequence (WW-NLS) localizes in the nucleus, whereas Htt100Q (residues 1–171) localizes in the cytoplasm. Cotransfection with Htt100Q causes redistribution of WW from the nucleus to the cytoplasm because 0 of speciﬁc interaction of WW and Htt PRR. Cells were stained with FLAG antibody for Htt100Q, HA antibody for WW-NLS, and 4 ,6-diamidino-2-phenylindole for nuclei. Scale bar, 10 mm.

(B) No colocalization of Htt100Q with PP2WW of HYPB in cytoplasm. PP2WW fused with a C-terminal NLS (PP2WW-NLS) localizes in the nucleus. Cotransfection with Htt100Q loses the ability to cause redistribution of PP2WW due to weakening of the interaction between WW and Htt PRR by the polyP stretch. (C) Proportions of the transfected cells with colocalization. Data are from WW-NLS (A) and PP2WW-NLS (B), and data are presented as means ± SEM (n = 30).

(D) Deletion of PRR in Htt100Q (Htt100QDPRR) abolishes colocalization of Htt100Q with WW or PP2WW. one. Thus, PP2WW is predominantly in the closed state, effec- loses the ability to be redistributed by Htt. PolyQ-expanded Htt tively inhibiting the WW domain from interacting with other can sequester HYPB to the cytoplasm, and thus the biological PRRs, such as Htt PRR (Figure 8), leaving a small fraction of function of HYPB might be altered by this regulatory interaction. the open state that has binding ability for other proteins. When HYPB is a H3K36 methyltransferase that is very important for Htt PRR binds to the open state of PP2WW, it may cause the cellular function (Hu et al., 2010), so it is speculative that HYPB equilibrium to shift from the closed to the open state. It is also is relevant to the disease caused by polyQ expansion of Htt. possible that the state population of the WW domain of HYPB Therefore, understanding of the molecular mechanism by which could be regulated intramolecularly by other PRRs or even by intramolecular interaction of HYPB, as well as how HYPC regu- other proteins or factors. Therefore, elucidating the mechanistic lates its binding with polyQ-expanded Htt, will provide implica- details of the regulatory interaction between Htt and HYPB is tions for the disease pathology of HD. very important to our understanding of HD, because this speciﬁc interaction is likely to be involved in the disease pathology (Pas- EXPERIMENTAL PROCEDURES sani et al., 2000). Proteins and Peptides Implication in Cellular Regulation and Disease The coding sequences of the WW domain (PP12WW, PP2WW, and its Pathology mutants) were ampliﬁed by PCR and cloned into the pET-32M plasmid using Escherichia Previous studies revealed that HYPB is associated with the the BamHI/XhoI cloning sites. The plasmid was transformed into coli BL21 (DE3) cells. When the cells harboring each plasmid were grown at disease-related deposits of N-terminal Htt (Passani et al., 37C to an optical density 600 of 0.7, protein expression was induced by 2000). Herein we demonstrate that the WW domain in HYPB adding isopropyl b-D-1-thiogalactopyranoside. After cell lysis, the supernatant can be redistributed from the nucleus to the cytoplasm by inter- was loaded onto a nickel-nitrilotriacetic acid column (QIAGEN). After washing acting with Htt PRR and that the autoinhibitory form of PP2WW followed by on-column thrombin cleavage to remove the thioredoxin tag, the

hibited no NOE violation >0.3 A˚ and no dihedral violation >5. The structural graphs were prepared using MOLMOL software (Koradi et al., 1996) or the UCSF Chimera program (Pettersen et al., 2004).

NMR Titration NMR titrations of WW and PP2WW with Htt PRR peptides were performed at 25C. 15N-labeled proteins (100 mM) in the above-mentioned NMR buffer were added in stepwise fashion with Htt PRR peptides to give the peptide/protein molar ratios, which ranged from 0:1 to 10:1, with each step monitored by 2D 1H-15N HSQC spectra. The combined average chemical shift changes 2 2 1/2 were calculated as Ddave = [(0.17DdN) +(DdH) ] , where DdH and DdN are the chemical shift changes in 1H and 15N dimensions, respectively. The titration curves were ﬁtted assuming a bimolecular binding event as described pre-

viously (Chang et al., 2006). The Kd values were obtained by analyzing the chemical shift changes of three typical residues upon addition of each peptide following a 1:1 binding mode.

Cell Culture and Immunoﬂuorescence Microscopy

FLAG-tagged Htt100Q (residues 1–171) and Htt100QDPRR (deletion of PRR segment) were constructed previously (Jiang et al., 2011). WW and PP2WW fused with an NLS sequence (KRKRR) were cloned into the pcDNA3.0 vector (with an inﬂuenza hemagglutinin (HA) tag). HEK293T cells were transfected with indicated plasmids using PolyJet reagent (SignaGen Laboratories) according to the manufacturer’s instructions. The cells were grown on glass coverslips for 72 hr after transfection. Anti-FLAG (Sigma-Aldrich) and anti- HA (Santa Cruz Biotechnology) antibodies were used for immunoﬂuorescence microscopy. The images were obtained using a Leica TCS SP2 confocal microscope (Leica Microsystems). Figure 8. Schematic Representation of the Intramolecular Regula- tion of the WW domain of HYPB Interacting with Htt ACCESSION NUMBERS The structure of the WW domain in PP2WW of HYPB between the closed and open conformational states is in equilibrium. This intramolecular polyP inter- The atomic coordinates and supporting data reported in this paper have been action with the WW moiety in PP2WW impedes interaction of the oncoming deposited in the Protein Data Bank under the ID codes 2MDC (WW), 2MDJ PRR of the Htt protein, which forms aggregates through b sheet structures by (PP2WW), and 2MDI (KPP2WW). the expanded polyQ tracts.

SUPPLEMENTAL INFORMATION protein was eluted and subjected to chromatography on a Superdex 75 column (GE Healthcare). The coding sequence for Htt PRR corresponding to res- Supplemental Information includes four figures and one table and can be idues 44–78 of Htt was amplified by PCR and cloned into the pGBTNH plasmid found with this article online at http://dx.doi.org/10.1016/j.str.2013.12.005. using the BamHI/XhoI cloning sites (Bao et al., 2006). The expression and purification of Htt-PRR was the same as other proteins described previously, ACKNOWLEDGMENTS except that the fused GB1 tag was not removed. 15N- and/or 15N/13C-labeled 15 proteins were prepared by using M9 minimal medium containing NH4Cl and/ The authors would like to thank Dr. Q.H. Huang for providing the cDNA encod- 13 or C6-D-glucose as the sole nitrogen and/or carbon resource, respectively. ing the C-terminal part of HYPB/SETD2 and Meng Wu for technical help in The peptides (Pept-1, Pept-2, and Pept-3) were obtained by automatic NMR data acquisition. This work was supported by grants from the National solid-phase peptide synthesis as described previously (Gao et al., 2006). Basic Research Program of China (2011CB911104 and 2012CB911003) and the National Natural Science Foundation of China (31200570). NMR Spectroscopy To solve the solution structures of WW, PP2WW, and KPP2WW from HYPB by Received: September 14, 2013 15 13 NMR, N/ C-labeled samples containing 500 mM protein in a phosphate Revised: November 19, 2013 buffer (20 mM sodium phosphate, 50 mM NaCl, 0.05% w/v sodium azide, Accepted: December 5, 2013 pH 6.5, and 95% H2O/5% D2O or 100% D2O) were used for data acquisition. Published: January 9, 2014 All NMR data were recorded at 25C on a Bruker 600-MHz AVANCE III spec- trometer equipped with a TCI CryoProbe proton-optimized triple resonance REFERENCES NMR inverse probe (Bruker). The backbone and side-chain 1H, 15N, and 13C resonances were assigned based on the spectra of 3D HNCO, HNHA, Bao, W.J., Gao, Y.G., Chang, Y.G., Zhang, T.Y., Lin, X.J., Yan, X.Z., and Hu, HNCACB, CBCA(CO)NH, C(CO)NH, and HCCH-TOCSY. NOE distance re- H.Y. (2006). Highly efficient expression and purification system of small-size 15 straints for structure calculations were obtained from 3D N-edited and protein domains in Escherichia coli for biochemical characterization. Protein 13 C-edited NOESY spectra. Expr. Purif. 47, 599–606. Bates, G., Harper, P., and Jones, L. (2002). Huntington’s Disease. (Oxford, UK: Structure Calculation and Analysis Oxford University Press). NMR data processing and structure calculation were performed as reported previously by investigators at our laboratory (Gao et al., 2006; Jiang et al., Bork, P., and Sudol, M. (1994). The WW domain: a signalling site in dystrophin? 19 2011). For WW, PP2WW, and KPP2WW, the restraints used for structure Trends Biochem. Sci. , 531–533. calculation and statistics are summarized in Table 1. The structural calculation Chang, Y.G., Song, A.X., Gao, Y.G., Shi, Y.H., Lin, X.J., Cao, X.T., Lin, D.H., was performed for nine cycles and a total of two hundred structures were and Hu, H.Y. (2006). Solution structure of the ubiquitin-associated domain of ultimately obtained. The 15 lowest-energy structures were selected, which ex- human BMSC-UbP and its complex with ubiquitin. Protein Sci. 15, 1248–1259.

Cheetham, G.M. (2004). Novel protein kinases and molecular mechanisms of Koradi, R., Billeter, M., and Wuthrich, K. (1996). MOLMOL: a program for autoinhibition. Curr. Opin. Struct. Biol. 14, 700–705. display and analysis of macromolecular structures. J Mol Graph 14, 51–55, Chen, H.I., and Sudol, M. (1995). The WW domain of Yes-associated protein 29-32. binds a proline-rich ligand that differs from the consensus established for Landles, C., and Bates, G.P. (2004). Huntingtin and the molecular pathogen- Src homology 3-binding modules. Proc. Natl. Acad. Sci. USA 92, 7819–7823. esis of Huntington’s disease. Fourth in molecular medicine review series. EMBO Rep. 5, 958–963. DiFiglia, M., Sapp, E., Chase, K., Schwarz, C., Meloni, A., Young, C., Martin, E., Vonsattel, J.P., Carraway, R., Reeves, S.A., et al. (1995). Huntingtin is a cyto- Laughlin, J.D., Nwachukwu, J.C., Figuera-Losada, M., Cherry, L., Nettles, plasmic protein associated with vesicles in human and rat brain neurons. K.W., and LoGrasso, P.V. (2012). Structural mechanisms of allostery and auto- 20 Neuron 14, 1075–1081. inhibition in JNK family kinases. Structure , 2174–2184. Li, J.Y., Plomann, M., and Brundin, P. (2003). Huntington’s disease: a synapt- Edmunds, J.W., Mahadevan, L.C., and Clayton, A.L. (2008). Dynamic histone opathy? Trends Mol. Med. 9, 414–420. H3 methylation during gene induction: HYPB/Setd2 mediates all H3K36 trimethylation. EMBO J. 27, 406–420. Li, S.H., and Li, X.J. (2004). Huntingtin-protein interactions and the pathogenesis of Huntington’s disease. Trends Genet. 20, 146–154. Faber, P.W., Barnes, G.T., Srinidhi, J., Chen, J., Gusella, J.F., and MacDonald, MacDonald, M.E., Ambrose, C.M., Duyao, M.P., Myers, R.H., Lin, C., Srinidhi, M.E. (1998). Huntingtin interacts with a family of WW domain proteins. Hum. L., Barnes, G., Taylor, S.A., James, M., Groot, N., et al.; The Huntington’s Mol. Genet. 7, 1463–1474. Disease Collaborative Research Group (1993). A novel gene containing a trinu- Gao, Y.G., Yan, X.Z., Song, A.X., Chang, Y.G., Gao, X.C., Jiang, N., Zhang, Q., cleotide repeat that is expanded and unstable on Huntington’s disease chro- and Hu, H.Y. (2006). Structural insights into the specific binding of huntingtin mosomes. Cell 72, 971–983. proline-rich region with the SH3 and WW domains. Structure 14, 1755–1765. Macias, M.J., Hyvo¨ nen, M., Baraldi, E., Schultz, J., Sudol, M., Saraste, M., and Goehler, H., Lalowski, M., Stelzl, U., Waelter, S., Stroedicke, M., Worm, U., Oschkinat, H. (1996). Structure of the WW domain of a kinase-associated pro- Droege, A., Lindenberg, K.S., Knoblich, M., Haenig, C., et al. (2004). A protein tein complexed with a proline-rich peptide. Nature 382, 646–649. interaction network links GIT1, an enhancer of huntingtin aggregation, to Modregger, J., DiProspero, N.A., Charles, V., Tagle, D.A., and Plomann, M. 15 Huntington’s disease. Mol. Cell , 853–865. (2002). PACSIN 1 interacts with huntingtin and is absent from synaptic varicos- Harjes, P., and Wanker, E.E. (2003). The hunt for huntingtin function: interac- ities in presymptomatic Huntington’s disease brains. Hum. Mol. Genet. 11, tion partners tell many different stories. Trends Biochem. Sci. 28, 425–433. 2547–2558. Holbert, S., Denghien, I., Kiechle, T., Rosenblatt, A., Wellington, C., Hayden, Neduva, V., and Russell, R.B. (2007). Proline-rich regions in transcriptional 2007 M.R., Margolis, R.L., Ross, C.A., Dausset, J., Ferrante, R.J., and Ne´ ri, C. complexes: heading in many directions. Sci. STKE , pe1. (2001). The Gln-Ala repeat transcriptional activator CA150 interacts with Passani, L.A., Bedford, M.T., Faber, P.W., McGinnis, K.M., Sharp, A.H., huntingtin: neuropathologic and genetic evidence for a role in Huntington’s Gusella, J.F., Vonsattel, J.P., and MacDonald, M.E. (2000). Huntingtin’s WW disease pathogenesis. Proc. Natl. Acad. Sci. USA 98, 1811–1816. domain partners in Huntington’s disease post-mortem brain fulfill genetic criteria for direct involvement in Huntington’s disease pathogenesis. Hum. Hu, M., Sun, X.J., Zhang, Y.L., Kuang, Y., Hu, C.Q., Wu, W.L., Shen, S.H., Du, Mol. Genet. 9, 2175–2182. T.T., Li, H., He, F., et al. (2010). Histone H3 lysine 36 methyltransferase Hypb/ Setd2 is required for embryonic vascular remodeling. Proc. Natl. Acad. Sci. Pettersen, E.F., Goddard, T.D., Huang, C.C., Couch, G.S., Greenblatt, D.M., USA 107, 2956–2961. Meng, E.C., and Ferrin, T.E. (2004). UCSF Chimera—a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612. Jiang, X., Kowalski, J., and Kelly, J.W. (2001). Increasing protein stability using ¨ a rational approach combining sequence homology and structural alignment: Sittler, A., Walter, S., Wedemeyer, N., Hasenbank, R., Scherzinger, E., Stabilizing the WW domain. Protein Sci. 10, 1454–1465. Eickhoff, H., Bates, G.P., Lehrach, H., and Wanker, E.E. (1998). SH3GL3 asso- ciates with the Huntingtin exon 1 protein and promotes the formation of Jiang, Y.J., Che, M.X., Yuan, J.Q., Xie, Y.Y., Yan, X.Z., and Hu, H.Y. (2011). polygln-containing protein aggregates. Mol. Cell 2, 427–436. Interaction with polyglutamine-expanded huntingtin alters cellular distribution Southwell, A.L., Khoshnan, A., Dunn, D.E., Bugg, C.W., Lo, D.C., and and RNA processing of huntingtin yeast two-hybrid protein A (HYPA). J. Biol. Patterson, P.H. (2008). Intrabodies binding the proline-rich domains of mutant Chem. 286, 25236–25245. huntingtin increase its turnover and reduce neurotoxicity. J. Neurosci. 28, Kay, B.K., Williamson, M.P., and Sudol, M. (2000). The importance of being 9013–9020. proline: the interaction of proline-rich motifs in signaling proteins with their Sun, X.J., Wei, J., Wu, X.Y., Hu, M., Wang, L., Wang, H.H., Zhang, Q.H., Chen, 14 cognate domains. FASEB J. , 231–241. S.J., Huang, Q.H., and Chen, Z. (2005). Identification and characterization of a Khoshnan, A., Ko, J., and Patterson, P.H. (2002). Effects of intracellular novel human histone H3 lysine 36-specific methyltransferase. J. Biol. Chem. expression of anti-huntingtin antibodies of various specificities on mutant hun- 280, 35261–35271. tingtin aggregation and toxicity. Proc. Natl. Acad. Sci. USA 99, 1002–1007. Trudeau, T., Nassar, R., Cumberworth, A., Wong, E.T., Woollard, G., and Koepf, E.K., Petrassi, H.M., Ratnaswamy, G., Huff, M.E., Sudol, M., and Kelly, Gsponer, J. (2013). Structure and intrinsic disorder in protein autoinhibition. J.W. (1999). Characterization of the structure and function of W —> F WW Structure 21, 332–341. domain variants: identification of a natively unfolded protein that folds upon Wagner, E.J., and Carpenter, P.B. (2012). Understanding the language of ligand binding. Biochemistry 38, 14338–14351. Lys36 methylation at histone H3. Nat. Rev. Mol. Cell Biol. 13, 115–126.