Structural Basis for Recognition of Arginine Methylated Piwi Proteins by the Extended Tudor Domain

Structural basis for recognition of arginine methylated Piwi proteins by the extended Tudor domain

Ke Liua,b,1, Chen Chenc,1, Yahong Guob,1, Robert Lamb, Chuanbing Bianb, Chao Xub, Dorothy Y. Zhaoc, Jing Jinc, Farrell MacKenzieb, Tony Pawsonc,d,2, and Jinrong Mina,b,e,2

aHubei Key Laboratory of Genetic Regulation and Integrative Biology, College of Life Science, Huazhong Normal University, Wuhan 430079, People’s Republic of China; bStructural Genomics Consortium, University of Toronto, 101 College Street, Toronto, ON, M5G 1L7, Canada; cSamuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, ON, Canada M5G 1X5; dDepartment of Molecular Genetics, University of Toronto, Toronto, ON, Canada M5S 1A8; and eDepartment of Physiology, University of Toronto, Toronto, ON, Canada M5S 1A8

Contributed by Tony Pawson, September 2, 2010 (sent for review August 23, 2010)

Arginine methylation modulates diverse cellular processes and The Tudor domain, together with Chromo, MBT, PWWP, and represents a molecular signature of germ-line-specific Piwi family Agenet-like domains, comprise the “royal family” of protein proteins. A subset of Tudor domains recognize arginine methylation domains that engage in protein–protein interactions (16, 17). modifications, but the binding mechanism has been lacking. Here The Tudor domain is characterized by a β-barrel core and in many we establish that, like other germ-line Tudor proteins, the ancestral cases an aromatic cage suitable for docking methylated lysine or staphylococcal nuclease domain-containing 1 (SND1) polypeptide is arginine (16, 18). The Tudor domain family is largely divided into expressed and associates with PIWIL1/Miwi in germ cells. We find two groups: a methyllysine binding group (i.e., JMJD2A, 53BP1) that human SND1 binds PIWIL1 in an arginine methylation-depen- and a methylarginine binding group (SMN, TDRD group [Tdrd1- dent manner with a preference for symmetrically dimethylated Tdrd12]) (19). Crystal structures of 53BP1 tandem Tudordomains arginine. The entire Tudor domain and a bifurcated SN domain in complex with histone H4K20me2 and a JMJD2A hybrid Tudor domain with histone H3K4me3 have provided the structural basis are required for this binding activity, whereas the canonical Tudor BIOCHEMISTRY domain alone is insufficient for methylarginine ligand binding. Crys- for Tudor domain recognition of methylated lysine residues tal structures show that the intact SND1 extended Tudor domain through aromatic cages (20, 21). On the contrary, much less is forms a wide and negatively charged binding groove, which can known about the binding mechanism of methylarginine-specific accommodate distinct symmetrically dimethylated arginine pep- Tudor domains, although several lines of biochemical evidence tides from PIWIL1 in different orientations. This analysis explains suggest they have a preference for binding proteins with symme- how SND1 preferentially recognizes symmetrical dimethylarginine trically dimethylated arginine (sDMA) (22). A unique structural via an aromatic cage and conserved hydrogen bonds, and provides a feature of Tudor domains from the TDRD group is an additional α β general paradigm for the binding mechanisms of methylarginine- -helix and two -strands N-terminal to the canonical Tudor core containing peptides by extended Tudor domains. domain (19), which may assist this subfamily of Tudor domain in binding to its methylated arginine ligands, and are absent from Piwi-interacting RNA ∣ Tudor domain-containing proteins methyllysine-binding Tudor domains. This extended sequence domain is annotated as a Maternal Tudor domain (23). Among 12 TDRD Tudor proteins, SND1 (TDRD11, staphylococcal rginine methylation, the covalent addition of methyl group to nuclease domain-containing 1, Tudor-SN, P100) is the most an- arginine residues, is a form of protein posttranslational mod- A cient molecule to possess this N-terminal extension, and we have ification that is increasingly recognized for its importance in med- – proposed that this Tudor domain may be a precursor for all germ- iating protein protein interactions (1, 2). Three types of arginine line TDRD Tudor domains (19). Therefore, elucidating the bind- methylation—namely, monomethylation, asymmetrical dimethy- — ing of the SND1 extended Tudordomain to its ligands is of general lation, and symmetrical dimethylation are catalyzed by type I importance in understanding the structural determinants under- or type II protein arginine methyltransferases (PRMTs). Among lying specific arginine methylation-dependent interactions and these, PRMT5 is considered the major PRMT that induces sym- the recognition modes of arginine methylated peptide motifs. metrical dimethylation of arginine residues on target proteins (2). In order to test whether a full extended Tudor domain is Piwi proteins, the germ-line-specific clade of Argonaute family required for methylarginine binding, and to establish a prototypic proteins, together with their associated Piwi-interacting RNAs binding mechanism for arginine methylation-dependent Piwi- (piRNAs) are evolutionarily conserved in animals and play Tudor interactions, we have investigated the crystal structures of pivotal roles in transposon silencing during gametogenesis (3, 4). the human SND1 extended Tudor domain in complex with sym- Mammalian Piwi proteins, which include PIWIL1/Miwi, PIWIL2/ metrically dimethylated arginine peptides derived from PIWIL1. Mili, PIWIL3 and PIWIL4/Miwi2, contain multiple arginine-gly- cine (RG) and arginine-alanine (RA) repeats at their N termini, and methylation of these arginine sites is important for recruiting Author contributions: K.L., C.C., Y.G., T.P., and J.M. designed research; K.L., C.C., Y.G., R.L., C.B., C.X., D.Y.Z., and F.M. performed research; K.L., C.C., Y.G., R.L., C.B., C.X., D.Y.Z., J.J., T.P., the Tudor domain-containing (TDRD) group of germ-line-en- and J.M. analyzed data; and C.C., T.P., and J.M. wrote the paper. riched Tudor domain proteins such as Tdrd1, Tdrd2/Tdrkh, The authors declare no conflict of interest. Tdrd4/RNF17, Tdrd5, Tdrd6, Tdrd7, Tdrd8/Stk31, and Tdrd9 Freely available online through the PNAS open access option. (5–12). Arginine methylation is a conserved posttranslational Data deposition: The atomic coordinates, crystallography, and structure factors have modification of Piwi family proteins of various species including been deposited in the Protein Data Bank, www.pdb.org (PBD ID codes 3OMC and 3OMG). Drosophila Xenopus , , and mammals (7, 13). Its importance has 1K.L., C.C., and Y.G. contributed equally to this work. been demonstrated by null mutants of drosophila PRMT5 (dart5, 2To whom correspondence may be addressed. E-mail: [email protected] or jr.min@ csul), in which arginine methylation of Piwi family members is utoronto.ca. compromised, leading to retrotransposon activation and defects This article contains supporting information online at www.pnas.org/lookup/suppl/ in germ-line development (7, 14, 15). doi:10.1073/pnas.1013106107/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1013106107 PNAS Early Edition ∣ 1of6 Downloaded by guest on September 29, 2021 These data delineate the structural basis for preferential binding To examine its potential association with Miwi, we immunopre- of symmetrically dimethylated arginine by an extended Tudor cipitated SND1 from testis lysates and immunoblotted with Miwi domain, which is of general relevance to the family of germ-line antibody and found that SND1 coprecipitated with Miwi as com- mammalian Tudor proteins, and identify distinct mechanisms by pared with a negative control IgG immunoprecipitation (Fig. 1B). which the SND1 extended Tudor domain recognizes arginine These results indicate a potential unique involvement of the mul- methylated peptide motifs. tifunctional SND1 in the Piwi/piRNA pathway in the germ cells. SND1 consists of four staphylococcal nuclease-like (SN) Results domains, and a Tudor domain, which is embedded in a fifth SN SND1 Associates with Miwi in Vivo, and Its Extended Tudor Domain domain (Fig. 2A). We refer to the Tudor domain and the fifth Preferentially Binds Piwi Peptides with sDMA. Several TDRD pro- SN domain together as an extended Tudor domain in this study. teins associate with Piwi proteins in mouse testes or in cotrans- To pursue the endogenous association with Piwi identified by fected cells, and their interactions are most likely arginine immunoprecipitation in the mouse, and to better understand the methylation-dependent (5). For example, we have established binding specificity of SND1, we used a fluorescence polarization that Tdrkh is a major binding partner of Miwi, and that the Tudor binding assay to measure the affinity of the human SND1 extended domain is critical in mediating Tdrkh–Miwi interaction (8). Tudor domain for an N-terminal peptide from human PIWIL1 Because the crystal structure of the apo-SND1 Tudor domain protein, which is 100% conserved with mouse Miwi. This peptide is highly similar to that of the Miwi-binding germ-line Tudor pro- contains multiple RA/RG repeats, and was synthesized with a tein Tdrkh, in terms of overall binding surface and aromatic cage symmetrical dimethylation modification at one of three different residue position (8, 24), we asked whether SND1 is capable of arginine sites (R4, R10, or R14) (Fig. 1 C–E). The SND1 extended D binding Miwi. Because it is not known whether the SND1 protein Tudor domain bound all three of these sDMA peptides (Fig. 1 E is expressed in male germ cells, we performed immunoblotting and ). Simultaneous methylation of all three arginine residues to examine its tissue distribution in mice. SND1 was highly did not markedly affect binding. SND1 bound to the R4me2s pep- K μ expressed in pancreas and liver, with moderate to weak levels in tide (symmetrical dimethylation at arginine 4) with a D of 10 M. other tissues including testis (Fig. S1). We next analyzed SND1 The binding affinity was about twofold weaker for a peptide that expression dynamics during spermatogenesis (Fig. 1A). SND1 was monomethylated at arginine 4 (R4me1), about fourfold weak- was undetectable or weakly expressed at postnatal week 1, but er for a peptide with asymmetrical dimethylation at arginine 4 (R4me2a), and about ninefold weaker for an unmodified R4 exhibited elevated protein levels from postnatal week 2 to adult, D E correlating with the onset of meiosis and Miwi expression (25). peptide at a salt concentration of 50 mM NaCl (Fig. 1 and ). Hence, the SND1 extended Tudor domain preferentially recognizes symmetrically arginine-dimethylated PIWIL1/Miwi pep- A Testis C ES 1wk 2wk 3wk 8wk DLD1 PIWIL1 N terminus tides. At a higher salt concentration of 150 mM NaCl, the 130 95 SND1 binding affinities were reduced roughly fourfold for each peptide 72 D (Fig. 1 D and E), indicating that electrostatic interactions play a Tubulin role in the SND1 interaction with PIWIL1 N terminus.

Overall Structures of the SND1 Extended Tudor Domain in Complex B with Two PIWIL1 Peptides, R4me2s and R14me2s. To gain insight into SND1 IP SND1 IP Miwi IP IgG IP IgG IP the molecular mechanism of selective binding of the SND1 extended Tudor domain to symmetrically arginine-dimethylated Miwi proteins, we determined the crystal structures of the SND1 IgG extended Tudor domain in complex with R4me2s and R14me2s WB: Miwi peptides (Fig. 2 and Fig. S2). The apostructure of the SND1 E extended Tudor domain has been determined before (24) and is almost identical to that in our structures of the domain complexed with peptides. Based on our structures of these complexes, we can divide the SND1 extended Tudor domain into three modules: the canonical Tudor domain colored in blue, the fifth SN-like domain colored in cyan and green, and the linker α-helix colored in yellow. The SN-like domain is split into two segments by the Tudor domain: the N-terminal two ß-strands colored in cyan and the C-terminal three α-helices and three ß-strands colored in green (Fig. 2). The canonical Tudor domain, the linker α-helix, and the N-terminal two ß-strands of the SN domain form the maternal Tudor domain sequence. In order to test the significance of the N- and C-terminal segments of the SN domain in Piwi protein binding, we made two deletion mutants by separately truncating the N- and C-term- Fig. 1. SND1 associates with Miwi in vivo and its extended Tudor domain inal parts of the SN domain. Binding data obtained from these preferentially binds Piwi peptides with sDMA. (A) Expression kinetics of mutants showed that deleting the N-terminal two ß-strands of SND1 in mouse testes of different developmental ages. ES, mouse embryonic the SN domain (construct covering residues 700–910) totally stem cells; DLD1, a human colon carcinoma cell line. (B) Association of disrupted binding, and deleting the C-terminal part of the SN SND1 with Miwi in adult testis. Testis lysates was immunoprecipitated using domain (residues 650–800) severely diminished binding (KD ¼ anti-SND1, anti-Miwi, and IgG control antibodies and subject to anti-Miwi 300 μM) (Table 1). Therefore, both segments of the SN domain antibody immunoblotting. (C) The N terminus sequence of human PIWIL1 are critical for binding of the extended Tudor domain to its or mouse Miwi. The methylation of those arginine residues colored in red is studied in this work. (D) The binding curves of fluorescence polarization ligands. TDRKH is a germ cell-specific Piwi-binding Tudor pro- measurements of different PIWIL1-R4 peptides to SND1 extended Tudor tein, which has a 25% sequence identity and 43% sequence simi- domain. (E) Binding affinities of SND1 extended Tudor domain to different larity with the extended Tudor domain of SND1, and is therefore PIWIL1 peptides. predicted to possess an extended Tudordomain (Fig. S2C). Totest

2of6 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1013106107 Liu et al. Downloaded by guest on September 29, 2021 A

Fig. 2. Crystal structures of human SND1 extended Tudor domain in complex with PIWIL1 peptides. (A) BIOCHEMISTRY Domain structure of SND1. SN, staphylococcal nuclease-like domain. (B) Complex structure of SND1 and R4me2s peptide: The SND1 extended Tudor domain is shown in cartoon, and the R4me2s peptide and residues interacting with the R4me2s peptide from SND are shown in stick models. Hydrogen bonds are shown in red dotted lines. (C) Complex structure of SND1 and R14me2s peptide: The SND1 extended Tudor domain is shown in cartoon, and the R14me2s peptide and residues interacting with the R14me2s peptide from SND are shown in stick models. Hydrogen bonds are shown in red dotted lines.

if the binding characteristics of the extended Tudor domain are that all three modules (the N- and C-terminal SN segments conserved, we measured the binding affinities of TDRKH Tudor and the Tudor core) are involved in binding the methylarginine domain with different boundaries to the R4me2s peptide and peptides (Fig. 2). Together the three modules form a wide nega- found that TDRKH (residues 290–561) binds R4me2s with a KD tively charged binding groove that accommodates the methylar- of 4 μM. As predicted, binding was severely diminished (KD ¼ ginine containing peptides. This finding is consistent with the loss 310 μM) following C-terminal truncation of the TDRKH Tudor of binding displayed by the deletion mutants, because the struc- domain (residues 290–430), and was undetectable for a construct tures predict that deleting any part of the extended Tudor domain covering just the canonical Tudor domain (residues 327–420) would severely diminish peptide binding. (Table 1). These truncation mutants are well behaved during purification and one truncation mutant was crystallized previously PIWIL1 R4me2s and R14me2s Peptides Bind the Extended Tudor (TDRKH, residues 327–420) (8). Therefore, both the N- and Domain of SND1 in Reverse Directions. Surprisingly, the R4me2s C-terminal parts of the split SN domain of SND1, in addition and R14me2s peptides bound the SND1 extended Tudor domain to the Tudorcore domain, are essential for ligand binding, and this in reverse directions, although the methyl-guanidinium group of arrangement is likely conserved in the extended Tudor domains of the symmetrical dimethylarginine could be well superimposed germ-line-specific Tudor proteins such as TDRKH and Tudor, a (Fig. 3A). Both R4me2s and R14me2s peptides form extensive notion that is supported by a recent study published during the interactions with SND1 extended Tudor domain (Fig. 2). In preparation of our manuscript (26). the crystal structure of the SND1-R4me2s complex, the Ala7 of Because the SN domain contains five β-strands, which resem- the peptide is inserted into a shallow pocket and Ala5 is pointed bles a Tudor domain, we overlaid the fifth SN domain and the to the sidewall of the binding groove (Fig. 2B and Fig. S3). Arg6 canonical Tudor domain of SND1, and found that the ß-barrel forms electrostatic and water-mediated hydrogen bonding inter- core of the SN domain can be superimposed with the Tudor actions with SND1 (Fig. 2B and Fig. S3). This conformation can domain (Fig. S2B), although the Tudor-like ß-barrel core of the explain why the RA/RG motif is preferred by SND1. However, in SN domain lacks an aromatic cage. Therefore we propose that the case of the SND1-R14me2s complex, if the R14me2s still the extended Tudor domain harbors unique interdigitated tan- adopted the same orientation as the R4me2s peptide, the dem Tudor domain folds. Glu17 residue would occupy the pocket filled by Ala7 in the case The mechanistic basis for the importance of the SN domain in of the SND1-R4me2s structure, which would be too shallow for ligand binding was revealed by the structures of the extended Glu17, and more importantly would cause charge expulsion SND1 Tudor domain with complexed peptides. These showed (Fig. 2 and Fig. S3). Therefore, from an energetic point of view,

Liu et al. PNAS Early Edition ∣ 3of6 Downloaded by guest on September 29, 2021 Table 1. Binding affinities of SND1 and TDRKH deletion mutants to ever, 53BP1 and L3MBTL1 both preferentially bind low methyla- a PIWIL1 peptide (TGRme2sARARARGRARGQE) tion states of lysine in histone tails (20, 29). The methyllysine- binding pocket consists of three to four aromatic residues and SND1 mutant name KD, μM an aspartic acid, and interacts with the dimethylammonium group SND1-deltaN No binding of the methyllysine. By comparing the pocket dimensions, espe- SND1-deltaC 300 ± 50 cially the parallel aromatic residues, we found that the distance TDRKH (290–530) 4 ± 1 – between the F740 and Y766 in SND1 is 1.2 Å narrower than that TDRKH (290 430) 310 ± 50 B C E TDRKH (327–420) No binding between the Y1502 and Y1523 in 53BP1 (Fig. 3 , , and ). The pocket size limits SND1 so that it can only accommodate the planar methyl-guanidinium group. This finding is confirmed by it is favorable for the R14me2s peptide to reside in a reverse di- our binding results, which showed that SND1 did not bind any rection in the wide binding groove (Fig. 3A), reminiscent of SH3 of the lysine-methylated histone peptides we tested (Fig. 1E). domains, which used a conserved binding surface to recognize On the other hand, histone N-terminal tails also harbor a few proline-rich peptides in opposite directions (27, 28). methylatable arginine sites, such as H3R2 and H4R3. We have By superimposing the R4me2s and R14me2s structures, we shown that SND1 exhibits some degree of plasticity to accommo- note that the peptides do not overlay well except for the methy- date various methylarginine peptides. Our binding results show larginine group (Fig. 3A). It is conceivable that the wide binding that SND1 bind symmetrically arginine-dimethylated histone pep- groove of the SND1 extended Tudor domain provides plasticity to tides, albeit a few times more weakly than PIWIL1 peptides accommodate methylarginine peptides in different binding (Fig. 1E and Fig. S4). modes (Fig. 3A). In addition to the first cluster of RA/RG repeats in PIWIL1, Piwi proteins harbor methylarginines in other se- Discussion quence contexts (5). The wide binding groove of the extended Protein arginine methylation, catalyzed by PRMTs, modulates cel- Tudor domain in SND1 potentially enables it to recognize multi- lular processes such as mRNA splicing, DNA repair, transcription ple different methylarginine motifs, a prediction that warrants regulation, and signal transduction. These functions are believed further investigation. to be mediated by methylarginine binding proteins. Thus far, a subset of Tudordomains from proteins, such as SMN and TDRDs, SND1 Extended Tudor Domain Preferentially Recognizes Symmetri- are the only modules identified as binding arginine methylation cally Arginine-Dimethylated Peptides. Like other royal family mem- marks. Here we reported the crystal structures of the extended bers, the binding of the symmetrical dimethylated arginine by the Tudor domain of SND1 in complex with two PIWIL1 N-terminal extended Tudor domain involves an aromatic cage. The aromatic peptides harboring single symmetrically dimethylated arginines, residues F740, Y746, Y763, and Y766 form a rectangle cuboid and provide the structural basis for the binding preference of cage with the aromatic rings of F740 and Y766 parallel to each the extended Tudor domain for symmetrically dimethylated argi- other, and those of Y746 and Y763 approximately perpendicular nine. Like methyllysine-binding domains, the SND1 extended to those of F740 and Y766 (Fig. 3 B and C). The parallel aromatic Tudor domain uses an aromatic cage to recognize methylated rings of F740 and Y766 flank the side chain of the methylarginine. residues. Interestingly, in their unliganded forms, the Tudor Besides the aromatic cage, the methylarginine forms a hydrogen domains of methylarginine-binding SMN and TDRD3 also have bond with N768 through the NH1/2 group, and interacts with such an aromatic cage, although in the SMN structure the aromatic N823 and T688 through water-mediated hydrogen bonds (Fig. 3 cage is not preformed because W102 occupies the binding pocket B and C). Residues F740, Y746, Y763, Y766, and N768 are from (Fig. S5). Presumably, W102 in SMN will flip out to allow the the canonical Tudor domain, and T688 and N823 are from the SN methylarginine residue to reside in the aromatic cage. Hence, domain (Fig. 3 B and C). These interactions are conserved in both our complex structures of SND1 with PIWIL1 peptides provide R4me2s and R14me2s peptides. Therefore, both the Tudor and a general binding mechanism for methylarginine recognition. SN domains are involved in methylarginine binding. The intact Arginine methylation sites on the N termini of Piwi proteins extended Tudor domain is critical in binding methylated arginine are evolutionarily conserved. The RG/RA-rich clusters provide peptides and the methylarginine is bound through cation-π, docking sites for proteins with Tudor domains (8), resembling hydrophobic, van der Waals, and hydrogen bond interactions. the recognition of histone modifications by methyllysine-binding A series of mutagenesis experiments were performed to verify domains. Given that there are six arginine sites in the first RG/ the importance of these residues in SND1 for methylarginine RA-rich cluster on the PIWIL1/Miwi N terminus, multiple argi- binding. Mutating F740, Y746, Y763, or Y766 to leucine in SND1 nine residues may be methylated simultaneously. Our binding reduced the binding to the R4me2s peptide by 11–22-fold data showed that a peptide with three symmetrically dimethylated (Table S1). Mutating N768 to aspartic acid increased the binding arginines (R4, R10, and R14) only showed slightly increased af- affinity twofold, presumably by introducing extra electrostatic finity toward the extended Tudor domain of SND1, indicating interactions with the methylarginine; in contrast, mutating N768 that one methylarginine is adequate for binding a single extended to glutamine or arginine disrupted binding (Table S1). N768R in- Tudor, and that there is no obvious requirement for multivalent troduces an unfavorable positive charge and the N768Q mutation interactions. However, it is conceivable that multiple methylation possibly disrupts the hydrogen bond with the methylarginine. sites may increase the local concentration of available methylated These data show that the methylarginine binding pocket observed ligands and enhance the possibility of initial recruitment of in the crystal structures is essential in binding a methylated the extended Tudor domain. On the other hand, some germ-line arginine peptide. Tudor domain proteins contain multiple Tudor domains. For Our structures can also explain why the SND1 extended Tudor instance, TDRD6 has eight different Tudor domains, which po- domain preferentially binds symmetrically dimethylated arginine tentially could bind multiple distinct arginine methylation marks peptides. Monomethylation would reduce hydrophobic interac- at the same time (30). Regardless, our finding that the binding tions with SND1, and based on our SND1-R4me2a model surface of the SND1 extended Tudor domain can bind methylated (Fig. 3D), arginine asymmetrical dimethylation would disrupt peptides in opposite orientations suggests a surprising plasticity in the hydrogen bond between the NH1 group of the methylarginine its ability to accommodate arginine methylated sites. During the and N768. preparation of our manuscript, a significant study has shown the The extended Tudor domain has a similar aromatic cage to that crystal structure of extended Tudor domain of Drosophila Tudor observed in the 53BP1 and L3MBTL1 structures (20, 29). How- with arginine methylated peptides of Drosophila Aubergine (26).

4of6 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1013106107 Liu et al. Downloaded by guest on September 29, 2021 R4me2 A s C R14me2s N N C Rme2s

B C

D742 Y746 Y766 Y766 Y746 N823 N823 F740 F740 T688

R4 R14me2s me2s T688 N768 N768 Y763 Y763

SND1-R4me2s SND1-R14me2s

D E BIOCHEMISTRY Fig. 3. Methylarginine binding pockets in SND1. (A) H4R19 Superposition of SND1-R4me2s and SND1-R14me2s Y746 complexes. R4me2s peptide is colored in yellow, and R14me2s peptide is colored in aquamarine. Y766 Symmetrically dimethylated arginines from both Y1502 complexes are shown in stick models. (B) Methylargi- H4K20me2 nine binding pocket residues and R4me2s peptide are F740 R4me2a D1521 shown in stick models. (C) Methylarginine binding W1495 pocket residues and R14me2s peptide are shown in stick models. (D) Model of asymmetrical dimethy- Y1523 Y763 F1519 lated arginine bound by SND1. Note the absence N768 of a hydrogen bond between Rme2s and N768. (E) Methyllysine-binding pocket in the 53BP1- SND1-Rme2a model 53BP1-H4K20me2: PDB 2IG0 H4K20me2 complex.

As predicted (19), the extended Tudor domain of the Drosophila pET28-MHL vector to generate N-terminal His-tagged fusion Tudor protein displays similar domain folds and conserved aro- protein. The recombinant protein was overexpressed in Escheri- matic cage residues for binding sDMA to the ancestral SND1 chia coli BL21 (DE3)-V2R-pRARE2 at 16 °C and purified by polypeptide. Together, these findings underscore a previously affinity chromatography on Ni-nitrilotriacetate resin (Qiagen) underappreciated requirement of the flanking sequences of cano- followed by TEV protease treatment to remove the tag. For crys- nical Tudor core domain in assisting the binding of methylargi- tallization experiments, purified protein was further treated with nine-containing peptides, a feature that is likely conserved Endoproteinase Glu-C. All of them were further purified by across germ-line Tudor proteins (19). anion-exchange column. Finally, the protein was concentrated A role for SND1 in small RNA pathways was first suggested by to 15 mg∕mL in a buffer containing 20 mM Tris • HCl, pH the discovery of SND1 as a component of RNA induced silencing 7.5, 0.2 M NaCl, and 1 mM DTT. complex of the siRNA pathway. Subsequently, SND1 has been Mutant proteins were also overexpressed in Escherichia coli shown to bind and cleave inosine-containing hyperedited mi- BL21 (DE3)-V2R-pRARE2 and purified using the same proce- cro-RNAs through its SN domains (31). However, the binding dure as described above, except further purification using size- substrates for SND1 Tudor domain in these physiological contexts exclusion chromatography (Hi Load 16∕60 Superdex 75, GE are still unclear. In this study, we showed the coexpression and Healthcare Biosciences). The molecular weight of all protein interaction of PIWIL1/Miwi and SND1, suggesting a previously samples was checked by mass spectrometry. undescribed function for SND1 in regulating piRNA pathways. It will be of considerable interest to investigate whether SND1 may Binding Assays. All peptides used for fluorescence polarization and affect piRNA biogenesis, processing, or stability. A mouse knock- isothermal titration calorimetry (ITC) measurements were out model is necessary to further understand the biological role of synthesized by TuftsUniversity Core Services. Fluorescence polar- SND1 in these processes. ization binding assay was performed in 10 μL at a constant fluorescence labeled-peptide concentration of 40 nM and increasing Experimental Procedures amounts of SND1 (residues 650–910) at concentrations ranging Protein Expression and Purification. The extended Tudor domain of from low to high micromolar in a buffer of 20 mM Tris • HCl, human SND1 (residues 650–910) was cloned into a modified pH 7.5, 50 mM NaCl or 150 mM NaCl, 1 mM DTT, and 0.01%

Liu et al. PNAS Early Edition ∣ 5of6 Downloaded by guest on September 29, 2021 Tween-X-100. The assay was performed in 384-well plates, using a Structure Determination. Diffraction data collected at 100 °K using Synergy 2 microplate reader (BioTek). An excitation wavelength CuKα radiation generated on a Rigaku FR-E SuperBright rotat- of 485 nm and an emission wavelength of 528 nm were used. The ing anode system equipped with a Saturn A200 CCD detector data were corrected for background of the free-labeled peptides. were integrated and scaled using HKL2000 (33). All peptide- K To determine the D values, the data were fit to a hyperbolic bound SND1 structures were solved by molecular replacement function using Sigma Plot software (Systat Software, Inc.). ITC method as implemented by MOLREP in the CCP4 program suite measurements were performed as reported previously (32). (CCP4, 1994) using the apostructure of human SND1 extended Tissue Immunoprecipitation and Western Blot Analysis. For tissue Tudor domain (Protein Data Bank ID code 2O4X) as a search extract, mouse tissues and testes from different aged animals model (24). Following several alternate cycles of restrained re- were lysed in 150 mM NaCl/50 mM Tris • HCl, pH 7.2/1% finement and manual rebuilding using COOT (34), the improved Triton X-100/1% sodium deoxycholate/0.1% SDS using Polytron models revealed clear electron densities allowing placement of tissue homogenizer. Immunoprecipitation was performed as pre- the bound peptides. All refinement steps were performed using viously described (8). Briefly, adult testes were homogenized in REFMAC (CCP4, 1994) in the CCP4 program suite. During the NP40 lysis buffer (10 mM Tris • HCl, pH7.5, 150 mM NaCl, 0.5% final cycles of model building, TLS parameterization (35, 36) 10 μ ∕ 10 μ ∕ NP40, 1 mM PMSF, g mL leupeptin, g mL aprotinin, was included in the refinement of all models, which comprised 10 μ ∕ g mL pepstatin) using a Dounce homogenizer. After centri- protein and solvent molecules. Data collection and refinement fugation, the supernatant was filtered through a 0.45 μm filter, statistics are summarized in Table S2. precleared with Protein G Sepharose for 2 h, and incubated with 1 μg of antibodies overnight at 4 °C. The immunoprecipitates were recovered by incubation with 80 μL 10% Protein A slurry ACKNOWLEDGMENTS. We thank Aled Edwards and Cheryl Arrowsmith for critical reading of the manuscript and Rick Bagshaw for reagents. This for 3 h. After extensive washing with the lysis buffer and eluted research was supported by the Structural Genomics Consortium, a registered using 1% SDS sample buffer for Western blotting. Antibodies charity (number 1097737) that receives funds from the Canadian Institutes used for immunoprecipitation and Western blotting were Anti- for Health Research (CIHR), the Canadian Foundation for Innovation, SND1 (Abcam), Anti-SND1 (C-17, Santa Cruz), Anti-Hiwi Genome Canada through the Ontario Genomics Institute, GlaxoSmith- (Abcam), and Anti-Miwi (G82, Cell Signalling). Kline, Karolinska Institute, The Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innova- Protein Crystallization. Purified protein treated with V8 protease tion, Merck & Company, Inc., the Novartis Research Foundation, the Swedish was mixed with PIWIL1 peptides by directly adding a three- or Agency for Innovation Systems, the Swedish Foundation for Strategic fivefold molar excess of peptide to the protein solution. The crys- Research, and the Wellcome Trust. T.P. is supported by grants from the CIHR (MOP-6849), the Canadian Cancer Society Research Institute, and the Ontario tals of SND1 appeared in a condition containing 30% PEG 1500 Research Foundation. C.C. and J.J. are recipients of CIHR fellowships. K.L. is at 18 °C overnight by the sitting drop vapor diffusion method. supported by a Central China Normal University scholarship. Open access Prior to the data collection, the SND1 crystals were freshly publication costs were defrayed by the Ontario Genomics Institute Genomics soaked in the mother liquor containing 15% glycerol and flash- Publication Fund. frozen in liquid nitrogen.

1. Bedford MT, Richard S (2005) Arginine methylation an emerging regulator of protein 19. Jin J, et al. (2009) Eukaryotic protein domains as functional units of cellular evolution. function. Mol Cell 18:263–272. Sci Signal 2(98):1–18. 2. Bedford MT, Clarke SG (2009) Protein arginine methylation in mammals: Who, what, 20. Botuyan MV, et al. (2006) Structural basis for the methylation state-specific recogni- and why. Mol Cell 33:1–13. tion of histone H4-K20 by 53BP1 and Crb2 in DNA repair. Cell 127:1361–1373. 3. Malone CD, Hannon GJ (2009) Small RNAs as guardians of the genome. Cell 21. Huang Y, Fang J, Bedford MT, Zhang Y, Xu RM (2006) Recognition of histone H3 lysine- 136:656–668. 4 methylation by the double tudor domain of JMJD2A. Science 312:748–751. 4. Thomson T, Lin H (2009) The biogenesis and function of PIWI proteins and piRNAs: 22. Cote J, Richard S (2005) Tudor domains bind symmetrical dimethylated arginines. J Biol Progress and prospect. Annu Rev Cell Dev Biol 25:355–376. Chem 280:28476–28483. 5. Siomi MC, Mannen T, Siomi H (2010) How does the royal family of Tudor rule the PIWI- 23. Bateman A, et al. (2004) The Pfam protein families database. Nucleic Acids Res 32: interacting RNA pathway? Genes Dev 24:636–646. D138–141. 6. Wang J, Saxe JP, Tanaka T, Chuma S, Lin H (2009) Mili interacts with tudor domain- 24. Shaw N, et al. (2007) The multifunctional human p100 protein “hooks” methylated containing protein 1 in regulating spermatogenesis. Curr Biol 19:640–644. ligands. Nat Struct Mol Biol 14:779–784. 7. Kirino Y, et al. (2009) Arginine methylation of Piwi proteins catalysed by dPRMT5 is 25. Deng W, Lin H (2002) Miwi, a murine homolog of piwi, encodes a cytoplasmic protein required for Ago3 and Aub stability. Nat Cell Biol 11:652–658. essential for spermatogenesis. Dev Cell 2:819–830. 8. Chen C, et al. (2009) Mouse Piwi interactome identifies binding mechanism of Tdrkh 26. Liu H, et al. (2010) Structural basis for methylarginine-dependent recognition of Tudor domain to arginine methylated Miwi. Proc Natl Acad Sci USA 106:20336–20341. Aubergine by Tudor. Genes Dev 24:1876–1881. 9. Vagin VV, et al. (2009) Proteomic analysis of murine Piwi proteins reveals a role for 27. Alexandropoulos K, Cheng G, Baltimore D (1995) Proline-rich sequences that bind to arginine methylation in specifying interaction with Tudor family members. Genes Src homology 3 domains with individual specificities. Proc Natl Acad Sci USA Dev 23:1749–1762. 92:3110–3114. 10. Reuter M, et al. (2009) Loss of the Mili-interacting Tudor domain-containing protein-1 28. Lim WA, Fox RO, Richards FM (1994) Stability and peptide binding affinity of an activates transposons and alters the Mili-associated small RNA profile. Nat Struct Mol SH3 domain from the Caenorhabditis elegans signaling protein Sem-5. Protein Sci Biol 16:639–646. 3:1261–1266. 11. Vasileva A, Tiedau D, Firooznia A, Muller-Reichert T, Jessberger R (2009) Tdrd6 is 29. Min J, et al. (2007) L3MBTL1 recognition of mono- and dimethylated histones. Nat required for spermiogenesis, chromatoid body architecture, and regulation of miRNA Struct Mol Biol 14:1229–1230. expression. Curr Biol 19:630–639. 30. Jozic D, et al. (2005) Cbl promotes clustering of endocytic adaptor proteins. Nat Struct 12. Nishida KM, et al. (2009) Functional involvement of Tudor and dPRMT5 in the piRNA Mol Biol 12:972–979. processing pathway in Drosophila germ lines. EMBO J 28:3820–3831. 31. Li CL, Yang WZ, Chen YP, Yuan HS (2008) Structural and functional insights into human 13. Kirino Y, et al. (2010) Arginine methylation of Aubergine mediates Tudor binding and Tudor-SN, a key component linking RNA interference and editing. Nucleic Acids Res germ plasm localization. RNA 16:70–78. 36:3579–3589. 14. Anne J, Ollo R, Ephrussi A, Mechler BM (2007) Arginine methyltransferase Capsuleen 32. Guo Y, et al. (2009) Methylation-state-specific recognition of histones by the MBT is essential for methylation of spliceosomal Sm proteins and germ cell formation in repeat protein L3MBTL2. Nucleic Acids Res 37:2204–2210. Drosophila. Development 134:137–146. 33. Otwinowski Z, Minor W (1997) Processing of X-ray diffraction data collected in oscilla- 15. Gonsalvez GB, et al. (2007) Two distinct arginine methyltransferases are required for tion mode. Method Enzymol 276:307–326. biogenesis of Sm-class ribonucleoproteins. J Cell Biol 178:733–740. 34. Emsley P, Cowtan K (2004) Coot: Model-building tools for molecular graphics. Acta 16. Maurer-Stroh S, et al. (2003) The Tudor domain “Royal Family”: Tudor, plant Agenet, Crystallogr, Sect D: Biol Crystallogr 60:2126–2132. Chromo, PWWP and MBT domains. Trends Biochem Sci 28:69–74. 35. Winn MD, Isupov MN, Murshudov GN (2001) Use of TLS parameters to model 17. Kim J, et al. (2006) Tudor, MBT and chromo domains gauge the degree of lysine methy- anisotropic displacements in macromolecular refinement. Acta Crystallogr, Sect D: Biol lation. EMBO Rep 7:397–403. Crystallogr 57:122–133. 18. Adams-Cioaba MA, Min J (2009) Structure and function of histone methylation 36. Winn MD, Murshudov GN, Papiz MZ (2003) Macromolecular TLS refinement in binding proteins. Biochem Cell Biol 87:93–105. REFMAC at moderate resolutions. Method Enzymol 374:300–321.

6of6 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1013106107 Liu et al. Downloaded by guest on September 29, 2021