Article

Oligomerization and Auto-methylation of the Human Lysine Methyltransferase SETD6

Lital Estrella Weil 1,2, Yulia Shmidov 3, Margarita Kublanovsky 1,2, David Morgenstern 5, Michal Feldman 1,2, Ronit Bitton 3,4 and Dan Levy 1,2

1 - The Shraga Segal Department of Microbiology, Immunology and Genetics, Ben-Gurion University of the Negev, P.O.B. 653, Be'er-Sheva 84105, Israel 2 - The National Institute for Biotechnology in the Negev, Ben-Gurion University of the Negev, P.O.B. 653, Be'er-Sheva 84105, Israel 3 - Department of Chemical Engineering, Ben-Gurion University of the Negev, P.O.B. 653, Be'er-Sheva 84105, Israel 4 - Ilse Katz Institute for Nanoscale Science and Technology, Ben-Gurion University of the Negev, P.O.B. 653, Be'er-Sheva 84105, Israel 5 - The Nancy and Stephen Grand Israel National Center for Personalized Medicine, Weizmann Institute of Science, Rehovot 7610001, Israel

Correspondence to Dan Levy: The Shraga Segal Department of Microbiology, Immunology and Genetics, Ben-Gurion University of the Negev, Be'er-Sheva 84105, Israel. [email protected] https://doi.org/10.1016/j.jmb.2018.08.028 Edited by Igor Stagljar

Abstract

Signaling via lysine methylation by lysine methyltransferases (PKMTs), has been linked to diverse biological and disease processes. The mono-methyltransferase SETD6 (SET-domain-containing protein 6) is a member of the PKMT family and was previously shown to regulate essential cellular processes such as the NF-κB, WNT and the oxidative stress pathways. However, on the biochemical level, little is known about the enzymatic mode of action of SETD6. Here we provide evidence that SETD6 forms high-molecular-weight structures. Specifically, we demonstrate that SETD6 monomeric, dimeric and trimeric forms are stabilized by the methyl donor, S-adenosyl-L-methionine. We then show that SETD6 has auto-methylation activity at K39 and K179, which serves as the major auto-methylation sites with a moderate auto-methylation activity toward K372. A point mutation at K179 but not at K39 and K372, located at the SET domain of SETD6, impaired SETD6 ability to form a trimer, strongly implying a link between the auto-methylation and the oligomerization state. Finally, by radioactive in vitro methylation experiments and biochemical kinetics analysis, we show that the auto-methylation at K39 and K179 increases the catalytic rate of SETD6. Collectively, our data support a model by which SETD6 auto-methylation and self-interaction positively regulate its enzymatic activity in vitro and may suggest that other PKMTs are regulated in the same manner. © 2018 Elsevier Ltd. All rights reserved.

Introduction methyl-donor for the methyl-transferase reaction, which gives rise to the methylated lysine and an Protein lysine methylation is a common and versatile S-adenosyl-L-homocysteine (SAH) product [4]. post-translational modification that has been shown to Self-modification and self-interaction are basic be involved in the regulation of many signaling mechanisms that are shared among many kinases. pathways [1]. Lysine methylation is catalyzed by These biochemical properties are important for the protein lysine methyltransferases (PKMTs). There enzymes stability, catalytic activity and their ability to are over 60 candidate members of this enzyme family, transduce down-stream signals [5–8]. While homo- the vast majority of which contain a conserved catalytic dimerization of PKMTs was not reported yet, there are SETdomainthatcatalyzestheadditionofmethyl several examples showing a link between auto- groups (mono, di or tri) to histone and non-histone methylation and PKMT enzymatic activity. The H3K9 [2,3]. S-adenosyl-L-methionine (SAM) is the PKMT G9a activity was shown to be regulated by auto-

0022-2836/© 2018 Elsevier Ltd. All rights reserved. J Mol Biol (2018) 430, 4359–4368 4360 SETD6 Oligomerization and Auto-methylation methylation at K239. The auto-methylation of G9a is an sulfosuccinimidyl-suberate (BS3), which covalently effective mediator for HP1 and G9a interaction similar crosslinks proteins 1.14 nm apart, via lysine residues to H3K9me3 binding [9]. A conformational switch [22]. Under untreated conditions, a single monomeric mediated by the auto-methylation of the yeast methyl- SETD6 band was observed. After 30 min of incubation, transferase Clr4 (Suv39h) at multiple lysines (K455 we could detect complexes of the expected dimers and and K472) was shown to enhance H3K9 methylation trimers sizes, reaching a saturation after 2 h. His- activity [10]. In addition, a recent study demonstrated SUMO was used as a negative control for the BS3 that the auto-methylation of SUV39H2 at K392 crosslinking assay (Fig. S1). reduces its enzymatic activity by blocking its interaction To examine SETD6 oligomerization state under with the substrates [11], whereas PRMT8 auto- native conditions, recombinant His-SETD6 was load- methylation reduces its interaction with SAM [8]. ed on a Superdex200 Increase size exclusion column While PKMTs are considered to be attractive thera- (SEC) (Fig. 1b). Consistent with the crosslinking peutic targets [12–14], the regulatory mechanisms of experiments, we observed higher amount of SETD6 their activity and specifically how their auto- monomers relative to the dimer–trimer forms (Fig. 1b, methylation regulates their activity are still unclear. upper panel), which was validated by Coomassie and The SET-domain-containing protein 6 (SETD6) is a Western blot analysis with anti-SETD6 antibody member of the PKMT family. SETD6 is located on the (Fig. 1b, bottom panel). To test whether constitutive long arm of 16 (16q21). The spans equilibrium between the monomeric, dimeric and 5049 bases with eight coding exons. Human SETD6 trimeric forms of SETD6 exists, the trimeric/dimeric has two splice variants: the long, 473-residue “isoform eluted fractions (fractions 13 and 14 from panel 1B), a” and the short, 449-residue “isoform b,” which lacks were loaded again on the SEC, resulting in a similar an in-frame segment (residues 40–63) [15].SETD6is distribution of the SETD6 native complexes (Fig. 1c). linked to the regulation of various major cellular This result suggests the existence of such equilibrium processes such as mono-methylation of the NF-κB between the different oligomerization states. Better subunit RelA at Lys 310, to suppress the activation of separation and enrichment of the dimer–trimer NF-κB target [16] and to promote RelA populations were obtained by loading crosslinked transcriptional activity in bladder cancer [17].In recombinant SETD6 on SEC (Fig. 1d). To test addition, SETD6 was shown to play a role in the SETD6 oligomerization state in cells, over-expressed NRF2 oxidative stress response by interacting with Flag-SETD6 was immunoprecipitated from HEK293T DJ1 [18], the WNT signaling pathway by methylation cells, eluted with Flag-peptides followed by separation of PAK4 [19], nuclear hormone receptor signaling [20] on native gel and Western analysis with anti-SETD6 and embryonic stem cell differentiation via methylation antibody (Fig. 1e). The result indicates that SETD6 of H2AZ [21]. forms dimeric and trimeric structures in a pattern and Despite the emerging importance of SETD6 in the ratio similar to recombinant His-SETD6. Taken togeth- regulation of diverse cellular processes, little is known er, our results suggest that SETD6 forms monomeric, about the biochemical properties which regulate its dimeric and trimeric structures in vitro andincells. activity. In this study, we show that SETD6 forms monomers, dimers and trimers and that these SAM stabilizes SETD6 monomeric, dimeric and structures are stabilized by SAM. We discovered trimeric structures that SETD6 is capable of auto-methylation and mapped the auto-methylation sites to K39, K179 and As SAM serves as the methyl donor during the K372. A point mutation at K179 impaired its ability to methylation reaction [1], we hypothesized that SAM form a trimer, implying a link between the auto- stabilizes the monomeric, dimeric and trimeric com- methylation and oligomerization properties. Finally, by plexes of SETD6. To address this hypothesis, biochemical kinetics analysis, we show that the auto- recombinant SETD6 was incubated in the absence methylation at K39 and K179 increases SETD6's or presence of SAM under methylation reaction affinity to its methylated substrate. conditions. Native gel electrophoresis, followed by Coomassie stain or Western blot analysis, was used to detect changes in the presence of SETD6 Results monomeric, dimeric, and trimeric as well as higher- molecular-weight structures (Fig. 2aandb).The addition of SAM resulted in stabilization of the SETD6 forms monomers, dimers and trimers as monomeric, dimeric and trimeric forms of SETD6 well as higher-molecular-weight structures and a significant reduction in the high-molecular- weight bands. However, in the absence of SAM, we In order to test if SETD6 forms strutures with could detect bands with significantly higher size molecular weight higher than monomers, we performed than monomers, dimers and trimers. Importantly, the a crosslinking experiment (Fig. 1a) with recombinant high-molecular-weight bands were not reduced after His-SETD6 in the presence of 68 μMbis- the addition of equivalent amount of SAH, which ED lgmrzto n Auto-methylation and Oligomerization SETD6

Fig. 1. SETD6 forms monomers, dimers and trimers as well as high-molecular-weight structures. (a) Recombinant His-SETD6 was treated with 68 μMBS3 for the indicated times followed by Coomassie stain. (b) SEC profile of His-SETD6. Eluted fractions (11–18) were loaded on SDS gel followed by Coomassie and Western blot analysis. Standard curve obtained with proteins of known masses is shown on the right. The position of the monomer, dimer and trimer on the curve is indicated with an arrow. (c) The dimer/trimer eluted fractions (13–14) were re-loaded again resulting in a similar distribution of the dimer/trimer versus monomer high-molecular-weight structures. (d) Recombinant His-SETD6 was treated with 68 μMBS3 for 30 min followed by SEC. (e) Flag-SETD6 was over-expressed in HEK293T followed by immunoprecipitation with Flag antibody. Flag-SETD6 was then eluted from the beads with Flag-peptide and loaded on native acrylamide gel followed by Western blot analysis using α-SETD6 antibody. Recombinant His-SETD6 serves as a positive control. 4361 4362 SETD6 Oligomerization and Auto-methylation

Fig. 2. SAM stabilizes SETD6 high-molecular-weight structures. (a) His-SETD6 was untreated or incubated with SAM or SAH for 2 h at 30 °C and was then loaded on native acrylamide gel followed by Coomassie stain. Coomassie stain under denaturation conditions is shown at the bottom. (b) Same as in panel a comparing SETD6 WT and SETD6 Y285A mutant, followed by Western blot analysis using α-His antibody. (c) Average protein size of His-SETD6 WT and SETD6 Y285A mutant in DDW solution with and without SAM, using DLS analysis.

suggests that this phenomenon is SAM-dependent population of aggregates with hydrodynamic diameter (Fig. 2a). To further illustrate the role of SAM in the range from 20 to 80 nm. This result is in agreement with enrichment of monomeric, dimeric and trimeric forms the results in Fig. 2a and b that shows existence of of SETD6, we used mutated SETD6 at Y285, which structures with various molecular weights (not distinct serves as the SAM binding site and leads to catalytic for DLS). Upon the addition of SAM to the SETD6 WT inactivation of the enzyme [16,23]. As shown in Fig. reaction, we obtained one population of small struc- 2b, the addition of SAM resulted in loss of the high- tures with hydrodynamic diameter of 4–10 nm, which molecular bands with SETD6 WT but not with the correlates with the monomer, dimer and trimer sizes. Y285A mutant. For SETD6 Y285A mutant with and without SAM, we To support SAM role in stabilizing the monomeric, obtained one population of very large aggregates with dimeric and trimeric structures of SETD6, we per- hydrodynamic diameter of 300–600 nm. This result formed dynamic light scattering (DLS) analysis com- indicates that addition of SAM to the SETD6 Y285A paring SETD6 WT to SETD6 Y285A mutant, after mutant had no effect on its aggregation state. Taken incubation at 30 °C for 2 h with and without SAM. As together, these results suggest that SAM stabilizes canbeseenfromFig. 2c, for SETD6 WT, there is one SETD6 monomeric, dimeric and trimeric structures. SETD6 Oligomerization and Auto-methylation 4363

K39 and K179 serve as the major K179, we hypothesized that these two biochemical auto-methylation sites of SETD6 characteristics are linked. To address this hypothe- sis, we performed native gel electrophoresis and SETD6 was shown before to have a strong auto- compared the oligomeric state between SETD6 methylation activity [16,19,24], however, the auto- WT and SETD6 mutants K39R, K179R and Y285A methylation site was not mapped yet. In order to (Fig. 4a) under methylation reaction conditions in the roughly map the auto-methylation site of SETD6, we presence of SAM. The results displayed a shift in the cloned and purified a truncated version of SETD6 oligomerization equilibrium of the K179R mutant which lacks the C′ terminus (Fig. 3a, upper panel). compared to SETD6 WT. Especially noticeable were The N′ truncation, containing the SET domain, was the increase in aggregation and decrease in the tested for its in vitro auto-methylation activity, alone trimer formation in K179R mutant, which suggest and in the presence of the full-length SETD6 protein that the oligomeric state and auto-methylation of (Fig. 3a). In contrast to the full-length protein, we SETD6 are linked. Indeed, a radioactive in vitro could not detect auto-methylation of the N′ trunca- methylation assay under native conditions (Fig. 4b) tion. However, when both proteins, WT and N′, were showed an increased auto-methylation signal in the incubated together, both were methylated. These WT enzyme in its trimer form in a time-dependent results suggest that the auto-methylation site is manner. Similar results were obtained when His- located within the N′-terminus fragment of SETD6. SETD6 was subjected to a radioactive methylation Furthermore, these findings also provide evidence assay in the absence or presence of the crosslinker for trans-auto-methylation, although we cannot BS3, displaying an enrichment of the trimeric form exclude the possibility of a cis-auto-methylation compared to the dimer (Fig. 4c). It is important to note activity as well. that the trimer and dimer bands intensity could not be To specifically map the auto-methylation site in compared to the monomer state and does not reflect SETD6, recombinant His-SETD6 was incubated the actual native equilibrium between the different under methylation reaction conditions with or without oligomeric states (non-crosslinked monomers versus SAM followed by mass spectrometry analysis dimers and trimers). This is due to a very low efficiency (Fig. 3b). The mass spectrometry identified two of the BS3 crosslinking reaction. potential mono-methylated sites: K39, which resides at the N′-terminus of SETD6, adjacent to the catalytic The auto-methylation at K39 and K179 increases SET domain, and K372 located toward the C′-terminus SETD6 enzymatic activity of the protein within the Rubisco domain, known to be important for protein-protein interaction [15] (Fig. 3b). The link between the oligomeric state and the auto- Using the Swiss-PDB-Viewer (sPDBv; 3QXY) of methylation of SETD6 at K39 and K179 raised the SETD6, we identified two exposed lysine residues: hypothesis that SETD6 auto-methylation affects K39, which was identified in the mass spectrometry SETD6 enzymatic activity. To address this hypothesis, experiment, and K179, which resides at the catalytic we performed an in vitro methylation assay and tested SET domain (Fig. 3c). Based on our rough mapping the ability of SETD6 WT and a double mutant of SETD6 experiments of the auto-methylation position to the at K39 and K179 (K39R/K179R) to methylate RelA or N′-terminus of SETD6 (Fig. 3a) together with the PAK4 (Fig. 5a), substrates we have previously shown mass spectrometry results (Fig. 3b), we hypothe- to be methylated by SETD6 [16,19,25]. As expected, a sized that K39, K179 and K372 serve as the auto- methylation of RelA and PAK4 was observed with methylation sites. To address this hypothesis, SETD6 WT. Interestingly, the methylation of both purified mutants were subjected to a radioactive substrates was dramatically reduced by the K39R/ in vitro methylation assay. We found only a K179R mutant (Fig. 5a). The differences in the moderate decrease in the auto-methylation of the enzymatic activity of WT and K39R/K179R mutant K372R mutant. In contrast, the auto-methylation were also measured by the MTase-Glo™ methyltrans- activity was dramatically decreased in the K39R ferase assay (Fig. S2A), which allowed us to acquire mutant and almost completely disappeared in the Michaelis–Menten parameters of the methylation K179R mutant compared to SETD6 WT (Fig. 3d). reaction (Figs. 5b and S2). In these experiments, we These results suggest that K39, K179 and K327 are used constant amounts of the enzymes with increasing important for the auto-methylation activity of SETD6 concentrations of RelA at two reaction times (10 and 15 and that K39 and K179 serve as the major auto- min, Fig. S2B and C), since in these conditions we methylation sites. observed a linear increase in the amount of the utilized SAH (the methylation reaction byproduct) (data not Auto-methylation at K179 is required for SETD6 shown). Significant difference in the KM of WT SETD6 trimeric state comparing to the K39R/K179R mutant was observed (Figs. 5b and S2D). These results suggest that the Having demonstrated that SETD6 exhibits self- auto-methylation of SETD6 at positions K39 and K179 dimerization and auto-methylation mainly at K39 and enhances the affinity of SETD6 towards RelA (Fig. 5a, 4364 ED lgmrzto n Auto-methylation and Oligomerization SETD6

Fig. 3. K39 and K179 serve as the major auto-methylation sites of SETD6. (a) In vitro methylation assay for His-SETD6 WT and N′ truncation using radioactive labeled SAM and Coomassie stain as loading control. Scheme of SETD6 full-length and N′ truncation is illustrated above. (b) MS/MS spectra showing mono-methylation of SETD6 at K179 and K372. (c) Swiss-PDB-Viewer (sPDBv; 3QXY) snapshot of SETD6 showing the two exposed lysine residues at K39 and K179 along with K372 and the catalytic in-active mutant Y285. (d) In vitro methylation assays in the presence of 3H-SAM and recombinant His-SETD6 (WT, K39R, K179R and K372R mutants). The methylated proteins were detected by autoradiogram. Coomassie stain of the recombinant proteins used in the reactions is shown on the bottom. Schematic diagram illustrating SETD6 three potential methylation sites identified using mass spectrometry and sPDBv analysis. SETD6 Oligomerization and Auto-methylation 4365

Fig. 4. Auto-methylation at K179 is linked to SETD6 oligomeric stability. (a) Native gel electrophoresis of His-SETD6 WT and mutants: K39R, K179R and Y285A followed by Western blot analysis using a-His antibody. Western blot analysis under denaturation conditions is shown at the bottom. (b) In vitro methylation assays in the presence of 3H-SAM and recombinant His-SETD6 WT for the indicated time points. Reactions were separated on native and SDS-PAGE gels electrophoresis followed by exposure to autoradiogram. Coomassie stain under denaturation conditions is shown at the bottom. (c) Recombinant His-SETD6 was untreated or treated with 68 μMBS3 for 2 h at 30°C and separated on SDS-PAGE followed by an in vitro methylation assay in the presence of 3H-labeled SAM. Membrane was exposed to autoradiogram. Coomassie stain of the recombinant proteins used in the reactions is shown at the bottom. and b). Taken together, our data indicate that SETD6 on histones and non-histone proteins [26,27].In oligomeric state and the auto-methylation at K39 and contrast to kinases, the biochemical mechanisms of K179 are tightly linked and positively regulate SETD6 PKMTs activity and specifically how their auto- catalytic activity in vitro (Fig. 5c). methylation properties regulate their activity are still unclear. Here we provide evidence that SETD6 oligomeric state and auto-methylation capability are Discussion critical for SETD6 enzymatic activity in vitro. Our data support a model by which SAM stabilizes PKMTs are key players in the regulation of many the mono-, di- and tri-oligomeric states of SETD6. signaling pathways via methylation of lysine residues SAM is the universal methyl donor for lysine,

Fig. 5. The auto-methylation at K39 and K179 increases SETD6 enzymatic activity. (a) In vitro methylation assay in the presence of 3H-labeled SAM with recombinant His-SETD6 WT compared to K39R/K179R mutant and GST-RelA or His-Sumo PAK4 as the substrate. Coomassie stain of the recombinant proteins used in the reactions is shown at the bottom. (b) KM values of His-SETD6 WT and K39R/K179R mutant toward GST-RelA achieved using the Methyltransferase-Glo™ assay. (c) Suggested model for how SETD6 auto-methylation and oligomerization properties regulate its catalytic activity. 4366 SETD6 Oligomerization and Auto-methylation arginine and DNA methylation and it is formed by the between SETD6 oligomeric along with K39 and condensation of the amino acid methionine and ATP K179 auto-methylation states. Taken together, our [28]. ATP is a molecular unit of currency of findings add a new dimension to our understanding intracellular energy for many physiological reactions. of SETD6 enzymatic properties and open a new It is generated in different cellular metabolic pro- avenue of research both in the biochemical and the cesses such as glycolysis, citric acid (Krebs) cycle functional cellular level and may suggest that other and oxidative phosphorylation. This may imply that PKMTs are regulated in a similar way. there is a correlation between the cellular metabolic state and the enzymatic activity of the different methyltransferases and specifically the catalytic Materials and Methods activity of SETD6. The cellular concentration of SAM and SETD6 access to SAM may dictate the oligomerization and auto-methylation states of Plasmids SETD6, and hence SETD6 cellular function and specifically its ability to interact and methylate its The human SETD6 isoform b WT and Y285A substrates. Nevertheless, it still remains to be mutant were subcloned into pET-Duet containing explored if these biochemical properties that were 6xHis-tag. pGEX-6P-1 RelA containing a GST-tag all measured in vitro occur under physiological was described in [16]. pETS-SUMO PAK4 contain- settings in a cellular context. ing a His-tag was described in Ref. [18]. SETD6 While our results demonstrate that SETD6 can mutants K39R, K179R and K39R/K179R were form dimeric and trimeric oligomers, our attempts to generated using specific primers for site-directed map the exact residues that mediate the different mutagenesis followed by DNA sequencing for oligomeric states were not successful. We have confirmation and cloned into pET-Duet containing ′ recently determined the structure of SETD6 in 6xHis-tag. SETD6-N was amplified and cloned into complex with a RelA peptide in the presence of pET-Duet plasmid containing 6xHis-tag as de- SAM at a resolution of 2.2Å [15]. However, the scribed in Ref. [25]. pcDNA Flag-SETD6 for over- information is not sufficient to map the interaction expression in HEK-293T cells was previously site. We hope that future structural biology ap- described by Chen et al. [18]. proaches such as crystallography, NMR and the usage of electronic microscopy to view and model Recombinant protein expression and purification large complexes of WT SETD6 and the mutants will allow us to further explore what are the structural pET-Duet containing SETD6 with 6xHis-tag was consequences of SETD6 oligomeric and auto- transformed to Escherichia coli, Rosetta. The bac- methylation states. Such information will enable us teria were grown in LB medium at 18 °C overnight, to uncouple these two phenomena to better under- after IPTG induction. Cell pellet was isolated from stand the contribution of each one of them. Such medium using centrifugation and lysed by sonication approaches may also provide structural evidence to on ice for 1.5 min in total, 10/5 on/off, 25% amplitude. distinguish between SETD6 cis- and trans-auto- Finally, the lysate was centrifuged (20 min, 4 °C, methylation. 18,000 rpm) and filtered, and the His-tagged proteins Lysine methylation signaling pathways and their were purified using an on a HisTrap column (GE) catalyzing enzymes are considered to be attractive with the ÄKTA gel filtration system. Protein were therapeutic targets. SETD6 auto-methylation and eluted by 0.5M Imidazole followed by dialysis to 10% oligomerization properties described here, which glycerol in PBS buffer. regulate SETD6 enzymatic activity, raise the possi- bility that specific inhibitors may block this mode of Size exclusion chromatography activation. In a recent paper, we designed a short peptide based on the sequence of RelA fused to a Purified His-SETD6 protein sample was loaded on cell penetrating peptide and showed a direct a SuperdexTM200 Increase 10/300 size exclusion interaction and inhibition of SETD6 enzymatic column (GE) in 0.5 ml/min flow rate at 4 °C, with the activity in vitro and in cells [29]. Further investigation eluent 250 mM NaCl2 in PBS buffer. is required to fully understand the biological and physiological roles of SETD6 auto-methylation and Crosslinking assay oligomeric conditions in normal and disease states. Such understanding will allow us to design-specific Crosslinking experiments were performed using inhibitors which are based on the newly discovered BS3 (bis(sulfosuccinimidyl)suberate) crosslinker SETD6 biochemical properties described here. (Thermo-Fisher). Twenty micrograms of the purified These inhibitors have the potential to be more His-SETD6 was incubated in a 25 μl reaction in the selective compared to other PKMT competitive presences 68 μMBS3 at room temperature for inhibitors because of the unique functional linkage varying durations. The reaction was then terminated SETD6 Oligomerization and Auto-methylation 4367

with protein sample buffer [250 mM Tris–HCl (pH 6.8), 1 mM MgCl2] at 30 °C. Samples were equilibrated for 10% SDS, 30% glycerol, 5% β-mercaptoethanol, 1 min at 4 °C prior to data collection. Correlograms bromophenol blue] heated at 95 °C for 5 min followed were collected at 173° for at least 10 runs of 10 s at by SDS-PAGE and Coomassie staining (Expedeon, 4 °C. The recorded correlograms were analyzed with InstantBlueTM) or followed by in vitro methylation the CONTIN procedure using the software provided assay described below. with the instrument.

Antibodies and Western blot analysis Mass spectrometry

Primary antibodies used were as follows: rabbit Samples of recombinant SETD6 were incubated polyclonal anti-SETD6 [16] and mouse monoclonal overnight with and without S-adenosyl-methionine in anti-His (Thermo Fisher, Cat. No. MA1-21315). HRP- PKMT buffer at 30 °C. Samples were then digested conjugated secondary antibodies used were as with trypsin and run on a Q-Exactive LC-MS/MS follows: goat anti-rabbit and goat anti-mouse (Jackson instrument in data-dependent mode (DDA). Data ImmunoResearch, 111-035-144 and 115-035-062, were searched initially using Preview (Protein Metrics respectively). The antibodies were diluted in 10% Inc) to determine search parameters, followed by a skim milk in PBST buffer, according to the manufac- thorough database search against the human prote- turer's recommendation. For Western blot analysis, ome appended with common contaminants using protein samples were heated at 95 °C for 5 min in FDR of b1%. denaturing or native sample buffer and separated by SDS-PAGE or native gel electrophoresis, respectively. In vitro methylation assay

Native gel electrophoresis The in vitro methylation assay reaction tube contained the recombinant protein, 2 mCi of 3H- The protein samples were mixed with native labeled S-adenosyl-methionine (AdoMet; PerkinElmer) sample buffer [250 mM Tris–HCl (pH 6.8), 30% and PKMT buffer in total volume of 25 μl. After glycerol, bromophenol blue], followed by separation incubation at 30 °C in varying times, the reactions on native 6% polyacrylamide gel electrophoresis at were resolved by SDS-PAGE or loaded on native gel 150 V for 60 min on ice. electrophoresis followed by autoradiography and/or Coomassie staining (Expedeon, InstantBlueTM). Cell line, transfection and immunoprecipitation Bioluminescent methyltransferase assay Human embryonic kidney cells (HEK-293T) were maintained in Dulbecco's modified Eagle's medium Methyltransferase assay for enzyme activity using (Sigma, D5671) with 10% fetal bovine serum the Methyltransferase-Glo™ assay was performed (Gibco), 2 mg/ml L-glutamine (Sigma G7513), according to the manufacturer's protocol. penicillin–streptomycin (Sigma, P0781) and non- Supplementary data to this article can be found essential amino acids (Sigma, M7145) were cultured online at https://doi.org/10.1016/j.jmb.2018.08.028. in humidified incubator with 5% CO2 at 37 °C. Cells were plated in 10-cm plates and transfected using Mirus transfection reagent (TransIT®-LT1) accord- ing to the manufacturer's instructions. Cells were lysed in RIPA lysis buffer [50 mM Tris–HCl (pH 8), Acknowledgments 150 mM, NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS (v/v), 1 mM dithiothreitol (DTT)] and Sigma We thank the Levy laboratory for technical assis- protease inhibitor cocktail (P8340, diluted 1:100). tance. We gratefully acknowledge Neta Sal-man and Cells lysate was incubated with FLAG-M2-affinity gel Niv Papo for their support and helpful discussions. beads (A2220; Sigma) for 1 h at 4 °C. After incubation, This work was supported by grants to D.L. from The beads were washed 4 times with lysis buffer and Israel Science Foundation (285/14), The Research incubated with 80 μg flag peptides (Sigma, F3290- Career Development Award from the Israel Cancer 4MG) for 1 h at 4 °C. Research Fund, the Israel Cancer Association and the National Institute for Biotechnology in the Negev. DLS Author Contributions: L.W., Y.S., M.K., M.F., R. B. and D.L. conceived and designed the exper- DLS measurements were performed using a iments. LW. performed the majority of the experi- Zetasizer Nano-ZS (Malvern, UK). Samples of ments. D.M. performed and analyzed the mass recombinant SETD6 were incubated overnight with spectrometry experiments. L.W. and D.L. wrote the and without S-adenosyl-methionine in PKMT buffer paper. All authors read and approved the final [10 mM Tris–HCl (pH 8), 2% glycerol, 0.8 mM KCl, manuscript. 4368 SETD6 Oligomerization and Auto-methylation

Conflict of Interest: The authors declare no [13] Z. Wang, D.J. Patel, Small molecule epigenetic inhibitors competing financial interests. targeted to histone lysine methyltransferases and demethy- lases, Q. Rev. Biophys. 46 (2013) 349–373. [14] F. Liu, X. Chen, A. Allali-Hassani, A.M. Quinn, T.J. Wigle, G.A. Received 22 March 2018; Wasney, et al., Protein lysine methyltransferase G9a inhibitors: Received in revised form 31 July 2018; design, synthesis, and structure activity relationships of 2,4- Accepted 23 August 2018 diamino-7-aminoalkoxy-quinazolines, J. Med. Chem. 53 Available online 4 September 2018 (2010) 5844–5857. [15] Y. Chang, D. Levy, J.R. Horton, J. Peng, X. Zhang, O. Keywords: Gozani, et al., Structural basis of SETD6-mediated regulation SETD6; of the NF-kB network via methyl-lysine signaling, Nucleic lysine methylation; Acids Res. 39 (2011) 6380–6389. auto-methylation [16] D. Levy, A.J. Kuo, Y. Chang, U. Schaefer, C. Kitson, P. Cheung, et al., Lysine methylation of the NF-kappaB subunit RelA by SETD6 couples activity of the histone methyltrans- Abbreviations used: ferase GLP at chromatin to tonic repression of NF-kappaB PKMTs, protein lysine methyltransferases; SETD6, SET- signaling, Nat. Immunol. 12 (2011) 29–36. domain-containing protein 6; S-adenosyl-L-methionine, [17] N. Mukherjee, E. Cardenas, R. Bedolla, R. Ghosh, SETD6 SAM; SAH, S-adenosyl-L-homocysteine; SEC, size regulates NF-kappaB signaling in urothelial cell survival: exclusion column; DLS, dynamic light scattering. implications for bladder cancer, Oncotarget 8 (2017) 15114–15125. [18] A. Chen, M. Feldman, Z. Vershinin, D. Levy, SETD6 is a References negative regulator of oxidative stress response, Biochim. Biophys. Acta 1859 (2016) 420–427. [1] J. Murn, Y. Shi, The winding path of protein methylation [19] Z. Vershinin, M. Feldman, A. Chen, D. Levy, PAK4 research: milestones and new frontiers, Nat. Rev. Mol. Cell methylation by SETD6 promotes the activation of the Wnt/ Biol. 18 (2017) 517–527. beta-catenin pathway, J. Biol. Chem. 25 291 (13) (2016) [2] C.H. Arrowsmith, C. Bountra, P.V. Fish, K. Lee, M. Schapira, 6786–6795 (PMID: 26841865). Epigenetic protein families: a new frontier for drug discovery, [20] D.J. O'Neill, S.C. Williamson, D. Alkharaif, I.C. Monteiro, M. Nat. Rev. Drug Discov. 11 (2012) 384–400. Goudreault, L. Gaughan, et al., SETD6 controls the [3] T. Kouzarides, Chromatin modifications and their function, expression of estrogen-responsive genes and proliferation Cell 128 (2007) 693–705. of breast carcinoma cells, Epigenetics 9 (2014) 942–950. [4] S.J. Mentch, J.W. Locasale, One-carbon metabolism and [21] O. Binda, A. Sevilla, G. LeRoy, I.R. Lemischka, B.A. Garcia, epigenetics: understanding the specificity, Ann. N. Y. Acad. S. Richard, SETD6 monomethylates H2AZ on lysine 7 and is Sci. 1363 (2016) 91–98. required for the maintenance of embryonic stem cell self- [5] J. Beenstock, N. Mooshayef, D. Engelberg, How do protein renewal, Epigenetics 8 (2013) 177–183. kinases take a selfie (autophosphorylate)? Trends Biochem. [22] J.C. Kermode, A.W. DeLuca, A. Zilberman, J. Valliere, S.M. Sci. 41 (2016) 938–953. Shreeve, Evidence for the formation of a functional complex [6] P. Geng, Y. Zhang, X. Liu, N. Zhang, Y. Liu, X. Liu, et al., between vasoactive intestinal peptide, its receptor, and Gs in Automethylation of protein arginine methyltransferase 7 and lung membranes, J. Biol. Chem. 267 (1992) 3382–3388. its impact on breast cancer progression, FASEB J. 31 (2017) [23] S.C. Dillon, X. Zhang, R.C. Trievel, X. Cheng, The SET- 2287–2300. domain protein superfamily: protein lysine methyltransfer- [7] X. Koh-Stenta, A. Poulsen, R. Li, J.L. Wee, P.Z. Kwek, S.Y. ases, Genome Biol. 6 (2005) 227. Chew, et al., Discovery and characterisation of the automethy- [24] D. Levy, C.L. Liu, Z. Yang, A.M. Newman, A.A. Alizadeh, P.J. lation properties of PRDM9, Biochem. J. 474 (2017) 971–982. Utz, et al., A proteomic approach for the identification of novel [8] M.B. Dillon, H.L. Rust, P.R. Thompson, K.A. Mowen, Auto- lysine methyltransferase substrates, Epigenetics Chromatin 4 methylation of protein arginine methyltransferase 8 (PRMT8) (2011) 19. regulates activity by impeding S-adenosylmethionine sensitivity, [25] L. Martin-Morales, M. Feldman, Z. Vershinin, P. Garre, T. J. Biol. Chem. 288 (2013) 27872–27880. Caldes, D. Levy, SETD6 dominant negative mutation in [9] H.G. Chin, P.O. Esteve, M. Pradhan, J. Benner, D. Patnaik, familial colorectal cancer type X, Hum. Mol. Genet. 26 (2017) M.F. Carey, et al., Automethylation of G9a and its implication 4481–4493. in wider substrate specificity and HP1 binding, Nucleic Acids [26] R. Hamamoto, V. Saloura, Y. Nakamura, Critical roles of non- Res. 35 (2007) 7313–7323. histone protein lysine methylation in human tumorigenesis, [10] N. Iglesias, M.A. Currie, G. Jih, J.A. Paulo, N. Siuti, M. Kalocsay, Nat. Rev. Cancer 15 (2015) 110–124. et al., Automethylation-induced conformational switch in Clr4 [27] E.L. Greer, Y. Shi, Histone methylation: a dynamic mark in (Suv39h) maintains epigenetic stability, Nature 560 (7719) health, disease and inheritance, Nat. Rev. Genet. 13 (2012) (2018) 504–508 (PMID: 30051891). 343–357. [11] L. Piao, M. Nakakido, T. Suzuki, N. Dohmae, Y. Nakamura, [28] W.A. Loenen, S-adenosylmethionine: jack of all trades and R. Hamamoto, Automethylation of SUV39H2, an oncogenic master of everything? Biochem. Soc. Trans. 34 (2006) 330–333. histone lysine methyltransferase, regulates its binding affinity [29] M. Feldman, D. Levy, Peptide inhibition of the SETD6 to substrate proteins, Oncotarget 7 (2016) 22846–22856. methyltransferase catalytic activity, Oncotarget 9 (2018) 4875. [12] L. Morera, M. Lubbert, M. Jung, Targeting histone methyl- transferases and demethylases in clinical trials for cancer therapy, Clin. Epigenetics 8 (2016) 57.