The Role of Trp79 in β‐Actin on Histidine Methyltransferase SETD3 Catalysis

Nτ‐methylation of His73 in actin by histidine methyltransferase SETD3 plays an important role in stabilising actin filaments in eukaryotes. Mutations in actin and overexpression of SETD3 have been related to human diseases, including cancer. Here, we investigated the importance of Trp79 in β‐actin on productive human SETD3 catalysis. Substitution of Trp79 in β‐actin peptides by its chemically diverse analogues reveals that the hydrophobic Trp79 binding pocket modulates the catalytic activity of SETD3, and that retaining a bulky and hydrophobic amino acid at position 79 is important for efficient His73 methylation by SETD3. Molecular dynamics simulations show that the Trp79 binding pocket of SETD3 is ideally shaped to accommodate large and hydrophobic Trp79, contributing to the favourable release of water molecules upon binding. Our results demonstrate that the distant Trp79 binding site plays an important role in efficient SETD3 catalysis, contributing to the identification of new SETD3 substrates and the development of chemical probes targeting the biomedically important SETD3.


Introduction
Microfilaments in the cytoskeleton of eukaryotic cells are composed of highly abundant actin proteins that enable the cells to carry out essential functions, including cell growth, division, contraction, and movement. [1]1b,4] The filaments represent a crucial part of the cellular architecture, and their length is controlled by the hydrolysis rate of the bound ATP, leading to depolymerisation of the actin filament. [5]Actin is subjected to various posttranslational modifications (PTMs), including acetylation, methylation and ubiquitylation, that are involved in a range of cell functions. [6]Among these, actin histidine 73 methylation is highly conserved and stabilises the actin filaments. [3]Methylation of β-actin (βA) was found to be associated with decreased ATP hydrolysis and increased polymerisation rate of actin monomers. [5,7]ETD3, a member of the Su(var)3-9, Enhancer of zeste and Trithorax (SET) domain family of enzymes, was recently identified as the actin specific N τ -methyltransferase that catalyses methylation of the imidazole ring of His73 in β-actin using cosubstrate S-adenosylmethionine (SAM) as the methyl donor, producing S-adenosylhomocysteine (SAH) as a product (Figure 1A). [3,8]7a, 9] Upon binding of β-actin, His73 is inserted into a deep and narrow catalytic pocket, where the nucleophilic N τ of imidazole is oriented in close proximity to the electrophilic methyl group of SAM required for an efficient methyl transfer reaction (Figure 1B). [9]7a,9] Furthermore, the aromatic ring of Tyr313 is involved in hydrophobic stacking contact with the side chain of His73, and the interaction with the hydroxyl group attains a partial neutralization of N τ of the imidazole ring of His73 (Figure 1B). [3,10]10a] Dysregulation of SETD3 in cells is associated with various human diseases, such as breast cancer and viral infections. [11]The installation of methyl group on His73 of βactin is an important regulatory mechanism, highlighting the need for βA-based substrates and inhibitors in the rational drug design targeting SETD3. [12]ET-domain-containing enzymes catalyse the methylation of diverse amino acids, including lysine, arginine and histidine. [13]10a] However, SETD3 was recently shown to methylate different histidine mimics in βA peptidomimetics. [12]Remarkably, βA peptidomimetics possessing methionine analogues at position 73 were shown to act as potent inhibitors of human SETD3. [14]In addition, the incorporation of Ile mimics at position 71 in the βA peptide demonstrated an important role of Ile71 in registering His73 at the catalytic pocket of SETD3. [15]Our objective for the current study is to investigate the role of distant Trp79 residue on the productive SETD3 histidine methyltransferase catalysis.

Substrate screening
To explore the role of Trp79 in βA on human SETD3 catalysis, we synthesised a panel of 12 βA peptides bearing tryptophan and its analogues at position 79 (βA, residues 66-88, sequence: TLKYPIEHGIVTNWDDMEKIWHH; Figure 2, Table S1 in the Supporting Information).These analogues include both natural and unnatural amino acids: i) d-tryptophan to probe stereochemistry, ii) smaller side chains to explore the influence of chain length and size (Gly, Ala, Ile), iii) five-and six-membered naturally occurring aromatic amino acids (His, Phe, Tyr), iv) steriically demanding and nonpolar side chains (NmeTrp, 1NapA and 2NapA), and v) βhTrp to explore the importance of the Trp backbone (Figure 2).All tryptophan analogues were incorporated at position 79 in βA by employing Fmoc-based solidphase peptide synthesis (SPPS).The synthetic βA peptides were purified by RP-HPLC and their purity was verified by MALDI-TOF MS and analytical RP-HPLC (Figures S1-S12).The His-tagged recombinant human SETD3 was expressed in Escherichia coli and purified by HisTrap FF column (Figure S13).
The biocatalytic potential of SETD3 towards our panel of synthetic βA-Trp*79 peptides was assessed by MALDI-TOF MS based enzymatic assays.To evaluate the substrate specificity of βA peptides, enzyme assays were carried out with 200 nM SETD3 in the presence of βA-Trp*79 peptides and SAM (10 μM βA peptide, 100 μM SAM) at 1 and 3 h at 37 °C, pH 9 (Figures 3  and S14).Control reactions in the absence of SETD3 were carried out in parallel to demonstrate the enzyme-dependent methyl transfer reactions on His73.The SETD3-catalysed reaction of βA-Trp79 yielded 81 % conversion to the methylated peptide after 3 h (Figure 3A).In contrast, the βA-DTrp79 peptide underwent 35 % methylation (Figure 3B), demonstrating the importance of l-stereochemistry of the tryptophan residue for efficient binding in the Trp79 pocket.βA peptides bearing small and/or hydrophobic side chains (Gly, Ala and Ile) were poorly methylated (9, 14 and 18 % after 3 h, respectively; Figure 3C-E), thus indicating that there is a need for significant bulkiness of the side chain to maintain the energetically favourable hydrophobic interactions in the Trp79 binding pocket.The βA peptides bearing more polar and larger residues such as His and Tyr yielded 18 % and 44 % methylated peptides after 3 h, respectively (Figure 3F, H).Substitution of the Trp79 residue by Phe showed comparable methylation to the βA-Trp79 peptide (73 % conversion; Figure 3G), again supporting the importance of bulkiness and hydrophobicity of the side chain at position 79 in the molecular recognition and productive enzyme catalysis.These results suggest that the secondary Trp79 pocket is mainly driven by van der Waals and hydrophobic interactions.Introducing a naphthalene ring flanking in the same direction as the benzene ring of the indole side chain of Trp (1NapA) was tolerated better than the ring flanking opposite (2NapA), yielding 75% and 50 % of methylated βA peptides after 3 h (Figure 3J, K).These results can be attributed to the similar positioning of the 1NapA ring system as for the indole of Trp, also defining limits of the side chain bulkiness and importance of orientation of the aromatic rings.The peptide possessing the N-methyl group on the indole ring (NmeTrp) showed efficient methylation (73 %), possibly replacing the hydrogen bond interaction with hydrophobic interactions instead (Figure 3I).Furthermore, βA-βhTrp79 underwent 57 % conversion to methylated peptide after 3 h, indicating some tolerance for extended backbone of the Trp residue (Figure 3L).Overall, these results demonstrate that the distant Trp79 residue is important to attain efficient His73 methylation by SETD3.
Human SETD3 was subsequently incubated in the presence of βA peptides and SAM under optimised conditions at higher enzyme concentration (1 μM SETD3, 10 μM βA peptide, 100 μM SAM) at 37 °C in reaction buffer, and the conversion to methylated peptides was measured at different time points (1 and 3 h) (Figures S15 and S16).SETD3 catalysed full methylation of βA-Trp79 already within 1 h, in agreement with previous finding (Figure S15 and S16A), [15] and the βA peptide containing the stereochemically inverted d-Trp within 3 h (Figure S15 and S16B).Enzymatic assays demonstrated that the substitution of Trp79 by Gly only yielded 52 % methylated peptide within 3 h under optimised conditions (Figure S15 and  S16C).In contrast, βA peptides possessing smaller residues such as Ala and Ile were well tolerated by SETD3, resulting in a higher degree of conversion (82% and 85 %) to methylated peptides after 3 h, respectively (Figures S15 and S16D, E). βA-His79 underwent 78 % methylation after 3 h (Figure S15 and S16F), whereas substitutions of βA-Trp79 by Phe, Tyr, NmeTrp, 1NapA, 2NapA and βhTrp resulted in complete conversion to the corresponding methylated peptides already after 1 h (Fig- ure S15 and S16G-L).These results demonstrate that βA peptides possessing simple tryptophan mimics underwent efficient SETD3-catalysed methylation at higher enzyme concentrations.

Enzyme kinetics
We then investigated the catalytic efficiencies of the βA peptides that were observed as good substrates at 200 nM SETD3 concentration.Enzyme kinetics analyses were carried out under steady-state conditions by incubation of saturating amounts of SAM (100 μM) and different concentrations of the βA peptides (Figures 4 and S17).By direct comparison of the catalytic efficiency (k cat /K m ) of SETD3 towards our panel of βA peptides, we observed that all βA peptides bearing the tryptophan analogues are worse SETD3 substrates compared to the βA-Trp79 peptide (Table 1).βA-d-Trp79, βA-Gly79 and βA-Ala79 were not evaluated since their turnover numbers appeared to be very low.βA-Ile79 was observed to be a poor substrate, where the substantial drop in the catalytic efficiency (22-fold) can be attributed to both decrease in k cat and significant increase in K m values (Figure 4, Table 1).The decreased catalytic efficiency (~3-fold) for βA-Tyr79 is due to increased K m value (Figure 4, Table 1), whereas βA-His79 dis-played nearly sevenfold drop, resulting from a decrease in k cat and increase in K m values.βA peptides bearing bulky residues Phe, 1NapA, 2NapA, βhTrp and NmeTrp showed slight loss of activity, which appeared again due to less favourable k cat and K m values.

Molecular dynamics simulations
To further study the molecular properties of the Trp79 binding to human SETD3, three 200 ns MD simulations were performed of SETD3 bound to βA-Trp79 and three of SETD3 bound to βA-Gly79 (Figure 5).In general, it was observed that Trp79 remained bound in the same conformation throughout most of the MD simulations, which is supported by the lower root mean squared deviation (RMSD) compared to Gly79 (Figure 5A-B and S19-S22).Interestingly, in the simulations of βA-Gly79, a noticeable increase in motion of the peptide backbone is observed for all the six terminal residues (Val76-Asp81, Figure S20).This confirms the importance of the Trp79 binding site as an anchoring point that stabilises the actin binding.Conversely, the most noticeable change in conformation in the βA-Trp79 simulations was observed after 174 ns in one of the simulations, where Trp79 temporarily rotates towards Met251 of SETD3.Except for this small movement, Trp79 was found to consistently remain in the same binding conformation inside the largely hydrophobic binding cavity, which indicates very favourable shape complementarity to the protein in maintaining favourable van der Waals interactions between nonpolar residues of SETD3 and βA-Trp79 (Figure S21).Based on the visual inspection of the SETD3-βA-SAH complex, it is evident that Trp79 engages in a hydrogen bond with a neighbouring water molecule, which forms additional hydrogen bonds to surrounding water and protein residues His323 and Thr77 (Figure 5C, Figure S22).Interestingly, the hydrogen bonding  between Trp79 and water remained very stable throughout most of the MD simulations, thus suggesting that positioning of water at this specific site is important for βA-Trp79 binding.These results support the superior binding of Trp79 over its analogues (as reflected in lower K m values, Table 1) due to crucial indole-NH hydrogen bonding interactions with water molecules inside the binding cavity of SETD3.
Because the Trp79 residue primarily engages in hydrophobic contacts inside the largely hydrophobic binding cavity of SETD3, it is expected that desolvation of this binding site might lead to a very favourable contribution to the free energy of binding.This hypothesis was investigated in an additional 50 ns restrained MD simulation of SETD3 without a bound βA, which was analysed using Grid inhomogeneous solvent theory (GIST) to estimate the thermodynamics of water molecules inside the cavity (Figures 5D and S23).As expected, solvation of the binding site is associated with a large entropic penalty, as water molecules inside the tight hydrophobic site have decreased rotational and translational entropy.This is to be expected for most sites because of the hydrophobic effect.The entropic penalty was estimated to be 8.5 kcal mol À 1 , which outweighed the favourable enthalpic contribution (À 5.0 kcal mol À 1 ) from formed between water molecules and the protein inside the binding site.The Trp79mediated release of high-energy water molecules that occupy the Trp79 binding pocket thus contributes to the total desolvation energy of approximately À 3.5 kcal mol À 1 .This result is in line with the observation that high-energy water molecules also occupy the SETD3's βA-Ile71 binding pocket [15] and hydrophobic pockets of biomedically important proteins. [16]

Conclusions
In conclusion, our enzymatic work on βA peptides containing Trp79 and its simplest analogues reveals that there is a need for significant bulkiness and a hydrophobic side chain in the molecular recognition of βA-Trp79 and efficient His73-methylation by SETD3.βA peptides with small side chains at position 79 appeared to be very poor substrates of SETD3.SETD3 was observed to efficiently catalyse the methylation of βA peptides with sterically demanding and hydrophobic amino acids at the distant position 79.Enzyme kinetics data showed that SETD3 displayed lower catalytic efficiency against βA substrates in our panel than did the natural sequence βA-Trp79; this makes the βA peptide with Trp79 the superior substrate for SETD3.Molecular dynamics simulations showed that the binding of Trp79 inside the binding pocket is very stable, and that therefore, the hydrophobic cavity is ideally shaped to accommodate the large and hydrophobic Trp79 residue.Overall, these results demonstrate that the association of Trp79 and its analogues with the distant "secondary" Trp79 binding pocket is driven by van der Waals interactions and hydrophobic interactions between SETD3 and Trp79-mimicking ligands, as manifested by energetically favourable desolvation of the ligand side chain and ligand-mediated release of high-energy water molecules that occupy the Trp79 binding pocket.Finally, these results provide important knowledge on the role of the distant Trp79 binding pocket on SETD3 catalysis, contributing to the identification of new SETD3 substrates and the development of chemical probes targeting the biomedically important SETD3.More generally, the work demonstrates the application of natural and unnatural amino acids in precise examinations of biologically important posttranslational modifications. [17]perimental Section

Synthetic procedures
All β-actin peptides (residues 66-88) were chain assembled on Rink amide resin (0.78 mmol g À 1 loading capacity) by Fmoc-based solidphase peptide synthesis (SPPS) chemistry using PurePep Chorus Peptide Synthesiser on a 0.05 mmol scale.Fmoc deprotection was achieved by a solution of 20 % (v/v) piperidine in DMF.Couplings were carried out by adding a mixture of the amino acid (3.0 equiv.),HATU (2.9 equiv.)and DIPEA (6.0 equiv.).General synthetic protocol included 10 min coupling at 75 °C, 4 min deprotection at 50 °C, except for His, which was double coupled.Unnatural amino acids NmeTrp, βhTrp and d-Trp were coupled manually using the same equivalents at room temperature overnight.1-NapA and 2-NapA were incorporated in βA by coupling for 3 h and 10 min deprotection at 75 °C using automated peptide synthesiser.After coupling of the last amino acid, the peptides were washed with methanol, dichloromethane and dried over diethyl ether.Peptides proceeded to cleavage from resin using 95 % TFA, 2.5 % TIPS and 2.5 % MQ for 4 h.The crude peptides were precipitated with cold diethyl ether (À 20 °C) and pelleted via centrifugation, lyophilised and purified by preparative HPLC.

SETD3 expression and purification
The recombinant human SETD3 fused to an N-terminal His 6 -tag was produced and purified by a slightly modified procedure described by Kwiatkowski et al. [3] SETD3 was overexpressed in E. coli BL21(DE3) overnight at 13 °C, 200 rpm, in the presence of 0.3 mM IPTG.The recombinant enzyme was then purified by employing HisTrap FF column (5 mL) and eluted in 20 mL of 50 mM HEPES pH 7.5, 400 mM NaCl, 10 mM KCl, 300 mM imidazole, 1 mM DTT.The elution buffer was exchanged by sequential dialysis of the enzyme preparation against 500 mL of dialysis buffer (20 mM Tris-HCl pH 7.5, 200 mM NaCl, 1 mM DTT and 6 % sucrose) overnight at 4 °C and then twice against 500 mL of the buffer for 3 h at room temperature.The enzyme purity was verified by SDS-PAGE (> 97 %), and the yield of recombinant enzyme was 3.4-6.2mg of protein per 950 mL of culture.The purified enzyme was aliquoted and stored at À 80 °C.

MALDI-TOF MS enzymatic assays
The conversion reactions of βA peptides were carried out with 10 μM of the βA peptide incubated with SAM (100 μM) and SETD3 (200 nM and 1 μM) in reaction buffer (25 mM Tris, 20 mM NaCl, pH 9.0) in a final volume of 50 μL.Reactions were incubated at 37 °C, shaken at 750 rpm and quenched with 10 % TFA in Milli-Q water at time points 1 and 3 h.The percentage of produced methylated peptide was quantified by calculating the integral with Flex Analysis software, considering all the ionic species.
The kinetics evaluation was carried out with dilution series of the βA peptides (150-3.125μM), incubated with SAM (100 μM) and SETD3 (150-700 nM) in reaction buffer (25 mM Tris, 20 mM NaCl, pH 9.0) in a final volume of 50 μL.Reaction mixtures were incubated for 30 min at 37 °C, shaken at 750 rpm, and quenched with 10 % TFA in Milli-Q water within the linear conversion of methylated peptides.Produced methylated peptides were analysed by MALDI-TOF-MS and quantified by calculating the integral, at any concentration point, with Flex Analysis software, considering all the ionic species.Kinetic parameters were obtained by fitting V 0 values and βA peptide concentrations to the Michaelis-Menten equation using GraphPad Prism software.Kinetic experiments were carried out in replicates (n = 2) and final values are reported as mean value � standard error (SE).

MALDI-TOF MS inhibition assays
The inhibition screening assays were performed with SETD3 (360 nM), SAM (100 μM) and βA peptide potential inhibitors (100 μM) with 20 min preincubation, followed by the addition of the βA-Trp79 substrate (10 μM), and incubated for additional 20 min in a final volume of 50 μL at 37 °C at pH 9. The reactions were quenched by the addition of 10 % TFA in Milli-Q water and mixed with CHCA to be analysed by MALDI-TOF-MS.SETD3 residual activity was determined by calculating the integral of methylated βA-Trp79 peptide and normalised to a control reaction in absence of potential inhibitory peptides.Experiments was performed in replicates (n = 2), error bars reported as SE.

Molecular dynamics simulations
The molecular dynamics simulations were setup following the protocol described previously. [14]In short, the SETD3 protein (PDB ID: 6ICV) was prepared in Maestro, [18] by determining protonation states and determining bond orders.The systems were set up using tleap, where the protein was solvated in TIP3P [19] water box with a 0.150 M NaCl concentration and a 12 Å buffer distance.Tleap was also used to introduce the W79G mutation.Simulations were performed using Amber18 [20] a 2 fs timestep, and with the previously determined GAFF [21] parameters for SAH, and the ff14SB force field parameters for the protein.The simulations of βA bound to SETD3 were performed after a short 1000 step minimization, by first heating the system to 300 K for 50 ps.The Berendsen barostat was then applied and an additional 550 ps simulation was performed to equilibrate the system, before the final 200 ns were simulated and the final 100 ns were used for analysis in three replicates.
An additional simulation series was performed to evaluate the water thermodynamics in SETD3 but without βA bound, as described elsewhere. [22]The system was first minimised while restraining water for 40000 steps, after which it was slowly heated to 50 K for 20 ps with pressure scaling using Langevin dynamics.The heating was then accelerated for an additional 20 ps until 300 K has reached.The system was then allowed to equilibrate for 10 ns while still scaling the temperature and pressure and restraining everything except water.Pressure scaling was then removed for 5 ns while restraints.Finally, while maintaining these settings, a 50 ns production run was performed.Results were analysed using the GIST tool in CPPTRAJ with a cubic grid-spacing of 0.5 Å and referencing the energies to those reported for TIP3P in the Amber manual.Hydration sites were iteratively assigned at sites with the highest water density within a 1.5 Å radius and so that no sites overlapped.Desolvation effect of Trp79 binding was calculated by summing grid-points within the van der Waals radius of Trp79 in the crystal structure.

Figure 2 .
Figure 2. A panel of amino acids including tryptophan analogues incorporated at position 79 in the βA peptide to explore the role of Trp79 on human SETD3 methyltransferase catalysis.

Figure 5 .
Figure 5. Molecular dynamics simulations of the SETD3-βA-SAH complex and thermodynamics of water inside the Trp79 pocket.A) Grey: representation of the average conformation of SETD3 from 200 ns MD simulation and blue: snapshot at 174 ns.B) The residue RMSD for the Trp79 conformations during the simulations.Transparent lines: individual simulations, solid lines: the average over all replicas.C) Highlighting the distance (purple dashed line) between the side chain NH of Trp79 and the oxygen of a neighbouring water.Yellow dashed lines: hydrogen bonds.D) Bar plot of the thermodynamic properties of water inside the Trp79 binding pocket of free, unbound SETD3.