Actin-specific histidine N-methyltransferase activity was assayed by measuring the incorporation of the [³H]methyl group of [³H]S-adenosyl methionine ([³H]SAM) into either homogeneous recombinant human β-actin produced in Escherichia coli or a H73A β-actin mutant (Figure 1—figure supplement 1), which served as a negative control, ensuring that only the H73-specific N-methyltransferase activity was measured (Figure 1—figure supplement 2). The use of recombinant human β-actin as the rat methyltransferase substrate was reasonable because the sequences of human and rat β-actin cannot be distinguished from one another, owing to an extreme evolutionary conservation. Following numerous purification steps, the N-methyltransferase activity was purified from rat leg muscle approximately 1200-fold (see Table 1). The methyltransferase was eluted as a single peak in each of the purification steps (Figure 1), indicating the presence of a single enzyme species. No methylation of H73A β-actin was detected at any stage of the purification, suggesting that the myofibrillar extract of rat muscle is devoid of other β-actin-methylating enzymes. The overall yield of the purification was only about 0.6% (see Table 1), largely because of the fact that in the last two steps of purification, only the most active fractions were used for the next step of the procedure.

Figure 1 with 3 supplements see all

Download asset Open asset

SDS-PAGE analysis of the activity peak fractions (F2 and F3) from the Reactive Red 120-agarose showed a protein concentration that was too low to allow the identification of the proteins coeluting with the enzymatic activity (Figure 1D). Thus, the proteins in these fractions were concentrated about 15-fold and reanalyzed by SDS-PAGE. Actin-methylating activity coeluted with three protein bands of about 65, 75 and 90 kDa, which were clearly visible in the concentrated fraction F3 (Figure 1D). These bands were cut out from the gel, digested with trypsin and analyzed by MS/MS. The sequences of the identified peptides were compared with the rat reference proteome from the NCBI Protein database. This comparison indicated that there was only one methyltransferase among these peptides, present in band ‘C’, and this was identified as histone lysine N-methyltransferase SETD3 (Table 2). Eighteen peptide sequences from the rat SETD3 (underlined in Figure 2) that had shared identity with other SETD3 proteins in the National Center for Biotechnology Information (NCBI) Protein database were found to cover about 39% of its sequence. However, to exclude the possibility of missing any potential methyltransferase as the result of a poor extraction of tryptic peptides from protein bands, we performed the MS/MS identification of all proteins present in the entire fraction 14 of Superdex 200 (see Figure 1B) and fraction F3 of Reactive red 120 (see Figure 1C), and SETD3 protein was again the only methyltransferase found in these fractions.

Figure 2

Download asset Open asset

As SETD3 protein was shown to methylate histones (Eom et al., 2011) and all of the SET-domain-containing enzymes characterized to date are protein lysine methyltransferases (Clarke, 2013), we were surprised to find that SETD3 was the protein histidine N-methyltransferase responsible for the measured activity. To double check our results, we performed another complete, though modified, round of purification of the actin-methylating enzyme from rat leg muscles, as well as MS/MS analysis of the most purified fractions from Superdex 200. Once again, we identified SETD3 protein as the only meaningful candidate for rat actin-specific histidine N-methyltransferase, suggesting we had identified the correct protein. Moreover, our results were in agreement with previous reports suggesting that the actin-specific histidine N-methyltransferase corresponded to a ≈70 kDa protein (Raghavan et al., 1992).

Protein BLAST searches (Altschul et al., 1997) with rat SETD3 and phylogenetic analysis of the resulting sequences indicated that orthologs of this protein were found in almost all eukaryotes including vertebrates (≈75–98% identity with the rat sequence), insects (≈35–40% identity), plants and fungi (≈30% identity) (Figure 3). No SETD3 orthologs were detected in the sequenced proteome of Saccharomyces cerevisiae or of Naegleria gruberi in agreement with previous reports, showing the absence of actin-specific histidine methyltransferase activity in these species (Kalhor et al., 1999). All taxa followed the expected lines of descent, indicating that the enzyme was present in a common eukaryotic ancestor (see Figure 3). A low similarity in amino-acid sequence between eukaryotic SETD3 orthologs and some bacterial proteins was also detected (e.g. ≈30% identity, Cystobacter fuscus, NCBI Reference Sequence: WP_095987699.1), suggesting that the eukaryotic enzyme might have been acquired from an ancestral prokaryote.

Figure 3

Download asset Open asset

All identified SETD3 proteins contain the SET and Rubisco LSMT substrate-binding domains at their N- and C-terminus, respectively. The SET domain is thought to interact directly with the protein substrate and most probably contributes indirectly to SAM binding too, whereas the Rubisco LSMT substrate-binding domain seems to interact with a protein substrate only (see Figures 2 and 4). Although the sequence identity between orthologs from distinct species such as rat and maize (Zea mays) was rather low (about 30%), the alignment of their sequences showed the presence of several motifs that were well conserved in both domains, suggesting that these orthologs may have similar substrates (see Figure 2).

Figure 4

Download asset Open asset

A partial protein structure of human SETD3 has been published (PDB 3SMT). It lacks about 100 amino acids at the C-terminus, but it shows both substrate-binding domains. As shown in Figure 4A, these two domains form a wide cleft in the center of SETD3, so that the enzyme adopts a cradle-like shape. The SAM molecule is bound in the bottom of the cleft, exposing its labile methyl group (see Figure 4A). It seems likely that the cradle-shape of the enzyme allows it to arch over a protein substrate, and that the position of SAM facilitates the subsequent transfer of the methyl group onto a specific amino acid (in the case of actin, a histidine), most probably located at the surface of a substrate protein.

As shown in Figure 4B, β-actin has a globular shape, flattened in one dimension. It consists of large and small domains that are further subdivided into subdomains 1, 2 and 3, 4, respectively. The nucleotide-binding cleft (for ATP or ADP binding) is located inside the cleft between subdomains 2 and 4, while a hydrophobic cleft is present between subdomains 1 and 3 at the ‘bottom’ of the actin molecule, and serves as a primary binding site for numerous actin-binding proteins. The methyl-accepting H73 residue is present in a nucleotide sensor loop (P70 to N78) in an insert between subdomains 1 and 2. The H73 is present at the surface of the actin monomers and seems to be easily accessed by the methyltransferase. Furthermore, the estimated dimensions of the actin monomer and of the substrate-binding cleft in SETD3 indicate that actin could be accommodated well at the cleft of the methyltransferase and that steric hindrance should not prevent interactions between these proteins.

To confirm the molecular identity of rat actin-specific histidine N-methyltransferase with that of rat SETD3 and to compare the activities of rat and human methyltransferases, these two SETD3 orthologs were expressed in COS-7 cells as fusion proteins with the C-terminal polyhistidine tag (Figure 5). The recombinant enzymes catalyzed the methylation of recombinant β-actin, but not of the H73A mutant (see Figure 5). In further studies, large amounts of recombinant enzymes were produced in COS-7 cells, purified to homogeneity (Figure 6—figure supplement 1) and shown to catalyze the methylation of actin (Figure 6). In addition, the recombinant human and rat SETD3 with the N-terminal His₆-tag were produced in E. coli, purified (Figure 7—figure supplement 1) and also shown to methylate actin (Figure 7). Their specific activity was similar to that determined for the enzymes produced in the COS-7 cells (≈5 nmol.min⁻¹.mg⁻¹ protein), which further confirms the identification of SETD3 as the actin-specific methyltransferase.

Figure 5

Download asset Open asset

Figure 5—source data 1

Radiochemical measurements of actin-specific histidine N-methyltransferase activity in lysates from COS-7 cells that overexpress recombinant rat SETD3.

https://doi.org/10.7554/eLife.37921.013

Download elife-37921-fig5-data1-v1.xls

Figure 6 with 1 supplement see all

Download asset Open asset

Figure 6—source data 1

Radiochemical measurements of β-actin methylation by purified recombinant SETD3 overexpressed in COS-7 cells.

https://doi.org/10.7554/eLife.37921.016

Download elife-37921-fig6-data1-v1.xls

Figure 7 with 1 supplement see all

Download asset Open asset

Figure 7—source data 1

Radiochemical measurements of β-actin methylation by purified recombinant SETD3 overexpressed in E. coli.

https://doi.org/10.7554/eLife.37921.019

Download elife-37921-fig7-data1-v1.xls

As partially purified rabbit actin-specific histidine N-methyltransferase (Raghavan et al., 1992) was shown to methylate a synthetic peptide (Peptide H: YPIEHGIVT), corresponding to the H73-containing nucleotide sensor loop of the actin molecule, we wanted to know whether SETD3 can methylate peptide H. As shown in Figure 8, the two SETD3 orthologs catalyzed the methylation of peptide H, but they did not methylate peptide A (YPIEAGIVT), a variant of the nucleotide sensor loop in which H73 is replaced by an alanine. Even if the rat and human SETD3 activities measured with peptide H were very poor (≈ 0.02 nmol.min⁻¹.mg⁻¹ protein) compared with the actual methylation of actin (≈ 5 nmol. min⁻¹.mg⁻¹ protein), these activities further indicated that SETD3 is the histidine-specific N-methyltransferase.

Figure 8 with 1 supplement see all

Download asset Open asset

Figure 8—source data 1

Radiochemical measurements of the SETD3-dependent methylation of synthetic peptides that incorporate the methylation site in β-actin.

https://doi.org/10.7554/eLife.37921.023

Download elife-37921-fig8-data1-v1.xls

To confirm the specific methylation of H73 in recombinant human β-actin, we incubated it with homogenous recombinant rat and human SETD3 enzymes in the presence of deuterated SAM ([²H]SAM). This resulted in approximately 35% of the actin being methylated, as measured using a radiochemical assay that was run in parallel. At the end of the reaction, the proteins in the assay mixture were separated by SDS-PAGE, and the bands corresponding to the β-actin polypeptide were analyzed by nanoUPLC-Q-TOF. At least 74% of the actin sequence coverage was achieved in each MS analysis, including all of the known methyl-accepting sites in human β-actin (K18, K68, H73, K84, K191 and K326, see http://www.phosphosite.org). Peptide YPIEHGIVTNWDDMEK (M + H = 1963.9355 Da) containing the H73 residue was the only tri-deuterium-methylated peptide detected by the PLGS 2.4 software (Waters, USA). Manual inspection of the MS/MS spectra of both the parent ion of this peptide and its fragments confirmed the presence of tri-deuterium-methylated H73 (Figure 9). Both rat and human SETD3 enzymes catalyzed tri-deuterium-methylation of H73 in β-actin, whereas no such modification was detected in YPIEHGIVTNWDDMEK peptide (M + H = 1946.9021 Da) or in any other tryptic peptide derived from β-actin protein incubated in the absence of SETD3 (see Figure 9).

Figure 9

Download asset Open asset

Figure 9—source data 1

Mass spectrometry (Q-TOF) identification of both non-modified and deuterium(D3)-methylated β-actin peptides produced in the absence or presence of the human SETD3 protein.

https://doi.org/10.7554/eLife.37921.025

Download elife-37921-fig9-data1-v1.xls

SETD3 was originally reported as a methyltransferase that methylates histone H3 in zebrafish and mouse (Kim et al., 2011; Eom et al., 2011). Specifically, the enzyme was shown to methylate a histone H3.3 peptide (H3N4: STGGVK), suggesting that it could methylate K36 of mouse histone H3. We therefore compared the SETD3-dependent methylation of actin peptide H (see Figure 8) and the histone peptide H3N4. We could see that SETD3 methylated both peptides (Figure 8—figure supplement 1), but peptide H was at least 10-fold more efficiently methylated than peptide H3N4, confirming that SETD3 is most probably a histidine N-methyltransferase rather than a lysine-specific methyltransferase.

The kinetic properties of recombinant SETD3 proteins were studied in detail using either recombinant human β-actin or the synthetic peptide H as substrates in the presence of E. coli  S-adenosyl-L-homocysteine (SAH) nucleosidase and Bacillus subtilis adenine deaminase. The two bacterial enzymes collaborated to break down SAH, a potent inhibitor of methyltransferases, preventing its accumulation in the reaction mixture (Dorgan et al., 2006). Rat and human SETD3 have a very high affinity towards both SAM and human β-actin (Table 3), with K_M values well below the physiological concentrations of SAM and β-actin monomers in vertebrate cells (≈ 30 µM and 50–100 µM, respectively (Clarke and Banfield, 2001; Pollard, 2017). Both enzymes had comparable estimates of V_max value at saturating concentrations of SAM, but the K_M for β-actin (≈0.8 µM) was about 4-fold lower for human SETD3 than for the rat enzyme (3 µM). Consequently, the catalytic efficiency (K_M/k_cat) of human SETD3 towards β-actin was 3-fold higher than that of the rat enzyme (see Table 3). Taking into account the intracellular concentrations of SAM and β-actin monomers, however, the two enzymes probably operate at similar V₀ in vivo, i.e. close to their V_max (≈10 nmol.min⁻¹.mg⁻¹ protein). By contrast, actin peptide H was a very poor substrate for SETD3 as rat and human SETD3 catalyzed methylation of the peptide with extremely low V_max values and showed several thousand–fold lower affinity towards the peptide in comparison to the values determined with human β-actin (see Table 3).

It is well-established that actin that is produced in bacteria forms non-native mis-folded species that aggregate into inclusion bodies (Stemp et al., 2005). This is because actin requires the assistance of the eukaryotic chaperonin CCT (TRiC) to enable it to fold into its native conformation (Martín-Benito et al., 2002; Stemp et al., 2005). Completely denatured actin is, however, largely capable of re-folding in a spontaneous manner into its native tertiary structure (Martín-Benito et al., 2002), and it was this re-folded recombinant β-actin that we used in the experiments described above. We next wanted to know whether STED3 would methylate natively folded actin, like that found in cells, equally well. To this end, recombinant human β-actin was produced in Saccharomyces cerevisiae, an eukaryotic species that is devoid of actin-specific histidine methyltransferase activity (Kalhor et al., 1999). Surprisingly, neither rat nor human SETD3 methylated ATP-β-actin or ADP-β-actin, as determined using the radiochemical assay (Figure 10, see Figures 6 and 7), suggesting that the endogenous substrate of SETD3 might not be the natively folded nucleotide-bound actin monomers. More importantly, yeast-produced human β-actin became a good substrate for SETD3 when purified in denaturing conditions and re-folded into nucleotide-free quasi-native actin, as did E. coli-produced recombinant human β-actin (see Figure 10). This effect was not due to the absence of free nucleotides in the reaction mixture because the addition of either 0.5 mM ATP or ADP (a 10-fold excess over the concentrations of nucleotides in the preparation of yeast-produced actin) did not impact the activity of SETD3 towards the refolded actin. To try to solve this puzzle, we prepared nucleotide-free yeast-produced human β-actin monomers in complex with either profilin or cofilin (two well known interacting partner proteins of β-actin [Lappalainen, 2016]) and tested whether the actin complexes became good substrates for SETD3. The methylation of profilin-β-actin was negligible and that of cofilin-β-actin complex was low (only ≈3–4% of the total protein in the complex could be methylated). Taken together, our results indicate that globular actin monomers (probably in complex with some actin-binding proteins) or possibly also F-actin can be methylated by SETD3 in vivo.

Figure 10

Download asset Open asset

Figure 10—source data 1

Radiochemical measurements of SETD3-dependent methylation of human β-actin overexpressed in S. cerevisiae.

https://doi.org/10.7554/eLife.37921.031

Download elife-37921-fig10-data1-v1.xls

To assess the importance of SETD3 in methylating the β-actin H73 residue in vivo, we used CRISPR/Cas9 technology to inactivate the Setd3 gene in HAP1 cells. Five distinct Setd3-KO clonal HAP1 cell-lines were generated, as described in the 'Materials and methods', and three of these were ultimately selected for further studies (KO-A1, KO-A3 and KO-A5). Sequencing of the DNA showed the presence of mutations, preventing the production of active SETD3 in these cell lines (see 'Materials and methods' for details of the mutations introduced), which was further confirmed by the absence of SETD3 protein in cell extracts from KO-A1, KO-A3 and KO-A5 (Figure 11).

Figure 11

Download asset Open asset

To analyze the methylation of H73 in actin from wildtype and SETD3-KO HAP1 cell lines, we partially purified the actin from these lines (Figure 12A) and analyzed these proteins by tandem mass spectrometry of the band corresponding to actin from the fraction containing the highest concentration of actin (Figure 12B). As shown in Figure 12C, methylation of the H73 residue was detected in almost all H73-containing β-actin peptides derived from the wildtype HAP1 cells, whereas close to 90% of the same β-actin peptides derived from two different Setd3-deficient cell lines were not methylated at H73. Importantly, the total number of β-actin peptides identified in MS/MS experiments was comparable for wildtype (300–400 Peptide Spectrum Matches) and KO cells (≈500 Peptide Spectrum Matches), indicating that similar amounts of actin protein were analyzed in all experiments.

Figure 12

Download asset Open asset

Figure 12—source data 1

Mass spectrometry (LTQ XL ion trap) identification of both non-modified and methylated β-actin peptides from HAP1 human cells.

https://doi.org/10.7554/eLife.37921.034

Download elife-37921-fig12-data1-v1.xls

We next took advantage of the recently published knock-out model of SETD3 in D. melanogaster (Tiebe et al., 2018) to investigate the methylation status of actin in vivo. The absence of SETD3 in the fly prevented the methylation of Drosophila actin on H74 (which corresponds to the H73 in the mammalian protein). Soluble proteins in homogenates from wildtype and Setd3-deficient Drosophila larvae were analyzed by tandem mass spectrometry. As shown in Table 4, the knock-out of SETD3 prevented the methylation of all three different forms of actin (actin-42A, −57B and −87E) present in extracts from the fly larvae, whereas all H74-containing actin peptides derived from wildtype larvae were methylated.

Taken together, we show that actin from wildtype HAP1 cells or Drosophila flies was > 90% methylated, whereas in the absence of SETD3, neither human cells nor Drosophila can methylate actin H73 or H74, respectively. This confirms the molecular identification of SETD3 as the evolutionarily conserved actin-specific histidine N-methyltransferase, and the extent of actin methylation suggests the high functional importance of this post-translational modification.

Methylation of the actin H73 residue has been shown to stabilize actin filaments, whereas the absence of such modification resulted in an increased rate of depolymerization of these filaments in vitro (Nyman et al., 2002). To determine the effect of Setd3-deficiency on the organization of actin filaments, we used confocal microscopy following F-actin staining with TRITC-phalloidin. To obtain a comprehensive picture of actin filaments, the morphology of F-actin was reconstructed in 3D with the use of Zeiss Zen software (Zeiss, Germany). This technique showed a clear decrease in F-actin content in SETD3-KO cells, where actin was largely non-methylated at H73 (Figure 13), suggesting an accelerated actin depolymerization and a loss of cytoskeleton integrity. These effects were more pronounced after 48 hr culture (compared to 24 hr), and are in agreement with the results of previous studies using purified recombinant actin (Nyman et al., 2002). To exclude artifacts associated with the clonal populations of HAP1 cells, we analyzed two other SETD3-KO clones that showed comparable results (Figure 13—figure supplement 1).

Figure 13 with 1 supplement see all

Download asset Open asset

Since Nyman et al. (2002) showed that the replacement of H73 by an alanine decreased the thermal stability of actin monomers in vitro, this suggested that the methylation of actin at H73 could contribute to its stability, in agreement with our results in HAP1 cells. Therefore, we wanted to compare the in vitro thermostability of actin in cell-free lysates of wildtype (methylated actin) and SETD3-KO (non-methylated actin) HAP1 cells. When cell extracts were heated at increasing temperatures, it became clear that the absence of methylation of H73 destabilized actin in the temperature range between 30°C to 46°C (Figure 14). Taken together, these results suggest that the loss of methylation in H73 leads to a structural instability of actin monomers that might, in turn, result in the accelerated depolymerization of actin filaments in SETD3-KO cells.

Figure 14

Download asset Open asset

Figure 14—source data 1

Quantification of western blots with Quantity One software.

https://doi.org/10.7554/eLife.37921.040

Download elife-37921-fig14-data1-v1.xls

A routine daily inspection of HAP1 cell cultures revealed that at about 80–90% culture confluency, Setd3-deficient HAP1 cells tended to acidify Iscove's Modified Dulbecco's Medium (IMDM) faster than the control cells. This phenotype was not attributable to the differences in cell culture density between control and SETD3-KO cell lines, suggesting an increased formation of lactate from glucose by Setd3-deficient cells. To test this possibility, both control and Setd3-deficient HAP1 cells (90% culture confluency) were further incubated in fresh IMDM medium for 20 hr, after which the medium was collected, and glucose consumption and lactate production were measured in a neutralized HClO₄ extract. As shown in Figure 15, both the formation of lactate and the consumption of glucose in SETD3-KO cell lines were increased by about 15–20% in comparison with those determined for the control cells, indicating a shift of their metabolism towards a more glycolytic phenotype. Taken together, these results suggest that the lack of methylation at H73 in HAP1 cells destabilizes F-actin, which possibly explains the increase in glucose consumption that occurs in these cells.

Figure 15

Download asset Open asset

Figure 15—source data 1

Quantification of glucose consumption in HAP1 human cells.

https://doi.org/10.7554/eLife.37921.042

Download elife-37921-fig15-data1-v1.xls

Figure 15—source data 2

Quantification of lactate production in HAP1 human cells.

https://doi.org/10.7554/eLife.37921.043

Download elife-37921-fig15-data2-v1.xls

The presence of two distinct histidine N-methyltransferases in rabbit leg muscle has been reported by Raghavan et al. (1992). The first enzyme, carnosine N-methyltransferase catalyzes the Nπ-methylation of the histidine imidazole ring of carnosine (β-alanyl-L-histidine), an abundant dipeptide in the skeletal muscle of vertebrates. The second enzyme was shown to methylate the Nτ position of the H73 residue of mammalian β-actin protein. The previous studies performed in our laboratory resulted in the identification of genes coding for carnosine N-methyltransferase in vertebrates (Drozak et al., 2013; Drozak et al., 2015), whereas the current investigation presents data that reveal the molecular identification of rat actin-specific histidine N-methyltransferase as the SET domain-containing protein 3 (SETD3), disclosing the molecular identity of the mammalian actin-methylating enzyme. This conclusion is supported by the following findings: (i) two different and extensive rounds of purification of the actin-methylating activity from rat skeletal muscle resulted in the identification of the protein SETD3 as the only logical candidate matching the enzymatic activity; (ii) recombinant rat SETD3 catalyzes the specific methylation of the H73 residue present in both the recombinant β-actin protein and the synthetic peptide corresponding to the nucleotide sensor loop of β-actin; (iii) in contrast to actin isolated from control HAP1 cells, the protein present in Setd3-deficient cells is essentially devoid of methylated H73 residues, as verified by MS/MS analysis; and (iv) the inactivation of SETD3 in Drosophila Setd3 knockout larvae results in the complete absence of H74 methylation in all three different actins found in Drosophila.

The methylation of H73 is a highly conserved modification among actins from many different eukaryotes, although it has been found to be absent from the actin of the amoeboflagellate N. gruberi and the yeast S. cerevisiae (Sussman et al., 1984; Nyman et al., 2002). Orthologues of the protein SETD3 are present in animals, plants and certain fungi, but not in the proteomes of the abovementioned proteozoan and yeast species, further supporting the identification of SETD3 as the actin-specific histidine N-methyltransferase.

SETD3 was originally reported as a methyltransferase that catalyzes the modification of histone H3 at K4 and K36 residues in zebrafish and mouse (Kim et al., 2011; Eom et al., 2011). The independent studies of Chen et al., 2013 confirmed the histone methyltransferase activity of SETD3 and suggested that the enzyme may have other non-histone substrates in the cytoplasm owing to the presence of the Rubisco LSMT substrate-binding domain in its structure. This view was further strengthened by the finding that SETD3 is predominantly localized in the cytosol of human cells, while only a small portion is found in the nucleus (Cheng et al., 2017). In the present study, we have not examined the activity of recombinant SETD3 towards histones because isolated histones are extremely poor substrates (Chen et al., 2013). When we investigated the methyltransferase activity of SETD3 with either the actin peptide H or the histone peptide H3N4, however, the activity with the histone peptide was at least a 4000-fold lower than that of recombinant β-actin. Consequently, we identified actin as the novel physiological substrate of SETD3, which is present in the cytoplasm and is likely to be the most abundant actin methyltransferase. It is possible however, that SETD3 may also be a dual methyltransferase able to methylate both the Lys and His residues of other proteins, but this still requires in vivo confirmation.

The molecular identification of the actin-specific histidine N-methyltransferase through its extensive purification from rat muscles was only possible because we were able to prepare large amounts of pure β-actin monomers to be used as the substrate for measuring the enzymatic activity in radiochemical assays. Although this β-actin was produced and subsequently purified from the inclusion bodies of E. coli in the nucleotide-free form, it remained a good substrate for the methyltransferase. Given that, on the one hand, it is not possible to produce natively folded actin in bacteria (Stemp et al., 2005; Martín-Benito et al., 2002) and that, on the other hand, the ATP/ADP-free actin denatures rapidly and irreversibly (Graceffa and Dominguez, 2003), it is likely that the actin preparation used in our study may comprise actin molecules with stable non-native spatial structures. Indeed, the actin methylation reaction was still substoichiometric, and we have observed the methylation of up to 35% of the total actin present in the reaction mixture. This observation was readily reproducible in the presence of bacterial SAH nucleosidase and adenine deaminase, suggesting that it was not due to a potential accumulation of SAH in the reaction mixture (unpublished data, Drozak et al.). This indicates that only a small portion of the actin monomers produced in bacteria and purified under denaturing conditions adopted a conformation that was recognized by the SETD3 enzyme. This observation is in line with a previous report, showing only a minor methylation of actin inclusion bodies (up to 1% protein) by rabbit actin methyltransferase (Raghavan et al., 1992). Furthermore, SETD3 was at least 200-fold less active towards the synthetic peptide H, corresponding to the amino acid sequence surrounding the H73, than was recombinant β-actin (see Table 3). This indicated that structural features of the actin molecule other than the primary sequence of the sensor loop are probably important for its recognition by the actin methyltransferase.

Both tested SETD3 orthologs are rather ‘lazy’ enzymes, with a k_cat of 0.6–0.8 min⁻¹, resembling the values estimated for protein methyltransferases rather than those measured for enzymes catalyzing the methylation of low-molecular-weight substrates (Clarke and Banfield, 2001; Osborne et al., 2007). Considering that the intracellular concentrations of SAM and β-actin monomers in vertebrate cells are about 30 µM and 50–100 µM, respectively (Clarke and Banfield, 2001; Pollard, 2017), both rat and human SETD3 have a high affinity for both SAM (K_M≈ 0.3 µM) and β-actin (K_M≈ 0.8–3 µM). These observations suggest that the enzymes may be fully saturated by substrates and operate at rates that are close to their V_max in vivo. It is important to keep in mind, however, that the recombinant β-actin used in the present kinetic studies is certainly not a physiological form of β-actin. Therefore, the affinity of SETD3 to physiological β-actin remains to be verified.

It is very intriguing to note that in our hands both the recombinant ATP- and ADP-β-actins produced in yeast (S. cerevisiae), which are expected to have a native conformation, were not methylated by the SETD3 enzyme. By contrast, the yeast-produced actin that was purified in denaturing conditions to yield nucleotide-free quasi native protein became a good substrate for the methyltransferase. This finding may suggest that the presence of ATP or ADP could create a structural hindrance that prevents SETD3 from reaching the H73 residue if the methyltransferase interacts with actin through the nucleotide-binding cleft (see Figure 4).

Taken together, these observations have led us to hypothesize that nucleotide-free actin monomers in a complex with one or more actin-binding proteins of yet unknown identity could be the physiological substrate for the SETD3 methyltransferase. If so, this resembles the activity of the yeast Yil110w protein, the only other known protein histidine N-methyltransferase, which methylates His243 of the ribosomal Rpl3 protein only when Rpl3 is associated with the ribosome (Webb et al., 2010; Al-Hadid et al., 2014).

This hypothesis fits with the current knowledge on the enzymes belonging to the family of SET-domain-containing proteins. Indeed, many of these methyltransferases are found in complexes with numerous other proteins that are essential for their catalytic activity and substrate specificity. For example, to exert its catalytic activity, the yeast histone H3 methyltransferase SET1 requires the presence of seven different proteins that form a complex named COMPASS (Complex of Proteins Associated with Set1), and similar COMPASS-like complexes are formed by other mammalian SET-domain-containing methyltransferases (for a review, see Herz et al., 2013).

Several reports describing the role of SETD3 in the regulation of the cell cycle and apoptosis (Kim et al., 2011), myocyte differentiation (Eom et al., 2011), cell response to hypoxic conditions (Cohn et al., 2016) and tumorigenesis (Pires-Luís et al., 2015; Cheng et al., 2017) have been published to date. Some of these studies identified the methylation of histones as a primary mechanism of SETD3 action (Kim et al., 2011; Eom et al., 2011), whereas others suggested that the enzyme may instead modify non-histone substrates that are present in the cytosol (Chen et al., 2013; Cohn et al., 2016). The existence of two distinct domains in SETD3, i.e. SET and Rubisco LSMT substrate-binding domains, has been suggested to be the structural basis for such dual substrate specificity (Chen et al., 2013).

Our results identify β-actin as a novel substrate of SETD3 and show that disruption of its activity leads to the depletion of F-actin and a loss of cytoskeleton integrity. This observation is probably explained by an increased rate of depolymerization of the non-methylated actin filaments. As a result, there is an accelerated conversion of a more structurally stable ATP-F-actin into a less stable ADP-F-actin, as also shown for a hypomethylated form of recombinant actin in vitro (Nyman et al., 2002). Thus, the biochemical role of the methylated H73 residue is likely to contribute to stabilizing the structure of monomers that build the ATP-F-actin and to delay the intermolecular conformational rearrangements of F-actin that complement the hydrolysis of ATP. This conclusion finds at least partial confirmation in the thermal stability assay, which shows that the methylated actin derived from the wildtype HAP1 cells is clearly more thermo-stable than the non-methylated one (SETD3-KO cells) (see Figure 14). Taken together, our results show that the phenotype that we have identified in Setd3-deficient human HAP-1 cells is caused by the absence of the catalytic activity of SETD3, rather than by a non-catalytic or actin-stabilizing interaction between SETD3 and actin filaments.

The inherent instability of F-actin in Setd3-deficient HAP1 cells is likely to be at the origin of the increased lactate production by these cells. Both the polymerization and the depolymerization of actin filaments occur continuously even in resting cells and constitute a major energy drain, causing up to 50% of total ATP consumption in vertebrate cells (Kudryashov and Reisler, 2013). Thus, the destabilization that leads to the accelerated degradation of hypomethylated F-actin probably drives an increasing demand for ATP that could be fulfilled by shifting the cell metabolism towards glycolysis. The phenotype of Setd3-deficient cells is therefore somewhat similar to that of malignant cells, suggesting that SETD3 activity may play a role in suppressing tumor development. Indeed, the hypomethylation of actin was observed in rat cells following their transformation with the Src oncogen (Chiou et al., 2012). Furthermore, the enzymatically inactive SET-domain-loss form of SETD3 was shown to function as a dominant-negative mutant, promoting oncogenesis in human cells (Chen et al., 2013), most probably by blocking the methylation of an unknown SETD3 substrate. It is tempting to speculate that actin could be this substrate.

In the current investigation, we have identified the rat actin-specific histidine N-methyltransferase as SETD3, an enzyme catalyzing the extremely well-conserved methylation of H73 in β-actin. The SETD3 enzyme is therefore the first protein-histidine methyltransferase to be identified in vertebrates and the first SET-domain-containing enzyme possibly displaying a dual methyltransferase specificity towards both Lys and His residues in proteins. SETD3 is responsible for the methylation of H73 in human and Drosophila actin, and a loss of SETD3 activity in human HAP1 cells induces phenotypic changes resembling those present in cancer cells, suggesting that the hypomethylation of actin might be involved in tumorigenesis. Finally, this work also shows that non-native protein substrates might be useful tools in the search for novel protein methyltransferases.

While this paper was in production, similar and complementary data describing the identification of SETD3 as the actin-specific histidine N-methyltransferase was accepted for publication (Wilkinson et al., 2018). Since the work was independently carried out by both research groups and equally contributes to reveal the biochemical importance of SETD3 enzyme as the actin-specific H73 N-methyltransferase, we ask the authors referring to the function SETD3 in their future publications to mention both this paper, as well as Wilkinson et al. (2018).

DEAE-Sepharose, Q-Sepharose, Phenyl-Sepharose, Superdex 200, HiScreen Blue FF, HisTrap HP (Ni²⁺ form) and PD-10 columns were obtained from GE Healthcare Life-Sciences (Little Chalfont, UK). Reactive Red 120-Agarose and Dowex 50W-X4 (200 mesh) resins came from Sigma-Aldrich, and Vivaspin-500 and 20 centrifugal concentrators were from Sartorius Stedim (Goettingen, Germany). All other enzymes and DNA-modifying enzymes as well as the TurboFect transfection reagent were obtained from Thermo-Fermentas (Waltham, USA), A and A Biotechnology (Gdynia, Poland) or BioShop (Burlington, Canada).

Plasmid pCOLD I encoding human β-actin (ACTB, GenBank: NM_001101.4, pCOLD I/β-actin) was a kind gift from Dr. Minoru Tamura (Ehime University, Japan) and was prepared as described by Tamura et al. (2011).

For β-actin production, E. coli BL21(DE3) (Agilent, USA) cells were transformed with the DNA construct and a single colony was selected to start an over-night pre-culture. 500 mL of LB broth (with 100 μg/mL ampicilin) was inoculated with 50 ml of the pre-culture and incubated at 37°C and 200 rpm until an OD600 of 0.5 was reached. The culture was placed on ice for 20 min (cold-shock) and isopropyl β-D-1-thiogalactopyranoside (IPTG) was added to a final concentration of 0.2 mM to induce protein expression. Cells were incubated for 20 hr at 15°C, 200 rpm, and harvested by centrifugation (6000 × g for 10 min). The cell paste was resuspended in 27.5 ml lysis buffer consisting of 20 mM Hepes, pH 7.5, 1 mM DTT, 1 mM ADP, 0.5 mM PMSF, 2 μg/ml leupeptin and 2 μg/ml antipain, together with 0.2 mg/ml hen egg white lysozyme (BioShop), and 1000 U Viscolase (A and A Biotechnology, Poland). The cells were lysed by freezing in liquid nitrogen and, after thawing and vortexing, the extracts were centrifuged at 4° C (20,000 × g for 30 min).

The pellet, containing inclusion bodies, was completely resuspended in buffer A (20 mM Hepes, pH 7.5, 2M urea, 0.5 M NaCl, 5 mM DTT and 2 mM EDTA) with the use of Potter-Elvehjem homogenizer and centrifuged at 4° C (20,000 × g for 10 min). The resulting pellet was then subjected to a further two rounds of sequential wash in buffer B (20 mM Hepes, pH 7.5, 0.5 M NaCl, 5 mM DTT and 2 mM EDTA) and buffer C (20 mM Hepes, pH 7.5, and 0.5 M NaCl). The washed inclusion bodies were finally solubilized in loading buffer (20 mM Tris-HCl, pH 7.5, 6 M guanidine HCl, 0.5 M NaCl, and 10 mM imidazole) and applied onto a HisTrap FF column (5 ml) equilibrated with the same buffer.

The column was washed with 20 ml of loading buffer and the bound protein was refolded by the use of a linear gradient of 6–0 M guanidine HCl in loading buffer (40 ml for 20 min). Next, the column was washed with 15 ml of loading buffer without guanidine HCl and the retained proteins were eluted with a stepwise gradient of imidazole (25 ml of 40 mM, 25 ml of 60 mM and 20 ml of 500 mM). The recombinant proteins were eluted with 500 mM imidazole in homogeneous form as confirmed by SDS-PAGE (see Figure 1—figure supplement 1). The β-actin preparation was immediately desalted on PD-10 columns equilibrated with 20 mM Tris-HCl, pH 7.5, 1 mM DTT, 6% sucrose, 2 μg/ml leupeptin and 2 μg/ml antipain. The purified β-actin was stored at −70°C.

Plasmid pCOLD I encoding human β-actin (pCOLD I/β-actin) was used as a template for PCR amplification of β-actin cDNA. The resulting ORF was then cloned into pESC-URA yeast expression vector (pESC-URA/β-actin). Saccharomyces cerevisiae BY4742 strain (EUROSCARF, Germany) was transformed with the above-mentioned plasmid using the high-efficiency lithium acetate transformation method (Gietz and Woods, 2002). The resulting yeast strain was designated as BY4742/β-actin.

The method for the purification of recombinant β-actin expressed in S. cerevisiae was that reported by Karlsson (1988). BY4742/β-actin was cultivated in 500 ml of SG medium (0.67% yeast nitrogen base without amino acids, 2% galactose, supplemented with histidine, leucine and lysine) for 24 hr. Cells were collected, washed with water, and suspended in 30 ml of buffer C (5 mM Tris-HCl pH 7.6, 0.2 mM CaCl₂, 0.2 mM DTT, and 0.3 mM PMSF). After the addition of 100 U of DNase I, cells were homogenized in Braun MSK cell homogenizer using glass beads (B. Braun Melsungen). Crude lysate was centrifuged (5000 × g, 5 min), diluted twice with the buffer, and applied to the Ni-NTA-agarose column (Sigma-Aldrich). Next, the column was washed with 30 ml of buffer C supplemented with 8 mM imidazole, and retained proteins were eluted with buffer C containing 500 mM imidazole (≈15 ml, 0.7 ml fractions). Protein concentration in each fraction was estimated by SDS-PAGE and the most concentrated samples were taken for further procedures.

To obtain a yeast-produced refolded human β-actin, pellet of yeast cells (1000 ml culture) was suspended in 30 ml of buffer D (20 mM Tris-HCl pH 7.2, 0.5 M NaCl, 1 mM DTT, and 100 U of DNase I) and homogenized in a Braun MSK cell homogenizer. Crude lysate was then supplemented with 6 M guanidine HCl and 10 mM imidazole, centrifuged (20,000 × g, 20 min) and applied on a HisTrap FF column (5 ml). The subsequent on-column refolding and purification of β-actin was performed as described for E. coli-derived protein.

For preparation of β-actin complexes with either cofilin or profilin (see below), an equimolar amount of the recombinant homogenous actin-binding protein was mixed with the yeast-produced β-actin and dialyzed twice against 1000 ml of dialysis buffer (20 mM Tris-HCl pH 7.6, 1 mM DTT, 6% sucrose, and 0.2 mM CaCl₂), for 2 hr in 250 ml and then overnight in 500 ml. When obtaining β-actin complexed with either ADP or ATP, samples eluted from Ni-NTA-agarose were dialyzed under the same conditions against dialysis buffer containing 50 μM of an appropriate nucleotide. In all cases, dialyzed protein preparation was concentrated using an Amicon 10K device (Merck, Germany).

Rat total RNA was prepared from 200 mg of leg muscles with the use of TriPure reagent according to the manufacturer’s instructions, whereas human skeletal muscle total RNA was purchased from Clontech (USA). cDNA was synthesized using Moloney murine leukemia virus-reverse transcriptase (Thermo-Fermentas), with oligo(dT)₁₈ primer and 2.5 μg total RNA according to the manufacturer’s instructions.

The open reading frames encoding rat (NCBI Reference Sequence: XM_002726774.2) and human (NM_032233.2) SETD3 protein were PCR-amplified using Pfu DNA polymerase in the presence of 1 M betaine. The SETD3 genes were amplified using 5′ primers containing the initiator codon preceded by the Kozak consensus sequence (Kozak, 1987) and an KpnI site, and 3′ primers in which the original stop codon was replaced by an amino-acid-coding codon flanked by a NotI site (for primer sequences, see Table 5). The amplified DNA products of the expected size were digested with the appropriate restriction enzymes, cloned in the pEF6/Myc-His A expression vector (Invitrogen, USA) (which allows the production of proteins with an C-terminal His₆-tag), and verified by DNA sequencing (Macrogen, The Netherlands). The resulting vector was designated as pEF6/SETD3. For transfections, COS-7 cells (Cell Lines Service, Germany) were plated in 100 mm Petri dishes at a cell density of 1.7 × 10⁶ or 2.1 × 10⁶ cells per plate in Dulbecco's minimal essential medium supplemented with 100 units/ml penicillin, 100 μg/ml streptomycin, and 10% (v/v) fetal bovine serum, and grown in a humidified incubator under 95% air and 5% CO₂ atmosphere at 37°C. After 24 hr, each plate was transfected with 7–8 μg of either unmodified pEF6/Myc-His A vector or pEF6/SETD3 using the TurboFect transfection reagent according to the protocol provided by the manufacturer. After 48 hr, the culture medium was removed, the cells were washed with 5 ml phosphate buffered saline and harvested in 1 ml of 20 mM Tris-HCl pH 7.2, containing 10 mM KCl, 1 mM DTT, 5 μg/ml leupeptin and 5 μg/ml antipain. The cells were lysed by freezing in liquid nitrogen and, after thawing and vortexing, the extracts were centrifuged at 4° C (20,000 × g for 30 min) to remove insoluble material.

For the purification of recombinant SETD3 proteins, the supernatant of the COS-7 lysate (13–18 ml) was diluted 3-fold with buffer A (50 mM Tris-HCl pH 7.4, 400 mM NaCl, 10 mM KCl, 30 mM imidazole, 3 μg/ml leupeptin and 3 μg/ml antipain) and applied onto a HisTrap HP column (1 ml) equilibrated with the same buffer. The column was washed with 6–10 ml buffer A and the retained protein was eluted with a stepwise gradient of imidazole (6 ml of 60 mM, 6 ml of 150 mM and 6 ml of 300 mM) in buffer A. The recombinant proteins were eluted with 150 mM imidazole in homogeneous form, as confirmed by SDS-PAGE (see Figure 6—figure supplement 1). The enzyme preparations were desalted on PD-10 columns equilibrated with 20 mM Tris-HCl pH 7.2, 50 mM KCl, 1 mM DTT, 6% sucrose, 2 μg/ml leupeptin and 2 μg/ml antipain. The yield of recombinant proteins was 0.5–0.6 mg of homogenous protein per about 25–40 mg of soluble COS-7 cell protein. The purified enzymes were aliquoted and stored at −70°C. Repeated freeze-thawing was avoided because it impairs SETD3 activity.

Bacterial expression plasmids for human and rat SETD3 fused to an N-terminal His₆-tag were constructed by sub-cloning the open reading frames from the pEF6/SETD3 plasmids into the pCOLD I vector (Takara Bio, Kusatsu, Japan), using KpnI and XbaI enzymes (pCOLD I/SETD3).

For protein production, E. coli BL21(DE3) cells were transformed with an appropriate DNA construct and a single colony was selected to start an over-night pre-culture. 500 mL of LB broth (with 100 μg/mL ampicilin) was inoculated with 50 ml of the pre-culture and incubated at 37°C and 200 rpm until an OD600 of 0.6 was reached. The culture was placed on ice for 20 min (cold-shock) and IPTG was added to a final concentration of 0.3 mM to induce protein expression. Cells were incubated for 16 hr at 13°C, 200 rpm, and harvested by centrifugation (6000 × g for 10 min). The cell paste was resuspended in 25 ml lysis buffer consisting of 25 mM Hepes pH 7.5, 300 mM NaCl, 10 mM KCl, 1 mM DTT, 2 mM MgCl₂, 1 mM PMSF, 0.25 mg/ml hen egg white lysozyme (BioShop, Canada) and 1250 U Viscolase (A and A Biotechnology, Poland). The cells were lysed by freezing in liquid nitrogen and, after thawing and vortexing, the extracts were centrifuged at 4° C (20000 × g for 20 min).

For the purification of recombinant SETD3 proteins, the supernatant of E. coli lysate (22 ml) was diluted 3-fold with buffer A (50 mM Hepes pH 7.5, 300 mM NaCl, 10 mM KCl, 30 mM imidazole, 1 mM DTT and 0.5 mM PMSF) and applied onto a HisTrap FF column (5 ml) equilibrated with the same buffer. The column was then washed with 35 ml buffer A, and the retained proteins were eluted with a stepwise gradient of imidazole (15 ml of 60 mM, 16 ml of 150 mM and 16 ml of 300 mM) in buffer A. The recombinant proteins were present in both 150 mM and 300 mM imidazole fractions (see Figure 7—figure supplement 1), but only SETD3 protein eluted at the highest concentration of imidazole, exhibiting >95% purity as confirmed by SDS-PAGE, was further processed. The enzyme preparation was desalted onto PD-10 columns equilibrated with 20 mM Tris-HCl pH 7.2, 50 mM KCl, 1 mM DTT and 6% sucrose. The yield of recombinant proteins was 6.6 mg and 8.2 mg of homogenous rSETD3 and hSETD3, respectively, per 500 ml of culture. The purified proteins were aliquoted and stored at −70°C.

Human skeletal muscle total RNA was obtained from Clontech (USA) and cDNA was synthesized as described above. The open reading frames encoding human cofilin-1 and profilin-1 (NCBI Reference Sequence: NM_005507.2 and NM_005022.3, respectively) were PCR-amplified using Pfu DNA polymerase in the presence of 1 M betaine. ORFs of both proteins were amplified using 5′ primers containing the initiator codon preceded by an NdeI site and 3′ primers flanked by a HindIII site (for primer sequences, see Table 5). The amplified DNA products of the expected size were digested with the appropriate restriction enzymes and cloned into the pCOLD I expression vector (Takara Bio, Kusatsu, Japan), which allows the production of proteins with an N-terminal His₆-tag (pCOLD I/Cofilin-1 and pCOLD I/Profilin-1), and verified by DNA sequencing (Macrogen, The Netherlands).

For protein production, E. coli BL21(DE3) cells were transformed with the appropriate DNA construct, and a single colony was selected to start an over-night pre-culture. 200 mL of LB broth (with 100 μg/mL ampicilin) was inoculated with 20 ml of the pre-culture and incubated at 37°C and 200 rpm until an OD600 of 0.6 was reached. The culture was placed on ice for 20 min (cold-shock) and IPTG was added to a final concentration of 0.25 mM to induce protein expression. Cells were incubated for 16 hr at 13°C, 200 rpm, and harvested by centrifugation (6000 × g for 10 min). The cell paste was resuspended in 10 ml lysis buffer consisting of 25 mM Hepes pH 7.5, 300 mM NaCl, 50 mM KCl, 1 mM DTT, 2 mM MgCl₂, 1 mM PMSF, 5 μg/ml leupeptin and 5 μg/ml antipain, together with 0.25 mg/ml hen egg white lysozyme (BioShop, Canada) and 50 U DNase I (Roche). The cells were lysed by freezing in liquid nitrogen and, after thawing and vortexing, the extracts were centrifuged at 4° C (20,000 × g for 30 min).

For the purification of recombinant actin-binding proteins, the supernatant of E. coli lysate (10 ml) was diluted 4-fold with buffer A (50 mM Hepes pH 7.5, 300 mM NaCl, 10 mM KCl, 30 mM imidazole and 1 mM DTT) and applied onto a HisTrap FF column (5 ml) equilibrated with the same buffer. The column was sequentially washed with 23 ml buffer A and 23 ml of the same buffer supplemented with urea at 1.8 M concentration. Next, the retained protein was eluted with a stepwise gradient of imidazole (23 ml of 60 mM, 23 ml of 150 mM and 22 ml of 300 mM) in buffer A. The recombinant proteins were eluted with 300 mM imidazole in homogeneous form, as confirmed by SDS-PAGE (not shown). The enzyme preparations were dialyzed three times against 400 ml of dialysis buffer consisting of 20 mM Tris-HCl pH 7.5, 200 mM NaCl, 1 mM DTT and 6% sucrose. The yield of recombinant proteins was 7.1 and 17.7 mg of homogenous cofilin-1 and profilin-1, respectively, per 200 ml of E. coli culture. The purified proteins were aliquoted and stored at −70°C.

E. coli DNA was extracted by heating 50 µl of over-night cultured BL21(DE3) cells at 95°C for 15 min, whereas Bacillus subtilis (ATCC 6633, Sigma-Aldrich) genomic DNA was purified from 100 mg of bacterial cells with the use of TriPure reagent according to the manufacturer’s instructions.

The open reading frames encoding E. coli SAH nucleosidase and B. subtilis adenine deaminase (NCBI Reference Sequence: NC_000913.3 and NC_000964.3, respectively) were PCR-amplified using either Pfu DNA polymerase alone or a mixture of Taq:Pfu polymerases (1:0.2), respectively, in the presence of 1 M betaine. SAH nucleosidase ORF was amplified using a 5′ primer containing the initiator codon preceded by an NdeI site and a 3′ primer with a HindIII site, whereas adenine deaminase DNA was amplified using a 5′ primer with the initiator codon preceded by an KpnI site and a 3′ primer with a BamHI site (for primer sequences, see Table 5). The amplified DNA products of expected size were digested with the appropriate restriction enzymes, cloned into the pCOLD I expression vector (pCOLD I/SAH nucleosidase and pCOLD I/adenine deaminase) and verified by DNA sequencing (Macrogen, The Netherlands).

For protein production, E. coli BL21(DE3) cells were transformed with the appropriate DNA construct and a single colony was selected to start an over-night pre-culture. 100 mL of LB broth (with 100 μg/mL ampicilin) was inoculated with 10 ml of the pre-culture and incubated at 37°C and 200 rpm until an OD600 of 0.6 was reached. The culture was placed on ice for 20 min (cold-shock) and IPTG was added to a final concentration of 0.25 mM to induce protein expression. Cells were incubated for 16 hr at 13°C, 200 rpm, and harvested by centrifugation (6000 × g for 10 min). The cell pellet was resuspended in 10 ml lysis buffer consisting of 25 mM Hepes pH 7.5, 300 mM NaCl, 50 mM KCl, 1 mM DTT, 2 mM MgCl₂, 1 mM PMSF, 5 μg/ml leupeptin and 5 μg/ml antipain, together with 0.25 mg/ml hen egg white lysozyme (BioShop) and 25 U DNase I (Roche). The cells were lysed by freezing in liquid nitrogen and, after thawing and vortexing, the extracts were centrifuged at 4°C (20,000 × g for 30 min).

For protein purification, the supernatant of the E. coli lysate (10 ml) was diluted 3-fold with buffer A (50 mM Tris-HCl, pH 7.2, 400 mM NaCl, 10 mM KCl, 30 mM imidazole, 1 mM DTT, 3 μg/ml leupeptin and 3 μg/ml antipain) and applied onto a HisTrap FF column (5 ml) equilibrated with the same buffer. The column was washed with 20–30 ml buffer A and the retained protein was eluted with a stepwise gradient of imidazole (25 ml of 60 mM, 20 ml of 150 mM and 20 ml of 300 mM) in buffer A. The recombinant proteins were eluted with 150–300 mM imidazole in homogeneous form, as confirmed by SDS-PAGE (not shown). The enzyme preparations were desalted onto PD-10 columns equilibrated with 20 mM Tris-HCl pH 7.2, 50 mM KCl, 1 mM DTT, 6% sucrose, 2 μg/ml leupeptin and 2 μg/ml antipain. The yield of recombinant proteins was 1.2 mg and 3.1 mg of homogenous adenine deaminase and SAH nucleosidase, respectively, per 200 ml of E. coli culture. The purified enzymes were aliquoted and stored at −70°C.

A preliminary SDS-PAGE analysis of the peak activity fractions from the Reactive Red 120-agarose purification step revealed the presence of very faint protein bands with a low protein content, consequently the fractions were 15-fold concentrated in a Vivaspin-500 ultrafiltration device and reanalyzed by SDS-PAGE. All bands present in the concentrated fractions were cut from a 10% gel and digested with trypsin. In-gel digestions of the peptides were performed as described (Shevchenko et al., 2006). Peptides were analyzed by nanoUPLC-tandem mass spectrometry employing Acquity nanoUPLC coupled with a Synapt G2 HDMS Q-TOF mass spectrometer (Waters, Milford, USA) fitted with a nanospray source and working in MSˆE mode under default parameters. Briefly, the products of in-gel protein digestion were loaded onto a Waters Symmetry C18 trapping column (20 mm × 180 μm) coupled to the Waters BEH130 C18 UPLC column (250 mm × 75 μm). The peptides were eluted from these columns in a 1–85% gradient of acetonitrile in water (both containing 0.1% formic acid) at a flow rate of 0.3 µl/min. The peptides were directly eluted into the mass spectrometer. Data were acquired and analyzed using MassLynx 4.1 software (Waters, USA) and ProteinLynx Global Server 2.4 software (PLGS, Waters, USA) with a False Discovery Rate ≤ 4%. To identify actin-specific histidine N-methyltransferase, the complete rat (Rattus norvegicus) reference proteome was downloaded from the NCBI Protein database, randomized and used as a databank for the MS/MS software.

Human-derived HAP1 cells (Horizon Discovery, Waterbeach, UK) have a haploid karyotype except for a disomy of chromosome 8 (25, XY,+8, Ph+) and are therefore considered near-haploid. The cell line expresses Setd3 gene at the level of 32.04 transcripts per kilobase million (https://www.horizondiscovery.com/human-setd3-knockout-cell-line-hzghc84193). Since Setd3 gene is on chromosome 14, all generated Setd3-defficent HAP1 cells will show the complete loss of SETD3 expression. The parental cells were therefore used to create HAP1 cell lines that are deficient in SETD3 by using CRISPR/Cas9 technology.

Two CRISPR/Cas9 constructs were generated to target exon 4 of the human Setd3 gene by ligating two different sets of annealed primer pairs (CRISPR-hSETD3-s1 CACCGgttctgtcgagggttttgaaa, CRISPR-hSETD3-as1 AAACtttcaaaaccctcgacagaacC and CRISPR-hSETD3-s2 CACCGgaggcccatttcattagatc, CRISPR-hSETD3-as2 AAACgatctaatgaaatgggcctcC). The annealed primer pairs were ligated into the vector pSpCas9n(BB)−2A-Puro (PX462) V2.0 (a gift from F. Zhang, Massachusetts Institute of Technology; Addgene plasmid no. 62987) (Ran et al., 2013) digested by BbsI. Constructs were validated by sequencing (Beckman Coulter Genomics). HAP1 cells were cultured in IMDM (Iscove’s Modified Dulbecco’s Medium) containing 10% FBS, 2 mM L-glutamine and penicillin/streptomycin (Thermo-Life Technologies). Cells were transfected with the CRISPR constructs exactly as described in Collard et al., 2016. Genomic DNA from puromycin-resistant clones was used to PCR-amplify the regions encompassing the targeted site and the PCR products were sequenced to evaluate the gene modification present in each clone. The three SETD3-KO clones that were ultimately selected for further studies presented the following modifications: clone KO-A1, a 16 bp deletion after Phe71, which introduces a frame shift and a premature stop codon at position 75; clone KO-A3, a 4 bp deletion and 6 bp insertion that introduce a frame shift after Met83 and a premature stop codon at position 109; and clone KO-A5, a 18 bp deletion and 107 bp insertion after Trp85, with a premature stop codon at position 292.

The absence of SETD3 protein from each SETD3-KO cell lines was confirmed by western blotting performed as described previously (Drozak et al., 2013), with the use of rabbit polyclonal primary antibody against human SETD3 (ab174662, Abcam, UK) and a horseradish peroxidase-conjugated goat anti-rabbit secondary IgG antibody (AS09602, Agrisera, Vannas, Sweden).

HAP1 wildtype (WT-A1 and WT-A2) and SETD3-KO cells (KO-A1 and KO-A3) were cultured in IMDM containing 10% FBS, 2 mM L-glutamine and penicillin/streptomycin (Life Technologies), each in four 10 cm diameter plates. At confluence, the medium was rapidly removed, the plates were washed with 5 ml cold PBS and each of them scraped in 0.8 ml of buffer A (25 mM Hepes pH 7.4, 20 mM NaCl, 2.5 µg/ml leupeptin and 2.5 µg/ml antipain) and combined to collect the cells. The four soluble cell extracts were prepared by two cycles of freezing the cell suspension in liquid nitrogen and thawing before vortex-mixing and centrifuging for 15 min at 15,000 × g and 4°C to recover soluble proteins.

To purify the actin partially, the cell lysates were diluted 5-fold in buffer A (20 ml final volume) and loaded onto a 1 ml-HiTrap Q HP (GE Healthcare) equilibrated with the same buffer. The column was washed with 6 ml of buffer A and proteins were eluted with a linear NaCl gradient (20–750 mM in 20 ml buffer A). The elution fractions (1 ml) were analysed by SDS-PAGE followed by Coomassie blue staining and by western blotting using a rabbit anti-actin antibody (A2066, Sigma-Aldrich). The actin band of the richest fraction was cut out of the gel and digested with trypsin.

The resulting peptides were analysed by LC-MS/MS with a LTQ XL (ThermoFisher Scientific) equipped with a microflow ESI source interfaced with a Dionex Ultimate Plus Dual gradient pump, a Switchos column switching device, and a Famos Autosampler (Dionex, Amsterdam, The Netherlands) as described (Houddane et al., 2017). Raw data files were then analyzed by bioinformatics using Proteome Discoverer (Thermo Fisher Scientific). Peak lists were generated using the application spectrum selector in the Proteome Discoverer 1.4 package. From raw files, MS/MS spectra were exported as individual files in the .dta format with the following settings: peptide mass range, 400–3500 Da; minimal total ion intensity, 500; and minimal number of fragment ions, 12. The resulting peak lists were searched using SequestHT against a human protein database obtained from Uniprot and containing only actin sequences. The following parameters were used: trypsin was selected with semi-proteolytic cleavage only after lysine and arginine; the number of internal cleavage sites was set to 2; the mass tolerance for precursor and fragment ions was 1.0 Da; and the considered dynamic modifications were + 15.99 Da for oxidized methionine,+14.00 Da for methylation on histidine and + 42.00 Da for acetylation of N-terminus or lysines. Peptide spectral matches (PSM) were filtered using charge-state versus cross-correlation scores (Xcorr) and methylation sites were validated manually.

To investigate whether the absence of SETD3 in D. melanogaster also prevented actin methylation, we prepared protein extracts by homogenizing approximately 20 larvae from wildtype or SETD3 KO flies (Tiebe et al., 2018) in 0.5 ml of 25 mM Hepes pH 7.5 with 2.5 µg/ml of leupeptin and antipain. After centrifugation (15 min at 16 000 ×g), the supernatant was recovered, protein concentration was measured (Bradford, 1976) and 10 µg of total proteins were digested with trypsin and used for LC-MS/MS analysis. Peptides were dissolved in solvent A (0.1% trifluoroacetic acid (TFA) in 2% acetonitrile (ACN)), directly loaded onto reversed-phase pre-column (Acclaim PepMap 100, Thermo Scientific) and eluted in backflush mode. Peptide (250 ng) separation was performed using a reversed-phase analytical column (Acclaim PepMap RSLC, 0.075 × 250 mm, Thermo Scientific) with a linear gradient of 4–27.5% solvent B (0.1% formic acid (FA) in 98% ACN) for 100 min, 27.5–40% solvent B for 10 min, 40–95% solvent B for 1 min and holding at 95% for the last 10 min at a constant flow rate of 300 nl/min on an EASY-nLC 1000 ultra-performance liquid chromatography (UPLC) system. The peptides were analyzed by an Orbitrap Fusion Lumos tribrid mass spectrometer (ThermoFisher Scientific). The peptides were subjected to NSI source followed by tandem mass spectrometry (MS/MS) in Fusion Lumos coupled online to the UPLC. Intact peptides were detected in the Orbitrap at a resolution of 120,000. Peptides were selected for MS/MS using a higher-energy C-trap dissociation (HCD) setting of 30; ion fragments were detected in the ion trap. A data-dependent procedure that alternated between one MS scan followed by 20 MS/MS scans was applied for the top 20 precursor ions above a threshold ion count of 5.0E3 in the MS survey scan with 30.0 s dynamic exclusion. The electrospray voltage applied was 2.1 kV. MS1 spectra were obtained with an automatic gain control (AGC) target of 4E5 ions and a maximum injection time of 50 ms, whereas MS2 spectra were acquired with an AGC target of 5E3 ions and a maximum injection time of 300 ms. For MS scans, the m/z scan range was 350 to 1500. The resulting MS/MS data were processed using the Sequest HT search engine within Proteome Discoverer 2.2 against a protein database containing Drosophila melanogaster sequences obtained from Swissprot (4810 entries, version 27 10 2017). Trypsin was specified as cleavage enzyme allowing up to two missed cleavages, four modifications per peptide and up to five charges. Mass error was set to 10 ppm for precursor ions and 0.6 Da for fragment ions. Oxidation on Met, carbamidomethyl on Cys and methylation on His (+14.015 Da) were considered as variable modifications. False discovery rate (FDR) was assessed using Percolator and thresholds for protein, peptide and modification site were specified at 1%.

Visualization of actin filaments was performed as described Domanski et al., 2016Domanski et al., 2016. Briefly, HAP1 cells, both control and Setd3-KO clones, were cultured in IMDM supplemented with 10% FBS, penicillin (100 U/ml) and streptomycin (100 µg/ml) at 37°C in a humidified atmosphere containing 5% CO₂. For confocal microscopy, 5 × 10⁵ cells were seeded on a 22 × 22 mm glass coverslip in a 35 mm plate. Either 24 or 48 hr after cell seeding, culture medium was removed and the cells were washed three times with PBS, fixed with 3% paraformaldehyde for 10 min at room temperature, permeabilized with 0.05% Triton X-100 in PBS for 3 min and washed again three times with PBS. F-actin was visualized by staining with TRITC-phalloidin (Sigma-Aldrich) in PBS for 1 hr at room temperature. Nuclei were localized with Hoechst 33342 dye (Sigma-Aldrich). After rinsing in PBS, coverslips were mounted using fluorescence mounting medium (DakoCytomation-Agilent, Santa Clara, USA). Samples were analyzed with the use of a Zeiss Axio Observer.Z1 LSM 700 confocal laser scanning microscope equipped with a Plan-Apochromat 63/1.40 Oil objective (Zeiss, Oberkochen, Germany). To detect the fluorescence of TRITC-phalloidin, a 555 nm excitation line was used, and for detecting blue fluorescence of Hoechst 33342, a 405 nm excitation line was employed. For 3D analysis, confocal Z-stacks comprising 10–13 optical slices were reconstructed into a 3D image using Zeiss Zen software with default settings.

For determination of the thermostability of actin protein present in cell-free lysates of HAP1 wildtype (WT-A2) and SETD3-KO cells (KO-A3), 40 µl of lysates (73.6 µg protein) were incubated for 5 min at 4, 30, 40, 43, 46, 49, 52, 55 and 60°C, followed by a 5 min incubation at 4°C. The denatured proteins were removed in the pellet following centrifugation (13,000 × g, 10 min) and the remaining non-denatured and still soluble proteins present in the supernatant were subsequently separated by SDS-PAGE. Soluble actin was detected by western blotting with the use of rabbit polyclonal primary antibody against vertebrate actin (A2066, Sigma-Aldrich, USA) and a horseradish peroxidase-conjugated goat anti-rabbit secondary IgG antibody (AS09602, Agrisera, Sweden). The quantitative analysis of the protein band corresponding to actin was performed by densitometry using Quantity One software (BioRad, Hercules, USA).

Sequences homologous to rat SETD3 protein (NCBI Reference Sequence: XM_002726774.2) were identified by Protein Blast searches. A phylogenetic analysis was performed on the Phylogeny.fr platform (www.phylogeny.fr) (Dereeper et al., 2008). Amino acid sequences were aligned with MUSCLE (v3.7) (Edgar, 2004). After alignment, ambiguous regions were removed with Gblocks (v0.91b) (Castresana, 2000). Phylogenetic trees were generated using the Phylogenetic estimation using Maximum Likelihood (PhyML) (Guindon and Gascuel, 2003) with the WAG model for amino acid substitution (Whelan and Goldman, 2001). The final tree was customized with the editing interface TreeDyn (Chevenet et al., 2006). A confidence level was assessed using the approximate likelihood ratio test (aLRT) (minimum of SH-like and χ²-based parametric) (Anisimova and Gascuel, 2006).

Protein concentration was determined spectrophotometrically according to Bradford (1976) using bovine γ-globulin as a standard. When appropriate, the His₆-tagged recombinant proteins were detected by western blot analysis, employing a mouse primary antibody against His₆-tag (27-4710-01, GE Healthcare, USA) and a horseradish peroxidase-conjugated goat anti-mouse antibody (A2554, Sigma-Aldrich, USA), as described previously (Drozak et al., 2013). All western-blotting analyses employed chemiluminescence and signals acquisition with Amersham Hyperfilm ECL, with the pattern of the prestained protein ladder being copied from the blotting membrane onto the film using a set of felt-tip pens.

V_max, K_M and k_cat for the methyltransferase activities of the studied enzymes were calculated with Prism 4.0 (GraphPad Software, La Jolla, USA) using a nonlinear regression.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

The authors would like to express their appreciation to Professor Jadwiga Bryla (University of Warsaw) who read the first draft of this article and provided numerous helpful suggestions.

© 2018, Kwiatkowski et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.