Lysine Ethylation by Histone Lysine Methyltransferases

Article abstract of DOI:10.1002/cbic.201900359

Biomedicinally important histone lysine methyltransferases (KMTs) catalyze the transfer of a methyl group from S-adenosylmethionine (AdoMet) cosubstrate to lysine residues in histones and other proteins. Herein, experimental and computational investigations on human KMT-catalyzed ethylation of histone peptides by using S-adenosylethionine (AdoEth) and Se-adenosylselenoethionine (AdoSeEth) cosubstrates are reported. MALDI-TOF MS experiments reveal that, unlike monomethyltransferases SETD7 and SETD8, methyltransferases G9a and G9a-like protein (GLP) do have the capacity to ethylate lysine residues in histone peptides, and that cosubstrates follow the efficiency trend AdoMet>AdoSeEth>AdoEth. G9a and GLP can also catalyze AdoSeEth-mediated ethylation of ornithine and produce histone peptides bearing lysine residues with different alkyl groups, such as H3K9meet and H3K9me2et. Molecular dynamics and free energy simulations based on quantum mechanics/molecular mechanics potential supported the experimental findings by providing an insight into the geometry and energetics of the enzymatic methyl/ethyl transfer process.

Full text of DOI:10.1002/cbic.201900359

Accepted Article
Title: Lysine Ethylation by Histone Lysine Methyltransferases
Authors: Abbas H. K. Al Temimi, Michael Martin, Qingxi Meng, Danny
C. Lenstra, Ping Qian, Hong Guo, Elmar Weinhold, and
Jasmin Mecinovic
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemBioChem 10.1002/cbic.201900359
Link to VoR: http://dx.doi.org/10.1002/cbic.201900359
A Journal of
www.chembiochem.org

10.1002/cbic.201900359
ChemBioChem
Lysine Ethylation by Histone Lysine Methyltransferases
Abbas H. K. Al Temimi,[a] Michael Martin,[b] Qingxi Meng,[c] Danny C.
Lenstra,[a] Ping Qian,[c] Hong Guo,*[d,e] Elmar Weinhold,*[b] and Jasmin
Mecinović*[a,f]
a Institute for Molecules and Materials, Radboud University, Heyendaalseweg 135, 6525 AJ
Nijmegen, The Netherlands.
b Institute of Organic Chemistry, RWTH Aachen University, Landoltweg 1, D-52056 Aachen,
Germany. elmar.weinhold@oc.rwth-aachen.de
c
Chemistry and Material Science Faculty, Shandong Agricultural University, Tai´an, P.R.
China.
d Department of Biochemistry and Cellular and Molecular Biology, University of Tennessee,
Knoxville, USA. hguo1@utk.edu
e UT/ORNL Center for Molecular Biophysics, Oak Ridge National Laboratory, USA.
f
Department of Physics, Chemistry and Pharmacy, University of Southern Denmark,
Campusvej 55, 5230 Odense, Denmark. mecinovic@sdu.dk
1
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
Biomedicinally important histone lysine methyltransferases (KMTs) catalyze the transfer
of methyl group from S-adenosylmethionine (AdoMet) cosubstrate to lysine residues in
histones and other proteins. Here we report experimental and computational
investigations on human KMT-catalyzed ethylation of histone peptides using S-
adenosylethionine (AdoEth) and Se-adenosylselenoethionine (AdoSeEth) cosubstrates.
MALDI-TOF MS experiments revealed that, unlike monomethyltransferases SETD7 and
SETD8, methyltransferases G9a and GLP do have a capacity to ethylate lysine residues
in histone peptides, and that cosubstrates follow the efficiency trend AdoMet > AdoSeEth
> AdoEth. G9a and GLP can also catalyze AdoSeEth-mediated ethylation of ornithine,
and produce histone peptides bearing lysine residues with different alkyl groups, such as
H3K9meet and H3K9me2et. Molecular dynamics and free energy simulations based on
quantum mechanics/molecular mechanics (QM/MM) potential supported the
experimental findings by providing an insight into the geometry and energetics of the
enzymatic methyl/ethyl transfer process.
Introduction
Histone proteins are subject of diverse posttranslational modifications (PTMs), including
methylation, acetylation, crotonylation, phosphorylation, citrullination and ubiquitination,
which regulate the activity of human genes via epigenetic mechanisms.[1] Methylation of lysine
residues in unstructured histone tails is associated with both gene activation and repression,
depending on the histone, methylation state and methylation site.[2] Histone lysine methylation
is catalyzed by S-adenosylmethionine (AdoMet or SAM)-dependent histone lysine
methyltransferases (KMTs) that install one (Kme), two (Kme2) or three (Kme3) methyl groups
on the Nε-amino group of lysine (Figure 1A).[3] Histone lysine methylation is removed by
flavin-dependent lysine specific demethylases and Fe(II)/2-oxoglutarate (2OG)-dependent
2
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
histone demethylases (KDMs),[4] and recognized by a large number of Nε-methyllysine binding
epigenetic reader proteins,[5] which collectively spread the epigenetic landscape of
posttranslational modifications.
Histone KMTs contain the conserved SET (Su(var)3-9, enhancer-of-zeste, trithorax)
domain responsible for the enzymatic activity; DOT1L is the only member of the histone KMT
family, known to date, that does not contain the SET domain.[6] Structural analyses revealed
that KMTs possess distinct binding pockets for AdoMet cosubstrate and histone substrate
(Figure 1B).[6e] In the ternary complex, the nucleophilic Nε-amino group of lysine is well
aligned with the electrophilic methyl group of AdoMet for an efficient SN2 reaction that takes
place in a narrow hydrophobic channel typically comprised of side chains of Tyr and Phe
residues (Figure 1B). The presence of Tyr and Phe in the active sites of KMTs appears to define
the methylation state of the product; Tyr to Phe substitutions result in the formation of higher
methylation states of lysine.[6e] The target lysine needs to be deprotonated for nucleophilic
attack and an active site Tyr may also be responsible for deprotonation of the protonated lysine,
although a water channel has also been suggested to play a role as a general base.[7]
Despite recent success in structural, mechanistic and inhibition studies on histone lysine
methyltransferases,[3a, 8] the biocatalytic potential of KMTs remains to be established.[9]
Enzymatic assays revealed that human KMTs exhibit a high degree of specificity for
methylation of lysine analogues that differ in stereochemistry, side chain length and main
chain.[10] One the other hand, KMTs appear to have a limited ability to catalyze other
alkylations of histones, including transfer of allyl, propargyl and larger alkyl groups from
AdoMet analogues bearing methyl group replacements.[11] Here we report investigations on
KMT-catalyzed ethylation of histone peptides employing S-adenosylethionine (AdoEth) and
Se-adenosylselenoethionine (AdoSeEth) cosubstrates (Figure 1C).
3
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
Results and Discussion
Analogues of AdoMet with methyl group replacements, including AdoEth and
AdoSeEth, can be enzymatically synthesized from -methionine derivatives and adenosine
L
triphosphate (ATP) using methionine adenosyltransferases (MAT) from different
organisms.[12] A pronounced product inhibition of the MAT enzymes, however, often limits
the synthesis to small amounts with isolated enzymes.[13] Larger amounts of cosubstrate
analogues can be obtained by chemical synthesis (Scheme 1). S-Adenosylhomocysteine
(AdoHcy)[14] or Se-adenosylselenohomocysteine (AdoSeHcy)[11a, b, h] are typically reacted with
alkylating agents under slightly acidic conditions using a mixture of formic and acetic acid.
These conditions guide regioselective alkylation of the sulfur or selenium atom because all
other nucleophilic positions are transiently protected by protonation. During the synthesis of
the more reactive selenonium analogue AdoSeEth, we noticed many byproducts. Fortunately,
formation of these byproducts could be efficiently suppressed by adding water to the mixture
of formic and acetic acid (Scheme 1).
Generally, alkylations of AdoHcy under acidic conditions lead to both diastereoisomers
(epimers) at sulfur in almost equal amounts.[14a] In the case of AdoEth the two epimers formed
in a 45:55 ratio (S-epimer to R-epimer). Both epimers were separated by reverse-phase HPLC
and only the S-epimer (corresponding to the biologically active S-epimer of AdoMet) was used
in this study. However, in the case of AdoSeEth separation of both epimers by reverse-phase
HPLC was not possible and AdoSeEth was used as an epimeric mixture.
We then carried out comparative enzymatic assays for KMT-catalyzed methylation
(using AdoMet) and ethylation (using AdoEth and AdoSeEth) of synthetic histone peptides
using MALDI-TOF MS, as recently described;[10a, b] histone H31-15 was used for studies with
SETD7 (also known as KMT7), G9a (also known as KMT1C and EHMT2) and GLP (also
4
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
known as KMT1D and EHMT1), and histone H413-27 was used for studies with SETD8 (also
known as KMT5A). MALDI-TOF MS data confirmed that human KMTs catalyze nearly
quantitative methylation of histone peptides in the presence of AdoMet: H3K4me, H4K20me,
H3K9me3 and H3K9me3 were formed in the presence of SETD7, SETD8, G9a and GLP,
respectively (Figure 2, top panels). Unlike monomethylation, SETD7 and SETD8 did not
catalyze ethylation of H3K4 and H4K20 with AdoEth or more reactive AdoSeEth within
detection limits (Figure 2A–B and Figures S1–S2). Human G9a and GLP, however, were able
to catalyze ethylation of H3K9, with AdoSeEth being a superior ethylation agent over AdoEth
(Figure 2C–D). Although G9a and GLP catalyze di- and trimethylation of H3K9, both enzymes
were only able to catalyze monoethylation of H3K9, with no di- and triethylation detected
(Figure 2C–D). A longer incubation time (5 h) with AdoEth led to a formation of 56% and 30%
of H3K9et by G9a and GLP, respectively (Figure S3), whereas almost complete (87%)
formation of H3K9et was observed after 5 h when AdoSeEth and G9a were used (Figure S4).
Control experiments in the absence of AdoEth/AdoSeEth or G9a/GLP verified that ethylation
reactions are KMT-catalyzed (Figures S5–S6).
Having shown that G9a and GLP have an ability to catalyze monoethylation of H3K9,
we next investigated potential enzymatic ethylation of biologically important methylated
histones H3K9me and H3K9me2. G9a and GLP both poorly catalyzed ethylation of H3K9me
(traces detected) in the presence of AdoEth, even upon prolonged incubation (Figure 3A–B
and Figure S7). Both enzymes, however, produced detectable amounts of H3K9meet in the
presence of more reactive AdoSeEth (Figure 3A–B and Figure S8). G9a, in particular,
produced significant amounts (55% after 3 hours and 75% after 5 hours) of H3K9meet (Figure
3A–B and Figure S8). An observation that G9a and GLP have a capacity to catalyze the
formation of H3K9meet is interesting, because functionally related histone lysine demethylases
PHF8, FBXL11 and JMJD2E were found to catalyze removal of methyl and ethyl groups in
5
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
H3K9meet, thus producing an unmodified H3K9 (a substrate for G9a and GLP).[15] Moreover,
our MALDI-TOF MS assays revealed that G9a and GLP also catalyze ethylation of H3K9me2,
producing approximately 25% of bulky H3K9me2et in the presence of AdoSeEth; only traces
of H3K9me2et were observed when AdoEth was used as a cosubstrate (Figure 3C–D).
Prolonged incubation (5 h at 37 °C) led to increased amounts (41%) of H3K9me2et in the
presence of G9a and AdoSeEth, whereas AdoEth did not enhance formation of the trialkylated
product (Figures S9–S10). Control experiments in the absence of G9a/GLP or AdoSeEth
showed no ethylation of H3K9me and H3K9me2, implying that the reactions are KMT-
catalyzed and that the ethyl group in the H3K9meet and H3K9me2et product derives from
AdoSeEth cosubstrate (Figures S11–S12).
Despite the fact that AdoSeEth and AdoEth can both act as ethylating agents in
enzymatic assays, they do exhibit significant differences with respect to reactivity. In analogy
with AdoMet/AdoSeMet,[11b, 16] AdoSeEth appears to be more reactive (i.e. better alkylation
agent) than AdoEth. One notable difference between the two molecules is the bond length; the
C-Se bond is longer (2.0 Å) than the C-S bond (1.8 Å) making the selenonium analogues more
reactive. Due to longer C-Se bond and higher reactivity of AdoSeEth, we hypothesized that
KMTs might have an ability to catalyze ethylation of ornithine, the one-methylene shorter
lysine analogue. In line with our earlier observation,[10b] we observed that G9a and GLP do not
methylate H3Orn9, in the presence of AdoMet, within limits of detection (Figure 4, top panels).
Similarly, no G9a/GLP-catalyzed ethylation of H3Orn9 was observed when AdoEth was used
as a cosubstrate, even upon longer incubation (Figure 4, middle panels and Figure S13).
Interestingly, our MALDI-TOF MS data showed that G9a and GLP predominantly catalyze
monoethylation of H3Orn9 in the presence of AdoSeEth; as in the case of H3K9, no
diethylation of H3Orn9 was observed (Figure 4, bottom panels). A longer incubation time (5 h
at 37 °C) led to nearly complete (96%) and significant (78%) formation of H3Orn9et with G9a
6
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
and GLP, respectively (Figure S14). Controls without G9a/GLP or AdoSeEth again verified
that ethylation reactions are KMT-catalyzed (Figure S15). We also examined potential
G9a/GLP-catalyzed ethylation of H3Dab9 peptide that possesses the two-methylene shorter
lysine analogue 2,4-diaminobutyric acid (Dab). However, we did not detect any ethylated
products in the presence of AdoEth and AdoSeEth within detection limits (Figure S16–S17).
Finally, we also carried out enzymatic assays with ornithine-containing histone peptides
H3Orn4 and H4Orn20 with human SETD7 and SETD8. Unlike monomethylation of H3K4,
human SETD7 did not catalyze monoethylation of H3Orn4 in the presence of AdoEth or
AdoSeEth (Figure S18). Similarly, despite high degree of monomethylation of H4K20 by
SETD8, the enzyme did not yield any H4Orn20et in the presence of AdoEth or AdoSeEth
(Figure S19). This is in line with the absence of ethylation of H3K4 and H4K20 with AdoEth
and AdoSeEth as discussed above (Figure 2A–B). Collectively, our enzymatic assays revealed
that human G9a and GLP possess a biocatalytic potential for ethylation of the shorter ornithine
residue.
To obtain a better understanding of the G9a- and GLP-catalyzed ethylation of H3K9 and
H3Orn9 we performed kinetic experiments with different cosubstrate concentrations (Table 1,
Figures S20 and S21). However, G9a- and GLP-catalyzed ethylation reactions with AdoEth
were so slow that no multiple turnovers were obtained within the extended reaction time.
Comparing the single turnover rate constant k = 0.046 min-1 of G9a in the presence of saturating
AdoEth concentrations with kcat = 11.6 min-1 obtained with AdoMet under multiple turnover
(steady state) conditions indicates that alkylation with the natural cosubstrate is at least 200-
fold faster. Very similar kinetics results were obtained with GLP. Such a two to three orders
reduced activity with AdoEth compared with AdoMet was also observed with DNA
MTases,[14a] and can be attributed to an increased steric strain in the SN2-type transition state
for ethyl compared to methyl transfer. With AdoSeEth the ethylation rate under saturating
7
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
cosubstrate concentrations is about 3-4 times faster for both G9a and GLP, however, the true
rate enhancement upon going from AdoEth to AdoSeEth might be even larger because
AdoSeEth was employed as an epimeric mixture at the selenonium center (separation of the
epimers by reverse-phase HPLC was not possible, vide supra) and the epimer with the non-
natural R-configuration might act as an inhibitor as observed for (R)-AdoMet.[17] Furthermore,
the ethylation rate of the side chain shortened substrate H3Orn9 with AdoSeEth is only reduced
by 50% compared to that of the H3K9 substrate with the natural target amino acid.
We then carried out quantum mechanics/molecular mechanics (QM/MM) investigations
to rationalize experimental observations on the KMT-catalyzed ethylation of lysine residues.
Because the parameters for selenium are still not available in the semi-empirical QM DFTB3
method, only the simulations for the ethylation reactions involving AdoEth were performed in
this work. The average active-site structures of the reactant complexes for the first methylation
and first ethylation in SETD8 are compared in Figure 5 (active-site structures near the transition
state are shown in Figure S22); the distribution map of r(CM-Nε)/r(CM1-Nε) and θ is also given
in each case. As can be seen from these figures, the alignment of the electron lone pair on Nε
of the target lysine with the transferable ethyl groups (Figure 5B) is significant worse than that
with the transferable methyl group (Figure 5A). For instance, the average distance between Nε
of lysine and CM1 is 4.3 Å in Figure 5B, while it is 3.4 Å between Nε and CM in Figure 5A. The
poor alignment between the ethyl donor and acceptor positions is also reflected in the
distribution map. The free-energy profiles for the first methylation and first ethylation reactions
in SETD8 are given in Figure 5C. Consistent with the structures of the reactant complexes, the
free energy barrier for the ethylation reaction (22.7 kcal mol-1) is considerably higher than that
of methylation (19.4 kcal mol-1). The simulation results indicate that although SETD8 can
catalyze the monomethylation of H4K20, it may not be able to do so for the ethylation reaction,
consistent with the experimental data (Figure 2B).
8
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
The average active-site structures of the reactant complexes for the first and second ethyl
transfers in GLP are given in Figure 6A–B, respectively (active-site structures near the
transition state are shown in Figure S23). Figure 6A–B shows that the alignment of the electron
lone pair on Nε of the target lysine with the transferable ethyl group in GLP is significantly
better for the first ethyl transfer (with a relatively shorter average CM1···Nε distance of 4.08 Å
and more populations of the near attack conformations) compared to that for the second ethyl
transfer (with the average CM1···Nε distance of 4.56 Å). This is in contrast with the cases for
the first and second methyl transfers in GLP where the target lysine and methyl lysine can be
well aligned, respectively, with the transferable methyl group (Figure S24). The free-energy
profiles for the first and second ethylation reactions in Figure 6C demonstrate that while GLP
may catalyze the first ethylation reaction (with the free energy barrier of 18.3 kcal mol-1 that is
higher than 17.0 kcal mol-1 for the first methylation reaction10b), it is unlikely to be able to
catalyze the second ethylation reaction using AdoEth due to the fact that the free energy barrier
becomes significantly higher (23.4 kcal mol-1). These results are in agreement with the
observed monoethylated product H3K9et and the lack of the diethylated product H3K9et2 in
MALDI-TOF assays, as shown Figure 2D.
Conclusions
We have demonstrated that human histone lysine methyltransferases G9a and GLP have a
capacity to catalyze monoethylation of H3K9 in the presence of AdoEth and AdoSeEth
cosubstrates. Enzymatic assays revealed that AdoSeEth is a superior ethylating agent over
AdoEth, but comparatively poorer cosubstrate than the natural AdoMet. An ability of AdoEth
to act as an ethylating agent for histone ethylation in cells[18] and in vitro as investigated here
might have some biological relevance, because its precursor ethionine is a toxic compound that
can be converted to AdoEth by eukaryotic MAT enzymes.[19] Computational work unveiled
9
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
that the molecular origin for the more efficient enzymatic methylation over ethylation of lysine
residues in histones lies in more optimal alignment of smaller methyl group of AdoMet when
compared to larger ethyl group of AdoEth. Our examinations also revealed that G9a and GLP
catalyze ethylation of histone H3 bearing biologically relevant methylated lysine residues
(H3K9me and H3K9me2), and histone H3 peptide possessing one-methylene shorter ornithine
(H3Orn9) in the presence of AdoSeEth. Collectively, this work highlights a biocatalytic
potential of selected human KMTs and expands the substrate scope for KMT-catalyzed
alkylation of histones. It is envisioned that this work, along with recent investigations of KMT-
catalyzed alkylation of proteins, including histones, will advance our basic understanding of
KMT catalysis.
Experimental Section
Materials
Se-Adenosylselenohomocysteine
(AdoSeHcy)[11h]
and
5'-[(S)-[(3S)-3-amino-3-
carboxypropyl]ethylsulfonio]-5'-deoxyadenosine (S-adenosylethionine, AdoEth)[14a] were
prepared as described before.
Synthesis
Ethyl triflate. Ethyl triflate was obtained by a slightly modified literature procedure.[20]
Polyvinylpyridine (1,21 g) was suspended in dichloromethane (27 mL) with ice cooling.
Trifluoromethanesulfonic anhydride (1,55 g, 5,49 mmol) was added dropwise and then ethanol
(240 mg, 5.21 mmol) in dichloromethane (2.75 mL) was added within 2 min. The ice bath was
removed and stirring continued for another 10 min at room temperature. The solid was filtered
off, and the solvent was removed under reduced pressure to yield a slightly brown liquid (754
10
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
1
mg, 4.23 mmol, 81 %). H NMR (300 MHz, CDCl3): δ = 1.45 (t, J = 7.0 Hz, 3H; H2), 4.55
ppm (q, J = 7.2 Hz, 2H; H1).
5'-[(Se)[(3S)-3-Amino-3-carboxypropyl]ethylselenonio]-5'-deoxyadenosine (Se-adenosyl-
selenoethionine, AdoSeEth). Se-Adenosylselenohomocysteine (AdoSeHcy) (20 mg, 46,4
µmol) was dissolved in a 1:1 mixture of formic and acetic acid (1,95 mL) and water (0.98 mL)
was added. The solution was supplemented with ethyl triflate (1,49 g, 8.37 mmol) and the
reaction mixture was stirred at room temperature for 1 h. The reaction was quenched by adding
water (6 mL) and the aqueous phase was extracted three times with diethylether (7.5 mL each
time). Purification of the product in the aqueous phase was performed by preparative reverse-
phase HPLC (Prontosil-ODS 5 µm, 120 Å, 250 × 20 mm, Bischoff, Leonberg, Germany).
Compounds were eluted with methanol (linear gradients from 0 % to 7.8 % in 15 min and to
78 % in 5 min) in aqueous trifluoroacetic acid (0.01%) and a flow of 10 mL/min. Compounds
were detected at 260 and 272 nm. The two epimers (at selenium) both eluted with a retention
time around 9.9 min and could not be separated. Product containing fractions were collected
and the solvents were removed by lyophilization. The remaining solid was dissolved in water
(1.5 mL) and the yield 8.32 mg (18.1 µmol, 39 %) was determined by UV spectroscopy (ε260
= 15,400 L mol-1 cm-1). ESI-MS m/z (relative intensity): 461.1 (100) [M]+, 360.2 (8) [5’-
ethylseleno-5’-deoxyadenosine + H]+, 250.4 (17) [M - ethionine]+, 136,2 (9) [adenine + H]+.
Expression and purification of KMTs
The four wild-type human histone lysine methyltransferases (SETD7, SETD8, G9a and GLP)
were expressed and purified as earlier described.[10a, b] The wild-type sequences of human
methyltransferases were as follows: SETD8 (aa 186–352), SETD7 (aa 1-366), G9a (aa 913-
1193), and GLP (aa 951-1235). Briefly, overnight cultures of E.coli Rosetta BL21
(DE3)pLysS harboring expression plasmids were made by growing at 37 °C with shaking for
11
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
18 h in LB media supplemented with kanamycin and chloramphenicol
. The expression was
induced at OD600 (approximately 0.5-0.6) by adding isopropyl-beta- -thiogalactopyranoside
D
(IPTG) and shaking continued at 16 °C for 20 h. Cells were harvested by centrifugation at
4 °C for 15 min, and the cell pellets were resuspended in lysis buffer. Cells were lysed using
a Soniprep 150 sonicator for 20 s (8 times) with 90 s intervals keeping the cells chilled in
an ice water bath at all times. After centrifugation the supernatant was loaded onto the Ni-
charged His-tag binding resin that was equilibrated with column buffer. Resins were washed
thoroughly with column buffer, followed by washing buffer and protein was eluted with elution
buffer using a linear gradient concentration of imidazole. The protein was then applied to
size exclusion chromatography (SEC) using a Superdex 75 column (GE Healthcare).
Purified proteins were concentrated employing Amicon Ultra Centrifugal Filter Units
(Millipore) with suitable molecular weight cutoffs (10 kD). Protein concentration was
determined employing the Nanodrop DeNovix DS-11 spectrophotometer and the purity was
monitored by SDS-PAGE on a 4–15 % gradient polyacrylamide gel (Bio-Rad). Enzymes
were aliquoted and stored at -80 °C for future use.
Enzymatic assays by MALDI-TOF MS
The reactions were carried out in total volume of 25 µL in an Eppendorf vial using
thermomixer. A typical enzymatic assay included histone peptides (40 µM), cosubstrate
AdoMet (200 µM with SETD8 and SETD7; 500 µM with GLP and G9a), AdoEth (1 mM) or
AdoSeEth (1 mM), and KMT enzyme (2 µM) in a reaction buffer of 50 mM Tris-HCl at pH
8.0. The reactions were incubated at 37 °C and aliquots were taken from the reaction vial at
different time points (1 h, 3 h, and 5 h) to measure the conversion of histone peptide substrates
to alkylated products. The reaction was stopped by mixing 3 µL of the reaction mixture with 3
µL of MeOH. An aliquot (3 µL) of these samples were directly mixed onto the MALDI target
12
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
plate using 3ꢀµL of α-cyano-4-hydroxycinnamic acid (CHCA) matrix (5ꢀmgꢀmL−1 in 50%
(v/v) acetonitrile/water) and air-dried. Peptide substrates masses were measured using the
positive-ion reflector mode. Full mass scans were acquired in the m/z range of 500-4000. Each
mass spectrum was generated from the data deriving from 3–5 single laser shots in 200-shot
steps from different positions of the sample spot. All the spectra were manually acquired by
using a Microflex mass spectrometer and the FlexControl software and the data were annotated
employing FlexAnalysis software (Bruker Daltonics, Germany). The methylation and
ethylation species were observed as follows; monomethylation (+ 14 Da), dimethylation (+ 28
Da), trimethylation (+ 42 Da), and monoethylation (+ 28 Da). All methylation and ethylation
experiments were performed in replicate.
Enzyme kinetics analyses
The kinetics parameters (k, K) were determined by incubating G9a or GLP (2 µM for AdoEth;
1 µM for AdoSeEth) and histone peptide (25 µM 15-mer H3K9 or H3Orn9) in 50 mM TRIS
pH 8.0 in the presence of various concentrations of AdoEth (0-250 µM) or AdoSeEth (0-250
µM). The reactions (final volume = 20 µL) were incubated at 37 °C (700 rpm) for 15 min, after
which they were quenched by adding an equal volume of MeOH. For analysis by MALDI-
TOF MS, 2 µL of the quenched reaction mixture was mixed with 6 µL of saturated CHCA
solution (1:1 v/v in MeCN/ddH2O + 0.1% TFA). From this, 1 µL was spotted onto the MALDI
plate for crystallization. The enzymatic activity was determined using the peak areas (including
all isotopes) for each alkylation state. To obtain the kinetics parameters, the data were plotted
and fitted to the non-linear regression enzyme kinetics kcat function using GraphPad PRISM
software (Figure S20). The kinetics profiles for AdoMet (Figure S21) were obtained by
incubation of G9a or GLP (50 nM), histone peptide (10 µM 15-mer H3K9) and AdoMet (0-15
13
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
µM) in 50 mM TRIS pH 8.0 buffer for 3 min at 37 °C (700 rpm). MALDI-TOF MS
measurements and data analysis was performed as for AdoEth and AdoSeEth described above.
QM/MM studies
QM/MM free energy (potential of mean force) and MD simulations were performed to study
the active-site dynamics of SETD8 and GLP and to calculate the free energy profiles of the
ethyl transfers from AdoEth to lysine and its ethylated form using the CHARMM program.[21]
The –CH2–CH2–S+(Et) –CH2– part of AdoEth and lysine/ethyl lysine chain were treated by
QM and the rest of the system by MM. The link-atom approach[22] was applied to separate the
QM and MM regions. A modified TIP3P water model[23] was employed for the solvent, and
the stochastic boundary molecular dynamics method[24] was used for the QM/MM simulations.
The system was separated into a reaction zone and a reservoir region, and the reaction zone
was further divided into a reaction region and a buffer region. The reaction region was a sphere
with radius r of 20 Å, and the buffer region extended over 20 Å ≤ r ≤ 22 Å. The reference
center for partitioning the system was chosen to be the Nε atom of the target lysine. The
resulting systems contained around 5800 atoms for GLP (or 5400 atoms for SETD8), including
about 700-900 water molecules. The DFTB3 method[24-25] implemented in CHARMM was
used for the QM atoms. The semi-empirical approach adopted here has been used previously
on a number of systems, and the results seem to be quite reasonable.[26] The all-hydrogen
CHARMM potential function (PARAM27)[27] was used for the MM atoms.
The initial coordinates for the reactant complexes of the methylation/ethylation were
based on the crystallographic complexes (PDB codes: 2BQZ and 3HNA for SETD8 and GLP,
respectively); a methyl/ethyl group was manually added to SAH to change it to
AdoMet/AdoEth and the methyl group on the methyl lysine was manually deleted to generate
the target lysine. The initial structures for the entire stochastic boundary systems were
14
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
optimized using the steepest descent (SD) and adopted-basis Newton-Raphson (ABNR)
methods. The systems were gradually heated from 50.0 to 298.15 K in 50 ps. A 1 fs time step
was used for integration of the equation of motion, and the coordinates were saved every 50 fs
for analyses. 1.5 ns QM/MM MD simulations were carried out for each of the reactant
complexes, and the similar approaches have been used previously.[7a, 28]
The umbrella sampling method[29] implemented in the CHARMM program along with
the Weighted Histogram Analysis Method (WHAM)[30] was applied to determine the changes
of the free energy (potential of mean force) as a function of the reaction coordinate for the
methyl/ethyl transfer from AdoMet/AdoEth to the target lysine/ethyl lysine in SETD8 and in
GLP. The reaction coordinate was defined as a linear combination of r(CM−Sδ) and r(CM−Nε)
for methylation [R = r(CM−Sδ) – r(CM−Nε)] or r(CM1−Nε) and r(CM1−Sδ) for ethylation [R =
r(CM1-Sδ) – r(CM1-Nε)] (see Figure 5 for the atom designation). 30 windows were used and for
each window 50 ps production runs were performed after 50 ps equilibration. The force
constants of the harmonic biasing potentials used in the PMF simulations were 50–400 kcal
mol–1 Å–2.
Keywords
Epigenetics, ethylation, histone, lysine, methyltransferase.
Acknowledgements
We thank the Netherlands Organization for Scientific Research (NWO, NCI-TA grant
731.015.202) for financial support. This work was supported in part by the Natural Science
Foundation of Shandong Province in China (No. ZR2017MB048 to P.Q.) and by the
Excellence Initiative of the German Federal and State Governments. We thank Professor
Martin Karplus for a gift of the CHARMM program.
15
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
Figure 1. Histone lysine methyltransferase (KMT)-catalyzed alkylation of histones. A) AdoMet-
mediated lysine methylation leading to Kme, Kme2 and Kme3. B) View from a crystal structure of GLP
(green) in complex with H3K9me (violet) and AdoHcy (yellow) (PDB ID: 3HNA). C) AdoEth- or
AdoSeEth-mediated lysine ethylation leading to Ket.
16
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
Scheme 1. Chemical synthesis of AdoEth[14a] and AdoSeEth.
17
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
Figure 2. MALDI-TOF MS assays showing methylation (top panel) and ethylation reactions (middle
and bottom panels) of A) H3K4 by SETD7; B) H4K20 by SETD8; C) H3K9 by G9a; D) H3K9 by GLP.
18
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
Figure 3. MALDI-TOF MS assays showing methylation (top panel) and ethylation reactions (middle
and bottom panels) of A) H3K9me by G9a; B) H3K9me by GLP; C) H3K9me2 by G9a; D) H3K9me2
by GLP.
19
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
Figure 4. MALDI-TOF MS assays showing methylation (top panel) and ethylation reactions (middle
and bottom panels) of A) H3Orn9 by G9a; B) H3Orn9 by GLP.
20
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
Figure 5. A) Representative active-site structure of the reactant complex of SETD8 for the first
methylation containing AdoMet and lysine (H4K20) obtained from the QM/MM MD simulations. The
distribution map on the right shows the alignment of NεH2 and the transferable methyl group in the
reactant complex in terms of the distance (r) between Nε and CM and the angle (θ) between the direction
of electron lone pair on Nε and the CM−S bond. SET8 is shown in sticks, and AdoMet and lysine are in
balls and sticks. Some average distances from the simulations are also given (in angstroms). B)
Representative active-site structure of the reactant complex of SETD8 for the first ethylation containing
AdoEth and lysine (H4K20) along with the r(CM1···Nε) and θ distribution map obtained from the
QM/MM MD simulations. C) Free energy (potential of mean force) profiles for the
methylation/ethylation reactions in SETD8 as a function of the reaction coordinate [R = r(CM−Sδ) –
r(CM−Nε)/R = r(CM1−Sδ) – r(CM1−Nε)]. The free energy profile for the first methylation: orange line
with a free energy barrier of 19.4 kcal mol-1 and the location of the transition state is at around 0.5. The
free energy profile for the first ethylation: blue line with a free energy barrier of 22.7 kcal mol-1.
21
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
Figure 6. A) Average structure obtained from molecular dynamics simulations for the first ethyl
transfer from AdoEth to H3K9 in GLP along with the distribution map during the MD simulations.
Some average distances of the potential hydrogen bonds between AdoEth/substrate and nearby residues
are shown with units of angstroms. B) Average structure obtained from molecular dynamics simulations
for the second ethyl transfer from AdoEth to H3K9et in GLP. C) Free energy changes for mono- and
diethylation of H3K9 in GLP. Monoethylation: orange line with a free energy barrier of 18.3 kcal mol-
1. Diethylation: blue line with a free energy barrier of 23.4 kcal mol-1.
22
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
Table 1. Rate constants k for G9a- and GLP-catalyzed methylation and ethylation of H3K9 and H3Orn9
histone peptides with saturating AdoMet, AdoEth and AdoSeEth concentrations. Reactions with
AdoMet were obtained under steady state conditions, while reactions with AdoEth and AdoSeEth were
too slow to reach steady state.
AdoMet1
H3K9
AdoEth2
H3K9
AdoSeEth3
H3K9 H3Orn9
k (min-1)
k (min-1)
11.6 0.61 0.046 0.005 0.21 0.01 0.12 0.01
G9a
9.90 0.44 0.058 0.003 0.16 0.01 0.10 0.003
GLP
1No methylation was observed with H3Orn9 and AdoMet. 2No ethylation was observed with H3Orn9
and AdoEth. 3AdoSeEth was employed as epimeric mixture at the selenonium center.
23
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
References
[1] a) A. J. Bannister and T. Kouzarides, Cell Res. 2011, 21, 381-395; b) T. Kouzarides, Cell
2007, 128, 693-705; c) B. D. Strahl and C. D. Allis, Nature 2000, 403, 41-45.
[2] J. C. Black, C. Van Rechem and J. R. Whetstine, Mol. Cell 2012, 48, 491-507.
[3] a) M. Luo, Chem. Rev. 2018, 118, 6656-6705; b) C. Martin and Y. Zhang, Nat. Rev. Mol.
Cell Biol. 2005, 6, 838-849.
[4] S. M. Kooistra and K. Helin, Nat. Rev. Mol. Cell Biol. 2012, 13, 297.
[5] S. D. Taverna, H. Li, A. J. Ruthenburg, C. D. Allis and D. J. Patel, Nat. Struct. Mol. Biol.
2007, 14, 1025-1040.
[6] a) J.-F. Couture, L. M. A. Dirk, J. S. Brunzelle, R. L. Houtz and R. C. Trievel, Proc. Natl.
Acad. Sci. USA 2008, 105, 20659-20664; b) P. A. Del Rizzo and R. C. Trievel, Epigenetics
2011, 6, 1059-1067; c) S. C. Dillon, X. Zhang, R. C. Trievel and X. Cheng, Genome Biol. 2005,
6, 227-227; d) H.-M. Herz, A. Garruss and A. Shilatifard, Trends Biochem. Sci. 2013, 38, 621-
639; e) M. Schapira, Curr. Chem. Genomics 2011, 5, 85-94.
[7] a) H.-B. Guo and H. Guo, Proc. Natl. Acad. Sci. USA 2007, 104, 8797-8802; b) X. Zhang
and T. C. Bruice, Proc. Natl. Acad. Sci. USA 2008, 105, 5728-5732.
[8] M. Luo, ACS Chem. Biol. 2012, 7, 443-463.
[9] M. R. Bennett, S. A. Shepherd, V. A. Cronin and J. Micklefield, Curr. Opin. Chem. Biol.
2017, 37, 97-106.
[10] a) R. Belle, A. H. K. Al Temimi, K. Kumar, B. J. G. E. Pieters, A. Tumber, J. E. Dunford,
C. Johansson, U. Oppermann, T. Brown, C. J. Schofield, R. J. Hopkinson, R. S. Paton, A.
Kawamura and J. Mecinović, Chem. Commun. 2017, 53, 13264-13267; b) A. H. K. Al Temimi,
Y. V. Reddy, P. B. White, H. Guo, P. Qian and J. Mecinović, Sci. Rep. 2017, 7, 16148; c) A.
H. K. Al Temimi, R. S. Teeuwen, V. Tran, A. J. Altunc, D. C. Lenstra, W. Ren, P. Qian, H.
Guo and J. Mecinović, Org. Biomol. Chem. 2019, 17, 5693-5697; d) A. H. K. Al Temimi, R.
van der Wekken-de Bruijne, G. Proietti, H. Guo, P. Qian and J. Mecinović, Bioconjugate Chem.
2019, 30, 1798-1804.
[11] a) I. R. Bothwell, K. Islam, Y. Chen, W. Zheng, G. Blum, H. Deng and M. Luo, J. Am.
Chem. Soc. 2012, 134, 14905-14912; b) I. R. Bothwell and M. Luo, Org. Lett. 2014, 16, 3056-
3059; c) K. Islam, I. Bothwell, Y. Chen, C. Sengelaub, R. Wang, H. Deng and M. Luo, J. Am.
Chem. Soc. 2012, 134, 5909-5915; d) K. Islam, W. Zheng, H. Yu, H. Deng and M. Luo, ACS
Chem. Biol. 2011, 6, 679-684; e) W. Peters, S. Willnow, M. Duisken, H. Kleine, T. Macherey,
K. E. Duncan, D. W. Litchfield, B. Lüscher and E. Weinhold, Angew. Chem. Int. Ed. 2010, 49,
5170-5173; f) T. Shimazu, J. Barjau, Y. Sohtome, M. Sodeoka and Y. Shinkai, PLoS ONE
2014, 9, e105394; g) R. Wang, K. Islam, Y. Liu, W. Zheng, H. Tang, N. Lailler, G. Blum, H.
Deng and M. Luo, J. Am. Chem. Soc. 2013, 135, 1048-1056; h) S. Willnow, M. Martin, B.
Lüscher and E. Weinhold, ChemBioChem 2012, 13, 1167-1173.
[12] a) F. Schlenk and J. L. Dainko, Biochim. Biophys. Acta 1975, 385, 312-323; b) S. Singh,
J. Zhang, T. D. Huber, M. Sunkara, K. Hurley, R. D. Goff, G. Wang, W. Zhang, C. Liu, J.
Rohr, S. G. Van Lanen, A. J. Morris and J. S. Thorson, Angew. Chem. Int. Ed. 2014, 53, 3965-
3969.
[13] J. R. Matos and C.-H. Wong, Bioorg. Chem. 1987, 15, 71-80.
24
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
[14] a) C. Dalhoff, G. Lukinavicius, S. Klimasauskas and E. Weinhold, Nat. Chem. Biol. 2006,
2, 31-32; b) C. Dalhoff, G. Lukinavičius, S. Klimašauskas and E. Weinhold, Nat. Prot. 2006,
1, 1879.
[15] R. J. Hopkinson, L. J. Walport, M. Münzel, N. R. Rose, T. J. Smart, A. Kawamura, T. D.
W. Claridge and C. J. Schofield, Angew. Chem. Int. Ed. 2013, 52, 7709-7713.
[16] D. F. Iwig, A. T. Grippe, T. A. McIntyre and S. J. Booker, Biochemistry 2004, 43, 13510-
13524.
[17] R. T. Borchardt and Y. S. Wu, J. Med. Chem. 1976, 19, 1099-1103.
[18] C. S. Baxter and P. Byvoet, Cancer Res. 1974, 34, 1424-1428.
[19] J. H. Alix, Microbiol. Rev. 1982, 46, 281-295.
[20] S. A. Ross, M. Pitié and B. Meunier, J. Chem. Soc., Perkin Trans. 1 2000, 571-574.
[21] B. R. Brooks, R. E. Bruccoleri, B. D. Olafson, D. J. States, S. Swaminathan and M.
Karplus, J. Comput. Chem. 1983, 4, 187-217.
[22] M. J. Field, P. A. Bash and M. Karplus, J. Comput. Chem. 1990, 11, 700-733.
[23] W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W. Impey and M. L. Klein, J. Chem.
Phys. 1983, 79, 926-935.
[24] C. L. Brooks, A. Brünger and M. Karplus, Biopolymers 1985, 24, 843-865.
[25] M. Elstner, D. Porezag, G. Jungnickel, J. Elsner, M. Haugk, T. Frauenheim, S. Suhai and
G. Seifert, Phys. Rev. B 1998, 58, 7260-7268.
[26] a) A. S. Christensen, T. Kubař, Q. Cui and M. Elstner, Chem. Rev. 2016, 116, 5301-5337;
b) Q. Cui, M. Elstner, E. Kaxiras, T. Frauenheim and M. Karplus, J. Phys. Chem. B 2001, 105,
569-585.
[27] A. D. MacKerell, D. Bashford, M. Bellott, R. L. Dunbrack, J. D. Evanseck, M. J. Field, S.
Fischer, J. Gao, H. Guo, S. Ha, D. Joseph-McCarthy, L. Kuchnir, K. Kuczera, F. T. K. Lau, C.
Mattos, S. Michnick, T. Ngo, D. T. Nguyen, B. Prodhom, W. E. Reiher, B. Roux, M.
Schlenkrich, J. C. Smith, R. Stote, J. Straub, M. Watanabe, J. Wiórkiewicz-Kuczera, D. Yin
and M. Karplus, J. Phys. Chem. B 1998, 102, 3586-3616.
[28] a) Y. Chu, Q. Xu and H. Guo, J. Chem. Theory Comput. 2010, 6, 1380-1389; b) Y. Chu,
J. Yao and H. Guo, PloS ONE 2012, 7, e37674; c) Q. Xu, Y.-z. Chu, H.-B. Guo, J. C. Smith
and H. Guo, Chem. Eur. J. 2009, 15, 12596-12599.
[29] G. M. Torrie and J. P. Valleau, Chem. Phys. Lett. 1974, 28, 578-581.
[30] S. Kumar, J. M. Rosenberg, D. Bouzida, R. H. Swendsen and P. A. Kollman, J. Comput.
Chem. 1992, 13, 1011-1021.
25
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900359
ChemBioChem
Table of Contents
Histone lysine methyltransferases, a biomedically important class of epigenetic enzymes that
catalyze metylation of histones in the presence of S-adenosylmethionine (AdoMet) cosubstrate,
have a limited capacity to catalyze ethylation of histones in the presence of S-adenosylethionine
(AdoEth) and Se-adenosylselenoethionine (AdoSeEth) cosubstrates.
26
This article is protected by copyright. All rights reserved.