A Selenium-Based Click AdoMet Analogue for Versatile Substrate Labeling with Wild-Type Protein Methyltransferases

Article abstract of DOI:10.1002/cbic.201100781

Protein methylation is catalyzed by S-adenosyl-L-methionine-dependent protein methyltransferases (MTases), and this posttranslational modification serves diverse cellular functions. Some MTases seem to exhibit broad substrate specificities and comprehensive methods for target profiling are needed. Here we report the synthesis of a new AdoMet analogue for enzymatic transfer of a small propargyl group and labeling of modified proteins through copper-catalyzed azide-alkyne cycloaddition (CuAAC). Replacement of sulfur by selenium strongly enhanced the stability of the progargylic cofactor, leading, in combination with better activation by the selenonium center, to higher enzymatic reactivity. A broad spectrum of wild-type protein MTases acting on lysine, arginine, and glutamine residues accept this cofactor and modified substrates can be efficiently labeled by CuAAC click chemistry.

Full text of DOI:10.1002/cbic.201100781

DOI: 10.1002/cbic.201100781
A Selenium-Based Click AdoMet Analogue for Versatile
Substrate Labeling with Wild-Type Protein
Methyltransferases
Sophie Willnow,[a, b] Michael Martin,[a] Bernhard Lꢀscher,[b] and Elmar Weinhold*[a]
Protein methylation is catalyzed by S-adenosyl-l-methionine-
dependent protein methyltransferases (MTases), and this post-
translational modification serves diverse cellular functions.
Some MTases seem to exhibit broad substrate specificities and
comprehensive methods for target profiling are needed. Here
we report the synthesis of a new AdoMet analogue for enzy-
matic transfer of a small propargyl group and labeling of mod-
ified proteins through copper-catalyzed azide–alkyne cyclo-
addition (CuAAC). Replacement of sulfur by selenium strongly
enhanced the stability of the progargylic cofactor, leading, in
combination with better activation by the selenonium center,
to higher enzymatic reactivity. A broad spectrum of wild-type
protein MTases acting on lysine, arginine, and glutamine resi-
dues accept this cofactor and modified substrates can be effi-
ciently labeled by CuAAC click chemistry.
Introduction
Proteins are post-translationally N-methylated at various amino
acids, predominantly at lysine and arginine, but also at gluta-
mine and histidine residues. These reactions are catalyzed by
protein methyltransferases (MTases), which naturally transfer
the activated methyl group from the ubiquitous cofactor S-ad-
enosyl-l-methionine (1, AdoMet or SAM; Scheme 1, below) to
their targets. The functions of some MTases and the conse-
quences of protein methylation have been established in the
past few years, but a complete insight into the significance of
protein methylation is still lacking.
tors RF1 and RF2, which can be methylated by PrmC (formerly
known as HemK), for example, show increased activities in
peptide chain release upon methylation.[16]
In the past it was assumed that every protein methylation
target has its own MTase. However, in recent years various sub-
strates have been observed for almost every single protein
MTase. In particular, in the case of histone MTases, many non-
histone substrates have been found. In addition to Set7/9,
which has its various substrates as mentioned above, human
H3K9 MTase G9a[17,18] and human arginine MTase PRMT1[19,20]
also methylate more than just their histone targets H3 and H4,
respectively. The substrates of G9a,[21] Set7/9,[22,23] and PRMT1[24]
seem to have consensus sequences surrounding the target
amino acid regions, so new substrates can be predicted
through sequence homologies.[24–26] This might not cover all
targets, however, as exemplified by Set7/9 substrates that lack
the consensus sequence.[14] A comprehensive and unbiased
method for identifying substrates of MTases would thus be
highly desirable.
In particular, methylation (and demethylation) of histone
proteins as part of epigenetic control mechanisms have been
intensively studied. Here, methylation of certain lysine and ar-
ginine residues, together with other post-translational modifi-
cations such as acetylation or phosphorylation, make up the
histone code.[1–3] These modifications are associated with tran-
scriptional regulation of genes, either directly through the ac-
cessibility of the adjacent DNA strand or through the binding
of methyl effectors that lead to downstream events.[4–6] The
yeast protein MTase Clr4, for example, catalyzes trimethylation
of Lys9 in histone H3 (H3K9), which is responsible for spread-
ing of heterochromatin and thus silencing of genes.[7,8] In con-
trast, methylation of H3K4, for example, by human MTase
Set7/9, is associated with transcriptional activity.[9]
Here we describe a selenium-based AdoMet analogue for
enzymatic transfer of a propargyl group suitable for subse-
quent click-labeling with biotin or fluorophores. The cofactor
analogue combines high stability (relative to the sulfur ana-
logue) and better activation by the selenonium center with
However, protein methylation can also have other functions
apart from transcription control.[10,11] Some methylations influ-
ence the activity of proteins: methylation of TAF10 by Set7/9,
for example, enhances its interaction with RNA-polymera-
se II.[12] In addition, it has been proposed that methyl groups
affect the stability of proteins.[13,14] One paradigm is the human
DNA MTase DNMT1, a Set7/9 substrate that is more prone to
proteasome-mediated degradation upon methylation.[15] Pro-
tein methylation is not restricted only to mammalian or eu-
karyotic proteins, however. The bacterial ribosomal release fac-
[a] S. Willnow, M. Martin, Prof. Dr. E. Weinhold
Institute of Organic Chemistry, RWTH Aachen University
Landoltweg 1, 52056 Aachen (Germany)
E-mail: elmar.weinhold@oc.rwth-aachen.de
[b] S. Willnow, Prof. Dr. B. Lꢀscher
Institute of Biochemistry and Molecular Biology
Medical School, RWTH Aachen University
Pauwelsstrasse 30, 52074 Aachen (Germany)
luescher@ rwth-aachen.de
ChemBioChem 0000, 00, 1 – 8
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

E. Weinhold et al.
Scheme 1. A) Structural formulas of the natural cofactor AdoMet (1) and the synthetic cofactors AdoEnYn (2), AdoYn (3), and SeAdoYn (4). B) Synthesis of the
selenium analogue SeAdoYn (4).
a small bio-orthogonal chemical group for transfer by several
different protein MTases.
CARM1, PRDM8, PRDM10, and PRDM16 were unable to accept
the cofactor analogue.[36]
In our search for new cofactor reactivities we synthesized
a propargylic AdoMet analogue in which sulfur is replaced by
selenium (Scheme 1B). This cofactor—SeAdoYn (4)—was ob-
tained by treating 5’-chloro-5’-deoxyadenosine with l-seleno-
homocysteine[37] under basic conditions. The product (Se-ade-
nosyl-l-selenohomocysteine) was then selectively alkylated at
selenium with propargyl triflate in a formic acid/acetic acid/
water (1:1:1) mixture. The acidic conditions are vital for proto-
nation and hence protection of all nucleophilic centers in the
molecule except for selenium.[38] SeAdoYn (4) was purified by
reversed-phase HPLC, but the epimers at the selenonium
center could not be separated.
Results and Discussion
We recently reported the use of the synthetic double-activated
AdoMet analogue AdoEnYn (2, Scheme 1A) for labeling pro-
teins through the action of protein MTases (Figure 1A).[27] Here,
the methyl group of AdoMet (1) is replaced by a pent-2-en-4-
ynyl side chain. The double bond in the b-position to the sulfo-
nium center compensates unfavorable steric effects in the SN2-
like transition by conjugative stabilization[28] and the terminal
alkyne serves as a reactive functionality for bio-orthogonal la-
beling through a copper-catalyzed azide–alkyne cycloaddition
(CuAAC) click reaction.[29,30] With this two-step labeling ap-
proach (Figure 1A) we were able to biotinylate the histone
substrates of fungal Dim-5 and human MLL (immunoprecipitat-
ed and in the ASH complex) specifically.[27] We also succeeded
in labeling RNA by use of tRNA MTase TRM1.[31]
Interestingly, the replacement of the sulfonium center by
a selenonium center had a very positive impact on cofactor
stability. In general, AdoMet (1) and its analogues are not very
stable molecules; three major decay pathways, depending on
the pH, have been described for AdoMet (1).[39] At neutral or
slightly basic pH an intramolecular nucleophilic attack of the
carboxylate onto Cg dominates, resulting in the formation of
a-amino-g-butyrolactone together with the thioether 5’-meth-
ylthio-5’-deoxyadenosine.
We analyzed the decay of the propargylic AdoMet analogues
at pH 7.5 by reversed-phase HPLC and found that AdoYn (3)
very rapidly decays by another pathway leading to a hydrated
product (Figure 2A), which is consistent with another study.[33]
For AdoYn (3), the two epimers at the sulfonium center can be
separated by reversed-phase HPLC and the more slowly elut-
ing epimer has a half-life of about 5 min. Decay of the more
rapidly eluting epimer was not observable by HPLC because
the cofactor and its hydrated product had the same retention
time (data not shown). However, from mass analysis, the signal
of the hydrated product was detectable after 5 min. After 1 h
the signal for the cofactor was completely gone. The hydrated
products of the two epimers further decay slowly to the corre-
sponding thioether, indicating that hydration occurs at the
Upon expanding the scope of AdoEnYn (2), we found that
its side chain can also be transferred by the yeast MTase Clr4
(Figure 1B). However, we and others[32,33] found that AdoEnYn
(2) is not readily accepted by other protein MTases. This could
be due to steric hindrance by the pent-2-en-4-ynyl side chain,
which would entail five carbon atoms in the active sites in-
stead of a single carbon methyl group. We thus reduced the
size of the side chain and tested AdoYn (3),[33–36] with a prop-
argyl (prop-2-ynyl) group containing only three carbon atoms.
Here, the terminal triple bond not only enhances cofactor re-
activity but also serves as a chemical handle in subsequent
CuAAC reactions after being transferred to the target. Unfortu-
nately, AdoYn (3) is quite unstable,[27,33] and the labeling effi-
ciency with Clr4 is drastically reduced in relation to AdoEnYn
(2, Figure 1B). This might also explain why in a recent study
only the H3K9 MTase SETDB1 was active with AdoYn (3),
whereas the other protein MTases Set7/9, PRMT1, SMYD2,
&2
&
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 8
ÝÝ
These are not the final page numbers!

Selenium Cofactor for Methyltransferases
pronounced for labeling with Dim-5 at a more basic pH of 9.8.
No labeling occurred with AdoYn (3, not shown) whereas the
reaction with SeAdoYn (4) still took place, although to a lesser
extent than with Clr4 at pH 7.5 (Figure 1C). In addition to co-
factor stability, better activation by the selenonium center in
SeAdoYn (4) than by the sulfonium center in AdoYn (3) should
enhance enzymatic transfer rates of the propargylic side chain.
Such an increased reactivity had been observed before with
Se-adenosyl-l-selenomethionine (SeAdoMet), the selenium ana-
logue of AdoMet.[40] Substitution of sulfur by selenium results
in a small but significant increase in methyl group transfer rate
by catechol O-methyltransferase and cyclopropane fatty acid
synthase. SeAdoYn (4) also compared favorably with the bulki-
er and probably less reactive sulfonium analogue AdoEnYn (2)
and led to a stronger signal with Clr4 after 30 min (Figure 1B).
After 120 min both signals had increased, but now labeling
with AdoEnYn (2) was more pronounced than with SeAdoYn
(4). This is in line with a lower stability and faster decay of
SeAdoYn (4) whereas AdoEnYn (2) remains active with Clr4 for
a longer period of time.
To validate SeAdoYn (4) further, we tested another labeling
and detection method for the alkynylated proteins. Instead of
using an azide-derivatized biotin for the CuAAC, we labeled
histone H3 with an azide-derivatized tetramethylrhodamine
(TAMRA) fluorophore. This was readily synthesized from com-
mercially available NHS-TAMRA and a poly(ethylene glycol)
bearing an amino group at one end and an azido group at the
other (see the Experimental Section). The fluorescent-labeled
proteins can be detected with a UV transilluminator (Fig-
ure 1C). Incubation of histone H3 with Clr4 and SeAdoYn (4)
for 60 min and labeling by CuAAC resulted in the strongest
signal (lane 1). Histone H3 modified with Clr4 and AdoEnYn (2)
also showed a detectable signal, but with lower intensity
(lane 4). With Dim-5, the use of SeAdoYn (4) also resulted in
a stronger signal (lane 8) than that of AdoEnYn (2, lane 11), but
both signals were weaker than with Clr4. The weaker signals of
Dim-5-mediated labeling might also be due to some loss of
histone H3 at the higher pH, which is visible in the Coomassie-
stained gel.
Figure 1. A) Schematic overview of the two-step labeling procedure. After
MTase-catalyzed transfer of the activated side chains from AdoMet ana-
logues to protein substrates, the terminal alkynes react specifically with
azides containing biotin or TAMRA fluorophore by click chemistry. B) Com-
parison of labeling efficiencies for histone H3 by the H3K9 MTase Clr4 with
the cofactor analogues AdoEnYn (2), AdoYn (3), and SeAdoYn (4). The modi-
fied histones were coupled with azide-derivatized biotin prior to their analy-
sis by SDS-PAGE and Western blotting with use of peroxidase-conjugated
avidin. C) Direct comparison of labeling efficiencies for histone H3 by Clr4
and Dim-5 with AdoEnYn (2) and SeAdoYn (4) by in-gel fluorescence. After
60 min the alkynylated histones were coupled with azide-derivatized
TAMRA, precipitated, and analyzed by SDS-PAGE and fluorescence detection
(excitation at l=312 nm and emission at l=540 nm).
To investigate the utility of SeAdoYn (4) as a versatile
AdoMet analogue for substrate labeling we tested several ad-
ditional protein MTases acting on lysine, glutamine, and argi-
nine residues of a variety of proteins. The human histone H3K4
MTase Set7/9 (Figure 3A, lane 3), as well as the human histone
H3K9 MTase G9a (Figure 3B), were able to modify their histone
substrates. To verify that modification by Set7/9 is not restrict-
ed only to histone H3, we also labeled another substrate of
Set7/9: namely TAF10 (Figure 3A, lane 2). The site specificity of
labeling was investigated by use of the histone H3 variants
K4A and K9A. The H3K4 MTase Set7/9 was fully proficient in la-
beling the K9A variant but was unable to achieve significant la-
beling of the K4A variant lacking the target lysine residue (Fig-
ure 4A). This result demonstrates that the site specificity of the
enzyme is not altered with the AdoMet analogue SeAdoYn (4).
Similarly, the H3K9 MTase Dim-5 labeled the K4A variant with
intensity comparable to that of the wild-type substrate, where-
as no significant labeling was observed with the K9A variant
propargylic side chain. This fully explains the low labeling effi-
ciencies with AdoYn (3) observed in Figure 1B. In contrast,
there was no hydration product detectable even after 1 h for
the selenium analogue SeAdoYn (4). This cofactor only decays
to the selenoether, with a half-life of about 1.5 h (Figure 2B).
We tested the novel SeAdoYn (4) cofactor analogue with
Clr4, and labeling with the selenium analogue indeed resulted
in a much more intense signal than with the sulfur analogue
AdoYn (3, Figure 1B), which directly points to different cofac-
tor stabilities at pH 7.5. This difference becomes even more
ChemBioChem 0000, 00, 1 – 8
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
&
3
&
ÞÞ
These are not the final page numbers!

E. Weinhold et al.
Figure 2. A) Scheme for the decomposition of AdoYn (3) at pH 7.5 and reversed-phase HPLC analysis with indicated masses (m/z). B) The same for SeAdoYn
(4). No hydrated decomposition product is observed for the selenium compound.
modify its substrates: the riboso-
mal release factors RF1 and RF2
(Figure 3C). To complete our in-
vestigation, we studied the
human arginine MTase PRMT1
and found that the substrates
histone H4 and hnRNPK were
modified by PRMT1 (Figure 3D).
SeAdoYn (4) thus has a broad
scope and functions as a cofactor
for different members of major
classes of protein MTases. In ad-
dition, SeAdoYn (4) is not only
restricted to protein MTases but
also serves as a cofactor for DNA
MTases (data not shown) and
RNA MTases.[41]
Conclusions
The abilities of different protein
MTases to use the selenium-
Figure 3. A) and B) Analysis of human lysine MTases Set7/9 and G9a with SeAdoYn (4). Set7/9 labels histone H3
based AdoMet analogue SeAdo-
Yn (4) as cosubstrate are very
promising for broad applications
of protein labeling. The better
and TAF10. G9a labels histone H3. C) and D) Labeling with MTases other than lysine MTases. Bacterial glutamine
MTase PrmC labels RF1 and RF2. Human arginine MTase PRMT1 labels histone H4 and hnRNPK.
lacking the target lysine residue in this case (Figure 4B). Again,
this result reveals unchanged site specificity with SeAdoYn (4).
In addition to these lysine-specific MTases, the bacterial-glu-
tamine-specific MTase PrmC was able to use SeAdoYn (4) to
reactivity relative to AdoEnYn (2) can be attributed to the
smaller side chain and to less steric hindrance in the active
sites, as well as to the stronger activation by the selenonium
center than by the sulfonium center. Furthermore, SeAdoYn (4)
&4
&
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 8
ÝÝ
These are not the final page numbers!

Selenium Cofactor for Methyltransferases
Analytical HPLC was carried out with a Waters Breeze System
HPLC, a Waters 1225 pump system, and a Waters 2487 dual l
absorbance detector with Prontosil C-18 columns (250ꢂ4.5 mm,
5 mm, 120 ꢃ) and C-18 precolumns (8.0 mmꢂ4.0 mm, 5 mm, 120 ꢃ).
Preparative HPLC was carried out with the same HPLC system, but
with preparative Prontosil C-18 columns (300ꢂ20 mm, 5 mm,
120 ꢃ) and C-18 precolumns (30 mmꢂ20 mm, 5 mm, 120 ꢃ). Lyo-
philization of aqueous samples was performed with a Savant spd
111 V SpeedVac concentrator (Thermo Scientific) and a Savant refri-
gerated vapor trap (RVT 100). For TLC, alumina plates (ALUGRAM
Xtra SIL G/UV254, Macherey–Nagel) were used. UV absorption meas-
urements were performed with a Varian Cary 3E spectrometer.
1H NMR spectra were recorded with a Mercury 300 (300 MHz)
spectrometer (Varian). Chemical shifts (d) are reported relative to
tetramethylsilane (0.00 ppm) and coupling constants are reported
in Hz. Electrospray ionization mass spectra (positive ion mode)
were recorded with a LCQ DECA XP Plus instrument (Finnigan).
Se-Adenosyl-l-selenohomocysteine (SeAdoHcy):
A
solution
(0.52 mL) of sodium hydroxide (1m) and 5’-chloro-5’-deoxyadeno-
sine (78 mg, 0.27 mmol, 1 equiv) was added to l-selenohomocys-
teine [freshly made from l-selenomethionine (58 mg, 0.29 mmol,
1.1 equiv)]. After the mixture had been stirred for 1.5 h at 808C,
the orange-brown solution was brought to a slightly acidic pH
(pH 5) with acetic acid (100 mL) and the solvent was removed
under high vacuum. Residual solvent was coevaporated with an
ethanol/toluene mixture and the product was subsequently recrys-
tallized twice from methanol to give the product (59 mg,
0.14 mmol, 51%). Rf =0.35 (isopropanol/water/acetic acid 12:5:3);
1H NMR (300 MHz, D2O): d=2.13–2.23 (m, 2H; Hb), 2.53 (t, J=
7.7 Hz, 2H; Hg), 2.81–2.91 (m, 2H; H5’), 3.69 (t, J=6.3 Hz, 1H; Ha),
4.19–4.24 (m, 1H; H4’), 4.24–4.30 (m, 1H; H3’), 4.74 (t, J=5.1 Hz,
1H; H2’), 5.93 (d, J=5.2 Hz, 1H; H1’), 8.12 (s, 1H; H8), 8.22 ppm (s,
1H; H2); MS (ESI): m/z (%): 455.18 (16) [M+Na]+, 433.18 (10)
[M+H]+, 387.10 (2) [MÀCO2]+, 235.27 (100) [MÀselenomethio-
nine]+, 136.07 (5) [adenine+H]+.
Figure 4. Analysis of site specificity for labeling with SeAdoYn (4) and
A) Set7/9, or B) Dim-5. After incubation variously with histone H3 wild type
(WT), histone variant K9A, or variant K4A for the indicated times, protein
modification was evaluated by CuAAC with azide-derivatized biotin and
Western blotting with use of peroxidase-conjugated avidin.
is much more stable than AdoYn (3), which is hydrated too
quickly to enable efficient transfer of the propargylic side
chain by most MTases. Finally, in contrast to specific AdoMet
analogues with even larger side chains, which are only active
with mutated forms of MTases,[32,33] SeAdoYn (4) is active with
wild-type enzymes; this allows versatile substrate labeling.
Moreover SeAdoYn (4) does not affect the selectivities of the
enzymes. Target amino acids can be lysine, arginine, and gluta-
mine residues in different MTase substrates.
Prop-2-ynyl triflate: Propargyl triflate was obtained by a literature
procedure[43] with slight modifications. Polyvinylpyridine (3.06 g,
29.17 mmol, 2.1 equiv) was suspended in dichloromethane (66 mL)
with ice cooling. Trifluoromethanesulfonic anhydride (4.11 g,
14.58 mmol, 1.05 equiv) and then prop-2-yn-1-ol (0.78 g,
13.9 mmol, 1 equiv) in dichloromethane (5 mL) were added drop-
wise. The reaction was allowed to warm to room temperature and
stirred for another 10 min at room temperature. The solid was
filtered off, and the solvent was removed under reduced pressure
We intend to use SeAdoYn (4) to analyze the substrate spe-
cificities of protein MTases. This labeling method can be used
on protein arrays to label modified target proteins with fluoro-
phores. In addition, the covalent attachment of affinity probes,
such as biotin, should enable the direct isolation and identifi-
cation of MTase substrates from cell lysates. The use of this
novel AdoMet analogue in MTase substrate profiling should be
a very powerful tool for determination of the global functions
of individual MTases. This promises to result in better under-
standing of the functions of protein MTases in epigenetics and
beyond.
to yield
a
brown liquid (2.17 g, 11.5 mmol, 83%). 1H NMR
(300 MHz,CDCl3): d=2.74 (t, J=2.47 Hz, 1H; H3), 4.98 ppm (d, J=
2.48 Hz, 2H; H1).
5’-[(Se)[(3S)-3-Amino-3-carboxypropyl]prop-2-ynylselenonio]-5’-
deoxyadenosine (SeAdoYn, 4): SeAdoHcy (20 mg, 46.3 mmol) was
dissolved in an acetic acid/formic acid/water mixture (1:1:1, 2.7 mL)
and prop-2-ynyl triflate (1.56 g, 8.33 mmol, 180 equiv) was added.
The reaction mixture was stirred at room temperature for 15 min.
The mixture was diluted with water (4 mL) and extracted thrice
with diethyl ether. The crude product was purified by preparative
HPLC with use of a methanol gradient (0–7.8% in 15 min, 7.8–78%
in 5 min, and 78% for 5 min) in aqueous trifluoroacetic acid
(0.01%) with a flow of 10 mLminÀ1 and detection at 260 nm and
272 nm. The two epimers (at selenium) both eluted with a retention
time around 14.4 min and could not be separated. Cofactor-con-
taining fractions were collected and the solvents were removed by
lyophilization. The solid was dissolved in water (1 mL) and trifluoro-
acetic acid (0.65 mL) was added to keep the solution acidic. The
Experimental Section
General: Chemical compounds were obtained from Fluka, Sigma–
Aldrich Chemicals, Alfa Aesar, Acros Organics, or Thermo Scientific
and used without further purification. HPLC-grade solvents were
purchased from Promochem and filtered and degassed under
vacuum prior to use. 5’-Chloro-5’-deoxyadenosine was synthesized
by chlorination of adenosine with SOCl2 as described in the litera-
ture.[42] l-Selenohomocysteine was obtained by a method de-
scribed by Zhou et al.[37] and AdoYn (3) was synthesized as de-
scribed previously.[34]
ChemBioChem 0000, 00, 1 – 8
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
&
5
&
ÞÞ
These are not the final page numbers!

E. Weinhold et al.
yield was 4.58 mg (9.72 mmol, 21%) as determined by UV spectros-
copy with use of an extinction coefficient of 15400 LmolÀ1 cmÀ1 at
260 nm for the adenine chromophore. 1H NMR (300 MHz, D2O):
d=2.24–2.30 (m, 2H; Hb), 2.96–3.11 (m, 1H; H3’’), 3.39–3.53 (m,
2H; Hg), 3.73–3.90 (m, 2H; H5’), 3.81–3.86 (m, 1H; Ha), 3.94–4.08
(m, 2H; H1’’), 4.50 (q, J=6 Hz, 1H; H4’), 4.54–4.60 (m, 1H; H3’),
4.69–4.74 (m, 1H; H2’), 5.99–6.05 (m, 1H; H1’), 8.31–8.33 ppm (m,
2H; H2, H8); MS (ESI): m/z (%): 471.09 (100) [M]+, 370.04 (26) [MÀ
aminobutyrolactone+H]+, 250.60 (5) [5’-deoxyadenosine]+.
was not removed. TAF10 expression plasmid was provided by
Laszlo Tora and the protein was purified with modifications to the
method described by Kouskouti et al.[12] TAF10 was expressed in
BL21 Gold(DE3) (Agilent/Stratagene) and purified on a Ni-NTA
column (Qiagen). Fractions containing the enzyme were pooled
and concentrated. The His tag was not removed. PrmC, RF1, and
RF2 expression plasmids were provided by Valꢄrie Heurguꢄ-
Hamard. PrmC was expressed and purified by the method de-
scribed by Graille et al.[16] and RF1 and RF2 by the method de-
scribed by Mora et al.[47]
Azide-derivatized TAMRA: An isomer mixture of 5- and 6-carboxy-
tetramethylrhodamine succinimidyl ester (5.3 mg, 10 mmol, 1 equiv)
was dissolved in dry DMSO. 11-Azido-3,6,9-trioxaundecan-1-amine
(21.8 mg, 23 mL, 100 mmol, 10 equiv) and triethylamine (10.1 mg,
16 mL, 100 mmol, 10 equiv) were added to a total volume of 250 mL.
The reaction mixture was protected from light and incubated at
room temperature for 1 h. It was then analyzed and purified by
preparative reversed-phase HPLC and compounds were eluted
with a methanol gradient (31.2% for 5 min, 31.2–66.3% in 7 min,
66.3–70.2% in 5 min, and 70.2–78% in 2 min) in aqueous trifluoro-
acetic acid (0.01%). The two product isomers were separated,
yielding 2.28 mg (36.2%) for the 5-isomer (retention time 14.5 min)
and 2.15 mg (34.1%) for the 6-isomer (retention time 15.0 min) as
determined by UV spectroscopy with use of an extinction coeffi-
cient of 80000 LmolÀ1 cmÀ1 at 552 nm for the TAMRA chromo-
phore. MS (ESI): m/z 5-isomer (%): 631.20 (100) [M]+; MS (ESI): m/z
6-isomer (%): 631.27 (100) [M]+.
Two-step methyltransferase assay: Alkynylation of proteins was
performed with equal molar amounts of substrates and MTases
(Dim-5 and Clr4: 6.5 mm; Set7/9, G9a, PRMT1 and PrmC: 10 mm), co-
factor analogues 2, 3, or 4 (600 mm, about 1:1 mixtures of diaste-
reoisomers at sulfur or selenium), and glycerol (5%) in the appro-
priate buffer (10 mL) either at 208C (Clr4 and Dim-5) or at 308C
(Set7/9, G9a, PRMT1 and PrmC). The following buffers were used.
Clr4: HEPES (100 mm, pH 7.5); Dim-5: glycine (50 mm, pH 9.8);
Set7/9, G9a, and PRMT1 with histone H4: Tris·HCl (50 mm, pH 9.0),
magnesium chloride (5 mm); PRMT1 with hnRNPK: HEPES (50 mm,
pH 8.0), potassium acetate (40 mm), BSA (0.2 mgmLÀ1), glycerol
(10%), EDTA (1 mm); PrmC: HEPES (10 mm, pH 7.5). Without further
purification, the modification reaction mixtures including the alky-
nylated proteins were labeled by CuAAC. The protein modification
reactions (10 mL) were supplemented with a premix solution
(2.5 mL) to give the following final concentrations: azide-derivatized
biotin[27] or azide-derivatized TAMRA (1.2 mm), copper(II) sulfate
(0.6 mm), sodium ascorbate (50 mm) and ligand l-proline
(1.2 mm)[27] or tris-(hydroxypropyltriazolylmethyl)amine (THPTA;
0.6 mm).[48] Reaction mixtures were incubated at 308C for 30 min
or at room temperature for 1 h. Biotin-labeled proteins were direct-
ly separated by SDS-PAGE, whereas TAMRA-labeled proteins were
precipitated with CHCl3/MeOH/H2O (1:4:4) prior to separation by
SDS-PAGE.
Analysis of cofactor analogue stability: Stock solutions (40 mL, ap-
proximately 10 mm) of AdoYn (3) or SeAdoYn (4) were added to
HEPES buffer (100 mm, pH 7.5, 500 mL) and incubated at room tem-
perature. Samples (30 mL) were injected into the HPLC system after
different incubation times and analyzed by use of an analytical re-
versed-phase column. Compounds were eluted with the gradient
described for the purification of SeAdoYn (4) but with a flow of
1 mLminÀ1. In parallel, samples (30 mL) were taken and treated
with trifluoroacetic acid until the pH was around 2. The solvent
was removed by lyophilization and the solid was dissolved in
water for injection into the mass spectrometer.
Biotinylated proteins were detected by Western blotting as de-
scribed previously.[27] Ponceau staining [Ponceau S (0.5%) in acetic
acid (1%), slow shaking for approximately 10 min)] as loading con-
trol for proteins was carried out prior to blocking with BSA (2.5%)
in PBS-T buffer [sodium chloride (140 mm), potassium chloride
(2.5 mm), disodium hydrogenphosphate (2 mm), potassium dihy-
drogenphosphate (1.45 mm), Tween-20 (0.05%), pH 7.5] overnight.
In-gel detection of TAMRA-labeled proteins was performed by im-
mediately placing the gel on an UV transilluminator (excitation at
312 nm) after SDS-PAGE and taking a photo through a filter (540Æ
50 nm). For loading control the gel was Coomassie-stained [Coo-
massie Brilliant Blue (0.05%), acetic acid (10%), ethanol (40%),
water (50%)] after in-gel fluorescence detection.
Collection and purification of protein MTases and protein sub-
strates: PRMT1 MTase and hnRNPK were gifts from Antje Ostareck-
Lederer. Purified G9a MTase was a gift from New England Biolabs
and histone H3.1 and histone H4 were purchased from New Eng-
land Biolabs. Histone H3 wild type, K4A, and K9A were purified by
a protocol by Luger et al.[44] with some alterations. Pelleted inclu-
sion bodies from a culture (500 mL) were solubilized with a
Dounce homogenizer in buffer [guanidinium hydrochloride (7m),
Tris·base (20 mm), dithiothreitol (10 mm), pH 7.5, 5 mL] and stirred
at room temperature for 2 h prior to centrifugation (20 min,
23000g). The supernatant was dialyzed against buffer [NaCl
(100 mm), Tris (10 mm), EDTA (1 mm), b-mercaptoethanol (5 mm),
phenylmethylsulfonyl fluoride (0.2 mm), pH 8] containing urea (7m,
three times, twice for 1 h and once overnight) and without urea
(twice, 6 h and overnight) and stored in aliquots at À208C. Dim-5
expression plasmid was provided by Xiaodong Cheng, and the pro-
tein was purified without removal of the GST tag.[45] Clr4D191 (resi-
dues 192–490) expression plasmid was provided by Rui-Ming Xu
and the protein was purified as described by Min et al.[7] but with-
out the use of a HiTrap-Q column. Set7/9 expression plasmid was
provided by Danny Reinberg and the protein was purified with the
following modifications to the literature method.[46] Induction in
BL21 Gold(DE3) (Agilent/Stratagene) was done at 378C for 3 h and
the enzyme was purified on a Ni-NTA column (Qiagen). Fractions
containing the enzyme were pooled and concentrated. The His tag
Acknowledgements
We thank Sriharsa Pradhan for inviting us to perform the G9a ex-
periments at New England Biolabs, Antje Ostareck-Lederer for
providing PRMT1 and hnRNPK proteins, Xiaodong Cheng for
Dim-5 plasmid, Rui-Ming Xu for Clr4 plasmid, Valꢁrie Heurguꢁ-
Hamard for PrmC, RF1, and RF2 plasmids, Danny Reinberg for
Set7/9, histone H3 wild-type and histone H3 variant K4A and K9A
plasmids and Laszlo Tora for the TAF10 plasmid. We thank Ker-
stin Glensk for assistance with protein purifications. This work
was supported by the Excellence Initiative of the German Federal
and State Governments.
&6
&
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 8
ÝÝ
These are not the final page numbers!

Selenium Cofactor for Methyltransferases
[26] A. Dhayalan, S. Kudithipudi, P. Rathert, A. Jeltsch, Chem. Biol. 2011, 18,
111–120.
Keywords: click chemistry · epigenetics · methyltransferases ·
protein modifications · selenium
[27] W. Peters, S. Willnow, M. Duisgen, H. Kleine, T. Macherey, K. E. Duncan,
D. W. Litchfield, B. Lꢀscher, E. Weinhold, Angew. Chem. 2010, 122, 5296–
5299; Angew. Chem. Int. Ed. 2010, 49, 5170–5173.
[28] C. Dalhoff, G. Lukinavicius, S. Klimasauskas, E. Weinhold, Nat. Chem. Biol.
2006, 2, 31–32.
[29] V. Hong, S. I. Presolski, C. Ma, M. G. Finn, Angew. Chem. 2009, 121,
10063–10067; Angew. Chem. Int. Ed. 2009, 48, 9879–9883.
[30] M. Meldal, C. W. Tornoe, Chem. Rev. 2008, 108, 2952–3015.
[31] Y. Motorin, J. Burhenne, R. Teimer, K. Koynov, S. Willnow, E. Weinhold, M.
Helm, Nucleic Acids Res. 2011, 39, 1943–1952.
[32] R. Wang, W. Zheng, H. Yu, H. Deng, M. Luo, J. Am. Chem. Soc. 2011, 133,
7648–7651.
[33] K. Islam, W. Zheng, H. Yu, H. Deng, M. Luo, ACS Chem. Biol. 2011, 6,
679–684.
[34] E. Weinhold, C. Dalhoff, S. Klimasauskas, G. Lukinavicius, WO 2006/
108678, 2006.
[35] H. Stecher, M. Tengg, B. J. Ueberbacher, P. Remler, H. Schwab, H. Griengl,
M. Gruber-Khadjawi, Angew. Chem. 2009, 121, 9710–9712; Angew.
Chem. Int. Ed. 2009, 48, 9546–9548.
[1] B. D. Strahl, C. D. Allis, Nature 2000, 403, 41–45.
[2] T. Jenuwein, C. D. Allis, Science 2001, 293, 1074–1080.
[3] S. S. Oliver, J. M. Denu, ChemBioChem 2011, 12, 299–307.
[4] M. Lachner, R. J. O’Sullivan, T. Jenuwein, J. Cell Sci. 2003, 116, 2117–
2124.
[5] S. D. Taverna, H. Li, A. J. Ruthenburg, C. D. Allis, D. J. Patel, Nat. Struct.
Mol. Biol. 2007, 14, 1025–1040.
[6] S. S. Ng, W. W. Yue, U. Oppermann, R. J. Klose, Cell. Mol. Life Sci. 2009,
66, 407–422.
[7] J. Min, X. Zhang, X. Cheng, S. I. S. Grewal, R.-M. Xu, Nat. Struct. Biol.
2002, 9, 828–832.
[8] K. Zhang, K. Mosch, W. Fischle, S. I. S. Grewal, Nat. Struct. Mol. Biol.
2008, 15, 381–388.
[9] K. Nishioka, S. Chuikov, K. Sarma, H. Erdjument-Bromage, C. D. Allis, P.
Tempst, D. Reinberg, Genes Dev. 2002, 16, 479–489.
[10] J. Huang, S. L. Berger, Curr. Opin. Genet. Dev. 2008, 18, 152–158.
[11] R. J. Sims III, D. Reinberg, Nat. Rev. Mol. Cell Biol. 2008, 9, 815–820.
[12] A. Kouskouti, E. Scheer, A. Staub, L. Tora, I. Talianidis, Mol. Cell 2004, 14,
175–182.
[36] O. Binda, M. Boyce, J. S. Rush, K. K. Palaniappan, C. R. Bertozzi, O.
Gozani, ChemBioChem 2011, 12, 330–334.
[37] Z. S. Zhou, A. E. Smith, R. G. Matthews, Bioorg. Med. Chem. Lett. 2000,
10, 2471–2475.
[13] S. Pradhan, H. G. Chin, P. O. Esteve, S. E. Jacobsen, Epigenetics 2009, 4,
383–387.
[38] C. Dalhoff, G. Lukinavicius, S. Klimasauskas, E. Weinhold, Nat. Protoc.
2006, 1, 1879–1886.
[39] J. L. Hoffman, Biochemistry 1986, 25, 4444–4449.
[40] D. F. Iwig, A. T. Grippe, T. A. McIntyre, S. J. Booker, Biochemistry 2004, 43,
13510–13524.
[14] X.-D. Yang, A. Lamb, L.-F. Chen, Epigenetics 2009, 4, 429–433.
[15] P.-O. Esteve, H. G. Chin, J. Benner, G. R. Feehery, M. Samaranayake, G. A.
Horwitz, S. E. Jacobsen, S. Pradhan, Proc. Natl. Acad. Sci. USA 2009, 106,
5076–5081.
[16] M. Graille, V. Heurgue-Hamard, S. Champ, L. Mora, N. Scrima, N. Ulryck,
H. van Tilbeurgh, R. H. Buckingham, Mol. Cell 2005, 20, 917–927.
[17] H. G. Chin, P.-O. Esteve, M. Pradhan, J. Benner, D. Patnaik, M. F. Carey, S.
Pradhan, Nucleic Acids Res. 2007, 35, 7313–7323.
[18] O. Pless, E. Kowenz-Leutz, M. Knoblich, J. Lausen, M. Beyermann, M. J.
Walsh, A. Leutz, J. Biol. Chem. 2008, 283, 26357–26363.
[19] A. Ostareck-Lederer, D. H. Ostareck, K. P. Rucknagel, A. Schierhorn, B.
Moritz, S. Huttelmaier, N. Flach, L. Handoko, E. Wahle, J. Biol. Chem.
2006, 281, 11115 – 11125.
ˇ
[41] M. Tomkuvienꢄ, B. Clouet-d’Orval, I. Cerniauskas, E. Weinhold, S. Klima-
ˇ
sauskas, Nucleic Acids Res. 2012, in press.
[42] R. W. Miles, L. P. C. Nielsen, G. J. Ewing, D. Yin, R. T. Borchardt, M. J.
Robins, J. Org. Chem. 2002, 67, 8258–8260.
[43] S. A. Ross, M. Pitie, B. Meunier, J. Chem. Soc. Perkin Trans. 1 2000, 571–
574.
[44] K. Luger, T. J. Rechsteiner, A. J. Flaus, M. M. Y. Waye, T. J. Richmond, J.
Mol. Biol. 1997, 272, 301–311.
[45] X. Zhang, H. Tamaru, S. I. Khan, J. R. Horton, L. J. Keefe, E. U. Selker, X.
Cheng, Cell 2002, 111, 117–127.
[46] R. C. Trievel, B. M. Beach, L. M. A. Dirk, R. L. Houtz, J. H. Hurley, Cell 2002,
111, 91–103.
[47] L. Mora, V. Heurgue-Hamard, S. Champ, M. Ehrenberg, L. L. Kisselev,
R. H. Buckingham, Mol. Microbiol. 2003, 47, 267–275.
[48] T. R. Chan, R. Hilgraf, K. B. Sharpless, V. V. Fokin, Org. Lett. 2004, 6, 2853–
2855.
[20] M. T. Bedford, S. G. Clarke, Mol. Cell 2009, 33, 1–13.
[21] P. Trojer, D. Reinberg, Nat. Chem. Biol. 2008, 4, 332–334.
[22] S. Chuikov, J. K. Kurash, J. R. Wilson, B. Xiao, N. Justin, G. S. Ivanov, K.
McKinney, P. Tempst, C. Prives, S. J. Gamblin, N. A. Barlev, D. Reinberg,
Nature 2004, 432, 353–360.
[23] J.-F. Couture, E. Collazo, G. Hauk, R. C. Trievel, Nat. Struct. Mol. Biol.
2006, 13, 140–146.
[24] K. L. Bicker, O. Obianyo, H. L. Rust, P. R. Thompson, Mol. BioSyst. 2011, 7,
48–51.
[25] P. Rathert, A. Dhayalan, M. Murakami, X. Zhang, R. Tamas, R. Jurkowska,
Y. Komatsu, Y. Shinkai, X. Cheng, A. Jeltsch, Nat. Chem. Biol. 2008, 4,
344–346.
Received: December 17, 2011
Published online on && &&, 0000
ChemBioChem 0000, 00, 1 – 8
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
&
7
&
ÞÞ
These are not the final page numbers!

FULL PAPERS
S. Willnow, M. Martin, B. Lꢀscher,
E. Weinhold*
&& – &&
A Selenium-Based Click AdoMet
Analogue for Versatile Substrate
Labeling with Wild-Type Protein
Methyltransferases
Element editing: S-Adenosyl-l-methio-
nine-dependent methyltransferases can
transfer extended chemical groups
(such as the propargyl group, which
combines smallness, activation, and the
potential for labeling after transfer)
from cofactor analogues containing un-
saturated bonds next to the reactive
carbon. Replacing the sulfur in an unsta-
ble propargylic analogue with selenium
gave a cofactor suitable for efficient
protein labeling.
&8
&
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 8
ÝÝ
These are not the final page numbers!