A sensitive mass spectrum assay to characterize engineered methionine adenosyltransferases with S-alkyl methionine analogues as substrates

Article abstract of DOI:10.1016/j.ab.2013.12.026

Methionine adenosyltransferases (MATs) catalyze the formation of S-adenosyl-l-methionine (SAM) inside living cells. Recently, S-alkyl analogues of SAM have been documented as cofactor surrogates to label novel targets of methyltransferases. However, these chemically synthesized SAM analogues are not suitable for cell-based studies because of their poor membrane permeability. This issue was recently addressed under a cellular setting through a chemoenzymatic strategy to process membrane-permeable S-alkyl analogues of methionine (SAAMs) into the SAM analogues with engineered MATs. Here we describe a general sensitive activity assay for engineered MATs by converting the reaction products into S-alkylthioadenosines, followed by high-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) quantification. With this assay, 40 human MAT mutants were evaluated against 7 SAAMs as potential substrates. The structure-activity relationship revealed that, besides better engaged SAAM binding by the MAT mutants (lower K_m value in contrast to native MATs), the gained activity toward the bulky SAAMs stems from their ability to maintain the desired linear S_N2 transition state (reflected by higher k_cat value). Here the I117A mutant of human MATI was identified as the most active variant for biochemical production of SAM analogues from diverse SAAMs.

Full text of DOI:10.1016/j.ab.2013.12.026

Analytical Biochemistry 450 (2014) 11–19
Contents lists available at ScienceDirect
Analytical Biochemistry
journal homepage: www.elsevier.com/locate/yabio
A sensitive mass spectrum assay to characterize engineered methionine
adenosyltransferases with S-alkyl methionine analogues as substrates
Rui Wang a,b, Weihong Zheng a, Minkui Luo a,
⇑
a Molecular Pharmacology and Chemistry Program, Memorial Sloan–Kettering Cancer Center, New York, NY 10065, USA
b Program of Pharmacology, Weill Graduate School of Medical Science, Cornell University, New York, NY 10021, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 20 August 2013
Methionine adenosyltransferases (MATs) catalyze the formation of S-adenosyl-L-methionine (SAM)
inside living cells. Recently, S-alkyl analogues of SAM have been documented as cofactor surrogates to
label novel targets of methyltransferases. However, these chemically synthesized SAM analogues are
not suitable for cell-based studies because of their poor membrane permeability. This issue was recently
addressed under a cellular setting through a chemoenzymatic strategy to process membrane-permeable
S-alkyl analogues of methionine (SAAMs) into the SAM analogues with engineered MATs. Here we
describe a general sensitive activity assay for engineered MATs by converting the reaction products into
S-alkylthioadenosines, followed by high-performance liquid chromatography–tandem mass spectrome-
try (HPLC–MS/MS) quantification. With this assay, 40 human MAT mutants were evaluated against 7
SAAMs as potential substrates. The structure–activity relationship revealed that, besides better engaged
SAAM binding by the MAT mutants (lower Km value in contrast to native MATs), the gained activity
toward the bulky SAAMs stems from their ability to maintain the desired linear SN2 transition state
(reflected by higher kcat value). Here the I117A mutant of human MATI was identified as the most active
variant for biochemical production of SAM analogues from diverse SAAMs.
Received in revised form 16 December 2013
Accepted 18 December 2013
Available online 27 December 2013
Keywords:
Epigenetics
Methyltransferase
MAT
S-Adenosyl-L-methionine
LC–MS/MS
Ó 2013 Elsevier Inc. All rights reserved.
S-Adenosyl-
L
-methionine (SAM),1 one of the most commonly
the effect of the weaker carbon chalcogen bond on catalytic turnover
[11]. As recently demonstrated by our laboratory and others, bulky
sulfonium-b-sp2/sp1-alkyl analogues of SAM can serve as cofactor
surrogates of native or engineered methyltransferases for substrate
labeling or profiling [9,12–23].
used enzyme cofactors, is found among all living organisms [1,2].
The biochemical reactivity of SAM is largely embedded within its
three carbon sulfonium bonds as well as their neighboring regions
[2–4]. For instance, the two alkyl sulfonium bonds can undergo en-
zyme-mediated homolytic cleavage to yield the canonical 50-deoxy-
adenosyl or noncanonical 3-amino-3-carboxypropyl radical [5,6]. In
contrast, the methyl sulfonium bond is often subjected to the hetero-
lytic cleavage to render a methyl electrophile for diverse acceptors
ranging from large DNA, RNA, and protein complexes to small-mol-
ecule metabolites [7]. In conjunction with the increased interest in
these transformations, multiple SAM analogues have been docu-
mented to probe reactions or mechanisms of SAM-using enzymes,
including methyltransferases [8–10]. Among the early examples
was to replace SAM’s sulfonium moiety with selenium to examine
To access bulky sulfonium-alkyl analogues of SAM, most previ-
ous efforts relied on chemical alkylation of S-adenosyl-L-homocys-
teine (SAH) with the corresponding alkyl electrophiles [12–23].
However, besides the limitation of less desirable yields and requir-
ing laborious high-performance liquid chromatography (HPLC)
purification [12–23], chemically synthesized SAM analogues often
consist of diastereomeric mixtures with only the sulfonium-(S)-
epimer bioactive [12–24]. More important, SAM analogues gener-
ally show poor membrane permeability, which limits their use in
a cellular setting [8].
Within living cells, SAM is mainly generated by methionine
adenosyltransferases (MATs) with endogenous methionine and
ATP as substrates and phosphate and pyrophosphate as byproducts
(Fig. 1) [2,16]. An alternative biosynthesis of SAM involves the SalL-
catalyzed reverse transformation with methionine and 50-chloro-
50-deoxyadenosine as substrates [22]. The general in vivo setting
of MATs also inspired us to develop a chemoenzymatic strategy
with engineered MATs to process membrane-permeable S-alkyl
⇑
Corresponding author.
E-mail address: luom@mskcc.org (M. Luo).
Abbreviations used: SAM, S-adenosyl-L-methionine; SAH, S-adenosyl-L-homocys-
1
teine; HPLC, high-performance liquid chromatography; MAT, methionine adenosyl-
transferase; SAAM, S-alkyl analogue of methionine (S-alkyl-L-homocysteine); NADPH,
nicotinamide adenine dinucleotide phosphate; NMR, nuclear magnetic resonance;
DMSO, dimethyl sulfoxide; UV, ultraviolet; LC–MS, liquid chromatography–mass
spectrometry; ESI, electron spray ionization; MRM, multiple reaction monitoring; MS/
MS, tandem mass spectrometry; MTA, 50-methylthioadenosine.
analogues of methionine (SAAMs or S-alkyl-L-homocysteines) into
0003-2697/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.ab.2013.12.026

12
Assay for MATs with SAM analogues as substrates / R. Wang et al. / Anal. Biochem. 450 (2014) 11–19
H
O
O
-O
P
O-
H
O-
O-
-O
O-
Mg2+
O-
O
P
O
P
O
O
O
-O
O
O
-O
O
O
O
NH3
V121
P
Mg2+
P
NH3
P
O
-O
(106)
O-
O-
O
V121(106)
S
(105)
120
S
G
I
HO
HO
G120(105)
I117(102)
Me
HO
Me
O
O
I322
(302)
N
I322
(302)
(102)
117
HO
N
N
N
N
N
N
N
NH2
NH2
Fig. 1. MAT-catalyzed SAM production with methionine and ATP as substrates and phosphate and pyrophosphate as byproducts. Here the residues adjacent to S-methyl
moiety of the substrate methionine are constructed and highlighted on comparing human MATI and MATII (blue, PDB 2OBV and 2P02) and Escherichia coli MAT (black in
parentheses) as reported previously [29].
the SAM analogues [16]. Despite the previous proof-of-concept use
of the engineered MAT in vitro and inside living cells [16], few
efforts have been made to systematically characterize these MAT
variants, likely because of the lack of a general activity assay for
MAT mutants with diverse SAAMs as substrates and SAM
analogues as products [16,25–27]. Given the potential use of the
chemoenzymatic strategy for multiple SAM-using enzymes, as
exemplified recently by methyltransferases [12–23], here we doc-
ument a sensitive, generally applicable mass-spectroscopy-based
assay to quantify SAM analogues (Fig. 2).
The conventional kinetic analysis of MATs was carried out
through HPLC-based spectroscopic quantification of the reaction
product SAM, a less sensitive assay format that often requires a
large amount of reaction materials [25]. Alternatively, less environ-
mentally friendly radiolabeled methionine or ATP can be used as a
substrate with the reaction product SAM isolated by chromatogra-
phy and quantified by a scintillation counter [26]. Recently, a more
sensitive spectroscopic MAT assay was developed by monitoring
the production of nicotinamide adenine dinucleotide phosphate
(NADPH) from the reaction byproduct pyrophosphate using three
coupling enzymes: uridine diphosphoglucose pyrophosphorylase,
phosphoglucomutase, and glucose 6-phosphate dehydrogenase
[27]. This continuous assay is sensitive but can be affected by po-
tential cellular components interfering with the coupling enzymes
or NADPH readout. Here we describe a sensitive mass-spectros-
copy-based assay by converting SAM and S-alkyl analogues of
SAM (the reaction products of native and engineered MATs) into
S-alkyl thioadenosines for quantification (Fig. 2). This assay is fur-
ther featured by its potential application to quantify S-alkyl ana-
logues of SAM in complex cellular settings. We implemented the
assay to evaluate activities of 40 MAT variants toward 7 SAAMs
as potential substrates. The structure–activity relationship re-
vealed that the difference of the overall catalytic efficiency (kcat
/
Km values) of engineered MATs toward bulky SAAMs largely arises
from the values of kcat reflected at the chemical step and enzymatic
transition state, whereas the values of Km for binding substrates
are comparable among diverse SAAMs. This observation argues
for the importance of engineering MATs to both bind bulky sub-
strates (the step reflected by small Km values) and efficiently pro-
cess them into products through active enzymatic transition
states (the step reflected by decent kcat values). Here the I117A
mutant of human MATI was identified to be the most active MAT
R =
SAAM 3 R =
SAAM 4 R =
SAAM 2
SAAM 6
SAAM 7 R =
SAAM 8
R =
R =
O
R =
SAAM 5
O
H2N
O
COOH
H2N
Native
MAT
NH2
N
N
NH2
N
N
NH2
COOH
H3C
N
N
N
N
Citrate buffer
S
S
S
H3C
O
H3C
O
ATP
Methionine 1
OH
Native SAM
HO
OH
MTA
HO
HPLC-MS/MS
MTA
SAM Analogues
COOH
Analogues
SAAMs
H2N
Engineered
MAT
NH2
N
N
NH2
NH2
Citrate buffer
N
R
N
N
COOH
ATP
S
S
N
S
N
O
N
O
O
R
R
H2N
OH
O
HO
OH
HO
Fig. 2. Schematic presentation of MS-based MAT activity assay. In this assay, SAM/SAM analogues were quantitatively converted into the corresponding MTA/MTA analogues
and lactone through intramolecular lactonization, followed by tandem HPLC–MS/MS quantification.

Assay for MATs with SAM analogues as substrates / R. Wang et al. / Anal. Biochem. 450 (2014) 11–19
13
variant for biochemical production of sulfonium-alkyl analogues of
SAM from diverse SAAMs.
The resultant white solid was dissolved in dry ethanol (10 ml) and
cooled down to 0 °C. A solution of tosylate or bromide (1.05 mmol)
in ethanol (3 ml) was added (bromides for the synthesis of SAAMs 3,
4, 7, and 8; tosylates for SAAMs 5 and 6) [16–19,21,22]. This reac-
tion mixture was stirred at 0 °C for 1 h and then brought to ambient
temperature (23 °C) overnight. Solvent was then removed by rotary
evaporation, and the resultant residue was redissolved in water
(5 ml) and purified using a self-packed 1.5 ꢀ 5-cm Dowex 50 (H+)
cation exchange column. The column was washed with water until
the pH of the eluent became neutral. The products were then eluted
from the column using 5% ammonia hydroxide, which was later re-
moved by rotary evaporation. The portions containing the expected
products (monitored by silica gel thin layer chromatography [TLC]
with the eluting solvent, n-BuOH/AcOH/H2O 4:1:1; the staining
protocol, 0.2% ninhydrin in 96% EtOH/AcOH/s-collidine 6:3:1 on
heating) were combined and concentrated on a rotary evaporator
to yield a white solid. The crude products were then dissolved in
0.1 N HCl and subject to further purification with a preparative re-
Materials and methods
General materials and methods
SAM, SAH,
L-methionine 1, S-ethyl analogue of methionine
(L-ethionine, 2), and the reagents for chemical synthesis were ob-
tained from Aldrich Chemical and used without further purifica-
tion. Optima-grade acetonitrile was obtained from Fisher
Scientific and degassed under vacuum prior to HPLC purification.
Citrate buffer was purchased from Sigma–Aldrich (cat. no.
83273). Cocktail of ethylenediaminetetraacetic acid (EDTA)-free
protease inhibitors was purchased from Roche Applied Science.
Aqueous solutions of SAAMs were concentrated with a Savant
Sc210A SpeedVac concentrator (Thermo) and then lyophilized with
a Flexi-Dry lP Freeze-Dryer (FTS system). Nuclear magnetic reso-
verse phase HPLC system (XBridge Prep C18 column, 5-lm, OBD,
nance (NMR) spectra were recorded on a Burke AVIII 500- or
600-MHz spectrometer. NMR chemical shifts are reported in
ppm; multiplicity is indicated by s (singlet), d (doublet), t (triplet),
q (quartet), m (multiplet), dd (doublet of doublet), and so forth;
coupling constants (J) are reported in Hz with peak integration pro-
19 ꢀ 150 mm). The final products were eluted by a solvent of a lin-
ear gradient from 10 to 60% of acetonitrile in aqueous trifluoroace-
tic acid (0.1%) for 15 min with a flow rate of 10 ml/min. The desired
fractions were combined and lyophilized to yield a white powder
(overall yields of 82, 67, 56, 56, 35, and 58% for SAAMs 3, 4, 5, 6,
7, and 9, respectively).
vided. Occasionally, formic acid-d2 of 5
l
l was added into 600
ll of
D2O as the NMR solvent to increase the solubility of the com-
1H NMR (500 MHz, D2O) of SAAM 3 (S-allyl-
L-homocysteine)
pounds containing an
a
-amino acid moiety. 1H/13C NMR chemical
shifts were referenced to solvent peaks (residual 1H in D2O and
dimethyl sulfoxide [DMSO]-d6 = 4.79 ppm and 3.50 ppm, respec-
tively; residual 13C in formic acid-d2 and DMSO-d6 = 166.2 ppm
and 39.7 ppm, respectively). Analytical HPLC was carried out on a
Waters 600 controller HPLC/2998 diode array detector using an
[28]: d 2.07–2.20 (m, 2H), 2.62 (t, 2H, J = 7.5 Hz), 3.23 (d, 2H,
J = 7.2 Hz), 3.84 (t, 1H, J = 6.3 Hz), 5.16–5.22 (m, 2H), 5.80–
5.88(m, 1H).
1H NMR (500 MHz, D2O) of SAAM 4 (S-crotyl-
L-homocysteine):
d 1.71 (d, 3H, J = 6.4 Hz), 2.15–2.20(m, 1H), 2.23–2.29 (m, 1H),
2.67 (t, 2H, d = 7.5 Hz), 3.20 (d, 2H, J = 7.3 Hz), 4.17 (t, 1H,
J = 6.3 Hz), 5.47–5.53 (m, 1H), 5.70–5.71 (m, 1H); 13C NMR
(125 MHz, D2O): d 16.88, 24.91, 29.45, 32.50, 52.15, 126.04,
129.92, 172.07; ESI–MS: 190 [M+H]+. HRMS: calculated for
C8H16NO2S ([M+H]+) 190.0902, found 190.0897.
XBridge Prep C18 reverse phase column (5
Preparative HPLC purification was carried out on a DELTA PAK
C18 column (15
m, 300 A, 300 ꢀ 3.9 mm) or an XBridge Prep
C18 reverse phase column (5
m, OBD, 19 ꢀ 150 mm) with ultravi-
lm, 4.6 ꢀ 150 mm).
l
l
olet (UV) detection at 260 nm. Mass spectral analysis was carried
out at the MSKCC Analytical Core Facility on a PE SCIEX API 100
or Waters Acuity SQD liquid chromatography–mass spectrometry
(LC–MS) system with electron spray ionization (ESI). LC–MS sam-
ples were analyzed by multiple reaction monitoring (MRM) modes
using the 6410 tandem LC–MS/MS system (Agilent Technologies)
coupled with a Zorbax Eclipse XDB-C18 column (2.1 ꢀ 50 mm,
1H NMR (500 MHz, D2O) of SAAM 5 (trans-pent-2-enyl-
L-homo-
cysteine): d 1.02 (t, 3H, J = 7.4 Hz), 2.09–2.13 (m, 2H), 2.15–2.19 (m,
1H), 2.22–2.23 (m, 1H), 2.68 (t, 2H, d = 7.3 Hz), 3.23 (d, 2H,
J = 7.2 Hz), 4.01 (t, 1H, J = 6.2 Hz), 5.50–5.54 (m, 1H), 5.76–5.79
(m, 1H); 13C NMR (150 MHz, D2O): d 12.81, 24.70, 24.94, 29.70,
32.41, 53.05, 123.73, 136.90, 173.15; ESI–MS: 204 [M+H]+. HRMS:
calculated for C9H18NO2S ([M+H]+) 204.1058, found 204.1056.
1H NMR (500 MHz, D2O+formic acid-d2) of SAAM 6 (S-(pent-2-
3.5
lm).
en-4-yny)-L-homocysteine): d 1.98–2.04 (m, 1H), 2.06–2.12 (m,
Synthesis, purification, and characterization of SAAMs 2 to 8
1H), 2.50 (t, 2H, J = 7.4 Hz), 3.12 (d, 2H, J = 7.4 Hz), 3.15 (s, 1H),
3.99 (t, 1H, J = 6.0 Hz), 5.50 (d, 1H, 15.7 Hz), 6.06–6.12 (m, 1H);
13C NMR (125 MHz, D2O + formic acid-d2): d 25.99, 30.15, 33.06,
52.86, 79.09, 82.71, 111.45, 142.06, 172.77; MS(ESI) m/z: 200
[M+H]+; HRMS: calculated for C9H14NO2S ([M+H]+) 200.0745,
found 200.0746.
S-Ethyl analogue of methionine (SAAM 2,
able from Sigma–Aldrich. S-Benzyl- -homocysteine and S-allyl-
homocysteine (SAAM 3) were prepared as reported previously
[28]. To prepare SAAMs 4 to 8 (Scheme 1), S-benzyl- -homocysteine
(255 mg, 1 mmol) [29] was placed in a round-bottom flask con-
nected to an ammonia cylinder and dissolved in approximately
20 ml of condensed liquid ammonia in a dry ice–ethanol bath. So-
dium metal (50 mg, 2.2 mmol) was then added gradually to afford
a dark blue solution. After the dark blue solution became colorless
a few minutes later, the dry ice–ethanol bath was removed to allow
the evaporation of ammonia. The remaining trace of ammonia was
removed with the aid of a flow of argon, followed by vacuum for 3 h.
L-ethionine) is avail-
L
L-
L
1H NMR (500 MHz, D2O) of SAAM 7 (S-(hex-2-en-5-yny)-
L-
homocysteine): 2.13–2.19 (m, 1H), 2.22–2.28 (m, 1H), 2.52 (t, 1H,
J = 2.4 Hz), 2.66 (t, 2H, J = 7.5 Hz), 3.00–3.02 (m, 2H), 3.23 (dd,
2H, J = 7.2, 0.7 Hz), 4.14 (t, 1H, J = 6.3 Hz), 5.64–5.70 (m, 1H),
5.75–5.81 (m, 1H); 13C NMR (150 MHz, D2O + formic acid-d2): d
21.20, 25.68, 30.14, 32.68, 52.78, 71.94, 83.23, 117.00 (q,
J = 289.78), 127.90, 128.37, 163.65 (q, J = 35.2 Hz), 172.76; MS(ESI)
NH2
NH2
NH2
COOH
a. Na,Liq. NH3
b. RX, ethanol
BnCl, HCl
R
Me
Bn
S
COOH
S
COOH
Reflux
S
R = alkyl, X = halide or tosylate
Scheme 1. General synthesis of SAAMs 3 to 8.

14
Assay for MATs with SAM analogues as substrates / R. Wang et al. / Anal. Biochem. 450 (2014) 11–19
m/z: 214 [M+H]+; HRMS: calculated for C10H16NO2S ([M+H]+)
214.0902, found 214.0898.
KCl, 2 mM MgCl2, 2.5 mM ATP, 7.5
and 2.5 mM methionine or SAAM in a final volume of 10
l
M native or engineered MATs,
ll. The ac-
1H NMR (500 MHz, DMSO-d6) of SAAM 8 (S-[4-(prop-2-ynyl-
tive mutants were incubated with SAAM in a 96-well plate at
ambient temperature (23 °C) for 8 to 10 h. The long incubation
time, although saturating the signals of more reactive substrate–
enzyme pairs, allowed maximizing the signals of less active sub-
strate–enzyme pairs (96-well polymerase chain reaction [PCR]
plates sealed with adhesive PRC sealing foil sheets should be used
to avoid potential evaporation in particular for the latter step in-
oxy)but-2-enylthio]-L-homocysteine): 1.74–1.82(m, 1H), 1.93–1.99
(m, 1H), 2.54 (t, 1H, J = 7.6 Hz), 3.15 (d, 2H, J = 5.8 Hz), 3.20–3.33
(m, 1H), 3.45 (d, 1H, J = 2.4 Hz), 3.98 (d, 2H, J = 4.3 Hz), 4.12 (d, 2H,
J = 2.4 Hz), 5.64–5.67 (m, 2H), 7.54 (brs, 2H); 13C NMR (150 MHz,
DMSO-d6): d 26.62, 31.06, 32.04, 53.14, 56.53, 68.79, 77.27, 80.32,
128.38, 129.49, 169.15; MS(ESI) m/z: 244 [M+H]+; HRMS: calculated
for C11H17NO3NaS ([M+Na]+) 266.0827, found 266.0822.
volved with heating). Subsequently, 1.0 ll of 1.0 M citrate buffer
was added into the reaction mixture, followed by incubation at
55 °C for 3.5 h, to convert the SAM/SAM analogues into the corre-
sponding 50-methylthioadenosine (MTA)/MTA analogues. This deg-
radation procedure was previously reported to give a consistent
yield of approximately 80% from SAM to MTA through intramolec-
ular lactonization [30]. The reaction mixture was then diluted by a
1:100 ratio with double-distilled H2O containing a known amount
Protein expression and purification
Human MAT plasmids (full-length MATI and MATII) were a gen-
erous gift from Udo Oppermann (University of Oxford, http://
www.sgc.ox.ac.uk/structures/MAT1A_2obv.html and http://www.
sgc.ox.ac.uk/structures/MM/MAT2AA_2p02_MM.html). To express
the N-terminal 6 ꢀ His MATI and MATII, the plasmids were trans-
formed into the Escherichia coli (DE3) Rosseta 2 strain and induced
of SAH as the internal reference for MS analysis. Then 2 ll of the
diluted reaction mixture was injected into the 6410 tandem
LC–MS/MS system for quantitative analysis. As a result, the
dilution factor of ‘‘110’’ was applied here on calculating the
concentrations of SAM/SAM analogues as described below.
The integrated peak areas for the MTA/MTA analogues and the
SAH standards were quantified by the corresponding MRM modes
using the 6410 tandem LC–MS/MS system. Prior to the MS analysis,
the sample was applied to a Zorbax Eclipse XDB-C18 column
with 0.5 mM isopropyl b- -1-thiogalactopyranoside (IPTG) at 17 °C
D
for 16 h before harvesting. The resultant cell pellets were lysed with
a buffer containing 50 mM Tris–HCl (pH 8.0), 50 mM NaCl, 5 mM b-
mercaptoethanol, 25 mM imidazole, and the cocktail of Roche pro-
tease inhibitors and 5% (v/v) glycerol. The MATI and MATII proteins
were then purified by Ni–NTA agarose resin (Qiagen) followed by a
5-ml HiTrap-Q Sepharose XL column (GE Healthcare). The fractions
containing MATI and MATII proteins were combined and concen-
trated using an Amicon Ultra-10 K centrifugal filter device. The
protein concentrations were determined with a Bradford assay kit
(Bio-Rad) using bovine serum albumin (BSA) as a standard. The
concentrated proteins were stored at ꢁ80 °C before use. The MAT
mutants were generated from the native plasmids with a Quik-
Change site-directed mutagenesis kit (Agilent Technologies) with
the vendor’s protocols. The mutation sites of the plasmids were
confirmed by DNA sequencing. All of the mutants were expressed
and purified as described above for the native MATs.
(2.1 ꢀ 50 mm, 3.5
lm) for LC separation using a gradient elution
system with buffer A (0.1% HCOOH in double-distilled H2O) and
buffer B (CH3CN) of the following gradient: 0 to 0.5 min, 5 to 30%
B; 0.5 to 5 min, 30% B; 5 to 7 min, 30 to 5% B. Integrated peak areas
for all MTA/MTA analogues were compared with those of exoge-
nously added SAH standards (see Fig. S1 in the online supplemen-
tary material as an example). Here SAM/SAM analogues in the
enzymatic reactions were quantified on the basis of the relative
peak ratios to the known amount of SAH rather than the absolute
peak areas of MTA/MTA analogues (see below for details), a proce-
dure that was implemented to avoid potential errors from the
multiple-step operation.
Conventional HPLC analysis of SAM production by native MATs
Prior to quantification of the SAM/SAM analogues, a standard
curve was generated by applying known concentrations of SAM/
SAM analogues and equal concentrations of SAH (after the step
of 1:100 dilution) under the conditions described above. The ratios
of the peak areas between the SAH standard and the resultant
MTA/MTA analogues were calculated to give the ‘‘correction fac-
tors’’ (see Fig. S1 as an example). The concentrations of SAM/SAM
analogues in the chemoenzymatic reactions were then quantified
from the converted MTA/MTA analogues according to the following
equation: [SAM]reaction = (peak area of MTA/MTA analogues)/(peak
area of SAH) ꢀ [SAH]standard ꢀ ‘‘correction factor’’ ꢀ ‘‘dilution fac-
tor,’’ where [SAM]reaction is the concentration of SAM or SAM ana-
logues generated in the enzymatic reaction, (peak area of MTA/
MTA analogues)/(peak area of SAH) is the peak ratio obtained by
the corresponding MRM MS, [SAH]standard is the concentration of
internal SAH standard added in the step of the 1:100 dilution,
the ‘‘correction factor’’ was obtained as described from the stan-
dard curves generated with known concentrations of SAM and
SAM analogues, and the ‘‘dilution factor’’ is ‘‘110’’ on counting
the 1:10 dilution with 1.0 M citrate buffer and the subsequent
1:100 dilution with double-distilled H2O as described above.
A prior HPLC-based MAT activity assay was used to characterize
the kinetics of native MATs [25]. This experiment was carried out
as an established standard to evaluate the robustness of the newly
developed LC–MS/MS-based assay in the current work. Briefly, the
activities of native MATs were measured in 2 ml of reaction mix-
ture containing 100 mM Tris–HCl (pH 8.0), 100 mM KCl, 2 mM
MgCl2, 8 mM glutathione, 2.5 mM ATP, 7.5 lM MATs, and varied
concentrations of methionine (up to 4 mM). The reaction mixture
was incubated at ambient temperature (23 °C) with a 4-min inter-
val within 20 min (a linear range of initial rates), and then a 300-
ll
reaction aliquot was quenched with 300 l of 20% HClO4 aqueous
l
solution. After centrifugation at 15,350g for 30 min, the superna-
tants containing SAM were resolved by reverse-phase HPLC using
a DELTA PAK C18 column (15
l
m, 300 ꢀ 3.9 mm) by monitoring
at UV 260 nm. The triethylamine-acetic acid buffer (50 mM, pH
5.0) and methanol were premixed with the ratios of 98:2 (buffer
A) and 50:50 (buffer B). SAM was eluted with buffer A for
30 min, followed by buffer B for 5 min, at a flow rate of 1 ml/min.
The integrated peak areas at 260 nm were used to generate
the standard curve with the known concentration of SAM and
to quantify the SAM produced in the kinetic assay
(e260
= 15,400 L molꢁ1 cmꢁ1 for SAM’s adenine moiety).
Characterization of kinetics of native and engineered MATs
LC–MS/MS-based MAT activity assay for heat map analysis
To determine the apparent kcat and Km values of native MATI
and MATII and their mutants with methionine or SAAMs as sub-
strates, the assays were carried out as described above (the LC–
The reactions of MATs and their mutants were carried out in a
reaction mixture containing 50 mM Tris–HCl (pH 8.0), 100 mM
MS/MS-based MAT assay) except that 100 ll of reaction mixture

Assay for MATs with SAM analogues as substrates / R. Wang et al. / Anal. Biochem. 450 (2014) 11–19
15
containing SAM or SAAMs (up to 4 mM) was used and linear initial
velocities were measured within 20 min with a 4-min interval for
native MATs with methionine as a substrate or within 60 min with
a 10-min interval for engineered MATs with SAAMs as substrates.
The steady-state kinetic parameters were obtained on fitting
the initial velocities (the production of SAM or SAM analogues vs.
the reaction time) against the concentrations of methionine or
SAAMs with GraphPad Prism 5 software according to the standard
Michaelis–Menten equation (v = kcat ꢀ [S]/(Km + [S]).
through the quantitative conversion of the reaction product SAM
(or SAM analogues) to MTA (or the corresponding S-alkylthioade-
nosines for SAM analogues) (55 °C in citrate buffer for 3.5 h as
implemented before for MTA synthesis; see Materials and Methods
and supplementary material) [30]. After the acidic degradation in a
96-well format, followed by centrifugation and adding SAH as an
internal standard, the reaction mixture was then analyzed by
HPLC–MS/MS. The amount of SAM (5 lM–1 mM in a 10-ll reaction
mixture) is linearly converted to MTA as quantified by HPLC–MS/
MS (data not shown). The LC–MS/MS-based assay was further val-
idated on comparing its kinetic data with those generated by the
conventional HPLC-based MAT assay (see results below).
Results and discussion
The newly developed LC–MS/MS-based activity assay for MATs
is highly sensitive and generally applicable. Under the current as-
Selection of SAAMs and engineered MATs
say setting, we were able to process a 2% aliquant of a 10-
tion sample (after 1:100 dilution) for the MS analysis and detected
as little as 0.1 pmol of MTA (equivalent to 2% aliquant of 5 M SAM
in a 10- l reaction mixture as described above). Such sensitivity is
3 orders of magnitude higher than that of the conventional
HPLC-coupled UV spectroscopy method, which typically requires
ll reac-
To systematically evaluate the substrate specificity of MAT vari-
ants, we prepared SAAMs 2 to 8 featured by their varied S-alkyl
substituents (Scheme 1). The less sterically hindered S-ethyl ana-
logue of methionine (SAAM 2) was previously reported to be an ac-
tive substrate of native MATs (yeast and rat), producing the
corresponding S-ethyl analogue of SAM [31,32]. Using a similar
synthetic method [28], SAAMs 3, 4, and 5 were obtained (see Mate-
rials and Methods and supplementary material). In addition, bulk-
ier SAAMs 6, 7, and 8 were also synthesized given that the
corresponding SAM analogues have been reported as SAM surro-
gates for multiple native or engineered methyltransferases
[17,19–21]. Briefly, the treatment of methionine with benzyl chlo-
l
l
a 300-ll reaction mixture for each time point [21,25]. The
LC–MS/MS-based assay format is convenient in comparison with
the conventional MAT assay using radioactive ATP or methionine
[26], which is sensitive but must deal with expensive and less envi-
ronmentally friendly materials. Our LC–MS/MS-based MAT activity
is also generally applicable to engineered MATs with diverse
SAAMs as substrates because the daughter ions of the correspond-
ing S-alkylthioadenosines can be readily selected with the readily
altered MS/MS detection mode (see Materials and Methods above
and the results below). In addition, our LC–MS/MS-based MAT
activity assay was designed in a mix-then-measure format and to
ride in concentrated HCl furnished S-benzyl-L-homocysteine [29].
Its benzyl moiety was then removed by sodium metal in liquid
ammonia. The resultant sodium thiolate was readily reacted with
the corresponding alkyl tosylates or bromides in dry ethanol to af-
ford SAAMs 3 to 8 (see Materials and Methods). The general syn-
thetic strategy for SAAMs 3 to 8 is featured by the use of the
handle a small volume of reaction samples (10 ll in the current
format), which can be further improved and adapted by automatic
liquid dispensers into a high-throughput format. In terms of sensi-
tivity and general application, our MAT activity assay is compara-
common precursor S-benzyl-L-homocysteine (Scheme 1). The col-
lection of methionine derivatives contains characteristic S-b-vinyl
substituents (e.g., 3–8) and terminal alkyne moiety (e.g., 6–8).
The former is essential for the efficient transalkylation from the
corresponding SAM analogues, whereas the latter, after transalky-
lation to the targets, is amenable to copper-catalyzed azide–alkyne
Huisgen cycloaddition (the click reaction) with azide-containing
probes for further characterization of the labeled targets
[9,17,21,23,33].
To identify and characterize potential MAT variants that can act
on SAAMs 2 to 8, we focused on the two isoforms of human MATs,
MATI and MATII, whose primary sequences are conserved across
species with liver-specific and ubiquitous localization (Fig. 3A),
respectively [34]. On aligning the MAT sequences of different spe-
cies, we can profile a cluster of conserved residues that account for
binding the methyl moiety of the substrate methionine (I117,
C120, V121, S247, and I322 for human MATI, PDB 2OBV; I117,
G120, V121, S247, and I322 for human MATII, PDB 2P02) (Figs. 1
and 3A) [16]. To expand this region to accommodate bulky SAAMs
for enzyme catalysis, we systematically replaced these residues
with less sterically hindered hydrophobic amino acids individually
(e.g., Ala, Gly) or in combination. Here a collection of 40 MAT mu-
tants was expressed and purified for further characterization.
ble to
a newly reported spectroscopic MAT assay, which
quantifies the enzyme-coupled production of NADPH from the
reaction byproduct pyrophosphate [27]. However, the latter assay
is not tolerant to the presence of interfering metabolites (e.g., pyro-
phosphate, glucose 6-phosphate) and, thus, is not suitable for eval-
uating native or engineered MATs in complex cellular settings. In
contrast, the LC–MS/MS-based MAT activity can be implemented
to examine native or engineered MATs under cellular settings by
selectively monitoring the ions of interest (see Materials and Meth-
ods). Admittedly, the current LC–MS/MS MAT assay requires access
to a specialized LC–MS/MS instrument, which might not be gener-
ally available. It remains to be determined whether the current
assay format can be adapted for other HPLC–MS instruments.
Identification of MAT variants recognizing SAAMs as substrates
After validating the MS-based activity MAT assay, we applied it
to screen the MAT mutants against SAAMs 2 to 8 in the presence of
a physiologically relevant concentration of ATP (2.5 mM) [35]. As
displayed in the heat map, the I117A variants of both MATI and
MATII demonstrated broad substrate specificity and high efficiency
in processing the SAAMs as substrates (Fig. 3B and C). Among the
examined mutant–substrate pairs, most of the MAT mutants
showed detectable activity with the less sterically hindered S-ethyl
analogue of methionine 2 as a substrate (Fig. 3B and C). Such wide
substrate specificity dropped dramatically from SAAM 3 to SAAM
5. In particular, many MAT mutants displayed barely detectable
ability to process bulkier methionine analogues such as SAAMs 6,
7, and 8. However, among the MAT mutants, the two I117A vari-
ants still showed relatively higher efficiency to act on all of the
examined SAAMs (Fig. 3B and C). Given that the native MATs show
Development of a sensitive, generally applicable LC–MS/MS-based
activity assay for MATs with SAAMs as substrates
To measure the activity of native MATs, many prior approaches
relied on HPLC to separate and quantify SAM via its characteristic
UV absorbance at 260 nm (e
260 = 15,400 L molꢁ1 cmꢁ1 for SAM’s
adenine moiety) [25]. However, given the inherent instability of
SAM (and of many SAM analogues as well) and the low sensitivity
of UV detection [18,19], we envisioned a more robust assay (Fig. 2)

16
Assay for MATs with SAM analogues as substrates / R. Wang et al. / Anal. Biochem. 450 (2014) 11–19
111
127
240
254
317
331
A
B
Human MATI EQQSPDIAQCVHLDRNE-----TVYHLQPSGRFVIGG-----QVSYAIGVAEPLSIS
EQQSPDIAQGVHLDRNE-----TIYHLQPSGRFVIGG-----QVSYAIGVSHPLSIS
EQQSPDIAQCVHLDRNE-----TIYHLQPSGRFVIGG-----QVSYAIGVAEPLSIS
EQQSPDIAQGVHLDRNE-----TIYHLQPSGRFVIGG-----QVSYAIGVSHPLSIS
Human MATII
Rat MATI
Rat MATII
O
O
C
=
R
t
trans
tarns
tarns
S
NH2
NH
2
HOOC
HOOC
MAT I
I117G
I117A
I117V
MAT II
I117G
I117A
I117V
I117L
I117L
C120G
C120A
V121G
V121A
S247G
S247A
I322G
I322A
I322V
I322L
V121G
V121A
V121L
S247G
S247A
I322G
I322A
I322V
I322L
I117AV121G
I117AV121A
I117AI322G
I117AI322A
I117AI322V
I117AI322L
I117AC120G
I117AC120A
I117AV121G
I117AV121A
100
0
Relative production
of SAM analogues
Fig. 3. Sequence alignment of MATs and reactivity heat map of MATI and MATII variants against SAAMs as substrates. (A) Sequence alignment of human and rat MATs.
Highlighted are the conserved residues (I117, C120, V121, S247, and I322 of human MATs), which account for binding the methyl moiety of the substrate methionine on the
basis of the structures of human MATI and MATII (PDB 2OBV and 2P02). (B,C) Native and MATI and MATII mutants were screened against SAAMs 2 to 8 in a combinatorial
manner. The results for MATI variants (B) and MATII variants (C) are presented in a heat map format. The amount of the produced SAM analogues was normalized to that of
the ethyl-SAM produced by the corresponding I117A mutants.
barely detectable activity toward SAAMs 6 to 8 (Fig. 3B and C),
mutating I117 is essential for MATs to gain the ability to recognize
these bulky SAAMs.
difference for the same set of SAAMs. On comparing the overall
catalysis efficiency (kcat/Km values), methionine 1 and most SAAMs
80
60
A
MATI/Methionine 1
Kinetic analysis of native and I117A MATs
40
20
Because the I117A mutants of MATI and MATII are most effec-
tive in processing SAAMs 2 to 8, we implemented the above-estab-
lished LC–MS/MS MAT activity assay to characterize their kinetics
(Fig. 4, S8, and Table 1, S1). Apparent kinetic parameters obtained
for native MATI, MATII, and methionine using the newly developed
assay are consistent with those obtained using the conventional
HPLC-based MAT assay except that a much greater amount of
materials needed to be used in the latter assay [25]. Thus, this con-
sistency validated the robustness of the current LC–MS/MS-based
MAT assay for the current application (Fig. 4, S8, and Table 1,
S1). Consistent with the heat map comparison (Fig. 3B and C),
the catalytic turnover (kcat values) of native MATI and MATII
dropped approximately 50-fold with SAAM 2 and was barely mea-
surable for the bulkier SAAMs 3 to 8. Although kcat values of the
I117A variants of MATI and MATII are 5-fold lower than those of
native MATI and MATII toward methionine, the I117A mutants
showed comparable kcat values between methionine and SAAM 2
(<2-fold decrease of its kcat value). The broad substrate specificity
of the MAT I117A variants is further maintained for the bulkier
SAAMs 3 to 5 and SAAM 7 (3- to 7-fold difference in their kcat val-
ues from 2). For the I117A variants, SAAMs 6 and 8, which contain
the sterically rigid alkynyl-trans-vinyl moiety and the bulkiest
propargyl-oxy-trans-butenyl moiety, respectively, are the least
reactive substrates, as shown by their 30- to 100-fold decrease of
kcat values, in contrast to methionine 1 and SAAM 2.
15
10
5
MATI I117A/Methionine1
MATI I117A/SAAM 2
MATI I117A/SAAM 7
0
0
1
2
3
4
[Analogue] (mM)
60
50
40
B
MATII/Methionine 1
30
20
15
10
5
MATII I117A/Methionine 1
MATII I117A/SAAM 2
MATII I117A/SAAM 7
0
0
1
2
3
4
[Analogue] (mM)
Fig. 4. Representative Michaelis–Menten kinetics of methionine analogues paired
with MATI variants (A) and MATII variants (B). Here kcat = 11.5 minꢁ1 and
Km = 0.73 mM for native MATI with methionine 1 as substrate; kcat = 1.9 0.1,
2.1 0.2, and 0.59 0.6 minꢁ1 and Km = 1.0 0.2, 2.0 0.5, and 1.4 0.4 mM for
MATI I117A with methionine 1, SAAM 2, and SAAM 7 as substrates, respectively;
kcat = 8.8 0.6 minꢁ1 and Km = 0.6 0.1 mM for native MATII with methionine 1 as
substrate; kcat = 1.9 0.1, 1.1 0.1, and 0.62 0.05 minꢁ1 and Km = 1.8 0.3,
1.0 0.4, and 2.5 0.4 mM for MATII I117A with methionine 1, SAAM 2, and SAAM
7 as substrates, respectively. See Fig. S8 in the supplementary material for the
complete set of kinetic data.
Despite up to 100-fold variation of the kcat values of native MATI
and MATII and their I117A variants with methionine and SAAM as
substrates, the corresponding Km values are within the 2- to 3-fold

Assay for MATs with SAM analogues as substrates / R. Wang et al. / Anal. Biochem. 450 (2014) 11–19
17
Table 1
Relative kcat and Km values of native MATI and MATII and their I117A variants with methionine 1 or SAAMs 2 to 8
.
SAAM:
1
2
3
4
5
6
7
8
Native MATI
kcat = 1.0
Km = 1.0
kcat = 0.017
Km = 2.3
ND
ND
ND
ND
ND
ND
kcat/Km = 1.0
kcat = 0.17
Km = 1.4
kcat/Km = 0.12
kcat = 1.0
kcat/Km = 0.0070
kcat = 0.18
Km = 2.8
kcat/Km = 0.064
kcat = 0.015
Km = 1.7
kcat/Km = 0.0081
kcat = 0.12
MATI I117A
Native MATII
MATII I117A
kcat = 0.052
Km = 1.2
kcat/Km = 0.042
ND
kcat = 0.060
Km = 0.58
kcat/Km = 0.11
ND
kcat = 0.024
Km = 1.0
kcat/Km = 0.024
ND
kcat = 0.010
Km = 0.81
kcat/Km = 0.013
ND
kcat = 0.051
Km = 1.9
kcat/Km = 0.027
ND
kcat = 0.0065
Km = 5.1
kcat/Km = 0.0013
ND
Km = 1.0
kcat/Km = 1.0
kcat = 0.22
Km = 3.0
kcat = 0.034
Km = 0.80
kcat = 0.062
Km = 8.5
kcat = 0.033
Km = 5.3
kcat = 0.0025
Km = 10
kcat = 0.071
Km = 4.3
kcat = 0.011
Km = 4.5
Km = 1.7
kcat/Km = 0.074
kcat/Km = 0.070
kcat/Km = 0.042
kcat/Km = 0.0074
kcat/Km = 0.0063
kcat/Km = 0.00025
kcat/Km = 0.016
kcat/Km = 0.0024
Note. Apparent kinetic parameters (kcat and Km values) were obtained in the presence of 2.5 mM ATP and varied concentrations of SAAMs as substrates. The relative kinetic
parameters were obtained by dividing their kcat and Km values by those of the I117A variants. Here kcat = 11.5 0.7 minꢁ1 and Km = 0.7 0.1 mM for native MATI and
methionine; kcat = 8.8 0.6 minꢁ1 and Km = 0.6 0.1 mM for native MATII and methionine. The kinetic parameters of native MATs with methionine were validated by the
conventional HPLC assay (see Fig. S9 in supplementary material). See Fig. 4 for Michaelis–Menten curves of the representative SAAM–mutant pairs. ND, not detectable. See
Table S1 for the complete set of kinetic data with error bars.
(2–7 but not 8) are suitable substrates of the I117A variant of MATI
(with 10-fold difference of the kcat/Km values for 2–7). In contrast,
the kcat/Km values of the I117A variant of MATII is more substrate
dependent with the higher preference for SAAMs 1 to 4 and SAAM
7 than for SAAMs 5, 6, and 8 (<5-fold vs. 30- to 100-fold drop of
their kcat/Km values). Among the SAAMs containing terminal-
alkyne amenable to the click reaction (6–8), SAAM 7 is the most
reactive substrate for both MATI and MATII I117A variants, with
a slight preference for the former (<2-fold difference of the
kcat/Km values between the two variants; 2- to 100-fold difference
of the kcat/Km values between SAAM 7 and SAAM 6 or 8).
mutants to act on these compounds as substrates. In contrast to
the less than 10-fold variation of Km values, which reflects the en-
zyme–substrate binding along the reaction path, the more than
100-fold fluctuation of kcat values (Fig. 4 and Table 1) plays a more
significant role in the overall enzyme catalysis. The MAT-catalyzed
adenylation on methionine is expected to go through a linear SN2
transition state with methionine and pyrophosphate as the nucle-
ophile and leaving group, respectively (Figs. 1 and 5) [36]. Given
that a kcat value likely reflects the enzymatic efficiency at this
chemical step, the activities of the MATI and MATII I117A mutants
on the SAAMs can be ranked as 2 > 3, 4, 5, 7 > 6, 8 according to their
respective kcat values. Such a trend features the SAAMs containing
smaller S-alkyl substituents as the more active substrates (e.g., 2,
7) versus those containing bulky or rigid S-alkyl substituents as
more inert substrates (e.g., 6, 8). We argued that this difference
likely arises from the ability of the former to maintain the linear
SN2 transition state of native MATI and MATII, as reflected by the
higher kcat value at the chemical step (Fig. 5). In contrast, the bulk-
iness or rigidity may restrict the latter from forming an optimized
linear SN2 transition state along the reaction path (Fig. 5).
Relevant steady-state parameters of engineered MATs with bulky
SAAMs as substrates
The methionine binding pocket of native human MATI and MA-
TII appears not to be spacious enough to accommodate bulky
SAAMs, as reflected by the significant decrease or complete loss
of enzymatic activities toward the bulky SAAMs 2 to 8. In contrast,
removal of several bulky amino acids (e.g., I117, C120, V121, S247,
and I332 in Fig. 3) adjacent to the S-methyl binding site of the sub-
strate methionine readily expands the binding pocket to accommo-
date these bulky SAAMs (Fig. 3). The slight variation of Km values
(<10-fold) of the most active MATI and MATII I117A mutants for
the SAAMs (Km values in Table 1) further argues that these MATI
and MATII variants gain comparable affinity to the structurally di-
verse SAAMs as substrates. This result is remarkable given the
small decrease of size from Ile to Ala versus the significant increase
of size from methionine to SAAM 8. Here we also noticed that the
heat maps for SAAMs agree well between native MATI and MATII,
their I117A mutants, but not their I117V mutants (Fig. 3B). The
striking gain-of-function activities of certain MATI and MATII mu-
tants toward SAAMs and the difference for others (e.g., I117V)
could be attributed to specific conformational changes associated
with each variant. However, more structural characterization is re-
quired to provide a molecular-level rationale about how a single
point mutation can lead to such difference among the closely
related MAT variants.
Most active MAT variants with matched SAAMs as substrates
On comparing kcat, Km, and kcat/Km values for the panel of sub-
strates (methionine or SAAMs), we concluded that the MATI and
MATII 17A mutants have the highest activities toward methionine
(Table 1). Here SAAM 7 is only 4-fold less reactive than methionine
as the substrate of the MATI and MATII I117A mutants. This result
is consistent with the prior observation that SAAM 7 can be pro-
cessed by the MATII I117A mutant to produce the corresponding
SAM analogue within living cells [16]. The prior work also showed
that SAAM 7 is a better substrate of the MATII variant than is SAAM
8 [16]. Such a finding is readily reflected by the 7-fold higher kcat
(or kcat/Km) value of SAAM 7 in comparison with SAAM 8. The intra-
cellular concentration of methionine varies from 150 to 280 pmol/
107 cells [37]. In previous work, we showed that the incubation of
1 mM SAAM 7 in combination with methionine depletion in the
growth medium is sufficient for the MATII I117A variant to pro-
duce the corresponding SAM analogue with its intracellular con-
centration comparable to that of SAM in a native setting [16].
Our data further suggest that the binding of SAAMs 2 to 8,
although necessary, is insufficient for the MATI and MATII I117A

18
Assay for MATs with SAM analogues as substrates / R. Wang et al. / Anal. Biochem. 450 (2014) 11–19
O
O
O
P
-O
H3N
H
H
V121
Active transition state of
S
O
MATI/II with native substrates
Me
O
G120
HO
N
I
117
N
N
I332
HO
N
NH2
O
O
O
O
O
P
-O
H3N
O
P
-O
H
H
H
H
H3N
S
V121
S
O
O
V121
R'
A117
R
O
G120
O
I332
HO
G120
HO
N
N
A117
332
I
N
NH2
N
N
N
HO
HO
N
N
NH2
Transition state of engineeredMATI/II
with less active substrate analogues
Transition state of engineered MATI/II
with active substrate analogues
Fig. 5. Expected transition state structures of MAT-catalyzed adenylation. Top panel: Proposed linear SN2 transition state of the native enzyme with methionine and ATP as
substrates; Bottom panel: Expected transition states of engineered MATs with methionine analogues and ATP as substrates. The active transition state (left) is featured by its
linear SN2 configuration, in contrast to the less active enzyme–substrate intermediate (right).
The comparable kcat, Km, and kcat/Km values between SAAMs and
methionine (4-fold difference for SAAM 7) for the MATI and MATII
I117A variants (Table 1) suggest that complete depletion of methi-
onine might not be necessary for in-cell production of the corre-
sponding SAM analogues.
methyltransferases [12,13,16–18], the revealed structure–activity
relationship of MAT variants and SAAMs can provide further guid-
ance to couple two sets of enzymes for more efficient substrate
labeling inside living cells. The high sensitivity and robustness of
the current MS-based MAT assay also permits the analysis of the
corresponding metabolites in cell lysates.
Conclusion
Acknowledgments
Here we have described a sensitive HPLC–MS/MS-based MAT
assay to characterize native and engineered MATs with methionine
and SAAMs as substrates. In contrast to the several conventional
MAT activity assays, which are of low sensitivity and have limited
application, our assay is characterized by its high sensitivity and
general applicability. With the aid of this robust assay, more than
40 MATI and MATII mutants were screened against SAM and 7
SAAMs. Here MATI I117A was identified to be the most active MATI
variant by acting on structurally diverse SAAMs. The relatively con-
stant Km values versus largely altered kcat values of MATI I117A
across SAAMs 2 to 8 further argue that, although the MAT variant
is able to bind bulky SAAMs with comparable affinity as native
MATI and MATII bind methionine, such binding is not sufficient
to promote enzymatic catalysis. The most reactive substrates, such
as 3 and 7 for the MATI and MATII I117A variants, may be featured
by their ready formation of the active SN2 transition state. In con-
trast, the less reactive substrates, such as 6 and 8, cannot be pro-
cessed because of the potential disruption of the enzyme
transition states by their structurally rigid or sterically hindered
S-alkyl substituents. These results collectively argue that the flex-
ibility and size of S-alkyl substituents are important parameters
for SAAMs to be recognized as substrates by the engineered MATs.
More important, the current work enabled us to identify the I117A
variant of MATI as a more general MAT variant to process SAAMs
into corresponding SAM analogues.
We thank Kabirul Islam for providing several SAAMs, Udo
Oppermann for providing MAT plasmids, and Tony Taldone for
developing the HPLC–MS/MS method. We thank Ian Bothwell for
comments on the manuscript. We are grateful for the financial sup-
port from the National Institute of General Medical Sciences
(1R01GM096056), the National Institute of Health (NIH) Director’s
New Innovator Award Program (1DP2-OD007335), the March of
Dimes Foundation (Basil O’Connor Starter Scholar Award), the
Starr Cancer Consortium, and the Alfred W. Bressler Scholars
Endowment Fund.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.ab.2013.12.026.
References
[1] S. Klimasauskas, E. Weinhold,
A new tool for biotechnology: AdoMet-
dependent methyltransferases, Trends Biotechnol. 25 (2007) 99–104.
[2] S.C. Lu, S-Adenosylmethionine, Int. J. Biochem. Cell Biol. 32 (2000) 391–395.
[3] J. Kim, H. Xiao, J.B. Bonanno, C. Kalyanaraman, S. Brown, X. Tang, N.F. Al-Obaidi,
Y. Patskovsky, P.C. Babbitt, M.P. Jacobson, Y.S. Lee, S.C. Almo, Structure-guided
discovery of the metabolite carboxy-SAM that modulates tRNA function,
Nature 498 (2013) 123–126.
[4] B.J. Landgraf, S.J. Booker, Biochemistry: the ylide has landed, Nature 498 (2013)
45–47.
[5] F. Yan, J.M. LaMarre, R. Röhrich, J. Wiesner, H. Jomaa, A.S. Mankin, D.G.
Fujimori, RlmN and Cfr are radical SAM enzymes involved in methylation of
ribosomal RNA, J. Am. Chem. Soc. 132 (2010) 3953–3964.
As the primary biological methyl donor, SAM plays a pivotal
role under a variety of physiological settings. For instance, protein
methyltransferases, which transfer the methyl group from SAM to
the specific Lys or Arg residues of substrates, are essential for epi-
genetic regulation [12,13]. Given the potential utility of the termi-
nal-alkyne-containing SAM analogues to label substrates of
[6] Y. Zhang, X. Zhu, A.T. Torelli, M. Lee, B. Dzikovski, R.M. Koralewski, E. Wang, J.
Freed, C. Krebs, S.E. Ealick, H. Lin, Diphthamide biosynthesis requires an
organic radical generated by an iron–sulphur enzyme, Nature 465 (2010) 891–
896.

Assay for MATs with SAM analogues as substrates / R. Wang et al. / Anal. Biochem. 450 (2014) 11–19
19
[7] A.W. Struck, M.L. Thompson, L.S. Wong, J. Micklefield, S-Adenosyl-
L
-
[22] J.M. Lipson, M. Thomsen, B.S. Moore, R.P. Clausen, J.J. La Clair, M.D. Burkart, A
tandem chemoenzymatic methylation by S-adenosyl- -methionine,
ChemBioChem 14 (2013) 950–953.
[23] C. Dalhoff, G. Lukinavicius, S. Klimasa˘uskas, E. Weinhold, Direct transfer of
extended groups from synthetic cofactors by DNA methyltransferases, Nat.
Chem. Biol. 2 (2006) 31–32.
[24] G. Cantoni, The nature of the active methyl donor formed enzymatically from
l-methionine and adenosinephosphate, J. Am. Chem. Soc. 74 (1952) 2942–
2943.
[25] V. Kamarthapu, K.V. Rao, P.N. Srinivas, G.B. Reddy, V.D. Reddy, Structural and
kinetic properties of Bacillus subtilis S-adenosylmethionine synthetase
expressed in Escherichia coli, Biochim. Biophys. Acta 1784 (2008) 1949–1958.
[26] G.D. Markham, E.W. Hafner, C.W. Tabor, H. Tabor, S-Adenosylmethionine
synthetase from Escherichia coli, J. Biol. Chem. 255 (1980) 9082–9092.
[27] S.P. Zano, P. Bhansali, A. Luniwal, R.E. Viola, Alternative substrates selective for
S-adenosylmethionine synthetases from pathogenic bacteria, Arch. Biochem.
Biophys. 536 (2013) 64–71.
methionine-dependent methyltransferases: highly versatile enzymes in
biocatalysis, biosynthesis, and other biotechnological applications,
ChemBioChem 13 (2012) 2642–2655.
L
[8] Q. Lin, F. Jiang, P.G. Schultz, N.S. Gray, Design of allele-specific protein
methyltransferase inhibitors, J. Am. Chem. Soc. 123 (2001) 11608–11613.
[9] W. Peters, S. Willnow, M. Duisken, H. Kleine, T. Macherey, K.E. Duncan, D.W.
Litchfield, B. Lüscher, E. Weinhold, Enzymatic site-specific functionalization of
protein methyltransferase substrates with alkynes for click labeling, Angew.
Chem. Int. Ed. 49 (2010) 5170–5173.
[10] J. Li, H. Wei, M.M. Zhou, Structure-guided design of a methyl donor cofactor
that controls a viral histone H3 lysine 27 methyltransferase activity, J. Med.
Chem. 54 (2011) 7734–7738.
[11] D.F. Iwig, A.T. Grippe, T.A. McIntyre, S.J. Booker, Isotope and elemental effects
indicate a rate-limiting methyl transfer as the initial step in the reaction
catalyzed by Escherichia coli cyclopropane fatty acid synthase, Biochemistry 43
(2004) 13510–13524.
[12] R. Wang, M. Luo,
A
journey toward bioorthogonal profiling of protein
[28] J. Jiracek, M. Collinsova, I. Rosenberg, M. Budesinsky, E. Protivinska, H.
Netusilova, T.A. Garrow, S-Alkylated homocysteine derivatives: new
inhibitors of human betaine–homocysteine S-methyltransferase, J. Med.
Chem. 49 (2006) 3982–3989.
methylation inside living cells, Curr. Opin. Chem. Biol. 17 (2013) 729–731.
[13] M. Luo, Current chemical biology approaches to interrogate protein
methyltransferases, ACS Chem. Biol. 7 (2012) 443–463.
[14] J.M. Winter, G. Chiou, I.R. Bothwell, W. Xu, N.K. Garg, M. Luo, Y. Tang,
Expanding the structural diversity of polyketides by exploring the cofactor
tolerance of an inline methyltransferase domain, Org. Lett. 15 (2013) 3774–
3777.
[29] B. Peric Simov, F. Wuggenig, K. Mereiter, H. Andres, J. France, P. Schnelli, F.
Hammerschmidt, Direct chemical synthesis of chiral methanol of 98% ee and
its conversion to [2H1,3H]methyl tosylate and [2H1,3H-methyl]methionine, J.
Am. Chem. Soc. 127 (2005) 13934–13940.
[15] K. Islam, Y. Chen, H. Wu, I.R. Bothwell, G.J. Blum, H. Zeng, A. Dong, W. Zheng, J.
Min, H. Deng, M. Luo, Defining efficient enzyme–cofactor pairs for
bioorthogonal profiling of protein methylation, Proc. Natl. Acad. Sci. USA 110
(2013) 16778–16783.
[30] J.A. Gutierrez, M. Luo, V. Singh, L. Li, R.L. Brown, G.E. Norris, G.B. Evans, R.H.
Furneaux, P.C. Tyler, G.F. Painter, D.H. Lenz, V.L. Schramm, Picomolar inhibitors
as transition-state probes of 50-methylthioadenosine nucleosidases, ACS
Chem. Biol. 2 (2007) 725–734.
[16] R. Wang, K. Islam, Y. Liu, W. Zheng, H. Tang, N. Lailler, G. Blum, H. Deng, M. Luo,
[31] J.H. Alix, Molecular aspects of the in vivo and in vitro effects of ethionine, an
analog of methionine, Microbiol. Rev. 46 (1982) 281–295.
Profiling
genome-wide
chromatin
methylation
with
engineered
posttranslation apparatus within living cells, J. Am. Chem. Soc. 135 (2013)
1048–1056.
[17] K. Islam, I. Bothwell, Y. Chen, C. Sengelaub, R. Wang, H. Deng, M. Luo,
Bioorthogonal profiling of protein methylation using azido derivative of S-
[32] W.R. Leopold, J.A. Miller, E.C. Miller, Comparison of some carcinogenic,
mutagenic, and biochemical properties of S-vinylhomocysteine and
ethionine, Cancer Res. 42 (1982) 4364–4374.
[33] G. Blum, I.R. Bothwell, K. Islam, M. Luo, Profiling protein methylation with
cofactor analog containing terminal alkyne functionality, Curr. Protoc. Chem.
Biol. 5 (2013) 67–88.
[34] J.M. Mato, L. Alvarez, P. Ortiz, M.A. Pajares, S-Adenosylmethionine synthesis:
molecular mechanisms and clinical implications, Pharmacol. Ther. 73 (1997)
265–280.
[35] F.M. Gribble, G. Loussouarn, S.J. Tucker, C. Zhao, C.G. Nichols, F.M. Ashcroft, A
novel method for measurement of submembrane ATP concentration, J. Biol.
Chem. 275 (2000) 30046–30049.
adenosyl-
[18] I.R. Bothwell, K. Islam, Y. Chen, W. Zheng, G. Blum, H. Deng, M. Luo, Se-
adenosyl-l-selenomethionine cofactor analogue as reporter of protein
methylation, J. Am. Chem. Soc. 134 (2012) 14905–14912.
L-methionine, J. Am. Chem. Soc. 134 (2012) 5909–5915.
a
[19] K. Islam, W. Zheng, H. Yu, H. Deng, M. Luo, Expanding cofactor repertoire of
protein lysine methyltransferase for substrate labeling, ACS Chem. Biol. 6
(2011) 679–684.
[20] R. Wang, G. Ibáñez, K. Islam, W. Zheng, G. Blum, C. Sengelaub, M. Luo,
Formulating
a
fluorogenic assay to evaluate S-adenosyl-
L
-methionine
(2011)
[36] G.D. Markham, F. Takusagawa, A.M. Dijulio, C.W. Bock, An investigation of the
catalytic mechanism of S-adenosylmethionine synthetase by QM/MM
calculations, Arch. Biochem. Biophys. 492 (2009) 82–92.
analogues as protein methyltransferase cofactors, Mol. BioSyst.
2970–2981.
7
[21] R. Wang, W. Zheng, H. Yu, H. Deng, M. Luo, Labeling substrates of protein
[37] R. Carmel, D.W. Jacobsen (Eds.), Homocysteine in Health and Disease,
Cambridge University Press, Cambridge, UK, 2001.
arginine methyltransferase with engineered enzymes and matched S-
adenosyl-L-methionine analogues, J. Am. Chem. Soc. 133 (2011) 7648–7651.