Rationally engineered variants of S-adenosylmethionine (SAM) synthase: Reduced product inhibition and synthesis of artificial cofactor homologues

Article abstract of DOI:10.1039/c4cc08478k

S-Adenosylmethionine (SAM) synthase was engineered for biocatalytic production of SAM and long-chain analogues by rational re-design. Substitution of two conserved isoleucine residues extended the substrate spectrum of the enzyme to artificial S-alkylhomocysteines. The variants proved to be beneficial in preparative synthesis of SAM (and analogues) due to a much reduced product inhibition. This journal is

Full text of DOI:10.1039/c4cc08478k

ChemComm
View Article Online
View Journal
COMMUNICATION
Rationally engineered variants of
S-adenosylmethionine (SAM) synthase: reduced
product inhibition and synthesis of artificial
cofactor homologues†
Cite this:DOI: 10.1039/c4cc08478k
Received 27th October 2014,
Accepted 2nd January 2015
M. Dippe, W. Brandt, H. Rost, A. Porzel, J. Schmidt and L. A. Wessjohann*
DOI: 10.1039/c4cc08478k
www.rsc.org/chemcomm
S-Adenosylmethionine (SAM) synthase was engineered for biocatalytic
production of SAM and long-chain analogues by rational re-design.
Substitution of two conserved isoleucine residues extended the sub-
strate spectrum of the enzyme to artificial S-alkylhomocysteines. The
variants proved to be beneficial in preparative synthesis of SAM (and
analogues) due to a much reduced product inhibition.
S-Adenosylmethionine synthase (SAMS, E.C. 2.5.1.6) is a key
enzyme in biological methyl transfer and catalyzes the formation of
S-adenosylmethionine (SAM) from L-methionine and ATP (Fig. 1).
Since only very few enzymes rely on other donors with reactive
methyl groups,1–4 SAM serves as the general cofactor for cellular
methyltransferases in all organisms.1,2,5 Thus, SAM is involved in a
multitude of biological processes such as DNA modification6 or
neurotransmitter metabolism in the brain.7 In plants, it is required
Fig. 1 Scheme of the SAMS reaction. From the two substrates, ATP and
L-methionine (left), SAM (right, top) is produced. The co-products phosphate
and diphosphate (right, below) originate from the triphosphate moiety of ATP.
for major metabolic pathways like lignin biosynthesis8 and also for leading to slow or incomplete methylation. This is relevant in
the modification of secondary metabolites.1–3
producer strains requiring SAM in high quantity. This limita-
In particular, phenylpropanoids, isoprenoids or alkaloids tion might be partially circumvented by over-expression of the
are often decorated by a specific methylation pattern which is SAMS enzyme of the expression host,9 or by (multiple) inte-
essential for their biological activity.1,3 Hence, synthetic routes gration of a recombinant enzyme with higher activity. However,
to these valuable compounds often necessitate at least one commonly SAMS enzymes are not suitable for this strategy due
methylation step. In many of these reactions, enzymatic methyl to low specific activity or inhibition by their product SAM, thus
transfer has proven to be superior to traditional chemical limiting its own production in vivo and in vitro, despite suffi-
methods because of strict chemo- and regioselectivity. However, cient availability of substrates and biocatalyst.10,11
the application of methyltransferases is still limited since the
In addition, this type of metabolic engineering does not allow
biologically active epimer of their costly cofactor SAM is hardly transfer of alkyl chains other than methyl residues. Most eubacterial
accessible, i.e. too expensive if produced by chemical synthesis. SAMS enzymes do not tolerate changes in the structure of their
Thus, methylation by living whole-cell biotransformation is a amino acid substrate, and thus they lack the ability to synthesize
promising alternative. The methyl acceptor is added to a cofactor homologues or analogues from artificial methionine
suspension of microorganisms which express the recombinant derivatives. From the large number of those SAMS enzymes which
enzymes and provide SAM from internal or fed L-methionine. have been characterized, only two proteins tolerably convert
However, this strategy has two major drawbacks. The intra- S-ethyl-L-homocysteine (ethionine) to S-adenosylethionine.11,12
cellular concentration of endogenously available SAM often can In contrast, an increasing number of enzymes from archaea
be flux-determining for subsequent methyl transfer reactions, and mammals were recently found to convert a broad range of
non-natural substrates, e.g. those from Sulfolobus solfataricus,13
Methanocaldococcus jannaschii14 or the human isoform MAT2A.14
Leibniz-Institute of Plant Biochemistry, Department of Bioorganic Chemistry,
These enzymes proved to be useful tools in the synthesis of SAM
Weinberg 3, D-06120 Halle, Germany. E-mail: wessjohann@ipb-halle.de
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4cc08478k derivatives from methionine homologues and the corresponding
This journal is ©The Royal Society of Chemistry 2015
Chem. Commun.

View Article Online
Communication
ChemComm
selenium compounds.13,14 The molecular basis of this unusual
promiscuity, however, is still subject of debate. Ten structures
of SAMS (including enzymes from eubacterial, archaeal and
mammalian origin) have been published so far,13 implicating that
relaxed substrate specificities might be related to the architecture
and dimension of the active site13,15 and/or to interaction with a
flexible neighboring loop region upon substrate binding.13,16,17
Due to this indeterminancy in structure-activity relationships,
studies on engineering of SAMS enzymes for activity, substrate
specificity or product inhibition are scarce.15,18
Based on the recent informations from literature, X-ray and
homology models (v.i.), we believed to have a basis for the
rational re-design of a SAMS to produce a practically useful
variant suitable for synthesis of SAM homologues. The SAMS
from Bacillus subtilis (strain ATCC 6051) was chosen because
this microbial protein has been shown to be efficiently pro-
duced in Escherichia coli.19 Accordingly, high yields (59 mg lÀ1
of culture) of active and chromatographically homogeneous
protein (see Fig. S1 in the ESI†) could be obtained after affinity
purification. The enzyme was active towards the L-enantiomer
of methionine, and towards a racemic mixture of both stereo-
isomeric forms of the amino acid. The reaction rates as a
function of the substrate concentration (Fig. 2a) are characteri-
zed by sigmoid kinetics, which can be well described by the Hill
function. The resulting kinetic parameters are shown in Table S1
in the (ESI†). The protein specifically converted the L-enantiomer
of methionine because the concentration of this substrate at half-
maximum saturation (S0.5) was nearly twofold lower than that for
D,L-methionine. However, the maximum rates (Vmax) and Hill
coefficients (h) observed for L-methionine and racemic methionine
were similar. Therefore, D-methionine is not inhibitory to SAMS.
The enzyme was also tested for conversion of a series of
methionine derivatives (Fig. 3) which differ from the natural
substrate by introduction of deuterium atoms (1), in the length
of their linear (2–4) or branched (5–7) thioether-bound alkyl
chain as well as in functionalization by double bonds (8, 9) or
hydrophilic groups (10, 11). As the enzyme was comparably
active towards D,L-methionine-(methyl-D3) (Fig. 3, 1) and non-
deuterated methionine (Fig. 2a), it proved to be well suitable for
the production of isotope-labeled SAM. Similar to other eubac-
terial SAMS enzymes, none of the other artificial amino acids
was accepted as substrate. This result could be explained by a
model of the tertiary structure of the enzyme which is based on
the crystal structure of the homologous protein from E. coli.16
Fig. 4a shows the ternary complex of the protein with its product
SAM. A detailed view of the bound SAM (Fig. 4b) indicates that it is
Fig. 2 Rates of conversion of methionine and derivatives catalyzed by the
SAMS enzyme from B. subtilis (a), and its variants I317V (b), I317A (c), I105V/
I317A (d) or I105A/I317A (e) as a function of the substrate concentration.
The reaction with L-methionine (J), D,L-methionine (K), D,L-methionine-
(methyl-D3) ( ), D,L-ethionine ( ), S-n-propyl-D,L-homocysteine (,), S-n-
butyl-D,L-homocysteine (.) or S-(prop-1-en-1-yl)-D,L-homocysteine ( ) was
assessed by determination of phosphate released from the co-substrate ATP
during the reaction. Curves were fitted according to the Hill model.
Fig. 3 S-Alkylhomocysteines used in this study. 1 D,L-methionine-(methyl-D3),
2
D,L-ethionine, 3 D,L-n-propionine, 4 D,L-n-buthionine, 5 D,L-isopropionine,
6 S-(1-methylpropyl)-D,L-homocysteine, 7 S-(2-methylpropyl)-D,L-homocysteine,
recognized by the formation of several hydrogen bonds with the 8 S-(prop-1-enyl)-D,L-homocysteine, 9 S-(propa-1,2-dienyl)-D,L-homocysteine,
10 S-(2-hydroxyethyl)-D,L-homocysteine, 11 S-(2-carboxymethyl)-D,L-
homocysteine.
enzyme. On a closer inspection of this binding site the two
isoleucines I105 and I317 appeared as residues which directly
interact with the methyl group of SAM. Therefore, it seems likely
that these residues disturb product formation from methionine which are conserved in enzymes from eubacteria, but also in the
derivatives with more bulky substituents. However, both iso- majority of archaeal and eukaryotic proteins (see Fig. S2 in ESI†)
leucines do not seem to be restrictive for the accommodation of may largely contribute to the inability of many SAMS to convert
the artificial amino acid substrate in the active site (Fig. 4c) prior extended methionine derivatives.
to adenosylation, which indicates a conformational shift of the
To prove this hypothesis, the two positions were exchanged
protein during catalysis. In conclusion, the isoleucine residues against less voluminous residues by site-directed mutagenesis.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2015

View Article Online
ChemComm
Communication
Fig. 4 Model of the SAM synthase from B. subtilis. (a) Tertiary structure of the tetramer with SAM (space fill representations) bound between two
monomer units. (b) Active site with the bound product SAM (green carbon atoms). The space for docking of SAM derivatives with larger functional groups
R is restricted by the two isoleucines I105 and I317 which were therefore subject of site-directed mutagenesis. (c) Arrangement of the substrate
analogues S-n-butyl-L-homocysteine (magenta carbon atoms) and adenylyl imidodiphosphate (green carbon atoms) in the active center. The red arrow
indicates the nucleophilic attack during catalysis. The reaction is accompanied with conformational change of the protein, which will lead to a steric clash
of the butyl with the isoleucine-317 side chain (black arrow).
First, the residue I317 which has the lowest distance to the
Subsequent to enzyme characterization, the suitability of the
methyl group of the SAM product (3.7 Å) was substituted to valine. wild-type protein and of variants with appropriate substrate
The resulting enzyme variant shows a markedly enhanced activity specificity (I317V and I317A) for the synthesis of SAM and
towards L-methionine and D,L-methionine (Fig. 2b). The S0.5 values analogues was evaluated by high-performance thin-layer chromato-
for both substrates (Table S1 in ESI†), however, were similar to the graphy (see Fig. S3 in the ESI†). Since the enzymes were not
wild-type protein. Due to its more spacious active site, the variant inhibited by the D-enantiomer of the amino acid substrates
was also tolerably active (Fig. 2b) towards D,L-ethionine (2). In a (Table S1, ESI†), methionine and its derivatives could be advanta-
second variant, I317 was exchanged by the much less voluminous geously applied as racemic mixture of its stereoisomers in these
alanine. Accordingly, the enzyme accepted an even broader reactions. The time course of SAM formation by the wild-type enzyme
spectrum of substrates (Fig. 2c) including methionine, ethionine is shown in Fig. 5a. As described for several other SAMS proteins, the
and the bulky homologues S-propyl- and S-butyl-homocysteine enzyme is inhibited by its product SAM10,11,20 which results in
(3 and 4, resp).‡ The conversion of all substrates showed almost stagnating conversion and low yield. In contrast, this product inhibi-
identical maximum rates (Table S1 in ESI†). Even the unsaturated tion is completely reduced in the variant I317V (Fig. 5b) and less
substrate S-(prop-1-enyl)-D,L-homocysteine (8) was accepted by the pronounced in case of the mutant enzyme I317A (Fig. 5c).
enzyme variant (Fig. 2c).
Accordingly, transformation of methionine and ethionine
Despite its relaxed substrate specificity, the enantioselectivity by SAMS-I317V in preparative synthetic reactions led to high
of the variant towards the amino acid substrate was similar to conversion rates for the amino acid substrates. The syntheses
that of the wild-type enzyme, e.g. in the conversion of methionine
(as judged from S0.5 values for L- and racemic methionine, see
Table S1 in ESI†). Likewise, from both stereoisomers of buthio-
nine the substrate with L-configuration was exclusively converted
(21.2 Æ 0.1 nmol minÀ1 mgÀ1 at a substrate concentration of
5 mM). To further probe the role of the proximate isoleucine
residue I105 in substrate recognition, the variants I105V/I317A
and I105A/I317A were generated. With increasing size of the
active center of the enzymes, conversion of the natural substrate
methionine (Fig. 2d and e) is progressively reduced. On the other
hand, the variants prefer long-chain substrates. Compared to the
wild-type enzyme (Fig. 2a), substrate specificity of the variant
I105A/I317A (Fig. 2e) is inverted (S-propylhomocysteine 4
S-butylhomocysteine 4 ethionine c methionine). Hence,
steric effects play a major role in substrate conversion by SAMS
enzymes, and selectivity for certain substrates can be engineered by
exchange of two amino acid positions only. However, unlike for the
promiscuous SAMS enzyme from Sulfolobus solfataricus,13 activity of
the variants towards methionine analogues with increased steric
Fig. 5 Formation of SAM and analogues by SAMS enzyme from B. subtilis
(a), or its variants I317V (b), and I317A (c). The conversion of racemic
requirements – such as branched-chain (5–7) or polyunsaturated (9)
methionine (K), ethionine ( ), S-propylhomocysteine (,) and S-butylhomo-
cysteine (.) was performed in the presence of equal amounts of substrate
(10 mM) and enzyme (10 mU mlÀ1). The reactions were analyzed by HPTLC as
compounds – was marginal (Z0.3 nmol minÀ1 mgÀ1). Likewise,
derivatives having a hydroxyl (10) or carboxyl (11) group were
not accepted.§
described in the ESI.†
This journal is ©The Royal Society of Chemistry 2015
Chem. Commun.

View Article Online
Communication
ChemComm
§ Engineering of specificity towards these hydrophilic compounds by
introduction of one (variant I105V/I317S) or two (variant I105S/I317S)
serine residues into the hydrophobic amino acid binding site was also
not successful and led to completely inactive enzymes (data not shown).
were performed similar to the kinetic experiments but in larger
scale (200 mmol). After prolonged incubation (18 hours) to reach
maximum product yield, 84% and 89% of the L-amino acid was
converted. In the production of the n-propyl- and n-butyl analogues
of SAM by the variant I317A, 43% and 28% of conversion was
reached after 8 hours.
1 L. A. Wessjohann, M. Dippe, M. Tengg and M. Gruber-Khadjawi, in
Cascade biocatalysis: Integrating stereoselective and environmentally
friendly reactions, ed. S. Riva and W.-D. Fessner, Wiley-VCH, Weinheim,
1st edn, 2014, pp. 393–426.
After product purification by cation exchange chromato-
graphy, SAM and its homologues could be isolated in final
yields of 25% (Me), 17% (Et), 8% (Pr) and 11% (Bu) based on
´
2 L. A. Wessjohann, H. F. Schreckenbach and G. N. Kalu:erovic, in
Biocatalysis in Organic Synthesis (Science of Synthesis), ed. K. Faber,
W.-D. Fessner and N. Turner, Thieme, Stuttgart, 1st edn, 2014, ISBN:
9783131741615.
1
racemic starting material. As proven by H NMR, the enzyma-
3 L. A. Wessjohann, J. Keim, B. Weigel and M. Dippe, Curr. Opin.
Chem. Biol., 2013, 17, 229.
4 R. G. Matthews, M. Koutmos and S. Datta, Curr. Opin. Struct. Biol.,
2008, 18, 658; P. M. Ueland, P. I. Holm and S. Hustad, Clin. Chem.
Lab. Med., 2005, 43, 1069.
5 M. A. Grillo and S. Colombatto, Amino Acids, 2008, 34, 187.
6 S. Bheemanaik, Y. V. Reddy and D. N. Rao, Biochem. J., 2006,
399, 177.
7 M. Matsumoto, C. S. Weickert, M. Akil, B. K. Lipska, T. M. Hyde,
M. M. Herman, J. E. Kleinman and D. R. Weinberger, Neuroscience,
2003, 116, 127.
tically produced SAM contained a high excess (Z90%) of the
biologically active (S,S)-epimer. Interestingly, preparations of
the S-adenosyl-L-ethionine, -propionine and -buthionine were
diastereomeric mixtures, racemic with respect to the chiral
sulfonium center. As described above, the enzymatic reaction
is stereoselective, and thus we assume that the faster epimeri-
sation of higher homologues is a result of the higher +I effect
and the purification process.
8 W. Boerjan, J. Ralph and M. Baucher, Annu. Rev. Plant Biol., 2003,
54, 519.
9 S. H. Sung, B. G. Kim and J. H. Ahn, J. Microbiol. Biotechnol., 2011,
21, 854.
10 G. D. Markham, E. W. Hafner, C. W. Tabor and H. Tabor, J. Biol.
Chem., 1980, 255, 9082; J. R. Matos, F. M. Raushel and C. H. Wong,
Biotechnol. Appl. Biochem., 1987, 9, 39.
11 Z. J. Lu and G. D. Markham, J. Biol. Chem., 2002, 277, 16624.
12 P. K. Chiang and G. L. Cantoni, J. Biol. Chem., 1977, 252, 4506.
13 F. Wang, S. Singh, J. Zhang, T. D. Huber, K. E. Helmich, M. Sunkara,
K. A. Hurley, R. D. Goff, C. A. Bingman, A. J. Morris, J. S. Thorson
and G. N. Phillips, FEBS J., 2014, 281, 4224.
14 S. Singh, J. Zhang, T. D. Huber, M. Sunkara, K. Hurley, R. D. Goff,
G. Wang, W. Zhang, C. Liu, J. Rohr, S. G. van Lanen, A. J. Morris and
J. S. Thorson, Angew. Chem., Int. Ed., 2014, 53, 3965.
15 R. Wang, K. Islam, Y. Liu, W. Zheng, H. Tang, N. Lailler, G. Blum,
H. Deng and M. Luo, J. Am. Chem. Soc., 2013, 135, 1048.
16 J. Komoto, T. Yamada, Y. Takata, G. D. Markham and F. Takusagawa,
Biochemistry, 2004, 43, 1821.
In conclusion, a detailed study of the active site of S-adenosyl-
methionine synthases and docking of substrates and products
allowed the targeted introduction of subtle variation of size and
hydrophilicity of the methionine binding site. This resulted in
dramatic changes in activity and in an altered substrate scope.
Most of all it gave an entry into enzymes with reduced or almost
absent product inhibition. SAMS variants suitable for larger scale
application in synthesis and biotechnology are now available.20
The research leading to these results has received funding
from the European Union’s Seventh Framework Programme FP7/
2007-2013 under grant agreement no 266025 (BIONEXGEN).
Notes and references
‡ The influence of position 317 on catalytic turnover and its crucial role
in determination of substrate spectrum was additionally confirmed by 17 J. Schlesier, J. Siegrist, S. Gerhardt, A. Erb, S. Blaesi, M. Richter,
introduction of other small- and medium-sized residues. Enzymes sub- O. Einsle and J. N. Andexer, BMC Struct. Biol., 2013, 18, 13.
stituted for aliphatic (I317G, I317P, I317L) or polar (I317E, I317D, I317N) 18 M. Dippe, W. Brandt, L. A. Wessjohann, H. Rost and A. Porzel,
amino acids show a strongly reduced activity (Z5.4 nmol minÀ1 mgÀ1).
Eur. Pat., EP 13005228.5, 2013.
On the other hand, substitution by cysteine resulted in an active enzyme 19 V. Kamarthapu, K. V. Rao, P. N. Srinivas, G. B. Reddy and V. D.
which – similar to the variant I317A – converted methionine and homo-
Reddy, Biochim. Biophys. Acta, 2008, 1784, 1949.
logues (59.4 Æ 0.4, 16.4 Æ 0.7, 8.4 Æ 0.3 and 6.0 Æ 0.3 nmol minÀ1 mgÀ1 20 Y. Perez-Pertejo, R. M. Reguera, H. Villa, C. Garcia-Estrada,
for conversion of 5 mM D,L-methionine, -ethionine, -propionine and
-buthionine, respectively).
R. Balana-Fouce, M. A. Pajares and D. Ordonez, Eur. J. Biochem.,
2003, 270, 28.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2015