Angewandte
Communications
Chemie
a key step in the biosynthesis of novobiocin.[18–20] A hallmark
of NovO is its substrate promiscuity (e.g., 1) and the ability to
utilize S-alkylated analogues of SAM to form products such as
2 (Figure 1a).[20,21]
previously was a rotation of the sidechain of Arg243, from the
solvent-exposed exterior of the protein to the interior of the
active site, enabling the formation of electrostatic interactions
between Arg243 and the carboxylate of SAM (Figure 2 and
Supporting Information, Figure S1), and the side chain of
Glu17 from the adjacent monomer. No associated changes in
the solvent-exposed channel were observed. These structures
were then used as a guide for the preparation of point mutants
in order to explore the roles of specific residues in catalysis.
Phe186Leu, Trp129Phe, Asp183Glu, Trp190Ala, Val12Met,
and Tyr70Met displayed a reduced level of activity relative to
wild-type SalL (Supporting Information, Table S2 and Fig-
ure S2). The Phe186Leu mutant was able to form SAM and
SAM analogues, albeit in slightly poorer conversions relative
to the wild-type (Supporting Information, Figure S15). In
contrast, enzymatic activity was abolished in the Asp183Ala,
Asn188Ala, Phe186Ala, and Phe228Ile/Ala mutants. This
suggests that p-stacking between Phe228 and the adenine
nucleobase, the electrostatic interaction between the Met
carboxylate and Asp183, and H-bonding to the Hoogsteen
face of the adenine nucleobase (Asn188) are essential for
catalysis.
Further supporting evidence for the importance of
Asp183 for catalysis was observed when a tetrazole carboxylic
acid bioisostere of methionine was used (Figure 3a). The use
of tetMet[27] resulted in 99% conversion to 4c, despite slower
reaction kinetics relative to SAM formation (4a, Figure 3b
and Supporting Information, Table S3 and Figures S8–12,
S14). However, an increase in steric bulk at the sulfur center
i.e., by replacing Met with l-ethionine formed SAE (4b)
albeit in lower conversion (41%) relative to SAM.
Replacing adenine with 7-deazaadenine (4d) resulted in
only 10% conversion (Figure 3c),[27] whereas a hypoxanthine
nucleobase did not form 4e. This suggests that the interaction
between the N7 of adenine and Asn188 is critical for the
catalytic function of SalL. High conversions to 4 f–j were
observed using analogues containing modifications to the 2-
and 6-position of adenine. Combining 2,6-diamino or 2-
chloro-6-aminoadenine modifications with tetMet produced
4k and 4l in greater than 99% and 41% conversion,
respectively (Figure 3d). Although no formation of 4m was
observed when tetethionine was used, SAE analogues 4n and
4o were formed in 78% and 37%, respectively. Finally, no
cofactor products were formed using ClDA substrates lacking
either 2’/3’ ribose hydroxyl groups (4p–r).
One limitation of this process is the need to prepare these
cofactors by chemical synthesis, which is laborious, low
yielding, and produces both epimers at the sulfur
center.[22–26] Furthermore, SAM analogues are inherently
unstable in buffered solution (t1/2 942 min for SAM at
pH 8).[27,28] A more step- and atom-efficient strategy is to
couple cofactor formation with C-alkyl transfer.[15,29,30] One
example of this one-pot process is the generation of cofactor
analogues in situ from either ClDA or ATP and (m)ethio-
nine,[15,29–33] followed by C-(m)ethyl transfer catalyzed by
a MTase (Figure 1b).[34] ClDA is a shelf-stable, atom-eco-
nomical adenosine source for such a process catalyzed by SalL
compared to ATP, which is a substrate for SAM production by
methionine adenosyltransferase (MAT).[15,27,29,35,36]
Although in-depth knowledge of the substrate promiscu-
ity of C-MTases has been garnered from structural and
mutagenesis studies,[17,19,20,37] little is known about how the
structural features of the SAM cofactor itself influences the
yield and scope of C-alkylation.[14,30] Herein, we showcase
a method to address these limitations by strategic modifica-
tions to SAM and S-adenosyl ethionine (SAE, Figure 1c).
An earlier structural study of SalL in complex with ClDA
and methionine revealed a solvent-exposed channel into the
active site.[38–40] To explore this in more detail, we obtained
two structures of wild-type SalL with SAM and chloride
(6RYZ, 1.50 ꢀ), and with ClDA alone (6RZ2, 1.77 ꢀ;
Figure 2, Supporting Information, Table S1). One significant
difference in our structures compared to those obtained
Inspection of the crystal structure of NovO in complex
with S-adenosylhomocysteine (SAH) revealed the presence
of a hydrophobic cleft with a volume of approximately
21 ꢀ3.[18] This is in the exact location of the 2-position of the
adenine nucleobase (Figure 4 and Supporting Information,
Figure S3). We surmised that SAM/SAE cofactors bearing
modifications of complementary steric volume at this position
would also be substrates for NovO. Our tandem enzymatic
process using purified SalL and NovO in the presence of
stoichiometric amounts of ClDA analogue, l-Met, and
coumarin (5) indeed demonstrated the enhanced conversion
of methylated coumarin (5a) via the in situ formation of
modified SAM analogues relative to SAM (Supporting
Information, Figure S70 and Table S5).
Figure 2. Wild-type SalL in complex with ClDA and l-Met (PDB
2Q6I[31]) superimposed with wild-type SalL in complex with SAM
(6RYZ, this study). Neighboring monomers and amino acid side-
chains of 2Q6I are shown in cyan and light blue. Arg243 and SAM
(6RYZ, coral) illustrate relocation of this side chain to form new ionic
interactions (coral) with the SAM carboxylate and the side chain of
Glu17.
2
ꢀ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2019, 58, 1 – 7
These are not the final page numbers!