Multienzyme One-Pot Cascades Incorporating Methyltransferases for the Strategic Diversification of Tetrahydroisoquinoline Alkaloids

Article abstract of DOI:10.1002/anie.202104476

The tetrahydroisoquinoline (THIQ) ring system is present in a large variety of structurally diverse natural products exhibiting a wide range of biological activities. Routes to mimic the biosynthetic pathways to such alkaloids, by building cascade reactions in vitro, represents a successful strategy and can offer better stereoselectivities than traditional synthetic methods. S-Adenosylmethionine (SAM)-dependent methyltransferases are crucial in the biosynthesis and diversification of THIQs; however, their application is often limited in vitro by the high cost of SAM and low substrate scope. In this study, we describe the use of methyltransferases in vitro in multi-enzyme cascades, including for the generation of SAM in situ. Up to seven enzymes were used for the regioselective diversification of natural and non-natural THIQs on an enzymatic preparative scale. Regioselectivites of the methyltransferases were dependent on the group at C-1 and presence of fluorine in the THIQs. An interesting dual activity was also discovered for the catechol methyltransferases used, which were found to be able to regioselectively methylate two different catechols in a single molecule.

Full text of DOI:10.1002/anie.202104476

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Multienzyme one-pot cascades incorporating methyltransferases
for the strategic diversification of tetrahydroisoquinoline alkaloids
Authors: Helen Claire Hailes, Fabiana Subrizi, Yu Wang, Benjamin
Thair, Daniel Mendez-Sanchez, Rebecca Roddan, Max
Cardenas-Fernandez, Jutta Siegrist, Michael Richter, Jennifer
Andexer, and John Ward
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.202104476
Link to VoR: https://doi.org/10.1002/anie.202104476

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
10.1002/anie.202104476
Angewandte Chemie International Edition
COMMUNICATION
Multienzyme one-pot cascades incorporating methyltransferases
for the strategic diversification of tetrahydroisoquinoline
alkaloids
Fabiana Subrizi,[a] Yu Wang,[a] Benjamin Thair,[a] Daniel Méndez-Sánchez,[a] Rebecca Roddan,[a] Max Cárdenas-
Fernández,[b] Jutta Siegrist,[c] Michael Richter,[d] Jennifer N. Andexer,[c] John M. Ward,[b] and Helen C. Hailes*[a]
[a] Dr. F. Subrizi, Yu Wang, B. Thair, Dr. D. Méndez-Sánchez, R. Roddan, Prof. H. C. Hailes
Department of Chemistry, University College London
20 Gordon Street, London WC1H 0AJ
E-mail: h.c.hailes@ucl.ac.uk
[b] Prof. J. M. Ward, Dr. M. Cárdenas-Fernández
Department of Biochemical Engineering, University College London
Bernard Katz Building, London WC1E 6BT
[c] Dr. J. Siegrist, Prof. J. N. Andexer
Institute of Pharmaceutical Sciences, University of Freiburg
Albertstr. 25, 79104 Freiburg (Germany)
[d] Dr. M. Richter
Fraunhofer Institute for Interfacial Engineering and Biotechnology (IGB)
Branch Biocat, Schulgasse 11a, 94315 Straubing (Germany)
Supporting information for this article is given via a link at the end of the document.
a number of natural and non-natural THIQs.[13–16] The diversity of
THIQs generated has been further extended by forming in vitro
cascades with enzymes upstream of NCS.[17] In plant BIA
biosynthesis, methyltransferase (MT) enzymes operate
downstream of NCS to methylate norcoclaurine.[18] Such natural
product methylations involve transfer of a methyl group from (S)-
adenosylmethionine (SAM) to specific nucleophilic sites on the
scaffold, and is widely associated with activation of the compound
or improvement of its biological properties.[19–24]
Abstract: The tetrahydroisoquinoline (THIQ) ring system is
present in a large variety of structurally diverse natural products
exhibiting a wide range of biological activities. Routes to mimic
the biosynthetic pathways to such alkaloids, by building cascade
reactions in vitro, represents a successful strategy and offers
better stereoselectivities than traditional synthetic methods. (S)-
Adenosylmethionine (SAM) dependent methyltransferases are
crucial in the biosynthesis and diversification of THIQs; however,
their application is often limited in vitro by the high cost of SAM
and low substrate scope. In this study, we describe the use of
methyltransferases in vitro in multi-enzyme cascades, including
for the generation of SAM in situ. Up to seven enzymes were used
for the regioselective diversification of natural and non-natural
THIQs on an enzymatic preparative scale. Regioselectivites of the
methyltransferases were dependent on the group at C-1 and
presence of fluorine in the THIQs. An interesting dual activity was
also discovered for the catechol methyltransferases used, which
were found to be able to regioselectively methylate two different
catechols in a single molecule.
(a) Previous work on the formation of THIQs.[11,12,17]
A SAM supply system[32]
R1
Adenine
SRH
R
HO
ATP
L-Met
enzymes
NH2
NH
3 R4
THIQs
R-OH
R-NH2
R-OMe
R-NHMe
HO
HO
R2
MT
EcMAT EcMTAN
R
dopamine or
L-tyrosine
(b) This work to produce methylated THIQs in vitro via one-pot cascades with up to 7-enzymes
HO
R1 enzymes
NH2
R
NH
3 R4
HO
R2
HO
R
MT
EcMAT EcMTAN
MeO
HO
ATP
L-Met
NH
3 R4
NH
HO
MeO
R2
R
Introduction
n = 1, 2
2 analogues
6 analogues
THIQs are an important group of bioactive alkaloids.
Alongside their applications as analgesics and antitussives,[1–4]
THIQs show promise as chemotherapeutics towards cancer,
neuropathologies and multi-drug resistant bacteria.[3,5–8] However,
supply of these compounds for studies and clinical use is limited.
Natural THIQs can be harvested from mixtures from certain
plants, but their chemical syntheses are complicated by the
presence of chiral centres and the density of functional groups.[9]
An enzyme of particular interest has been norcoclaurine synthase
(NCS), which catalyses a Pictet-Spengler (PS) reaction between
dopamine and 4-hydroxyphenylacetaldehdye, to initiate
benzylisoquinoline alkaloid (BIA) biosynthesis. Most usefully,
using natural and engineered NCS variants,[10–12] the β-
arylethylamine and carbonyl substrates can be varied to generate
MeO
MeO
NH
NR
MeO
HO
HO
HO
MeO
NH
HO
HO
HO
(S)-coclaurine
(S)-nororientaline R = H
(S)-orientaline R = Me
MeO
(S)-norreticuline
Figure 1. (a) Formation of THIQs using enzymes and O- or N-methylation using
methyltranserases (MTs) and an S-adenosylmethionine (SAM) supply/(S)-
adenosyl homocysteine (SAH) degradation system; (b) The route to biologically
relevant methylated THIQs prepared in this work via cascades with up to 7-
enzymes. Abbreviations: E. coli (Ec); methionine adenosyl transferase (MAT);
methylthioadenosine nucleosidase (MTAN); L-methionine (L-Met); adenosine
triphosphate (ATP); (S)-ribosyl homocysteine (SRH).
1
This article is protected by copyright. All rights reserved.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
10.1002/anie.202104476
Angewandte Chemie International Edition
COMMUNICATION
(a)
(b)
Figure 2. Substrates used with the selected MTs indicated. (a) Structures of the THIQs tested. (b) Reaction conversions are based on the consumption of the THIQ
analogues 1-14 used, by HPLC analysis against standard calibration curves. Reaction conditions: THIQ 0.5 mM, SAM 3 mM, and selected MTs (clarified lysate, 10
% v/v) which showed conversions after 90 min: RnCOMT in blue, MxSafC in purple, and Cj-6-OMT in pink. THIQs (R,S)-12, (S)-13, and (S)-14 were not accepted
by any of the MTs tested and are not reported. All reactions were performed in duplicate. See SI for representative HPLC chromatograms and calibration curves.
MTs are therefore an essential addition for the diversification of
THIQs. However, while the O-MTs involved in producing natural
BIAs are regioselective,[25,26] they lack the substrate scope in vivo
needed to be components of a flexible enzyme cascade to larger
sets of THIQs. While in vitro biocatalytic cascades take inspiration
directly from natural biosynthetic pathways, they allow more
flexibility over reaction design offering control in the order and
quantity of enzymes added to the cascade as well as the enzyme
used. Importantly, the in vitro approach contributes to reduce
background reactions such as the phosphate catalysed PS,[27]
and can simplify product isolation and purification. Here we
present the use of three O-MTs in vitro, catechol-O-MTs from
Rattus norvegicus (RnCOMT),[28] Coptis japonica Cj-6-OMT,[25,29]
and SafC from Myxococcus xanthus (MxSafC),[30] and an N-MT
from Coptis japonica (CjCNMT)[31] to establish potential
reactivities towards a range of THIQs. We report the integration
of these enzymes into up to 7-step cascades for the strategic
diversification of bioactive compounds (Figure 1). The cost of
SAM has also previously limited the scale of MT-catalysed
reactions, but this challenge is resolved with the incorporation of
5) and C-1 substituted THIQs with linear aliphatic or cyclic
moieties (6-11). BIAs lacking the catechol group on the
isoquinoline scaffolds or those bearing a halide group (12-14)
were also included. All substrates 1-14 were tested with COMT
enzymes RnCOMT and MxSafC. Cj-6-OMT was also used for
comparison as it is known to selectively methylate in vivo the 6-
OH of norcoclaurine (S)-2.[29] Initially, purified enzymes were used
for reactions with substrates 1-3. With the aim of scaling up the
process and making it more viable in an industrial setting, the
same reactions were also carried out with clarified lysates,
showing regioselectivities that were comparable to the purified O-
MTs (SI section 3.2). Clarified lysates were then used with all the
substrates. We were delighted to see that all THIQs bearing a
catechol were readily accepted and new peaks were detected by
HPLC within the first hour (Figure 2b and SI Figure S5). In general,
RnCOMT and MxSafC showed broader substrate scope relative
to Cj-6-OMT, with excellent activities towards THIQs with an
aliphatic side chain (6-10), which were only slightly or not
converted by Cj-6-OMT under these conditions. Notably,
complete conversions were obtained within the first 90 min with
substrates 1-11 with RnCOMT, while MxSafC showed some
starting material left only in the case of substrates 9-11. As
expected, none of the THIQs lacking a catechol moiety (12 and
14) were accepted by the COMTs, including derivative 13 where
a catecholic hydroxyl group has been replaced with a bromine.
a
reported modular in vitro cofactor supply system,[32]
demonstrating the capacity of such an approach in ambitious
biocatalytic applications.
Results and Discussion
Initial O-MT screening
Building the enzyme cascades and O-MT regioselectivities
Establishment of the regiochemistries was also key, which
required scale-up of the reactions for product characterisation and
several factors were considered. Firstly, the high cost of the
cofactor SAM restricts its use in larger scale biocatalytic reactions
In the initial screening, a variety of THIQs were prepared
using Thalictrum flavum NCS (TfNCS), in high enantiomeric
excesses (e.e.s) where the single isomer is indicated, and
purified.[11,16,17] The substrate panel (Figure 2a) included BIAs (1-
2
This article is protected by copyright. All rights reserved.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
10.1002/anie.202104476
Angewandte Chemie International Edition
COMMUNICATION
in vitro: although it has been reported in the N-methylation of
THIQs,[31] this is not practicable for wider applications. In previous
work, integration of MTs into a linear cascade with two further
enzymes overcame this issue using a methionine adenosyl
transferase (MAT E.C. 2.5.1.6) and a methylthioadenosine
nucleosidase (MTAN, E.C. 3.2.2.9).[33] MATs generate SAM from
ATP and L-methionine, both of which are less expensive and
importantly more stable than SAM. When SAM is used for the
methylation reaction, (S)-adenosyl homocysteine (SAH) is formed
as a by-product. SAH, which would otherwise inhibit the
methylation, can then be cleaved by the MTAN into (S)-ribosyl
homocysteine (SRH) and adenine (Figure 1a). For the
supply/SAH degradation system, enzymes MAT and MTAN, both
from E. coli (EcMAT and EcMTAN, respectively) were selected.[33]
Secondly, clarified lysates were used instead of pure enzymes to
avoid the requirement for enzyme purification steps. Thirdly, to
further streamline the process, the methylation cascade reaction
was coupled directly to the NCS-catalysed PS reaction in a one-
pot two-step cascade, obviating the need for purification of the
THIQ intermediates. The methylation of representative substrates
(S)-1, (S)-6, 9, 10 and (S)-11 was explored first (Scheme 1: e.e.s
for (S)-1 (>97%) and (S)-6 (>98%) are shown in the SI). Here it
was essential that all the dopamine was consumed in the TfNCS
catalysed reaction before the MT step to avoid the methylation of
unreacted dopamine,[33] which could interfere with the
regiochemical outcome.
(Table 1). As expected, the regiochemistry of the reaction was
strongly dependent on the nature of the side chain, which directed
the methylation preferentially at the 6-OH for substrates (S)-1, (S)-
6, and (S)-11 and surprisingly preferably on the 7-OH for
substrates 9 and 10. Interestingly, when 9 or 10 underwent the
same reaction conditions with MxSafC, the 6-OH was methylated
instead preferentially and derivatives 6-OMe-9 or 6-OMe-10
respectively were obtained in good yield and excellent
regioselectivities (Table 1). This unique performance with a switch
in the regioselectivities for both RnCOMT and MxSafC further
confirms the complementarity of these two enzymes.[33]
With the challenge to demonstrate the potential of this
approach, the methylation with RnCOMT was further coupled with
a multistep in vitro enzyme cascade for the in situ generation of
norcoclaurine (S)-2 (generated in >97% e.e.) using Candidatus
nitrosopumilus salaria BD31Q tyrosinase (CnTYR) and
Enterococcus faecalis DC32 tyrosine decarboxylase (EfTyrDC),
together with a versatile transaminase from Chromobacterium
violaceum[34] (CvTAm) and wild-type TfNCS enzyme.[17] A total of
seven enzymes as clarified cell-lysates were used in the same
pot, achieving the conversion of L-tyrosine into (S)-coclaurine (S)-
6-OMe-2 in an unique one-pot three step cascade (Scheme 2a).
Table 1. Isolated yields and regioisomeric ratios for the preparative scale
cascade reactions described in Schemes 1-3.
Regioisomeric
ratio[b]
(6-OMe:7-
OMe)
(a)
Substrate
Product
MT
Yield[a]
HO
RnCOMT MeO
HO
HO
TfNCS
NH2
NH
NH
RCHO
MAT
MTAN
HO
HO
55%
(96%)
(S)-1
(S)-2
(S)-6-OMe-1
(S)-6-OMe-2
(S)-6-OMe-2
RnCOMT
RnCOMT
95:5
95:5
R
R
dopamine
(S)-6-OMe-1 R = CH2Ph
(S)-6-OMe-6 R = C5H11
(S)-6-OMe-11 R =CH2CH2Ph
(S)-1 R = CH2Ph
(S)-6 R = C5H11
55%
(92%)
(S)-11 R = CH2CH2Ph
RnCOMT
43%
(89%)
(S)-2
(S)-2
MxSafC[c]
MxSafC[c]
60:40
na
(b)
or
R1O
R2O
HO
HO
HO
MxSafC
TfNCS
6%
(9%)[d]
NH
n
NH2
(S)-6,7-(OMe)2-2
NH
n
O
MAT
MTAN
HO
dopamine
34%
(64%)[e]
RnCOMT
n
(S)-3
(S)-3
(S)-6,3'-(OMe)
(S)-6,4'-(OMe)
2
-3
-3
RnCOMT
95:5
100
9 n = 2
10 n = 1
7-OMe-9 n = 2, R1 = H, R2 = Me
7-OMe-10 n = 1, R1 = H, R2 = Me
MxSafC
Cj-6-OMT
then
MxSafC
6-OMe-9 n = 2, R1 = Me, R2 = H
6-OMe-10 n = 1, R1 = Me, R2 = H
27%
(70%)[e]
2
19%
(27%)
Scheme 1. Four-enzyme cascades for the synthesis of: (a) THIQs (S)-6-OMe-
1, (S)-6-OMe-6, and (S)-6-OMe-11, and (b) THIQs 6-OMe-9, 7-OMe-9, 6-OMe-
10, and 7-OMe-10 from 9 and 10 respectively.
(S)-5
(S)-5
(S)-6-OMe-5
(S)-6-OMe-5
RnCOMT
MxSafC
95:5
95:5
20%
(31%)
44%
(90%)
With the architecture of the cascades defined, a biocatalytic
scale-up of the methylated THIQs was attempted in a single pot.
Starting from dopamine (5-20 mM) and an excess of the
corresponding carbonyl compound, the THIQs (S)-1, (S)-6, 9,10
and (S)-11 were generated with high enantiopurity (for single
isomer compounds) and the methylation step initiated by the
addition of RnCOMT clarified lysate (10% v/v) in the presence of
ATP, L-methionine and the SAM supply/SAH degradation system
enzymes EcMAT (10% v/v) and EcMTAN (2.5% v/v). Using
RnCOMT the methylated products (S)-6-OMe-1, (S)-6-OMe-6, 7-
OMe-9, 7-OMe-10, and (S)-6-OMe-11 were isolated with good
yields and high regioselectivities with only 5-10% of the opposite
regioisomer detected by 1H NMR spectroscopy in most cases
(S)-6
(S)-6-OMe-6
7-OMe-9
RnCOMT
RnCOMT
90:10
10:90
9
40%
56%
9
6-OMe-9
MxSafC
90:10
10:90
(94%)[f]
45%[g]
(90%)
10
7-OMe-10
RnCOMT
30%[g]
(89%)
10
6-OMe-10
MxSafC
90:10
85:15
33%
(40%)
(S)-11
(S)-6-OMe-11
RnCOMT
3
This article is protected by copyright. All rights reserved.