Engineering Dirhodium Artificial Metalloenzymes for Diazo Coupling Cascade Reactions**

Article abstract of DOI:10.1002/anie.202107982

Artificial metalloenzymes (ArMs) are commonly used to control the stereoselectivity of catalytic reactions, but controlling chemoselectivity remains challenging. In this study, we engineer a dirhodium ArM to catalyze diazo cross-coupling to form an alkene that, in a one-pot cascade reaction, is reduced to an alkane with high enantioselectivity (typically >99 % ee) by an alkene reductase. The numerous protein and small molecule components required for the cascade reaction had minimal effect on ArM catalysis. Directed evolution of the ArM led to improved yields and E/Z selectivities for a variety of substrates, which translated to cascade reaction yields. MD simulations of ArM variants were used to understand the structural role of the cofactor on ArM conformational dynamics. These results highlight the ability of ArMs to control both catalyst stereoselectivity and chemoselectivity to enable reactions in complex media that would otherwise lead to undesired side reactions.

Full text of DOI:10.1002/anie.202107982

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Engineering Dirhodium Artificial Metalloenzymes for Diazo
Coupling Cascade Reactions
Authors: David M. Upp, Rui Huang, Ying Li, Max J. Bultman, Benoit
Roux, and Jared C Lewis
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.202107982
Link to VoR: https://doi.org/10.1002/anie.202107982

10.1002/anie.202107982
Angewandte Chemie International Edition
Engineering Dirhodium Artificial Metalloenzymes for Diazo Coupling Cascade Reactions
David M. Upp1,†, Rui Huang1,†, Ying Li2, Max J. Bultman1, Benoit Roux2,3,*, Jared C. Lewis1,*
1Department of Chemistry, Indiana University, Bloomington, IN 47405, USA
2Department of Biochemistry and Molecular Biology, University of Chicago, Chicago, IL 60637,
USA
3Department of Chemistry, University of Chicago, Chicago, IL 60637, USA
†These authors contributed equally to this study.
Abstract
Artificial metalloenzymes (ArMs) are commonly used to control the stereoselectivity of catalytic
reactions, but controlling chemoselectivity remains challenging. In this study, we engineer a
dirhodium ArM to catalyze diazo cross-coupling to form an alkene that, in a one-pot cascade
reaction, is reduced to an alkane with high enantioselectivity (typically >99% e.e.) by an alkene
reductase. The numerous protein and small molecule components required for the cascade reaction
had minimal effect on ArM catalysis. Directed evolution of the ArM led to improved yields and
E/Z selectivities for a variety of substrates, which translated to cascade reaction yields. MD
simulations of ArM variants were used to understand the structural role of the cofactor on ArM
conformational dynamics. These results highlight the ability of ArMs to control both catalyst
stereoselectivity and chemoselectivity to enable reactions in complex media that would otherwise
lead to undesired side reactions.
Introduction
The ability of transition metals to bind and react with a wide range of species underpins their utility
as catalysts, but it also necessitates methods to ensure that a given metal species catalyzes a desired
reaction at the correct site on a target substrate.[1] In the laboratory, this challenge is dramatically
simplified by excluding all but the necessary components for a desired reaction. Transition metal
chemoselectivity can then be tuned using ligands that modulate the steric and electronic properties
of a metal center (its primary coordination sphere).[2] Significant effort has also been devoted to
incorporating attractive substrate-catalyst interactions distal to a metal center (its secondary
coordination sphere) to control catalyst selectivity.[3,4] Similar primary and secondary sphere
effects are exploited by metalloenzymes to modulate transition metal reactivity,[5] and these
examples have inspired many of the efforts to recapitulate enzyme-like secondary sphere effects
in small molecule complexes.[6]
The remarkable activities and selectivities of metalloenzymes are all the more impressive given
that they operate in a complex cellular milieu. This capability suggests that the molecular
recognition imparted by second sphere interactions in enzymes enables far greater control over
transition metal reactivity than can currently be achieved with small molecule ligands.[7] Similar
control over synthetic metal complexes could enable reactions in complex environments, including
enzymatic and chemoenzymatic cascades containing multiple catalysts, reagents, and
intermediates.[8,9] Artificial metalloenzymes (ArMs) have been explored as a means to merge the
reactivity of synthetic catalysts with the selectivity and evolvability of protein scaffolds.[10]
Moreover, streptavidin-,[11,12] LmrR-,[13] and albumin-based[14] ArMs have been used for in vivo
This article is protected by copyright. All rights reserved.

10.1002/anie.202107982
Angewandte Chemie International Edition
catalysis, and streptavidin-,[15] FhuA-,[16] and P450-based[17] ArMs have been used for in vitro
cascade reactions. In each of these examples, however, inherent cofactor reactivity alleviates the
need for scaffold-controlled chemoselectivity. The ArMs produce the same products as the metal
cofactors alone, albeit with impressive rate acceleration and enantioselectivity.
Our group has explored the design[18] and evolution[19] of dirhodium ArMs comprising dirhodium
cofactor (1)[20] covalently linked to a prolyl oligopeptidase (POP) scaffold containing a genetically
encoded azidophenylalanine (Z) residue at position 477 (Figure 1A). Dirhodium complexes react
with donor-acceptor diazo compounds to generate carbene complexes that react with water, thiols,
amines, olefins, silanes, and even sp3 C-H bonds.[21,22] We envisioned that this broad reactivity
would allow studies on the extent to which a protein scaffold, rather than reaction conditions or
the primary coordination sphere, can be engineered to control transition metal chemoselectivity.
Previous studies in our laboratory established that ArM chemoselectivity can be evolved to favor
cyclopropanation over formal carbene insertion into water O-H bonds.[18] While this side reaction
is not typically observed using other natural and unnatural heme carbene transferases,[23–25] we
reasoned that this level of control over dirhodium reactivity could enable cascade reactions
involving a variety of additional species in solution. For example, dirhodium-catalyzed diazo
cross-coupling has been used to generate fumaric acid esters that are converted by alkene
reductases to 2-substituted succinate derivatives (Scheme 1B).[26] In this study, however,
dirhodium catalysis was conducted in organic solvent at cryogenic temperatures, and the solvent
was evaporated to enable biocatalytic reduction in aqueous buffer. Herein, we evolve a dirhodium
ArM that catalyzes diazo cross-coupling with high chemo- and stereoselectivity and demonstrate
that the resulting ArM can be interfaced with an alkene reductase in a one-pot cascade reaction to
produce substituted succinate derivatives with high enantioselectivity (Scheme 1C).
A
O
O
H
H
O
O
O
O
O
Rh Rh
O
O
O
O
1
M
N3
N
M
N
N
B Reported one-pot, two-step sequence using dirhodium tetrapivalate catalyst:
O
O
N2
H
OMe
EtO
Ph
EtO
Ph
GDH
Rh2
ER
OMe
OMe
O
Ph
+
NADP+, glucose
buffer
O
DCM
-78°C
O
O
H
E/Z >10
Evaporate and
redissolve in DMSO
EtO
up to 62% isolated
yield over 2 steps
up to >99% e.e.
N2
C Cascade sequence involving dirhodium ArM catalyst:
N2
O
OMe
Ar
EtO
ER
ArM
GDH
O
OMe
Ar
+
O
NADP+, glucose
buffer
O
H
EtO
up to 60% HPLC yield
up to >99% e.e.
N2
Figure 1. Dirhodium ArM formation and catalysis. A) Cofactor 1 and covalent bioconjugation via
azidophenylalanine. B) Synthesis of chiral aryl succinates through a one-pot, two-step reaction.[26]
C) POP-1/ER cascade catalysis.
This article is protected by copyright. All rights reserved.

10.1002/anie.202107982
Angewandte Chemie International Edition
Results and Discussion
ArM Directed Evolution
The reactivities of several previously-evolved dirhodium ArMs were evaluated using a model
diazo coupling reaction (Entries 1-3, Table 1). Because the ene reductases investigated accept E-
alkenes,[26] it was important to establish whether ArMs could provide this isomer in good yield
under conditions suitable for biocatalysis. ArM variant 0-ZA4 (Table S1) catalyzed diazo coupling
with 46% yield and a 2.9:1 E/Z ratio (4/5), while variant 3-VRVH[19] provided a 53% yield and
3.5:1 E/Z ratio. Variant 1-SGH,[19] which contains only the three mutations in 3-VRVH that are
necessary for the high selectivity of the latter, provided a similar yield and E/Z ratio as 3-VRVH,
so reaction conditions were optimized using this ArM. ArM loading could be reduced to 0.1 mol%
with minimal change in yield, but excess 3 was required (Table 1, Entries 4 and 5, Table S2). We
previously established that high salt concentrations are required for high ArM activity and
selectivity,[18] and 0.7 M NaBr was sufficient in this regard (Entries 4, 6, and 7). In all cases, the
observed selectivity is lower than the analogous reaction in organic solvent under cryogenic
conditions (>10/1)[26], so improving this was a key goal of directed evolution. A significant amount
(39%) of the donor-acceptor diazo substrate reacted with water, forming OH insertion product 6,
so minimizing this side reaction would also be required as in our previous evolution efforts[18,19]
.
Table 1. Scaffold selection and reaction optimization.[a]
O
N2
EtO
OMe
E-alkene: 4
Z-alkene: 5
OMe
OMe
O
MeO
ArM
O
2
MeO
MeO
+
+
50 mM PIPES, pH 7.4
5% THF, NaBr
O
OH
H
OEt
22 hrs, 4°C
6
N2
O
3
% Yieldc
ArM
Variant
[ArM] [NaBr]
Entrya
4
5
6
E/Z
(µM)
(M)
1
2
ZA4
50
50
50
5
0.7
0.7
0.7
0.7
0.7
0.1
1.75
34
41
43
44
39
34
46
12
12
13
12
13
19
8
48
35
39
31
37
13
33
2.9
3.4
3.3
3.7
3.1
1.8
5.7
3-VRVH
1-SGH
1-SGH
1-SGH
1-SGH
1-SGH
3
4
5b
6
5
5
7
5
a Standard reaction conditions: 5 mM 2, 25 mM 3, 50 mM PIPES pH 7.4, 5% cosolvent, 22 hours at 4°C with shaking. b Standard conditions using 5 mM
3. c Determined by SFC analysis using 1,3,5-trimethoxybenzene as internal standard. Reported yields and E/Z values are the average of triplicate
reactions.
ArM evolution was conducted similarly to our earlier efforts.[19] Scaffold libraries containing the
Z-477 mutation were expressed in 96-well plates and covalently modified in lysate using cofactor
This article is protected by copyright. All rights reserved.

10.1002/anie.202107982
Angewandte Chemie International Edition
1. The resulting ArMs were immobilized on Ni-NTA resin in filter plates in which the diazo
coupling reactions were conducted. Following reaction, the catalyst was removed by centrifugal
filtration, and the reaction products were analyzed by SFC. Previous efforts[18,19] revealed that
mutations in a b-strand across from a putative Rh-binding histidine residue significantly affected
ArM-catalyzed cyclopropanation activity and selectivity. Site-saturation libraries were therefore
constructed for several residues (98-101) in this b-strand of 1-SGH using degenerate NNK codons.
This effort revealed that Q98P improved the yield of 4, increased the E/Z ratio, and decreased the
yield of 6 (variant 2-P, Table 2). Site-saturation mutagenesis of S99 in 2-P gave variant 3-H (Table
2), which possessed a similar yield as 2-P but increased E/Z selectivity.
Similar mutagenesis of residues F100 and T101 did not lead to further improvements.
Combinatorial codon mutagenesis (CCM)[27] of 25 active site residues projecting into the active
site of 3-H was therefore conducted using degenerate NDT codons. The mutation V71G (4-G,
Table 2) was identified using this approach, but a subsequent CCM library did not yield a positive
variant. MD simulations (vide infra) were used to identify CCM library residues proximal to the
cofactor. Six of these were analyzed in more depth using site saturation (NNK) libraries, resulting
in variant 5-G (E283G, Table 2), which displayed further improvements in diazo coupling yield
and selectivity. Decreasing the ArM loading 100-fold with respect to 2 to 0.001 mol% substantially
increased the TTN, with 5-G catalyzing 44,612 turnovers to 4, which is among the highest TTN
reported for an ArM[28]. If turnovers associated with formation of 5 and 6 are included, a
remarkable 72,196 TTN is observed, highlighting the high activity of the dirhodium cofactor
within 5-G. The significant decrease in E/Z selectivity in these high TTN conditions may result
from active site modification via carbene insertion during catalysis, which was previously
characterized in POP-1 ArMs following cyclopropanation catalysis[19]
.
Table 2. Directed evolution of ArM for diazo coupling.
O
N2
EtO
OMe
E-alkene: 4
Z-alkene: 5
OMe
OMe
O
MeO
ArM
O
2
MeO
+
+
50 mM PIPES, pH 7.4
5% THF, 0.7M NaBr
22 hrs, 4°C
O
OH
H
OEt
6
N2
O
MeO
3
% Yieldc
Muta-
Mutations
genesis from previous
Entrya Variant method
generation
Parent
Q98P
4
5
6
E/Z TTN of 4
1
2
1-SGH
2-P
-
44 12 31
55 11 33
3.7
4.9
7.3
8.6
388
545
Q98NNK
S99NNK
CCM
3
3-H
S99H
51
72
76
7
8
5
24
21
511
4
4-G
V71G
717
5
5-G E283NNK
5-G E283NNK
E283G
E283G
16 14.9
2.8
761
6b
45 16 11
44612
This article is protected by copyright. All rights reserved.

10.1002/anie.202107982
Angewandte Chemie International Edition
a Standard reaction conditions: 0.1 mol% ArM, 5 mM 2, 25 mM 3, 50 mM PIPES pH 7.4, 5% THF, 22 hours at 4°C with shaking. b Standard conditions
using 0.001 mol% (50 nM) ArM and 96 hour reaction time. c Determined by SFC analysis using 1,3,5-trimethoxybenzene as internal standard. Reported
yields and E/Z values are the average of triplicate reactions.
The substrate scope of 5-G was next evaluated under optimized reaction conditions. Improved
yields of the desired E-alkenes were observed in all cases using 5-G relative to 1-GSH (Table 3),
indicating that mutations accumulated during directed evolution generally improved the scaffold
for diazo coupling. Steric and electronic perturbation of the para substituent (R1) and ester (R2) of
the aryl diazoacetate coupling partner were tolerated, and both ethyl diazoesters and amides could
be used (R3). With the exception of the previously unreported amide substrate, these substrates
are in line with the known scope for dirhodium-catalyzed diazo coupling,[26,29] indicating that the
ArM enables the desired dirhodium activity while significantly reducing undesired side reactions
such as water O-H insertion.
Table 3. Substrate scope of ArM-catalyzed diazo cross-coupling.
N2
O
R2
ArM
H
O
R3
R2
O
O
R1
50 mM PIPES, pH 7.4
0.7M NaBr, 5% THF
22 hrs, 4°C
H
+
R3
R1
N2
% Yieldb
Entrya
R1
R2
R3
1-GSH
44
5-G
76
75
68
64
58
37
66
1
2
3
4
5
6
7
OMe
H
OMe
OMe
OMe
OMe
OMe
OMe
Me
OEt
OEt
OEt
OEt
NEt2
OBn
OEt
38
Cl
23
Br
38
OMe
OMe
Cl
33
18
30
a
b
0.1 mol% ArM, 5 mM donor-acceptor diazo, 25 mM acceptor-only diazo, 50 mM PIPES pH 7.4, 0.7M NaBr, 5% THF, 22 hrs at 4°C with shaking.
Determined by HPLC analysis using 1,3,5-trimethoxybenzene as internal standard. Reported yields are the average of triplicate reactions.
ArM/ER Cascade Catalysis
The synthesis of enantioenriched succinate derivatives via cascade catalysis involving an ER was
then explored. The activities of several ERs,[30,31] including alkene reductase from Yersinia
bercovieri (YersER), enoate reductase-1 from Kluyveromyces lactis (KYE1) and 1,2-
oxophytodienoate reductase from Lycopersicum esculentum (OPR1), were evaluated on the 2-aryl
fumaric acid derivatives produced via ArM catalysis to select the optimal ER for each substrate
(Table S3). The ArM/ER cascade requires that the ArM, the ER, and a glucose dehydrogenase
(GDH, to supply the ER with reduced cofactor) all tolerate one another, in addition to glucose, a
terminal reductant that is converted to gluconic acid, and NADP(H). This challenge is particularly
notable given that dirhodium carbenoid species react readily with water, proteins,[32] and a range
of small molecule nucleophiles.[21] Remarkably, however, 1-GSH and 5-G catalyzed diazo
This article is protected by copyright. All rights reserved.

10.1002/anie.202107982
Angewandte Chemie International Edition
coupling with only slightly reduced yields even in the presence of all cascade components, and the
ERs successfully converted the majority of the fumaric ester intermediates to the reduced products
in good yields (Table 4). In some cases, the cascade reaction scope was limited by the ERs
examined; bulky R2 groups (e.g. OiPr) were coupled in good yields by the ArM but not reduced
by the ERs. Within the known ER substrate scope, however, 5-G catalyzed the desired alkene
formation with up to 723 turnovers, again highlighting the capacity of the ArM scaffold to protect
the dirhodium center from deactivation and side reactions.
Table 4. Substrate scope of ArM/ER cascade reactions.
N2
O
ER
ArM
GDH
R2
R3
50 mM PIPES, pH 7.4,
25 mM glucose, 0.2 mM NADP+,
0.7M NaBr, 5% Dioxane
R2
O
O
R1
H
+
R3
O
R1
1 hr, 4°C then 23 hrs, 23°C
N2
% Yieldb (e.e)c
Entrya
ER
KYE1
R1
R2
R3
1-GSH
5-G
1
2
3
4
5
6
7
OMe
H
OMe
OMe
OMe
OMe
OMe
OMe
Me
OEt
OEt
OEt
OEt
NEt2
OBn
OEt
25 (>99%)
35 (>99%)
18 (>99%)
32 (>99%)
22 (>99%)
9 (>99%)
34 (79%)
61 (>99%)
56 (>99%)
47 (>99%)
60 (>99%)
40 (>99%)
12 (>99%)
52 (78%)
YersER
YersER
YersER
OPR1
Cl
Br
OMe
OMe
Cl
OPR1
YersER
a 0.1 mol% ArM, 5 mM donor-acceptor diazo, 25 mM acceptor-only diazo, 50 mM PIPES pH 7.4, 0.7M NaBr, 5% dioxane, 1 hr at 4°C with shaking
followed by 23 hrs shaking at 23 °C. b Determined by HPLC analysis using 1,3,5-trimethoxybenzene as internal standard. Reported yields are the average
of triplicate reactions. c Enantioselectivity determined by chiral HPLC analysis.
The yield of the final succinic acid derivatives tracked with the ArM alkene yields, suggesting that
while mutations gained throughout evolution increased diazo coupling performance, tolerance to
cascade conditions was present from the outset of ArM evolution. On the other hand, diazo
coupling reactions catalyzed by an acetyl-substituted cofactor in aqueous buffer provided the OH
insertion product 6 almost exclusively and only trace 4 or 5 (Table S4). Moreover, in the presence
of glucose, formal OH insertion involving both water and glucose[33] was observed by mass
spectrometry, but the latter is completely absent in the ArM catalyzed reaction (Figure S1). Finally,
while dirhodium catalysts are capable of modifying surface-exposed protein residues,[32] no such
modifications were observed by mass spectrometry for the enzymes used in the cascade reactions
(Figure S1). These results suggest that the ArM provides a hydrophobic environment that excludes
polar nucleophiles like bulk water and glucose to enable selective diazo cross-coupling.[34]
ArM Conformational Dynamics
To gain insight into how the POP scaffold might accomplish this level of substrate specificity, MD
simulations were conducted on models of 5-G that involved different starting coordination states
This article is protected by copyright. All rights reserved.

10.1002/anie.202107982
Angewandte Chemie International Edition
of cofactor 1. We previously reported that apo-POP undergoes inter-domain opening and closing
to form a solvent-exposed cleft,[35] and similar behavior was observed for apo-5-G (Figure 2A, B).
We speculated that analogous dynamics in POP ArMs would facilitate cofactor bioconjugation in
the open state and provide a more compact, hydrophobic environment for chemoselective catalysis
in the closed state.[7] We further hypothesized that the improved chemoselectivity of ArMs
containing specific active site His residues, such as H326 in the lineage examined in this study,
might result from an internal Rh-His cross-link[36] favoring the closed form of the ArM. Supporting
this notion, MD simulations of 5-G with the Rh-His bond intact showed that POP opening/closing
was greatly reduced (Figure 2C). Despite its flexible linker, cofactor 1 holds 5-G closed when the
Rh-His bond is present (Figure 2D). Interestingly, a simulation starting from a state lacking a Rh-
His bond was able to access an open structure for much of the simulation (676/1000 ns, Figure
2C), though this system did not open to the same extent as apo 5-G. This constraint appears to
result from a persistent hydrophobic interaction between 1 and a number of residues on the interior
surface of the 5-G scaffold (Figure S2). This finding could explain the improved selectivity of
POP-based ArMs lacking an interior His residue,[18,19] but further free energy calculations and
experimental validation will be required to establish this possibility.
This article is protected by copyright. All rights reserved.

10.1002/anie.202107982
Angewandte Chemie International Edition
Figure 2. Domain dynamics of apo-POP and POP-1 ArMs. A) Open state of apo-5-G showing the
interdomain angle, θ. B) The interdomain angle of apo-5-G. The yellow bar (17-23°) indicates the
open/closed transition. C) The interdomain angle of 5-G-1 with (green) and without (red) a Rh-
His bond. D) Representative trajectory showing the rhodium-histidine interaction in 5-G with the
mutations found in this study highlighted in yellow.
Conclusion
Controlling the reactivity and selectivity of a transition metal catalyst requires precise tuning of its
primary and secondary coordination spheres. The numerous potential interactions that can occur
between a metal catalyst, a substrate, and a protein scaffold make ArMs a promising platform for
selective catalysis.[10] In this study, we showed that dirhodium ArMs can use first and second
sphere interactions to catalyze diazo cross-coupling reactions in complex cascade reaction
mixtures. Others have established that ArM scaffolds can protect catalysts from poisons such as
glutathione, which reversibly bind to and deactivate metal centers,[11,14] but the current study
highlights rare examples of chemoselectivity. Building on earlier observations for water tolerance
by dirhodium ArMs,[18,19] this capability enables selective reaction of one functional group over
many others on different substrates in solution, presumably by regulating substrate access to and
orientation within the ArM active site. Previously evolved ArM variant 1-SGH was submitted to
further directed evolution to improve diazo coupling yield and selectivity, increasing the yield of
the desired E-alkene over the Z-alkene and OH insertion side products. These improvements
carried over to the cascade reaction with an ER. While it is likely that most of the four mutations
found in this study are involved in outer-sphere control and substrate positioning, H326 and H99
can bind the rhodium cofactor. This interaction affected POP dynamics in MD simulations, helping
it to maintain a closed state. These models suggest a number of mechanisms by which the cofactor
can affect the structural dynamics of the ArM scaffold, adopting dual catalytic and structural roles
just as natural metalloenzyme cofactors do.[37] Further studies on this system will help clarify
different ways that cofactor-scaffold interactions can give rise to emergent properties in ArMs.
Acknowledgements
We thank Prof. Andreas Bommarius for plasmids pET28a-YersER and pET28a-KYE1 and Prof.
Kurt Faber for plasmid pET21a-OPR1. This study was supported by the U.S. Army Research
Laboratory and the U.S. Army Research Office under Contracts/Grants W911NF-18-1-0034 and
W911NF-15-1-0334 (to J.C.L.) and under Grant Number W911NF-18-1-0200 (to J.C.L and B.R.).
D.U. was supported by NSF under the CCI Center for Selective C–H Functionalization (CCHF,
CHE-1700982).
References
[1] N. A. Afagh, A. K. Yudin, Angew. Chem., Int. Ed. 2010, 49, 262–310.
[2] J. Mahatthananchai, A. M. Dumas, J. W. Bode, Angew. Chem., Int. Ed. 2012, 51, 10954–
10990.
This article is protected by copyright. All rights reserved.

10.1002/anie.202107982
Angewandte Chemie International Edition
[3] R. R. Knowles, E. N. Jacobsen, Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 20678–20685.
[4] F. D. Toste, M. S. Sigman, S. J. Miller, Acc. Chem. Res. 2017, 50, 609–615.
[5] S. W. Ragsdale, Chem. Rev. 2006, 106, 3317–3337.
[6] R. Breslow, Acc. Chem. Res. 1995, 28, 146–153.
[7] J. C. Lewis, Acc. Chem. Res. 2019, 52, 576–584.
[8] V. Köhler, N. J. Turner, Chem. Comm. 2014, 51, 450–464.
[9] C. A. Denard, J. F. Hartwig, H. Zhao, ACS Catal. 2013, 3, 2856–2864.
[10] F. Schwizer, Y. Okamoto, T. Heinisch, Y. Gu, M. M. Pellizzoni, V. Lebrun, R. Reuter, V.
Kohler, J. C. Lewis, T. R. Ward, Chem. Rev. 2017, 118, 142–231.
[11] M. Jeschek, R. Reuter, T. Heinisch, C. Trindler, J. Klehr, S. Panke, T. R. Ward, Nature
2016, 537, 661–665.
[12] A. D. Liang, J. Serrano-Plana, R. L. Peterson, T. R. Ward, Acc. Chem. Res. 2019, 52, 585–
595.
[13] S. Chordia, S. Narasimhan, A. L. Paioni, M. Baldus, G. Roelfes, Angew. Chem., Int. Ed.
2021, 60, 5913–5920.
[14] S. Eda, I. Nasibullin, K. Vong, N. Kudo, M. Yoshida, A. Kurbangalieva, K. Tanaka, Nat.
Catal. 2019, 2, 780–792.
[15] V. Köhler, Y. M. Wilson, M. Dürrenberger, D. Ghislieri, E. Churakova, T. Quinto, L.
Knörr, D. Häussinger, F. Hollmann, N. J. Turner, T. R. Ward, Nat. Chem. 2012, 5, 93–99.
[16] M. A. S. Mertens, D. F. Sauer, U. Markel, J. Schiffels, J. Okuda, U. Schwaneberg, Catal.
Sci. Technol. 2019, 9, 5572–5576.
[17] J. Huang, Z. Liu, B. Bloomer, D. Clark, A. Mukhopadhyay, J. Keasling, J. Hartwig, 2021,
DOI 10.21203/rs.3.rs-130318/v1.
[18] P. Srivastava, H. Yang, K. Ellis-Guardiola, J. C. Lewis, Nat. Comm. 2015, 6, 7789.
[19] H. Yang, A. M. Swartz, H.-J. Park, P. Srivastava, K. Ellis-Guardiola, D. M. Upp, G. Lee, K.
Belsare, Y. Gu, C. Zhang, R. E. Moellering, J. C. Lewis, Nat. Chem. 2018, 10, 318–324.
[20] H. Yang, P. Srivastava, C. Zhang, J. C. Lewis, ChemBioChem 2014, 15, 223–227.
[21] D. Gillingham, N. Fei, Chem. Soc. Rev. 2013, 42, 4918–4931.
This article is protected by copyright. All rights reserved.

10.1002/anie.202107982
Angewandte Chemie International Edition
[22] H. M. L. Davies, D. Morton, Chem. Soc. Rev. 2011, 40, 1857–1869.
[23] Y. Yang, F. H. Arnold, Acc. Chem. Res. 2021, 54, 1209–1225.
[24] G. Sreenilayam, E. J. Moore, V. Steck, R. Fasan, ACS Catal. 2017, 7, 7629–7633.
[25] L. Villarino, K. E. Splan, E. Reddem, L. Alonso-Cotchico, C. G. de Souza, A. Lledós, J.
Maréchal, A. W. H. Thunnissen, G. Roelfes, Angew. Chem., Int. Ed. 2018, 57, 7785–7789.
[26] Y. Wang, M. J. Bartlett, C. A. Denard, J. F. Hartwig, H. Zhao, ACS Catal. 2017, 7, 2548–
2552.
[27] K. D. Belsare, M. C. Andorfer, F. S. Cardenas, J. R. Chael, H.-J. Park, J. C. Lewis, ACS
Synth. Biol. 2017, 6, 416–420.
[28] P. Dydio, H. M. Key, A. Nazarenko, J. Y. E. Rha, V. Seyedkazemi, D. S. Clark, J. F.
Hartwig, Science 2016, 354, 102–106.
[29] J. H. Hansen, B. T. Parr, P. Pelphrey, Q. Jin, J. Autschbach, H. M. L. Davies, Angew.
Chem., Int. Ed. 2011, 50, 2544–2548.
[30] Y. Yanto, C. K. Winkler, S. Lohr, M. Hall, K. Faber, A. S. Bommarius, Org. Lett. 2011, 13,
2540–2543.
[31] C. Stueckler, M. Hall, H. Ehammer, E. Pointner, W. Kroutil, P. Macheroux, K. Faber, Org.
Lett. 2007, 9, 5409–5411.
[32] Z. Chen, F. Vohidov, J. M. Coughlin, L. J. Stagg, S. T. Arold, J. E. Ladbury, Z. T. Ball, J.
Am. Chem. Soc. 2012, 134, 10138–10145.
[33] J. Wu, X. Li, X. Qi, X. Duan, W. L. Cracraft, I. A. Guzei, P. Liu, W. Tang, J. Am. Chem.
Soc. 2019, 141, 19902–19910.
[34] R. Breslow, S. Bandyopadhyay, M. Levine, W. Zhou, ChemBioChem 2006, 7, 1491–1496.
[35] K. Ellis-Guardiola, H. Rui, R. L. Beckner, P. Srivastava, N. Sukumar, B. Roux, J. C. Lewis,
Biochemistry 2019, 58, 1616–1626.
[36] Z. T. Ball, Acc. Chem. Res. 2013, 46, 560–570.
[37] Y. Valasatava, A. Rosato, N. Furnham, J. M. Thornton, C. Andreini, J. Inorg. Biochem.
2018, 179, 40–53.
This article is protected by copyright. All rights reserved.

10.1002/anie.202107982
Angewandte Chemie International Edition
TOC Graphic:
Directed evolution was used to improve dirhodium artificial metalloenzyme (ArM)-catalyzed
diazo cross-coupling yield and selectivity. The engineered ArM was then used in cascade
reactions where the alkene product was reduced by an ene reductase, with the ArM scaffold
controlling chemoselectivity in the reaction mixture. MD simulations of the scaffold show
that the cofactor modulates the domain dynamics and favors a more closed, hydrophobic
pocket conducive to selective catalysis.
This article is protected by copyright. All rights reserved.