Effect of proximal ligand substitutions on the carbene and nitrene transferase activity of myoglobin

Article abstract of DOI:10.1016/j.tet.2019.03.009

Engineered myoglobins were recently shown to be effective catalysts for abiological carbene and nitrene transfer reactions. Here, we investigated the impact of substituting the conserved heme-coordinating histidine residue with both proteinogenic (Cys, Ser, Tyr, Asp) and non-proteinogenic Lewis basic amino acids (3-(3′-pyridyl)-alanine, p-aminophenylalanine, and β-(3-thienyl)-alanine), on the reactivity of this metalloprotein toward these abiotic transformations. These studies showed that mutation of the proximal histidine residue with both natural and non-natural amino acids result in stable myoglobin variants that can function as both carbene and nitrene transferases. In addition, substitution of the proximal histidine with an aspartate residue led to a myoglobin-based catalyst capable of promoting stereoselective olefin cyclopropanation under nonreducing conditions. Overall, these studies demonstrate that proximal ligand substitution provides a promising strategy to tune the reactivity of myoglobin-based carbene and nitrene transfer catalysts and provide a first, proof-of-principle demonstration of the viability of pyridine-, thiophene-, and aniline-based unnatural amino acids for metalloprotein engineering.

Full text of DOI:10.1016/j.tet.2019.03.009

Accepted Manuscript
Effect of proximal ligand substitutions on the carbene and nitrene transferase activity
of myoglobin
Eric J. Moore, Rudi Fasan
PII:
S0040-4020(19)30284-4
https://doi.org/10.1016/j.tet.2019.03.009
TET 30195
DOI:
Reference:
To appear in: Tetrahedron
Received Date: 26 December 2018
Revised Date: 2 March 2019
Accepted Date: 8 March 2019
Please cite this article as: Moore EJ, Fasan R, Effect of proximal ligand substitutions on the carbene
and nitrene transferase activity of myoglobin, Tetrahedron (2019), doi: https://doi.org/10.1016/
j.tet.2019.03.009.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Effect of proximal ligand substitutions on the carbene
and nitrene transferase activity of myoglobin
Eric J. Moorea and Rudi Fasan*,a
Department of Chemistry, University of Rochester, Rochester, NY 14627, United States. Email:
rfasan@ur.rochester.edu
____________________________________________________________________________________
Abstract: Engineered myoglobins were recently shown to be effective catalysts for abiological
carbene and nitrene transfer reactions. Here, we investigated the impact of substituting the
conserved heme-coordinating histidine residue with both proteinogenic (Cys, Ser, Tyr, Asp) and
non-proteinogenic Lewis basic amino acids (3-(3’-pyridyl)-alanine, p-aminophenylalanine, and
β-(3-thienyl)-alanine), on the reactivity of this metalloprotein toward these abiotic
transformations. These studies showed that mutation of the proximal histidine residue with both
natural and non-natural amino acids result in stable myoglobin variants that can function as both
carbene and nitrene transferases. In addition, substitution of the proximal histidine with an
aspartate residue led to a myoglobin-based catalyst capable of promoting stereoselective olefin
cyclopropanation under nonreducing conditions. Overall, these studies demonstrate that proximal
ligand substitution provides a promising strategy to tune the reactivity of myoglobin-based
carbene and nitrene transfer catalysts and provide a first, proof-of-principle demonstration of the
viability of pyridine-, thiophene-, and aniline-based unnatural amino acids for metalloprotein
engineering.
1

ACCEPTED MANUSCRIPT
Heme-containing proteins have recently emerged as promising biocatalysts for promoting
an array of abiological carbene and nitrene transfer reactions useful for the construction of new
carbon-carbon1-14 and carbon-nitrogen bonds,15-20 respectively. In particular, our group has
previously reported the versatility and synthetic value of engineered myoglobins for catalyzing
stereoselective olefin cyclopropanation reactions in the presence of diazo-based carbene
donors3,4,21,11,22 and C—H amination reactions in the presence of azide-based nitrene donors,17
among other abiological reactions.23 The reactivity of myoglobin in these transformations derives
from its hemin cofactor (iron-protoporphyrin IX), whose iron center is coordinated at the axial
position by a conserved ‘proximal’ histidine residue (His93 in sperm whale myoglobin).24 The
latter is implicated in mediating binding of the cofactor to the protein as well as in modulating
the oxygen binding properties of this metalloprotein.24
Modification of the metalloporphyrin cofactor environment in myoglobin (Mb)
represents a promising approach toward tuning or enhancing the reactivity of myoglobin in the
context of nitrene and carbene transfer reactions. In previous studies, we and others have
demonstrated that substitution of the iron center with other metals (e.g., Mn, Co, Rh, Ru, Ir) led
to Mb variants with altered nitrene17 or carbene transferase activity,25-27 also enabling reactions
that cannot be achieved with the iron-containing counterparts.25,26 In another study, the
combination of metal substitution with a proximal ligand mutation (serine-ligated Co-Mb)
enabled the development of a Mb-based catalyst capable of favoring olefin cyclopropanation
over the Y—H carbene insertion reactions.11 Furthermore, replacement of the protoporphyrin IX
ligand with other porphyrin derivatives has yielded Mb-based cyclopropanation catalysts with
distinct catalytic properties compared to wild-type Mb.28,9 In the context of P450 enzymes,
substitution of the axial cysteine ligand coordinating heme with serine was reported to favor
2

ACCEPTED MANUSCRIPT
nitrene transfer reactions such as aziridination and intermolecular C—H amination15,18,20 or favor
cyclopropanation over S—H carbene insertion.29 Substitution of the cysteine axial ligand with
selenocysteine also affected the reactivity of P450s as monooxygenases.30 More recently,
substitution of the proximal histidine ligand in Mb with Nδ-methylhistidine (NMH) was found to
increase its peroxidase activity31 and the same modification in an engineered Mb-based carbene
transferase3 resulted in increased cyclopropanation activity under non-reducing conditions.32
Despite this progress, the potential value of altering the iron first coordination sphere residue in
engineered myglobins toward modulating their non-native carbene and nitrene transferase
reactivity have remained largely undefined. In addition, the feasibility of exploiting non-
canonical amino acids for this purpose beyond N-methyl-histidine has not been established.
Here, we addressed these questions through the preparation of a panel of proximal ligand Mb
variants (Figure 1) and their characterization in terms of their stability and spectroscopic
properties and reactivity in representative carbene and nitrene transfer reactions.
Figure 1. Overview of myoglobin structure and chemical structures of the natural and non-
canonical amino acid residues introduced at the heme proximal axial ligand position (His93).
3

ACCEPTED MANUSCRIPT
For these studies, a series of proximal ligand variants were prepared based on
Mb(H64V,V68A), an engineered active site variant of sperm whale myoglobin that exhibits high
activity and stereoselectivity in olefin cyclopropanation with ethyl α-diazoacetate (EDA)3,4 and
other acceptor-only diazo compounds,21,22 along with intramolecular C—H amination activity in
the presence of sulfonyl azides.17 A such, this variant was deemed useful for assessing the impact
of altering the heme coordination environment on both Mb-mediated carbene and nitrene transfer
activity and on subtle properties such as diastereo- and enantioselectivity. Accordingly, a first set
of Mb(H64V,V68A)-based variants were generated upon substituting the proximal histidine
(His93) for another proteinogenic nucleophilic amino acid capable of metal coordination,
namely, Ser, Cys, Tyr, and Asp. Cys and Tyr are involved in heme ligation in P450s33 and
catalase,34 respectively, whereas Asp is implicated in metal coordination in various
metalloenzymes.35,36 To further extend the range of nucleophilic groups to be introduced at the
proximal site of the hemoprotein beyond that allowed by proteinogenic amino acids, we also
targeted the preparation of Mb(H64V,V68A) variants in which His93 is replaced with the non-
canonical amino acid (ncAA) p-aminophenylalanine (pAmF),37 3-(3’-pyridyl)-alanine (3PyA),38
and β-(3-thienyl)-alanine (3ThA)38 (Figure 1). While sharing a similar size with histidine, these
amino acids incorporate a side-chain group with significantly different basicity/nucleophilicty, as
supported by the different pKa and/or Lewis basicity of aniline, pyridine and thiophene
compared to imidazole (or N-methyl-imidazole).39 To generate these variants, a
Mb(H64V,V68A) construct containing an amber stop codon (TAG) at position 93 was expressed
in the presence of the desired ncAA and orthogonal aminoacyl-tRNA synthetases previously
developed for the genetic incorporation of these ncAAs37,38 via the amber stop codon suppression
technology.40,41 Gratifyingly, all of these ncAA-containing Mb variants could be successfully
4

ACCEPTED MANUSCRIPT
produced and isolated from E. coli. In addition, selective incorporation of the corresponding
ncAA at the desired position (93) was confirmed by LC-MS and MS/MS analysis of peptide
fragments obtained after proteolytic digestion of these proteins (Figure S1).
Analysis of the electronic absorption spectrum of the purified Mb(H64V,V68A) variants
showed the presence of a prominent Soret band, indicating proper incorporation of the hemin
cofactor in each case, including the ncAA-containing variants. At the same time, side-by-side
comparison of the UV-vis spectra of the proximal ligand-substituted Mb variant in their ferric,
ferrous and CO-bound forms show distinguishable features compared to the histidine-ligated
Mb(H64V,V68A) (Figure 2, Table S1). Ferric Mb(H64V,V68A) shows a Soret band with λmax
at 409 nm and Q bands at 530 and 636 nm, characteristic of a hexacoordinated high-spin iron
species.24 Reduction to the ferrous state with sodium dithionite results in a red shift of the Soret
band to 432 nm, which undergoes a blue-shift to 424 nm upon complexation with CO. As
summarized in Table S1, differences in the absorption maxima for the Soret band and/or Q
bands are observable for the proximal ligand-substituted variants when compared to each other
and to Mb(H64V,V68A), for one or more forms of the hemoprotein (i.e., ferric, ferrous, CO-
bound form). For example, the Soret bands of Mb(H64V,V68A,H93(3PylA)) have λmax at 412
(ferric)/419(ferrous)/418(CO-bound) compared to 410/424/420 for Mb(H64V,V68A,H93pAmF),
406/421/420 for Mb(H64V,V68A,H93S), and 409/432/424 for Mb(H64V,V68A) (Table S1).
Moreover, the spectroscopic features of these proximal ligand variants are distinct from those of
Mb(H64V,V68A) variants in which His93 was substituted with a non-nucleophilic Phe residue
(405/417/415)25 or N-methyl-His, the latter being reported to exhibit a UV-vis spectrum
indistinguishable from that of Mb(H64V,V68A).32 Altogether, these spectroscopic analyses
5

ACCEPTED MANUSCRIPT
indicated that the heme cofactor environment in Mb(H64V,V68A) is altered upon mutation of
the proximal histidine residue with both the natural and unnatural amino acids.
Figure 2: Electronic absorption spectra for representative proximal ligand Mb variants. (A)
Mb(H64V,V68A,H93D), (B) Mb(H64V,V68A,H93(3PylA)), (C) Mb(H64V,V68A,H93(pAmF))
and (D) Mb(H64V,V68A,H93(3ThA)). Spectra correspond to the proteins in their ferric (blue
line), ferrous (orange line) (solid line), and CO-bound form (grey line). The λmax for the Soret
band are indicated. See Table S1 for additional spectroscopic data on these and the other Mb
variants.
To examine the impact of these mutations on protein stability, the Mb variants were
further characterized with respect to their stability against thermal denaturation. Consistent with
6

ACCEPTED MANUSCRIPT
previous observations,42 Mb(H64V,V68A) was determined to have an apparent Tm of 66.0 °C
using variable temperature circular dichroism spectroscopy (Figure 3, entry 1). In comparison,
the proximal ligand variants were found to show a general reduction in thermostability, the
majority exhibiting an approximately 10-12 °C reduction in Tm value. These results are not
entirely surprising considering the structural role of the proximal histidine residue in Mb.24 The
largest change in thermostability was observed with the H93D substitution (Tm: 47.9 vs. 66.0
°C), whereas smallest decrease in thermostability was observed for the H93Y mutation (Tm: 60.9
vs. 66.0 °C). Notably, the mutations involving the non-canonical amino acids pAmF, 3PylA, and
3ThA were well tolerated by the myoglobin scaffold, resulting in folded proteins (Figure 1) with
melting temperatures of 55-57 °C (Figure 3). Thus, despite the reduction in Tm compared to the
parent protein, the stability of these and the other proximal ligand Mb variants remain
comparable or higher than that of heme-dependent enzymes from mesophilic organisms (Tm = 45
– 55 °C).43,44
Mb(H64V,V68A)
Mb(H64V,V68A,H93Y)
Mb(H64V,V68A,H93(3PyA))
1
Mb(H64V,V68A,H93D)
0.8
0.6
0.4
0.2
0
Entry Variant
Apparent Tm (°C)
66.0 ± 1.0
55.3 ± 0.8
55.6 ± 0.4
57.3 ± 1.5
58.5 ± 0.5
60.9 ± 0.8
55.1 ± 0.6
47.9 ± 2.0
1
2
3
4
5
6
7
8
Mb(H64V,V68A)
Mb(H64V,V68A,H93(3PyA))
Mb(H64V,V68A,H93(pAmF))
Mb(H64V,V68A,H93(3ThA))
Mb(H64V,V68A,H93C)
20
40
60
80
100
Mb(H64V,V68A,H93Y)
Temperature (°C)
Mb(H64V,V68A,H93S)
Mb(H64V,V68A,H93D)
Figure 3. Thermal denaturation curves for Mb(H64V,V68A) and selected proximal ligand
mutants as determined by variable temperature circular dichroism spectroscopy. The Table
reports the apparent Tm values for Mb(H64V,V68A) and its variants as determined by fitting raw
7

ACCEPTED MANUSCRIPT
CD data to a sigmoidal equation via nonlinear regression analysis. Mean values and standard
errors are derived from experiments performed in triplicate.
Encouraged by these results, we examined the carbene transfer activity of the proximal
ligand Mb variants using the cyclopropanation of styrene (1) in the presence of ethyl α-
diazoacetate (2) as a model reaction. Under ‘standard’ (i.e., anaerobic and reducing) conditions,3
Mb(H64V,V68A) catalyzes the nearly quantitative conversion of the olefin to the
cyclopropanation product 3 (97%, 1,940 TON) with high enantio- and diasteroselectivity (99%
de and ee; Table 1, Entry 1). Under these conditions, nearly all of the proximal ligand-
substituted variants, including those incorporating the non-canonical amino acids pAmF, 3PylA
and 3ThA, were found to be viable cyclopropanation catalysts, supporting between 900 and
1,100 catalytic turnovers (TON) in this reaction (Table 1). As an exception, the serine-
containing variant Mb(H64V,V68A,H93S) shows a significant reduction in cyclopropanation
activity compared to the parent protein (25% vs. 98% conv.). Among the proximal ligand
variants, Mb(H64V,V68A,H93D) emerged as the most efficient and selective cyclopropanation
biocatalyst, producing 3 with nearly identical turnovers and diastereoselectivity (1,930 TON,
95% de) and only moderately decreased enantioselectivity (75% ee) compared to
Mb(H64V,V68A) (Table 1, Entry 8).
We next performed the cyclopropanation reactions in the presence of oxygen
(‘aerobic/reducing’ conditions) and in the absence of reductant (‘anaerobic/nonreducing
conditions’), the latter being beneficial for reducing the ferric protein to the catalytically active
ferrous form.3 Under aerobic conditions, all of the proximal ligand variants, except
Mb(H64V,V68A,H93D), showed a ~ 2-fold reduction in cyclopropanation reactivity compared
8

ACCEPTED MANUSCRIPT
to the anaerobic reaction, likely due to inhibition resulting from oxygen binding to the heme
cofactor. In stark contrast, Mb(H64V,V68A,H93D) maintains high cyclopropanation activity
also in the presence of oxygen, producing 3 in 70% conversion and 1,400 turnovers. Notably,
Mb(H64V,V68A,H93D) was found to exhibit even higher cyclopropanation activity along with
higher diastereo- and enantioselectivity under non-reducing conditions (78% conv., 1,590 TON,
99% de, 92% ee). This result is in contrast to Mb(H64V,V68A) and the other proximal ligand
variants, which show significantly reduced (27% conv.) and negligible cyclopropanation activity
(≤ 6% conv.), respectively, under the same reaction conditions. Thus, the introduction of Asp at
the proximal ligand site of the hemoprotein has a unique effect toward enabling Mb to function
as a carbene transferase under non-reducing conditions. Moreover, this mutation can recapitulate
the effect previously observed for the same Mb variant upon substitution of His93 with N-
methyl-histidine.32 To investigate the involvement of ferric vs. ferrous species in
Mb(H64V,V68A,H93D)-catalyzed cyclopropanation, the reaction under anaerobic/non-reducing
conditions was also performed in the presence of CO-saturated buffer, which resulted in no
product formation. As CO binds with high affinity only to ferrous heme, these results suggested
that the ferrous form of Mb(H64V,V68A,H93D) is responsible for catalysis. However, since it is
known that CO is able to slowly reduce ferric hemoproteins to their ferrous form,45 further
experiments were performed to rule out the possibility that ferric Mb(H64V,V68A,H93D) may
be reduced by CO and then inhibited by it over the course of these reactions. Incubation of ferric
Mb(H64V,V68A,H93D) with CO showed no formation of the ferrous CO-bound species up to
40-45 min, and formation of a only small amount of this species (<10%) at more extended
incubation times (> 60 min) (Figure S3). On the other hand, analysis of ferric
Mb(H64V,V68A,H93D)-catalyzed cyclopropanation of styrene with EDA after short reaction
9

ACCEPTED MANUSCRIPT
times (i.e., 5 and 10 min, within which period no ferrous Mb(H64V,V68A,H93D) is formed
upon reduction by CO; Figure S3) showed formation of the cyclopropanation production in the
absence of CO, but not in the presence of it. Based on these results, we conclude that the H93D
mutation enables Mb(H64V,V68A,H93D) to be effectively reduced to the catalytically active
ferrous form by the diazo compound in situ,46-48 resulting in efficient olefin cyclopropanation
without the need of an exogeneous reductant.
Table 1: Catalytic activity and selectivity of proximal ligand-substituted Mb variants in the
cyclopropanation of styrene with ethyl α-diazoacetate (EDA).[a]
Reaction Conditions:
Entry Catalyst
Anaerobic / Reducing
Aerobic / Reducing
TON Conv.[b] % de (ee)[c]
Anaerobic / Non-reducing
TON
1940
1045
1115
845
Conv.[b]
97%
52%
56%
42%
45%
45%
25%
96%
% de (ee)[c]
TON Conv.[b]
% de (ee)[c]
>99 (98)
95 (62)[d]
90 (36)[d]
93 (60)[d]
88 (29)
1
2
3
4
5
6
7
8
Mb(H64V,V68A)
>99 (>99)
84 (8)
1560
625
905
590
540
475
270
1400
78%
31%
45%
29%
27%
24%
13%
70%
97 (96)
85 (11)
84 (5)
545
105
50
27%
5%
3%
2%
6%
5%
2%
79%
Mb(H64V,V68A,H93(3PyA))
Mb(H64V,V68A,H93(pAmF))
Mb(H64V,V68A,H93(3ThA))
Mb(H64V,V68A,H93C)
Mb(H64V,V68A,H93Y)
Mb(H64V,V68A,H93S)
Mb(H64V,V68A,H93D)
84 (5)
83 (8)
84 (10)
90 (22)
90 (31)
89 (18)
94 (68)
45
905
90 (27)
90 (26)
90 (19)
95 (75)
125
110
40
910
93 (42)
490
86 (14)
1930
1590
99 (92)
[a] Reaction Conditions: 0.01 M styrene, 0.02 M EDA, 5 µM Mb catalyst, 50 mM potassium phosphate buffer (pH 7)
containing 5% ethanol, room temperature.
[b] As determined by GC using calibration curves with isolated 3.
[c] For trans-(1S,2S) stereoisomer as determined by chiral GC.
[d] Diastereomeric and enantiomeric ratios may be overestimated due to low product formation in these reactions.
Mb(H64V,V68A) was previously shown to catalyze the cyclization of sulfonylazides to
give the corresponding sultam,17 a reaction involving a nitrene C(sp3)—H insertion via a heme-
10

ACCEPTED MANUSCRIPT
nitrenoid intermediate.16,19 To test the effect of the mutations on the nitrene transfer reactivity of
this protein, a model C—H amination reaction with 2,4,6-triisopropyl-benzenesulfonyl azide (4)
was used (Table 2). Under anaerobic and reducing conditions, Mb(H64V,V68A) catalyzes the
conversion of 4 into the sultam 5 with 32% yield (GC) and 635 catalytic turnovers (Table 2,
Entry 1). Interestingly, all of the proximal ligand variants proved to be functional and efficient
nitrene transferases, supporting between 290 and 650 turnovers (Table 2). This catalytic activity
compares well with that of synthetic organometallic catalysts previously reported for this
transformation (50-100 TON).49 Furthermore, the ncAA-containing variants
Mb(H64V,V68A,H93(3PylA)) and Mb(H64V,V68A,H93pAmF) as well as
Mb(H64V,V68A,H93D) were found to exhibit comparable efficiency as nitrene transferases as
the histidine-ligated Mb(H64V,V68A) (635-650 vs. 635 TON, Table 2), thus indicating a
significantly broader tolerance of this non-native reactivity to alteration of the heme coordinating
residue compared to the carbene transferase reactivity (Table 1). In contrast to the results
obtained with the cyclopropanation reaction (Table 1), a significant reduction in C—H
amination activity was observed for all the Mb variants in the aerobic reactions, with
Mb(H64V,V68A,H93pAmF) being the most efficient biocatalyst under these conditions. No C—
H amination activity was observed instead in the absence of reductant with all the variants,
indicating that, unlike for Mb(H64V,V68A,H93D) in the presence of the diazo reagent, the
catalytically active ferrous form of these proteins is not produced in the presence of the azide
reagent. Altogether, the results further highlight the differential impact of proximal ligand
substitutions on Mb reactivity in the context of the two group transfer reactions.
11

ACCEPTED MANUSCRIPT
Table 2. Catalytic activity of proximal ligand-substituted Mb variants for C-H amination with
2,4,6 triisopropylbenzenesulfonyl azide.[a]
Reaction Conditions:
Entry Catalyst
Anaerobic + Reductant
Aerobic + Reductant
TON
635
650
640
535
290
475
405
635
Conv.[b]
32%
33%
32%
27%
14%
24%
20%
32%
TON
76
70
95
72
48
44
24
15
Conv.[b]
1
2
3
4
5
6
7
8
Mb(H64V,V68A)
4%
Mb(H64V,V68A,H93(3PyA))
Mb(H64V,V68A,H93pAmF)
Mb(H64V,V68A,H93(3ThA))
Mb(H64V,V68A,H93C)
Mb(H64V,V68A,H93Y)
Mb(H64V,V68A,H93S)
Mb(H64V,V68A,H93D)
3%
5%
4%
2%
2%
1%
1%
[a] Reaction conditions: 0.01 M triisopropylbenzenesulfonyl azide, 5 µM Mb catalyst, 50 mM phosphate buffer (pH
7) containing 2.5% ethanol, room temperature.
[b] As determined by GC calibration curves using isolated 5.
In summary, this work demonstrates how modification of the heme cofactor environment
through substitution of the proximal axial ligand can be exploited to alter and modulate the
carbene and nitrene transfer reactivity of myoglobin-based catalysts. Extending upon our
12

ACCEPTED MANUSCRIPT
previous findings,25,28 these results highlight the robustness and functional plasticity of the
myoglobin scaffold toward tolerating structural alterations of the heme cofactor environment for
tuning such non-native reactivities. Most notably, the present strategy proved effective for
obtaining a Mb catalyst, Mb(H64V,V68A,H93D), that is capable of promoting efficient and
stereoselective olefin cyclopropanation under nonreducing conditions. This reactivity, which was
uniquely imparted by replacement of the proximal histidine ligand with an aspartate residue,
results for the enhanced ability of this Mb variant to undergo conversion to the catalytically
active, ferrous form directly upon reaction with the diazo substrate. Finally, the functionality of
the ncAA-containing Mb variants as carbene and nitrene transferases demonstrate the utility of
p-aminophenylalanine (pAmF), 3-(3’-pyridyl)-alanine (3PyA), and β-(3-thienyl)-alanine (3ThA)
for re-engineering the first coordination sphere of this metalloprotein. Although these Mb
variants did not outperform those generated by proximal ligand substitutions via proteinogenic
amino acids in the context of the present reactions, these findings provide support to the
feasibility of exploiting these pyridine-, thiophene-, and aniline-based unnatural amino acids for
myoglobin engineering, and more broadly, for the engineering of metalloenzymes and
metalloproteins.
Materials and Methods
General Information
All chemicals and reagents were purchased from commercial suppliers and used without any
further purification unless otherwise stated. EDA was purchased from Sigma-Aldrich as 87%
13

ACCEPTED MANUSCRIPT
m/v solution in dichloromethane. The unnatural amino acids were purchased from Chem-Impex.
LC-MS grade trypsin was purchased from Pierce.
Cloning and Mutagenesis
The cloning of a pET22-based vector for the expression of Mb variant Mb(H64V,V68A) was
described previously.3 In this vector, the protein gene is under a IPTG-inducible T7 promoter.
Additional mutations were incorporated into the Mb(H64V,V68A) template using SOE PCR and
the primers listed in Table S2. The preparation of a pEVOL-based plasmid50 for the expression
of an engineered M. jannaschii tyrosyl-tRNA synthetase (TyrRS) and cognate amber suppressor
tRNA (MjtRNACUATyr) for the genetic incorporation of pAmF was described previously.51 Using
a similar procedure, a second pEVOL-based plasmid was prepared for encoding for a
polyspecific AARS derived from Methanosarcina barkeri pyrrolysyl-tRNA synthetase and the
cognate amber suppressor tRNA (MbtRNACUAPyl) for the genetic incorporation of 3-
pyridylalanine and 3-thienylalanine.38 This gene was prepared by introducing the mutations
L270I, Y271F, L274G, C313F, and Y349F into the MbPylRS gene via site-directed mutagenesis
and the resulting gene was re-introduced into the pEVOL vector using the Bgl II/Sal I cassette. In
these vectors, the AARS gene is under an arabinose-inducible araBAD promoters and the tRNA
is under a constitutive (proK) promoter.
Growth media
Cell cultures were grown in enriched M9 medium, which was prepared as follows. For 1 L, 740
mL of deionized H2O was added with 200 mL of M9 salts (5×) solution, 50 mL of 10% (w/v)
yeast extract, 10 mL of glycerol, 1 mL of MgSO4 (2 M), and 100 ꢀL of CaCl2 (1 M). The M9
salt (5×) solution was prepared by dissolving 34 g of Na2HPO4, 15 g of K2HPO4, 2.5 g of NaCl,
14

ACCEPTED MANUSCRIPT
5 g of NH4Cl in 1 L of deionized H2O and sterilized by autoclaving. The glycerol and MgSO4
solutions were autoclaved separately. The CaCl2 stock solution was sterilized by filtration.
Enriched M9 agar plates were prepared by adding 17 g of agar to 1 L of enriched M9 media
containing all of the aforementioned components at the specified concentrations with the
exception of CaCl2, which was added immediately prior to plating. To media and plates was
added ampicillin to a final concentration of 100 mg/L and chloramphenicol was added to a final
concentration of 34 mg/L.
Protein expression
The Mb variants containing all natural amino acids were expressed from pET22-based vectors in
E. coli BL21(DE3) as described previously.3 The ncAA-containing Mb variants were expressed
from E. coli BL21(DE3) cells co-transformed with the pET22-based vector encoding for
Mb(H64V,V68A,H93(TAG)) and the appropriate pEVOL-based vector for the expression of
pAmF-RS (pAmF) or PylHRS (3PyA and 3ThA). After selection on M9 agar plates containing
ampicillin (100 mg L-1) and chloramphenicol (34 mg L-1), single colonies of the transformed
cells were used to inoculate 5 mL overnight cultures of enriched M9 medium (37°C, 200 rpm, 12
hours). The overnight cultures were used to inoculate 1L of enriched M9 medium. When OD600
reached 0.4 – 0.6, cell cultures were condensed by centrifugation (4,000 rpm, 4°C, 20 min)
followed by resuspension of the cell pellet in 200 mL of fresh enriched M9 medium. Arabinose,
δ-aminolevulinic acid, and the desired amino acid were added to the expression media at final
concentrations of 4 mM, 0.3 mM, and 5 mM respectively followed by incubation for 30 minutes
with shaking (27°C, 200 rpm) after which protein expression was induced with 0.5 mM IPTG.
Following induction, cultures were shaken at 200 rpm and 27°C and harvested after 20 hours by
centrifugation (4000 rpm, 4°C, 20 minutes). Cell pellets were resuspended in 20 mL of Ni-NTA
15

ACCEPTED MANUSCRIPT
Lysis Buffer (50 mM KPi, 250 mM NaCl, 10 mM histidine, pH 8.0) and protein extraction and
purification were performed immediately.
Protein Purification and Characterization
Cell suspensions were lysed by sonication and clarified by centrifugation (14,000 rpm, 20 min,
4°C). The clarified lysate was transferred to a Ni-NTA column equilibrated with Ni-NTA Lysis
Buffer (50 mM KPi, 250 mM NaCl, 10 mM histidine, pH 8.0). The resin was washed with 50
mL of Ni-NTA Lysis Buffer followed by 50 mL of Ni-NTA Wash Buffer (50 mM KPi, 250 mM
NaCl, 20 mM histidine, pH 8.0). Proteins were eluted with Ni-NTA Elution Buffer (50 mM KPi,
250 mM NaCl, 250 mM histidine, pH 7.0). After elution from the Ni-NTA column, the protein
was buffer exchanged against 50 mM KPi buffer (pH 7.0) using 10 kDa Centricon filters. The
concentration of the Mb variants was determined using extinction coefficients as determined
using the hemochrome assay52 (Table S3). UV-vis absorbance spectra were recorded using a
protein solution at 2 - 5 µM in 50 mM KPi (pH 7.0). For LC-MS analyses, the ncAA-containing
amino acids were digested with trypsin in 50 µL reactions containing 2.5 µg trypsin, 10 mM
CaCl2, and 60 – 70 µg of myoglobin in 50 mM NH4HCO3 (pH 8.0). Digestion reactions
incubated at 37 °C without shaking for 12 – 15 hours and were cleaned up using the C18 ZipTip
protocol. Digested samples were given to the URMC Mass Spectrometry Resource Facility for
LC-MS/MS analysis.
Cyclopropanation and C—H Amination Reactions
The cyclopropanation reactions with styrene and EDA and the C—H amination reactions with
2,4,6-triisopropyl-benzenesulfonyl azide were carried out as described previously.3,17 Briefly,
cyclopropanation reactions under anaerobic/reducing conditions were carried out at a 400 µL
16

ACCEPTED MANUSCRIPT
scale using 5 µM myoglobin catalyst, 10 mM styrene, 20 mM EDA, and 10 mM dithionite in 50
mM KPi (pH 7.0). For analysis of C—H amination activity, reactions were typically carried out
at a 400 µL scale using 5 µM myoglobin catalyst, 10 mM 2,4,6 triisopropylbenzenesulfonyl
azide, and 10 mM dithionite in 50 mM KPi (pH 7.0). Under anaerobic conditions, reactions
performed in an anaerobic chamber and were stirred for 16 hours at room temperature and
aerobic reactions were performed in vessels open to the atmosphere. The reactions under non-
reducing conditions were performed in a similar manner without the addition of sodium
dithionite.
Product Analysis
The products of the cyclopropanation and C—H amination reactions were analyzed by chiral GC
and HPLC, respectively, using synthetically prepared (racemic) standards of 34 and 517 as
references. The cyclopropanation reactions were analyzed by adding 20 ꢀL of internal standard
(benzodioxole, 50 mM in ethanol) to the reaction mixture followed by extraction with 400 ꢀL of
dichloromethane and analysis by gas chromatography (GC). GC analyses were carried out using
a Shimadzu GC-2010 gas chromatograph equipped with an FID detector and a Chiral Cyclosil-B
column (30 m × 0.25 mm × 0.25 ꢀm film). Separation method: 1 ꢀL injection, injector temp: 200
°C, detector temp: 300 °C. Gradient: column temperature set at 120 °C for 3 min, then to 150 °C
at 0.8 °C/min, then to 245 °C at 25 °C/min. Total run time was 46.30 min. For the reactions with
2,4,6 triisopropylbenzenesulfonyl azide, the DCM extraction solution was evaporated, and the
residue was redissolved in 400 ꢀL of methanol followed by HPLC analysis using a Shimadzu
LC-2010A instrument equipped with a UV−vis detector. Analytical conditions: Thermo Fisher
Scientific Hypersil GOLD C18 column (250 × 4.6 mm); UV detection: 254 nm; flow rate: 1.0
mL/ min: solvent A: 1% trifluoroacetic acid in H2O; solvent B: 1% trifluoroacetic acid in
17

ACCEPTED MANUSCRIPT
acetonitrile; gradient: 0−3 min: 30% B; 3−21 min: 30% B to 90% B; 21−24 min: 90% B;
24−24.5 min: 90% B to 30% B; 24.5−30 min: 30% B. Retention times: 17.35 min for 5; and
24.54 min for 4. Calibration curves for quantification of each product were constructed using
authentic standards prepared synthetically. All measurements were performed at least in
duplicate.
Analysis of thermostability
Circular dichroism (CD) spectroscopy and thermal denaturation analysis were performed using a
JASCO J-1100 CD spectrophotometer equipped with variable temperature/wavelength
denaturation software. Far UV CD spectra (250−190 nm) were obtained using 3 ꢀM solutions of
purified Mb variant in 50 mM potassium phosphate buffer (pH 7.0) and recorded at 20 °C at a
scan rate of 50 nm/min with a bandwidth of 1 nm and an averaging time of 10 s per
measurement. Thermal denaturation curves were measured by monitoring the change in molar
ellipticity at 222 nm (θ222) over a temperature range from 20 to 100 °C. The temperature increase
was set to 0.5 °C per minute with an equilibration time of 10 s. Data integration time for the melt
curve was set to 4 s with a bandwidth of 1 nm. Linear baselines for the folded (θf) and unfolded
state (θu) were generated. Melting temperature (Tm) was determined using the low temperature
(θf = mfT + bf ) and high temperature (θu = muT + bu) equations fitted to the experimental data
before and after global unfolding, respectively. These data were normalized and converted to
fraction of folded protein (Ff) vs. temperature plots, and the resulting curve was fitted to a
sigmoidal equation (θfit) via nonlinear regression analysis in SigmaPlot from which apparent
melting temperatures were derived. See Figure S for additional data.
ACKNOWLEDGEMENTS
18

ACCEPTED MANUSCRIPT
This work was supported by the U.S. National Science Foundation grant CHE-1609550 and in
part by the U.S. National Institute of Health (NIH) grant GM098628. E.J.M. acknowledges
support from the NIH Graduate Training Grant T32GM118283. The authors are grateful to Dr.
Jermaine Jenkins and the Structure Biology & Biophysics Facility at the University of Rochester
for providing access to the CD instrumentation.
References
1.
2.
Coelho, P. S.; Brustad, E. M.; Kannan, A.; Arnold, F. H. Science 2013, 339, 307-310.
Coelho, P. S.; Wang, Z. J.; Ener, M. E.; Baril, S. A.; Kannan, A.; Arnold, F. H.; Brustad,
E. M. Nat. Chem. Biol. 2013, 9, 485-487.
3.
4.
Bordeaux, M.; Tyagi, V.; Fasan, R. Angew. Chem. Int. Ed. 2015, 54, 1744–1748.
Bajaj, P.; Sreenilayam, G.; Tyagi, V.; Fasan, R. Angew. Chem. Int. Ed. 2016, 55, 16110–
16114.
5.
6.
Tyagi, V.; Fasan, R. Angew. Chem. Int. Ed. 2016, 55, 2512-2516
Tyagi, V.; Sreenilayam, G.; Bajaj, P.; Tinoco, A.; Fasan, R. Angew. Chem. Int. Ed. 2016,
55, 13562-13566.
7.
Wang, Z. J.; Renata, H.; Peck, N. E.; Farwell, C. C.; Coelho, P. S.; Arnold, F. H. Angew.
Chem. Int. Ed. 2014, 53, 6810-6813.
8.
Weissenborn, M. J.; Low, S. A.; Borlinghaus, N.; Kuhn, M.; Kummer, S.; Rami, F.;
Plietker, B.; Hauer, B. Chemcatchem 2016, 8, 1636-1640.
9.
Ohora, K.; Meichin, H.; Zhao, L. M.; Wolf, M. W.; Nakayama, A.; Hasegawa, J.;
Lehnert, N.; Hayashi, T. J. Am. Chem. Soc. 2017, 139, 17265-17268.
Knight, A. M.; Kan, S. B. J.; Lewis, R. D.; Brandenberg, O. F.; Chen, K.; Arnold, F. H.
ACS Central Sci. 2018, 4, 372-377.
10.
11.
12.
Moore, E. J.; Steck, V.; Bajaj, P.; Fasan, R. J. Org. Chem. 2018, 83, 7480-7490.
Chen, K.; Huang, X. Y.; Kan, S. B. J.; Zhang, R. K.; Arnold, F. H. Science 2018, 360,
71-+.
13.
14.
Brandenberg, O. F.; Prier, C. K.; Chen, K.; Knight, A. M.; Wu, Z.; Arnold, F. H. ACS
Catal. 2018, 8, 2629-2634.
Villarino, L.; Splan, K. E.; Reddem, E.; Alonso-Cotchico, L.; Gutiérrez de Souza, C.;
Lledós, A.; Maréchal, J.-D.; Thunnissen, A.-M. W. H.; Roelfes, G. Angew. Chem. Int.
Ed. 2018, 57, 7785-7789.
15.
McIntosh, J. A.; Coelho, P. S.; Farwell, C. C.; Wang, Z. J.; Lewis, J. C.; Brown, T. R.;
Arnold, F. H. Angew. Chem. Int. Ed. 2013, 52, 9309-9312.
Singh, R.; Bordeaux, M.; Fasan, R. ACS Catal. 2014, 4, 546-552.
Bordeaux, M.; Singh, R.; Fasan, R. Bioorg. Med. Chem. 2014, 22, 5697-5704.
Hyster, T. K.; Farwell, C. C.; Buller, A. R.; McIntosh, J. A.; Arnold, F. H. J. Am. Chem.
Soc. 2014, 136, 15505-15508.
16.
17.
18.
19

ACCEPTED MANUSCRIPT
19.
20.
Singh, R.; Kolev, J. N.; Sutera, P. A.; Fasan, R. ACS Catal. 2015, 5, 1685-1691.
Prier, C. K.; Zhang, R. J. K.; Buller, A. R.; Brinkmann-Chen, S.; Arnold, F. H. Nature
Chem. 2017, 9, 629-634.
21.
22.
23.
24.
25.
Tinoco, A.; Steck, V.; Tyagi, V.; Fasan, R. J. Am. Chem. Soc. 2017, 139, 5293-5296.
Chandgude, A. L.; Fasan, R. Angew. Chem. Int. Ed. 2018, 57, 15852-15856.
Brandenberg, O. F.; Fasan, R.; Arnold, F. H. Curr. Opin. Biotech. 2017, 47, 102-111.
Springer, B. A.; Sligar, S. G.; Olson, J. S.; Phillips, G. N. Chem. Rev. 1994, 94, 699-714.
Sreenilayam, G.; Moore, E. J.; Steck, V.; Fasan, R. Adv. Synth. Cat. 2017, 359, 2076–
2089.
26.
27.
28.
29.
Key, H. M.; Dydio, P.; Clark, D. S.; Hartwig, J. F. Nature 2016, 534, 534-537.
Wolf, M. W.; Vargas, D. A.; Lehnert, N. Inorg. Chem. 2017, 56, 5623-5635.
Sreenilayam, G.; Moore, E. J.; Steck, V.; Fasan, R. ACS Catal. 2017, 7, 7629-7633.
Chen, K.; Zhang, S. Q.; Brandenberg, O. F.; Hong, X.; Arnold, F. H. J. Am. Chem. Soc.
2018, 140, 16402-16407.
30.
31.
Onderko, E. L.; Silakov, A.; Yosca, T. H.; Green, M. T. Nat. Chem. 2017, 9, 623-628.
Pott, M.; Hayashi, T.; Mori, T.; Mittl, P. R. E.; Green, A. P.; Hilvert, D. J. Am. Chem.
Soc. 2018, 140, 1535-1543.
32.
Hayashi, T.; Tinzl, M.; Mori, T.; Krengel, U.; Proppe, J.; Soetbeer, J.; Klose, D.; Jeschke,
G.; Reiher, M.; Hilvert, D. Nature Catal. 2018, 1, 578-584.
Poulos, T. L. Chem. Rev. 2014, 114, 3919-3962.
Fita, I.; Rossmann, M. G. Proc. Natl. Acad. Sci. USA 1985, 82, 1604-1608.
Krebs, C.; Fujimori, D. G.; Walsh, C. T.; Bollinger, J. M. Accounts Chem. Res. 2007, 40,
484-492.
33.
34.
35.
36.
37.
White, M. D.; Flashman, E. Curr. Opin. Chem. Biol. 2016, 31, 126-135.
Santoro, S. W.; Wang, L.; Herberich, B.; King, D. S.; Schultz, P. G. Nature Biotechnol.
2002, 20, 1044-1048.
38.
Xiao, H.; Peters, F. B.; Yang, P. Y.; Reed, S.; Chittuluru, J. R.; Schultz, P. G. Acs
Chemical Biology 2014, 9, 1092-1096.
39.
40.
41.
42.
Joule, J. A.; Mills, K.; Smith, G. F. Heterocyclic Chemistry; Springer; 3rd edition 1995.
Wang, L.; Schultz, P. G. Angew. Chem. Int. Ed. 2004, 44, 34-66.
Dumas, A.; Lercher, L.; Spicer, C. D.; Davis, B. G. Chem. Sci. 2015, 6, 50-69.
Moore, E. J.; Zorine, D.; Hansen, W. A.; Khare, S. D.; Fasan, R. Proc. Natl. Acad. Sci.
USA 2017, 114, 12472-12477.
43.
44.
Anzenbacher, P.; Hudecek, J.; Struzinsky, R. FEBS Lett. 1982, 149, 208-210.
Prieto, G.; Suarez, M. J.; Gonzalez-Perez, A.; Ruso, J. M.; Sarmiento, F. Phys. Chem.
Chem. Phys. 2004, 6, 816-821.
45.
46.
Bickar, D.; Bonaventura, C.; Bonaventura, J. J. Biol. Chem. 1984, 259, 777-783.
Wolf, J. R.; Hamaker, C. G.; Djukic, J. P.; Kodadek, T.; Woo, L. K. J. Am. Chem. Soc.
1995, 117, 9194-9199.
47.
Lai, T. S.; Chan, F. Y.; So, P. K.; Ma, D. L.; Wong, K. Y.; Che, C. M. Dalton T. 2006,
4845-4851.
48.
49.
50.
51.
Salomon, R. G.; Kochi, J. K. J. Am. Chem. Soc. 1973, 95, 3300-3310.
Ruppel, J. V.; Kamble, R. M.; Zhang, X. P. Org. Lett. 2007, 9, 4889-4892.
Young, T. S.; Ahmad, I.; Yin, J. A.; Schultz, P. G. J. Mol. Biol. 2010, 395, 361-74.
Frost, J. R.; Jacob, N. T.; Papa, L. J.; Owens, A. E.; Fasan, R. ACS Chem. Biol. 2015, 10,
1805-16.
20

ACCEPTED MANUSCRIPT
52.
Berry, E. A.; Trumpower, B. L. Analyt. Biochem. 1987, 161, 1-15.
Graphical Abstract
21