1
Supporting Information for
Diverse engineered heme proteins enable stereodivergent cyclopropanation of
unactivated alkenes
Anders M. Knight
1
, S. B. Jennifer Kan
2
, Russell D. Lewis
1
, Oliver F. Brandenberg
2
, Kai Chen
2
,
Frances H. Arnold
1,2
*
1
Division of Biology and Bioengineering and
2
Division of Chemistry and Chemical Engineering,
California Institute of Technology,
1200 East California Boulevard, MC 210
-
41, Pasadena, CA 91125, United States
*Corresponding author: frances@cheme.caltech.edu
Table of contents
:
Previous literature for
stereocomplementary enzymatic cyclopropanation and
enantioselective, intermolecular unactivated alkene cyclopropanation
..........
........
0
2
Safety statement.....................................................................................
............
0
3
Nomenclature for compound labeling
................................................................
0
4
Materials and
methods...........................................................................
..........
0
4
Protein engineering strategies
...........................................................................
13
Protein homology models
..................................................................................
15
Cyclopropanation a
ctivity and selectivity of
whole
-
cell
variants, purified
protein
s
,
and
cell
lysates...........................
..................................................................
..
1
6
Compound synthesis and
characterization
.......................................................
..
1
9
Compound chiral separation conditions and representative traces
.......................
2
8
Calibration curves for analytical
-
scale TTN determination
....................................
3
7
Sequences of primers and variants
..................
................................................
4
2
NMR characterization
of cyclopropane products..................................................
5
1
References...............................................................
.............
.............................
7
3
2
Previous literature for
stereocomplementary enzymatic cyclopropanation and
enantioselective, intermolecular unactivated alkene cyclopropanation
Supplemental Table 1a.
Literature precedent for stereocomplementary cyclo
propanation of
styrenyl alkenes via carbene transfer.
Reference numbers are given for the main text reference
numbers.
Trans
-
cyclopropane
Cis
-
cyclopropane
(
R
,
R
)
(
S
,
S
)
(
R
,
S
)
(
S
,
R
)
Ref 1
1
,
10.1126/science.1231434
Not
reported
Up to 98% de,
96%
e.e.
Up to 84% de,
97%
e.e.
Not
reported
Ref 1
3
,
doi/10.1002/cbic.201600528
Up to 86% de,
32%
e.e.
Up to 98% de,
97%
e.e.
Up to 78% de,
99%
e.e.
Up to
42% de
,
95%
e.e.
Ref 1
7
,
10.1002/anie.201608680
Up to 99.9% de,
95%
e.e.
Up to 99.9%
de, 99.9%
e.e.
Not
reported
Not
reported
Ref
34
,
10.1021/jacs.7b00768
Up to 99.5% de,
92%
e.e.
Up to 99.9%
de, 99.9%
e.e.
Not
reported
Not
reported
Supplemental Table
1b
. Current state
-
of
-
the
-
art methods for the enantioselective, intermolecular
cyclopropanation
of unactivated alkenes.
Because
iron
-
catalyzed
asymmetric
examples are not
known, achiral examples are listed for iron.
Entry
Metal
Catalyst
Ref.
Note
1
Fe
1
Iron porphyrin
-
catalyzed,
achiral
.
Two
examples involved
unactivated alkenes.
Fe(PFP)Cl
(0.02
-
0.05 mol%) catalyzed the cyclopropanation of 2
-
ethyl
-
1
-
butene (100 eq
.
) with EDA in 390 TON,
d.r.
not reported.
PFP =
meso
-
tetrakis(pentafluoropheny1)porphyrin
2
Fe
2
Iron porphyrin
-
catalyzed,
achiral
.
One
example involved
unactivated alkenes.
Fe(TPP)Cl (3 mol%) catalyzed the cyclopropanation of allyl
benzene with
in situ
generated trifluoromethyl diazomethane (1.5
eq.) in water with 3.3 TON,
d.r.
not reported.
TPP = 5,10,15,20
-
tetraphenyl
-
21
H
,23
H
-
porphine
3
Fe
3
Iron porphyrin
-
catalyzed,
ach
iral
.
One
example involved
unactivated alkenes.
This example is similar to entry
2
, but a different
in situ
method was
used to generate the trifluoromethyl diazomethane. No product was
detected.
4
Fe
4
Iron porphyrin
-
catalyzed,
achiral
.
Three
examples involved
unactivated alkenes.
3
Fe(TPP)Cl (10 mol%) catalyzed the cyclopropanation of aliphatic
alkenes using
in situ
generated difluoromethylcarbene (2 eq.) in 6
-
8
TONs, 7:1 to 13: 1
d.r.
(
trans
).
5
Rh
5
-
13
All Rh
-
carboxylate or Rh
-
carboxamidate
-
catalyzed (0.2
-
2 mol% Rh)
examples of unactivated alkene cyclopropanation reported are
trans
-
selective
.
The most efficient and selective examples are the cyclopropanation
of 1
-
hexene (72 TTN, >20:1
d.r.
(
trans
), 96%
e.e.
), and 1
-
octene
(70 TTN, >20:1
d.r.
(
trans
), 90 %
e.e.
)
.
7
Ir
14
Ir
-
salen (1 mol%) catalyzed the cyclopropanation of >10
unactivated alkenes with up to 93 TTN, 98:2
d.r.
(
cis
major), 99%
e.e.
.
8
Ir
15
Ir(Me)PPIX in engineered myoglobin (0.5 mol%) catalyzed the
cyclopropanation of 1
-
octene in 40 TON, 91:9 er, 40:1
d.r.
(
trans
major). 6 equivalents of EDA were added
via
syringe pump over 12
h.
9
Ir
16
Ir(Me)PPIX in engineered CYP119 enzymes catalyzed the
cyclopropanation of a variety of unactivated alkenes (terminal,
internal and 1,1
-
disubstituted) with up to 1300 TON and 99%
e.e.
. 3
equivalents of EDA were added
via
syringe pump over 3 h.
10
Ru
17, 1
8
Ru
-
(iminophosphoranyl)ferrocene (2 mol%) catalyzed the
cyclopropanation of 3 unactivated alkenes with up to 37 TON,
72:28
d.r.
(
cis
major), 98%
e.e.
.
11
Co
19, 20
All Co
-
catalyzed (1
-
5 mol% Co) examples of unactivated alkene
cyclopropanation reported are
cis
-
selective
and based on Co
-
porphyrins.
The most efficient and selective example is the cyclopropanation of
phenylbutene (90 TTN, >99:1
d.r.
(
cis
), 96%
e.e.
).
12
Cu
21, 22
All Cu
-
catalyzed (1 mol% Cu) examples of unactivated alkene
cyclopropanation reported are
trans
-
selective
.
The most efficient and selective example is the cyclopropanation of
1
-
octene (80 TTN, 93:7
d.r.
(
trans
), 90%
e.e.
).
Safety statement
No unexpected or unusually high safety hazards were encountered in these methods. While ethyl
diazoester (EDA) has well
-
defined usage and risks, use of more volatile and reactive diazo
compounds (e.g. diazoalkanes) should be performed with caution (for ex
ample procedures, see
reference
23
).
4
Nomenclature for compound labeling
Alkenyl substrates are named
2x
(where x is a
-
m). The corresponding cyclopropyl esters (from
reaction with ethyl diazoacetate
1
) are named
3x
.
S
ingle diastereomer compounds are named
cis
-
3x
or
trans
-
3x
.
The 1
-
octene cyclopropane products, whose absolute configuration is known
(see
Compound chiral separation conditions and representative traces
),
are labeled (
1
(
R
/
S
),
2
(
R
/
S
))
-
3
a
. Alkenyl substrates with functional groups protected (e.g. 7
-
octen
-
1
-
ol, 7
-
octen
-
1
-
oic
acid) are named
2xa
, and their corresponding cyclopropane products are named
3xa
.
These are
depicted in Supplemental Figure 1.
Supplemental
Figure
1
.
C
ompound nom
encl
ature used in this work.
Materials and Methods
Solvents and reagents were ordered from Sigma Aldrich, TCI, CombiBlocks, or Alfa Aesar and
used without further purification. GC
-
FID data were collected on a Shimadzu GC
-
17A, Agilent
6850 GC system, and Agilent 7820A GC system. GC
-
MS data were collected on
a Shimadzu
GCMS
-
QP2010 SE. Screening HLPC
-
UV data were taken on an Agilent 1200 series HPLC.
Normal
-
phase chiral HPLC data were taken on an Agilent 1100 series HPLC. NMR spectra were
recorded on a Bruker Prodigy 400 MHz instrument
or Varian 300 MHz instru
ment
with CDCl
3
as
solvent.
1
H NMR spectra were recorded at 400 MHz and
13
C NMR spectra were recorded at 100
5
MHz. Chemical shifts were normalized to the chloroform solvent’s protio impurity (
1
H NMR 7.26
ppm,
13
C NMR 77.16 ppm).
Optical rotation data were c
ollected on a JASCO P
-
2000 Polarimeter.
Proteins tested in enzyme discovery
Genes encoding e
leven
heme
-
binding proteins were ordered as codon
-
optimized gBlocks
(Integrated DNA Technologies, Coralville, Iowa)
and assembled into pET22b(+) with the pelB
lea
der sequence removed and a C
-
terminal 6xHis tag. As the putative distal ligand could interfere
with the initially low substrate binding affinities, the proteins were ordered with the distal
axial
ligands mutated to smaller, nonpolar residues
, found to be b
eneficial for vinylarene
cyclopropanation in myoglobin
24
.
Some gBlocks
(Supplemental Table 2)
were also ordered with
a mutation in a putative entrance tunnel residue previously found to enhance styrene
cyclopropanation activity in myoglobin
.
24
These heme p
roteins were first tested for
cyclopropanation activity using 1,7
-
octadiene and EDA as substrates (1,7
-
octadiene was chosen
for its higher effective concentration of terminal olefin). The most active and selective proteins
identified from these experiments
,
protoglobin from
Aeropyrum pernix
(
ApePgb
)
and
nitric oxide
dioxygenase from
Rhodothermus marinus
(
RmaNOD
)
, were subsequently tested as wild
-
type
proteins against 1
-
octene
2a
, a commonly used model substrate for unactivated alkene
cyclopropanation studies.
6
Supplemental Table 2.
Heme
-
binding proteins tested for unactivated alkene cyclopropanation
activity using 1,7
-
octadiene and EDA as substrates.
UniProt ID
Organism
Annota
tion
Mutation
(
s
)
from WT
Cyclopropane
product formation
detected
Q3IDI7
Pseudoalteromonas
haloplanktis
Putative hemoglobin
-
like
oxygen
-
binding protein
Y42V F69A
No
Q7CX73
Agrobacterium fabrum
Uncharacterized protein
Y26V F53A
No
Q5L1S0
Geobacillus
kaustophilus
Hypothetical conserved
protein
Y29V Q50A
No
Q9NPG2
Homo sapiens
Neuroglobin
F28V F61I
H64A
No
Q0PB48
Campylobacter jejuni
Truncated hemoglobin
none
No
B3DVC3
Methylacidiphilum
infernorum
Hemoglobin IV
H71V L93A
Yes
D0MGT2
Rhodothermus
marinus
Nitric oxide dioxygenase
Y32V Q52A
Yes
G7VHJ7
Pyrobaculum ferrireducens
Protoglobin
Y58V
Yes
Q9YFF4
Aeropyrum pernix
Protoglobin
Y60V
Yes
O66586
Aquifex aeolicus
Thermoglobin
Y29V Q50A
Yes
Screening p
reviously engineered cytochromes P411
A
composite plate of 36 P411
-
CIS proteins from lineages engineered for non
-
natural reactions
such as aziridination, sulfimidation, amination, and non
-
styrenyl, activated alkene
cyclopropanation
25
was screened for activity and selectivity in
3a
product forma
tion. Most
reactions showed trace activity and moderate diastereoselectivity, but the highest activities by a
large margin were found in the lineage engineered for
N
-
vinyl amide cyclopropanation (O. F.
Brandenberg et al., unpublished results). Screening th
is lineage for stereoselectivity showed that
P411
-
CIS L437F T438Q L75Y L181I, referred to in this study as P411
-
UA, had the highest activity
and enantioselectivity
. It was
therefore used as a starting point in unactivated alkene
cyclopropanation.
7
Homolo
gy model
s
ApePgb was modeled with the
Methanosarcina acetivorans
protoglobin (PDB ID: 3ZJL) and
RmaNOD was modeled with
Alcaligenes eutrophus
flavohemoglobin (PDB ID: 1CQX) using
SWISS
-
MODEL
.
26
A homology model of P411
-
UA was generated through side
-
chain m
utations
to the P411
-
CIS crystal structure (PDB ID: 4H23). Figures generated from protein homology
models and crystal structures were made with PyMOL (Schrödinger, Inc.).
Subcloning and transformation of genes
Genes for
Aeropyrum pernix
protoglobin (ApeP
gb) and
Rhodothermus marinus
nitric oxide
dioxygenase (RmaNOD) were ordered as codon
-
optimized gBlocks (Integrated DNA
Technologies, Coralville, Iowa). The gBlocks were amplified via polymerase chain reaction (PCR)
and the PCR products were gel extracted a
nd purified with Zymoclean Gel DNA Recovery Kit
(Zymo Research Corp, Irvine, CA). The PCR product was subcloned into pET22b(+) via Gibson
assembly
.
27
Gibson assembly products were transformed into electrocompetent
E. cloni
EXPRESS BL21(DE3) cells (Lucigen,
Middleton, WI) with a Gene Pulser Xcell (Bio
-
Rad,
Hercules, CA).
Aliquots of SOC medium
(750 μ
L) were added and the cells were incubated at
37°C and 230 rpm for 45 minutes before being plated on LB
-
ampicillin (100
μ
g mL
1
) agar plates.
Overnight cultures (5 mL LB
-
amp in culture tubes) were grown at 37°C and 230 rpm for 12
-
18
hours. Overnight cultures were used to inoculate flask cultures, prepare glycerol stocks, and
isolate plasmids. Plasmids were isolated with Qiagen Min
iprep kits and the genes were sequence
verified (T7 promoter / terminator sequencing primers, Laragen, Inc.).
Protein expression
Cultures of Hyperbroth (
HB,
AthenaES) with 100 μg mL
1
ampicillin in unbaffled Erlenmeyer flasks
were inoculated 1% (v/v) with
stationary
-
phase
overnight cultures and shaken in an In
n
ova 42
8
shaker at 230 rpm, 37
°
C. At OD
600
= 0.8, cultures were chilled on ice for 20 minutes. Protein
expression was induced wit
h 0.5 mM
isopropyl β
-
D
-
1
-
thiogalactopyranoside
(
IPTG
)
and heme
production was enhanced with supplementation of 1 mM 5
-
aminolevulinic acid
(ALA)
. The
cultures were shaken at 180 rpm and 22
°
C overnight (18
-
24 hours). Cells were pelleted via
centrifugation at
4000 g for 10 minutes at 4
°
C. The supernatant was decanted and the cells were
resuspended in M9
-
N buffer supplemented with 25 mM glucose.
Site
-
saturation library construction
Site
-
saturation mutagenesis was performed using the 22
-
codon method
.
28
Briefly,
oligonucleotides were ordered with NDT, VHG, and TGG codons in the coding strand at the amino
acid position to be saturated. A reverse primer complementary to all three forward primers was
also ordered. Two PCRs were performed for each library, t
he first containing a mixture of forward
primers (12:9:1 NDT:VHG:TGG) and a pET22b(+) internal reverse primer and the second
containing the complementary reverse primer and a pET22b(+) internal forward primer. The two
PCR products were gel
-
purified with Zy
moclean Gel DNA Recovery Kit (Zymo Research Corp,
Irvine, CA) and ligated together via Gibson assembly. The Gibson assembly product was
transformed into electrocompetent
E. cloni
EXPRESS BL21(DE3) cells (Lucigen, Middleton, WI).
Aliquots
of SOC medium
(750
μ
L) were added and the cells were incubated at 37°C for 45 minutes
before being plated on LB
-
ampicillin (100
μ
g mL
1
) agar plates.
Site
-
saturation library expression
Single colonies from the LB
-
ampicillin agar plates were picked using sterile toothpicks
and grown
in 300 μL LB
-
ampicillin in 2 mL 96 deep
-
well plates at 37°C, 250 rpm, 80% humidity overnight
(12
-
18 hours). Multi
-
channel pipettes were used to transfer 30 μL of starter culture into deep
-
well
plates containing 1 mL HB
-
amp per well. Glycerol stoc
ks of these plates were prepared in parallel
by adding starter culture (100 μL) and 50% (v/v) sterile glycerol (100 μL) to a 96
-
well microplate,
9
which was then stored at
80°C. The deep
-
well expression culture plate was incubated at 37°C,
250 rpm, 80% humi
dity for 2.5 hours. The plate was then chilled on ice for 30 minutes. The
cultures were induced with 0.5 mM IPTG and supplemented with 1 mM ALA to increase cellular
heme production. The plate was incubated at 22°C and 250 rpm overnight. The plate was
centr
ifuged at 4000×g for 10 minutes at 4°C.
Site
-
saturation library reactions and screening
The pellets in the
site
-
saturation library deep
-
well plates were resuspended i
n nitrogen
-
free M9
minimal medium
(47.7 mM Na
2
HPO
4
, 22.0 mM KH
2
PO
4
, 8.6 mM NaCl, 2.0 mM
MgSO
4
, and 0.1
mM CaCl
2
, abbreviated as
M9
-
N, 400 μL). In an anaerobic chamber, 50 μL reactant mixture in
ethanol (final concentrations in a 450 μL reaction: 20 mM EDA
1
, 20 mM 4
-
phenyl
-
1
-
butene
2b
)
were added to the reaction plate. The reaction plate was
covered with a
piercea
ble foil cover (USA
Scientific) and shaken at 500 rpm for 3 hours. To quench the reaction and extract the substrates,
400 μL of a mixture of acetonitrile (49 mL) and 3
M HCl (1 mL) was added to each well. The
reaction plate was shaken
for an additional 30 minutes, followed by centrifugation (4000×g, 10
minutes, 4°C). The supernatant was filtered through a 0.2 μm PTFE 96
-
well filter plate into a 96
-
well microplate (4000×g, 1 minute, RT). The microplate was sealed with a pierc
e
able foil
cover.
The wells were screened for activity and diastereoselectivity of
3b
formation via HPLC using a
Kromasil 100
-
5
-
C18 column, 4.6x50 mm with a 71%
acetonitrile
isocratic method (3 minutes). In
later screening with higher enzymatic activity, the separati
on of
cis
-
and
trans
-
isomers of
3b
was
improved with the use of an Eclipse XDB
-
C18 column, 5 μm pa
rticle size, 4.6x150 mm and a 6
-
minute 71%
acetonitrile
isocratic method. Wells with improved activity relative to the parent
protein were streaked out from
the glycerol stock onto LB
-
amp plates. A single colony was picked
and grown in 5 mL LB
-
amp overnight (230 rpm, 37°C). These overnight cultures were used in
flask protein expression and small
-
scale biocatalytic reactions to verify enhanced activity and/or
s
electivity relative to the parent sequence.