Cytochrome P450-catalyzed insertion of carbenoids into N-H bonds

Article abstract of DOI:10.1039/c3sc52535j

Expanding nature's catalytic repertoire to include reactions important in synthetic chemistry will open new opportunities for 'green' chemistry and biosynthesis. We demonstrate the first enzyme-catalyzed insertion of carbenoids into N-H bonds. This type of bond disconnection, which has no counterpart in nature, can be mediated by variants of the cytochrome P450 from Bacillus megaterium. The N-H insertion reaction takes place in water, provides the desired products in 26-83% yield, forms the single addition product exclusively, and does not require slow addition of the diazo component.

Full text of DOI:10.1039/c3sc52535j

Chemical
Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
Cytochrome P450-catalyzed insertion of
carbenoids into N–H bonds†‡
Cite this: Chem. Sci., 2014, 5, 598
*
Z. Jane Wang, Nicole E. Peck, Hans Renata and Frances H. Arnold
Expanding nature’s catalytic repertoire to include reactions important in synthetic chemistry will open new
opportunities for ‘green’ chemistry and biosynthesis. We demonstrate the first enzyme-catalyzed insertion
of carbenoids into N–H bonds. This type of bond disconnection, which has no counterpart in nature, can be
mediated by variants of the cytochrome P450 from Bacillus megaterium. The N–H insertion reaction takes
place in water, provides the desired products in 26–83% yield, forms the single addition product exclusively,
and does not require slow addition of the diazo component.
Received 9th September 2013
Accepted 1st November 2013
DOI: 10.1039/c3sc52535j
www.rsc.org/chemicalscience
Introduction
The insertion of carbenoids into amine N–H bonds is a powerful
method of forming C–N bonds.1 This type of reaction has been
used for synthesis of many bioactive molecules such as unnat-
ural amino acids, alkaloids, and N-heterocyclic compounds.2
For instance, N–H insertion by rhodium catalysts was used in
the synthesis of thienamycin, which is one of the most potent
antibiotics ever identied and can only be accessed in signi-
cant quantities via organic synthesis.3 Carbene insertion reac-
tions typically utilize diazo compounds, which readily generate
reactive carbenoid intermediates with loss of dinitrogen in the
presence of a catalyst.4 In recent years, a variety of catalysts
employing transition metal complexes,5 ureas6 or articial
metalloenzymes7 have been developed. However, no enzyme
or antibody has ever been reported to exhibit this mode of
activation.
Recently we showed that variants of the cytochrome P450
from Bacillus megaterium (P450-BM3) are excellent catalysts for
intermolecular cyclopropanation of styrenes using synthetic
diazo compounds in water.8 We hypothesize that this reaction
proceeds through an iron-carbenoid intermediate that forms at
the hemin prosthetic group when the protein is treated with
ethyl diazoacetate (EDA) under anaerobic conditions. The yields
and selectivities exhibited by the P450-BM3 variants are
comparable to the best cyclopropanation catalysts based on
rhodium.8 Given that metal-carbenoids react with other
electron-rich partners like amines,9 we hypothesized that P450-
BM3 may also be a competent catalyst for carbenoid N–H
Fig. 1 Proposed carbenoid N–H insertion reaction by cytochrome
P450.
insertion (Fig. 1). N–H insertion with metal porphyrins has been
reported in organic solvents, but these reactions oen form a
mixture of the single and double N–H insertion products when
primary amines are used.10 An enzyme, on the other hand,
should be able to sterically select for one product compatible
with the binding pocket.
Although we have postulated that P450-BM3 forms iron-
carbenoids in the presence of diazo compounds,8 this carbe-
noid has not been characterized or observed directly. Thus, the
ability of engineered P450s to catalyze classical carbenoid
insertions provides additional evidence that an iron-carbenoid
intermediate is formed at the heme prosthetic group.
Results and discussion
We combined aniline with EDA in the presence of a reductant
(Na2S2O4) under argon atmosphere and tested the mixture with
seven P450-BM3 variants previously identied as competent
catalysts for cyclopropanation.8a In choosing this set of P450s,
we hypothesized that cyclopropanation activity would correlate
with ability to generate the iron-carbenoid intermediate that is
also necessary for N–H insertion. Whereas wildtype BM3 (WT)
provided only trace amounts of the desired product (1, Table 1,
entry 1), a few variants gave 1 in good yields aer 12 h at room
temperature. In particular, variant H2-5-F10, which is derived
from a thermostable P450-BM3 lineage11 and differs from WT by
Division of Chemistry and Chemical Engineering, California Institute of Technology,
1200 E. California Blvd., MC-210-41, Pasadena, CA 91125. E-mail: frances@cheme.
caltech.edu
† A provisional patent based on the results presented here has been led through
Caltech.
‡ Experimental procedures for protein catalyzed reactions. See DOI:
10.1039/c3sc52535j
598 | Chem. Sci., 2014, 5, 598–601
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
Table 1 N–H insertion using variants of cytochrome P450-BM3
reaction mixture altogether resulted in no product formation
(ESI, Table S.1‡). Furthermore, only 2% conversion was
observed when we switched from Na2S2O4 to a biological
reductant like NADPH. This is likely due to inefficient reduction
of Fe(III) to the active Fe(II) catalyst for carbene transfer, as the
redox potential of NAD+/NADH (Eꢀ0 ¼ ꢁ320 mV; all potentials
versus standard hydrogen electrode) is less negative than that of
low spin Fe(III)/Fe(II) in P450-BM3 (Eꢀ0 ¼ ꢁ430 mV). Because
P450s exist in both the ferric and ferrous forms in intact
Escherichia coli cells,12 we posited that catalysis could also be
achieved using intact cells. Addition of EDA and aniline to cells
expressing H2-5-F10 provided 1 in 38% yield (see ESI,‡ section II
for details).
Under the current optimized reaction conditions, complete
conversion to the N–H insertion product was never observed.
We reasoned that incomplete conversion could be due to
depletion of EDA in competing side reactions, inhibition of the
enzyme by 1, or deactivation of the enzyme over time. When an
equimolar ratio of EDA and aniline was used, unreacted aniline
could be detected by GC at the end of reaction but dimeric side
products like diethyl fumarate or maleate were not observed.
When we combined EDA and aniline in a 1 : 2 ratio, then added
a second equivalent of EDA aer 2 h, no further 1 was produced.
This indicates that the moderate yield of 1 is not due to
consumption of the limiting reagent, EDA, by competing side
reactions. To test for product inhibition, 1.0 equivalents of ethyl
2-(p-tolylamino) acetate (4), the N–H insertion product from the
reaction of EDA with toluidine, was added to a mixture of
enzyme, reductant, aniline and EDA. The yield of 1 remained
unchanged (ESI, Table S.2‡). These experiments suggest that
deactivation of the P450 catalyst may be responsible for the
moderate conversion to 1. Efforts to elucidate the mechanism of
this deactivation are currently ongoing.
We examined a variety of substrates for N–H insertion using
P450 variant H2-5-F10 and found that this catalyst is fairly
general and can catalyze N–H insertion with both primary and
secondary anilines (Fig. 2). Substitution is well-tolerated on the
aniline partner on both the ortho and para positions. Reaction
of aniline and EDA was also performed in milligram scale
without any organic cosolvent to provide 1 in 65% isolated yield.
On the other hand, alkyl amines like morpholine, cyclohexyl
amine and benzyl amine were poor substrates for this enzyme,
which gave <3% conversion to the insertion products (ESI,
Table S.3‡).
Diazo amides can be used as the carbene precursor for this
reaction, and product 8 can be formed in moderate yields.
Products such as 8 are valuable because they are precursors to
important synthetic building blocks such as amino amides and
1,2-diamines. Only one transition metal-based method has ever
been described for N–H insertion with diazo amides; TTNs of
16–32 were reported.13 In comparison, P450 variant H2-5-F10
catalyzes the insertion reaction with more than 150 TTN.
As P450-BM3s are competent catalysts for both N–H inser-
tion and cyclopropanation, we performed competition experi-
ments with styrene and aniline to determine the relative rates of
these two reactions. When 1.0 equivalent of EDA was combined
with 1.0 equivalent of aniline and 1.0 equivalent of styrene, the
Entry
Catalysta,b
Conditions
% yieldc
TTN
1
2
3
4
5
6
7
8
WT
T268A
Anaerobicd
Anaerobic
Anaerobic
Anaerobic
Anaerobic
Anaerobic
Anaerobic
Anaerobic
Anaerobic
Aerobic
1.7
16
43
34
24
47
14
9.5
0
17
160
433
340
238
473
136
95
BM3-CIS
H2-4-D4
H2-A-10
H2-5-F10
P411-CIS
P411-T268A
H2-5-F10 + CO
H2-5-F10
H2-5-F10e
9
10
11
0
20
291
2
68
Anaerobic
a
Reactions were carried out with protein (10 mM), EDA (8.5 mM), aniline
(20 mM) and Na2S2O4 (10 mM) in phosphate buffer (pH 8) with shaking
b
at room temperature for 12 h. See ESI for specic amino acid
substitutions from WT for each mutant. Yields were determined by
c
d
GC calibrated for 1. Anaerobic reactions performed under Ar
atmosphere. e Reaction performed with 20 mM protein.
15 mutations, formed 1 in 47% yield and 473 turnovers (TTN)
using 0.1% protein relative to EDA (entry 6).
The Cys at position 400 in P450-BM3 coordinates axially to
the iron-heme. Interestingly, variants possessing a mutation of
Cys at position 400 to Ser (P411-CIS and P411-T268A, entries 7
and 8) were less active than their Cys analogues (BM3-CIS and
T268A, entries 3 and 2). This contrasts with the much greater
activity of the Ser-ligated P411 enzymes observed in styrene
cyclopropanation.8b The Cys to Ser substitution does not result
in substantial geometric changes at the active site.8b It is
possible that a less electron-rich ligand in the axial position of
the heme prosthetic group lowers the activity of the catalyst in
N–H insertion reactions.
When CO was bubbled gently through the protein solution
before the addition of EDA and aniline, no product formation
was observed (Table 1, entry 9), presumably due to complexa-
tion of CO to the iron center. Performing the reaction under
aerobic conditions signicantly reduced the yield of 1, sug-
gesting that O2 also inhibits the transformation (Table 1, entry
10). Additionally, variants BM3-CIS, H2-4-D4, and H2-5-F10
(entries 3, 4, and 6, see ESI Table S.5‡ for amino acid
sequences) differ by only 1–2 active site amino acids from
variant H2-A-10 (entry 5), yet exhibit a range of activity (24–47%
yield). This demonstrates that changes in the protein sequence,
and presumably the geometry around the active site, can
modulate N–H insertion activity.
We have previously observed that reductant is not consumed
stoichiometrically in carbene transfer reactions with P450-BM3
and is only needed to reduce the Fe(III) resting state to the Fe(II)
active catalyst.7a Accordingly, decreasing the amount of Na2S2O4
from 10 mM to 2.5 mM did not signicantly affect the efficiency
of the N–H insertion reaction, but omitting Na2S2O4 from the
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 598–601 | 599

View Article Online
Chemical Science
Edge Article
Fig. 4 Product ratio of N–H insertion catalyzed by hemin versus P450
variant H2-5-F10.
and double insertion products were produced in 2.5 : 1 to 3.5 : 1
ratios (Fig. 4). This product distribution is consistent with what
has been observed with other metalloporphyrins such as iron
tetraphenylporphyrin (FeTPP).10 The ability of P450 variants to
make the single insertion product exclusively agrees with our
hypothesis that the protein binding pocket can control reaction
selectivity and constitutes a major advantage of protein-
mediated catalysis over catalysis by isolated metalloporphyrins.
Fig. 2 Substrate scope for cytochrome P450-catalyzed N–H inser-
tion. a Reactions were carried out with protein (20 mM), EDA (8.5 mM),
aniline (20 mM) and Na2S2O4 (10 mM) in phosphate buffer (pH 8) and
allowed to shake at room temperature for 12 h. Yields are reported in
Conclusions
b
parentheses and TTN are presented in Table S.1 in ESI.‡ Yields were
c
Enzymes offer an advantage over transition metal catalysts
because they operate under mild conditions in aqueous solvent,
are highly active, and can be highly selective for one product
when multiple are possible.14 By screening variants of P450-BM3
for activity on unnatural substrates, we have uncovered a new
mode of enzyme-catalyzed amine synthesis and report one of
the rst examples of N–H insertion by metal carbenoids cata-
lyzed in water. Variants of P450-BM3 reacted readily with diazo
compounds and amines to give a-amino esters and amides.
These reactions proceeded to good yields without the need for
slow addition of the diazo reagent. In contrast to N–H insertion
mediated by other metalloporphyrins, no double insertion
products were observed in the P450-catalyzed reaction, even
when equal molar ratios of amine and diazo compounds were
used. Our results support the hypothesis that an iron-carbenoid
is formed when P450-BM3 reacts with diazo compounds and
suggest that many other insertion reactions will also be cata-
lyzed by P450s.
determined by GC calibrated for each product. Isolated yield based
on milligram scale reaction: 25 mM EDA, 25 mM aniline, 25 mM protein,
and 25 mM Na2S2O4. d Reactions carried out with protein (20 mM), EDA
(10.0 mM), aniline (20 mM) and Na2S2O4 (10 mM).
Fig. 3 (a) Competition reaction of styrene and aniline for cyclo-
propanation and N–H insertion, respectively, with EDA. (b) Reaction of
bifunctional substrate 9 with EDA.
N–H insertion and cyclopropanation products were formed in
63% and 9% yield, respectively (Fig. 3a). This suggests that N–H
insertion reaction with H2-5-F10 is signicantly faster than the
cyclopropanation reaction (ESI, Table S.4‡). When EDA is
combined with substrate 9 containing both amine and olen
functionalities, the single N–H insertion product 10 is formed
in 32% yield, with only trace amounts of the single cyclo-
propanation product (Fig. 3b). This chemoselectivity for N–H
insertion is consistent with our observation that N–H insertion
is faster than cyclopropanation.
Acknowledgements
This research is funded by the Gordon and Betty Moore Foun-
dation through Grant GBMF2809 to the Caltech Programmable
Molecular Technology Initiative. Z. J. Wang was supported by a
Ruth M. Kirschstein Fellowship from the National Institute of
Health, award number F32EB015846-01. The authors thank Dr
P. S. Coelho, Dr J. McIntosh, and Mr C. Farwell for useful
discussions.
The enzyme is also selective. No double insertion products
were ever observed in the enzyme-catalyzed N–H insertion
reactions, as determined by GC-MS and 1H NMR of the products
in milligram-scale reactions. In contrast, when the isolated
hemin prosthetic group alone was used as catalyst, the single
Notes and references
1 (a) P. J. Yates, J. Am. Chem. Soc., 1952, 74, 5376; (b)
T. Saegusa, Y. Ito, S. Kobayashi, K. Hirota and T. Shimizu,
Tetrahedron Lett., 1966, 7, 6131; (c) M. P. Doyle,
600 | Chem. Sci., 2014, 5, 598–601
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
M. A. McKervey and T. Ye, Modern Catalytic Methods for
Chemical Science
B. Neelima and C. Reddy, J. Mol. Catal. A: Chem., 2006,
256, 269; (f) Z. Zhang and J. Wang, Tetrahedron, 2008, 64,
6577.
Organic Synthesis with Diazo Compounds, Wiley, New York,
1998, ch. 8.2.
2 (a) C. Bolm, A. Kasyan, K. Drauz, K. Gunther and G. Raabe,
Angew. Chem., Int. Ed., 2000, 39, 2288; (b) F. Davis, T. Fang
and R. Goswami, Org. Lett., 2002, 4, 1599; (c) K. Yamazaki
and Y. Kondo, Chem. Commun., 2002, 210; (d) K. E. Bashford,
A. L. Cooper, P. D. Kane, C. J. Moody, S. Muthusamy and
E. Swann, J. Chem. Soc., Perkin Trans. 1, 2002, 1672; (e)
S.-H. Lee, B. Clapham, G. Koch, J. Zimmerman and
K. D. Janda, J. Comb. Chem., 2003, 5, 188; (f)
A. C. B. Burtoloso and C. R. D. Correia, Tetrahedron Lett.,
2004, 45, 3355; (g) J. R. Davies, P. D. Kane and C. J. Moody,
Tetrahedron, 2004, 60, 3967; (h) H. Matsushita, S.-H. Lee,
K. Yoshida, B. Clapham, G. Koch, J. Zimmerman and
K. Janda, Org. Lett., 2004, 6, 4627; (i) I. Aviv and Z. Gross,
Chem. Commun., 2006, 4477; (j) B. Liu, S.-F. Zhu, W. Zhang,
C. Chen and Q.-L. Zhou, J. Am. Chem. Soc., 2007, 129, 5834.
3 (a) R. Ratcliffe, T. Salzmann and B. Christensen, Tetrahedron
Lett., 1980, 21, 31; (b) T. N. Salzmann, R. W. Ratcliffe,
B. G. Christensen and F. A. Bouffard, J. Am. Chem. Soc.,
1980, 102, 6161.
6 S. S. So and A. E. Mattson, J. Am. Chem. Soc., 2012, 134, 8798.
7 (a) Z. T. Ball, Acc. Chem. Res., 2013, 46, 560–570; (b)
D. Gillingham and N. Fei, Chem. Soc. Rev., 2013, 42, 4918.
8 For P450-catalyzed cyclopropanation, see (a) P. Coelho,
E. Brustad, A. Kannan and F. H. Arnold, Science, 2013, 339,
307; (b) P. Coelho, Z. J. Wang, M. Energ, S. A. Baril,
A. Kannan, F. H. Arnold and E. Brustad, Nat. Chem. Biol.,
2013, 9, 485; (c) As a comparison for Rh cyclopropanation,
see: H. M. L. Davies and C. Venkataramani, Org. Lett.,
2003, 5, 1403.
9 (a) T. Ye and M. A. McKervey, Chem. Rev., 1994, 94, 1091; (b)
C. J. Moody, Angew. Chem., Int. Ed., 2007, 46, 9148; (c)
S.-F. Zhu and Q.-L. Zhou, Acc. Chem. Res., 2012, 45, 1365,
and references therein.
10 (a) L. K. Baumann, H. M. Mbuvi, G. Du and L. K. Woo,
Organometallics, 2007, 26, 3995; (b) H. M. Mbuvi,
E. R. Klobuskowski, G. Roberts and L. K. Woo,
J. Porphyrins Phthalocyanines, 2010, 14, 284, also see ref. 4b.
11 J. C. Lewis, S. M. Mantovani, Y. Fu, C. D. Snow, R. S. Komor,
C. H. Wong and F. H. Arnold, ChemBioChem, 2010, 11, 2502.
4 (a) S.-F. Zhu, B. Xu, G.-P. Wang and Q.-L. Zhou, J. Am. Chem.
Soc., 2012, 134, 436; (b) B. J. Anding and L. K. Woo, 12 W. A. Johnston, D. J. B. Hunter, C. J. Noble, G. R. Hanson,
Organometallics, 2013, 32, 2599–2607.
J. E. Stok, M. A. Hayes, J. J. De Voss and E. M. J. Gillam,
5 For selected recent examples see: (a) S. Bachmann,
J. Biol. Chem., 2011, 286, 40750.
D. Fielenbach and K. A. Jorgensen, Org. Biomol. Chem., 13 S. Chanthamath, S. Thongjareun, K. Shibatomi and S. Iwasa,
2004, 2, 3044; (b) E. C. Lee and G. C. Fu, J. Am. Chem. Soc., Tetrahedron Lett., 2012, 53, 4862.
2007, 129, 12066; (c) M. E. Morilla, M. M. Dıaz-Requejo, 14 (a) K. Faber, Biotransformation in Organic Chemistry,
´
´
T. R. Belderrain, M. C. Nicasio, S. Tromenko and P. Perez,
Springer, New York, 2004; (b) J. D. Rozzell, Bioorg. Med.
Chem., 1999, 7, 2253; (c) A. J. J. Straathof, S. Panke and
A. Schmid, Curr. Opin. Biotechnol., 2002, 13, 548.
Chem. Commun., 2002, 2998; (d) M. Moyer, P. Feldman and
H. Rapoport, J. Org. Chem., 1985, 50, 5223; (e) M. Kantam,
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 598–601 | 601