C(sp3)-H bond hydroxylation catalyzed by myoglobin reconstituted with manganese porphycene

Article abstract of DOI:10.1021/ja409404k

Myoglobin reconstituted with manganese porphycene was prepared in an effort to generate a new biocatalyst and was characterized by spectroscopic techniques. The X-ray crystal structure of the reconstituted protein reveals that the artificial cofactor is located in the intrinsic heme-binding site with weak ligation by His93. Interestingly, the reconstituted protein catalyzes the H₂O₂-dependent hydroxylation of ethylbenzene to yield 1-phenylethanol as a single product with a turnover number of 13 at 25 C and pH 8.5. Native myoglobin and other modified myoglobins do not catalyze C-H hydroxylation of alkanes. Isotope effect experiments yield KIE values of 2.4 and 6.1 for ethylbenzene and toluene, respectively. Kinetic data, log k _obs versus BDE(C(sp³)-H) for ethylbenzene, toluene, and cyclohexane, indicate a linear relationship with a negative slope. These findings clearly indicate that the reaction occurs via a rate-determining step that involves hydrogen-atom abstraction by a Mn(O) species and a subsequent rebound hydroxylation process which is similar to the reaction mechanism of cytochrome P450.

Full text of DOI:10.1021/ja409404k

Communication
pubs.acs.org/JACS
C(sp3)−H Bond Hydroxylation Catalyzed by Myoglobin Reconstituted
with Manganese Porphycene
Koji Oohora, Yushi Kihira, Eiichi Mizohata, Tsuyoshi Inoue, and Takashi Hayashi*
Department of Applied Chemistry, Graduate School of Engineering, Osaka University, Suita, 565-0871, Japan
S
* Supporting Information
porphyrinoids, whereas the number of C−H bond activations
ABSTRACT: Myoglobin reconstituted with manganese
porphycene was prepared in an effort to generate a new
biocatalyst and was characterized by spectroscopic
techniques. The X-ray crystal structure of the reconstituted
protein reveals that the artificial cofactor is located in the
intrinsic heme-binding site with weak ligation by His93.
Interestingly, the reconstituted protein catalyzes the H2O2-
dependent hydroxylation of ethylbenzene to yield 1-
phenylethanol as a single product with a turnover number
of 13 at 25 °C and pH 8.5. Native myoglobin and other
modified myoglobins do not catalyze C−H hydroxylation
of alkanes. Isotope effect experiments yield KIE values of
2.4 and 6.1 for ethylbenzene and toluene, respectively.
Kinetic data, log kobs versus BDE(C(sp3)−H) for ethyl-
benzene, toluene, and cyclohexane, indicate a linear
relationship with a negative slope. These findings clearly
indicate that the reaction occurs via a rate-determining
step that involves hydrogen-atom abstraction by a Mn(O)
species and a subsequent rebound hydroxylation process
which is similar to the reaction mechanism of cytochrome
P450.
for a simple alkane with a large bond dissociation energy
(BDE) value for the C(sp3)−H bond is quite limited. However,
there have been no examples of clear enhancement of the
enzymatic activity of hemoproteins as a result of manganese
porphyrinoid substitution.7 Here, we report the first example of
C(sp3)−H bond hydroxylation catalyzed by myoglobin
reconstituted with manganese porphycene as shown in Figure
1.
yoglobin, an oxygen storage protein, has a heme b
Mcofactor which is also used by heme-dependent enzymes
such as horseradish peroxidase and cytochrome P450s.1
However, myoglobin has very weak peroxidase and mono-
oxygenase activities and does not catalyze alkane hydroxylation
via a C−H bond activation upon the addition of H2O2 or
peracids as oxygen donors. Therefore, molecular engineering of
myoglobin for enhancement of enzymatic activities is a
challenging study. Advances in this area are expected to lead
to generation of new biocatalysts and to improve our
understanding of the catalytic mechanisms of native enzymes.
Methods used to modify hemoproteins include site-directed
mutagenesis of amino acid residues in the heme pocket2 and
replacement of native heme with an artificial cofactor.3 We have
recently enhanced the peroxidase activity of myoglobin by the
latter method using artificial cofactors including an iron
porphyrinoid4 and a propionate-modified heme.5 Furthermore,
a hemoprotein reconstituted with a manganese porphyrinoid is
expected to provide a good candidate for modulating the
function of myoglobin, because manganese porphyrins and
porphyrinoids such as manganese corrole and corrolazine have
been studied as catalysts of C−H bond activation.6 For
example, 9,10-dihydroanthracene, cyclohexane-1,4-diene, and
their derivatives are often employed as substrates for MnV(O)
Figure 1. (a) Molecular structures of the metal porphycenes and
porphyrins. (b) Schematic representation for the preparation of
rMb(MnPc).
Porphycene, a constitutional porphyrin isomer, has remark-
ably different physicochemical properties than those of
porphyrin.8 We inserted a manganese ion into the previously
reported porphycene (Pc) macrocycle and hydrolyzed the
propionate diester moieties to yield the manganese(III)
complex, MnPc. This cofactor was then incorporated into the
apomyoglobin obtained by removal of heme from horse heart
myoglobin. The reconstituted protein, rMb(MnPc), was
purified and characterized by ESI-MS and UV−vis spectro-
scopic methods. The detected mass number was assigned as a
multiply ionized species of rMb(MnPc) (found m/z = 1953.4,
calculated m/z (z = +9) = 1953.3). This indicates that MnPc
makes a stable complex with the apoprotein. The UV−vis
spectrum showed the red shift in the Q-region (λmax = 568 and
Received: September 11, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja409404k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
631 nm) due to imidazole ligation of histidine compared to free
MnPc (Figure S1).
Figure 3 shows that rMb(MnPc) promotes hydroxylation and
provides only 1-phenylethanol as a product with a total
The crystal structure of rMb(MnPc) was further obtained at
2.2 Å resolution and revealed His93-ligation to Mn with a bond
length of 2.58 Å in the heme pocket (Figure 2a). The
Figure 3. Time course plots of catalytically produced 1-phenylethanol
by Mbs and MnPc. Conditions: [protein or MnPc] = 20 μM, [H2O2]
= 10 mM, [ethylbenzene] = 8 mM, in 100 mM potassium phosphate
buffer, pH 8.5, containing 8% CH3CN at 25 °C. In the case of MnPc,
DMSO (1% v/v) was added to the solution for complete dissolution
of the complex.
turnover number (TON) of 13 and initial turnover frequency
(TOF) of 33 h−1.12 In addition, no byproducts (such as phenol
or ketone derivatives) were detected in this reaction.
Furthermore, the protein matrix plays an important role in
the hydroxylation because free MnPc does not produce any
products under the same conditions. It is noteworthy that other
myoglobins do not have ethylbenzene hydroxylation activity.
Table 1 summarizes H2O2-dependent hydroxylation and
Figure 2. (a) Stereoview of the crystal structure of rMb(MnPc)
around MnPc (this study, PDB ID: 3WI8). The bonds to Mn are not
shown for clarity. The distal His64 residue has two conformers. (b)
Superimposition of rMb(MnPc) (green) and rMb(MnPor) (gray,
PDB ID: 2O58).
Table 1. Hydroxylation and Sulfoxidation Activities of
Myoglobins
a
b
c
ethylbenzene
thioanisole
d
e
d
f
substrate
protein
TON
TOF (h−1
)
TON
TOF (h−1
)
interaction between two propionate side chains and surround-
ing amino acid residues was different from that of native
myoglobin (nMb(FePor)) and myoglobin reconstituted with a
protoporphyrin IX manganese complex (rMb(MnPor)), but
generally similar to that of myoglobin reconstituted with iron
porphycene, rMb(FePc) (Figure S2).9 The polypeptide
structures of rMb(MnPc) and rMb(MnPor)10 are super-
imposable with an RMSD (root-mean-square deviation) value
of 0.37 Å for Cα atoms. It is apparent that helix F moves to the
opposite side of the cofactor. The other helices are positioned
similar to those of nMb(FePor) (Figure 2b). The coordination
bond between Mn and His93 in rMb(MnPc) is approximately
0.36 Å longer than that of rMb(MnPor) (Table S2), and the
temperature factors (B-factors) of the helix F and MnPc are
notably higher (Figure S3). These results suggest that MnPc is
relatively loosely bound to apoMb.11 On the distal side of the
heme pocket, the His64 residue clearly adopts two different
conformations. The occupancy of the coordinated water
molecule to Mn is small (roughly 0.5), and the O−Mn bond
of rMb(MnPc) is longer with a distance of 2.79 Å. According to
this crystal structure, the presence of MnPc in the heme pocket
will provide both six- and five-coordinate species.
rMb(MnPc)
rMb(FePc)
13
0
33
0
61
157
0
31
142
0
rMb(MnPor)
nMb(FePor)
0
0
0
0
89
61
a
Reactions were monitored in 100 mM potassium phosphate buffer at
b
25 °C. Conditions: [protein] = 20 μM, [H2O2] = 10 mM,
[ethylbenzene] = 8.0 mM in the buffer, pH 8.5, containing 8%
c
CH3CN. Conditions: [protein] = 8 μM, [H2O2] = 10 mM,
[thioanisole] = 2.0 mM in the buffer, pH 7.0, containing 1%
d
CH3CN. Turnover number after completion of the reaction.
e
f
Turnover frequency at 5 min after starting the reaction. Turnover
frequency at 30 min after starting the reaction. TON or TOF = 0
indicates that the product was not detected by GC analysis.
sulfoxidation promoted by myoglobins. Interestingly, the
activity of rMb(MnPc) toward thioanisole sulfoxidation is
much lower than that of nMb(FePor), whereas the TON and
TOF values of rMb(FePc) for the sulfoxidation are significantly
higher than those for other myoglobins. These results indicate
that high oxidation potential is not necessary for alkane
hydroxylation13 and appears rather to provide a disadvantage in
the case of a myoglobin-based system.14 On the other hand,
rMb(MnPor) has no hydroxylation or sulfoxidation activity due
to poor reactivity with H2O2.7a For rMb(MnPc), the formation
Myoglobin-catalyzed hydroxylation of ethylbenzene was
carried out upon addition of H2O2 at pH 8.5 and 25 °C.
B
dx.doi.org/10.1021/ja409404k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
of the compound I-like species, a two electron-oxidized
MnV(O) species,6 was preliminarily identified in a reaction
with cumene hydroperoxide, which is a useful reagent to
evaluate the O−O bond cleavage in peroxide. The ratio of two
products, cumyl alcohol and acetophenone, which are derived
from heterolytically and homolytically peroxide-cleaved prod-
ucts, respectively, is determined to be 7.0 at pH 7.0. This value
is 2.5-fold larger than that of nMb(FePor), and generally similar
to the value obtained for rMb(FePc).4a,15 Given this finding, it
appears that the addition of H2O2 to rMb(MnPc) predom-
inantly generates a compound I-like intermediate in the
catalytic cycle.
Next, to determine whether the mechanism promoted by
rMb(MnPc) includes the C−H activation, deuterated ethyl-
benzene, C8D10, was employed as a substrate and the initial rate
constants for the hydroxylation of C8H10 and C8D10, kH and kD,
respectively, were evaluated. The obtained kinetic isotope
effect, KIE (kH/kD), determined at 25 °C was 2.4 (Figure S4).
This value is lower than the theoretical value (ca. 7), but the
difference in kinetics between the normal substrate and the
deuterated substrate is clearly observed.16 Furthermore, the
initial rates of the hydroxylations of ethylbenzene, toluene, and
cyclohexane were logarithmically plotted against BDE(C(sp3)−
H) for each substrate, and a linear relationship was identified
with a negative slope as shown in Figure 4a.17,18 These findings
transition state for C−H activation partially includes the
electron transfer process.19 Moreover, the isotope-labeling
study using H218O2 yields the 18O-labeled alcohol quantita-
tively. This effectively rules out the oxygen atom transfer from a
water molecule or dissolved O2 and indicates that the OH-
rebound mechanism occurs after the H-abstraction from the
substrate alkane C−H bond by the MnV(O) species of
rMb(MnPc). Taken together, all of these findings demonstrate
that the catalytic hydroxylation reaction proceeds through the
cytochrome P450-like C−H bond activation process as follows:
MnV(O)Pc + R−H → MnIV(OH)Pc + R· → MnIIIPc + R−
OH.20
In conclusion, myoglobin reconstituted with manganese
porphycene catalyzes H2O2-dependent alkane hydroxylation in
the heme pocket. This is remarkable because native myoglobin
does not have hydroxylation activity. One of the reasons
proposed for the lack of such activity is that myoglobin has no
definite substrate-binding site such as the binding sites present
in the P450s. Watanabe and co-workers have reported the
hydroxylation of Trp43 which is close to heme within the heme
pocket of the F43W/H64L myoglobin mutant.21 In contrast,
our work presented here first demonstrates that the catalytic
activity of myoglobin toward substrate hydroxylation can be
dramatically improved only by the insertion of MnPc into
apomyoglobin. This finding obviously indicates that a MnV(O)
porphycene complex is a unique intermediate capable of
promoting C(sp3)−H bond activation followed by H-
abstraction and OH-rebound to give a corresponding alcohol.
The myoglobin protein matrix is essential for this reaction, and
it is of interest that the C−H bond activation mechanism is
similar to that of cytochrome P450s. Further mechanistic study
and optimization of the catalytic reaction conditions are now in
progress.
ASSOCIATED CONTENT
* Supporting Information
■
S
Synthetic procedures, experimental details, and characterization
data. This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
thayashi@chem.eng.osaka-u.ac.jp
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by Grants-in-Aid for Scientific
Research and Innovative Areas (specifically, the “Molecular
Activation Area”) provided by JSPS and MEXT. The authors
are grateful to the staff for their excellent support during data
collection on the BL44XU at SPring-8.
Figure 4. (a) Plots of log kobs for the catalytic hydroxylation of
ethylbenzene, toluene, and cyclohexane against each bond dissociation
energy value (BDE).18 (b) Plots of log kobs for the catalytic
hydroxylation of para-substituted toluene against the Hammett
constant, σp.19 The values for kobs, initial rate, were obtained as TOF
in the reaction at 5 min.
REFERENCES
■
(1) Phillips, G. N., Jr. In Handbook of Metalloproteins; Messerschmidt,
A., Huber, R., Poulos, T., Wieghardt, K., Eds.; John Wiley & Sons:
Chichester, U.K., 2001; Vol. 1, pp 5−15.
(2) (a) Watanabe, Y.; Nakajima, H.; Ueno, T. Acc. Chem. Res. 2007,
40, 554−562. (b) Yeung, N.; Lin, Y.; Gao, Y.; Zhao, X.; Russell, B.;
Lei, L.; Miner, K.; Robinson, H.; Lu, Y. Nature 2009, 462, 1079−1082.
(3) (a) Hayashi, T. In Handbook of Porphyrin Science; Kadish, K. M.,
Smith, K., Guilard, R., Eds.; World Scientific; Singapore, 2010; Vol. 5,
pp 1−69. (b) Hayashi, T. In Coordination Chemistry in Protein Cages:
provide support for our proposal that hydrogen-atom
abstraction from the alkane C−H bond by the compound I-
like species of rMb(MnPc) is the rate-determining step. In
addition, the substituent effect of toluene derivatives is
evaluated using para-methoxy-, methyl-, iodo-, bromo-, and
chloro-toluenes. The Hammett plot of σp values for each
substrate shows the linear relationship with the ρ value of
−1.65 (Figure 4b). The negative value implies that the
C
dx.doi.org/10.1021/ja409404k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Principles, Design, and Applications; Ueno, T., Watanabe, Y., Eds.; John
Wiley & Sons: Hoboken, NJ, 2013; pp 87−110.
D.; Ohki, T.; Ueno, T.; Hara, I.; Adachi, S.; Makino, Y.; Ueyama, N.;
Lu, Y.; Watanabe, Y. J. Biol. Chem. 2005, 280, 12858−12866.
(4) (a) Hayashi, T.; Murata, D.; Makino, M.; Sugimoto, H.; Matsuo,
T.; Sato, H.; Shiro, Y.; Hisaeda, Y. Inorg. Chem. 2006, 45, 10530−
10536. (b) Matsuo, T.; Murata, D.; Hisaeda, Y.; Hori, H.; Hayashi, T.
J. Am. Chem. Soc. 2007, 129, 12906−12907. (c) Matsuo, T.; Hayashi,
A.; Abe, M.; Matsuda, T.; Hisaeda, Y.; Hayashi, T. J. Am. Chem. Soc.
2009, 131, 15124−15125.
(5) (a) Hayashi, T.; Hisaeda, Y. Acc. Chem. Res. 2002, 35, 35−43.
(b) Sato, H.; Hayashi, T.; Ando, T.; Hisaeda, Y.; Ueno, T.; Watanabe,
Y. J. Am. Chem. Soc. 2004, 126, 436−437. (c) Matsuo, T.; Fukumoto,
K.; Watanabe, T.; Hayashi, T. Chem.Asian J. 2011, 6, 2491−2499.
(6) (a) Borovik, A. S. Chem. Soc. Rev. 2011, 40, 1870−1874.
(b) Goldberg, D. P. Acc. Chem. Res. 2007, 40, 626−634. (c) Latifi, R.;
Tahsini, L.; Karamzadeh, B.; Safari, N.; Nam, W.; de Visser, S. P. Arch.
Biochem. Biophys. 2011, 507, 4−13. (d) Lansky, D. E.; Goldberg, D. P.
Inorg. Chem. 2006, 45, 5119−5125. (e) Arunkumar, C.; Lee, Y.-M.;
Lee, J. Y.; Fukuzumi, S.; Nam, W. Chem.Eur. J. 2009, 15, 11482−
11489. (f) Song, W. J.; Seo, M. S.; George, S. D.; Ohta, T.; Song, R.;
Kang, M.-J.; Tosha, T.; Kitagawa, T.; Solomon, E. I.; Nam, W. J. Am.
Chem. Soc. 2007, 129, 1268−1277. (g) Srour, H.; Maux, P. L.;
Simonneaux, G. Inorg. Chem. 2012, 51, 5850−5856. (h) Liu, W.;
Huang, X.; Cheng, M.-J.; Nielsen, R.-J.; Goddard, W. A., III; Groves, J.
A. Science 2012, 337, 1322−1325. (i) Okun, Z.; Gross, Z. Inorg. Chem.
2012, 51, 8083−8090.
(7) (a) Cai, Y.-B.; Li, X.-H.; Jing, J.; Zhang, J.-L. Metallomics 2013, 5,
828−835. (b) Gelb, M. H.; Toscano, W. A.; Sligar, S. G. Proc. Natl.
Acad. Sci. U.S.A. 1982, 79, 5758−5762. (c) Makris, T. M.; von Koenig,
K.; Schlichting, I.; Sligar, S. G. J. Inorg. Biochem. 2006, 100, 507−518.
(8) Sessler, J. L.; Gebauer, A.; Vogel, E. In Porphyrin Handbook;
Kadish, K. M., Smith, K. M., Guilard, R., Eds.; Academic Press: San
Diego, CA, 2000; Vol. 2, pp 1−54.
(9) One of the propionate side chains interacts with the His 97 and
Lys 96 residues at distances of 3.01 and 4.22 Å, respectively.
(10) Zahran, Z. N.; Chooback, L.; Copeland, D. M.; West, A. H.;
Richter-Addo, G. B. J. Inorg. Biochem. 2008, 102, 216−223.
(11) Support for the weak coordination of His93 to Mn(III) also
comes from acid titration experiments (see Table S3).
(12) An enantiomeric excess (ee) value for produced 1-phenyl-
ethanol was 14%(S).
(13) Green, M. T. Curr. Opin. Chem. Biol. 2009, 13, 84−88.
(14) The highly active intermediate in the catalytic oxidation is
immdiately reduced by amino acid residues near the cofactor in the
heme pocket: Lardinois, O. M.; Ortiz de Montellano, P. R.
Biochemistry 2004, 43, 4601−4610.
(15) Adachi, S.; Nagano, S.; Ishimori, K.; Watanabe, Y.; Morishima,
I.; Egawa, T.; Kitagawa, T.; Makino, R. Biochemistry 1993, 32, 241−
252.
(16) For rMb(MnPc), the KIE values clearly depend on substrates.
For example, the value derived from toluene and deuterated toluene
(C7D8) is 6.1 (Figure S4). In the case of hydroxylation by manganese
porphyrin, it was reported that the KIE value is ca. 2-fold lower than
that of iron porphyrin. This may be caused by a different conformation
of the transition state: Sorokin, A.; Robert, A.; Meunier, B. J. Am.
Chem. Soc. 1993, 115, 7293−7299.
(17) TONs for hydroxylation of toluene and cyclohexane are 8.5 and
0.9, respectively.
(18) Man, W.-L.; Lam, W. W. Y.; Kwong, H.-K.; Yiu, S.-M.; Lau, T.-
C. Angew. Chem., Int. Ed. 2012, 51, 9101−9104.
(19) Nakano, T.; Agatsuma, N.; Kodama, S.; Kakuda, H.; Dolphin, D.
Bull. Chem. Soc. Jpn. 1996, 69, 3513−3512.
(20) (a) Green, M. T.; Dawson, J. H.; Gray, H. B. Science 2004, 304,
1653−1656. (b) Denisov, I. G.; Makris, T. M.; Sligar, S. G.;
Schlichting, I. Chem. Rev. 2005, 105, 2253−2277. (c) Ortiz de
Montellano, P. R. Chem. Rev. 2010, 110, 932−948. (d) Rittle, J.;
Green, M. T. Science 2010, 330, 933−937.
(21) (a) Hara, I.; Ueno, T.; Ozaki, S.; Itoh, S.; Lee, K.; Ueyama, N.;
Watanabe, Y. J. Biol. Chem. 2001, 276, 36067−36070. (b) Pfister, T.
D
dx.doi.org/10.1021/ja409404k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX