A De Novo Heterodimeric Due Ferri Protein Minimizes the Release of Reactive Intermediates in Dioxygen-Dependent Oxidation

Article abstract of DOI:10.1002/anie.201707637

Metalloproteins utilize O₂ as an oxidant, and they often achieve a 4-electron reduction without H₂O₂ or oxygen radical release. Several proteins have been designed to catalyze one or two-electron oxidative chemistry, but t

Full text of DOI:10.1002/anie.201707637

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
International Edition Chemie
www.angewandte.org
Accepted Article
Title: A De Novo Heterodimeric Due Ferri Protein Minimizes the
Release of Reactive Intermediates in Dioxygen-dependent
Oxidation
Authors: Marco Chino, Linda Leone, Ornella Maglio, Daniele D'Alonzo,
Fabio Pirro, Vincenzo Pavone, Flavia Nastri, and Angela
Lombardi
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.201707637
Angew. Chem. 10.1002/ange.201707637
Link to VoR: http://dx.doi.org/10.1002/anie.201707637
http://dx.doi.org/10.1002/ange.201707637

Angewandte Chemie International Edition
10.1002/anie.201707637
COMMUNICATION
A De Novo Heterodimeric Due Ferri Protein Minimizes the Release
of Reactive Intermediates in Dioxygen-dependent Oxidation
+
+
Marco Chino, Linda Leone, Ornella Maglio, Daniele D’Alonzo, Fabio Pirro, Vincenzo Pavone,
Flavia Nastri, and Angela Lombardi*
Abstract: Metalloproteins utilize O
2
as an oxidant, and they often
or oxygen radical
phenol-oxidase that catalyzes the O
aminophenol (4AP) to 4-benzoquinone
The released highly reactive product is rapidly
quenched in solution by reaction with m-phenylendiamine (MPD)
(Figure 1) to form the aminoindoaniline dye,
spectrophotometrically detectable at 486 nm, pH 7.
2
-dependent oxidation of 4-
achieve a 4-electron reduction without H
2
O
2
monoimine
[
13,14]
release. Several proteins have been designed to catalyze one or
(4BQM).
two-electron oxidative chemistry, but the de novo design of a protein
that catalyzes the net 4-electron reduction of O
reported yet. We report here the construction of a diiron-binding four-
helix bundle, made up of two different covalently linked α
2
has not been
[
15]
2
monomers, through click chemistry. Surprisingly, the prototype
protein, DF-C1, showed a large divergence in its reactivity from
earlier DFs. DFs release the quinone imine and free H
2 2
O in the
III
oxidation of 4-aminophenol in the presence of O , whereasFe -DF-
2
C1 sequesters the quinone imine into the active site, and catalyzes
inside the scaffold an oxidative coupling between oxidized and
reduced 4-aminophenol. The asymmetry of the scaffold allowed a
fine-engineering of the substrate binding pocket, that ensures
selectivity.
2
Figure 1. Reaction scheme depicting the O -dependent oxidation of 4AP
catalyzed by Fe -DF3 (top) and Fe -DF-C1 (bottom). In the case of DF3, the
oxidized 4BQM is released in solution and subsequently quenched with MPD.
Nature accomplishes the activation of the highly stable dioxygen
bond using different enzymes that often employ transition metal
III
III
[
1-3]
ions, such as copper or iron.
Metalloenzymes are
, and for catalyzing
4BQM is retained into the active site of DF-C1, where it undergoes coupling
spectacularly organized for processing O
2
with another molecule of 4AP prior to be released.
oxidation reactions of a variety of organic substrates with high
selectivity, while avoiding diffusion of deleterious reactive
intermediates.
The goal of the current work was to re-engineer the DF3
scaffold, in order to sequester the reactive quinoneimine
intermediate inside the protein, to change the 4AP oxidation
pathway. We targeted toluene mono-oxygenase proteins
Insights into structure-function relationships of natural
dioxygen-activating metalloenzymes have partly derived from
the study of biomimetic systems of different sizes and
[
16,17]
(TMOs)
and tried to transplant the asymmetry of their active
[
4-6]
complexities.
reactive and unstable species as sacrificial oxidants (such as
peroxides or peracids) rather than O .Progress in this field still
The majority of these models uses highly
site into DF3. Using a design method, recently developed by
[
18]
us,
we obtained an asymmetric four-helix bundle (DF-C1)
[
19]
2
through click chemistry,
design.
a strategy rarely used in protein
We chose the Cu catalyzed azide-alkyne
[
20,21]
I
represents a great challenge, for both mechanistic studies and
practical applications.
[
22]
cycloaddition (CuAAC) to selectively link two different α-helical
hairpins. To this aim, we inserted propargyl glycine (Pra) and 6-
azido-hexanoic acid (6-aha) into the DF3 sequence (Figure 2a),
leading to two clickable chains (K and Z, respectively). Pra and
6-aha positions in each chain were evaluated by computational
Along these lines, we have developed the DF (Due Ferri, two-
iron) family of de novo designed metalloproteins, to mimic the
[
7]
natural diiron-oxo-proteins. The DF scaffold is made up by a
stable four-helix bundle, able to tolerate several mutations for
[
8-12]
[18]
[13]
hosting different activities.
In particular, DF3 is an artificial
screening, based on the NMR structure of DF3 (PDB 2KIK)
Figure S1).
Beside the clicking residues, we introduced into the DF3
(
sequence a total of nine mutations, which can be categorized
into three classes, intended to: i) modulate active site access,
substrate recognition and catalysis; ii) form H-bond network; iii)
improve packing (Figures 2b, S2, S3, Supporting Information).
We first considered that four Gly lined the active site cavity in
DF3, creating a wide access channel; this was made narrower
[
*]
Dr. M. Chino, Dr. L. Leone, Dr. D. D’Alonzo, Dr. F. Pirro, Prof. V.
Pavone, Prof. F. Nastri, Prof. A. Lombardi
Department of Chemical Sciences
University of Napoli “Federico II”
Via Cintia, 80126 Napoli (Italy)
E-mail: angela.lombardi@unina.it
Dr. O. Maglio
IBB - National Research Council
Via Mezzocannone 16, 80134 Napoli (Italy)
These authors contributed equally to this work
Supporting Information for this article is given via a link at the end of
the document.
and more hydrophobic in DF-C1, by mutating G13
K
T and G9
Z
F
(
subscript K or Z indicates the chain). Remarkably, these two
[+]
Z Z
mutations play additional roles: G9 F, together with Y17 L and
[
23]
Y17
K Z
F, mimic the Phe-rich pocket found in TMOs; Y17 L
mutation also prevents the off-pathway tyrosinate-iron complex
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201707637
COMMUNICATION
[
10]
formation, previously reported by Reig et al.; G13
K
T, together
binding, with the expected 2:1 metal:protein stoichiometry
(Figures 3b and S9).
with L33 Q, could stabilize a water molecule involved in proton
K
[
24,25]
transfer.
Next, a new second-shell H-bond was introduced by L43
mutation, which, together with L29 Q mutation, enabled the
formation of an H-bond network comprising E10 -N43 -Q29
N26 (Figure 2b). Finally, two additional mutations were inserted
to optimize packing (Q16 L, I14 A).
K
N
Z
K
K
Z
-
Z
K
K
II
Figure 3. Folding and Fe -binding analyses of DF-C1.a) CD spectral changes
II
II
upon Fe addition. b) CD titration curve of apo-DF-C1 with Fe . The fraction
II
bound (x
b
) is plotted versus Fe equivalents. The smooth curve represents the
II
best fit for a binding isotherm with a 2:1 (Fe :Peptide) stoichiometry, and pKdiss
2
5
.61 ± 0.03 (R = 0.996; see Supporting Information).
Insight into the coordination geometry was obtained from the
II
absorption spectrum of the Co
complex (Supporting
Information). Position and intensities of the bands (Figure S10)
are indicative of either bis-five coordinate or five/six coordinate
[
26]
II
geometry,
as typically observed in Co -reconstituted diiron
[
27]
proteins (Table S1).
II
Finally, the NMR spectrum of the diamagnetic Zn -complex
exhibits features of a well-folded protein, with narrow line-widths
and well-dispersed proton chemical shifts, in both the amide and
aliphatic regions (Figure S11). Furthermore, Diffusion-Ordered
NMR spectroscopy (DOSY) gave a hydrodynamic radius (R
H
)
value of 24±2 Å, comparable to that calculated for the model
[
28,29]
(
23.4 Å) (Figure S12), confirming the intended folded state.
Fe -DF-C1 reacts with O
II
2
to form the oxo-bridged diferric
species, as monitored by the appearance of the ligand-to-metal-
-1
-1
charge transfer (LMCT) band at 350 nm (ε: 5500 M cm , as
expected for this class of proteins, see Figure S13 and Table
[
14]
S2).
To test whether the active site re-design was successful in
III
driving a different reaction pathway respect to Fe -DF3, we
investigated the reactivity of Fe -DF-C1in the oxidation of 4AP.
Figure 2. a) DF-C1 click reaction scheme; K- and Z-chains are in blue and
orange, respectively. b) Energy minimized model of Zn -DF-C1. All the
III
II
Differently from DF3, no significant formation of the
aminoindoaniline dye was observed when the 4AP oxidation
was performed in the presence of MPD (data not shown). The
mutated residues respect to DF3 are in sticks. The three classes of mutations
are highlighted in different colors: i) active site access substrate recognition
and catalysis (cyan); ii) H-bond network (magenta); iii) packing improvement
(green). The hydrogen bond network, spanning from E10(K) to N26(Z), is
III
addition of sole 4AP to a solution of Fe -DF-C1, in the presence
reported as dashed yellow lines.
of ambient O
2
, led to the formation of two absorption bands at
370 nm and 470 nm, whose intensities increased over time
The synthetic K and Z chains were coupled through CuAAC in
buffered solution at pH 8.0 (Figures 2a, S4-S7, Supporting
Information).
(Figure 4a).
III
The corresponding reaction with Fe -DF3 revealed a different
spectroscopic behavior, with the appearance of two bands at
3
70 nm and 500 nm (Figure 4b). Both bands achieved their
Spectroscopic characterization by CD, UV/Vis and NMR
indicated that DF-C1 binds two metal ions per protein, within a
well-folded structure.
maximum intensities within h, and subsequently, their
intensities decreased over several hours, with the concomitant
formation of a brownish precipitate.
1
The CD spectrum of the apo-protein at pH 7.0 is typical of a α-
II
helical structure (Figure S8). Addition of Fe (Figure 3a) (as well
II
II
as Zn and Co , Figure S9) to apo-DF-C1 led to an increase in
the ellipticity at 208 and 222 nm, indicating a further increase in
the helical content. CD titrations indicated a reasonably tight
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201707637
COMMUNICATION
4
BQM or the other reaction intermediates. Upon addition of
III
catalase (200 U/mL) to remove the diffusing H
produced only 1, similarly to Fe -DF-C1 (Figures S22).
2
O
2
, Fe -DF3
III
III
III
Differently from Fe -DF3, Fe -DF-C1 prevents the diffusion in
solution of the reactive 4BQM intermediate (Figure 6).The
designed second-shell interactions would favor 4AP oxidative
dimerization coupled to O
2
reduction in two successive two-
electron steps inside the protein. The 4AP would thus be kept
tightly bound into the protein scaffold after being oxidized to
4BQM, leading to dimer formation (oxidation), with in situ
peroxide consumption.
In this scenario, the rate-limiting step would be the release of
the oxidized dimer in solution, which in turns, by coupling with
another equivalent of 4AP, affords 1 (Figure 6).
III
III
Figure 4. UV/Vis spectra of the oxidation of 4AP by: a) Fe -DF-C1; b) Fe -
III
DF3. Fe -DF3 begun giving product precipitation after 1 h. Experimental
conditions: 40 μM catalyst, 400 μM 4AP; 10 mM HEPES, 10 mM Na SO pH
2 4
7)
To identify the compounds formed in the reactions catalyzed
by the two enzymes, the mixtures were analyzed by HPLC, 1
hour after mixing (Figure 5). A single peak, at r =15 min, was
t
III
observed in the chromatogram, when Fe -DF-C1 was used as
catalyst (Figures S15, S16). The product was identified as the
oxidation/condensation derivative of 4AP (Figure 5, 1, Figure
[
30]
S17, Table S4), analog of the Bandrowski’s base. The HPLC
III
profile of 4AP oxidation, catalyzed by Fe -DF3, revealed instead
the formation of multiple products, with compound
representing only a minor component (≈ 20% of total yield;
Figure 5, S18). control reaction, carried out using
1
A
II
stoichiometric amount of Fe ammonium sulfate, revealed no
product formation (Figures S20, S21).
III
Figure 6. Proposed catalytic cycle for the Fe -DF-C1catalyzed 4AP oxidation.
The proposed mechanism is supported by the following
evidences. Firstly, catalase or superoxide dismutase addition
does not cause any change in the product identity (Figure S23).
Moreover, no product formation was observed upon addition of
III
Figure 5. HPLC traces at 254 nm of 4AP oxidation in presence of: Fe -DF-C1
(
III
III
II
blue), Fe -DF3 (red), Fe ammonium sulfate (green). All traces were acquired
different substrates and hydrogen peroxide to Fe -DF-C1, thus
after 1 hour of reaction progress. Peak eluted at 15 min has been identified as
compound 1. Inset: ESI-IT-TOF mass spectrum of compound 1.
ruling out the involvement of peroxide in the mechanism (see
Supporting Information).
Secondly, addition of one equivalent of 4AP to Fe -DF-C1,
under aerobic conditions, caused the appearance of a broad
band around 500 nm (Figure S24a). These spectral changes
III
III
These results provide evidence that Fe -DF3 is an
III
“
imperfect” catalyst, respect to Fe -DF-C1, in the 4AP oxidation,
being uncontrolled and thereby less selective. This allowed us to
propose different reaction pathways for the two catalysts. 4AP is
oxidized at the diiron site of DF3 with release of the reactive
suggest that 4AP binds to the diferric center, giving rise to a
[9,31]
ligand to metal charge-transfer transition.
Further, addition of
10 eq. of 4AP under anaerobic conditions led to the reduction of
2 2
4BQM and free H O in solution (Figure 1). In the absence of
the di-ferric to the diferrous center, as ascertained by the
decrease of the band at 350 nm (Figure S24b). Upon dioxygen
addition, formation of compound 1 was observed.
MPD, 4BQM spontaneously couples in solution with 4AP, to give
compound 1, with the rate-limiting step being the first coupling
[
30]
reaction. Formation of additional by-products (Figure 5) could
derive from -dependent degradation of compound 1,
2 2
H O
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201707637
COMMUNICATION
Finally, analysis of the docking model between the 4AP dimer
and DF-C1 revealed that 4AP dimer properly fits into the
designed pocket, suggesting a key role of the designed Phe9 in
guiding the dimer formation by π-stacking interaction (Figure
S25a and S26). Interestingly, when 4AP dimer was docked to
DF3, no specificity was observed in the wider binding pocket
[5] Y. Lin, Coord. Chem. Rev. 2017, 366, 1.
[
[
[
6] L. A. Churchfield, A. George, F. A. Tezcan, Essays Biochem. 2017, 61,
245.
7] M. Chino, O. Maglio, F. Nastri, V. Pavone, W. F. DeGrado, A. Lombardi,
Eur. J. Inorg. Chem.2015, 2015, 3371.
8] S. Geremia, L. Di Costanzo, L. Randaccio, D. E. Engel, A. Lombardi, F.
Nastri, W. F. DeGrado, J. Am. Chem. Soc. 2005, 127, 17266.
(
Table S6 and Figure S25b), suggesting the intrinsically lower
[9] J. Kaplan, W. F. DeGrado, Proc. Natl. Acad. Sci. USA 2004, 101, 11566.
[10] A. J. Reig, et al., Nat. Chem. 2012, 4, 900.
driving force for the 4AP coupling to take place. Further
mechanistic studies are however required to carve out
completely all the intermediates involved in the reaction pathway.
This will lead to a deeper understanding of the mechanism
herein hypothesized, and eventually will open to unconsidered
alternatives.
[11] R. A. Snyder, S. E. Butch, A. J. Reig, W. F. DeGrado, E. I. Solomon, J.
Am. Chem. Soc. 2015, 137, 9302.
[12] R. A. Snyder, J. Betzu, S. E. Butch, A. J. Reig, W. F. DeGrado, E. I.
Solomon, Biochemistry 2015, 54, 4637.
[13] M. Faiella, C. Andreozzi, R. Torres Martin de Rosales, V. Pavone, O.
Maglio, F. Nastri, W. F. DeGrado, A. Lombardi, Nat. Chem. Biol .2009, 5,
In conclusion, we successfully applied a new method to
modify the DF3 scaffold and generate the asymmetricDF-C1, by
covalent ligation of two functionalized α peptides through click
2
chemistry. This approach, never used before for the design of
four helix bundle proteins, allowed us to engineer carefully the
second-shell active site interactions and to sequester a reactive
intermediate inside the protein. With this in hand, we succeeded
in steering the reaction mechanism and product distribution of a
di-iron protein. The present study shows that in 4AP oxidation
8
82.
[14] R. Torres Martin de Rosales, M. Faiella, E. Farquhar, L. Que, C.
Andreozzi, V. Pavone, O. Maglio, F. Nastri, A. Lombardi, J. Biol. Inorg.
Chem. 2010, 15, 717.
[
15] J. F. Corbett, J. Chem. Soc. B Phys. Org. 1969, 823.
16] M. S. McCormick, M. H. Sazinsky, K. L. Condon, S. J. Lippard, J. Am.
Chem. Soc. 2006, 128, 15108.
[
[
17] L. J. Bailey, B. G. Fox, Biochemistry 2009, 48, 8932.
18] M. Chino, L. Leone, O. Maglio, A. Lombardi, MethodsEnzymol. 2016,
[
580, 471.
the diiron site performs four-electron O
successive two-electron steps inside the protein. This finding fills
2
reduction in two
[19] H. C. Kolb, M. G. Finn, K. B. Sharpless, Angew. Chem. Int. Ed. 2001, 40,
2004; Angew. Chem. 2001, 113, 2056.
[
32]
the gap between de novo designed and natural proteins. The
efficiency of the new catalyst is still far from natural oxidases
and oxygenases; however, the asymmetry of the scaffold, and
the boost offered by the synthetic method will allow for a fine-
tuning of the active site steric and electronic properties, thus
affecting the overall reactivity and specificity.
[20] W. S.Horne, M. K. Yadav, C. D. Stout, M. R. Ghadiri, J. Am. Chem. Soc.
2004, 126, 15366.
[21] J. M. Beierle, W. S. Horne, J. H. van Maarseveen, B. Waser, J. C. Reubi,
M. R. Ghadiri, Angew. Chem. Int. Ed. 2009, 48, 4725; Angew. Chem.
2
009, 121, 4819.
22] V. D. Bock, H. Hiemstra, J. H. van Maarseveen, Eur. J. Org. Chem. 2006,
006, 51.
23] M. H. Sazinsky, J. Bard, A. Di Donato, S. J. Lippard, J. Biol. Chem. 2004,
79, 30600.
24] N. L. Elsen, L. J. Bailey, A. D. Hauser, B. G. Fox, Biochemistry 2009, 48,
838.
[
[
[
[
2
2
Acknowledgements
3
25] W. J. Song, M. S. McCormick, R. K. Behan, M. H. Sazinsky, W. Jiang, J.
The authors thank W.F. DeGrado for helpful suggestions and
Fabrizia Sibillo for technical assistance. The Cost Action
CM1003 and the Scientific Research Department of Campania
Region (STRAIN Project, grant number B25B0900000000,
postdoctoral fellowship to M.C.) are also acknowledged.
Lin, C. Krebs, S. J. Lippard, J. Am. Chem. Soc. 2010, 132, 13582.
[26] I. Bertini, C. Luchinat, Adv. Inorg. Biochem. 1984, 6, 71.
[27] M. H. Sazinsky, M. Merkx, E. Cadieux, S. Tang, S. J. Lippard,
Biochemistry 2004, 43, 16263.
[
28] J. A. Jones, D. K. Wilkins, L. J. Smith, C. M. Dobson, J. Biomol. NMR
997, 10, 199.
29] J. Garcıa de la Torre, M. L. Huertas, B. Carrasco, J. Magn. Reson.2000,
47, 138.
[30] J. F. Corbett, E. P. Gamson, J. Chem. Soc. Perkin Trans. 2 1972, 1531.
1
[
Keywords:bioinorganic chemistry•de novo protein
1
design•oxidoreductase•diiron-oxo proteins•click chemistry
[31] J. R. Calhoun, C. B. Bell, T. J. Smith, T. J. Thamann, W. F. DeGrado, E.
I. Solomon, J. Am. Chem. Soc. 2008, 130, 9188.
[1]
[2]
[3]
A. Decker, E. I. Solomon, Curr. Opin. Chem. Biol. 2005, 9, 152.
S. Sahu, D. P. Goldberg, J. Am. Chem. Soc. 2016, 138, 11410.
M. V. Doble, A. C. C. Ward, P. J. Deuss, A. G. Jarvis, P. C. J. Kamer,
Bioorg. Med. Chem. 2014, 22, 5657.
[
32] F. Yu, V. M. Cangelosi, M. L. Zastrow, M. Tegoni, J. S. Plegaria, A. G.
Tebo, C. S. Mocny, L. Ruckthong, H. Qayyumand, V. L. Pecoraro, Chem.
Rev. 2014, 114, 3495.
[4] F. Nastri, M. Chino, O. Maglio, A. Bhagi-Damodaran, Y. Lu, A. Lombardi,
Chem. Soc. Rev. 2016, 45, 520.
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201707637
COMMUNICATION
Table of Contents
COMMUNICATION
Not just a four helix bundle: The
use of four helix bundle scaffold with
an asymmetric active site leads to an
enhancement in selectivity of the
iron-catalyzed oxidative coupling of
phenols. The stabilization of the
Marco Chino, Linda Leone, Ornella
Maglio, Daniele D’Alonzo, Fabio Pirro,
Vincenzo Pavone, Flavia Nastri,
Angela Lombardi*
Page 1 – Page 4
oxidized
engineered binding pocket enables
the net four-electron O reduction,
without release of any detectable
. The newly engineered activity
intermediate
in
the
A De Novo Heterodimeric Due Ferri
Protein Minimizes the Release of
Reactive Intermediates in Dioxygen-
dependent Oxidation
2
2 2
H O
is unprecedented for this family of
catalysts.
This article is protected by copyright. All rights reserved.