You are on page 1of 10

Archives of Biochemistry and Biophysics 602 (2016) 106e115

Contents lists available at ScienceDirect

Archives of Biochemistry and Biophysics

journal homepage:

Subtle structural changes in the Asp251Gly/Gln307His P450 BM3

mutant responsible for new activity toward diclofenac, tolbutamide
and ibuprofen
Giovanna Di Nardo a, b, Valentina Dell'Angelo a, Gianluca Catucci a, Sheila J. Sadeghi a,
Gianfranco Gilardi a, b, *
Department of Life Sciences and Systems Biology, University of Torino, Via Accademia Albertina 13, Torino, Italy
CrisDi, Interdepartmental Center for Crystallography, University of Torino, Via Pietro Giuria 7, Torino, Italy

a r t i c l e i n f o a b s t r a c t

Article history: This paper reports the structure of the double mutant Asp251Gly/Gln307His (named A2) generated by
Received 9 October 2015 random mutagenesis, able to produce 40 -hydroxydiclofenac, 2-hydroxyibuprofen and 4-
Received in revised form hydroxytolbutamide from diclofenac, ibuprofen and tolbutamide, respectively. The 3D structure of the
10 December 2015
substrate-free mutant shows a conformation similar to the closed one found in the substrate-bound wild
Accepted 14 December 2015
Available online 21 December 2015
type enzyme, but with a higher degree of disorder in the region of the G-helix and F-G loop. This is due to
the mutation Asp251Gly that breaks the salt bridge between Aps251 on I-helix and Lys224 on G-helix,
allowing the G-helix to move away from I-helix and conferring a higher degree of exibility to this
Cytochrome P450
element. This subtle structural change is accompanied by long-range structural rearrangements of the
Biocatalysis active site with the rotation of Phe87 and a reorganization of catalytically important water molecules.
Protein engineering The impact of these structural features on thermal stability, reduction potential and electron transfer is
Conformational equilibrium investigated.
The data demonstrate that a single mutation far from the active site triggers an increase in protein
exibility in a key region, shifting the conformational equilibrium toward the closed form that is ready to
accept electrons and enter the P450 catalytic cycle as soon as a substrate is accepted.
2015 Elsevier Inc. All rights reserved.

1. Introduction poor stability and activity. However, bacterial counterparts such as

cytochrome P450 BM3 (CYP102A1) from Bacillus megaterium, can
Cytochrome P450 enzymes are heme-thiolate monooxygenases be directly used [3] or engineered to metabolise drugs and to
interesting for biocatalytic purposes for their ability of CeH bond produce the same metabolites as the human enzymes [4e14]. This
functionalization [1]. They have a large potential in elds such as protein is a self-sufcient fatty acids monooxygenase containing a
drug discovery, ne chemicals synthesis, bioremediation, and NADPH-dependent reductase (BMR) and a P450 catalytic domain
biotechnology. In particular, those from human liver are respon- (BMP) fused in a single polypeptidic chain [15,16]. The high catalytic
sible for Phase I metabolism of drugs, that are converted in one or efciency makes it an excellent template for protein engineering
more metabolites. The toxicity and biological activity of such me- aimed at improving substrate versatility and catalytic activity to-
tabolites have to be fully characterized during the development of wards drugs for the generation of metabolites [14,17e20].
new drugs as they can cause adverse reactions in the human body Random mutagenesis is a potentially successful approach to
[2]. It is necessary therefore to synthesize preparative amounts of introduce new functions in enzymes for biocatalytic purposes.
these compounds to assess their toxicity, a process that is difcult However, only the structural characterization of the improved
to achieve by classical synthetic methods. The direct use of human variants can elucidate the effect of the randomly introduced mu-
cytochrome P450 is limited by the need of a redox partner and the tations on protein structure and function, offering the basis for
understanding how and why an enzyme has evolved to recognize
and turn over new molecules. Here, we report the crystal structure
* Corresponding author. Department of Life Sciences and Systems Biology, Uni- of the heme domain of the double mutant D251G/Q307H of
versity of Torino, Via Accademia Albertina 13, Torino, Italy.
0003-9861/ 2015 Elsevier Inc. All rights reserved.
G. Di Nardo et al. / Archives of Biochemistry and Biophysics 602 (2016) 106e115 107

cytochrome P450 BM3, named A2, previously generated and protein concentration using as absorbance value the DAbs450e490.
characterized from a catalytic point of view for its ability to
generate metabolites of diclofenac, ibuprofen and tolbutamide [9]. 2.1.3. Crystallization and determination of protein structures
The information from the crystal structure of the mutant also in Crystallization was achieved by vapor diffusion method in 24-
complex with the substrate N-palmitoylglycine (NPG) together well sitting drop plates at 4  C. Using a protein:reservoir ratio of
with docking experimets and comparison with the available 1:1, P450 BMP A2 (20 mg mL1) was mixed with crystallization
structures of the wild type enzyme [21,22] were then used for a cocktails containing 14% PEG 3350, 100 mM cacodylic acid in a pH
more detailed and comparative functional characterization, giving range from 5.5 to 6.8 and MgCl2 from 100 to 160 mM. For the NPG-
the rational basis to understand why the A2 mutant has evolved bound form, the protein was saturated with the ligand and co-
toward the recognition and the turnover of new substrates due to crystallized. NPG (up to 150 mM) was titrated into BMP sample
two mutations far from its active site. until full shift to 398 nm was observed. The protein was immedi-
ately concentrated by ultraltration in the presence of saturating
2. Experimental section ligand and used for crystallization purposes. Red crystals were
obtained after 3e4 weeks and were ash-frozen in liquid nitrogen
2.1. Materials using 10% w/v PEG200 as a cryoprotectant. X-ray diffraction data
were collected on the ID29 at the European Synchrotron Research
All chemicals used for protein expression, purication, charac- Facility (Grenoble, France) [25] and on the BL13-XALOC beamline at
terization and crystallization were purchased from SigmaeAldrich ALBA synchrotron (Cerdanyola del Valle s, Spain) [26]. Data were
and were analytical grade. Restriction enzymes were purchased processed using XDS [27] and scaled using AIMLESS [28]. The
from New England Biolabs. structure was solved by molecular replacement using Molrep [29]
from the CCP4 suite [30] and the coordinates of the heme domain
2.1.1. Generation of the heme domain (BMP) of the mutant A2 of P450 BM3 structure (PDB 1JPZ) [22]. The models were adjusted
The fragment of the gene encoding for the heme domain of using Coot [31] and rened using Refmac5 [32] and PHENIX [33].
cytochrome P450 BM3 A2 was extracted by digestion with specic Coordinates and structure factors have been deposited in the Pro-
restriction enzymes (BamHI and MscI) and ligated into a pT7 tein Data Bank (PDB) with accession code 5DYP and 5DYZ.
expression vector. The correct insertion of the fragment containing
the mutations was veried by DNA sequencing. 2.1.4. Molecular docking
The chemical structures of diclofenac, ibuprofen and tolbuta-
2.1.2. Protein expression, purication and quantication mide were designed using the software MarvinSketch. The struc-
Expression was carried out in E. coli BL21 (DE3) cells in Lur- tures were then subjected to molecular geometry optimization on
iaeBertani (LB) medium containing 100 mg mL1 of ampicillin. Cells Yasara software. The ligands were docked into the crystal structure
were grown until the optical density at 600 nm (OD600) was 0.7 in of P450 BMP A2, the BMP WT substrate-free structure (PDB ID 1
500 mL ask containing 100 mL of LB medium. Then, the culture BU7) and the BMP WT of substrate-bound structure (PDB 1 JPZ)
was diluted again 1:100 transferring 10 mL of the cells grown up to where the substrate was removed before docking. The YASARA
OD600 value of 0.7 into 2 L asks containing 1 L of medium. Once embedded AutoDock V4 alghorithm [34] was used to predict
reached an OD600 value of 0.7, protein expression was induced by protein-ligand interactions. In order to dock the ligands in the
the addition of 1 mM IPTG. The expression was carried out for 19 h protein active site, a simulation cell was built around the heme iron
at 28  C, 180 rpm shaking. Following cell growth, bacterial cells atom and 25 runs of global docking were performed allowing
were recovered by centrifugation at 4  C and resuspended in Buffer ligand exibility. The binding energies were predicted using the
A (30 mM KPi pH 7.4, 2 mM DTT, 0.1 mM EDTA) containing 0.1 mM scoring function included in AutoDock. The binding energy is ob-
phenyl methyl sulfonyl uoride (PMSF) and 0.2 mg mL1 lysozyme. tained by calculating the energy at innite distance (between the
Cells were lysed by sonication and the soluble fraction was sepa- selected object and the rest of the scene, i.e. the unbound state) and
rated form cell debris by centrifugation at 40,000 rpm for 250 at subtracting the energy of the scene (i.e. the bound state). The more
4  C. Purication was carried out with some modications as positive the binding energy, the more favorable the interaction.
described before [23]. Briey, a double step anion-exchange chro-
matography using a diethylaminoethyl (DEAE-Sepharose Fast-Flow, 2.1.5. Differential scanning calorimetry (DSC)
GE healthcare) and a Q-Sepharose columns (Q-Sepharose Fast Flow, DSC was carried out on a Microcal VP-DSC instrument. Data
GE Healthcare). The fractions containing A2 mutant were pooled analysis was performed using Microcal Origin software. The pa-
and loaded into a 150 mL DEAE column and eluted using a rameters used were as follows: 25e105  C temperature gradient,
50e250 mM NaCl gradient (500 mL). The protein was then 90  C/h scan rate, 10 min prescan thermostat. Background scans
exchanged into Buffer A using an Amicon centrifugal lter (Milli- were carried out with degassed 100 mM KPi pH 8.0 buffer and
pore, 30 KDa cut-off) in order to remove salt excess and loaded into saturated with 150 mM NPG, 300 mM diclofenac, ibuprofen and
a Q-Sepharose chromatography column (70 mL). The protein was tolbutamide for the substrate-bound samples. All samples were run
eluted as reported previously for the DEAE and the buffer con- using 10 mM proteins and saturating amounts of the substrate,
taining NaCl was exchanged with storage buffer 100 mM KPi pH 8.0. when present.
The fractions containing A2 mutant were analysed for purity by
SDS-PAGE gel and pooled according to the purity ratio (A418/ 2.1.6. Fatty acids binding and NADPH consumption studies
A280 1.3). A CO-binding assay was performed on the puried P450 Dissociation constants (KD) for binding of the substrate NPG
BM3 WT and A2 proteins in order to estimate their concentration in were determined by UVeVis spectrophotometer (Agilent Tech-
a UVeVis spectrophotometer (Agilent Technologies) using a 1 cm nologies) in a 1 cm path length quartz cuvette by following the
pathlength quartz cuvette, according to the method described by typical low-to-high-spin transitions. The reaction mixture contains
Omura and Sato [24]. The protein concentration used to perform 1 mM of enzyme and substrate in a concentration range from 0.05 to
the assay was 1 mM. After the formation of the Fe(II)CO complex at 15 mM. Each spectrum was recorded after 3 min of equilibration
450 nm by bubbling CO to the sodium dithionite-reduced protein, upon each substrate addition at 25  C in a 100 mM KPi pH 8.0
the extinction coefcient of 91,000 M1 cm1 was used to estimate buffer. Difference spectra were generated by subtraction of the
108 G. Di Nardo et al. / Archives of Biochemistry and Biophysics 602 (2016) 106e115

spectrum for ligand-free protein from spectra recorded after each mediators (methylene blue, O-safranine, phenazine methosulfate,
addition of NPG. KD values were determined by plotting the anthraquinone-2,6-disulfonic acid, benzyl viologen and methyl
maximal absorbance changes calculated from each difference viologen) were also added to the sample to facilitate electro-
spectrum as a function of substrate concentration and by tting the chemical communication between the protein and the electrode.
data to the equation: The concentration of the mediators was 30 times less compared to
.  that of the protein, 3 mM protein versus 0.1 mM mediator. All the
DA398418 DAmax
398418 $Sfree Kd Sfree (1) redox mediators (except benzyl viologen and methyl viologen that
were diluted in water) were prepared in methanol to a nal con-
where [S]free is [S]totale[E$S] and [E$S] DA398e418 [E]total/ centration of 5 mM. Then a 0.05 mM solution of each mediator was
DAmax398e418. prepared using 100 mM KPi pH 8.0. Finally, putting all the media-
Steady-state kinetics of NADPH consumption for P450 BM3 WT tors together we prepared the mix used in the experiment.
and A2 mutant were measured by a Agilent 8453 UVevis spec- Reduction potentials (Em), expressed relative to the standard
trophotometer by monitoring the NADPH oxidation at 340 nm and hydrogen reference electrode (NHE), were obtained by plotting the
using an extinction coefcient at 340 nm of 6220 M1 cm1. Re- percentage of oxidized fraction of the protein (calculated using the
actions were performed in 100 mM KPi pH 8.0 buffer at 25  C using absorbance values at 419 nm for the ligand-free and 398 nm for the
a protein concentration of 0.5 mM, 150 mM NADPH and, when pre- substrate-bound protein) versus the potential and by tting to the
sent, saturating amounts of the substrate. Each reaction was initi- Nernst equation:
ated by the addition of NADPH to the protein and performed in
E Em RT=nF lnox=red (3)
triplicate. The rate of NADPH consumption was calculated tting
the rst linear part of the reaction to a linear regression curve.
where [ox] and [red] are the concentrations of oxidized and
reduced heme, respectively, E is the solution reduction potential at
2.1.7. Stopped-ow experiments equilibrium, n is the number of electrons, Em is the midpoint po-
Stopped-ow kinetic experiment were performed on cyto- tential in the conditions given above, F is the Faraday constant, R is
chrome P450 BM3 WT and A2 mutant in the absence and in the the gas constant, and T is the absolute temperature.
presence of the substrate NPG in order to determine the rate con- Each plot is the result of at least three experiments.
stant for the formation of the FeII(CO) complex which is strictly
dependent on the rate of electron transfer from NADPH to heme
through the reductase domain (BMR). Experiments were per- 3. Results and discussion
formed at 25  C in 100 mM KPi pH 8.0 buffer using an Hi-Tech
Scientic SF-61 stopped-ow system in a sealed glove box with 3.1. Crystal structures of the heme domain of A2
O2 concentration <10 ppm. Prior to analysis, all buffers and the
solution used were saturated with carbon monoxide (CO) by The crystal structure of the heme domain of P450 BM3 A2 in the
bubbling with the gas for at least 5 min. The enzyme (10 mM) substrate-free and ebound forms was solved at 2.4 and 1.97
present in the rst syringe was mixed with NADPH (200 mM) in a resolution, respectively. Data collection and renement statistics
second syringe in a 1:1 ratio. For the substrate bound form, P450 are reported in Table 1.
BM3 WT and A2 mutant were saturated with the ligand NPG Fig. 1A shows a superimposition of the crystal structure of P450
(15 mM) prior to be analysed. Absorption changes at 450 nm BMP A2 and the substrate-free structure of the wild type protein
(reecting formation of the FeII(CO) adduct) were monitored for 5 s (PDB ID 1 BU7) [21]. The overall RMSD is 0.63e0.72 , depending on
after stopped-ow mixing. Each reaction was performed at least in the monomer taken into account and superimposed, and the most
quadruplicate. Data were tted using Kinetic Studio data analysis signicant changes are seen in F- and G-helices, that are part of the
software to the equation: substrate access channel in P450 BM3 [21]. In line with previously
published crystal structures of different cytochromes P450, the F-G
Abs450 A1 *exp  k1 *X A2 *exp  k2 *X C (2) loop and part of the G-helix are unresolved in A2 mutant or present
the highest B factors, suggesting a high degree of disorder that has
where Abs450 is in the increase in the absorbance at 450 nm, A1 and
A2 are the pre-exponential factors, k1 and k2 are the two rate
Table 1
constants, expressed as reciprocals of the X axis time units.
X-ray data collection and renement statistics for variant A2.


2.1.8. Redox potentiometry PDB ID 5DYP PDB ID 5DYZ
Spectroelectrochemical experiment were carried out at 25  C in
Data collection
a quartz cuvette placed in a cuvette holder adaptor connected by a Space group P212121 P212121
ber optic cable to a spectrophotometer that recorded the spectra Cell dimensions a/b/c/ 61.13 61.21
and simultaneously to a potentiostat to measure the redox poten- 118.45 115.71
tial. The measurements were performed in anaerobiosis (O2 con- 146.95 143.28
Wavelength () 0.98 0.98
centration was less than 10 ppm) by a continuous ow of nitrogen Resolution () 2.40 1.97
gas within the otherwise sealed glove box using an Ag/AgCl refer- Average I/o(I) 13.2 (2.6)a 12.0 (2.1)a
ence electrode and a laminar gold working electrode. Sodium Completeness (%) 98.6 (99.6) 99.6 (94.7)
dithionite was used as the reducing agent for the titrations, aliquots Redundancy 5.5 (5.5) 6.5 (6)
of 1.25 mM were added each time to the protein solution until Rmerge 0.07 (0.572) 0.09 (0.864)
complete reduction was observed. Protein and sodium dithionite Average B-factor (2) 73.0 40.0
solutions were prepared in 100 mM KPi pH 8.0 buffer. After each R work/R free (%) 22.0/25.7 18.6/22.4
addition of sodium dithionite the sample was stirred and the sys- r.m.s.d. bond lengths () 0.006 0.010
tem was allowed to equilibrate for at least 15 min. After equili- r.m.s.d. bond angles ( ) 1.228 1.309
bration at each potential the optical spectrum was recorded. Redox Values in parentheses are for the highest-resolution shell.
G. Di Nardo et al. / Archives of Biochemistry and Biophysics 602 (2016) 106e115 109

Fig. 1. Overall crystal structure superimposition of the heme domain (BMP) of P450 BM3 A2 (brown) with A) substrate-free BMP WT (cyan, PDB ID 1BU7) [21] and B) NPG-bound
BMP WT (magenta, PDB ID 1JPZ) [22]. (For interpretation of the references to colour in this gure legend, the reader is referred to the web version of this article.)

been attributed to a high level of exibility of this part of the P450 Interestingly, the same effect was also found for other mutants of
fold [35e37]. The high degree of exibility in A2 mutant is the cytochrome P450 BM3 with enhanced catalytic properties but it is
result of the replacement of Asp251 with a glycine, that signi- was mainly due to active site mutations, or by a combination of ve
cantly affects the structure since the salt bridge involving Asp251 mutations [10,39]. Here, the same conformational outcome is
and Lys224 is broken, resulting in re-positioning of helix G (Fig. 2A). achieved by only one mutation breaking a key salt bridge on the
However, the movement of helix G is restricted by the new protein surface. Furthermore, the mutations have a long-range ef-
conformation adopted by Arg255 side-chain that allows the con- fect on the structure of the active site of the protein. In fact, a close
servation of a salt bridge with Asp217 found also in the WT protein look in the catalytic pocket shows some subtle changes that are
(Fig. 2A). This side chain elongation of Arg255 was also found in usually triggered by substrate binding. The rst event of the cyto-
another mutant of P450 BM3, named KT2, carrying 5 mutations and chrome P450 catalytic cycle is the substrate-induced displacement
with enhanced activity compared to the WT protein [38,39]. of the water molecule present as sixth ligand of the heme iron. This
The second mutation of A2, Gln307His, is located on the protein displacement induced a low-to-high spin transition of the metal
surface and involves the last residue of helix J0 . This mutation atom that can be observed by visible spectroscopy as a shift of the
removes the hydrogen bond present between Gln307 and Asp300 Soret peak from 419 to 392 nm. In BMP A2, the water molecule
and introduces a new hydrogen bond between Gln307 and Ser304 present as sixth ligand of the heme iron atom is in an off-axial
(Fig. 2B). As a result, the local structure is not signicantly altered position by 12  C and at a longer distance (2.8 ) compared to
by this mutation compared to the wild type protein and the BMP WT structure (Fig. 3A). This is consistent with the unstable
structural changes can therefore be attributed to Asp251Gly predominantly low spin state observed by visible spectroscopy in
mutation. the variant that is however very sensitive to slight changes of ionic
When the BMP A2 mutant structure is superimposed to that of strength and pH. Secondly, the residue Phe87, a residue that un-
substrate-bound BMP wild type, the RMSD decreases to dergoes a rotational movement by around 90 when the substrate
0.56e0.62 and only minor differences in helix G and some is bound, is already almost completely rotated in the substrate-free
external loops are present (Fig. 1B). Thus, the mutations introduced form of A2 mutant, as in the substrate-bound WT structure
in A2 mutant seems to cause a conformational change similar to (Fig. 3A). Since it has been proposed that the spin equilibrium in
that induced by substrate binding in the wild type protein or they P450 BM3 is also affected by Phe87, this is an important change
alter the equilibrium between the open and closed conformation. since it facilitates both the access of substrates to the active site of

Fig. 2. Structural overlay of BMP WT and A2 in the regions where the mutation of A2 are located. A) Asp251 is part of the I-helix in BMP WT (cyan) and forms a salt bridge with Lys
224 of G-helix. In A2, the salt bridge cannot be formed because of the Asp251Gly mutation and, as a result, the G-helix moves far away from I-helix. This movement is however
restricted by the presence both in WT and A2 of the salt bridge between Asp217 and Arg255 that is in a more extended conformation in A2. B) Local effect of the Gln307His
mutation. Gln307 forms a hydrogen bond with Asp300 in the WT (cyan). This hydrogen bond is not more present in A2. A new hydrogen bond is formed between His307 and Ser403
(brown). The local structure is not perturbed by this mutation. (For interpretation of the references to colour in this gure legend, the reader is referred to the web version of this
110 G. Di Nardo et al. / Archives of Biochemistry and Biophysics 602 (2016) 106e115

Fig. 3. Active site of BMP A2. A) Structural overlays showing the rotation of Phe87 in BMP WT (cyan) as a consequence of substrate binding (magenta). In the active site of BMP A2
(tan), Phe87 is almost completely rotated to the substrate-bound position and this movement probably favors substrate access to the active site. In A2 (brown), the I-helix distortion
do to the presence of a water molecule that disrupt the regular intra-helical H-bond pattern (Wat2, cyan), is not present as in the substrate-bound (magenta) WT protein. A water
molecule (Wat3) is instead located where it is in the substrate-bound structure (magenta). B) Hydrogen bond pattern involving the water molecules in A2. The off-axial water
molecule is hydrogen bonded to the CO group of Ala264 that is hydrogen bonded to the water molecule that interrupts the intrahelical hydrogen bond pattern. (For interpretation of
the references to colour in this gure legend, the reader is referred to the web version of this article.)

the protein and the water displacement [40e42]. Moreover, some slightly different position in A2 mutant compared to the WT
residue side chains such as Phe261 and His266 that undergo major enzyme [22] and bound to Tyr51, Gln73 and Ala74 through
changes upon substrate binding, are already in the substrate-bound hydrogen bonds (Fig. 4A). In our structure, Arg47 is not involved in
conformation. Finally, the so-called kink region of helix I, typical of NPG binding whereas a previously published crystal structure BMP
cytochromes P450 where an interruption of the helical H-bond WT in complex with NPG shows that this residue is involved in
pattern is present, is also more similar to the substrate-bound WT. binding [43].
In fact, the water molecule present in the groove formed by the The axial water ligand is now 3 far away from the heme iron in
disruption of the intrahelical hydrogen bond pattern in the an off-axial location increased by 14 compared to substrate-free
substrate-free WT enzyme is missing in A2 mutant (Wat2 in BMP A2 and the water molecule in the helix I groove overlays
Fig. 3A). As a result, the kink is less pronounced compared to the with Wat501 in BMP WT-NPG complex structure (PDB ID 1JPZ) and
substrate-bound WT enzyme and this is due to the disruption of the with the water molecule found in substrate-free A2 mutant
hydrogen bonds between Leu262CO-His266NH and Ala264CO- (Fig. 4B). Thus, upon substrate binding, even if with differences
Thr268NH (Fig. 3B). Moreover a water molecule, corresponding to compared to the WT enzyme, there is a relocation of the water
Wat501 in the structure of BMP WT in complex with NPG (PDB ID molecules responsible for the spin shift and thought to be impor-
1JPZ), is present in a similar position and forms hydrogen bonds tant for oxygen binding and catalysis [22].
with Ala264, Thr268 and Thr269 (Wat3 in Fig. 3A). Ala264 also Since the mutant A2 has been shown to have new catalytic
forms a hydrogen bond with the axial water ligated to the heme abilities toward diclofenac, ibuprofen and tolbutamide [9], co-
iron as in the WT enzyme (Fig. 3B). crystallization experiments were also performed by complexing
The crystal structure of A2 mutant in complex with NPG, a the enzyme with the three drugs at saturating concentrations
known substrate of P450 BM3 [22], was also solved and the elec- before crystallization and monitoring the spin shift associated to
tron density of the substrate was visible in one of the monomer the drug binding. For the A2-diclofenac complex, diffraction data
(Fig. 4A). The overall structure is very similar to that of substrate- were collected at 2.0 resolution and the structure solved. A
free A2 mutant (RMSD 0.58 ). The substrate NPG resulted in a positive density in the active site and in proximity of Arg47 in the

Fig. 4. Structure of BMP A2 in complex with the substrate NPG. A) Active site of the protein showing an unbiased difference electron density (contoured at 2.5 s) for the bound NPG,
and the residues involved in hydrogen bonds with the substrate. B) Structural overlay between substrate-free (brown) and ebound (green) BMP A2, showing the movement of the
water molecule Wat1, the rotation of Phe87 and the xed position of Wat3. (For interpretation of the references to colour in this gure legend, the reader is referred to the web
version of this article.)
G. Di Nardo et al. / Archives of Biochemistry and Biophysics 602 (2016) 106e115 111

difference electron density map was observed but tting a diclo- Table 2
fenac molecule resulted in an increase in the Rfree factor during Binding energies for the three drugs docked into the active site of BMP WT in the
substrate-free (1BU7), in the substrate-bound (1JPZ) conformations and BMP A2.
renement even if alternative conformations of the drug were used
to t the electron density. This result suggests that the drug enters 1BU7 1JPZ A2
the active site of the protein but it is highly mobile and does not Diclofenac 3.59 5.89 7.42
form stable and specic interactions with active site residues of the Ibuprofen 0.79 5.56 5.58
protein. Indeed, active site residues are not changed in A2 mutant Tolbutamide 8.28 5.56 6.68

compared to the wild type protein that is not able to bind and turn
over the drug. In fact, the KD values measured for the substrate NPG
is similar for WT and A2 mutant (0.73 0.07 and 0.54 0.07 mM, hydroxylate different compounds similarly placed in the binding
respectively). pocket has been explained by NMR studies showing that the sub-
The structural data, taken together, suggest that the mutation strate moves within 3 distance from the heme iron upon reduc-
Asp251Gly introduces a higher degree of exibility of part of the tion from Fe(III) to Fe(II) [44].
access channel compared to WT, that can increase the probability to In Fig. 5, it is possible to see that in BMP WT in the substrate-free
introduce new molecules to the active site of the protein. The conformation, Phe87 blocks the access of the drug close to the
mutation is responsible for a change in the overall open-to-close heme iron whereas in BMP A2, the already rotated conformation of
conformational equilibrium with a long-range effect on the active this residue allows the drug to go closer to metal center. Also in the
site, where the rotation of Phe87 facilitates the entrance of sub- BMP WT closed conformation, where Phe87 is rotated, the sub-
strates close to the heme iron. Moreover, a rearrangement of the strate can go even closer to heme. Moreover, the binding energies
active site water molecules, similar to the one induced by substrate for the three drugs are positive in the cases where the drugs were
binding in the WT protein, is seen in A2 mutant. docked in the active site of BMP WT in the closed conformation and
in the mutant BMP A2 (Table 2). They are negative when the sub-
strates are docked in the open conformation of BMP WT, suggesting
3.2. Docking of diclofenac, ibuprofen and tolbutamide in the active poor binding to the open conformation of the wild type enzyme.
site of BMP WT and A2 These results suggest that the three drugs, being non-physiological
substrates, are not able to induce the rotation of Phe87 and they
The three drugs that BMP A2 is able to turn over were docked in need to meet this residue in an already rotated conformation in
its active site as well as in the active site of BMP WT (PDB ID 1BU7). order to arrive close enough to heme iron.
Moreover we also docked the three drugs in the active site of BMP
WT in complex with NPG (PDB ID 1JPZ) where the substrate NPG
was removed before the docking experiments. This experiment was 3.3. Differential scanning calorimetry (DSC)
carried out to check whether the rotation of Phe87, present in the
closed conformation and induced by substrate binding in BMP WT, In order to investigate the role of the conformational change of
is responsible for substrate accommodation close to the heme in the protein in relation to its stability and exibility, differential
the active site of the protein. For this reason, in order to keep Phe87 scanning calorimetry experiments were performed fro BMP WT
in the three possible rotamers found in BMP WT (open and closed and A2 in the absence and presence of the NPG substrate. The re-
conformers) and BMP A2, we perform the docking analysis with the sults are shown in Fig. 6 and the melting temperatures (Tm)
protein xed in the crystallographic conformations and the sub- calculated are reported in Table 3.
strate exible. The DSC prole of the WT protein in the absence of substrate
Fig. 5 shows the results obtained for the drug ibuprofen, but shows a wide peak that can be deconvoluted into two components
similar results were obtained for diclofenac and tolbutamide. In with Tm1 of 55.8 and Tm2 of 61.4  C. Upon NPG binding, the two
particular, the distance between the heme iron and the carbon transitions become better resolved with a Tm1 of 61.4 and a Tm2 of
atom known to be hydroxylated (9) is within the range reported in 66.5  C. The presence of two transitions could be ascribed to the
the crystal structures of P450 BMP WT in complex with palmitoleic presence in BMP of two sub-domains [45] while the increase of the
acid (42) and NPG (7e8 ) (22). Although this distance appears to Tm values observed upon substrate addition is consistent with the
be too long for catalysis to occur, the ability of P450 BM3 to additional interactions between the protein and the substrate

Fig. 5. Docking of ibuprofen into the active site of A) BMP WT crystallized in the absence of substrate and in the open conformation (cyan, PDB 1BU7), B) BMP A2 (brown, PDB ID
5DYP) and C) BMP WT crystallized in complex with NPG in the closed conformation and with the substrate removed before docking (magenta, PDB ID 1JPZ). (For interpretation of
the references to colour in this gure legend, the reader is referred to the web version of this article.)
112 G. Di Nardo et al. / Archives of Biochemistry and Biophysics 602 (2016) 106e115

Fig. 6. Differential scanning calorimetry (DSC) proles for A) BMP WT, B) BMP WT in complex with NPG, C) BMP WT in complex with ibuprofen, D) BMP A2, E) BMP A2 in complex
with NPG F) BMP A2 in complex with ibuprofen.

Table 3
DSC parameters for BMP WT and BMP A2 in the substrate-free form and in complex with NPG and the three drugs.

Sample Tm1 DH1 (kcal mol1) Tm2 DH2 (kcal mol1) Tm3 DH3 (kcal mol1)
BMP WT 55.8 0.3 8.3E4 5.3E3 61.4 0.1 7.4E4 5.1E3
BMP WT NPG 61.4 0.1 1.7E5 4.5E3 66.5 0.1 1.3E5 4.3E3
BMP WT Diclofenac 55.6 0.1 9.6E4 441 61.2 0.1 1.2E5 418
BMP WT Ibuprofen 55.2 0.1 1.0E5 673 60.9 0.1 1.3E5 634
BMP WT Tolbutamide 55.6 0.1 9.9E4 579 60.9 0.1 1.2E5 543
BMP A2 53.6 0.1 1.1E5 578 61.7 0.1 6.5E4 554
BMP A2 NPG 53.0 0.2 9.2E4 4.6E3 60.0 0.1 1.7E5 7.4E3 66.0 0.1 2.8E4 3.3E3
BMP A2 Diclofenac 53.8 0.1 1.3E5 668 61.6 0.1 1.5E5 643
BMP A2 Ibuprofen 54.6 0.1 1.5E5 819 62.1 0.1 1.4E5 830
BMP A2 Tolbutamide 54.1 0.1 1.4E5 534 62.2 0.1 1.5E5 524

Table 4
Reduction potentials, rst electron transfer rates in the absence of oxygen and presence of CO and NADPH consumption rates in the presence of oxygen for P450 BM3 WT and

Sample Em (vs NHE)/mV FeII(CO) complex formation rates (s1) NADPH consumption rate (mol NADPH/mol protein/min1)
(Only heme domain)

BM3 WT 387 2 k1 11.8 0.8 32.3 2.0

k2 1.0 0.2
BM3 A2 342 3 k1 19.8 0.9 129.0 1.0
k2 1.3 0.1
BM3 WT NPG 345 5 k1 23.4 1.2 451.6 3.0
k2 1.4 0.1
BM3 A2 NPG 326 6 k1 17.6 1.5 645.2 3.0
k2 1.2 0.1

leading to a conformational change known to decrease the exi- with the structure of BMP A2, this suggests a higher degree of
bility of the protein [21,22]. exibility in this protein as indicated by the presence of disordered
Based on the knowledge of the BMP A2 structure described in regions. The effect of substrate addition to A2 mutant on the DSC
the previous section, it might be expected that the DSC prole of A2 prole is shown in Fig. 5E. The apparent less resolved peak prole is
mutant is similar to that of the substrate-bound WT. Fig. 5D indeed due to the presence of a new component (Tm3) at 66.0  C, while Tm1
shows that the substrate-free A2 mutant displays two transitions and Tm2 remain comparable to that found in the substrate-free A2.
even better resolved than those of the substrate-bound WT. How- This is consistent with a more rigid protein conformation imposed
ever, the Tm1 and Tm2 values are found to be 53.6 and 61.7  C by the presence of the substrate as also observed for WT. However,
respectively, that are lower than the corresponding values found in substrate-bound A2, the presence of a exible component rep-
for substrate-bound WT. This can be ascribed to the absence of the resented by Tm1 still persists.
stabilizing interactions of A2 mutant with the substrate. Moreover, The effect of the three drugs on the DSC prole of BMP WT and
the comparison of Tm1 and Tm2 values of the substrate-free proteins, A2 was also tested. Fig. 5C and F shows the proles obtained from
show a lower Tm1 value for the substrate-free A2 mutant. In line BMP WT and BMP A2 in the presence of ibuprofen whereas Table 3
G. Di Nardo et al. / Archives of Biochemistry and Biophysics 602 (2016) 106e115 113

summarizes the results for the effect of the three drugs. The pres-
ence of the drugs does not affect the DSC proles of BMP WT as
expected, whereas a small stabilizing effect in terms of Tm and/or
DH is seen on BMP A2 in complex of the three drugs. This effect is
lower than in the case of BMP A2 with NPG, where at least part of
the protein was stabilized (Tm3). This is consistent with our dif-
culties to see a dened electron density map of the drugs in the
active site of the protein, where ligand exibility still persists. The
results are also consistent with a change of the conformational
equilibrium in BMP A2 rather than a mutation-induced conforma-
tional change.
Our data are also in line with the results from the crystal
structures of other mutants of cytochrome P40 BM3 with catalytic
new abilities compared to WT [10,39] where the effect of the mu-
tations is to promote a change in the conformational equilibrium
and the structural rearrangements that bring the protein to adopt a
conformation similar to the substrate-bound form of the WT pro-
tein. However, for those variants the conformational rearrange-
ments are induced either by single and double active site mutations
or by a combination of mutations far from the active site [10,39]. For
some of them, the catalytic improvements of the variants were
explained in terms of a decrease in the free energy barrier neces-
sary to the protein for the transition from the substrate-free to the
ebound conformations [10,39]. In fact, DSC experiments showed
that the addition of substrate induces only a small thermal stabi-
lization in the mutants (already in the closed conformation)
compared to the WT protein, where a stabilization by 5e6  C is
achieved by substrate addition upon the change from the open to
the closed conformation [10]. Our data data suggest that the
removal of a salt bridge in A2 mutant has an effect on the confor-
mation of the protein and, from the crystallographic data, this can
be explained with the increase in the conformational exibility of
the G-helix found in the mutant A2.

3.4. Impact of structure on P450 BMP A2 functional properties

The structure of the A2 mutant shows a rearrangement of the Fig. 7. Spectroelectrochemical titrations of the heme domain of P450 BM3 A2 in A)
water molecules within the active site cavity (Fig. 3). As this could substrate-free and B) NPG-bound forms. The black traces represent the initial fully-
affect the redox properties of the metal center, spectroelec- oxidised spectra whereas the red traces represent the fully-reduced nal spectra.
The insets show the plot of the percentage of oxidized fraction, calculated from the
trochemistry was used to determine the reduction potential of the absorbance changes at A) 419 and B) 398 nm as a function of the potential measured
mutant (Fig. 7). The results are summarized in Table 4. Though both and the tting to the Nernst equation with the number of electrons xed at 1. (For
WT and A2 show a shift of the reduction potential (Em) toward interpretation of the references to colour in this gure legend, the reader is referred to
positive values upon substrate binding, the Em value of substrate- the web version of this article.)
free A2 is within the error the same as that of the substrate
bound WT. This can be explained by the analysis of the x-ray
heme domain of P450 BM3 WT and A2. This is possible because the
structure of substrate-free A2 where the position of the water
reduced form of the heme iron atom is able to bind CO and form a
molecule is close to that found in the substrate-bound WT. Addition
complex with a typical absorption band at 450 nm.
of the substrate leads to a 42 mV shift toward positive values for the
Fig. 8 shows the results obtained from the two proteins in the
WT, while in the A2 mutant this leads to only a 26 mV shift. The
absence and presence of the substrate NPG. The increase at 450 nm
lower positive shift in reduction potential observed for the mutant
was better tted to a bi-exponential as a function of time that led to
A2 is also coherent with the different position of the water
the rate constants (k1 and k2) reported in Table 4. This behavior was
previously observed for BMP WT and other mutants and can be
All together these results can be explained by the partially off-
attributed to the presence of a spin equilibrium leading to two
axial position of the water molecule found in A2, as well as the
populations of protein in high and low spin, respectively (39). The
pattern of hydrogen bond involving the residues of the helix I and
rst rate for the mutant A2 is 2-folds higher compared to the WT
the water molecules (Fig. 3B) and, more in general, the changed
enzyme. When the substrate NPG is added, as expected, the elec-
structural environment compared to the WT protein. Moreover, the
tron transfer rate increases by 2-folds for the WT protein and this is
reduction potential of the mutant A2 moves toward even more
consistent with the shift of the reduction potential toward positive
positive values when the substrate NPG is added, consistent with
values previously observed. For A2, the value remains unchanged
the movement of the water molecule to a more distant and off-axial
upon substrate addition as the reduction potential is similar for the
position (Fig. 4B).
substrate-free and ebound proteins, suggesting that the substrate-
Since the substrate-induced shift of the reduction potential to-
free conformation of the enzyme is already prone to accept the
ward positive values triggers the rst electron transfer in the cat-
alytic cycle of cytochrome P450, stopped ow measurements were
The results suggest that the conformational change observed in
used to follow the rst electron transfer kinetics from NADPH to the
114 G. Di Nardo et al. / Archives of Biochemistry and Biophysics 602 (2016) 106e115

Fig. 8. Stopped ow kinetics of CO binding for A) BMP WT, B) BMP WT in complex with NPG, C) BMP A2 and D) BMP A2 in complex with NPG. In all cases, the black trace represents
the initial spectrum whereas the red trace represents the spectrum after 1 s. The insets represent the plots of the absorbance at 450 nm as a function of time and tting to double
exponential curves. (For interpretation of the references to colour in this gure legend, the reader is referred to the web version of this article.)

BMP A2 induce a shift of the reduction potential toward positive The benet of the mutations can be seen in terms of a higher
values favoring the rst electron transfer. This means that, when a degree of exibility of the mutant compared to WT that facilitates
substrate is accepted in the active site, the mutant is more ready the access of substrates also due to the reorientation of specic side
than the WT protein to go though the catalytic cycle. This is chains such as Phe87. The rotation of this residue is achieved in
conrmed by the results shown in Table 4, where the NADPH basal BMP A2 without altering the active site of the protein but with a
consumption of P450 BM3 WT and A2 in the presence of oxygen long-range effect of the Asp251Gly mutation. Once the substrates
measured at 340 nm is reported. As it can be seen, the rates of are accepted, the protein is in a catalytically ready conformation,
NADPH consumption is 4-folds higher in P450 BM3 A2 compared to where the structural rearrangements are key to trigger the rst
the WT protein, suggesting that electrons are owing through the events of the P450 catalytic cycle. Thus, the typical cascade of
system to oxygen faster than in WT, even if a substrate is not pre- events usually associated with substrate binding (i.e. conforma-
sent. This last feature is usually associated to the uncoupling of tional change, water displacement, shift of the reduction potential
reducing equivalents for reactive oxygen species production rather and rst electron transfer) that represents a ne control mecha-
than for product formation found in cytochromes P450 [1,46]. This nism in P450 BM3 WT is at least partially lost in mutant A2 that
may be the reason why the turnover rate for drugs in A2 mutant is behaves similar to a loaded gun catalyst. Indeed, as expected from
much lower than the NADPH consumption rate, suggesting a high a rst round of random mutagenesis, the mutations produced the
level of uncoupling [9]. Furthermore, these data conrm that a so-called generalist [50] that can be used for other rounds of
slower electron transfer is usually associated to a higher level of directed evolution to generate variants more specic toward the
coupling [46e49]. substrates of interest. As a matter of the fact, this variant, that was
also shown to turn over other substrates such as polycyclic aro-
4. Conclusions matic hydrocarbons, was already used as a template for a second
round of mutagenesis and variants with improved activities to-
In conclusion, the crystal structure of P450 BMP A2 together wards chrysene and pyrene were successfully generated [51].
with the functional data highlights the importance for catalysis of
mutations far from the active site but having long-range effects on Acknowledgments
it. From a structural point of view, the ability to recognize new
substrates can be achieved by a specic mutation in a plastic region This work was supported by the grant FIRB (project
of substrate access channel that becomes more exible and facili- RBFR12FI27_004, awarded to G. Di Nardo) and Progetto Ateneo-San
tates the change in the conformational equilibrium between the Paolo 2012 (project SADJATEN12, awarded to S. Sadeghi).
open to the closed conformation of the enzyme as well as substrate
entrance. The change in the conformational equilibrium is References
responsible for a series of events in the active site that trigger the
shift of the reduction potential toward positive values that facili- [1] P.R. Ortiz de Montellano, Plenum Publisher, Springer, New York, NY, 2015.
[2] F.P. Guengerich, Chem. Res. Toxicol. 22 (2) (2009) 237e238.
tates the electron ow through the system compared to the WT [3] G. Di Nardo, A. Fantuzzi, A. Sideri, P. Panicco, C. Sassone, C. Giunta, G. Gilardi,
protein. J. Biol. Inorg. Chem. 12 (3) (2007) 313e323.
G. Di Nardo et al. / Archives of Biochemistry and Biophysics 602 (2016) 106e115 115

[4] B.M. van Vugt-Lussenburg, E. Stjernschantz, J. Lastdrager, C. Oostenbrink, [28] P.R. Evans, G.N. Murshudov, Acta Crystallogr. D. Biol. Crystallogr. 69 (7) (2013)
N.P. Vermeulen, J.N. Commandeur, J. Med. Chem. 50 (3) (2007) 455e461. 1204e1214.
[5] M.C. Damsten, B.M. van Vugt-Lussenburg, T. Zeldenthuis, J.S. de Vlieger, [29] A. Vagin, A. Teplyakov, J. Appl. Cryst. 30 (1997) 1022e1025.
J.N. Commandeur, N.P. Vermeulen, Chem. Biol. Interact. 171 (1) (2008) [30] M.D. Winn, C.C. Ballard, K.D. Cowtan, E.J. Dodson, P. Emsley, P.R. Evans,
96e107. R.M. Keegan, E.B. Krissinel, A.G.W. Leslie, A. McCoy, S.J. McNicholas,
[6] A.M. Sawayama, M.M. Chen, P. Kulanthaivel, M.S. Kuo, H. Hemmerle, G.N. Murshudov, N.S. Pannu, E.A. Potterton, H.R. Powell, R.J. Read, A. Vagin,
F.H. Arnold, Chem. Eur. J. 15 (43) (2009) 11723e11729. K.S. Wilson, Acta Crystallogr. D Biol. Crystallogr. 67 (2011) 235e242.
[7] J. Reinen, J.S. van Leeuwen, Y. Li, L. Sun, P.D. Grootenhuis, C.J. Decker, [31] P. Emsley, K. Cowtan, Acta Crystallogr. D Biol. Crystallogr. 60 (2004)
J. Saunders, N.P. Vermeulen, J.N. Commandeur, Drug Metab. Dispos. 39 (2011) 2126e2132.
1568e1576. [32] G.N. Murshudov, A.A. Vagin, E.J. Dodson, Acta Crystallogr. D Biol. Crystallogr.
[8] A. Rentmeister, T.R. Brown, C.D. Snow, M.N. Carbone, F.H. Arnold, Chem- 53 (3) (1997) 240e255.
CatChem. 3 (6) (2011) 1065e1071. [33] P.D. Adams, P.V. Afonine, G. Bunkoczi, V.B. Chen, I.W. Davis, N. Echols,
[9] G.E. Tsotsou, A. Sideri, A. Goyal, G. Di Nardo, G. Gilardi, Chemistry 18 (12) J.J. Headd, L.W. Hung, G.J. Kapral, R.W. Grosse-Kunstleve, A.J. McCoy,
(2012) 3582e3588. N.W. Moriarty, R. Oeffner, R.J. Read, D.C. Richardson, J.S. Richardson,
[10] C.F. Butler, C. Peet, A.E. Mason, M.W. Voice, D. Leys, A.W. Munro, J. Biol. Chem. T.C. Terwilliger, P.H. Zwart, Acta Crystallogr. D Biol. Crystallogr. 66 (2010)
288 (35) (2013) 25387e25399. 213e221.
[11] H. Venkataraman, M.C. Verkade-Vreeker, L. Capoferri, D.P. Geerke, [34] G.M. Morris, D.S. Goodsell, R.S. Halliday, R. Huey, W.E. Hart, R.K. Belew,
N.P. Vermeulen, J.N. Commandeur, Bioorg. Med. Chem. 22 (20) (2014) A.J. Olson, J. Comput. Chem. 19 (1998) 1639e1662.
5613e5620. [35] H. Li, T.L. Poulos, Acta Crystallogr. D Biol. Crystallogr. 51 (1) (1995) 21e32.
[12] D.H. Kim, K.H. Kim, D.H. Kim, K.H. Liu, H.C. Jung, J.G. Pan, C.H. Yun, Drug [36] C.A. Hasemann, R.G. Kurumbail, S.S. Boddupalli, J.A. Peterson, J. Deisenhofer,
Metab. Dispos. 36 (11) (2008) 2166e2170. Structure 3 (1) (1995) 41e62.
[13] D.H. Kim, T. Ahn, H.C. Jung, J.G. Pan, C.H. Yun, Drug Metab. Dispos. 37 (5) [37] T.W. Ost, J. Clark, C.G. Mowat, C.S. Miles, M.D. Walkinshaw, G.A. Reid,
(2008) 932e936. S.K. Chapman, S. Daff, J. Am. Chem. Soc. 125 (49) (2003) 15010e15020.
[14] G. Di Nardo, G. Gilardi, Int. J. Mol. Sci. 13 (12) (2012) 15901e15924. [38] C.J. Whitehouse, S.G. Bell, H.G. Tufton, R.J. Kenny, L.C. Ogilvie, L.L. Wong, Chem.
[15] L.O. Narhi, A.J. Fulco, J. Biol. Chem. 261 (16) (1986) 7160e7169. Commun. Camb. 28 (8) (2008) 966e968.
[16] S.S. Boddupalli, T. Oster, R.W. Estabrook, J.A. Peterson, J. Biol. Chem. 267 (15) [39] C.J. Whitehouse, W. Yang, J.A. Yorke, H.G. Tufton, L.C. Ogilvie, S.G. Bell,
(1992) 10375e10380. W. Zhou, M. Bartlam, Z. Rao, L.L. Wong, Dalton Trans. 40 (40) (2011)
[17] V.B. Urlacher, M. Girhard, Trends Biotechnol. 30 (1) (2012) 26e36. 10383e10396.
[18] R. Bernhardt, J. Biotechnol. 124 (1) (2006) 128e145. [40] M.G. Joyce, H.M. Girvan, A.W. Munro, D. Leys, J. Biol. Chem. 279 (22) (2004)
[19] C.J. Whitehouse, S.G. Bell, L.L. Wong, Chem. Soc. Rev. 41 (3) (2012) 23287e23293.
1218e1260. [41] K.G. Ravichandran, S.S. Boddupalli, C.A. Hasermann, J.A. Peterson,
[20] S.T. Jung, R. Lauchli, F.H. Arnold, Curr. Opin. Biotechnol. 22 (6) (2011) J. Deisenhofer, Science 261 (5122) (1993) 731e736.
809e817. [42] H. Li, T.L. Poulos, Nat. Struct. Biol. 4 (2) (1997) 140e146.
[21] I.F. Sevrioukova, H. Li, H. Zhang, J.A. Peterson, T.L. Poulos, Proc. Natl. Acad. Sci. [43] J. Catalano, K. Sadre-Bazzaz, G.A. Amodeo, L. Tong, A. McDermott, Biochem-
USA 96 (5) (1999) 1863e1868. istry. 52 (39) (2013) 6807e6815.
[22] D.C. Haines, D.R. Tomchick, M. Machius, J.A. Peterson, Biochemistry 40 (45) [44] Modi, S.; Sutcliffe, M.J.; Primrose, W.U.; Lian, L.Y.; Roberts, G.C. (1996) 3(5):
(2011) 13456e13465. 414e417.
[23] F. Valetti, S.J. Sadeghi, Y.T. Meharenna, S.R. Leliveld, G. Gilardi, Biosens. Bio- [45] S.E. Graham, J.A. Peterson, Arch. Biochem. Biophys. 369 (1) (1999) 24e29.
electron. 13 (6) (1998) 675e685. [46] D. Degregorio, S.J. Sadeghi, G. Di Nardo, G. Gilardi, S.P. Solinas, J. Biol. Inorg.
[24] T. Omura, R. Sato, J. Biol. Chem. 239 (1964) 2370e2378. Chem. 16 (1) (2011) 109e116.
[25] D. De Sanctis, A. Beteva, H. Caserotto, F. Dobias, J. Gabadinho, T. Giraud, [47] V.R. Dodhia, A. Fantuzzi, G. Gilardi, J. Biol. Inorg. Chem. 11 (7) (2006) 903e916.
A. Gobbo, M. Guijarro, M. Lentini, B. Lavault, T. Mairs, S. McSweeney, [48] G. Di Nardo, S. Castrignano , S.J. Sadeghi, R. Baravalle, G. Gilardi, Electrochem.
S. Petitdemange, V. Rey-Bakaikoa, J. Surr, P. Theveneau, G.A. Leonard, Comm. 52 (2015) 25e28.
C. Mueller-Dieckmann, J. Synchrotron Radiat. 19 (3) (2012) 455e461. [49] S.J. Sadeghi, G. Gilardi, Biotechnol. Appl. Biochem. 60 (1) (2013) 102e110.
[26] J. Juanhuix, F. Gil-Ortiz, G. Cun, C. Colldelram, J. Nicola s, J. Lido
 n, E. Boter, [50] C.A. Tracewell, F.H. Arnold, Curr. Opin. Chem. Biol. 13 (1) (2009) 3e9.
C. Ruget, S. Ferrer, J. Benach, J. Synchrotron Radiat. 21 (4) (2014) 679e689. [51] A. Sideri, A. Goyal, G. Di Nardo, G.E. Tsotsou, G. Gilardi, J. Inorg. Biochem. 120
[27] W. Kabsch, Acta Crystallogr. D. Biol. Crystallogr. 66 (2) (2010) 125e132. (2013) 1e7.