X-ray crystal structure of human heme oxygenase-1 in complex with 1-(adamantan-1-yl)-2-(1H-imidazol-1-yl)ethanone: A common binding mode for imidazole-based heme oxygenase-1 inhibitors

Article abstract of DOI:10.1021/jm800505m

Development of inhibitors specific for heme oxygenases (HOs) should aid our understanding of the HO system and facilitate future therapeutic applications. The crystal structure of human HO-1 complexed with 1-(adamantan-1-yl)-2-(1H- imidazol-1-yl)ethanone (3) was determined. This inhibitor binds to the HO-1 distal pocket such that the imidazolyl moiety coordinates with heme iron while the adamantyl group is stabilized by a hydrophobic binding pocket. Distal helix flexibility, coupled with shifts in proximal residues and heme, acts to expand the distal pocket, thus accommodating the bulky inhibitor without displacing heme. Inhibitor binding effectively displaces the catalytically critical distal water ligand. Comparison with the binding of 2-[2-(4-chlorophenyl)ethyl]-2-[1H- imidazol-1-yl)methyl]-1,3-dioxolane (2) revealed a common binding mode, despite differing chemical structures beyond the imidazolyl moiety. The inhibitor binding pocket is flexible, yet contains well-defined subpockets to accommodate appropriate functional groups. On the basis of these structural insights, we rationalize binding features to optimize inhibitor design.

Full text of DOI:10.1021/jm800505m

J. Med. Chem. 2008, 51, 5943–5952
5943
X-ray Crystal Structure of Human Heme Oxygenase-1 in Complex with
1-(Adamantan-1-yl)-2-(1H-imidazol-1-yl)ethanone: A Common Binding Mode for
Imidazole-Based Heme Oxygenase-1 Inhibitors#
Mona N. Rahman,† Jason Z. Vlahakis,‡ Walter A. Szarek,‡ Kanji Nakatsu,§ and Zongchao Jia*,†
Departments of Biochemistry, Pharmacology & Toxicology, and Chemistry, Queen’s UniVersity, Kingston, Ontario K7L 3N6, Canada
ReceiVed May 1, 2008
Development of inhibitors specific for heme oxygenases (HOs) should aid our understanding of the HO
system and facilitate future therapeutic applications. The crystal structure of human HO-1 complexed with
1-(adamantan-1-yl)-2-(1H-imidazol-1-yl)ethanone (3) was determined. This inhibitor binds to the HO-1 distal
pocket such that the imidazolyl moiety coordinates with heme iron while the adamantyl group is stabilized
by a hydrophobic binding pocket. Distal helix flexibility, coupled with shifts in proximal residues and heme,
acts to expand the distal pocket, thus accommodating the bulky inhibitor without displacing heme. Inhibitor
binding effectively displaces the catalytically critical distal water ligand. Comparison with the binding of
2-[2-(4-chlorophenyl)ethyl]-2-[1H-imidazol-1-yl)methyl]-1,3-dioxolane (2) revealed a common binding mode,
despite differing chemical structures beyond the imidazolyl moiety. The inhibitor binding pocket is flexible,
yet contains well-defined subpockets to accommodate appropriate functional groups. On the basis of these
structural insights, we rationalize binding features to optimize inhibitor design.
Introduction
stances, such as diabetes, inflammation, heart disease, hyperten-
Heme oxygenases (HOsa) are involved in the regioselective,
oxidative cleavage of heme at the R-meso carbon to release
equimolar amounts of ferrous iron (Fe2+), carbon monoxide
(CO), and biliverdin, the last of which is transformed into
bilirubin by biliverdin reductase (Supporting Information Figure
S1).1,2 Indeed, this reaction is the main source of CO in
mammals; heme degradation accounts for ∼85% of the CO
produced in humans under normal physiological conditions. To
date, three HO isozymes have been identified. HO-1, the
inducible isoform, is a 32 kDa stress protein found to be induced
by a variety of stimuli including heat shock, heavy metals, heme,
and reactive oxygen species; HO-1 is predominantly expressed
in the spleen.3 The constitutive isoform, HO-2, is an ∼36.5 kDa
protein that is widely distributed and has its highest concentra-
tion in the brain and testes. A third isozyme, HO-3, shares ∼90%
homology with HO-2 but has been shown to be inactive.4
The role of HO in heme catabolism is well understood as is
its catalytic mechanism. However, there is increasing evidence
regarding the role of CO as a cellular regulator in the brain,
blood vessels, and immune systems. Indeed, the CO produced
by HO-1 has been shown to have a variety of cellular regulatory
actions including anti-inflammatory, antiapoptotic, antiprolif-
erative, and vasodilatory effects.5-8 CO has been shown to exert
many of its effects through the activation of soluble guanylyl
cyclase (sGC) and/or mitogen-activated protein kinase (MAPK)
signaling pathways.5,9-12 Furthermore, the HO system has been
found to have a protective role in a wide variety of circum-
sion, neurological disorders, organ transplantation, and endot-
oxemia.13 HO-1 is also a potential target for cancer therapy
because the growth of most tumors is dependent on HO-1
activity.14 Inhibition of HO-1 using a metalloporphyrin inhibitor
(pegylated zinc protoporphyrin) has been associated with
antitumor activity15 as well as enhancement of the effects of
contemporary anticancer drugs.16 Thus, there is significant
interest in investigating and understanding the CO/HO system.
The development of inhibitors selective for the HO enzymes
aims to provide powerful tools for the dissection of the CO/
HO system as well as the mechanisms underlying its physi-
ological effects in various pathologies, in addition to future
therapeutic applications. Moreover, they may serve as candidates
for novel therapies against these diseases. To date, studies have
exploited HO inhibitors belonging to the metalloporphyrin class
(e.g., tin/zinc/chromium protoporphyrin). However, their close
structural similarity to heme results in some nonselectivity
because heme plays a key role in a number of biologically
relevant enzymes including cytochromes P450, nitric oxide
synthase, and soluble guanylyl cyclase. Thus, there have been
concerns regarding the validity of the interpretation of the results
of these studies.17-19 The discovery of azalanstat (1) (Figure
1), a compound not based on the porphyrin nucleus, as an HO
inhibitor led to a program in our laboratory concerned with
systematically modifying the structure of 1, resulting in a series
of novel compounds, several of which are HO-1 inhibitors.20-23
In continuation of our program, a structural study of complexes
formed between human HO-1 and some of our selective HO-1
inhibitors was initiated. During the course of these studies, the
results of an X-ray crystallographic study of one of our
compounds, namely, (2-[2-(4-chlorophenyl)ethyl]-2-[(1H-im-
idazol-1-yl)methyl]-1,3-dioxolane (2) were reported (Figure 1).24
In that study, compound 2 was cocrystallized with heme-
conjugated rat HO-1, and it appeared that compound 2 inhibited
HO-1 activity by binding to the distal side of the heme iron,
thus preventing the oxidation of Fe2+ by disrupting a critical
#
The hHO-1 complex structure has been deposited into the Protein Data
Bank (PDB code 3CZY).
* To whom correspondence should be addressed. Phone: (613) 533-6277.
Fax: (613) 533-2497. E-mail: jia@queensu.ca.
†
Department of Biochemistry.
Department of Chemistry.
Department of Pharmacology & Toxicology.
‡
§
a Abbreviations: HO, heme oxygenase; sGC, soluble guanylyl cyclase;
MAPK, mitogen-activated protein kinase; PMA, phosphomolybdic acid;
CHESS, Cornell High Energy Synchrotron Source; MR, molecular replace-
ment; HP, hydrophobic pocket.
24
hydrogen-bond network and interfering with O2 binding. In
10.1021/jm800505m CCC: $40.75
2008 American Chemical Society
Published on Web 09/18/2008

5944 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 19
Rahman et al.
4 was obtained by a nucleophilic displacement reaction on
1-bromo-4-(4-bromophenyl)-2-butanone22 using 1H-1,2,4-tria-
zole in N,N-dimethylformamide. The compounds were isolated,
characterized, and evaluated as the hydrochloride salts.
Initial screens using a CO formation assay with rat spleen
microsomal fractions gave IC50 values of 7 ( 1 µM for 3
and 4 ( 2 µM for 2.21 The binding of 3 to heme-conjugated
recombinant hHO-1 was examined by spectral analysis
(Figure 2A). In the absence of inhibitor, the heme-bound
hHO-1 gave a characteristic spectrum with a Soret peak at
404 nm. With increasing amounts of 3 (5-50 µM), the Soret
peak shifted to longer wavelengths up to 412 nm, at which
point it was presumed to have been saturated with inhibitor
as seen with the rat HO-1 complexed with 2.24 Furthermore,
a more prominent shoulder centered at 355 nm was apparent.
Secondary peaks centered at 535 and 560 nm increased with
increasing concentrations of inhibitor, while a third minor
peak at 630 nm decreased and then disappeared at high
inhibitor concentrations. Parallel experiments using inhibitor
2 gave similar results. The shift in the Soret peak was
suggestive of a change in the heme environment and, thus,
putative binding of the inhibitor. Heme degradation in the
presence of increasing amounts of inhibitor 3 was then
initiated by the addition of L-ascorbic acid (1 mM) and
monitored for 90 min. Under these conditions, the reaction
reached completion within 60 min in the absence of inhibitor
(data not shown). This rate of reaction was chosen to identify
differences in inhibition that might otherwise not be observ-
able to any significant extent at a faster rate. Initial rates
were determined for each condition by calculating the rate
of change of absorbance over the initial 3 min, according to
its respective Soret peak. As seen in Figure 2A, the Soret
peak shifts maximally at an inhibitor concentration of 25 µM.
At this concentration, the initial rate of reaction (as measured
at 412 nm) was diminished to 41 ( 7% of the control rate
without any inhibitor (as measured at 404 nm) (Figure 2B).
As a comparison, the rate was inhibited to 15 ( 4% of control
by 2 at this concentration. The Soret peak shifted to 412 nm
at a lower concentration of 2 (10 µM) at which the initial
rate of reaction was 39 ( 2% that of the control. Spectral
analysis following the heme degradation assay revealed the
disappearance of the Soret peak in the absence of inhibitor
(Figure 2C). As another means of measurement, the height
of the Soret peak following the reaction was compared to
that in a control experiment in which ascorbic acid was
absent. The absorbance of the Soret peak in the absence of
inhibitor fell to 24 ( 1%, whereas the height only diminished
to 44 ( 5% and 54 ( 5% in the presence of 25 and 50 µM
of inhibitor 3, respectively, further indicating inhibition of
heme degradation. In the presence of inhibitor 2 at these
concentrations, the absorbance only fell to 65 ( 4% and 80
( 5%, respectively.
Figure 1. Structures of 1, 2, and 3 depicting regions of interest.
Scheme 1a
a Reagents and conditions: (a) imidazole, N,N-dimethylformamide, room
temp, 7 days; (b) 37% aqueous HCl, EtOH; (c) 1H-1,2,4-triazole, N,N-
dimethylformamide, K2CO3, room temp, 2 h; (d) 37% aqueous HCl,
2-propanol.
contrast, the metalloporphyrins would compete with heme for
binding to the enzyme.
In the present study, we report the crystal structure of human
HO-1 in complex with heme and the imidazole-based inhibitor
1-(adamantan-1-yl)-2-(1H-imidazol-1-yl)ethanone (3) (Figure 1).
It is recognized that the structure of 3 is different from that of
2, yet the ease of formation of the complex of 3 with human
HO-1 prompted us to undertake this X-ray crystallographic
study. Interestingly, in spite of the structural differences between
3 and the imidazole-dioxolane series of inhibitors, it would
appear that there is a common mode of binding, and this
information may be useful in the design of more potent inhibitors
of HO-1. Our results, in conjunction with those obtained from
the previously reported24 structural study of the complex
between rat HO-1 and compound 2, indeed suggest a common
binding mode for our series of structurally related imidazole-
based inhibitors20-23 on the distal side of the heme, in which
the flexible distal helix of HO-1 opens up to accommodate the
bulky inhibitor. The imidazolyl group serves as an anchor by
coordinating with the iron moiety, while the bulky adamantyl
group fits into a hydrophobic pocket and is stabilized by
hydrophobic interactions. Details of the binding mode were
assessed in order to identify factors that could be useful in the
future structure-based design of more effective HO-1 inhibitors
based on the consideration of different functional groups within
the binding pocket.
Crystallization and Structure Determination of hHO-1
in Complex with Inhibitor 3. In order to confirm binding of
inhibitor 3 prior to attempting cocrystallization trials, spectral
analyses were undertaken on both the native protein and that
incubated with 3 at a molar ratio of 1:3. After a 15 min
incubation of the native protein with inhibitor, the expected shift
in the Soret peak from 404 to 412 nm was observed, thus
verifying binding. This mixture was subsequently utilized for
crystallization trials based on previously known conditions for
the native protein.
Results and Discussion
Synthesis and Characterization of Inhibitor 3. In this
article, we report the synthesis of compounds 3 and 4-(4-
bromophenyl)-1-(1H-1,2,4-triazol-1-yl)-2-butanone (4). As shown
in Scheme 1, compound 3 was synthesized by a nucleophilic
displacement reaction on commercially available 1-adamantyl
bromomethyl ketone (Sigma-Aldrich) using imidazole in N,N-
dimethylformamide. By use of a similar approach, compound
Cocrystals of hHO-1 bound to 3 were reminiscent of native
hHO-1 crystals, growing as clusters of thin plates, a feature that

Structure of hHO-1 in Complex
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 19 5945
Figure 2. Spectral analysis of 3 binding to hHO-1. (A) Heme-conjugated hHO-1 (10 µM) in 20 mM potassium phosphate (pH 7.4) was incubated
with increasing concentrations of 3 (blue) or 2 (green) for 5 min at room temperature. Absorbances were measured over a range of 300-700 nm
at intervals of 1 nm, and values were corrected for buffer (20 mM potassium phosphate, pH 7.4). The assays were performed in triplicate and the
values averaged. The wavelength corresponding to the Soret peak for each spectrum is indicated. The Soret peak gradually shifted from 404 to 412
nm with increasing concentrations of 3 or 2, with a more prominent shoulder centered at 355 nm. Secondary peaks centered at 535 and 560 nm
were observed to be increased with increasing concentrations of inhibitor, while a third minor peak at 630 nm was observed to decrease until it was
no longer detectable at high inhibitor concentrations. (B) Heme degradation rates in the presence of inhibitor 3. Heme degradation was subsequently
initiated by the addition of 1 mM L-ascorbic acid and allowed to proceed for 90 min at room temperature. Absorbances were measured at 404, 408,
410, and 412 nm at 90 s intervals. Initial rates were determined for each condition by calculating the rate of change of absorbance over the initial
3 min, at the wavelength corresponding to its respective Soret peak. For each replicate, values were normalized to the respective control condition
and subsequently averaged. Parallel reactions were also performed for heme-conjugated hHO-1 in the absence of both inhibitor and electron donor
(L-ascorbic acid) as a negative control (i.e., no oxidative degradation). Shaded bars depict rates in the presence of 3; open bars depict rates in the
presence of 2. (C) Spectral analysis following heme degradation. Absorbances were measured and analyzed as described in (A). Heme degradation
was observed as the disappearance of the Soret peak. Increasing concentrations of inhibitor resulted in increasing attenuation of the loss of the Soret
peak as well as of the secondary peaks at 535 and 560 nm but the appearance of a peak at 699 nm.
necessitated careful and challenging separation in order to obtain
single crystals. Diffraction data were obtained to a resolution
of 1.54 Å for the cocrystal which belongs to the P21 space group.
Because of the new crystal form, molecular replacement was
carried out using the native heme-conjugated hHO-1 (PDB code
1N3U) from which heme had been omitted. Two unambiguous
solutions were obtained corresponding to the two molecules in
the asymmetric unit. The structure of the heme-conjugated
hHO-1 bound to 3 was refined to an R factor of 0.19 and a free
R factor of 0.23. The Ramachandran plot showed that 100% of

5946 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 19
Table 1. Diffraction and Refinement Statisticsa
Rahman et al.
space group
P21
a (Å)
52.00
b (Å)
52.45
c (Å)
74.49
ꢀ (deg)
89.99
2
34
0.623
50-1.54
374747
57555
97.1 (82.8)
36.5 (4.4)
0.086 (0.356)
6.5 (3.7)
molecules in the asymmetric unit
solvent content (%)
mosaicity (deg)
resolution range (Å)
total reflections
unique reflections
completeness (%)
I/σ
b
Rmerge
average redundancy
Refinement Statistics
σ cutoff for refinement
none
0.194
0.226
54625
2899
4115
0.010
1.217
536
c
Rcrys
c
Rfree
number of reflections used
number of reflections in test set
number of non-hydrogen atoms used in refinement
rmsd bond lengths (Å)
rmsd bond angles (deg)
number of water molecules
ligand occupancy
1.00
Ramachandran Plot
most favored (%)
95.1
4.9
0.0
additional allowed (%)
generously allowed (%)
disallowed regions (%)
Figure 3. (A) Ribbon diagram of the inhibitor binding site of heme-
conjugated hHO-1 in complex with 3 at 1.54 Å resolution. Heme (light
orange) and 3 (yellow) are depicted as stick models. An omit map
contoured at 3.5σ is superimposed. Dashed lines indicate coordination
of imidazole nitrogens of 3 and His25 with the heme Fe. Residues
involved in inhibitor binding are indicated (black) as are secondary
residues (blue) involved in the Asp140-hydrogen bond network (Tyr58,
Tyr114, Arg136, Asn210),25 and those adjacent to the coordinating
His25 residue (Glu29, Gln38, Phe207) which were found to shift in
order to accommodate both heme and the inhibitor. The Thr135 residue
found to be involved in a hydrogen-bond network with a water molecule
and the dioxolane group of 2 in its cocrystal structure with rat HO-1
(PDB code 2DY5)24 is also indicated. Only water molecules in the
distal pocket have been included for simplicity. Backbone monomer
traces illustrate the structural alignment of 3 bound to hHO-1 (green)
with (B) native hHO-1 holoenzyme (PDB code 1N45, purple) and with
(C) 2 bound to rat HO-1 (PDB code 2DY5, cyan). Alignment was
performed by (B) fitting all atoms or by (C) secondary structure
matching (SSM).
0.0
a Values in parentheses are for the outermost shell (1.60-1.54 Å). b Rmerge
) ∑|Iobs - I |/∑Iobs, where Iobs is the intensity measurement and I is the
mean intensity for multiply recorded reflections. c Rcryst and Rfree ) (∑|Fobs
Fcalc|)/(∑|Fobs|) for reflections in the working and test sets, respectively.
-
the residues were in the allowed region. The final statistics are
given in Table 1, and the structure has been deposited in the
Protein Data Bank (3CZY).
Overall Structure of hHO-1 Bound to 3. The two molecules
in the asymmetric unit (A and B) were found to be virtually
identical with a CR rmsd of 0.06 Å (upon alignment of all the
atoms). The maximum displacement found is in the flexible,
distal helix: Gly144 with 0.39 Å and Gly143 with 0.11 Å. There
were also slight differences in the number and placement of
water molecules near the active site. Given the closeness in
structure of the two molecules in the asymmetric unit, further
discussion will focus on the A chain unless otherwise stated.
The overall structure of the hHO-1 complex with 3 was very
similar to that of the native heme-conjugated hHO-1,25 being
mostly R-helical with the heme sandwiched between the
proximal and distal helices (parts A and B of Figure 3,
Supporting Information Figure S2). Inhibitor 3 was bound to
the distal side of the heme, as clearly demarcated by the electron
density map, with the imidazole ring coordinating the heme iron,
i.e., serving as an anchor, at a distance of 2.06 Å (Figure 3A).
Alignment of the structure with that of the complex of 2 with
the rat HO-1 shows that they are also very close with an rmsd
of 0.76 Å (Figure 3C). As seen by the electrostatic surface
representation of the molecule, the adamantyl moiety is bound
within a hydrophobic pocket that extends back toward the distal
side of the heme binding pocket (Figure 4A). Residues lining
the distal pocket, which may contribute to hydrophobic interac-
tions, include Phe37, Phe167, Leu54, Met34, and Leu147. A
detailed table of residues within van der Waals distance of the
inhibitor, which may contribute to its binding, is listed in
Supporting Information Table S1. Compared to the native
structure and that of the rat HO-1 bound to 2 (parts B and C of
Figure 4, respectively), the distal pocket into which 3 binds
appears more open than that of either the native or the rat HO-1
bound to 2. The carbonyl group of Gly139 is close enough to
form hydrogen bonds with the coordinating nitrogen of the
imidazole group (2.68 Å) as well as weakly interacting with
the other nitrogen (3.68 Å), if they were protonated. However,
given that the neighboring His25 is a neutral imidazole,26 it is
likely that this is also the case for the imidazole ring of the
inhibitor in this environment, so there would be no hydrogen
bonds possible. Moreover, there appears to be a secondary
hydrophobic pocket adjacent to that binding the adamantyl group
which is not as prominent in either the structure of rat HO-1
bound to 224 or the native structure.25,27
Changes in the Distal Helix. Comparison of the inhibitor-
bound structure to the high resolution (1.5 Å) structure of the
presumably more active, “closed” conformation of hHO-125
revealed very little change in overall structure with a CR rmsd
of 0.74 Å. The maximum displacement of 3.92 Å (Gly144)
corresponds to changes in the distal helix within the heme
binding pocket (Figure 3B). Relative to the native structure, in

Structure of hHO-1 in Complex
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 19 5947
Leu138 (2.67 Å). Furthermore, there is a resultant break in the
distal helix from Ser142 to Gly144 implying a greater flexibility
in this region. A comprehensive analysis of the temperature
factors in this region revealed higher than average values for
both the native and inhibitor 3-complexed structures of hHO-
1, as well as for rat HO-1 in complex with 2. However, a
difference plot in which the normalized values of the closed,
i.e., putatively more active, conformation of the native structure
(PDB code 1N45, chain A) were subtracted, revealed no
significant increase in the thermal factor values in the distal
helix (data not shown). Thus, we conclude that the distortion
of the helix in order to accommodate the bulky inhibitor could
be accomplished by the inherent helical flexibility, without any
further increase in flexibility. It should be noted that in the
structure of the native, heme-conjugated HO-1, Ser142 is
involved in stabilizing distal helix distortion (i.e., a kink) via
hydrogen-bonds with the peptide backbones of Gly143 and
Leu141, a distortion that is maintained in the apo-form of the
enzyme.25 Close comparison of hydrogen-bonding patterns with
the native structure in this region reveals the loss of backbone
hydrogen-bonds involving Ser142, Gly143, Gly144, the amide
groups of Leu147 and Lys148, as well as the carbonyl groups
of Leu141 and Asp140. Moreover, potential hydrogen-bonds
form with water molecules involving the carbonyl groups of
Leu141 (2.69 Å) and Asp140 (2.70 Å and 2.71 Å), as well as
the amide group of Gly143 (3.12 Å). This last water molecule,
which interacts with both Gly143 and Asp140, is situated 4.08
Å from the active nitrogen on the imidazolyl group. In the B
molecule, this water may be close enough to form a weak
hydrogen-bond (3.86 Å) that may contribute to the stabilization
of the imidazolyl group. Interestingly, the structure of rat HO-1
bound to 224 also resulted in the loss of hydrogen-bonds
involving Gly139, Asp140, Leu141, Ser142, Gly143, Gly144,
Gln145, Leu147, Lys148, and Lys149, yet did not result in
breakage of this distal helix, implying that the breakage was
caused by the need to accommodate the bulkier adamantyl group
of 3. Thus, there is an inherent flexibility in the distal helix, as
already seen with the open and closed forms of the holo- and
apo-enzymes,25,27 that appears to be the basis by which 2 and
3 are able to bind, noncompetitively, to the heme-binding pocket.
Changes within the Distal Pocket. Within the hydrophobic
pocket that contains the western region (Figure 1) of 3, the side
chain of Leu147 is shifted by 2.58 Å, relative to the native
structure, while the phenyl moiety of Phe37 tilts toward the
adamantyl group by 1.3-1.4 Å (Figure 3A). Moreover, both
the heme moiety and the coordinating His25 residue are pushed
back to further open the distal pocket, with the His25 shifted
by 0.91 Å while the Fe moiety is shifted by 0.85 Å. As a result,
the distance between the coordinating nitrogen of His25 and
Fe is increased to 2.04 Å from 1.99 Å. Similarly, the His25-Fe
distance in the B molecule of the inhibitor complex is increased
to 2.01 Å from 1.90 Å. Adjacent to the coordinating His25, the
phenyl group of Phe207 is also shifted back by 0.4-0.75 Å,
while the side chains of Glu29 and Gln38 residues in the
proximal side of the heme-binding pocket are shifted back by
1.35 and 2.61 Å, respectively, maintaining the hydrogen-bond
between Glu29 and His25 at a distance of 2.83 Å but
lengthening the potential hydrogen-bond distance between it and
the His25 carbonyl from 2.68 to 2.97 Å. The shift in Gln38
results in it not being able to directly form a hydrogen-bond
with Glu29, similar to what is seen in the apo-structure of HO-1
as well as the open form of the holo-enzyme.25
Figure 4. Electrostatic surface potentials of (A) the complex of in-
hibitor 3 with hHO-1 revealing a hydrophobic pocket into which the
adamantyl moiety fits. Comparison of the electrostatic surface potentials
of (B) the native human HO-1 holo-enzyme (PDB code 1N45) and
(C) the complex of 2 with rat HO-1 (PDB code 2DY5) reveals a larger
distal pocket in the enzyme bound to 3. Blue and red colors indicate
positive and negative electrostatic potentials, respectively, as calculated
using PyMOL. The adamantyl moiety of 3 binds to the hydrophobic
pocket in a manner similar to that of the chlorophenyl moiety of 2. A
secondary hydrophobic pocket is also more prominent in hHO-1 bound
to 3.
the presence of 3, the distal helix is shifted outward from Ser142
to Ile150 to open up the hydrophobic distal pocket. While the
shift maintains hydrogen bonding of the alcohol moiety of
Ser142 with the ammonium moiety of Lys179 (2.72 Å in the
native to 2.79 Å) in the ligated complex, it appears to shift closer
to the inhibitor by 1.97 Å. As a result, hydrogen-bonds between
the Ser142 alcohol and the heme propionate moiety (3.05 Å)
and a solvent water molecule (3.14 Å) are lost, while a new
hydrogen bond would be formed with the carbonyl group of
Changes in Heme Structure. The heme in the native
structure of hHO-1 is puckered, while that in the inhibitor

5948 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 19
Rahman et al.
Table 2. Inhibitory Potency of Imidazole-Dioxolanes, Imidazole-Ketones, and Imidazole-Alcohols
IC50 (µM)a
substituent in the
western region
electronegativityb of
halogen atom
dioxolane
ketone
hydroxyl
p-fluorophenyl
p-chlorophenyl
p-bromophenyl
p-iodophenyl
phenyl
4.0
3.0
2.8
2.5
3.8 ( 1.1c
4.3 ( 2.1d (2)
1.9 ( 0.2c
3.7 ( 0.9c
0.7 ( 0.4c
2.7 ( 0.9c
1.4 ( 1.1c
0.5 ( 0.1 µMc
0.14 ( 0.06c
0.06 ( 0.03c
6.2 ( 0.8c
4.7 ( 0.5c
1.7 ( 0.7c (5)
0.11 ( 0.06c
4.0 ( 1.8c
adamantyl
3 ( 1 µM (3)
a IC50 values calculated from CO activity assays using spleen (HO-1) microsomal preparations derived from rats. b Electronegativity values as derived by
Linus Pauling.30 c Values obtained from Roman et al.22 d Value obtained from Vlahakis et al.21
complex is more planar. As a result there is an “uneven” shift
of the R-, ꢀ-, δ-, and γ-meso-carbons by 0.82, 2.22, 0.67, and
0.45 Å, respectively. Interestingly, the structure of a D140A
mutant of human HO-1 was also found to have a “flatter” heme
than the wild-type structure.28 Heme ruffling has been proposed
to increase spin density on the meso carbon atoms to promote
electrophilic attack by the ferric hydroperoxo intermediate.29
Thus, similar to the D140A mutant of HO-1, the binding of
inhibitor may alter the ability of HO-1 to adopt the proper heme
geometry and electronic state for the rapid electrophilic addition
of the peroxide OH group to the R-meso carbon.28 Of note, in
the refinement of the rat HO-1 structure with inhibitor 2, the
structure of the tetrapyrrole moiety of heme was restrained in
a planar conformation.24 It would be interesting to determine
whether this change in heme geometry is also a consequence
of binding to 2 by refinement without these additional restraints.
Central Region of the Inhibitor. The critical Asp140 residue
involved in the hydrogen-bonding network that anchors the
heme distal water ligand deemed essential for catalysis is not
perturbed by the binding of 3; each of the oxygen atoms of the
carboxylic acid moiety was shifted by only ∼0.2 Å.25 In the
native enzyme, when heme binds, the active site closes and
critical water residues are trapped to form this well-ordered
solvent structure. The hydrogen-bond network also involves
Asn210 and Arg136 to stabilize the critical water residue, with
a second tier including Tyr58 and Tyr114.27 The carbonyl group
of Gly139 was also observed to hydrogen-bond with the distal
water ligand. Binding of 3 does not perturb the positioning of
any of these residues (Figure 3A). However, it does displace
the hydrogen-bonded water network such that the distal water
ligand is displaced, thus presumably inhibiting heme oxidation,
in a manner similar to that proposed for 2.24 Close inspection
of the electrostatic surface shows that, in contrast to the other
structures, a potential channel is apparent alongside Asp140 that
is proposed to be involved in proton shuttling via a hydrogen-
bonded network of water molecules (Figure 4). The opening of
such a channel-like passage may also disrupt the ability of HO-1
to trap water molecules in the heme-binding site. This would
be compounded further by the more open conformation of the
distal pocket.
associate weakly with the ketone group of 3 (3.83 Å). This water
molecule is also close to the amino group of Arg136 (3.40 Å)
and a carboxyl group of Asp140 (3.78 Å). The differences
between molecules A and B do diminish the importance of this
hydrogen-bond network to the binding of this type of inhibitor
to HO-1. Thus, the presence of a ketone or dioxolane group in
the central region of the inhibitor may not be a significant mode
of binding for this type of inhibitor to HO-1. This latter point
was further emphasized upon comparison of the potency of a
variety of compounds (i.e., 2-oxy-substituted 1-(1H-imidazol-
1-yl-4-phenylbutanes), which had been screened using an in
Vitro assay for CO production using microsomal preparations
obtained from rat spleen.22 Manual docking of a few of these
compounds in the hHO-1 structure confirmed that their binding
would be superimposable to that of 2 and 3 (data not shown).
A correlation of the IC50 values obtained (Table 2), as well as
those similarly obtained for 2 and 3, revealed that, generally,
derivatives containing a hydroxyl in the central region of the
molecule were more potent inhibitors than their ketone and
dioxolane counterparts.22 This may be due to the increased
hydrogen-bonding potential of the alcohol group to form
hydrogen bonds with trapped waters, as well as other residues
in the region. For example, its increased rotational flexibility
may allow it to be close enough to form a hydrogen-bond with
Asp140 and further disrupt the hydrogen-bond network. Given
these observations, it may be possible to enhance the potency
of the inhibitor by introducing a hydrogen-bond donor func-
tionality that would form a more stable interaction with Asp140,
disrupt the hydrogen-bond network, and thus stabilize binding.
Western Region of the Inhibitor. A cursory comparison of
various substituents on the phenyl moiety in the western region
revealed that, at least for the alcohol derivatives, potency is
inversely correlated with electronegativity of the halogen
substituent (Table 2).30 This is to be expected given that this
region would bind within the hydrophobic pocket, although the
presence of the polar thiol group of Met34 therein may explain
why electronegative derivatives in this region are still potent.
The observed inhibitory potencies reflected by the electrone-
gativity trend are not as apparent in the dioxolane/ketone series
of compounds and may be a result of the increased rigidity of
these derivatives. An alternative explanation could be that the
larger halogens provide more and better contact. Finally, despite
the hydrophobicity of the adamantyl moiety of 3, its potency is
less than that of 2 (Figure 2). Comparison of the two structures
reveals that the adamantyl group does not extend as far into
the hydrophobic pocket as the chlorophenyl group (Figures 3C
and 4). As such, increasing the size of the linker between the
ketone and the adamantyl group may increase the potency of
this compound.
In the rat HO-1 structure with 2, a water molecule hydrogen-
bonded to the carbonyl of Thr135 is also weakly bonded to the
oxygen atom of the dioxolane group of 2. However, in the
human structure with 3, the closest water molecule to the ketone
functionality is situated 4.02 Å away. It should be noted that
the water molecules in this region are probably not static,
perhaps because of opening of the cavity to accommodate
inhibitor 3. The fluidity can be seen by comparison of this
structure (A molecule) with that of the B molecule which does
contain a similar water molecule hydrogen-bonded to the Thr135
carbonyl (2.89 Å), which subsequently forms a hydrogen-bond
network with another water molecule (2.67 Å) that may
Secondary Hydrophobic Pocket. The secondary hydropho-
bic pocket that is apparent upon binding of 3 (Figure 4) may
be the region in which the thio-(4-aminophenyl) moiety of

Structure of hHO-1 in Complex
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 19 5949
bond involving a hydrogen-donating group in this position.
Moreover, given the flexibility of the distal helix, it may also
be possible to take advantage of a potential hydrogen-bond with
the hydroxyl group of Ser142. Of note, the distance between
the Cꢀ of Ser142 and the carbon at position 5 of the imidazole
ring is 4.03 Å. Merely rotating the hydroxyl group of Ser142
toward the imidazole ring would allow it to be close enough to
form a strong hydrogen-bond. Thus, modification of the
imidazolyl moiety to enable this hydrogen-bond formation may
be a means by which the potency of the inhibitor can be
enhanced. One possibility is the introduction of an amino
substituent at position 5 of the imidazole ring, alternatively,
replacement of the imidazole ring by a 1,2,4-triazole ring in
which the C atom at position 5 of the imidazole is replaced by
an N atom. Indeed, we have recently developed a 1,2,4-triazole
derivative, namely, 4-(4-bromophenyl)-1-(1H-1,2,4-triazol-1-yl)-
2-butanone (4) (Scheme 1), that takes advantage of the afore-
mentioned hydrogen-bond. As a result, compound 4 exhibited
a greater than 4-fold increase in potency compared with its
imidazole analogue,22 namely, 4-(4-bromophenyl)-1-(1H-imi-
dazol-1-yl)-2-butanone (5) (IC50 ) 0.39 ( 0.07 µM compared
to 1.7 ( 0.7 µM, respectively) (Supporting Information Fig-
ure S3).
Figure 5. Model of 1 bound to hHO-1. Virtual docking simulation of
1 was performed using AutoDock (version 3.0). The macromolecule
used for docking was the A chain of hHO-1 in complex with 3, in
which 3 had been omitted. Electrostatic surface potentials were
calculated for hHO-1 bound to 3 as described in Figure 4. Blue and
red colors indicate positive and negative electrostatic potentials,
respectively. The imidazole moiety coordinates with the heme iron atom,
while the chlorophenyl moiety interacts with the primary hydrophobic
pocket (1° HP) as with compounds 2 and 3, while the thio-(4-
aminophenyl) group fits easily into the secondary hydrophobic pocket
(2° HP) as postulated. The coordinating His25 residue of hHO-1 is
shown.
Conclusion
Inhibitor 3 binds to hHO-1 in a manner similar to that of
inhibitor 2 to rat HO-1 despite very different structural features
of the inhibitors. Given this, as well as information from
modeling other imidazole-containing compounds (data not
shown), it appears that all these imidazole-containing com-
pounds bind to HO-1 using a common binding mode. The
flexibility of the hHO-1 distal helix allows the heme-binding
pocket to expand to accommodate binding of the inhibitor
without displacement of the heme moiety, hence in a noncom-
petitive manner; indeed, a helical break can ensue in the distal
helix to accommodate bulkier groups such as the adamantyl
moiety. Shifts in the proximal side of the pocket, as well as
with heme itself, also contribute to the opening of the distal
pocket. As a result, the solvent structure involving Asp140
becomes less ordered. The imidazole group of the inhibitor
serves as an anchor with the nitrogen at position 3 coordinating
with the heme iron atom, in a manner similar to that of the
imidazole group of His25, to result in a hexacoordinate heme
Fe. The western region of the inhibitor fits into a hydrophobic
pocket on the distal side of the heme-binding pocket, to further
stabilize the binding. The central region, while close to the
critical Asp140 residue, does not perturb the position of this
residue. Rather, inhibition of heme oxidation appears to occur
via disruption of the ordered hydrogen-bond network and
displacement of the critical water residue required for catalysis
by the inhibitor. Further, a secondary hydrophobic pocket may
serve as a binding pocket for substituents attached to the C-4
of the dioxolane ring in the imidazole-dioxolane series of
compounds. The insights gleaned from the present study may
be useful in the design of more potent inhibitors of HO-1.
compound 1 analogues interacts.20,21 The region contains two
residues containing polar groups, Ser53 and Met111, that have
been replaced by Ala in hHO-2.31 This may, in part, explain
the selectivity of these analogues for HO-1 over HO-2.20 Thus,
while this proximal side of the heme-binding pocket tends to
be more rigid than the distal pocket, there is still some flexibility
that results in the opening up of the secondary hydrophobic
pocket to accommodate such a bulky group in the thio-(4-
aminophenyl) region of 1. Indeed, virtual docking simulation
of 1 in the inhibitor binding site of hHO-1 supports this putative
interaction (Figure 5).
Implication for Structure-Based Inhibitor Design. As
indicated in the Introduction, our initial series of inhibitors are
based on 1, a compound containing an imidazole-dioxolane
structural feature. It appears that the results obtained in this study
could be useful for the design of such analogues, particularly
as regard the choice of substituents of the C4-substituent of the
dioxolane ring. The binding of 2 to the heme-binding pocket
of HO-1 is merely through hydrophobic interactions and
coordination bonding,24 as in the case of 3. The flexibility of
the heme-binding pocket, mainly through the dynamic nature
of the distal helix but also through some shifts of proximal
residues and heme itself, allows for the accommodation of these
bulky groups by opening up the binding pocket in such a way
to bind the inhibitor without displacing heme. As stated before,
modifications may be made to take advantage of functional
groups within the pocket such as the polar groups in the thio-
(4-aminophenyl) region of 1, disruption of the Asp140-hydrogen
bond network, increasing flexibility in the central region, as well
as optimizing the length of the central linker to maximize
interactions of the western region with the hydrophobic pocket.
However, the imidazole moiety of the compounds, by coordina-
tion of the heme Fe atom, can be seen as the major force, or
the anchor, to this interaction. As such, one means by which
potency of binding may be enhanced significantly is by sta-
bilizing this anchor.
Experimental Methods
1
Synthesis. The H and 13C NMR spectra were recorded on a
Bruker Avance 400 MHz spectrometer in CD3OD or CDCl3. The
signals owing to residual protons in the deuterated solvents were
used as internal standards. Chemical shifts (δ) are reported in ppm
downfield from tetramethylsilane.32 High-resolution electrospray
mass spectra were recorded on an Applied Biosystems/MDS Sciex
QSTAR XL spectrometer with an Agilent HP1100 Cap-LC system.
Samples were run in 50% aqueous MeOH at a flow rate of 6 µL/
The distance between the carbon at position 5 (Scheme 1) of
the imidazole ring of 3 and the carbonyl group of Gly139 is
3.21 Å, which is close enough to form a potential hydrogen-

5950 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 19
Rahman et al.
min. Elemental analyses were performed by MHW Laboratories
(Phoenix, AZ). Melting points were determined on a Mel-Temp II
melting point apparatus and are uncorrected. Thin-layer chroma-
tography was performed using glass- or aluminum-backed silica
gel 60 F254 plates (Silicycle). Plates were viewed under UV light
or by charring after spraying with phosphomolybdic acid (PMA)
in EtOH.
of LB media supplemented with 100 mg/L ampicillin following
inoculation from a miniculture. After 18 h of growth, cells were
harvested and centrifuged, resulting in green pellets that were stored
at -80 °C. Pellets were subsequently thawed on ice, resuspended
in lysis buffer (100 mM Tris, pH 8.0; 2 mM EDTA; 2 mM PMSF),
and lysed by sonication. Cell debris was separated by centrifugation
for 40 min at 15 000 rpm (Beckman JA-20 rotor) at 4 °C. A
35-60% ammonium sulfate pellet was prepared by adding am-
monium sulfate to 35% saturation and rocking at 4 °C for 40 min
followed by centrifugation for 20 min at 15 000 rpm (JA-20).
Ammonium sulfate was added to the supernatant to a final
concentration of 60% saturation and rocked for 1 h at 4 °C followed
by centrifugation as before. The 35-60% saturated ammonium
sulfate pellet was cleared of ammonium sulfate by dialysis against
10 mM potassium phosphate (pH 7.4) prior to further purification.
Up to this point, the hHO-1 could be monitored by the presence of
green pigment indicative of biliverdin bound to the enzyme, owing
to conversion of endogenous heme by the recombinant protein.
The protein was purified further using FPLC over a MonoQ (HR
10/10, Amersham Biosciences) anion-exchange column using a
gradient of 0-100 mM KCl in 10 mM potassium phosphate (pH
7.4). Fractions containing the apo hHO-1, as monitored by
absorbance (280 nm) and SDS-PAGE analysis, were pooled, and
the protein concentration was determined using the Bradford method
(Bio-Rad). Apo protein was subsequently combined with a 1.2:1
molar ratio of hemin (Fluka), rocked at 4 °C for 20-30 min, and
stored at -20 °C. The excess heme was removed by passage over
a PD-10 size-exclusion column (Amersham Biosciences) that had
been equilibrated with 20 mM potassium phosphate (pH 7.4).
Protein concentration was determined by absorbance (ε405 ) 140
mM -1 cm-1). Purity was assessed by measurement of the Rz ratio
(A405/A280 > 2.1) and by SDS-P AGE analysis.
CO Activity Assay. HO activity in rat spleen microsomal
fractions was determined by the quantitation of CO formed from
the degradation of methemalbumin (heme complexed with
albumin).38,39 Incubations for HO activity analysis were done under
conditions for which the rate of CO formation (pmol CO min-1
mg protein-1) was linear with respect to time and microsomal
protein concentration. Briefly, reactions mixtures (150 µL) each
consisting of 100 mM phosphate buffer (pH 7.4), 50 µM meth-
emalbumin, and 1 mg/mL protein were preincubated with inhibitor
(as the HCl salt) at final concentrations ranging from 0.1 to 100
µM for 10 min at 37 °C. Reactions were initiated by adding
NADPH at a final concentration of 1 mM, and incubations were
performed for an additional 15 min at 37 °C. Blanks for which
NADPH was omitted were subtracted to correct for CO that may
have formed extraneous to HO activity. Reactions were stopped
by instantly freezing the reaction mixture on dry ice, and CO
formation was monitored by gas chromatography according to the
method described by Vreman and Stevenson.38
Spectral Analyses. Inhibitor binding and heme degradation
assays were performed by measuring absorbance data using a
microplate spectrophotometer (Bio-Tek Instruments). Inhibitor
binding was determined by incubation of 10 µM heme-conjugated
hHO-1 with increasing concentrations of 3 (0-50 µM) in 20 mM
potassium phosphate buffer (pH 7.4) for 5 min at room temperature.
Spectra were recorded between 300 and 700 nm at 1 nm intervals.
As a reference, assays were also performed in parallel with 2 in
place of 3. After the determination of the Soret peak position, heme
degradation was initiated by the addition of L-ascorbic acid to a
final concentration of 1 mM and followed by measuring absorbance
at the wavelength corresponding to the respective Soret peaks (404,
408, 410, and 412 nm, respectively) at 1.5 min intervals for a period
of 90 min. Initial rates of decrease of the Soret band were calculated
at the wavelengths of the respective Soret peaks. Spectra were
measured from 300 to 700 nm following the reaction period.
Crystallization. Crystallization conditions were based on previ-
ously published protocols.25,27,33 Briefly, sitting-drop vapor diffu-
sion trials were performed at room temperature using a reservoir
solution containing 100 mM HEPES (pH 7.5), 2.04-2.24 M
ammonium sulfate, and 0.8-1.1% 1,6-hexanediol. The heme-hHO-1
1-(Adamantan-1-yl)-2-(1H-imidazol-1-yl)ethanone Hydrochlo-
ride (3 ·HCl). Imidazole (1.56 g, 22.91 mmol, 8 equiv) was added
to a solution of 1-(adamantan-1-yl)-2-bromoethanone (735 mg, 2.86
mmol) in N,N-dimethylformamide (9 mL) at 0 °C, and the mixture
was stirred at 0 °C for 0.5 h, then stirred at room temperature for
7 days (Scheme 1). The mixture was diluted with a saturated
aqueous solution of Na2CO3, extracted four times with ethyl acetate;
the combined organic extracts were washed with brine (2×) and
then dried (MgSO4). The solution was dried under high-vacuum to
give a pink solid. The solid was ground under H2O (15 mL) and
the mixture stirred at room temperature for 0.5 h. The solid was
removed by filtration and washed with H2O (10 × 10 mL). The
pink-white solid was dried under high vacuum to afford the clean
free base (545 mg, 2.23 mmol, 78%). To a solution of the free
base in warm ethanol (5 mL) was added a solution of 37% aqueous
HCl (250 mg, 2.54 mmol, 1.1 equiv) in ethanol (2 mL). The mixture
was concentrated and dried under high vacuum. The beige solid
was dissolved in a minimum amount of hot ethanol (∼4 mL), the
solution cooled at -20 °C, and the product allowed to crystallize
overnight. The solid was removed by filtration and washed with
ethanol (2 × 1 mL). High-vacuum drying afforded 3·HCl as a beige
solid (478 mg, 1.70 mmol, 59%): mp 261-262 °C. 1H NMR (400
MHz, CD3OD) δ 1.76-1.88 (m, 6H), 1.96-2.00 (m, 6H),
2.06-2.12 (m, 3H), 5.52 (s, 2H), 7.51 (∼t, J ) 1.6 Hz, 1H), 7.58
(∼t, J ) 1.6 Hz, 1H), 8.87 (s, 1H). 13C NMR (100 MHz, CD3OD):
δ 29.3, 37.4, 38.9, 46.9, 54.7, 120.4, 124.8, 137.9, 207.3. HRMS
(ES) [M - Cl]+ calcd for C15H21N2O: 245.1654. Found: 245.1646.
4-(4-Bromophenyl)-1-(1H-1,2,4-triazol-1-yl)-2-butanone Hy-
drochloride (4·HCl). A mixture of 1-bromo-4-(4-bromophenyl)-
2-butanone22 (306 mg, 1 mmol), 1H-1,2,4-triazole (172.5 mg, 2.5
mmol), anhydrous potassium carbonate (345 mg, 2.5 mmol), and
anhydrous N,N-dimethylformamide (3 mL) was stirred at room
temperature for 2 h. The solvent was removed under reduced
pressure, and water (30 mL) was added. The mixture was extracted
with chloroform (15 mL), and the organic extract was washed
sequentially with water (30 mL) and brine (10 mL), dried (sodium
sulfate), and concentrated. Purification of the residue by flash
column chromatography on silica gel (ethyl acetate) gave the free
base 4 as a white solid (279 mg, 0.95 mmol, 95%): mp 118-119
°C. Rf ) 0.32 (ethyl acetate). 1H NMR (CDCl3, 400 MHz): δ 2.74
(t, J ) 7.2 Hz, 2H), 2.87 (t, J ) 7.2 Hz, 2H), 4.93 (s, 2H), 7.02 (d,
J ) 8.0 Hz, 2H), 7.38 (d, J ) 8.0 Hz, 2H), 7.95 (s, 1H), 8.07 (s,
1H). 13C NMR (CDCl3, 100 MHz): δ 28.6, 41.4, 57.8, 120.4, 130.2,
131.8, 139.1, 144.6, 152.3, 201.0. The free base 4 (147 mg, 0.50
mmol) was dissolved in 2-propanol (2.5 mL), treated with 37%
HCl (99 mg, 1.00 mmol), and kept for 2 h at room temperature.
The solid that precipitated was removed by filtration, washed with
diethyl ether, and then recrystallized from 2-propanol-diethyl ether
to afford the hydrochloride salt of 4 as a white solid (119 mg, 0.36
1
mmol, 72%): mp 169-171 °C. H NMR (CD3OD, 400 MHz): δ
2.89-3.01 (m, 4H), 5.46 (s, 2H), 7.17 (d, J ) 8.4 Hz, 2H), 7.42
(d, J ) 8.4 Hz, 2H), 8.71 (s, 1H), 9.46 (s, 1H). 13C NMR (CD3OD,
100 MHz): δ 29.3, 41.9, 60.0, 121.0, 131.5, 132.6, 142.2, 146.3,
201.4. HRMS (ESI) calcd for C12H13BrN3O: 294.0242 [M - Cl]+.
Found: 294.0243.
Expression and Purification of hHO-1. For crystallization
purposes, a truncated, soluble version of hHO-1 was used that
contains 233 amino acids (hHO1-t233) and has been employed
successfully to solve the high-resolution crystal structure for
hHO-1.25,27,33 Bacterial expression and purification of hHO-1 were
based on previously published protocols.34-36
Briefly, DH5R cells transformed with the hHO1-t233/pBace
expression plasmid35,37 (a generous gift from Dr. Ortiz de Mon-
tellano, University of San Francisco) were grown in 4 L (4 × 1 L)

Structure of hHO-1 in Complex
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 19 5951
2006, 86, 583–650.
complex (431 µM in 20 mM potassium phosphate) was mixed with
3 at a molar ratio of 1:3 and incubated for 15 min at room
temperature. Binding was confirmed using spectral analysis by
examining changes in the native Soret peak (described above) prior
to setting up crystallization plates. Drops consisted of 2 µL
protein-inhibitor solution mixed with 2 µL of reservoir solution.
Data Collection and Structure Determination. X-ray diffrac-
tion measurements were performed at the F1 beamline of the
Cornell High Energy Synchrotron Source (CHESS). For data
collection a cryoprotectant comprising 100 mM HEPES (pH 7.5),
2.32 M ammonium sulfate, 0.9% 1,6-hexanediol, and 20% (v/v)
glycerol was used. Crystals were subsequently flash-cooled in a
stream of N2 at 100 K. Data were collected for 360° with an
oscillation of 1° and exposure time of 5 s and processed using
HKL2000.40 The structure of the protein-inhibitor complex was
solved by molecular replacement (MR) using Phaser. The
heme-hHO-1 complex (PDB code 1N3U) in which the heme had
been omitted was the initial probe. An initial template of inhibitor
3 was generated using CS Chem3D Ultra (version 6.0, copyright
2000, CambridgeSoft.com) and the Dundee PRODRG2 server.41
Following initial refinement with Refmac, the structures, first of
heme and then of inhibitor 3, were manually inserted and
subsequently refined using iterative cycles of Xfit and Refmac5 in
the CCP4 suite.42 Standard parameters for heme were used as
described in its library entry in the program during refinement, with
no additional restraints on planarity, bond lengths, and bond angles.
Structural alignments were performed using CCP442 or DaliLite.43
Electrostatic surface potentials were calculated using PyMOL.44
All images were prepared using PyMOL.
(6) Lee, T. S.; Chau, L. Y. Heme oxygenase-1 mediates the anti-
inflammatory effect of interleukin-10 in mice. Nat. Med. 2002, 8, 240–
246.
(7) Lee, T. S.; Tsai, H. L.; Chau, L. Y. Induction of heme oxygenase-1
expression in murine macrophages is essential for the anti-inflamma-
tory effect of low dose 15-deoxy-delta 12,14-prostaglandin J2. J. Biol.
Chem. 2003, 278, 19325–19330.
(8) Imuta, N.; Hori, O.; Kitao, Y.; Tabata, Y.; Yoshimoto, T.; Matsuyama,
T.; Ogawa, S. Hypoxia-mediated induction of heme oxygenase type
I and carbon monoxide release from astrocytes protects nearby cerebral
neurons from hypoxia-mediated apoptosis. Antioxid. Redox. Signaling
2007, 9, 543–552.
(9) Brouard, S.; Otterbein, L. E.; Anrather, J.; Tobiasch, E.; Bach, F. H.;
Choi, A. M.; Soares, M. P. Carbon monoxide generated by heme
oxygenase 1 suppresses endothelial cell apoptosis. J. Exp. Med. 2000,
192, 1015–1026.
(10) Brouard, S.; Berberat, P. O.; Tobiasch, E.; Seldon, M. P.; Bach, F. H.;
Soares, M. P. Heme oxygenase-1-derived carbon monoxide requires
the activation of transcription factor NF-kappa B to protect endothelial
cells from tumor necrosis factor-alpha-mediated apoptosis. J. Biol.
Chem. 2002, 277, 17950–17961.
(11) Kim, H. P.; Ryter, S. W.; Choi, A. M. CO as a cellular signaling
molecule. Annu. ReV. Pharmacol. Toxicol. 2006, 46, 411–449.
(12) Liu, X. M.; Chapman, G. B.; Wang, H.; Durante, W. Adenovirus-
mediated heme oxygenase-1 gene expression stimulates apoptosis in
vascular smooth muscle cells. Circulation 2002, 105, 79–84.
(13) Abraham, N. G.; Kappas, A. Pharmacological and clinical aspects of
heme oxygenase. Pharmacol. ReV., in press.
(14) Fang, J.; Akaike, T.; and Maeda, H. Antiapoptotic role of heme
oxygenase (HO) and the potential of HO as a target in anticancer
treatment. Apoptosis 2004, 9, 27–35.
(15) Fang, J.; Sawa, T.; Akaike, T.; Akuta, T.; Sahoo, S. K.; Khaled, G.;
Hamada, A.; and Maeda, H. In vivo antitumor activity of pegylated
zinc protoporphyrin: targeted inhibition of heme oxygenase in solid
tumor. Cancer Res. 2003, 63, 3567–3574.
(16) Fang, J.; Sawa, T.; Akaike, T.; Greish, K.; Maeda, H. Enhancement
of chemotherapeutic response of tumor cells by a heme oxygenase
inhibitor, pegylated zinc protoporphyrin. Int. J. Cancer 2004, 109,
1–8.
(17) Luo, D.; Vincent, S. R. Metalloporphyrins inhibit nitric oxide-
dependent cGMP formation in vivo. Eur. J. Pharmacol. 1994, 267,
263–267.
(18) Meffert, M. K.; Haley, J. E.; Schuman, E. M.; Schulman, H.; and
Madison, D. V. Inhibition of hippocampal heme oxygenase, nitric
oxide synthase, and long-term potentiation by metalloporphyrins.
Neuron 1994, 13, 1225–1233.
(19) Grundemar, L.; Ny, L. Pitfalls using metalloporphyrins in carbon
monoxide research. Trends Pharmacol. Sci. 1997, 18, 193–195.
(20) Vlahakis, J. Z.; Kinobe, R. T.; Bowers, R. J.; Brien, J. F.; Nakatsu,
K.; Szarek, W. A. Synthesis and evaluation of azalanstat analogues
as heme oxygenase inhibitors. Bioorg. Med. Chem. Lett. 2005, 15,
1457–1461.
(21) Vlahakis, J. Z.; Kinobe, R. T.; Bowers, R. J.; Brien, J. F.; Nakatsu,
K.; and Szarek, W. A. Imidazole-dioxolane compounds as isozyme-
selective heme oxygenase inhibitors. J. Med. Chem. 2006, 49, 4437–
4441.
(22) Roman, G.; Riley, J. G.; Vlahakis, J. Z.; Kinobe, R. T.; Brien, J. F.;
Nakatsu, K.; Szarek, W. A. Heme oxygenase inhibition by 2-oxy-
substituted 1-(1H-imidazol-1-yl)-4-phenylbutanes: effect of halogen
substitution in the phenyl ring. Bioorg. Med. Chem. 2007, 15, 3225–
3234.
(23) Kinobe, R. T.; Vlahakis, J. Z.; Vreman, H. J.; Stevenson, D. K.; Brien,
J. F.; Szarek, W. A.; Nakatsu, K. Selectivity of imidazole-dioxolane
compounds for in vitro inhibition of microsomal haem oxygenase
isoforms. Br. J. Pharmacol. 2006, 147, 307–315.
Molecular Modeling of 1 in the Inhibitor Binding Site of
hHO-1. An initial template of compound 1 was generated as
described above for 3. Virtual docking simulation of 1 was
performed using AutoDock (version 3.0).45 The macromolecule
used for docking was the A chain of hHO-1 in complex with 3, in
which 3 had been omitted.
Acknowledgment. This research was supported by CIHR
Grant MOP64305 held by Dr. Kanji Nakatsu and Dr. Walter
Szarek. Dr. Zongchao Jia holds a Canada Research Chair in
Structural Biology. The X-ray data were collected at the Cornell
High Energy Synchrotron Source. We thank Dr. Paul Ortiz de
Montellano and Dr. John Evans (both from the University of
San Francisco) for the generous gift of the hHO1-t233 plasmid
and advice regarding its expression. We are grateful to Dr. Qilu
Ye and Jimin Zhang for technical expertise in structural
determination, Tracy Gifford for her assistance with the biologi-
cal evaluations, and John G. Riley for the synthesis of compound
4. We also thank Dr. Gheorghe Roman for helpful discussions.
Supporting Information Available: (1) Oxidative degradation
of heme in the carbon monoxide/heme oxygenase (CO/HO)
pathway, (2) structure of heme-conjugated hHO-1 in complex with
3 at 1.54 Å resolution, (3) structures of the HO-1 inhibitors, (4)
contacts between heme conjugated hHO-1 and 3, (5) elemental
analysis results. This material is available free of charge via the
Internet at http://pubs.acs.org.
(24) Sugishima, M.; Higashimoto, Y.; Oishi, T.; Takahashi, H.; Sakamoto,
H.; Noguchi, M.; Fukuyama, K. X-ray crystallographic and biochemi-
cal characterization of the inhibitory action of an imidazole-dioxolane
compound on heme oxygenase. Biochemistry 2007, 46, 1860–1867.
(25) Lad, L.; Schuller, D. J.; Shimizu, H.; Friedman, J.; Li, H.; Ortiz de
Montellano, P. R.; Poulos, T. L. Comparison of the heme-free and
bound crystal structures of human heme oxygenase-1. J. Biol. Chem.
2003, 278, 7834–7843.
References
(1) Maines, M. D. The heme oxygenase system: a regulator of second
messenger gases. Annu. ReV. Pharmacol. Toxicol. 1997, 37, 517–554.
(2) Vreman, H. J.; Wong, R. J.; Stevenson, D. K. In Carbon Monoxide
and CardioVascular Function; Wang, R., Ed.; CRC Press: Boca Raton,
FL, 2002; p 273.
(26) Sun, J.; Wilks, A.; Ortiz de Montellano, P. R.; Loehr, T. M. Resonance
Raman and EPR spectroscopic studies on heme-heme oxygenase
complexes. Biochemistry 1993, 32, 14151–14157.
(27) Schuller, D. J.; Wilks, A.; Ortiz de Montellano, P. R.; Poulos, T. L.
Crystal structure of human heme oxygenase-1. Nat. Struct. Biol. 1999,
6, 860–867.
(28) Lad, L.; Wang, J.; Li, H.; Friedman, J.; Bhaskar, B.; Ortiz de
Montellano, P. R.; Poulos, T. L. Crystal structures of the ferric, ferrous,
and ferrous-NO forms of the Asp140Ala mutant of human heme
(3) Maines, M. D. Heme oxygenase: function, multiplicity, regulatory
mechanisms, and clinical applications. FASEB J. 1988, 2, 2557–
2568.
(4) Hayashi, S.; Omata, Y.; Sakamoto, H.; Higashimoto, Y.; Hara, T.;
Sagara, Y.; and Noguchi, M. Characterization of rat heme oxygenase-3
gene. Implication of processed pseudogenes derived from heme
oxygenase-2 gene. Gene 2004, 336, 241–250.
(5) Ryter, S. W.; Alam, J.; Choi, A. M. Heme oxygenase-1/carbon
monoxide: from basic science to therapeutic applications. Physiol. ReV.

5952 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 19
Rahman et al.
oxygenase-1: catalytic implications. J. Mol. Biol. 2003, 330, 527–
538.
oxidation of 5- and 15-phenylhemes. J. Biol. Chem. 2004, 279, 42593–
42604.
(37) Wilks, A.; Medzihradszky, K. F.; Ortiz de Montellano, P. R. Heme
oxygenase active-site residues identified by heme-protein cross-linking
during reduction of CBrCl3. Biochemistry 1998, 37, 2889–2896.
(38) Vreman, H. J.; Stevenson, D. K. Heme oxygenase activity as measured
by carbon monoxide production. Anal. Biochem. 1988, 168, 31–38.
(39) Cook, M. N.; Nakatsu, K.; Marks, G. S.; McLaughlin, B. E.; Vreman,
H. J.; Stevenson, D. K.; Brien, J. F. Heme oxygenase activity in the
adult rat aorta and liver as measured by carbon monoxide formation.
Can. J. Physiol. Pharmacol. 1995, 73, 515–518.
(40) Otwinowski, Z.; Minor, W. Processing of X-ray diffraction data
collected in oscillation mode. Macromol. Crystallogr., Part A 1997,
276, 307–326.
(41) Schuttelkopf, A. W.; van Aalten, D. M. PRODRG: a tool for high-
throughput crystallography of protein-ligand complexes. Acta Crys-
tallogr., Sect. D.: Biol. Crystallogr. 2004, 60, 1355–1363.
(42) Collaborative Computational Project, Number 4. The CCP4 suite:
programs for protein crystallography. Acta Crystallogr., Sect. D.: Biol.
Crystallogr. 1994, 50, 760-763.
(43) Holm, L.; Park, J. DaliLite workbench for protein structure comparison.
Bioinformatics 2000, 16, 566–567.
(44) DeLano, W. L. The PyMOL Molecular Graphics System, version
0.99rc6; DeLano Scientific: Palo Alto, CA, 2002.
(45) Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart, W. E.;
Belew, R. K.; Olson, A. J. Automated docking using a Lamarckian
genetic algorithm and an empirical binding free energy function.
J. Comput. Chem. 1998, 19, 1639–1662.
(29) Rivera, M.; Caignan, G. A.; Astashkin, A. V.; Raitsimring, A. M.;
Shokhireva, T. K.; Walker, F. A. Models of the low-spin iron(III)
hydroperoxide intermediate of heme oxygenase: magnetic resonance
evidence for thermodynamic stabilization of the d(xy) electronic
state at ambient temperatures. J. Am. Chem. Soc. 2002, 124, 6077–
6089.
(30) Ebing, D. D. General Chemistry; Houghton Mifflin Company: Boston,
MA, 1996.
(31) Bianchetti, C. M.; Yi, L.; Ragsdale, S. W.; Phillips, G. N., Jr.
Comparison of apo- and heme-bound crystal structures of a truncated
human heme oxygenase-2. J. Biol. Chem. 2007, 282, 37624–37631.
(32) Gottlieb, H. E.; Kotlyar, V.; Nudelman, A. NMR chemical shifts of
common laboratory solvents as trace impurities. J. Org. Chem. 1997,
62, 7512–7515.
(33) Schuller, D. J.; Wilks, A.; Ortiz de Montellano, P. R.; Poulos, T. L.
Crystallization of recombinant human heme oxygenase-1. Protein Sci.
1998, 7, 1836–1838.
(34) Wilks, A.; Ortiz de Montellano, P. R. Rat liver heme oxygenase. High
level expression of a truncated soluble form and nature of the meso-
hydroxylating species. J. Biol. Chem. 1993, 268, 22357–22362.
(35) Wilks, A.; Black, S. M.; Miller, W. L.; Ortiz de Montellano, P. R.
Expression and characterization of truncated human heme oxygenase
(hHO-1) and a fusion protein of hHO-1 with human cytochrome P450
reductase. Biochemistry 1995, 34, 4421–4427.
(36) Wang, J.; Niemevz, F.; Lad, L.; Huang, L.; Alvarez, D. E.; Buldain,
G.; Poulos, T. L.; Ortiz de Montellano, P. R. Human heme oxygenase
JM800505M