The reactivity of a-hydroxyhaem and verdohaem bound
to haem oxygenase-1 to dioxygen and sodium dithionite
Hiroshi Sakamoto
1
, Yoshiaki Omata
1
, Shunsuke Hayashi
1
, Saori Harada
1
, Graham Palmer
2
and Masato Noguchi
1
1
Department of Medical Biochemistry, Kurume University School of Medicine, Japan;
2
Department of Biochemistry and Cell Biology,
Rice University, Houston, Texas, USA
Recently we have shown that ferric a-hydroxyhaem bound
to haem oxygenase-1 can be converted to ferrous verdohaem
by approximately an equimolar amount of O
2
in the absence
of exogenous electrons [Sakamoto, H., Omata, Y., Palmer,
G., and Noguchi, M. (1999) J. Biol. Chem. 274, 18196–
18200]. Contrary to those results, other studies have claimed
that the conversion requires both O
2
andanelectron.More
recently, Migita et al. have reported that the major reaction

reducing equivalents from NADPH-cytochrome P450
reductase, and produces biologically active molecules:
biliverdin, CO, and iron, which display both beneficial
and deleterious effects, depending on the circumstances
[1,2]. Two isoforms of HO, HO-1 and HO-2, exist and are
different gene products. HO-1 is highly expressed in the
spleen and liver, and is inducible not only by haem itself but
also by a variety of agents causing oxidative stress. HO-1 is
also known as heat shock protein 32 and may protect cells
from oxidative damage through antioxidant activity of
bilirubin produced through the subsequent reduction of
biliverdin by biliverdin reductase. HO-2, on the other hand,
is constitutively expressed in the brain, testes and vascular
systems. Evidence accumulated in the past decade suggests
that the principal role of HO-2 is the production of CO as a
signal mediator (reviewed in [2–5]). However, this role of
CO as a signaling agent remains controversial [6].
The HO reaction consists of three sequential oxidation
steps [7,8]. Haem bound to the enzyme is first hydroxylated
at the a-meso-carbon, yielding a-hydroxyhaem. The second
step is the conversion of a-hydroxyhaem to verdohaem
with the concomitant release of the a-meso-carbon as CO
[9]. Finally, the oxygen bridge of verdohaem is cleaved to
produce iron and biliverdin. Recently, the crystal structures
of both human [10] and rat [11] HO-1 in complex with
haem have been determined and provide important
information on the selective a-meso-hydroxylation. The
haem is sandwiched between two helices: the proximal
helix bears His-25 as the ligand to the haem iron and the
distal helix lies above the b-, c-, and d-meso-carbon atoms

mechanisms for the subsequent steps, especially the conver-
sion of a-hydroxyhaem to verdohaem, are still unclear. On
the basis of spectroscopic studies of a-hydroxyhaem in
complex with chemical ligands [18–21] and with proteins
such as apomyoglobin [22] or HO [23,24], the ferric
a-hydroxyhaem is deprotonated and assumes the resonance
structure (Scheme 1), a high-spin ferric enolate (1a)-keto
(1b) tautomer and a low-spin ferrous p neutral radical (1c)
(reviewed in [15,25]). It is widely accepted that direct binding
of O
2
to 1c at the haem edge triggers CO extrusion and
verdohaem formation. However, the requirement for redu-
cing equivalents as well as the oxidation state of the
resultant verdohaem remains controversial.
Thus, Matera et al. [26], and subsequently Migita et al.
[27], have reported that this conversion requires both O
2
and
an exogenous electron to yield the ferrous verdohaem
(Scheme 2A). They chemically synthesized a-hydroxyhae-
min and prepared its complex with a soluble rat HO-1
protein. According to their results, most (% 70%) of ferric
a-hydroxyhaem bound to the enzyme was converted to a
ferric porphyrin cation radical upon exposure to O
2
, along
with generation of the ferrous verdohaem as a minor
product (less than 30%). The major product, the ESR-silent
porphyrin cation radical, can be returned to the ferrous

the absence of reducing equivalents and that the resultant
verdohaem appeared to be in the ferrous state (Scheme 2C).
In contrast to the previous studies, the pathway we have
proposed does not require any reducing equivalents, and
instead postulates generation of an oxidizing equivalent
with the formal stoichiometry of (1/2) H
2
O
2
[28].
In the present study, we have attempted to clarify the
reason(s) for these discrepancies. At first, we carefully
compared the absorption spectra of our ferric, ferrous and
CO-ferrous a-hydroxyhaem-HO complexes with those
prepared by Matera et al. [26]. Next, we compared the
conversion of a-hydroxyhaem to verdohaem and of verdo-
haem to biliverdin under various conditions including those
used by Migita et al. [27].
EXPERIMENTAL PROCEDURES
Materials
A truncated version of rHO-1 lacking the 22-amino acid
C-terminal hydrophobic stretch was expressed in Escheri-
chia coli and purified as described [29]. The catalytic activity
of this rHO-1 was comparable to that of the wild type.
NADPH-cytochrome P450 reductase was purified from rat
liver as previously described [30]. Gases of high purity,
argon (99.999%), O
2
(99.99%) and N
2

M
NaOH saturated with
argon was added, with the molar ratio to rHO-1 being
% 0.9. The final concentration of rHO-1 was % 70 l
M
.The
amounts of a-hydroxyhaem added were determined as the
pyridine haemochrome of a-hydroxyhaemin dimethyl ester
using e
422 nm
¼ 153.6 m
M
)1
Æcm
)1
[22]. By this procedure,
i.e. addition of the alkaline a-hydroxyhaem solution into the
rHO-1 protein in the potassium phosphate buffer (pH 7.4),
it was found that complex formation was easily accom-
plished (within 3 min). Hence, we adopted this procedure as
the routine method for preparing the a-hydroxyhaem-rHO-
1 complex.
Formation of ferrous verdohaem-rHO-1 complex
The ferrous verdohaem-rHO-1 complex was prepared in
two different ways. In the first method, the complex was
prepared from the a-hydroxyhaem-rHO-1 complex by
addition of an approximately equimolar amount of O
2
[24]. In the second method, the complex was prepared by
reconstitution from synthetic a-verdohaem and rHO-1 [31].

(e
446 nm
¼ 11.3 m
M
)1
Æcm
)1
) with the dithionite solution,
assuming that 1 mol of dithionite reduces 1 mol of lumi-
flavin 3-acetate by 2-electron reduction [32,33]. The con-
centration of NADPH was determined using e
339 nm
¼
6.2 m
M
)1
Æcm
)1
.TheO
2
-saturated buffer (1.25 m
M
as O
2
)
was prepared by bubbling O
2
into the buffer solution for
3 h. Each of these titrants (dithionite, NADPH and O
2

whereas the Soret bands of the ferrous and ferrous-CO
forms were narrow (Fig. 1B). The ratios of the intensities of
the Soret maximum of the ferric (spectrum aÕ), and ferrous
Fig. 1. Comparison of the absorption spectra of several a-hydroxyhaem-HO-1 complexes. (A) The ferric (spectrum a, –), ferrous (spectrum b, ÆÆ Æ), and
CO-ferrous a-hydroxyhaem-HO-1 complex (spectrum c, - - -) prepared in this study. The final concentrations of rHO-1 and a-hydroxyhaemin were
69 and 61 l
M
, respectively. See Experimental procedure for details. (B) The ferric (spectrum aÕ, –), ferrous (spectrum bÕ, ÆÆÆ), and CO-ferrous
a-hydroxyhaem-HO-1 complex (spectrum cÕ, - - -) reported by Matera et al. [26]. Figure 1 from [26] is replotted in order to make the apparent
heights of absorption maxima of CO-ferrous complex of both preparations the same.
Ó FEBS 2002 Conversion of a-hydroxyhaem to verdohaem by HO (Eur. J. Biochem. 269) 5233
complexes (spectrum bÕ) to that of the CO-ferrous complex
(spectrum cÕ) were 0.55 and 0.78, respectively. The optical
spectra of the three forms of our a-hydroxyhaem-rHO-1
complex are shown in Fig. 1A; the Soret maxima at 407,
433, and 421 nm for the ferric (spectrum a), ferrous
(spectrum b), and CO-ferrous (spectrum c) forms, respect-
ively, agree with those reported by Matera et al.[26].
However, the sharpness of the Soret band of our ferric form
was remarkable in that the ratios of intensities of the Soret
maximum of our ferric and ferrous forms to that of the
CO-ferrous form were 0.65, and 0.81, respectively. Com-
parison of these ratios to those of Matera et al.suggested
that formation of their complex between ferric a-hydroxy-
haem and HO was incomplete, and this led us to reexamine
the process of complex formation.
Upon reversing the order of additions, i.e. concentrated
rHO-1 solution was added to the neutral a-hydroxyhaem
solution in the same pH 7.4 buffer, complex formation took
over 6 h for completion (data not shown). We further noticed

haem [20,34]. As reduction of the haem iron allows the
meso-oxygenatomtobeprotonated[26],thedimeric
structure could not be maintained in the ferrous state.
Hence, we suggest that in these experiments the free
a-hydroxyhaem rapidly dimerized and subsequently aggre-
gated and precipitated and that reduction of ferric
a-hydroxyhaem by sodium dithionite reversed this aggre-
gation and consequently facilitated complex formation with
the enzyme. This in turn explains why the spectra of both
our and Matera’s ferrous and CO-ferrous complexes are
almost identical (Fig. 1).
Reaction of ferric a-hydroxyhaem with excess O
2
Migita et al. [27] reported that upon exposure of ferric
a-hydroxyhaem bound to HO-1 to air, it was oxidized to a
ferric porphyrin cation radical. However, we have demon-
strated that the anaerobic addition of an equimolar amount
of O
2
converts the ferric a-hydroxyhaem to the ferrous
verdohaem [24]. To investigate the effect of excess O
2
on
ferric a-hydroxyhaem, we again titrated the a-hydroxy-
haem-rHO-1 complex with O
2
under strictly anaerobic
conditions (Fig. 3A). As previously reported [24], the
addition of an approximately equimolar amount of O
2

, respectively. Spectra were recorded immediately
(spectrum a, –) and after 1-h incubation (spectrum b, - - -). (B) Spectra
were recorded immediately (–), and at 30 min, 1 h (ÆÆÆ),and2h( )
after addition of 1.4 eq of sodium dithionite (54 nmol) to the sample of
spectrum b.
5234 H. Sakamoto et al. (Eur. J. Biochem. 269) Ó FEBS 2002
Spectrum b (Fig. 3A) resembles the spectrum that Migita
et al. assigned to the ferric a-hydroxyhaem p cation radical,
and is different from that of the ferric verdohaem-rHO-1
complex [24,31]. The ferric verdohaem complex exhibits a
rhombic ESR spectrum with g-values of 2.54, 2.14, and 1.88
that are typical of low-spin haemproteins possessing
hydroxide as the sixth ligand [24,31], but the product giving
spectrum b gave a poorly resolved axial signal at g ¼ 2.02
and 1.98 (data not shown). The reaction product(s) of the
ferrous verdohaem with excess O
2
(hereafter referred to as
the O
2
-oxidized verdohaem) has not yet been fully clarified.
Figure 3B shows the sequential reactions of the
O
2
-oxidized verdohaem with sodium dithionite and then
with CO and O
2
, according to the procedures reported by
Migita et al. [27]. The anaerobic addition of sodium dithi-
onite yielded spectrum c showing an asymmetric Soret band

Àe
À
; ÀH
þ
ÀÀÀÀÀÀÀÀÀ!
ÀÀÀÀÀÀÀÀ
þe
À
; ÀH
þ
fðpyÞ
2
Fe
III
ðOEPOÞg
Àe
À
ÀÀÀÀÀÀÀÀÀ!
ÀÀÀÀÀÀÀÀ
þe
À
ðpyÞ
2
Fe
III
ðOEPOÞ
ÂÃ
þ
ð1Þ
The solution of {(py)

We next explored the degradation of the ferrous verdohaem
bound to rHO-1 that was obtained from the reaction of the
ferric a-hydroxyhaem complex with an equimolar amount of
O
2
, and compared the behavior of two reducing systems,
namely: sodium dithionite and NADPH-cytochrome P450
reductase. When sodium dithionite was added gradually to
the verdohaem complex under anaerobic conditions, decrea-
ses in absorbance at 400, 535 and 690 nm took place and
broad bands appeared at 431 and 795 nm (Fig. 4A, spectrum
a). The absorption around 795 nm initially increased and
then decreased during the addition of dithionite. The spectral
changes appeared to be completed with about 4 eq of sodium
dithionite. The possibility of that the production of hydrogen
peroxide caused degradation of verdohaem was ruled out
because the dithionite solution was made anaerobically and
was used anaerobically. These findings thus suggested that
the ferrous verdohaem had been converted to a further
reduced form by the dithionite, because no similar spectra
have been previously observed in the physiological degrada-
tion of haem by HO (vide infra). The subsequent addition of
CO yielded an asymmetric Soret band with shoulders at 421
and 435 nm (Fig. 4B, spectrum b). As described above,
absorption at 421 nm implies that a small amount of the CO-
ferrous a-hydroxyhaem complex was produced by reduction
of unbound ferric a-hydroxyhaem. The additional absorp-
tion at 435 nm may be related to a CO-adduct of the
dithionite-reduced verdohaem. Exposure to air led to a
decrease in the Soret band, resulting in the production of a

except for an increase in absorption at 340 nm due to the
added NADPH (Fig. 5A). Subsequent addition of CO gave
a spectrum (Fig. 5B, spectrum b) with absorption maxima
at 409 and 638 nm that are characteristic of the CO-ferrous
verdohaem complex [31,37]. Further exposure to air caused
a loss of the absorption maxima at 340, 409 and 638 nm,
and subsequent increases in the absorbances around 380
and 680 nm (Fig. 5C, spectrum c) indicative of biliverdin
formation [36]. For the ring opening that occurs in the
conversion of verdohaem to biliverdin, O
2
and reducing
equivalents are clearly necessary.
In order to confirm the results obtained above, we further
explored the degradation of the bound synthetic verdo-
haem. Additions of dithionite to the synthetic verdohaem-
rHO-1 complex caused spectral changes (Fig. 6A) similar to
those shown in Fig. 4A. The addition of CO produced an
absorption maximum at 438 nm (Fig. 6B, spectrum b) that
was considered to be a form of ferrous verdohaem further
Fig. 5. Spectral change during the degradation of verdohaem bound to
rHO-1 caused by the NADPH-cytochrome P450 reductase system. (A)
Spectra were recorded after additions of 0 (–), 1.0, 2.0, 3.0 (ÆÆÆ), and
3.7 eq (spectrum a, - - -) of NADPH, in the presence of NADPH-
cytochrome P450 reductase, to the verdohaem-rHO-1 complex
obtained as in Fig. 4A. (B) The spectrum (–) is the same as spectrum a.
Spectrum b (- - -) was recorded after addition of CO. (C) The spectrum
(–) is the same as spectrum b. Several spectra (ÆÆÆ) were recorded after
exposure to air and spectrum c (- - -) is the final product.
Fig. 4. Spectral change during the degradation of verdohaem bound to

Fe
II
(OEOP),
and formulated it as the p neutral radical {(py)
2
Fe
II
(OEOP)
•
}. They also reported that, when the electro-
chemical reduction was carried out in the presence of
dioxygen, ring rupture took place and an open-chain poly
pyrrole iron complex was produced. The absorption
spectrum of the p neutral radical showing peaks at 439
and 740 nm closely resembles the dithionite-reduced form
of the verdohaem-HO-1 complex, which shows absorption
peaks at 431 and 795 nm (Figs 4A and 6A). Exposure of the
dithionite-reduced form to air gave a biliverdin-iron chelate-
like compound. Thus, the dithionite-reduced form is
thought to correspond to the p neutral radical obtained in
the (py)
2
Fe
II
(OEOP) system. Ishizu and coworkers pro-
posed that the p neutral radical may be an intermediate in
the haem decomposition process. However it should be
noted that such a species is never seen in the physiological
haem degradation driven by cytochrome P450 reductase
(Figs 5 and 7).

O
2
. These results confirm the reaction mechanism previ-
ously proposed [24], in which O
2
first attacks the ring carbon
adjacent to the a-oxy group of species 1c, resulting in a
dioxygen adduct that is then rearranged to produce CO and
ferrous verdohaem. This intramolecular rearrangement is
accompanied by concomitant expulsion of an oxidizing
equivalent (Scheme 2C). At present we cannot explain the
fate of the oxidizing equivalent; some may react with HO
protein, the rest may be eliminated by continuous electron
flow from NADPH cytochrome P450 reductase. The
dithionite-reduced form of ferrous verdohaem is considered
to be a p neutral radical species that cannot be an
intermediate in the physiological degradation of haem.
Hence, the use of sodium dithionite should be avoided in the
study of the haem oxygenase reaction.
ACKNOWLEDGEMENTS
This work was supported in part by Grant-in-aid for Scientific
Research on Priority Areas (Biological Machinery (No. 13033041))
from the Ministry of Education, Culture, Sports, Science and
Technology of Japan, Grant-in-aid for Scientific Research (C) (No.
12670125) from the Japan Society for the Promotion of Science, Grant
00K1100 from the Ichiro Kanehara Foundation, Grant GM 55807
from the National Institutes of Health, and Grant C636 from the Welch
Foundation.
REFERENCES
1. Maines, M.D. (1997) The heme oxygenase system: a regulator of

T.L. (1999) Crystal structure of human heme oxygenase-1. Nat.
Struct. Biol. 6, 860–867.
11. Sugishima, M., Omata, Y., Kakuta, Y., Sakamoto, H., Noguchi,
M. & Fukuyama, K. (2000) Crystal structure of rat heme oxyge-
nase-1 in complex with heme. FEBS Lett. 471, 61–66.
12. Noguchi, M., Yoshida, T. & Kikuchi, G. (1983) A stoichiometric
study of heme degradation catalyzed by the reconstituted heme
oxygenase system with special consideration of the production of
hydrogen peroxide during the reaction. J. Biochem. (Tokyo) 93,
1027–1036.
13. Wilks, A. & Ortiz de Montellano, P.R. (1993) Rat liver heme
oxygenase. High level expression of a truncated soluble form and
nature of the meso-hydroxylating species. J. Biol. Chem. 268,
22357–22362.
14. Wilks, A., Torpey, J. & Ortiz de Montellano, P.R. (1994) Heme
oxygenase (HO-1). Evidence for electrophilic oxygen addition to
the porphyrin ring in the formation of a-meso-hydroxyheme.
J. Biol. Chem. 269, 29553–29556.
15. Ortiz de Montellano, P.R. & Wilks, A. (2001) Heme oxygenase
structure and mechanism. In Advances in Inorganic Chemistry,
Vol. 51 (Sykes, A.G., ed.), pp. 359–407. Academic Press, San
Diego.
16. Davydov, R.M., Yoshida, T., Ikeda-Saito, M. & Hoffman, B.M.
(1999) Hydroperoxy-heme oxygenase generated by cryoreduction
catalyzes the formation of a-meso-hydroxyheme as detected by
EPR and ENDOR. J. Am. Chem. Soc. 121, 10656–10657.
17. Liu, Y., Lightning, L.K., Huang, H., Moe
¨
nne-Loccoz, P., Schuller,
D.J., Poulos, T.L., Loehr, T.M. & Ortiz de Montellano, P.R. (2000)

25. Balch, A.L. (2000) Coordination chemistry with meso-hydroxyl-
ated porphyrins (oxophlorins), intermediates in heme degradation.
Coord. Chem. Rev. 200–202, 349–377.
26. Matera, K.M., Takahashi, S., Fujii, H., Zhou, H., Ishikawa, K.,
Yoshimura, T., Rousseau, D.L., Yoshida, T. & Ikeda-Saito, M.
(1996) Oxygen and one reducing equivalent are both required for
the conversion of a-hydroxyhemin to verdoheme in heme
oxygenase. J. Biol. Chem. 271, 6618–6624.
5238 H. Sakamoto et al. (Eur. J. Biochem. 269) Ó FEBS 2002
27. Migita, C.T., Fujii, H., Matera, K.M., Takahashi, S., Zhou, H. &
Yoshida, T. (1999) Molecular oxygen oxidizes the porphyrin ring
of the ferric a-hydroxyheme in heme oxygenase in the absence of
reducing equivalent. Biochim. Biophys. Acta 1432, 203–213.
28. Fuhrhop, J H., Besecke, S., Subramanian, J., Mengersen, C. &
Riesner, D. (1975) Reactions of oxophlorines and their, p.radi-
cals. J. Am. Chem. Soc. 97, 7141–7152.
29. Omata, Y., Asada, S., Sakamoto, H., Fukuyama, K. & Noguchi,
M. (1998) Crystallization and preliminary X-ray diffraction stu-
dies on the water soluble form of rat heme oxygenase-1 in complex
with heme. Acta Crystallogr. D54, 1017–1019.
30. Yasukochi, Y. & Masters, B.S.B. (1976) Some properties of a
detergent-solubilized NADPH-cytochrome c (cytochrome P-450)
reductase purified by biospecific affinity chromatography. J. Biol.
Chem. 251, 5337–5344.
31. Sakamoto, H., Omata, Y., Adachi, Y., Palmer, G. & Noguchi, M.
(2000) Separation and identification of the regioisomers of ver-
doheme by reversed-phase ion-pair high-performance liquid
chromatography, and characterization of their complexes with
heme oxygenase. J. Inorg. Biochem. 82, 113–121.
32. Burleigh Jr, B.D., Foust, G.P. & Williams, S. (1969) A method for

Raman spectroscopic characterization of a-hydroxyheme and
verdoheme complexes of heme oxygenase. Biochemistry 36, 1402–
1410.
38. Tajima, K., Tada, K., Yasui, A., Ohya-Nishiguchi, H. & Ishizu,
K. (1993) Detection of octaethylverdohaem p-neutral radical by
electrochemical-optical absorption and -ESR measurements.
J. Chem. Soc. Chem. Commun. 282–284.
Ó FEBS 2002 Conversion of a-hydroxyhaem to verdohaem by HO (Eur. J. Biochem. 269) 5239