Báo cáo khoa học: Critical roles of Leu99 and Leu115 at the heme distal side in auto-oxidation and the redox potential of a hemeregulated phosphodiesterase from Escherichia coli pptx

Critical roles of Leu99 and Leu115 at the heme distal side
in auto-oxidation and the redox potential of a heme-
regulated phosphodiesterase from Escherichia coli
Nao Yokota, Yasuyuki Araki, Hirofumi Kurokawa, Osamu Ito, Jotaro Igarashi and Toru Shimizu
Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai, Japan
The heme-regulated phosphodiesterase from Escheri-
chia coli (Ec DOS) is a heme-based sensor enzyme
composed of two functional domains: an N-terminal
domain with a PAS structure that contains the heme
iron; a C-terminal domain that contains the phospho-
diesterase catalytic domain [1–8]. Ec DOS hydrolyzes
cAMP when the heme iron is in the ferrous [Fe(II)]
state, whereas it is inactive when the heme iron is in
the ferric state [Fe(III)] [2,4,7]. Determination of the
X-ray crystal structure resolved some aspects of how
changes in the N-terminal sensor domain are intra-
molecularly transduced to regulation of the catalytic
domain [6]. Specifically, the X-ray crystal structure
of the isolated heme-bound PAS domain (Ec DosH)
Keywords
auto-oxidation; CO binding; heme-sensor
protein; O
2
binding; phosphodiesterase
Correspondence
T. Shimizu, Institute of Multidisciplinary
Research for Advanced Materials, Tohoku
University, 2-1-1 Katahira, Aoba-ku,
Sendai 980-8577, Japan
Fax: +81 22 217 5604, 5390
Tel: +81 22 217 5604, 5605

E-mail:
(Received 13 December 2005, revised 14
January 2006, accepted 18 January 2006)
doi:10.1111/j.1742-4658.2006.05145.x
The heme-regulated phosphodiesterase from Escherichia coli (Ec DOS),
which is a heme redox-dependent enzyme, is active with a ferrous heme but
inactive with a ferric heme. Global structural changes including axial ligand
switching and a change in the rigidity of the FG loop accompanying the
heme redox change may be related to the dependence of Ec DOS activity
on the redox state. Axial ligands such as CO, NO, and O
2
act as inhibitors
of Ec DOS because they interact with the ferrous heme complex. The
X-ray crystal structure of the isolated heme-bound domain (Ec DosH)
shows that Leu99, Phe113 and Leu115 indirectly and directly form a
hydrophobic triad on the heme plane and that they should be located at
or near the ligand access channel of the heme iron. We generated L99T,
L99F, L115T, and L115F mutants of Ec DosH and examined their
physicochemical characteristics, including auto-oxidation rates, O
2
and CO
binding kinetics, and redox potentials. The Fe(III) complex of the L115F
mutant was unstable and had a Soret absorption spectrum located 5 nm
lower than those of the wild-type and other mutants. Auto-oxidation rates
of the mutants (0.049–0.33 min
)1
) were much higher than that of the wild-
type (0.0063 min
)1
). Furthermore, the redox potentials of the former three

mutants (23.1–34.6 mV versus SHE) were also significantly lower than that
of the wild-type (63.9 mV versus SHE). Interaction between O
2
and the
L99F mutant was different from that in the wild-type, whereas CO binding
rates of the mutants were similar to those of the wild-type. Thus, it appears
that Leu99 and Leu115 are critical for determining the characteristics of
heme iron. Finally, we discuss the roles of these amino-acid residues in the
heme electronic states.
Abbreviations
BjFixL, oxygen sensor heme protein from Bradyrhizobium japonicum; Ec DOS, a heme-regulated phosphodiesterase of Escherichia coli;
Ec DosH, the isolated heme-bound PAS domain of Ec DOS; SHE, standard hydrogen electrode; SmFixL, oxygen sensor heme protein from
Sinorhizobium meliloti.
1210 FEBS Journal 273 (2006) 1210–1223 ª 2006 The Authors Journal compilation ª 2006 FEBS
indicates that global structural changes accompany the
redox change. Specifically, there is a switch in the heme
axial ligand and changes in the flexibility of the FG
loop (residues 86–96) in the protein when the redox
state of the heme iron is changed [6]. For the Fe(III)
heme, the axial ligands are His77 and hydroxide anion,
whereas those for the Fe(II) heme are His77 and
Met95. In addition, the FG loop is very flexible and
disordered and could not be resolved in the crystal
structural for the Fe(III) heme complex, whereas the
loop was rigid and could be resolved for the Fe(II)
heme complex. It is expected that these structural
changes in the heme-bound PAS domain are related to
intramolecular signal transduction to the catalytic
domain.
Interestingly, CO and NO bind to the Fe(II) heme

complex, inactivating the enzyme [2]. O
2
also binds to
the Fe(II) heme complex and easily oxidizes the heme
iron to the Fe(III) heme complex, terminating cata-
lysis. Therefore, these gases should act as inhibitors
by axially coordinating to the Fe(II) heme complex.
Met95 is the axial ligand of the Fe(II) heme complex
(Fig. 1A). Mutations at Met95 of Ec DosH markedly
change the kinetic parameters for CO and O
2
binding
to the Fe(II) heme complex as well as the redox poten-
tial of the heme iron [5,8]. The rate of CN binding to
the Fe(III) heme complex of Ec DosH is also remark-
ably accelerated by the M95A and M95L mutations by
8–11-fold [9].
The crystal structure (Fig. 1B) of the Fe(II)–O
2
com-
plex of Ec DosH indicates that Arg97 is hydrogen-
bonded to the molecular oxygen on the heme plane
[10]. A hydrophobic triad observed for other cor-
responding heme-bound PAS enzymes, oxygen sensor
heme protein from Bradyrhizobium japonicum (BjFixL)
and oxygen sensor heme protein from Sinorhizobium
meliloti (SmFix), is also observed for Ec DOS.
Although the triad is composed of Ile215 (BjFixL) ⁄
Ile209 (SmFixL), Leu236 (BjFixL) ⁄ Leu230 (SmFixL),
and Ile238 (BjFixL) ⁄ Val232 (SmFixL) for BjFixL and

SmFixL, there is no amino acid that spatially corres-
ponds to Ile215 (BjFixL) ⁄ Ile209 (SmFixL) in Ec DosH
(Figs 1 and 2) [10]. Phe113 and Leu115 of Ec DosH
correspond to two other members, Leu236 (BjFixL) ⁄
Leu230 (SmFixL) and Ile238 (BjFixL) ⁄ Val232
(SmFixL), respectively, of the hydrophobic triad.
Leu99 serves as a third hydrophobic heme contact in
Ec DosH [10]. The corresponding amino acids at posi-
tion 99 of Ec DOS for the two other heme-bound PAS
proteins, BjFixL and SmFxL, are Gly225 and Gly218,
respectively (Fig. 2). It is well known that CO-binding
and O
2
-binding access channels of myoglobin and he-
moglobin are composed of hydrophobic amino acid
residues [11].
The hydrophobic characteristics of these axial lig-
ands facilitate their binding to the hydrophobic pocket
on the heme distal side. Although the heme distal
structure of the heme-bound PAS domain for Ec DOS
AB
Fig. 1. Structure of the heme distal side of (A) the Fe(II) (PDB code 1V9Z) and (B) the Fe(II)–O
2
(PDB code 1VB6) complexes of Ec DosH
([6,10]; our unpublished data). In (B), Met95 [yellow in (A)] is omitted to clearly illustrate the binding of O
2
to Arg97 [blue in (A) and (B)].
Figures were obtained by
MOLFEAT version 2.1 (Fiatlux, Tokyo).
N. Yokota et al. Heme electronic states of Leu mutants of Ec DOS

FEBS Journal 273 (2006) 1210–1223 ª 2006 The Authors Journal compilation ª 2006 FEBS 1211
is different from that of myoglobin and hemoglobin
(which have a globin fold), it has been speculated that
the hydrophobic characteristic of the heme distal side
contributes substantially to the ligand binding kinetics
of the PAS domain of Ec DOS, as it does in myoglo-
bin and hemoglobin. Examination of the effects of
these hydrophobic amino acids on the kinetics of O
2
and CO binding to the Ec DosH heme is important
because both diatomic molecules act as inhibitors of
Ec DOS.
Therefore, we expected that it would be worth while
to determine how one (Leu115) of the hydrophobic
amino acids that directly participates in the hydro-
phobic triad and another (Leu99) in the second hydro-
phobic contact contribute to the physicochemical
characteristics of Ec DOS. We generated L99T, L99F,
L115T, and L115F mutants of Ec DosH to understand
how these hydrophobic amino acids contribute to O
2
and CO binding kinetics and other physicochemical
characteristics such as auto-oxidation and the redox
potential of the heme iron. We found that mutation of
these hydrophobic amino acids substantially influences
the rate of auto-oxidation and the redox potential of
the heme iron, but, surprisingly, they had less effect
on the characteristics of O
2
and CO binding. Finally,

we discuss the roles of these hydrophobic amino acids
in the structure surrounding heme and the electronic
states of heme.
Results and Discussion
Optical absorption spectra of the Fe(III)–CO,
Fe(II)–CO, and Fe(II)–CO complexes of the
Ec DosH mutants
The optical absorption spectra of the Fe(III)–CO,
Fe(II)–CO, and Fe(II)–CO complexes of the Leu99
and Leu115 mutants of Ec DosH were essentially the
same as those of the wild-type protein, except for the
L115F mutant (Fig. 3, Table 1). It is thought that
these mutations (except for L115F) did not alter the
structure of the heme surroundings, including the heme
co-ordination structure. The Fe(III) complex of the
L115F mutant has the Soret absorption at 413 nm,
which is lower than those of other proteins (417–
418 nm). It has been suggested that the Fe(III) com-
plexes of most of the mutant proteins are in a low-spin
state with His77 and hydroxide anion as the axial lig-
ands, as in the wild-type protein [3,5,6,8]. However, it
appears that the L115F mutant contains a different
axial ligand trans to His77. Introducing a group with a
large side chain, i.e. the phenyl group in L115F, may
have substantially changed the heme distal side struc-
ture and led to the changes in the heme co-ordination
structure and ⁄ or movement of the heme plane such as
sliding, twisting, or doming. When the distal axial lig-
and of myoglobin was changed from OH
–

to water
or acetate anion, the Soret peak position moved to a
lower wavelength by 4–5 nm [11]. Thus, it seems that
the axial ligand of the Fe(III) complex of Ec DosH
switched from OH
–
to the water molecule as a result
of the L115F mutation.
The Fe(II) complexes of all of the mutant proteins
in this study should have His77 and Met95 as axial lig-
ands, whereas the Fe(II)–CO complexes should have
CO and His77 as the axial ligands, as in the wild-type.
This suggests that the structures surrounding heme,
including the heme coordination structure, are essen-
tially the same in the wild-type protein and the
mutants generated here, except for the L115F mutant.
The over-expression efficiency of the L115F mutant
protein in the bacteria was comparable to that of the
wild-type protein, but the heme-bound L115F mutant
protein was more difficult to purify than the wild-type
and other mutant proteins. Purification yield of the
heme-bound L115F mutant was low, less than 10% of
other proteins. The heme content of the L115F mutant
was % 30%, which is significantly lower than those
(60–70%) of the wild-type and other mutant proteins,
suggesting that the L115F mutant has a lower ability
to bind heme than the wild-type and the other mutant
proteins. The L115F mutant of Ec DosH was not used
for further determination of the physicochemical char-
acteristics, such as the kinetics of O

2
or CO binding or
Fig. 2. Partial sequences of amino acids of the PAS domains of Ec DOS and other related heme-bound PAS proteins. The bold amino acids
represent those discussed in the text.
Heme electronic states of Leu mutants of Ec DOS N. Yokota et al.
1212 FEBS Journal 273 (2006) 1210–1223 ª 2006 The Authors Journal compilation ª 2006 FEBS
the redox potential (Tables 3, 4, 5), because strong
heme binding and stability of the heme-bound protein
are needed to obtain precise values.
The Fe(II)–O
2
complexes and auto-oxidation
We attempted to obtain spectra of the Fe(II)–O
2
heme
complexes of the Ec DosH mutant proteins. The spec-
tral maxima of the Fe(II)–O
2
complexes of all mutants
except the L99F mutant were located at 417 nm, essen-
tially the same as that of the wild-type protein (Fig. 4
and Table 2). However, we could not obtain the opti-
cal absorption spectrum of the stable Fe(II)–O
2
com-
plex of the L99F mutant. A similar result was also
obtained for the R220I mutant of BjFixL [19].
The rates of auto-oxidation for the Ec DosH mutant
proteins generated in this study were more than eight-
fold higher than that of the wild-type protein (Fig. 4C,

Table 2). In earlier studies, Ala and Leu substitut-
ions at Met95, an axial ligand in the Fe(II) complex,
Table 1. Optical absorption maxima (nm) and millimolar absorption coefficients (mM
)1
Æcm
)1
) of the wild-type and mutant proteins of
Ec DosH. The millimolar absorption coefficients (shown in parentheses) were determined using the pyridine hemochromogen method [26].
Fe(III) Fe(II) Fe(II)–CO
Soret baSoret baSoret ba
Wild-type 418 (110) 529 565 428 (149) 533 563 424 (149) 541 571
L99T 417 (112) 527 565 428 (158) 532 564 423 (178) 541 574
L99F 418 (112) 530 565 429 (160) 532 563 425 (207) 543 574
L115T 418 (105) 533 566 428 (137) 531 562 423 (168) 540 570
L115F 413 (117) 540 429 (141) 535 563 423 (240) 541 572
A
C
B
D
Fig. 3. Optical absorption spectra of Fe(III) (black), Fe(II) (red), and Fe(II)–CO (blue) complexes of the mutant proteins of Ec DosH. (A) L99T
(7.0 l
M per heme) (B) L99F (5.4 lM per heme) (C) L115T (7.8 lM per heme) and (D) L115F (3.4 lM per heme).
N. Yokota et al. Heme electronic states of Leu mutants of Ec DOS
FEBS Journal 273 (2006) 1210–1223 ª 2006 The Authors Journal compilation ª 2006 FEBS 1213
markedly decreased the rate of auto-oxidation [5],
whereas Ala and Asn substitutions at Asp40, an
amino-acid residue that interacts via two water mole-
cules with the proximal ligand His77, markedly
increased the rate of auto-oxidation [13] (Table 2). The
rates of auto-oxidation appear to be influenced by

the hydrogen-bonding network to O
2
, the polarity of
the heme distal side, the direction of the heme-bound
O
2
molecule, and the redox potential of the heme iron
[5]. The reason why the auto-oxidation rate of the
Leu99 and Leu115 mutants was much higher than that
of the wild-type protein is currently uncertain. It is
clear, however, that the polarity of the O
2
-binding site
changed for the L99T and L115T mutants, whereas
the stronger hydrophobic contact, strain, or compact-
ness (or some combination thereof) of the O
2
-bound
space may have contributed to the increase in the rates
of auto-oxidation for the L99F and L115F mutants.
O
2
binding kinetics
The O
2
binding kinetics of the L99T, L99F, and
L115T mutants was examined by using a stopped-flow
spectrometer under anaerobic conditions (Fig. 5 and
Table 3). The spectral changes accompanying O
2

bind-
ing monitored at 429 nm were composed of two phases
(% 1 : 1 ratio) for the wild-type and the L99T, L99F,
and L115T mutant proteins. Both the fast and slow
phases were dependent on the concentration of O
2
for
all proteins except the L99F mutant, for which the
slow phase was independent of the O
2
concentration.
The rate constants for the fast phase of O
2
binding to
the three mutant proteins [(49–75) · 10
)3
lm
)1
Æs
)1
]
were comparable to those of the wild-type protein
[(31–81) · 10
)3
lm
)1
Æs
)1
] [5]. Similarly, those rates for
the slow phase of O

2
binding to the two mutant
proteins [(6.8–7.2) · 10
)3
lm
)1
Æs
)1
] were comparable
to that for the wild-type protein (8.3 · 10
)3
lm
)1
Æs
)1
)
(Table 3). Therefore, it appears that mutations of
Leu99 and Leu115, except for the L115F mutation,
had little effect on the kinetics of O
2
binding.
In our previous report, we observed only the fast
phase [5], whereas only the slow phase was observed
by others [1], leading to conflicting results. In this
study, we used the new stopped-flow spectrometer
A
B
Fig. 4. (A) Optical absorption spectrum of the Fe(II)–O
2
complex of

the L99T mutant (7.1 l
M per heme). Arrows designate spectral
changes from the Fe(II)–O
2
complex (black) to the Fe(III) complex
(red). (B) Time-dependent changes in intensity at 580 nm accom-
panied by the change from the Fe(II)–O
2
to the Fe(III) complexes of
the wild-type (black) and L99T mutant (red) proteins of Ec DosH.
Experimental data (dotted lines) are fitted to the calculated lines
with the auto-oxidation rate constants of 0.0063 and 0.049 min
)1
,
for the wild-type and L99T mutant, respectively.
Table 2. Optical absorption spectral maxima (nm) of the Fe(II)–O
2
complexes and auto-oxidation rates (k
ox
) of the wild-type and
mutant Ec DosH proteins. Half-lives of the Fe(II)–O
2
complexes are
also described in the right column. ND, no data.
Soret
(nm)
b
(nm)
a
(nm)

k
ox
(min
)1
)
t
1 ⁄ 2
(min) References
Wild-type 417 541 579 0.0063 110 This study
– – – 0.0058 120 [5]
– – – 0.0053 130 [13]
L99T 417 541 579 0.049 14 This study
L99F
a
ND 0.37 1.9 This study
L115T 417 541 579 0.065 11 This study
L115F 417 541 579 0.33 2.1 This study
M95A – – – 0.0013 530 [5]
M95L – – – 0.0017 410 [5]
M95H – – – 0.016 43 [5]
D40A – – – 0.051 14 [13]
D40N – – – 0.033 21 [13]
a
Fast O
2
dissociation (or fast auto-oxidation) may have hampered
the determination of exact values because the O
2
binding rate was
similar to that of the wild-type (Table 3).

Heme electronic states of Leu mutants of Ec DOS N. Yokota et al.
1214 FEBS Journal 273 (2006) 1210–1223 ª 2006 The Authors Journal compilation ª 2006 FEBS
with new PC and software, which allowed us to mon-
itor broader time domains simultaneously and thus
probably to detect both fast and slow phases. The
reason why the kinetics of O
2
binding is composed of
two phases is not currently clear. The crystal struc-
ture of the Fe(II)–O
2
complex in Ec DosH forms a
homodimer [2], and one O
2
molecule can bind to
only one subunit of the dimer [10]. Therefore, it is
possible that the first O
2
molecule binds quickly to
one of the dimers, followed by binding of the second
O
2
molecule to the other subunit of the dimer, result-
ing in two phases. However, further study is needed
to probe this possibility, as there is no obvious struc-
tural origin for the differences in binding to the
monomers. Another possibility is that once the O
2
molecule binds to the heme distal site and the distal
Arg97 binds to the O

2
molecule [10], it may allosteri-
cally influence the binding of the second O
2
molecule.
(Table 3).
It is not clear why a stable Fe(II)–O
2
complex was
not observed for the L99F mutant (Table 2). The O
2
binding rate to the L99F mutant was similar to those
of the wild-type and the other mutant proteins
Table 3. Rates for O
2
association with the wild-type and mutant
Ec DosH proteins as measured by the stopped-flow method. Rates
determined by the stopped-flow method were dependent on the
O
2
concentration. At least three experiments were conducted to
obtain each rate constant. Experimental errors were less than
20%. ND, no data.
k
on
(· 10
)3
lM
)1
Æs

)1
)
References
Fast phase Slow phase
Wild-type 81 8.3 This study
31 – [5]
– 2.6 [1]
L99T 49 6.8 This study
L99F
a
75 ND This study
L115T 55 7.2 This study
L115F
b
ND ND This study
M95A > 1000 – [5]
M95L > 1000 – [5]
M95H > 1000 – [5]
M95I > 1000 – [5]
a
The slow phase of the L99F mutant was independent of the O
2
concentration.
b
Measurement of the rate of O
2
binding was not
feasible because of low heme binding affinity and instability of the
protein.
A

C
B
D
Fig. 5. (A) Changes in the optical spectra of the Fe(II)–O
2
complex formation of the wild-type protein after mixing solutions of protein (8 lM
per heme) and O
2
(488 lM) in the stopped-flow spectrometer. (B) The spectral changes monitored at 429 nm accompanied by the O
2
associ-
ation with the Fe(II) complex were composed of two phases. Experimental data (red dotted line) were fitted using a two-phase model (black
solid line). Rates for the O
2
association for both the fast (C) and slow (D) phases were dependent on the O
2
concentration.
N. Yokota et al. Heme electronic states of Leu mutants of Ec DOS
FEBS Journal 273 (2006) 1210–1223 ª 2006 The Authors Journal compilation ª 2006 FEBS 1215
(Table 3). In addition, the rate of auto-oxidation for
L99F was fast but comparable to that of the L115F
mutant (Table 2). The high rate constant for the disso-
ciation of O
2
from the L99F mutant compared with
the other proteins along with the high rate of auto-
oxidation explains why we did not monitor the stable
Fe(II)–O
2
complex.

CO binding kinetics
We next examined the kinetics of CO binding for the
mutant proteins of Ec DosH using laser flash photo-
lysis (Fig. 6 and Table 4). The changes in the Soret
region of the absorption spectra did not exhibit simple
isosbestic points, but were apparently composed of
two sets of spectral changes (Fig. 6A). Specifically, the
spectral changes in the fast phase had isosbestic points
around 414, 430, and 462 nm (red in Fig. 6B), whereas
those in the slow phase had isosbestic points around
396, 425, and 450 nm (blue in Fig. 6C). We overlapped
the spectral changes associated with the fast (isosbestic
points at 430 nm) and slow (isosbestic point at
425 nm) reactions (Fig. 6B,C). The spectral changes
associated with the fast phase occurred within a time
scale of microseconds, whereas those of the slow phase
were over a millisecond time scale. The pattern of CO
association with Ec DosH monitored at 420 nm was
also composed of fast and slow phases (Fig. 7A). The
fast phases, which occurred of the order of microsec-
onds (Fig. 7B), were independent of the CO concentra-
tion, whereas the slow phases, which were of the order
of milliseconds, were dependent on the CO concentra-
tion (inset of Fig. 7B). The rate constants of the fast
phase of the mutant proteins, except for the L115F
mutant, were (5.7–6.3) · 10
4
s
)1
, which is slightly

higher than that of the wild-type protein
(3.2 · 10
4
s
)1
). The rate constants of the CO-depend-
ent slow phase were (29–44) · 10
)3
lm
)1
Æs
)1
, which is
slightly higher than that of the wild-type protein
(26 · 10
)3
lm
)1
Æs
)1
) (Table 4). Therefore, the kinetics
of CO binding was not markedly influenced by the
mutation of Leu99 or Leu115. These findings are in
contrast with the fact that mutations at the Met95 resi-
due, the distal axial ligand in the Fe(II) complex,
A
B
C
Fig. 6. Transient spectra accompanying CO binding to the wild-type
enzyme. (A) Difference absorption spectral changes of the wild-type

protein (8 l
M per heme) observed after flash photolysis. Spectra of
the fast phase (red) obtained 0.6 ls, every few microseconds, and
200 ls after a flash are selected, and those of the slow phase
(blue) obtained 0.2 ms, every few milliseconds and 80 ms after a
flash are selected. Arrows designate spectral changes observed
accompanied by the CO binding to the Fe(II) complex. Spectral
changes were composed of fast (red) and slow (blue) components
with different isosbestic points. (B) Difference spectral changes for
the fast phase (of the order of microseconds) with isosbestic points
at 414, 431, and 462 nm. (C) Difference spectral changes for the
slow phase (of the order of milliseconds) with isosbestic points at
396, 425, and 450 nm. (B) and (C) were separately extracted from
(A). Arrows in (B) and (C) designate the same as in (A).
Heme electronic states of Leu mutants of Ec DOS N. Yokota et al.
1216 FEBS Journal 273 (2006) 1210–1223 ª 2006 The Authors Journal compilation ª 2006 FEBS
markedly enhance the rate constants (Table 4) [5]. This
was surprising because other physicochemical proper-
ties such as the auto-oxidation rates and redox poten-
tials (see next section) of the Leu99 and Leu115
mutant and wild-type proteins differed significantly. A
possible explanation for these findings is that O
2
and
CO have different modes of ligand binding but share a
ligand access channel in Ec DosH.
Intermediate species generated by flash
photolysis
As shown in Fig. S2 of Supplementary material, we
observed spectral changes containing Soret peaks of

two intermediate complexes [complexes b (435-nm spe-
cies) and c (428-nm species)] and of the final Fe(II)–
CO complex [complex a (423-nm species)] in the differ-
ence spectra obtained from the absorbance before and
after flash photolysis. The complex changes in the opti-
cal absorption for the Fe(II)–CO complex after flash
photolysis were similar to those previously reported
for Ec DosH [16,18]. These previous studies proposed
that the following species are obtained by flash photo-
lysis. In the spectral changes for the fast phase, the
peak around 438 nm is ascribed to the 5-coordinated
intermediate complex (complex b in Fig. S2). The
results suggest that the rate of binding of Met95 to the
5-coordinated Fe(II) complex (His77 as an axial lig-
and; complex b in Fig. S2) is fast and occurs in the
order of microseconds, whereas the rate of CO binding
to the 6-coordinated Fe(II) complex (Met95 and His77
as axial ligands; complex c in Fig. S2) is slow and
occurs in the order of milliseconds because CO must
push the axial ligand out of the heme. For this slow
CO binding process, CO is probably dissociated and
moves to a position relatively far from the heme iron
(complex b in Fig. S2). In addition to this slow CO
binding process, the results indicate that the very fast
CO binding should occur on a nanosecond timescale.
For this very fast CO binding process, it is likely that
CO does not move away but is located very close to
the heme iron (complex d in Fig. S2). In this case, CO
may not have sufficient time to move away after disso-
ciation by flash photolysis, leading to the very fast

recombination. Indeed, ultrafast ligand rebinding to
Ec DosH has been reported [15,16].
A
B
Fig. 7. Optical spectral changes for the Fe(II)–CO complex of the
wild-type protein (8 l
M per heme) monitored at 422 nm after flash
photolysis. Spectral changes were composed of two phases, fast
(A) and slow (B) phases. Spectral changes for the slow phase were
dependent on the CO concentration (inset), and the rate was evalu-
ated as 26 · 10
)3
lM
)1
Æs
)1
.
Table 4. Rates for CO association with the wild-type and mutant
Ec DosH proteins as determined by the flash photolysis method.
The fast phases were independent of the CO concentration,
whereas the slow phases were dependent on the CO concentra-
tion. Note that the fast phase is associated with rebinding of
Met95 to the heme. At least three experiments were conducted to
obtain each rate constant. Experimental errors were less than
20%. ND, no data.
Fast phase
k
on
(· 10
4

s
)1
)
Slow phase
k
on
(· 10
)3
lM
)1
Æs
)1
) References
Wild-type 3.2 26 This study
– 7.8 [5]
– 1.1 [1]
L99T 6.3 29 This study
L99F 5.7 44 This study
L115T 6.3 41 This study
L115F ND
a
ND
a
This study
M95A – 9300 [5]
M95L – 3400 [5]
M95H – 6200 [5]
M95I – 1100 [5]
a
Measurement of the rate of CO binding was not feasible because

of low heme binding affinity and instability of the protein.
N. Yokota et al. Heme electronic states of Leu mutants of Ec DOS
FEBS Journal 273 (2006) 1210–1223 ª 2006 The Authors Journal compilation ª 2006 FEBS 1217
Redox potentials
We obtained both the reductive and oxidative poten-
tials for Ec DosH (Fig. 8 and Table 5). Both the
reductive [23.1–34.6 mV versus the standard hydrogen
electrode (SHE)] and oxidative potentials (17.7–
25.2 mV versus SHE) of the mutant proteins, except
for the L115F mutant, were significantly lower than
those of the wild-type protein (63.9 and 52.7 mV ver-
sus SHE, respectively). It is surprising that the redox
potential was decreased by the mutations at the Leu99
and Leu115 residues of Ec DosH because these resi-
dues are distant from the heme plane and do not have
direct contact with the heme iron. In our previous
studies, mutations at Met95 and Asp40 of Ec DosH
markedly changed the redox potentials (Table 4)
[8,13]. Because Met95 is an axial ligand, Met95 muta-
tions would be expected to give rise to marked changes
in the redox potential. Also, Asp40 indirectly interacts
with the proximal axial ligand, His77, of Ec DosH
through ionic interactions via two water molecules.
However, Leu99 and Leu115 have neither direct con-
tact via ionic interaction nor covalent interaction.
Leu99 should be located slightly farther from the heme
iron than Leu115, but despite this, the changes in the
redox potentials for the Leu99 mutants were larger
than that for the L115T mutant.
Intriguingly, the redox potentials for the Leu99 and

Leu115 mutants of Ec DosH are in the opposite direc-
tion from the Asp40 mutants, suggesting that the
effects of mutations on the redox potential at the distal
side differ from those at the heme proximal side. In
this sense, it is reasonable that the Leu99 and Leu115
mutants decreased the redox potentials to extents sim-
ilar to those observed for the Met95 mutations,
although the effects of the Leu99 and Leu115 muta-
tions were modest compared with those of Met95
mutations.
It seems likely that potential data based on our data
and that of others have accuracies of % 3–5 mV. Even
taking the accuracy into consideration, redox poten-
tials of Ec DosH proteins have an apparent hysteresis
in the data, being different between the reductive and
oxidative potentials (Table 5). The apparent hysteresis
was small for the L99T mutant compared with the
wild-type and other mutants. We do not know whether
the hysteresis simply reflects differences in equilibration
when data are recorded in reductive and oxidative
directions. However, it is possible that it is due to the
axial ligand switching between hydroxide anion and
Met95 accompanied by the redox change as demon-
A
B
Fig. 8. (A) Spectral changes of the L99F mutant accompanied by
reduction by sodium dithionite. Arrows designate changes of the
spectrum of the Fe(III) complex (black) to that of the Fe(II) complex
(red). (B) Electrochemical reductive (black open circle) and oxidative
(red filled circle) titrations of the L99F mutant of Ec DosH. Experi-

mental data (dotted lines) are fitted to the calculated lines.
Table 5. Redox potentials (mV versus SHE) of the wild-type and
mutant Ec DosH proteins. We speculate that accuracies are % 3-
5 mV based on data of ours and others [2,8,13]. ND, no data.
Reductive Oxidative References
Wild-type 63.9 (n ¼ 0.88) 52.7 (n ¼ 0.98) This study
70 63 [2]
67 – [13]
L99T 23.1 (n ¼ 0.92) 20.3 (n ¼ 1.01) This study
L99F 24.2 (n ¼ 1.11) 17.7 (n ¼ 1.01) This study
L115T 34.6 (n ¼ 0.93) 25.2 (n ¼ 1.08) This study
L115F
a
ND ND This study
M95L )1 – [8]
M95A )26 – [8]
M95H )122 – [8]
D40A 95 – [13]
D40N 114 – [13]
a
Measurements of redox potentials were not feasible because of
low heme binding affinity and instability of the protein.
Heme electronic states of Leu mutants of Ec DOS N. Yokota et al.
1218 FEBS Journal 273 (2006) 1210–1223 ª 2006 The Authors Journal compilation ª 2006 FEBS
strated from the crystal structure of this protein [6]. In
CooA, it was suggested that the hysteresis observed in
the redox titrations reflects the axial ligand switching
between Cys75 and His77 [25].
NO complexes
Because the catalytic activity of Ec DOS may be

strongly inhibited by NO binding to the heme iron [2],
we examined the optical absorption spectra of the NO
complex of the Ec DosH mutant proteins (Fig. 9 and
Table 6). For the Fe(III)–NO complexes of the mutant
proteins, we did not observe any significant differences
between the wild-type and mutant proteins. Also, by
adding sodium dithionite to the Fe(III) species, we
confirmed that auto-reduction of the Fe(III)–NO com-
plex to the Fe(II)–NO complex does not occur for the
mutant proteins even under anaerobic conditions,
which is consistent with the characteristics of the wild-
type protein. Notably, however, the Soret absorption
peak of the Fe(II)–NO complex of the L115F mutant
was at 399 nm, which is different from those of the
wild-type and other mutant proteins (421–423 nm).
The Soret peak position of the Fe(II)–NO heme com-
plex below or near 400 nm indicates a 5-coordinated
heme–NO complex [14].
It is interesting to note that the Fe(II)–NO com-
plex of the L115F mutant of Ec DosH is a 5-co-
ordinated NO–heme complex. The heme-sensor
enzyme, soluble guanylate cyclase, is activated by the
formation of the 5-coordinated NO–heme complex
[24]. A similar 5-coordinated NO–heme complex is
formed for other heme-sensor enzymes and proteins,
including CooA, cystathionine b-synthase, bacterial
cytochrome c¢, and heme-regulated eukaryotic initi-
ation factor 2a kinase [14,24 and references therein].
Many other heme proteins, including myoglobin,
hemoglobin, peroxidases, and cytochromes P450,

have 6-coordinated NO–heme complexes [21–24 and
references therein]. The protein structures on the
heme proximal side and ⁄ or the bond length between
the heme iron and the proximal ligand may contrib-
ute to the formation of the 5-coordinated NO–heme
complex. We therefore speculate that the bond
strength between the heme iron and His77 for the
L115F mutant Ec DosH is weaker than those of the
Ec DosH wild-type and other Leu99 and Leu115
mutant proteins, probably because of an indirect
effect from the L115F mutations trans to the prox-
imal side. Nitrite and nitrate anions have been impli-
cated as being important in signal transduction by
NO [17]. However, these anions did not change the
optical absorption spectra of the wild-type or mutant
proteins.
A
B
Fig. 9. Optical absorption spectra of the Fe(III)–NO (black) and
Fe(II)–NO (red) complexes of the wild-type (6.8 l
M per heme) (A)
and L115F (5.8 l
M per heme) (B) mutant proteins of Ec DosH. The
spectra of L99F and L115T are essentially the same as those of
the wild-type.
Table 6. Optical absorption maxima (nm) and millimolar absorption
coefficients (m
M
)1
Æcm

)1
) of the NO complexes of the wild-type and
mutant Ec DosH proteins. The millimolar absorption coefficients
(shown in parentheses) were determined using the pyridine hemo-
chromogen method.
Fe(III)–NO Fe(II)–NO
Soret baSoret ba
Wild-type 420 (133) 532 568 421 (108) 540 578
L99T 420 (124) 534 568 421 (109) 543 579
L99F 421 (135) 533 568 423 (120) 547 580
L115T 420 (140) 533 567 421 (78) 544 576
L115F 421 (163) 534 568 399 (64) – –
N. Yokota et al. Heme electronic states of Leu mutants of Ec DOS
FEBS Journal 273 (2006) 1210–1223 ª 2006 The Authors Journal compilation ª 2006 FEBS 1219
Summary
In this study, we examined the roles of the Leu99 and
Leu115 residues, which are located on the heme distal
side and form a hydrophobic pocket, in the physico-
chemical properties of the heme iron in Ec DosH. The
binding affinity of heme to the L115F mutant of
Ec DosH was significantly lower than that of the wild-
type enzyme. Substitution of the phenyl groups with
Leu115 substantially changed the heme binding char-
acteristics and caused a different heme coordination
structure, leading to a blue-shifted Soret band for the
Fe(III) complex and to formation of the 5-coordinated
NO–Fe(II) heme complex. Earlier studies on SmFixL
mutants suggested that Leu230 was one of the import-
ant residues for heme incorporation [20]. In contrast,
coordination structures of the L99T, L99F, and L115T

mutants were similar to that of the wild-type enzyme.
We further found that the rates of auto-oxidation and
the redox potentials were significantly changed by the
mutations of these two residues. It is interesting that
these physicochemical values were changed by the
mutations, because the mutated residues, Phe99 and
Leu115, are distant from the heme iron and do not
directly interact with the O
2
molecule by ionic interac-
tion. Surprisingly, however, the mutations at Leu99
and Leu115 did not significantly influence the kinetics
of O
2
or CO binding. Taken together, these results
suggest that Leu99 and Leu115 play significant roles in
determining the electronic states of the heme iron but
are not important in the mode of O
2
or CO binding
and ⁄ or the structure of the ligand access channel.
Experimental procedures
Materials
1-hydroxy-2-oxo-3-(N-methyl-3-aminopropyl)-3-methyl-1-tri-
azene was obtained from Dojindo (Kumamoto, Japan).
Restriction and modification enzymes were acquired from
Takara Bio, Toyobo (Osaka, Japan), New England Biolabs
(Beverly, MA, USA), and Nippon Roche (Tokyo, Japan).
Other chemicals were purchased from Wako Pure Chemi-
cals (Osaka, Japan).

Construction of the expression plasmids for
Ec DosH and Ec DosH mutants
Expression plasmids for the wild-type and Ec DosH mutant
proteins were constructed as described in our previous
reports [2,4]. Briefly, the sequence corresponding to
Ec DosH was amplified by PCR using genomic DNA isola-
ted from Escherichia coli strain JM109 as a template. The
clones obtained were inserted into the E. coli expression
vector pET 28a(+), which introduces a (His)
6
tag at the
N-terminus of the expressed proteins. Site-directed mutants
L99T, L99F, L115T, and L115F were constructed with a
QuikChange
TM
mutagenesis kit (Stratagene, La Jolla, CA,
USA) with pET28a(+)-wild-type Ec DosH as a template
and using the following respective 5¢-sense primers: 5¢-gatga
gtcgggagACCcagctggagaaaaaag-3¢,5¢-gatgagtcgggagTTTcag
ctggagaaaaaag-3¢,5¢-ggacccgttttgcgACCtcgaaagtgagc-3¢,
and 5¢-ggacccgttttgcgTTTtcgaaagtgagc-3¢.
Preparation of Ec DosH
The (His)
6
-tagged Ec DosH proteins (wild-type and L99T,
L99F, L115T, and L115F mutants) were expressed in
E. coli BL21 (DE3) and purified as described previously
[2,4]. Heme synthesis was induced with 5-aminolevulinate
(450 lm). After purification by 30–70% ammonium sulfate
fractionation and dialysis, the protein was subjected to

Ni
2+
⁄ nitrilotriacetate ⁄ agarose chromatography (Qiagen,
Valencia, CA, USA). Final purification was by Sephadex
G-75 column chromatography. SDS ⁄ PAGE and subsequent
staining with Coomassie Brilliant Blue R-250 revealed that
the purified protein was more than 95% homogeneous
(Supplementary Material, Fig. S1).
Optical absorption spectra
Experiments under aerobic conditions were performed on a
Shimadzu UV-2500 spectrophotometer maintained at
25 °C. Anaerobic spectral experiments were conducted on a
Shimadzu UV-1600 spectrophotometer at 15 °Cinan
anaerobic glove box (Hirasawa, Tokyo, Japan) under a
nitrogen atmosphere with an O
2
concentration below
50 p.p.m. Spectral experiments under anaerobic conditions
were also performed in a gas-tight capped cell
(10 · 10 mm) under a nitrogen atmosphere using a Shim-
adzu UV-2500 spectrophotometer maintained at 25 °C with
a temperature controller. Before the addition of O
2
, CO, or
NO, Ec DosH (8 lm) was reduced with sodium dithionite
(% 1mm)in50mm Tris ⁄ HCl (pH 7.5). Excess dithionite
was removed using a Sephadex G-25 column in the glove
box. The protein was added to the CO-saturated buffer.
The NO gas solution was prepared using 1-hydroxy-2-oxo-
3-(N-methyl-3-aminopropyl)-3-methyl-1-triazene as a nitro-

gen donor.
Stopped-flow method
Fe(II) Ec DosH proteins (containing nearly 10 lm heme)
were generated by reduction with sodium dithionite in
50 mm Tris ⁄ HCl (pH 7.5), and excess dithionite was
removed with a Sephadex G-25 column in the glove box.
To obtain the O
2
binding rates, protein solutions were rap-
Heme electronic states of Leu mutants of Ec DOS N. Yokota et al.
1220 FEBS Journal 273 (2006) 1210–1223 ª 2006 The Authors Journal compilation ª 2006 FEBS
idly mixed with 160–1300 lm O
2
solutions using an RSP-
1000 stopped-flow spectrophotometer (Unisoku Scientific
Instruments, Osaka, Japan) under anaerobic conditions at
25 °C. To obtain anaerobic conditions in the protein solu-
tion, the solution in the reservoir was saturated by purging
with N
2
for more than 30 min. The formation of O
2
com-
plexes was monitored at 429 nm. To avoid photodissocia-
tion, the intensity of the monitoring light was reduced as
much as 80% with a filter. The time course for O
2
binding
was monitored between 0 and 500 ms. Within this time
scale, O

2
binding curves for all proteins examined in this
study were biphasic and fitted well to double-exponential
functions. O
2
binding was analyzed between 0 and 500 ms
using the following equation:
DAbsðtÞ¼Ae
Àkt
þ Be
Àkt
ð1Þ
where DAbs is the total intensity changes at a certain time,
t, after mixing, A and B are initial intensities for each
phase, and k is the Boltzman constant. The rates of O
2
recombination were plotted against the concentration of
O
2
. Both phases were dependent on the O
2
concentration.
EDTA had no significant effect on the ligand binding kinet-
ics (data not shown). At least three experiments were con-
ducted to obtain each rate constant. Regression analyses
were performed, and lines representing an optimal correla-
tion coefficient were selected.
Auto-oxidation rate
To obtain auto-oxidation rates, solutions of the Fe(III)
Ec DosH proteins (containing % 10 lm heme) were reduced

with sodium dithionite in 50 mm Tris ⁄ HCl (pH 7.5) con-
taining 1 mm EDTA. Excess dithionite was removed using
a Sephadex G-25 column in the glove box. Next, the pro-
tein solution was mixed in a cuvette with air-saturated
buffer to form the Fe(II)–O
2
complex. Changes in the
UV ⁄ visible spectra were monitored with respect to time at
25 °C using a UV-2500 spectrometer (Shimadzu). The
effects of EDTA on the rates of auto-oxidation were also
evaluated, but no significant differences were observed
between samples in the presence and absence of EDTA.
Laser flash photolysis
Laser flash photolysis experiments for the Fe(II)–CO com-
plexes were performed at 25 °C in a gas-tight capped cell
(10 · 10 mm). We used the second harmonic light (532 nm)
of an Nd-YAG laser (Surelite II-10; Continuum, Santa
Clara, CA, USA) with the fundamental radiation of
1064 nm. The monitoring light was produced using a
150 W xenon lamp (Hamamatsu Photonics, Hamamatsu,
Japan). The peak power of the laser was 10 mJ with a pulse
width of 6 ns (a repetition rate of 10 Hz) from a xenon
lamp with the light intensity reduced by as much as 80% at
wavelengths below 380 nm. Details of this equipment are
given elsewhere [5,11]. For the change from the Met95–
Fe(II)–His77 complex to the CO–Fe(II)–His77 complex,
recovery of the Ec DosH–CO complex was monitored with
continuous light from a xenon lamp (422 nm) using an
oscilloscope (inirinium; Hewlett–Packard). We also monit-
ored the absorption decrease at 444 nm, but the change

from the Fe(II)–His77 complex to the Met95–Fe(II)–His77
was not observed because of the proximity to the isosbestic
point. First-order rate constants were calculated using igor
pro software (WaveMetrics, Lake Oswego, OR, USA).
Approximately 10–20% of bound CO was dissociated upon
laser photolysis. Experiments were performed at least twice
for each complex. Data obtained from five or more points
were reproducible within a 20% margin of experimental
error. Transient absorptions of the bimolecular association
of CO and reduced Ec DosH were the average of 64 meas-
urements. The time course for the 5-coordinated to 6-
coordinated complex was monitored between 0 and 80 ls,
whereas that for the CO rebinding was monitored between
0 and 80 ms. Within this time scale, CO rebinding to all
proteins used in this study was monophasic and fitted well
to single-exponential functions. CO rebinding was analyzed
between 0 and 100 ls using eqn (1). The apparent rate
constants of CO recombination were plotted against the
concentration of CO, and were dependent on the concen-
tration of CO.
Redox potentials
Redox potentials were determined as previously reported
[2,8,13].
Acknowledgements
We are grateful to Dr Tokiko Suzuki-Yoshimura for
helpful discussions at the initial stages of this work.
This work was supported in part by Grants-in-Aid
from the Ministry of Education, Culture, Sports,
Science and Technology of Japan to H.K. and T.S.
References

1 Delgado-Nixon VM, Gonzalez G & Gilles-Gonzalez
MA (2000) Dos, a heme-binding PAS protein from
Escherichia coli, is a direct oxygen sensor. Biochemistry
39, 2685–2691.
2 Sasakura Y, Hirata S, Sugiyama S, Suzuki S, Taguchi
S, Watanabe M, Matsui T, Sagami I & Shimizu T
(2002) Characterization of a direct oxygen sensor
heme protein from Escherichia coli. Effects of the
heme redox states and mutations at the heme-binding
site on catalysis and structure. J Biol Chem 277,
23821–23827.
N. Yokota et al. Heme electronic states of Leu mutants of Ec DOS
FEBS Journal 273 (2006) 1210–1223 ª 2006 The Authors Journal compilation ª 2006 FEBS 1221
3 Sato A, Sasakura Y, Sugiyama S, Sagami I, Shimizu T,
Mizutani Y & Kitagawa T (2002) Stationary and time-
resolved resonance Raman spectra of His77 and Met95
mutants of the isolated heme domain of a direct oxygen
sensor from Escherichia coli. J Biol Chem 277, 32650–
32658.
4 Yoshimura T, Sagami I, Sasakura Y & Shimizu T (2003)
Relationships between heme incorporation, tetramer
formation, and catalysis of a heme-regulated phos-
phodiesterase from Escherichia coli: a study of deletion
and site-directed mutants. J Biol Chem 278, 53105–
53111.
5 Taguchi S, Matsui T, Igarashi J, Sasakura Y, Araki Y,
Ito O, Sugiyama S, Sagami I & Shimizu T (2004) Bind-
ing of oxygen and carbon monoxide to a heme-regu-
lated phosphodiesterase from Escherichia coli. Kinetics
and infrared spectra of the full-length wild-type enzyme,

isolated PAS domain, and Met-95 mutants. J Biol Chem
279, 3340–3347.
6 Kurokawa H, Lee DS, Watanabe M, Sagami I, Mikami
B, Raman CS & Shimizu T (2004) A redox-controlled
molecular switch revealed by the crystal structure of a
bacterial heme PAS sensor. J Biol Chem 279, 20186–
20193.
7 Sasakura Y, Yoshimura-Suzuki T, Kurokawa H &
Shimizu T (2005) Structure–function relationships of
Ec DOS, a heme-regulated phosphodiesterase from
Escherichia coli. Acc Chem Res 39, 37–43.
8 Hirata S, Matsui T, Sasakura Y, Sugiyama S, Yoshim-
ura T, Sagami I & Shimizu T (2003) Characterization of
Met95 mutants of a heme-regulated phosphodiesterase
from Escherichia coli. Optical absorption, magnetic cir-
cular dichroism, circular dichroism, and redox poten-
tials. Eur J Biochem 270, 4771–4779.
9 Watanabe M, Matsui T, Sasakura Y, Sagami I & Shi-
mizu T (2002) Unusual cyanide bindings to a heme-
regulated phosphodiesterase from Escherichia coli: effect
of Met95 mutations. Biochem Biophys Res Commun 299,
169–172.
10 Park HJ, Suquet C, Satterlee JD & Kang CH (2004)
Insights into signal transduction involving PAS domain
oxygen-sensing heme proteins from the X-ray crystal
structure of Escherichia coli Dos heme domain
(Ec DosH). Biochemistry 43, 2738–2746.
11 Antonini E & Brunori M (1971) Hemoglobin and Myo-
globin in Their Reactions with Ligands. Elsevier ⁄ North-
Holland Biomedical Press, Amsterdam.

12 D’Souza F, Smith PM, Zandler ME, McCarty AL, Itou
M, Araki Y & Ito O (2004) Energy transfer followed by
electron transfer in a supramolecular triad composed of
boron dipyrrin, zinc porphyrin, and fullerene: a model
for the photosynthetic antenna-reaction center complex.
J Am Chem Soc 126, 7898–7907.
13 Watanabe M, Kurokawa H, Yoshimura-Suzuki T,
Sagami I & Shimizu T (2004) Critical roles of Asp40 at
the haem proximal side of haem-regulated phosphodies-
terase from Escherichia coli. Redox potential, auto-
oxidation and catalytic control. Eur J Biochem 271,
3937–3942.
14 Igarashi J, Sato A, Kitagawa T, Yoshimura T, Yamau-
chi S, Sagami I & Shimizu T (2004) Activation of heme-
regulated eukaryotic initiation factor 2a kinase by nitric
oxide is induced by the formation of a five-coordinate
NO-Heme complex: optical absorption, electron spin
resonance, and resonance Raman spectral studies. J Biol
Chem 279, 15752–15762.
15 Liebl U, Bouzhir-Sima L, Ne
´
grerie M, Martin JJ & Vos
MH (2002) Ultrafast ligand rebinding in the heme
domain of the oxygen sensors FixL and Dos: general
regulatory implications for heme-based sensors. Proc
Natl Acad Sci USA 99, 12771–12776.
16 Liebl U, Bouzhir-Sima L, Kiger L, Marden MC,
Lambry JC, Ne
´
grerie M & Vos MH (2003) Ligand

binding dynamics to the heme domain of the oxygen
sensor Dos from Escherichia coli. Biochemistry 42,
6527–6535.
17 Gladwin M, Schechter AN, Kim-Shapiro DB, Patel RP,
Hogg N, Shiva S, Cannon RO, III, Kelm M, Wink DA,
Graham M, et al. (2005) The emerging biology of the
nitrite anion. Nat Chem Biol 1, 308–314.
18 Gonzalez G, Dioum E, Bertolucci CM, Tamita T,
Ikeda-Saito M, Cheesman MR, Watmough NJ &
Gilles-Bonzalez MA (2002) Nature of the displaceable
heme-axial residue in the Ec DOS protein, a heme-
based sensor from Escherichia coli. Biochemisty 41,
8414–8421.
19 Balland V, Bouzhir-Sima L, Kiger L, Marden MC, Vos
MH, Liebl U & Mattioli TA (2005) Role of arginine
220 in the oxygen sensor FixL from Bradyrhizobium
japonicum. J Biol Chem 280, 15279–15288.
20 Nakamura H, Saito K, Ito E, Tamura K, Tsuchiya
T, Nishigaki K, Shiro Y & Iizuka T (1998) Identifica-
tion of the hydrophobic amino acid residues required
for heme assembly in the rhizobial oxygen sensor pro-
tein FixL. Biochem Biophys Res Commun 247, 427–
431.
21 O’Keeffe DH, Ebel RE & Peterson JA (1978) Studies of
the oxygen binding site of cytochrome P-450. Nitric
oxide as a spin-label probe. J Biol Chem 253, 3509–
3516.
22 Scholler DM, Wang MY & Hoffman BM (1979) Reso-
nance Raman and EPR of nitrosyl human hemoglobin
and chains, carp hemoglobin, and model compounds.

Implications for the nitrosyl heme coordination state.
J Biol Chem 254, 4072–4078.
23 Quaroni LG, Seward HE, McLean KJ, Girvan HM,
Ost TW, Noble MA, Kelly SM, Price NC, Cheesman
MR, Smith WE, et al. (2004) Interaction of nitric oxide
with cytochrome P450 BM3. Biochemistry 43, 16416–
16431.
Heme electronic states of Leu mutants of Ec DOS N. Yokota et al.
1222 FEBS Journal 273 (2006) 1210–1223 ª 2006 The Authors Journal compilation ª 2006 FEBS
24 Boon EM & Marletta MA (2005) Ligand discrimination
in soluble guanylate cyclase and H-NOX family of heme
sensor proteins. Curr Opin Chem Biol 9, 441–446.
25 Aono S (2003) Biochemical and biophysical properties
of the CO-sensing transcriptional activator CooA. Acc
Chem Res 36, 825–831.
26 Klatt P, Schmidt K, Werner ER & Mayer B (1996)
Determination of nitric oxide synthase cofactors: heme,
FAD, FMN, and tetrahydrobiopterin. Methods Enzymol
268, 358–365.
Supplementary material
The following supplementary material is available
online:
Fig. S1. SDS ⁄ PAGE for the wild-type (1), L99T (2),
L99F (3), L115T (4), and L115F (5) mutants of
Ec DosH. Molecular mass markers are shown at both
ends (M). Approximately 1–2 lg protein was loaded.
Fig. S2. Proposed intermediate heme complexes
formed during flash photolysis for the Fe(II)–CO com-
plexes [16,18]. The 6-co-ordinated Fe(II)–CO complex
(a) and Fe(II) (c) complexes have Soret peaks at 423

and 428 nm, respectively, whereas the 5-coordinated
Fe(II) complex (His77 as an axial ligand) (b) has a
Soret peak at 435 nm. However, when CO is close to
the heme iron for the 5-coordinated Fe(II) complex
(d), CO rebinding should be very fast (less than nano-
seconds); therefore, we were unable to monitor the
spectral changes.
This material is available as part of the online article
from
N. Yokota et al. Heme electronic states of Leu mutants of Ec DOS
FEBS Journal 273 (2006) 1210–1223 ª 2006 The Authors Journal compilation ª 2006 FEBS 1223