X-ray crystallography, CD and kinetic studies revealed the essence
of the abnormal behaviors of the cytochrome
b
5
Phe35fiTyr mutant
Ping Yao
1
, Jian Wu
2
, Yun-Hua Wang
1
, Bing-Yun Sun
1
, Zong-Xiang Xia
2
and Zhong-Xian Huang
1
1
Chemical Biology Laboratory, Department of Chemistry, Fudan University, Shanghai, People’s Republic of China;
2
State Key Laboratory of Bio-organic and Natural Products Chemistry, Shanghai Institute of Organic Chemistry,
Chinese Academy of Sciences, Shanghai, People’s Republic of China
Conserved phenylalanine 35 is one of the hydrophobic patch
residues on the surface of cytochrome b
5
(cyt b
5
). This patch
is partially exposed on the surface of cyt b
5
while its buried

been studied by CD for the first time, revealing the existence
of the folding intermediate.
Keywords: cytochrome b
5
; folding; mutagenesis; stability;
structure.
Cytochrome b
5
(cyt b
5
) is a membrane-bound hemoprotein.
It consists of a water-soluble, heme-containing domain and
a short hydrophobic tail of approximate 40 amino acid
residues that anchors the protein to the microsomal
membrane [1]. The water-soluble domain functions as an
electron mediator in the cytochrome P450 reductase system
[2] and in the fatty acid desaturation system [3], etc. In
erythrocytes, cyt b
5
also exists as a soluble heme-binding
protein lacking the hydrophobic tail where its physiological
role is to reduce methemoglobin [4].
On the surface of cyt b
5
, there is a cluster of negatively
charged residues surrounding the exposed heme edge. These
acidic residues have been proved to bind to the basic
residues of the protein redox partners, such as cytochrome c
[5,6], cytochrome P450 [7], metmyoglobin [8] and methe-
moglobin [9]. On the surface of cyt b

obviously more stable towards heat and chemical denatur-
ationthanwild-typecytb
5
[13]. We also studied electron
transfer reactions of cyt b
5
Phe35fiTyr and Phe35fiLeu
variants with cytochrome c, with the wild-type and the
Tyr83Phe, Tyr83Leu variants of plastocyanin, and with the
inorganic complexes [Fe(EDTA)]
–
,[Fe(CDTA)]
–
and
[Ru(NH
3
)
6
]
3+
. The change at Phe35 of cyt b
5
did not affect
the second-order rate constants of the electron transfer
Correspondence to Z X.Huang,ChemicalBiologyLaboratory,
Department of Chemistry, Fudan University, Shanghai 200433,
China. Fax: + 86 21 65641740, Tel.: + 86 21 65643973,
E-mail:
Z X. Xia, State Key Laboratory of Bio-organic and Natural Products
Chemistry, Shanghai Institute of Organic Chemistry,

theessentialdifferencebetweenthewild-typeandmutant
proteins and to give a proper interpretation. In this paper,
the secondary structural changes of cyt b
5
and its Phe35fi
Tyr mutant towards heat have been characterized by CD.
Meanwhile, the heme dissociation and transfer reactions also
provide a good means of examining the subtle local
conformation changes around the heme group under natural
conditions. Therefore, the heme dissociation kinetics at
different urea concentrations and the heme transfer reactions
between the wild-type cyt b
5
or its Phe35fiTyr mutant and
apo-myoglobin (Mb) were studied to demonstrate the
affinity changes of the heme with cyt b
5
polypeptide chain.
In this paper the crystal structure of the cyt b
5
Phe35fiTyr
mutant has been determined by X-ray analysis. Based on the
molecular structure and the above detailed studies the
essence of these unusual behaviors is discussed.
MATERIALS AND METHODS
Protein preparation
Bovine liver cyt b
5
and its mutants were prepared and
purified as described previously [13]. The concentrations of

[20], respectively.
X-ray analysis of cytochrome
b
5
Phe35fiTyr mutant
Crystallization. Single crystals of the Phe35fiTyr mutant
of trypsin-solubilized bovine liver microsomal cytochrome
b
5
(Tb
5
) were grown by the vapor diffusion method in
hanging drops containing 10 mgÆmL
)1
protein solution in
3.1–3.2
M
phosphate buffer (pH 7.5) at 20 °C. This is
similar to the crystallizing condition used for wild-type Tb
5
[21] and lipase-solubilized bovine liver microsomal cyto-
chrome b
5
(Lb
5
) [22]. The typical size of the single crystals
was  0.6 · 0.5 · 0.3 mm. Crystals of wild-type Tb
5
[21]
and the Tb

packages
X
-
PLOR
[25] and
CNS
[26] successively on a Silicon
Graphics Indigo 2 workstation. All the data up to 1.8 A
˚
were used for structural refinement at the
CNS
refinement
stage. A random sample of 10% of the X-ray data was
excluded from the refinement and was taken as the test data
set, and the agreement between the calculated and observed
structure factors of the test data set was monitored
throughout the course of the refinement. The graphics
software
TURBO
-
FRODO
[27] was used for the model
rebuilding.
The initial structural model of the Phe35fiTyr mutant
was determined using the difference Fourier method based
on the crystal structure of the Val61His mutant of cyt b
5
at
2.1 A
˚

2
were removed from the final model.
The structure was further refined by using the more
powerful program package
CNS
. The simulated annealing
refinement starting from 2500 K with a cooling rate of 25 K
per cycle was carried out, followed by the individual
temperature factor refinement.
Thermal denaturation of cyt
b
5
monitored by CD
CD spectra of cyt b
5
and its variants were recorded with a
Jasco J-715 spectropolarimeter equipped with a Naslab
temperature controller. The path length was 0.1 cm in the
190–250 nm region and 1 cm in the 250–500 nm region,
Table 1. Crystal data and data collection statistics.
Space group C2
Cell dimensions
a(A
˚
) 70.71
b(A
˚
) 40.39
c(A
˚

The numbers in the
parentheses correspond to the data in the highest resolution shell
(1.80–1.84 A
˚
).
c
Mean signal-to-noise ratio.
4288 P. Yao et al. (Eur. J. Biochem. 269) Ó FEBS 2002
respectively. The ellipticity was recorded at 100 nmÆmin
)1
speed, 0.2 nm resolution, five accumulations, 1.0 nm
bandwidth. Cyt b
5
or its mutant was dissolved in the
phosphate buffer (100 m
M
pH 7.0). The protein concentra-
tions were 25 l
M
in the 190–250 nm region and 12.5 l
M
in
the 250–500 nm region, respectively. At each given tem-
perature, the protein sample was allowed to equilibrate for
20 min before the spectrum was recorded. The temperature
was increased stepwise over the range 30–95 °Candthe
temperature accuracy was within ± 0.1 °C.
Urea- and guanidine hydrochloride-mediated
denaturation of cyt
b

5
and apo-myoglobin:
CD spectroscopy. Thetransferofhemefromcytb
5
to apo-
Mb was examined in the 190–250 nm and 250–500 nm
regions separately (10 m
M
sodium acetate buffer, pH 5.5,
room temperature). Equal volumes of cyt b
5
and apo-Mb
were mixed at a final concentrations of 25 l
M
and 30 l
M
for
cyt b
5
and apo-Mb, respectively. The spectrum recording
conditions were the same as described above.
UV–visible spectroscopy. Kinetic analysis of heme disso-
ciation from the wild-type and the mutants of cyt b
5
were
performed as described by Hargrove et al.[30].Theheme
transfer reaction was monitored with a HP 8452A diode-
array spectrophotometer. The temperature was controlled
at ± 0.1 °C with a Neslab RTE-5B circulating bath
instrument. The reaction was initiated by rapidly mixing

)1
)andthe
first step is the rate-determining step for the whole reaction
[31], the heme transfer reaction from cyt b
5
to apo-Mb
could be treated as a first-order reaction. The kinetic trace
can be described mathematically by the equation
DA
t
¼ DA
eq
(1–e
–kt
)whereDA
t
is the increase in absorbance
at time t, DA
eq
is the increase in absorbance at equilibrium,
and k is the rate constants for heme transfer.
The activation energy of the heme transfer reaction
was obtained by measuring the rate constant over the
temperature range of 20–37 °C(10m
M
sodium acetate
buffer pH 5.5). The activation free energy was calculated
from the equation [32,33] k ¼ k
B
T/h exp(– DG°

PROCHECK
[34]. The Luzzati plot
shows that the estimated error of the refined coordinates is
 0.21 A
˚
.
Fig. 1 shows the electron density of Tyr35 and the heme
group in the Phe35fiTyr mutant. The overall structure of
the Phe35fiTyr mutant is basically the same as that of the
wild-type Tb
5
. The r.m.s. deviation for a total of 82 Ca
atoms between the two molecules is 0.07 A
˚
. The secondary
structures of the wild-type protein and its Phe35fiTyr
mutant are the same. Fig. 2A and B shows a part of the
heme-binding pocket of the Phe35fiTyr mutant in two
different views. In wild-type cyt b
5
, the residue Phe35 is
located at helix II, which is a part of the heme-binding
pocket of cyt b
5
, and its side-chain points toward the heme.
The mutation from the nonpolar residue Phe35 to the polar
residue Tyr35 makes slight changes in the side chain
conformation of this residue. The shift of the Ca atom of
Tyr35 of the Phe35fiTyr mutant from that of Phe35 of the
wild-type cyt b

(Eur. J. Biochem. 269) 4289
aromatic rings is 0.45 A
˚
, i.e., the distance from the atom CZ
(Fig. 1) of Tyr35 to that of Phe35. The crystal structure of
the Phe35fiTyr mutant shows that the side chain of Tyr35
makes strong van der Waals’ contacts with the heme, and
the shortest distance is 3.21 A
˚
, i.e., from the phenol oxygen
atom of Tyr35 to the carbon atom CHB (Fig. 1) of the
heme. In addition, the hydroxyl group of Tyr35 forms a
hydrogen bond (2.86 A
˚
) to a water molecule located outside
the heme pocket, as shown in Fig. 2A. This water molecule
forms another hydrogen bond (2.79 A
˚
) with the atom ND1
of the His26 side chain in a symmetry-related molecule
(Fig. 2A). This water molecule was also found in the
structure of wild-type cyt b
5
as well as in other mutants.
When Phe35 is mutated to Tyr35, this water molecule
moves toward the hydroxyl group of Tyr35 by 0.43 A
˚
,and
the side chain conformation of His26 correspondingly
moves a little bit (for example, the atom ND1 of His26

5
.
Those in the mutant are shown as thick lines
and large spheres, and those in wild-type
cyt b
5
are shown as thin lines and small
spheres. (His26 # of the wild-type cyt b
5
is very
close to that in the mutant and cannot be
seen).
Fig. 1. Stereo view of the (2Fo-Fc) electron
density of Tyr35 and heme in the Phe35fiTyr
mutant, contoured at 1.0 r. The atoms CHB
of heme as well as OH and CZ of Tyr35
are labeled. This diagram was prepared using
the graphics program
TURBO
-
FRODO
.
4290 P. Yao et al. (Eur. J. Biochem. 269) Ó FEBS 2002
conformation of the heme in the Phe35fiTyr mutant is
basically the same as that in wild-type cyt b
5
. One of the two
propionates is hydrogen bonded to the main- and side chain
atoms of Ser64, while the other one extends into the solvent
and does not form any hydrogen bond with the protein

250–500 nm region at 30 °C, 67 °C, 69.5 °C, 75 °C, 85 °C,
and 95 °C, respectively. The CD spectra of the Phe35fiTyr
mutant have a similar pattern and are not shown here. In
the Soret region, the peak positions of the two proteins are
basically similar at 30 °C, consistent with those reported in
literatures [35,36]. At the near UV region, the negative CD
peak at 268 nm derived from the four tyrosyl residues of
wild-type cyt b
5
[35] shows a different shape for the
Phe35fiTyr mutant, which has five tyrosyl residues. The
peak at 299.4 nm derived from the single tryptophan
residue for the wild-type protein shifts to 297.6 nm for the
mutant. The spectra in the 267–299.4 nm region are only
slightly different for these two proteins.
Thermal denaturation of wild-type cyt b
5
and its
Phe35fiTyr mutant show similar CD behavior. A negative
peak at 418 nm with strong intensity and a positive peak at
390 nm at room temperature are characteristic of low-spin
state of ferric cyt b
5
[36]. When the temperature was
increased the negative peak at 418 nm was blue-shifted with
a gradual reduction of its intensity. Simultaneously, the
intensity of the positive peak at 390 nm decreased. We
found that with increasing temperature to 69.5 °C, a new
peak around 398 nm with a negative intensity appeared.
The intensity of the peak at 398 nm increased dramatically

aromatic components of the core 2 of cyt b
5
. The pattern of
the absorption changes around 267 nm and 299.4 nm
implies that even though core 2 is largely intact after the
removal of the heme from the protein as reported by
Falzone et al. [38] core 2 experiences significant structural
fluctuation and gradually undergoes complete unfolding.
This study clearly shows the whole process of unfolding and
is an important supplement to the results reported by Pfeil
[39] by means of second derivative spectra and heat capacity
of apo- and holo-cyt b
5
.
Fig. 4A demonstrates the transitional CD curves of
wild-type cyt b
5
monitored at 222 nm, 299 nm, 398.4 nm
and 418.8 nm. The curves of 222 nm, 299 nm and 418.8 nm
possess a similar pattern suggesting that dissociation of the
Fig. 3. CD spectra of the wild-type cyt b
5
from 30 °Cto95°Cat(A)
195–250 nm and (B) 250–500 nm (for clarity of comparison, only part of
the spectra are shown.)
Ó FEBS 2002 Mutation at Phe35 of cytochrome b
5
(Eur. J. Biochem. 269) 4291
Fe–His bond is accompanied by the a-helix unfolding of the
peptide chain and the destroying of Trp22 asymmetrical

constant increased sharply after the urea concentration
exceeded 5
M
. These results reflect the tightness of the heme
attaching to the polypeptide of cyt b
5
. For the Phe35fiTyr
mutant the heme pocket traps the heme even more strongly
than the wild-type protein. Obviously, for the Phe35fiLeu
mutant the interactions between the heme and its pocket are
much weaker, only a moderate concentration of urea is
needed to speed up the release of heme from the pocket.
The heme transfer from cyt
b
5
or its Phe35fiTyr mutant
to apo-Mb
The kinetic parameters of heme dissociation from cyt b
5
were determined under nondenaturation conditions by
measuring the spontaneous release of the heme from cyt b
5
to apo-Mb, which is used as a heme trap. Although the CD
spectra could show the reaction process clearly, the protein
concentration required is much higher than for the
UV–visible method. Because the high concentration of
protein could cause denaturation of the apo-protein during
the long assay time, all of the heme transfer reactions were
monitored only by the UV–visible spectra. The rates of
heme transfer reaction from the wild-type or the Phe35fi

spectra of cyt b
5
, there is a negative peak at 418 nm with
strong intensity [36]. In contrast, Mb has a strong positive
absorption at 408 nm [41]. Hence, the heme transfer reaction
from cyt b
5
to apo-Mb could be easily and precisely
Fig. 4. The transitional curves of the CD spectra on heating at
222 nm, 299 nm, 398.4 nm, and 418.8 nm. (A) Wild-type cyt b
5
.
(B) Phe35fiTyr mutant of cyt b
5
.
Fig. 5. The rate constants of heme dissociation of cyt b
5
as function of
urea concentration.
4292 P. Yao et al. (Eur. J. Biochem. 269) Ó FEBS 2002
monitored by CD which clearly demonstrates that the heme
transfer reaction under the conditions used proceeded to
completion (data not shown). Meanwhile, from the concen-
tration changes of the holo-Mb in the reaction monitored by
UV–visible spectroscopy, the same conclusion ) that this
reaction is entirely completed ) can be drawn.
DISCUSSION
Protein folding studied by CD spectra
Up to now, CD spectra of cyt b
5

present at significant concentrations [28,44]. It was
thought that the heme-binding domain of cyt b
5
was
denatured simultaneously with heme dissociation. The
UV–visible spectrum study of cyt b
5
in response to heat
and urea did display several isosbestic points in the
absorbance curves, and the denaturation curves really
showed that the denaturation followed the two-state
mechanism.
The denaturation curves of CD absorption at 222 nm,
299 nm and 418.8 nm shown in Fig. 4A and B indicate that
unfolding of the a-helices, b-sheets and breaking of the His–
Fe bonds of the heme follow the two-state mechanism. It is
noted that a new absorption peak that appeared at
398.4 nm displays slightly different denaturation behav-
iours. Definitely, the absorption at 398.4 nm is derived from
a heme derivative. As heme is a symmetrical chromophore,
it exhibits no inherent optical activity itself [45,46]. Our
experiment also shows that heme in the buffer solution itself
does not exhibit any CD absorption in the region of 250–
500 nm at 30–95 °C. Apo-cyt b
5
has no CD absorption in
the Soret band too, but shows the absorption contributed
from aromatic amino acids in the near UV region [35]. The
concurrent existence of the Soret band absorption at
418.8 nm and 398.4 nm at 69.5 °CshowninFig.3B

E
a
(kJÆmol
)1
) 110.4 135.4
a
T ¼ 25 ± 0.1 °C.
Fig. 6. Kinetic traces for heme transfer reaction from the wild-type, or
Phe35fiTyr mutant of cyt b
5
to apo-myoglobin. (A) Experimental data.
(B) Fitted curve.
Ó FEBS 2002 Mutation at Phe35 of cytochrome b
5
(Eur. J. Biochem. 269) 4293
with breaking of the His–Fe bonds. It is known that the
apo-cyt b
5
prepared under mild conditions could generally
maintain the holo-like structures except for some confor-
mational fluctuations observed in the local regions [47].
However, as indicated by molecular dynamics simulations
all a-helices in core 1 are highly mobile, and the tertiary
structure in core 2 of cyt b
5
is rather rigid [48]. Thus, the
denaturation curve of the wild-type protein monitored at
398.4 nm and 67–75 °C by CD implied that there was
probably a collapse of core 1 accompanied by partially
unfolding of the a-helices and breaking of Fe–His bonds.

normally Ôthe heme iron is ligated axially by the side chains
of Met7 and His102. It is likely that one of these ligands
remains attached to the heme in the unfolded stateÕ [49,50].
Here, we provide the detailed CD spectra evidencing the
existence of the intermediate and a reasonable explanation.
The reason why our results do not agree with those
obtained for the rabbit liver cyt b
5
[36] is not yet known.
But, it is noted that the bovine liver Tb
5
used in this work is
more stable than rabbit liver cyt b
5
.TheT
m
(transition
midpoint of the heat denaturation curve of the UV–visible
spectrum at 412 nm) is 66.9 °C for bovine liver Tb
5
and
55.0 °C for rabbit liver cyt b
5
[13,36]. Maybe a detailed
structural study, similar to the comparison between the
microsomal cyt b
5
and the outer membrane liver mitochon-
dria cyt b
5

3.3 kJÆmol
)1
more stable than the wild-type protein
towards heat denaturation and is 4.3 kJÆmol
)1
more stable
in urea denaturation. The CD spectra of heat denaturation
also show that the structure transition temperature for the
Phe35fiTyr mutant is higher than that for the wild-type.
Kinetically, the rate constant of heme transfer reactions
from cyt b
5
to apo-Mb for the wild-type protein is 10 times
faster than that for the Phe35fiTyr mutant. The urea-
mediated heme dissociation reactions of various cyt b
5
variants also demonstrate that the heme is trapped in the
heme pocket with different degrees of tightness. Recently,
Silchenko et al. [55] found that cyt b
5
from the outer
mitochondrial membrane of rat liver is substantially more
stable against thermal and chemical denaturation than
bovine liver cyt b
5
. Their study demonstrated that the
enhanced stability of outer mitochondrial membrane cyt b
5
is in large part due to slow heme release, where the heme is
kinetically trapped in the heme pocket of hemoproteins. As

hydrophilicity of the histidine residue, the side chain
volume decreases by 36 A
˚
3
compared to the wild-type
cyt b
5
which would effectively reduce the van der Waals’
contact between the histidine and the heme. So, the
Phe35fiHis mutant is 11.8 kJÆmol
)1
less stable than the
wild-type protein [13].
There is a stabilization effect of the heme ring binding to
Phe35 and Phe58 by hydrophobic aromatic interactions. It
has been reported that an edge-to-face orientation between
two aromatic groups is energetically favorable [57].
Sakamoto et al. [45] have studied the effect of amino acids
substitution of hydrophobic residues on heme-binding
properties in the designed two-a-helix peptides. Their
studies demonstrated that the edge-to-face interactions
between the aromatic side chain of the phenylalanine
residues and the porphyrin plane might contribute to the
conformation of peptide–heme conjugates. They also
4294 P. Yao et al. (Eur. J. Biochem. 269) Ó FEBS 2002
proved that the phenylalanine residue located at i ±4
relative to the axial ligand histidine residue in the a-helix was
critical to the edge-to-face interaction between the phenyl-
alanine side chain and the porphyrin ring, providing
stabilization of peptide–heme conjugates [45,46,58]. The

conserved water molecule shown in the crystal structure of
the Phe35fiTyr mutant enhances significantly hydrophilic
influence on the heme causing great alteration of the protein
properties [59].
ACKNOWLEDGMENTS
This work was supported by two grants from the National Natural
Science Foundation of China. We are grateful to Prof. Li-Wen Niu,
Prof. Mai-Kun Teng and Dr Xue-Yong Zhu of the University of
Science and Technology of China for their support and help with the
X-ray data collection.
REFERENCES
1. Spatz, L. & Strittmatter, P. (1971) A form of cytochrome b
5
that
contains an additional hydrophobic sequence of 40 amino acid
residues. Proc. Natl. Acad. Sci. USA 68, 1042–1046.
2. Bonfils, C., Balny, C. & Maurel, P. (1981) Direct evidence for
electron transfer from ferrous cytochrome b
5
to the oxyferrous
intermediate of liver microsomal cytochrome P-450 LM2. J. Biol.
Chem. 256, 9457–9465.
3. Strittmatter, P., Spatz, L., Corcoran, D., Rogers, M.J., Setlow, B.
& Redline, R. (1974) Purification and properties of rat liver
microsomal stearyl coenzyme A desaturase. Proc. Natl Acad. Sci.
USA 71, 4565–4569.
4. Hegesh, E., Hegesh, J. & Kaftory, A. (1986) Congenital methe-
moglobinemia with a deficiency of cytochrome b
5
. N. Engl. J.

5
and subunits of methemoglobin.
J. Biol. Chem. 258, 7369–7373.
10. Mathews, F.S. & Czerwinski, E.W. (1976) Cytochrome b
5
and
Cytochrome b
5
Reductase from a Chemical and X-ray Diffraction
Viewpoint. Wiley, New York.
11. Lederer, F. (1994) The cytochrome b
5
-fold: an adaptable module.
Biochimie 76, 674–692.
12. Dangi, B., Sarma, S., Yan, C., Banville, D.L. & Guiles, R.D.
(1998) The origin of differences in the physical properties of the
equilibrium forms of cytochrome b
5
revealed through high-
resolution NMR structures and backbone dynamic analyses.
Biochemistry 37, 8289–8302.
13. Yao, P., Wang, Y.H., Sun, Y.L., Huang, Z.X., Xie, Y. & Xiao,
G.T. (1997) Importance of a conserved phenylananine-35 of
cytochrome b
5
to the protein’s stability and redox potential.
Protein Eng. 10, 578–581.
14. Yao, P., Wang, Y H., Xie, Y. & Huang, Z X. (1998) Spectro-
electro-chemical studies of cytochrome b
5

Huang, Z.X. (2000) Crystal structure of recombinant trypsin-
solubilized fragment of cytochrome b
5
and the structural com-
parison with Val61His mutant. Proteins: Structure, Function Genet
40, 249–257.
22. Durley, R.C.E. & Mathews, F.S. (1996) Refinement and structural
analysis of bovine cytochrome b
5
at 1.5 A
˚
resolution. Acta Crys-
tallogr. D52, 65–76.
23. Xue, L.L. Wang, Y.H. Xie, Y. Yao, P. Wang, W.H. Qian, W.
Huang, Z.X. Wu, J. & Xia. Z.X. (1999) Effect of mutation at
valine 61 on the three-dimensional structure, stability, and redox
potential of cytochrome b
5
. Biochemistry 38, 11961–11972.
24. Otwinowski, Z. & Minor, W. (1997) Processing of X-ray diffrac-
tion data collected in oscillation mode. Methods Enzymol. 276,
307–326.
Ó FEBS 2002 Mutation at Phe35 of cytochrome b
5
(Eur. J. Biochem. 269) 4295
25. Brunger, A.T. (1992) X-PLOR: a System for X-Ray Crystallo-
graphy and NMR. Version 3.1. New Haven: Yale University Press.
26. Brunger, A.T. Adams, P.D. Clore, G.M. et al. (1998) Crystal-
lography and NMR System (CNS): a new software system for
macromolecular structure determination. Acta Crystllogr. D54,

. J. Biol. Chem. 247, 4641–4647.
36. Sugiyama, T., Miura, R., Yamano, T., Shiga, K. & Watari, H.
(1980) A reversible spin conversion of cytochrome b
5
at high
temperatures. Biochem. Biophys. Res. Commun. 97, 22–27.
37. Luzzati, P.V. (1952) Traitment statistique des erreurs dans la de-
ternination des structures cristallines. Acta Crystallogr. 5, 802–810.
38. Falzone,C.J.,Mayer,M.R.,Whiteman,E.L.,Moore,C.D.&
Lecomte, J.T. (1996) Design challenges for hemoproteins: the
solution structure of apocytochrome b
5
. Biochemistry 35, 6519–
6526).
39. Pfeil, W. (1993) Thermodynamics of apocytochrome b
5
unfolding.
Protein Sci. 2, 1497–1501.
40. Kawamura-Konishi, Y., Kihara, H. & Suzuki, H. (1988) Recon-
stitution of myoglobin from apoprotein and heme, monitored
by stopped-flow absorption, fluorescence and circular dichroism.
Eur. J. Biochem. 170, 589–595.
41. Hsu, M C. & Woody, R.W. (1971) The origin of the heme Cotton
effects in myoglobin and hemoglobin. J. Am. Chem. Soc. 93, 3515–
3525.
42. Shawver, L.K., Siedel, S.L., Krieter, P.A. & Shires, T.K. (1984) An
enzyme-linked immunoadsorbent assay for measuring cyto-
chrome b
5
and NADPH-cytochrome P-450 reductase in rat liver

48. Storch, E.M. & Daggett, V. (1996) Structural consequences of
heme removal: molecular dynamics simulations of rat and bovine
apocytochrome b
5
. Biochemistry 35, 11569–11604.
49. Wittung-Stafshede, P., Gray, H.B. & Winkler, J.R. (1997) Rapid
formation of four-helix bundle. Cytochrome b562 folding trig-
gered by electron transfer. J. Am. Chem. Soc. 119, 9562–9563.
50. Pasher, T., Chesick, J.P., Winkler, J.R. & Gray, H.B. (1996)
Protein folding triggered by electron transfer. Science 271, 1558–
1560.
51. Altuve, A., Silchenko, S., Lee, Kyung-Hoon, Kuczera, K.,
Terzyan, S., Zhang, X J., Benson, D.R. & Rivera, M. (2001)
Probing the differences between Rat Liver Mitochondrial
membrane Cytochrome b
5
and Microsomal Cytochrome b
5
.
Biochemistry 40, 9469–9483.
52. Newbold, R.J., Hewson, R. & Whitford, D. (1992) The thermal
stability of the tryptic fragment of bovine microsomal cytochrome
b
5
and a variant containing six additional residues. FEBS Lett.
314, 419–422.
53. Caffrey, M.S. & Cusanovich, M.A. (1994) Site-specific mutagen-
esis studies of cytochrome c. Biochim. Biophys. Acta 1187,277–
288.
54. Sandberg, W. & Terwilliger, T. (1989) Influence of interior pack-

11, 134–138.
4296 P. Yao et al. (Eur. J. Biochem. 269) Ó FEBS 2002