Protein Expression and Purification 30 (2003) 43–54

www.elsevier.com/locate/yprep

Heterologous expression of enzymopenic

methemoglobinemia variants using a novel NADH:cytochrome
c reductase fusion proteinq
C. Ainsley Davis and Michael J. Barber*
Department of Biochemistry and Molecular Biology, College of Medicine, University of South Florida, Tampa, FL 33612, USA

Received 15 November 2002, and in revised form 8 January 2003

Abstract

Hereditary enzymopenic methemoglobinemia is a rare disease that predominantly results from defects in either the erythrocytic
(type I) or microsomal (type II) forms of the enzyme NADH:cytochrome b5 reductase (EC 1.6.2.2). All 25 currently identified type I
and type II methemoglobinemia mutants have been expressed in Escherichia coli using a novel six histidine-tagged rat cytochrome
b5 /cytochrome b5 reductase fusion protein designated NADH:cytochrome c reductase (H6 NCR). All 25 H6 NCR variants were
isolated and demonstrated to result in two groups of expression products. The first group of 16 mutants, which included the majority
of the type I mutants, included K116Q, P131L, L139P, T183S, M193V, S194P, P211L, L215P, A245T, A245V, C270Y, E279K,
V305R, V319M, M340), and F365), and yielded full-length fusion proteins that retained variable levels of NADH:cytochrome c
reductase (NADH:CR) activity, ranging from approximately 2% (M340)) to 92% (K116Q) of that of the wild-type fusion protein.
In contrast, the remaining nine mutants that represented the majority of the type II variants, comprised a second group that in-
cluded Y109*, R124Q, Q143*, R150*, P162H, V172M, R226*, C270R, and R285*, and resulted in truncated H6 NCR variants that
retained the amino-terminal cytochrome b5 domain but were devoid of NADH:CR activity due to the absence of the cytochrome b5
reductase flavin domain. Kinetic analyses of the first group of full-length mutant fusion proteins indicated that values for both kcat
NADH
and Km were decreased and increased, respectively, indicating that the various mutations affected both substrate affinity and/or
turnover. However, for the second group, the truncated products were the result of incomplete production of the carboxyl-terminal
flavin-containing domain or instability of the expression products due to improper folding and/or lack of flavin incorporation.
Ó 2003 Elsevier Science (USA). All rights reserved.

Enzymopenic hereditary methemoglobinemia is a identified worldwide, while variants of the enzyme,

recessively inherited disorder due, primarily, to a de-
creased activity of the flavoprotein, cytochrome b5 re-
ductase (EC 1.6.2.2), which catalyzes the intermolecular
transfer of reducing equivalents from the physiological
electron donor NADH1 to two molecules of cytochrome
b5 [1]. Deficiencies of cytochrome b5 reductase have been
characterized by abnormal electrophoretic mobility with
retention of normal catalytic properties, may have an
incidence as high as 1:100.
Four clinical-biochemical classifications of enzym-
openic hereditary methemoglobinemia have been pro-
posed based upon important differences in the

q
This work was supported by Grants GM 32696 from the National Institutes of Health and 0150936B from the American Heart Association,
Florida/Puerto Rico Affiliate.
*
Corresponding author. Fax: 813-974-7357.
E-mail address: mbarber@hsc.usf.edu (M.J. Barber).
1
Abbreviations used: NCR, NADH:cytochrome c reductase; cb5r, cytochrome b5 reductase; cb5, cytochrome b5 ; NADH:FR, NADH:ferricy-
anide reductase; NADH:CR, NADH:cytochrome c reductase; GST, glutathione-S-transferase; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide
gel electrophoresis; FPLC, fast protein liquid chromatography; MOPS, 3-(N-morpholino)propanesulfonic acid; PCR, polymerase chain reaction;
IPTG, isopropyl b-D -thiogalactopyranoside; MALDI-TOF, matrix-assisted laser desorption ionization time-of-flight; m=z, mass to charge ratio;
EDTA, ethylenediaminetetraacetic acid; PMSF, phenylmethylsulfonyl fluoride; TB, terrific broth; *, stop codon; ), codon deletion.

1046-5928/03/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved.
doi:10.1016/S1046-5928(03)00046-9
44 C. Ainsley Davis, M.J. Barber / Protein Expression and Purification 30 (2003) 43–54

disorderÕs pathophysiology. The type I disease, exhib-
ited by the majority of patients, is due to a deficiency
in erythrocytes of the soluble form of NADH:cyto-
chrome b5 reductase. In the type II disease, which
represents a more severe form of the disorder that
occurs in only 10–15% of patients, the defect is due to
a generalized NADH:cytochrome b5 reductase defi-
ciency, which also includes the membrane-associated
form present in somatic cells. The type III disorder,
which occurs with a substantially lower frequency, has
been identified as due to a generalized hematopoietic
deficiency of NADH:cytochrome b5 reductase in ery-
throcytes, platelets, lymphocytes, and granulocytes.
Finally, the type IV disease, which may also result in
methemoglobinemia, has been ascribed to a deficiency
of cytochrome b5 [1].
In mammalian systems, NADH:cytochrome b5 re-
ductase has been identified as present in two forms, both
of which are transcribed from a single gene assigned to
chromosome 22q13-qter for the rat enzyme [2,3]. The
first corresponds to a large (35 kDa) membrane-associ-
ated enzyme that participates in fatty acid desaturation
and elongation [4], cholesterol biosynthesis [5], and the
metabolism of xenobiotics [6] while the second corre-
sponds to a truncated, soluble 30 kDa protein that
participates in the reduction of methemoglobin to he-
moglobin in erythrocytes [7,8].
A number of specific mutations, shown in Fig. 1, have
been identified for the human cb5r that result in a va-
riety of pathological conditions; however, little infor-
mation is available concerning the effects of these
mutations on the structure and function of the various

Fig. 1. Comparison of the primary sequences of H6 NCR and cytochrome b5 reductase. The primary sequences of rat H6 NCR (Rat NCR) and the
soluble, catalytic domain of human cytochrome b5 reductase (Human cb5r) were aligned using the ClustalW algorithm. The rat cytochrome b5 heme-
binding domain portion of the NCR fusion protein is shown in red while the flavin- and NADH-binding lobes of rat and human cytochrome b5
reductase, identified from the rat X-ray structure [22], are shown in black and blue, respectively. In addition, the ‘‘linker’’ region joining the two cb5r
lobes is shown in green while the six-histidine tag and the unique glycine residue that results from the cloning process in NCR [14] are shown in italics.
Also shown are the positions of the known human cb5r methemoglobinemia mutations (shown in boldface) and the corresponding amino acid
substitutions, deletions or stop codon insertions (Human mHb). A ‘‘/’’ indicate the positions where multiple mutations have been identified; ‘‘*’’
denote the positions of stop codon mutations; and ‘‘-’’ represent residue deletions. For the human cb5r sequence, residues are numbered with respect
to their position in the full-length, membrane-associated form of the enzyme.
 C. Ainsley Davis, M.J. Barber / Protein Expression and Purification 30 (2003) 43–54
cb5r isoforms. The identification of the TCT ! CCG
codon change at residue 127 (S ! P)2 provided the first
example of a molecular defect in cb5r, resulting in he-
reditary methemoglobinemia [9]. Recently, a number of
additional mutations have been identified that involved
either radical amino acid substitutions, such as V105M
[10], residue deletions, including F298) [11] or the in-
sertion of premature stop codons (R218*) [12], resulting
in termination of the cb5r polypeptide chain. Analyses
of some of these methemoglobinemia mutations have
indicated that the specific amino acid substitutions re-
sulted in proteins that were either catalytically and/or
structurally defective, the latter variants exhibiting de-
creased stability with associated decreased half-lives in
various cell types [13].
We have previously engineered a gene that codes for
the expression in Escherichia coli, of a fusion protein
comprising the microsomal variants of the soluble heme-
containing fragment of cytochrome b5 and the soluble
flavin-containing fragment of cytochrome b5 reductase
[14], respectively. The fusion protein, designated
NADH:cytochrome c reductase (NCR), has been dem-
onstrated to retain the properties of the individual heme
and FAD prosthetic groups and to efficiently catalyze
the transfer of reducing equivalents from NADH to
oxidized cytochrome c. This fusion protein provides a
novel system for examining factors influencing the in-
teraction between the flavin and heme prosthetic groups
of both cytochrome b5 reductase and cytochrome b5 .
As an initial step in examining the effects of the various
known naturally occurring enzymopenic methemoglobi-
nemia mutations on the properties of cytochrome b5 re-
ductase, we have used the NADH:cytochrome c reductase
(NCR) fusion protein as an expression system to distin-
guish between cb5r mutants with either compromised
catalytic activity and/or decreased structural stability.
Experimental procedures
Materials
NADH, ferric citrate, riboflavin, PMSF, Tris base,
and horse heart cytochrome c (Type VI) were purchased
from Sigma Chemical Co. (St. Louis, MO). MOPS was
purchased from Calbiochem (San Diego, CA). Native
2
For the H6 NCR variants, the individual methemoglobinemia
mutants are designated by the position of the mutation in the full-
length H6 NCR primary sequence. This sequence is composed of 268
residues (T101-F367) that comprise the amino acid sequence of the
hydrophilic, FAD-containing domain of cytochrome b5 reductase
together with a 100 residue amino-terminal extension that includes the
polyhistidine tag (M1-H7) and the 92 residues (M7-I99) that form the
cytochrome b5 heme domain. Thus, the methemoglobinemia mutant
S127P in cb5r corresponds to the mutant S194P in H6 NCR, as
indicated in Table 1.
45
Pfu and Pfu Turbo polymerases as well as the E. coli
BL21-CodonPlus (DE3)-RIL cells were obtained from
Stratagene (La Jolla, CA). The various restriction en-
zymes, including NdeI, BamHI, and those identified in
Table 1, were obtained from New England Biolabs
(Beverley, MA) and the pET23B expression vector was
obtained from Novagen (Madison, WI). T4 DNA ligase
and M-MLV Rnase (–H) reverse transcriptase were
purchased from Promega (Madison, WI). Triton X-100
and Hot Start Micro 50 PCR tubes were obtained from
Molecular-Bio Products (San Diego, CA). IPTG was
obtained from RPI (Mt. Prospect, IL) while tryptone
and yeast extract were obtained from EM Science
(Gibbstown, NJ). Ni–NTA agarose and kits for plasmid
preparation and agarose gel extraction were purchased
from Qiagen Inc. (Valencia, CA). Nucleotide sequencing
was performed by the Molecular Biology Core Facility
of the H. Lee Moffitt Cancer Center and Research In-
stitute at the University of South Florida.
pH6 NCR plasmid construction
The following oligonucleotide primers (Integrated
DNA Technologies, Coralville, IA) were designed to
engineer the insertion of an amino-terminal six-histidine
affinity tag into the existing NCR expression plasmid
[14]:
Primer 1 (amino-terminus)
(50 –30 ) ATG CAC CAT CAC CAT CAC CAC
ATG GCT GAA CAA AGC
Primer 2 (6H tag and NdeI restriction site)
(50 –30 ) GGA ATT CCA TAT GCA CCA TCA
CCA TCA CCA CAT G
Primer 3 (carboxy-terminus and BamHI restriction
site)
(50 –30 ) CGC GGA TCC CCG TCA GAA GGT
GAA GCA TCG
Primers 1 and 3 were used in the initial PCR and primers
2 and 3 were used in the final PCR, resulting in the
amplification of the full-length 325 bp NCR sequence
together with the insertion of the 6-histidine tag and the
required NdeI and BamHI restriction sites. Simultaneous
restriction endonuclease double digests utilizing both
NdeI and BamHI and the final H6 NCR PCR product
and the pET23b plasmid containing the original non-
histidine tagged NCR gene were performed in separate
reactions to yield the pET23b vector H6 NCR PCR
product. The two products were ligated at 22 °C using T4
DNA ligase to yield the H6 NCR-pET23b construct. This
construct was transformed into E. coli DH5a cells and
the resulting colonies were screened by agarose gel elec-
trophoresis. The fidelity of this construct was determined
by dideoxy nucleotide sequencing in both the forward
and reverse directions and the resulting plasmid was used
to transform E. coli BL21-CodonPlus (DE3)-RIL com-
46 C. Ainsley Davis, M.J. Barber / Protein Expression and Purification 30 (2003) 43–54

Table 1
Oligonucleotide primers used to generate the various hereditary methemoglobinemia mutants
Mutation Primer 50 ! 30 Restriction site
H6 NCR cb5r
Y109* Y42* GAACCCCGACATCAAGTAA ACCTCTGCGGCTCATC n/aa
K116Q K49Q CTGCGGCTCATCGACCCAGGAGATTATCAGCCAT n/a
R124Q R57Q ATCAGCCATGACACTCAAGCGATTC CGATTTGCACTCCCT +TfiI
P131L P64L TTCCGATTTGCACTCCTTC TCGCCCCAGCACATC +BseRI
L139P L72P CAGCACATCCTGGGCCCC TCCTATCGGCCAGCAC +ApaI
Q143* Q76* GGCCTTCCTATCGGCTT AGCACATCTACCTCTCC -EaeI
R150* R83* CATCTACCTCTCCACCTTGAATC GATGGCAACTTGG +TfiI
P162H P95H GGTCATTCGTCCCTACACCCACGTGT CTAGTGATG +AflIII
V172M V105M GATGACAAGGGCCTTAATGGACCTGGT GGTCAAGGTT +SexAI
T183S T116S GTTTACTTCAAGGAT
TTC GCATCCCAAGTTTCCAGC +TfiI
M193V M126V CCAGCTGGAGGGAAGGTCTC TCAGTACCTGGAA +BsaI
S194P S127P CCAGCTGGAGGGAAAATGC CCGCAATACCTG GAAAAC +AlwN I
P211L P144L ATTGAATTCCGGGGCCTT CAATGGGCTACTGGTC -BanII
L215P L148P GGCCCCAATGGGCTACCCGGT CTACCAGGGCAAA +AgeI
R226* R159* GGGAAGTTCGCCATCTTGAGCAGACAAGAAGTCC -BsgI
A245T A178T TCTGTAGGCATGATC
CACTGGAGGGACAGGCATC +BclI
A245V A178V TCTGTAGGCATGATTGTTGGGAGGGACAGGCATC n/a
C270R C203R CAGATATC TGCTCTTCGCC
AACGACCACACTGTC +EcoR V
C270Y C203Y CGAACGACCACACTGTCCTACTATCTGCTCTTCGCC -DraIII
E279K E212K CTTCGCCAACCAGTCCAAAAAAAGACATCCTGCTGC n/a
R285* R218* GAAAGACATCCTGCTGTTAGCCTGAGCTGGAGGAAC n/a
V305R V238R AAGCTCTGGTACACACCGTGACAAAGCCCCCGAT +AflIII
V319M V252M GCCAAGGA
ATTCATGAATGAGGAGATGATCAGGGAC +TfiI
M340- M272- CGGTGGGGGTCCACACAGTATCAGTGTCTCCTC n/a
F365- F298- CCCAAGCGATGCACCTTCTGACGGGGATCCG n/a
Base pairs indicating the restriction site are underlined and base pairs different from wild-type are in bold, except for M340) and F366) where the
entire codon has been deleted.
a
n/a denotes primers that did not introduce or delete a restriction site and where the mutations were verified by DNA sequencing.

petent cells for subsequent recombinant H6 NCR protein constructs were then used to transform competent E.
production. coli BL21-CodonPlus (DE3)-RIL cells. Selected methe-
moglobinemia mutants were also produced for the
Site-directed mutagenesis discrete cb5r flavin domain as previously described [15].

The various methemoglobinemia variants of H6 NCR
were generated using a modification of the ‘‘Quick-Change’’
protocol (Stratagene, La Jolla, CA), as previously de-
scribed [15]. The various oligonucleotide primers used to
generate the 25 H6 NCR mutants are listed in Table 1.
Vector PCR was performed using Pfu Turbo poly-
merase (1.25 U) in the presence of cloned Pfu buffer
containing 10 ng H6 NCR vector DNA as template,
125 ng of each synthetic primer, 5% DMSO, 0.1% Triton
X-100, and 200 lM dNTPs with cycling parameters (20
cycles) of 1 min at 94 °C, 1 min at 55 °C, and 10 min at
68 °C. DpnI digestion of the reaction mixture following
PCR was utilized to cleave only the methylated template
DNA and the PCR products were subsequently purified
by agarose gel electrophoresis and used for direct
transformation of competent E. coli DH5a cells. Colo-
nies were screened for the specific mutation by restric-
tion enzyme digestion and visualization of the products
by acrylamide gel electrophoresis. The fidelity of the
mutant constructs was verified by nucleotide sequencing
in both the forward and reverse directions. Positive
Cell culture and protein isolation
Escherichia coli cells harboring the expression con-
structs for either pH6 NCR or the various methemoglo-
binemia variants were grown at 37 °C to an OD340 of 0.6 in
TB media, supplemented with both riboflavin (100 lM)
and ferric citrate (100 lM), following which protein pro-
duction was induced by addition of IPTG to a concen-
tration of 0.4 mM and the cells were allowed to grow for a
further 6 h at 25 °C. The cells were harvested by centri-
fugation at 400g, resuspended in 50 mM Tris–Cl buffer,
containing 300 mM NaCl and 10 mM imidazole, pH 8.0
(4 ml/g cells), and disrupted by sonication in the presence
of PMSF (1.2 mg/ml). The lysate was clarified by centri-
fugation at 30,000g for 20 min at 4 °C following which
NTA–agarose was added to the supernatant and the
suspension was incubated with gentle agitation at 4 °C for
1 h. The NTA–agarose suspension was transferred to a
chromatography column (1  10 cm) and the matrix was
extensively washed with 25 mM phosphate buffer, con-
taining 300 mM NaCl and 10 mM imidazole, pH 8.0,
 C. Ainsley Davis, M.J. Barber / Protein Expression and Purification 30 (2003) 43–54
following which the bound protein was eluted with
25 mM phosphate buffer, containing 300 mM NaCl and
100 mM imidazole, pH 8.0. Protein-containing fractions
with a reddish coloration were pooled, pressure concen-
trated (Amicon YM5 membrane), and purified further
using FPLC gel filtration (Superdex 200, 1  30 cm).
Pooled fractions containing the various expression
products were concentrated, beaded, and stored under
liquid N2 until required. Recombinant rat H4 -cyto-
chrome b5 reductase flavin domain was isolated as pre-
viously described by Marohnic and Barber [15] while
recombinant soluble rat cytochrome b5 was isolated as
described by Beck-von Bodman et al. [16]. SDS–poly-
acrylamide gel electrophoresis was performed as de-
scribed by Laemmli [17].
Spectroscopic analyses
UV/visible spectra were obtained using a Hewlett–
Packard (Agilent Technologies, Palo Alto, CA) 8453
diode-array spectrophotometer. H6 NCR concentrations
were estimated using A413 ¼ 124 cm1 mM1 [14].
MALDI-TOF mass spectrometric analysis of H6 NCR
samples was performed as described by Barber and
Quinn [14].
Enzyme activities
All assays were performed in triplicate and were
routinely determined at 25 °C in 116 mM MOPS buffer
(l ¼ 0:05), containing 0.1 mM EDTA, pH 7.0.
NADH:FR activities were performed, as previously
described by Barber and Quinn [14]. NADH:CR activ-
ities were determined at 550 nm using 100 lM NADH
and 40 lM cytochrome c. Enzyme activities are ex-
pressed in terms of lmol of NADH consumed per min/
nmol of heme. Initial-rate data were analyzed using the
software ‘‘Enzfit’’ (Elsevier Biosoft, Ferguson, MO) to
yield individual values for the apparent kcat and Km .
Results
pH6 NCR expression vector construction and protein pro-
duction
Previous work on the isolation of the non-histidine
tagged form of NCR utilized affinity chromatography
on 50 -ADP Sepharose, a NADþ analog, as the major
purification step. However, due to the potentially dele-
terious effects of the various cytochrome b5 reductase-
associated methemoglobinemia mutations on either the
stabilities of the resultant proteins or their affinity for
the physiological reducing substrate, NADH, we ini-
tially engineered a modified NCR expression construct
that included the addition of an amino-terminal, six-
47
histidine affinity tag, that would facilitate the rapid
isolation of the various methemoglobinemia mutants
without relying on their affinity for a NADþ analog. In
addition, we opted to produce the various methemo-
globinemia mutants using the H6 NCR expression sys-
tem, since the presence of the amino-terminal cb5
heme-containing domain provided a convenient visual
and sensitive spectrophotometric assay that could be
used to confirm recombinant protein expression while
the presence of a codon-optimized cb5 nucleotide se-
quence had previously been demonstrated to result in
highly efficient cb5 production in E. coli [16].
Recombinant expression and isolation of the H6 NCR
protein resulted in a soluble hemoflavoprotein with a
molecular mass of approximately 42 kDa as indicated by
SDS–PAGE analysis whereas MALDI-TOF mass
spectrometry revealed an m=z ratio of 41,869, in excel-
lent agreement with the theoretical value of 41,906 cal-
culated from the 367 residue primary sequence. Mass
spectrometry also confirmed that the isolated H6 NCR
protein retained both heme and FAD prosthetic groups
and exhibited both NADH:FR and NADH:CR ac-
tivities. UV/visible absorption spectra of the oxidized
protein revealed a spectrum identical to that of the non-
his-tagged variant with a UV maximum at 272 nm to-
gether with a Soret band at 413 nm that dominated the
visible spectrum. Anaerobic reduction with NADH re-
sulted in a shift in the Soret peak to 423 nm and the
appearance of the characteristic a and b peaks at 527 and
556 nm, respectively, indicative of conversion of the heme
from the Fe3þ to the Fe2þ oxidation state. Initial-rate
kinetic studies of both the NADH:FR (kcat ¼ 8:3  0:1
102 s1 , KmNADH
¼ 4  1 lM, and Km FeCN6
¼ 12 2 lM)
and NADH:CR (kcat ¼ 3:0  0:1  102 s1 , Km NADH
¼
cyt:c
2  1 lM, and Km ¼ 1  1 lM) activities yielded kinetic
constants that were equivalent to those previously ob-
tained for the non-his-tagged variant [14]. In addition, the
kcat obtained for the NADH:FR activity of the fusion
protein was identical to that previously obtained for the
individual wild-type cb5r flavin domain [15]. These results
indicated that the addition of the amino-terminal histi-
dine tag had no adverse effects on either the spectro-
scopic or kinetic properties of the fusion protein and
confirmed that the H6 NCR expression system could
provide a suitable approach to examine expression of the
various cb5r-related methemoglobinemia mutants.
Production of the enzymopenic methemoglobinemia
H6 NCR variants
Escherichia coli BL21(DE3)-RIL cells harboring the
various H6 NCR expression constructs for the 25 known
enzymopenic methemoglobinemia mutations were
grown in TB medium to mid log phase and recombinant
protein expression was induced by the addition of IPTG
to the growth medium. To minimize inclusion body
48 C. Ainsley Davis, M.J. Barber / Protein Expression and Purification 30 (2003) 43–54

formation due to elevated rates of recombinant protein
expression and incorrect protein folding, the growth
temperature was reduced to 25 °C following induction.
Despite the lower temperature, sufficient fusion protein
products were produced for all the H6 NCR variants to
retain the dramatic pink coloration of the E. coli cell
pellets when compared to their non-induced counter-
parts. However, while visual analysis of all the cell su-
pernatants obtained following harvesting and cell
disruption indicated production of at least the H6 NCR
amino-terminal cb5 heme-containing domain, the iso-
lated crude cell extracts exhibited substantial variability
in their associated NADH:CR activities, indicative of
the production of either the full length, mature proteins
with compromised catalytic activities or truncated,
H6 NCR non-functional fragments.
All 25 H6 NCR methemoglobinemia mutants were
isolated using the same two-step purification protocol
that initially involved NTA–agarose affinity chroma-
tography, followed by FPLC gel filtration. Representa-
tive results obtained from 4 L of cell culture, for the
isolation of the wild-type H6 NCR protein, four selected
individual mutants, corresponding to Y109*, L139P,
P162H, and P211L, respectively, are given in Table 2
together with their associated FPLC gel filtration chro-
matograms in Fig. 2. For the wild-type H6 NCR, FPLC
gel filtration of the pooled fractions eluted from the
NTA–agarose column indicated efficient production of
the heme-containing, functional fusion protein. The
chromatogram indicated a single, dominant elution
peak, which exhibited significant absorption at both 280
and 413 nm, the latter typical of a heme-containing
protein while activity assays of the individual column
fractions indicated that the UV absorbing fractions also
exhibited both NADH:CR and NADH:FR activities,
suggesting the isolation of a catalytically active
NADH:cytochrome c reductase fusion protein. The gel
filtration profile also indicated the absence of any sig-
nificant contaminating proteins as well as confirming the
absence of any protein fragments, corresponding to ei-
ther the discrete heme- or flavin-containing domains
that may have arisen as a result of partial proteolysis of
the fusion protein. The 38% yield of H6 NCR corre-
sponded to approximately 60 mg protein with a
NADH:CR specific activity of 78 lmol NADH con-
sumed/min/mg protein.
As anticipated, in contrast to the wild-type protein,
insertion of a premature stop codon in the cb5r gene
corresponding to the Y109* mutation primarily resulted
in the production of a substantially truncated, low mo-
lecular weight H6 NCR protein fragment that was de-
void of both NADH:FR and NADH:CR activities due

Table 2
Purification of recombinant H6 NCR and representative methemoglobinemia mutants
Fraction Total protein Volume Activity Activity Specific Specific Yield (%)
(mg) (ml) NADH:CR NADH:FR activity activity
(U) (U) NADH:CR NADH:FR
(U/mg) (U/mg)
H6 NCR
Lysate 538 43.1 12,348 130,593 23 243 100
NTA–agarose 114 10.2 7586 113,730 67 998 61
FPLC 60 2.4 4689 90,023 78 1500 38
Y109* (Y42*)
Lysate 448 45.1 0 4624 0 10 100a
NTA–agarose 55 10.1 0 127 0 2 51
FPLC 34 1.8 0 0 0 0 24
L139P (L72P)
Lysate 514 44.2 6771 23,689 13 46 100
NTA–agarose 84 10.0 3698 16,482 44 196 55
FPLC 36 2.4 2112 8669 59 240 31
P162H (P95H)
Lysate 468 43.4 0 4927 0 11 100a
NTA–agarose 52 10.2 0 95 0 2 48
FPLC 30 1.8 0 0 0 0 20
P211L (P144L)
Lysate 572 42.2 9972 34,139 17 60 100
NTA–agarose 120 11.0 6602 26,752 55 243 66
FPLC 64 2.4 3590 18,742 56 292 36
U ¼ lmol NADH consumed/min.
a
Yields for Y109* and P162H were determined using their molar extinction coefficients of 124 mM1 cm1 at 413 nm for the heme prosthetic
group.
 C. Ainsley Davis, M.J. Barber / Protein Expression and Purification 30 (2003) 43–54 49

Fig. 2. FPLC size exclusion chromatography of H6 NCR and representative methemoglobinemia mutants. FPLC size exclusion chromatography was
performed using a Superdex 200 column (1  30 cm) in 10 mM phosphate buffer, containing 0.1 mM EDTA, pH 7.0. (A) H6 NCR; (B) P211L; (C)
L139P; (D) P162H; (E) Y109*; and (F) cytochrome b5 . Absorbance at 280 nm (d); absorbance at 413 nm (j); NADH:FR activity (.); and
NADH:CR activity (N). All activities are reported as lmol NADH consumed/min/ll.

to the absence of the functional carboxyl-terminal cb5r
flavin-containing domain. However, the N-terminal
heme-containing portion of the fusion protein was
produced, although with a significant degree of hetero-
geneity, as indicated by the extensive number of column
fractions retaining appreciable absorbance at 413 nm.
For the L139P mutant, the gel filtration profile in-
dicated production of two major protein components.
The initial, early eluting minor component retained
absorbance at both 280 and 413 nm and was also as-
sociated with retention of both NADH:FR and
NADH:CR activities, indicative of the production of
the full-length L139P mutant H6 NCR fusion protein. In
contrast, the second late eluting low molecular weight
component retained absorbance at 280 and 413 nm but
was devoid of either NADH:FR or NADH:CR activi-
ties, suggesting the protein comprised only the heme
domain of the fusion protein. The pooled fractions
containing the full-length mutant fusion protein corre-
sponded to approximately 36 mg protein with a
NADH:CR specific activity of 59 lmol NADH con-
sumed/min/mg protein.
Isolation of the H6 NCR variant resulting from the
expression of the P162H mutant construct indicated
only production of a late-eluting low molecular weight
protein that retained absorbance at 413 nm but which
50 C. Ainsley Davis, M.J. Barber / Protein Expression and Purification 30 (2003) 43–54

was devoid of both NADH:FR and NADH:CR activi-

ties, suggesting production of only the amino-terminal
heme-containing portion of the fusion protein.
Gel filtration of the P211L mutant revealed an
elution profile very similar to that obtained for the
wild-type H6 NCR, suggesting production of the full-
length fusion protein. Fractions comprising the major
protein peak exhibited near equal absorbance at 280
and 413 nm, respectively, and retained both
NADH:FR and NADH:CR activities, at levels com-
parable to those of the wild-type fusion protein. No
significant late-eluting protein peaks were detected,
contraindicating either production of the discrete
amino-terminal heme domain or any other flavin do-
main truncation products.
To examine the compositions and estimate the
molecular masses of the various protein products ob-
tained from the FPLC gel filtration studies of the
various H6 NCR methemoglobinemia mutant expres-
sion studies, selected pooled fractions were analyzed
using SDS–PAGE. As shown in Fig. 3, all 25 mutants
were analyzed in addition to representative samples
containing either the wild-type fusion protein
(H6 NCR) or the soluble recombinant forms of rat
cytochrome b5 reductase (flavin domain) and cyto-
chrome b5 (heme domain). As indicated, SDS–PAGE
analysis of the various mutant proteins revealed sig-
nificant protein heterogeneity between the different
mutants with sample compositions effectively divided Fig. 3. SDS–PAGE analysis of all known hereditary methemoglobi-
into two classes. The first consisted of those mutants nemia mutations. Samples (2 lg total protein) from the individual steps
involved in the isolation of the soluble rat H6 NCR and representative
that predominantly yielded the mutant variant of the hereditary methemoglobinemia mutants were analyzed using a 15%
full-length fusion protein (Mr approximately 42 kDa) polyacrylamide gel as described in ‘‘Experimental procedures.’’ The
comprising both the heme- and FAD-containing do- individual lanes correspond to: Mr , protein molecular weight markers
mains, respectively, while the second class consisted of with the indicated molecular masses; NCR, wild-type H6 NCR; cb5r,
the mutants that resulted in production of various wild-type H4 cb5r; cb5, wild-type cb5; 1, Y109* (Y42*); 2, K116Q
(K49Q); 3, R124Q (R57Q); 4, P131L (P64L); 5, L139P (L72P); 6,
truncated, low molecular mass proteins that primarily Q143* (Q76*); 7, R150* (R83*); 8, P162H (P95H); 9, V172M
contained only the discrete heme domain (Mr ap- (V105M); 10, T183S (T116S); 11, M193V (M126V); 12, S194P (S127P);
proximately 12 kDa). Examples of mutants comprising 13, P211L (P144L); 14, L215P (L148P); 15, R226* (R159*); 16, A245T
the first class included K116Q, L139P, and V305R, all (A178T); 17, A241V (A178V); 18, C270R (C203R); 19, C270Y
(C203Y); 20, E279K (E212K); 21, R285* (R218*); 22, V305R
of which primarily produced proteins with masses of
(V238R); 23, V319M (V252M); 24, M335) (M272)); and 25, F365)
approximately 42 kDa, equivalent to that of native (F298)).
H6 NCR and indicating that the various amino acid
substitutions did not significantly affect the stability of
the resulting heme-and flavin-containing fusion pro- An advantage of using the H6 NCR expression
tein. In contrast, examples of the second class in- system was our ability to compare, for the first time,
cluded all the mutations that introduced an additional, the expression products generated for all five mutants
premature stop codon together with the mutants that involved the insertion of premature stop codons
R124Q, P162H, and C270R, all of which produced in the cb5r flavin domain sequence. Comparison of
proteins that were either slightly larger or equivalent the SDS–PAGE results obtained for the Y109*,
to the mass of the soluble heme-containing domain of Q143*, R150*, R226*, and R285* variants (Fig. 3)
cytochrome b5 (12 kDa). For the latter mutations, the indicated that both the Y109* and Q143* protein
lack of production of the full-length fusion protein products were slightly larger than the cytochrome b5
indicated that these mutations significantly decreased portion of the fusion protein, suggesting that the small
the stability of the flavin domain, resulting in altered cb5r fragment was expressed, resulting in products
folding and subsequent proteolysis. that retained the heme domain and a small part of the
flavin domain. In contrast, the expression products
 C. Ainsley Davis, M.J. Barber / Protein Expression and Purification 30 (2003) 43–54 51

obtained for the R150*, R226*, and R285* mutations
resulted in proteins that were smaller than anticipated
and of comparable size to that of the cb5 control,
suggesting that the small cb5r flavin domain fragment
was not stable in these variants.
The results of the SDS–PAGE analysis of several of
the mutants, such as V172M and A245V, also indicated
the presence of small quantities of protein of equivalent
mass to that of the flavin domain of native cb5r, sug-
gesting that limited proteolysis of some of the fusion
proteins was occurring during the isolation procedure.

Spectroscopic studies

To examine the cofactor composition of selected

Fig. 4. UV/visible spectra of oxidized H6 NCR and representative
methemoglobinemia mutants. UV/visible absorption spectra were re-
corded for oxidized samples of H6 NCR (1.5 lM heme) in 10 mM
MOPS buffer, containing 0.1 mM EDTA, pH 7.0. Individual spectra
correspond to wild-type H6 NCR (—); Y109* [Y42*] (– –); L139P
[L72P] (- - -); P162H [P95H] (––); P211L [P144L] (––); and cb5
(     ). The insert shows an expanded region of the visible spectrum
where the flavin prosthetic group makes a major contribution.
H6 NCR methemoglobinemia mutants, UV/visible
spectra were obtained for the Y109*, L139P, P162H,
and P211L mutants together with both the wild-type
fusion protein and purified cb5, and are shown in Fig. 4.
All the spectra were typical of b5 -type cytochrome-
containing proteins and exhibited major peaks at 273
and 413 nm with additional broad peaks at 530 and
560 nm, respectively. In addition, the wild-type fusion
protein and the L139P and P211L mutant variants had
increased absorbance in the 450–490 nm range due to
the presence of the additional flavin chromophore. None
of the spectra obtained from the various full-length ex-
pression products indicated the absence of flavin incor-

Table 3
NADH:cytochrome c reductase kinetic constants obtained for the various methemoglobinemia mutations
Mutant NADH:CR Activity Methemoglobinemia
type
H6 NCR cb5r kcat KmNADH Kmcyt:c kcat =KmNADH
(s1 ) (lM) (lM) (s1 M1 )
Y109* Y42* NDa ND ND ND II
K116Q K49Q 200  6 21 21 1:00  0:05  108 I
R124Q R57Q ND ND ND ND I
P131L P64L 200  6 41 21 5:00  0:06  107 I
L139P L72P 117  4 31 11 3:90  0:04  107 I
Q143* Q76* ND ND ND ND II
R150* R83* ND ND ND ND II
P162H P95H ND ND ND ND I
V172M V105M ND ND ND ND I
T183S T116S 117  5 10  1 21 1:17  0:05  107 I
M193V M126V 150  6 21 21 7:50  0:06  107 II
S194P S127P 17  2 10  1 21 1:70  0:02  106 II
P211L P144L 150  5 21 21 7:50  0:05  107 I
L215P L148P 117  4 21 21 5:85  0:04  107 I
R226* R159* ND ND ND ND II
A245T A178T 150  5 91 11 1:66  0:05  107 I
A245V A178V 150  5 21 11 7:50  0:05  107 I
C270R C203R ND ND ND ND II
C270Y C203Y 167  5 21 21 8:35  0:05  107 I
E279K E212K 133  4 21 41 6:65  0:04  107 I
R285* R218* ND ND ND ND II
V305R V238R 100  4 80  2 21 1:25  0:04  106 I
V319M V252M 200  5 61 21 3:33  0:05  107 I
M340) M272) 31 53  2 21 5:66  0:01  104 II
F365) F298) 17  2 61 11 2:83  0:02  106 II
Wild-type 217  5 21 11 1:09  0:05  108
a
ND indicates that NADH:CR activity was not detectable for these mutants since expression did not yield a full-length variant of the NCR fusion
protein.
52 C. Ainsley Davis, M.J. Barber / Protein Expression and Purification 30 (2003) 43–54
poration, as judged by the absence of the broad spectral
peaks in the 450–490 nm wavelength range, indicating
that apoflavoproteins were not generated to any ap-
preciable extent.
Kinetic properties of the various H6 NCR methemoglobi-
nemia variants
Initial-rate studies indicated that the majority of the
purified H6 NCR methemoglobinemia variants retained
some level of NADH:CR activity. Under controlled
conditions of pH (7.0) and ionic strength (l ¼ 0:05),
Lineweaver–Burk plots of NADH:CR activities, where
both reducing (NADH) and oxidizing (cytochrome c)
substrates were varied individually, were linear over the
entire ranges of substrate concentrations examined for
the individual mutants that retained NADH:CR activ-
ity. Kinetic constants derived from these initial-rate
NADH:CR activity studies are given for the respective
mutants in Table 3. While none of the mutants exhibited
NADH:CR activities at levels greater than or equivalent
to the wild-type fusion protein, several of them retained
relatively high levels of activity, including K116Q,
P131L, and V319M, indicating that these mutations
resulted in only modest reductions in the catalytic effi-
ciencies of these individual variants. In contrast, several
mutants, including S194P, M340), and F365) showed
drastically reduced activities, which were often accom-
panied by elevated Km s for NADH, indicating that in
these mutants, both substrate affinity and utilization had
been compromised. Comparisons of the specificity
NADH
constants (kcat =Km ) obtained for the individual
mutants indicated that the M340) variant corresponded
to the most compromised form of the protein, retaining
only 0.05% of that of the wild-type H6 NCR protein.
A comparison of the affinities of the various mutants
for the acceptor, cytochrome c, indicated that only one
of the mutants, corresponding to E279K, exhibited an
elevated Km (4 lM) for this oxidizing substrate, sug-
gesting that this residue may be located at the interface
between the heme and flavin domains and may impact
cytochrome c binding.
Discussion
This work represents the first successful construction
and expression of all the currently known cytochrome b5
reductase mutations that have been identified as result-
ing in either the type I or type II forms of enzymopenic
methemoglobinemia in humans [1], using a single type of
expression system.
While hereditary enzymopenic methemoglobinemia
(MIM 250800) has been identified as a relatively rare,
recessive disease [1], a variety of specific alterations at
the genetic level have been recognized that result in
mutant forms of cb5r. Mutations in the cb5r (DIA1)
gene sequence, which represent the most frequently
documented forms of enzymopenic methemoglobine-
mia, have been described for both intron and exon re-
gions, with the latter resulting in a variety of cb5r
sequences containing either specific amino acid substi-
tutions or deletions. Several mutations have also in-
volved the inclusion of premature stop codons in the
cb5r sequence, resulting in substantially truncated pro-
teins. The mutations examined in this work have been
classified as causing either the type I or type II forms of
the disease in humans, which are associated with
markedly different physiological changes. For the type I
disease, which represents 15 of the 25 mutations and is
restricted to the soluble form of cb5r, the only symptom
is a well-tolerated cyanosis due to increased methemo-
globin concentrations. In contrast, in the more severe
and lethal type II disease, which corresponds to defects
in both the soluble and membrane-bound forms of cb5r,
the cyanosis is accompanied by progressive neurological
abnormalities that include mental retardation, which
may result from impaired fatty acid desaturation.
Previous studies of expressed methemoglobinemia
mutants have been limited to the R57Q [10], P95H [18],
V105M [10], S127P [9], L148P [19], A178V [20], and
F298) [11] variants. These mutants have been produced
as fusion proteins with either the eight residue amino-
terminal portion of b-galactosidase (R57Q, V105M,
S127P, L148P, and F298)) [21] or in combination with
GST (P95H and A178V). While these studies have pri-
marily compared the spectroscopic, catalytic, and
structural stabilities of these mutants with the corre-
sponding properties of the wild type cb5r, none have
attempted to examine the properties of any of the
truncated protein variants or to compare the kinetic
properties of all the variants produced using a single
expression system. None have been crystallized, which
could yield a detailed structural model of a methemo-
globinemia variant. In addition, neither X-ray nor
NMR structures are currently available for the native
human form of the enzyme, which severely inhibits de-
tailed analyses of the resultant structural changes caused
by the various amino acid substitutions. In contrast, we
had recently solved the high-resolution structures for the
recombinant form of the rat cb5r catalytic domain in the
absence and presence of NADþ [22], providing a model
system with which to interpret the potential structural
changes resulting from the various mutations.
We chose to initially generate the various methemo-
globinemia mutants using the rat H6 NCR hemoflavo-
protein expression system, since this would provide a
number of advantages over the conventional cb5r ex-
pression system. These included the presence of the cb5
heme domain as a convenient visible indicator of re-
combinant protein expression, the potentially enhanced
protein yields due to the efficient expression of the co-
 C. Ainsley Davis, M.J. Barber / Protein Expression and Purification 30 (2003) 43–54
don-optimized gene for the amino-terminal heme do-
main [16], and the NCR-associated NADH:CR activity
[14], which is not endogenous to E. coli in the absence of
the fusion protein expression. Thus, we anticipated this
expression system would provide a facile approach for
identifying the various cb5r mutations that would gen-
erate stable protein products for the flavin domain in
either its full-length or truncated form, while the pres-
ence of functional expression products could be readily
verified by their retention of NADH:CR activity, which
requires the presence of both the functional cb5r flavin-
containing domain and the cb5 heme-containing domain
[14].
To enhance production of the NCR fusion protein
and significantly ease the isolation of both the wild-type
and the numerous methemoglobinemia variants, we
initially developed an alternate E. coli expression system
for NCR to that previously published for the protein
[14]. The alternate expression system involved the ad-
dition of an amino-terminal six-histidine affinity tag for
utilization in metal chelate chromatography coupled
with the use of BL21-CodonPlus (DE3)-RIL cells. These
cells, which carry extra copies of the argU, ileY, and
leuW tRNA genes, have been designed to enhance
bacterial expression of eukaryotic proteins involving
rare codons. This expression system resulted in pro-
duction of the wild-type H6 NCR in higher yields, typi-
cally at levels 25–40% greater than those obtained with
the previous expression system and which could be pu-
rified to homogeneity using a simple two-step isolation
procedure. The resulting H6 NCR fusion protein re-
tained the same spectroscopic and kinetic properties as
those of the previous developed non-tagged variant [14].
The results of the SDS–PAGE analysis combined
with the extensive ‘‘pink’’ coloration of the various cell
supernatants confirmed the fact that all 25 mutant ex-
pression vectors resulted in at least the minimal pro-
duction of the amino-terminal cb5 domain together with
correct heme insertion in E. coli. Production of this
domain was used to provide both confirmation of ade-
quate levels of protein expression and verify the gener-
ation of the small protein fragments due to the
incorporation of premature stop codons in the sequence
for mutations such as Y109*.
The results of our expression studies suggest that the
products of the various methemoglobinemia type I and
type II mutations could be conveniently divided into
two main classes. The first comprised the mutations that
caused only moderate changes in the folding of the cb5r
domain, resulting in the production of the full-length
fusion protein together with flavin incorporation in the
carboxyl-terminal cb5r domain and retention of cata-
lytic activity. Sixteen of the 25 mutations, corresponding
to K116Q, P131L, L139P, T183S, M193V, S194P,
P211L, L215P, A245T, A245V, C270Y, E279K, V305R,
V319M, M340), and F365) yielded full-length fusion
53
proteins that retained variable levels of NADH:CR ac-
tivity, ranging from 2 to 92% of that of the wild-type
H6 NCR fusion protein. The majority of these mutations
(75%) have been previously classified as type I mutations
in human cb5r, resulting in only the mild form of met-
hemoglobinemia.
In contrast, the second class of mutations, which in-
cluded the remaining nine variants and correspond to
the majority of the amino acid substitutions that have
been previously identified as type II mutations and
which result in the more severe form of methemoglobi-
nemia, failed to produce the full-length H6 NCR fusion
protein together with the associated absence of any de-
tectable NADH:CR activity. While most of these mu-
tations corresponded to the insertion of premature stop
codons that would be expected to effectively terminate
the correct assembly of the flavin domain, four variants,
corresponding to R124Q, P162H, V172M, and C203R,
also failed to yield full-length fusion proteins, suggesting
that the amino acid substitutions radically perturbed
either correct domain folding or the incorporation of the
flavin prosthetic group, both of which significantly de-
creased domain stability resulting in premature prote-
olysis during expression.
The results obtained for the expression of selected
methemoglobinemia mutations in the flavin-containing
domain of rat H6 NCR were directly comparable to the
same mutations that have been previously generated and
characterized using the human cb5r expression systems,
suggesting that the rat cb5r domain accurately reflects
the properties of the human enzyme. For example, the
NADH
kcat =Km ratio obtained for the recombinant form of
the human cb5r S127P mutant has been shown to be
approximately 2.8% of that of the wild-type enzyme [9],
which may be compared to a value of 1.6% obtained for
the corresponding H6 NCR S194P variant. Similarly, a
comparison of the value obtained for the specificity
NADH
constant (kcat =Km ) of the human L148P mutant [19]
of 60% of the wild-type value was directly comparable to
the value of 54% determined for the H6 NCR L215P
variant.
Previous studies of the flavin-containing domain of
rat cb5r have indicated that some amino acid mutations
do not result in the production of the appropriate
flavoprotein [15]. Site-directed mutagenesis studies of
R91, a residue involved in FAD binding, indicated that
while a wide variety of mutants could be generated and
purified, including R91L and R91D, the R91H variant
did not yield a stable protein product. These differences
in mutant expression have suggested that cb5r variants
that fail to incorporate the flavin prosthetic group are
unstable and rapidly degraded during production.
The results of these studies have aided the prediction
of which of the known human cb5r mutations should
yield stable flavoproteins for further analysis and have
confirmed that rat cb5r provides a suitable model system
54 C. Ainsley Davis, M.J. Barber / Protein Expression and Purification 30 (2003) 43–54
with which to compare the properties of the various
human methemoglobinemia variants. Thus, combining
these results with the information gained from the high- [11]
resolution X-ray structure of the rat cb5r diaphorase
domain [22] should enhance our understanding of
structural factors that regulate cb5r turnover and sta-
bility. Using the insights gained from the expression
[12]
studies of the various H6 NCR methemoglobinemia
variants, we have successfully produced the discrete cb5r
S127P mutant using a modified his-tagged expression
system [15] and obtained crystals that diffract to 2 A  [13]
resolution. The solution of this structure will provide the
first insight into the structural perturbations that pro-
duce this type II variant of methemoglobinemia.
[14]
References [15]
[1] E.R. Jaffe, D.E. Hultquist, in: C.R. Scriver, A.L. Beaudet, W.S.
Sly, D. Valle (Eds.), Metabolic and Molecular Basis of Inherited [16]
Disease, vol. III, eighth ed., 2001, pp. 4555–4570.
[2] G. Pietrini, P. Carrera, N. Borgese, Two transcripts encode rat
cytochrome b5 reductase, Proc. Natl. Acad. Sci. USA 85 (1988)
7246–7250. [17]
[3] P.C. Bull, E.A. Shephard, S. Povey, I. Santisteban, I.R. Phillips,
Cloning and chromosomal mapping of human cytochrome b5
reductase (DIA1), Ann. Hum. Genet. 52 (1988) 263–268. [18]
[4] S.R. Keyes, D.L. Cinti, Biochemical properties of cytochrome b5 -
dependent microsomal fatty acid elongation and identification of
products, J. Biol. Chem. 255 (1980) 11357–11364.
[5] V.V.R. Reddy, D. Kupfer, E. Capsi, Mechanism of C-5 double
bond introduction in the biosynthesis of cholesterol by rat liver [19]
microsomes, J. Biol. Chem. 252 (1977) 2797–2801.
[6] A. Hildebrandt, R.W. Estabrook, Evidence for the participation
of cytochrome b5 in hepatic microsomal mixed function oxidation
reactions, Arch. Biochem. Biophys. 143 (1971) 66–79. [20]
[7] D.E. Hultquist, P.G. Passon, Catalysis of methemoglobin reduc-
tion by erythrocyte cytochrome b5 and cytochrome b5 reductase,
Nature 299 (1971) 252–254.
[8] F. Kuma, Properties of methemoglobin reductase and kinetic study
of methemoglobin reduction, J. Biol. Chem. 256 (1981) 5518–5523. [21]
[9] Y.T. Yubisui, K. Shirabe, M. Takeshita, Y. Kobayashi, Y.
Fukumaki, Y. Sakaki, T. Takano, Structural role of serine 127 in
the NADH-binding site of human NADH-cytochrome b5 reduc-
tase, J. Biol. Chem. 266 (1991) 66–70. [22]
[10] K. Shirabe, T. Yubisui, N. Borgese, C.Y. Tang, D.E. Hultquist,
M. Takeshita, Enzymatic instability of NADH-cytochrome b5
reductase as a cause of hereditary methemoglobinemia type I (red
cell type), J. Biol. Chem. 267 (1992) 20416–20421.
K. Shirabe, Y. Fujimoto, T. Yubisui, M. Takeshita, An in-frame
deletion of codon298 of the NADH-cytochrome b5 reductase gene
results in hereditary methemoglobinemia type II (generalized
type). A functional implication for the role of the COOH-
terminal region of the enzyme, J. Biol. Chem. 269 (1994) 5952–
5957.
L.M. Vieira, J.-C. Kaplan, A. Kahn, A. Leroux, Four new
mutations in the NADH-cytochrome b5 reductase gene from
patients with recessive congenital methemoglobinemia type II,
Blood 85 (1995) 2254–2262.
J. Dekker, M.H.M. Eppink, R. van Zwieten, T. de Rijk, A.F.
Remacha, L.K. Law, A.M. Li, K.L. Cheung, W.J.H. van Berkel,
D. Roos, Seven new mutations in the nicotinamide adenine
dinucleotide reductase-cytochrome b5 reductase gene leading to
methemoglobinemia type I, Blood 97 (2001) 1106–1114.
M.J. Barber, G.B. Quinn, Production of a recombinant hybrid
hemoflavoprotein: engineering a functional NADH:cytochrome c
reductase, Protein Expr. Purif. 23 (2001) 348–358.
C.C. Marohnic, M.J. Barber, Arginine 91 is not essential for flavin
incorporation in hepatic cytochrome b5 reductase, Arch. Biochem.
Biophys. 389 (2001) 223–233.
S.B. Beck-von Bodman, M.A. Schuler, D.R. Jollie, S.G. Sligar,
Synthesis, bacterial expression and mutagenesis of the gene coding
for mammalian cytochrome b5 , Proc. Natl. Acad. Sci. USA 83
(1986) 9443–9447.
U.K. Laemmli, Cleavage of structural proteins during the
assembly of the head of bacteriophage T4, Nature 227 (1970)
680–685.
J. Manabe, R. Arya, H. Sumimoto, T. Yubisui, A.J. Bellingham,
D.M. Layton, Y. Fukumaki, Two novel mutations in the reduced
nicotinamide adenine dinucleotide (NADH)-cytochrome b5 re-
ductase gene of a patient with generalized type, hereditary
methemoglobinemia, Blood 88 (1996) 3208–3215.
T. Nagai, K. Shirabe, T. Yubisui, M. Takeshita, Analysis of
mutant NADH-cytochrome b5 reductase: apparent ‘‘type III’’
methemoglobinemia can be explained as type I with an unstable
reductase, Blood 81 (1993) 808–814.
K. Higasa, J-I. Manabe, T. Yubisui, H. Sumimoto, P. Pung-
Amritt, V.S. Tanphaichitr, Y. Fukumaki, Molecular basis of
hereditary methemoglobinemia, types I and II: two novel muta-
tions in the NADH-cytochrome b5 reductase gene, Br. J.
Haematol. 103 (1998) 922–930.
K. Shirabe, T. Yubisui, M. Takeshita, Expression of human
erythrocyte NADH-cytochrome b5 reductase as an a-thrombin-
cleavable fused protein in Escherichia coli, Biochem. Biophys.
Acta 1008 (1989) 189–192.
M.C. Bewley, C.C. Marohnic, M.J. Barber, The structure and
biochemistry of cytochrome b5 reductase are now consistent,
Biochemistry 46 (2001) 13574–13582.