Mycobacterium tuberculosis
FprA, a novel bacterial
NADPH-ferredoxin reductase
Federico Fischer, Debora Raimondi, Alessandro Aliverti and Giuliana Zanetti
Dipartimento di Fisiologia e Biochimica Generali, Universita
`
degli Studi di Milano, Milano, Italy
The gene fprA of Mycobacterium tuberculosis, encoding a
putative protein with 40% identity to mammalian adreno-
doxin reductase, was expressed in Escherichia coli and the
protein purified to homogeneity. The 50-kDa protein
monomer contained one tightly bound FAD, whose fluor-
escence was fully quenched. FprA showed a low ferric
reductase activity, whereas it was very active as a NAD(P)H
diaphorase with dyes. Kinetic parameters were determined
and the specificity constant (k
cat
/K
m
)forNADPHwastwo
orders of magnitude larger than that of NADH. Enzyme full
reduction, under anaerobiosis, could be achieved with a
stoichiometric amount of either dithionite or NADH, but
not with even large excess of NADPH. In enzyme titration
with substoichiometric amounts of NADPH, only charge
transfer species (FAD-NADPH and FADH
2
-NADP
+
)
were formed. At NADPH/FAD ratios higher than one, the

availability and on iron-containing cofactors for growth
and survival [3]. It is well-known that iron availability in the
host plays a very important role in promoting the infection
by mycobacteria. Interestingly, it has been reported that
Nramp1 (natural resistance-associated macrophage protein)
protein of mouse macrophages confers resistance to myco-
bacterial infection in mice [4]. Recently, a hyphothesis has
been proposed based on the homology of Nramp1 to
DCT1, a metal-ion transporter [5]. Thus, the action of
Nramp1 in the phagosomal membrane may be to deplete
Fe
2+
or other divalent cations from the phagosome, thus
hampering the pathogen growth. Among possible strategies
to effectively interfere with the pathogen metabolism, the
blockage or limitation of Fe
2+
availability inside the
mycobacterium seems a promising target to pursue. Redox
systems called ferric reductases use intracellular redox
cofactors to reduce the ferric Fe to the ferrous form for
biosynthesis of iron-proteins. A NAD(P)H:ferrimycobactin
oxidoreductase activity was measured in M. smegmatis cell
extract [6]. In Escherichia coli, enzymes of the ferredoxin-
NADP
+
reductase (FNR) protein family showing iron
reductase activity, such as the flavin reductase, sulfite
reductase and flavohemoglobin, have been implicated in
such metabolism [7]. Searches of the M. tuberculosis

(Received 1 February 2002, revised 11 April 2002,
accepted 2 May 2002)
Eur. J. Biochem. 269, 3005–3013 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02989.x
and to possess some properties similar to those of the bovine
adrenodoxin reductase [9].
MATERIALS AND METHODS
Materials
All chemicals and pyridine nucleotides were purchased from
Sigma–Aldrich Chemical Co. Cytochrome c (Sigma C2506)
was further purified by ion-exchange chromatography on
SP-Sepharose (Pharmacia Biotech.). Restriction endonuc-
leases, DNA polymerase and DNA modifying enzymes
were supplied by Amersham Pharmacia Biotech. M. tuber-
culosis cosmid MTCY164 was kindly provided by S. T.
Cole, Institut Pasteur, France. pGEM-T and pET11a were
from Promega and Novagen, respectively. Bovine Adx
1
was
a generous gift from F. Bonomi, University of Milano,
Italy. Recombinant spinach ferredoxin I (Fd I) was purified
as described previously [10]. M. smegmatis ferredoxin has
been purified by a modification of the procedure described
by Imai et al. [11]. DEAE-cellulose and Sepharose 4B steps
were replaced by chromatoraphy on HiLoad Q-Sepharose
High-Performance and HiLoad phenyl-Sepharose High-
Performance columns (Pharmacia Biotech). Ferredoxin was
eluted at about 0.7
M
NaCl from the first column using a
0–1

grown in flasks under vigorous shaking at various temper-
atures in 2 · YT medium supplemented with 100 mgÆL
)1
ampicillin. For enzyme purification, E.colicells were grown
in a New Brunswick 12 L fermentor at 25 °C to midlog
phase (D
600
¼ 1.2–1.5). The culture, after cooling to 15 °C,
was induced with 0.1 m
M
isopropyl thio-b-
D
-galactoside.
Cells were harvested after 15–17 h.
Purification of FprA
All purification steps were performed at 4 °Cexceptfor
FPLC, which was carried out at room temperature. E.coli
cell paste were resuspended in 2 mLÆg
)1
of buffer A
(50 m
M
Na-phosphate, pH 7.0, containing 1 m
M
EDTA
and 1 m
M
2-mercaptoethanol) supplemented with 1 m
M
phenylmethanesulfonyl fluoride and disrupted by sonica-

50 m
M
Hepes/KOH, pH 7.0, containing 10% glycerol and
1m
M
DTT. The enzyme stored at )80 °Cretaineditsfull
activity for more than 1 year.
Molecular characterization methods
SDS/PAGE was carried out on 10% polyacrylamide gels.
Microsequencing was performed on an Applied Biosystems
477/A protein sequencer equipped with an on-line HPLC
system. Analytical gel-filtration analyses were performed on
a HPLC apparatus (Waters) equipped with either Superdex
75 or Superose 12 columns (Pharmacia Biotech) in 50 m
M
Hepes/KOH, pH 7.0, containing 0.15
M
ammonium acetate
and 2 m
M
2-mercaptoethanol. FprA and ferredoxin (10 and
40 l
M
, respectively) were cross-linked by treatment with
5m
M
N-ethyl-3-(3-dimethylaminopropyl)carbodiimide in
25 m
M
Na-phosphate, pH 7.0 [12].

spinach Fd I, bovine Adx or M. smegmatis ferredoxin,
3006 F. Fischer et al. (Eur. J. Biochem. 269) Ó FEBS 2002
using 50 l
M
cytochrome c as the terminal electron acceptor.
Unless otherwise stated, the NADPH concentration
was kept constant by regeneration with 2.5 m
M
glucose
6-phosphate and 2 lgÆmL
)1
glucose 6-phosphate dehydro-
genase. Steady-state kinetic parameters for the diaphorase
activities and for the cytochrome c reductase activity with
mycobacterial ferredoxin were determined by varying the
concentrations of the substrates. Double-reciprocal plots of
the data yielded parallel lines. Initial rate data (v)werefitted
by nonlinear regression using
GRAFIT
4.0 (Erythacus
Software Ltd, Staines, UK) to a ping-pong Bi-Bi mechan-
ism equation (Eqn 1):
v ¼ V Â A Â B=ðK
a
 B þ K
b
 A þ A  BÞð1Þ
where A and B,andK
a
and K

GRAFIT
4.0 (Erythacus Software
Ltd, Staines, UK).
DA
¼ De Â
L þ P þ K
d
À
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
L þ P þ K
d
ðÞ
2
À 4 Â L Â P
q
2
ð2Þ
DA is the value of the difference spectrum at a selected
wavelength; De is the difference extinction coefficient at that
wavelength of the protein-ligand complex; L is the total
molar concentration of added ligand; P is the total molar
concentration of FprA.
All reduction experiments were carried out in anaerobic
cuvettes at 15 °C. Solutions were made anaerobic by
successive cycles of equilibration with O
2
-free nitrogen
and evacuation. Reductive titrations with Na-dithionite,
NADPH, or NADH were carried out using 15–50 l
M

can be estimated
taking into account that SQ does not absorb at 750 nm [18]
and that a A
625
/A
750
value of 2.79 for CT species could be
determined from experiments in which no SQ was formed.
RESULTS
Identification of fprA and fprB
The search of M. tuberculosis genome [1] for enzymes
potentially involved in iron metabolism led to the identifi-
cation of two genes, fprA and fprB, whose predicted protein
products are related to each other. They share a domain
with significant similarity (% 40% identity) with mamma-
lian AdR (Table 1). FprB contains a C-terminal domain
homologous to FprA (42% identity) plus an N-terminal
moiety comprising an iron-sulfur binding region signature
typical of bacterial 7Fe ferredoxins. It is remarkable that
AdR homologs are present in very few bacteria, whereas
two such proteins are found in mycobacteria (Table 1). To
our knowledge, the fusion protein does not have a counter-
part in other organisms, except for other mycobacteria.
Production of FprA
We tried to heterologously express both cloned genes, yet we
were only successful in obtaining FprA in a soluble active
form. In a preliminary series of experiments, E.coli
BL21(DE3) strain harboring pETfprA was grown at
37 °C. Upon induction, a novel protein band of 50-kDa
was clearly visible in SDS/PAGE, but most of the protein

H 100
Ó FEBS 2002 M. tuberculosis NADPH-ferredoxin reductase (Eur. J. Biochem. 269) 3007
Q-Sepharose, yielded about 2 mg of FprA per gram of cells,
with an overall yield of 25% and a purification factor of 18.
SDS/PAGE of the various fractions of the purification is
showninFig.1.
FprA is a flavoprotein
The visible absorption spectrum of the purified protein is
presented in Fig. 2. The absorbance in the visible region is
that typical of a flavoprotein with bands centered at 381 and
452 nm and shoulders at 422 and 473 nm. Maximal
absorbance in the ultraviolet region was at 272 nm. A value
of 7.0 for the A
272
/A
452
ratio was calculated from the
spectrum. Flavin fluorescence was almost completely
quenched. The non covalently bound flavin in FprA was
shown to be FAD. The flavin fluorescence of the released
cofactor increased about 10-fold after phosphodiesterase
treatment, as expected for the conversion from FAD to
FMN. The extinction coefficient of the enzyme at 452 nm
was calculated to be 10 600
M
)1
Æcm
)1
from the amount of
FAD released after protein denaturation by SDS. A

to those deduced from the gene sequence: MRPYYIAIVG
SGPSAFFAAAS. The M
r
of the recombinant FprA in
solution was determined in several conditions. Gel filtration
experiments in FPLC, either on Superose 12 or Superdex
75, allowed the determination of a value of 53 ± 5 kDa,
when the protein was maintained in 10% glycerol and 1 m
M
dithiothreitol, indicating that under these conditions the
protein is a monomer. The addition of glycerol and
2-mercaptoethanol were required to avoid formation of
aggregates.
Catalytic properties
The ferric reductase activity of the purified protein was
investigated by using Fe
3+
-EDTA in the presence of the
Fe
2+
-chelator ferrozine [14]. The activity was very low both
in the presence and absence of oxygen and/or FAD:
0.5–1 (mol NADPH)Æmin
)1
Æ(mol FAD)
)1
.Furthermore,
addition of 1 l
M
7Fe ferredoxin from M. smegmatis (see

varying the reduced pyridine nucleotide at various fixed
levels of the artificial dye showed a pattern of parallel lines.
Data were fitted to Eqn (1). For the K
3
Fe(CN)
6
reductase
activity, the experiments revealed that ferricyanide concen-
trations above 1 m
M
were inhibitory. A 100- to 150-fold
lower K
m
values for NADPH with respect to NADH were
observed in the diaphorase reactions, whereas similar values
of k
cat
were obtained with both coenzymes, thus the
specificity constant ratio NADPH/NADH was 225 in the
ferricyanide reaction and 116 in the DPIP one. The catalytic
efficiencies of FprA with respect to the acceptors differed by
10-fold with preference for the one-electron reducible
substrate, i.e. ferricyanide. To study the interaction with
pyridine nucleotides in details, FprA was titrated with both
NADP
+
and NAD
+
. In both cases, the visible spectrum of
the enzyme was perturbed. The difference spectra elicited by

d
value for NAD
+
was in the millimolar range.
Identification of a physiological electron acceptor
The physiological activity of the mammalian homolog of
FprA is to reduce the [2Fe)2S] iron–sulfur protein Adx
[9,19,20]. Nevertheless, there are no genes coding for
[2Fe)2S] ferredoxins in the M. tuberculosis genome [1]. At
first, we studied the interaction of the recombinant enzyme
with the bovine Adx and with another [2Fe)2S]protein,the
spinach leaf Fd I. Cytochrome c was used as final electron
acceptor in these reactions. Its reduction was observed only
when either Adx or Fd I was added in the assay, indicating
that FprA was able to interact productively with both these
electron carrier proteins. FprA was 10-fold more active with
the plant type Fd I than with Adx under the same
conditions. In the mean time, we cloned M. tuberculosis
genes coding for 7Fe and 3Fe ferredoxins, but failed in
obtaining the overexpression in E.coli. Several years ago, a
7Fe ferredoxin was purified from M. smegmatis [11]. By
using a similar procedure, we obtained a reasonable amount
of the M. smegmatis 7Fe ferredoxin in homogeneous form
as judged by several criteria (native and denaturing PAGE,
protein determination/molarity determined by using the
reported extinction coefficient at 406 nm). N-Terminal
analysis of the purified protein confirmed its identity with
Fig. 3. Spectral perturbations elicited by ligand binding to FprA. All
measurements were performed in 10 m
M

–
Æs
)1
Æl
M
)1
)
K
acceptor
m
(lM)
k
cat
/K
m
(e
–
Æs
)1
Æl
M
)1
)
NADPH
K
3
Fe(CN)
6
63.0 ± 1.3 0.45 ± 0.02 140 ± 0.1 22 ± 2 2.9 ± 0.1
DPIP 25.6 ± 0.8 0.89 ± 0.08 29 ± 0.1 58 ± 3.8 0.44 ± 0.07

Figure 3B shows the difference spectrum obtained at
saturating concentration of Fd I. Two positive peaks
appeared centered around 450 and 380 nm, where FprA
has absorption maxima. An approximate K
d
value of 2 l
M
was obtained by titration. The interaction between the two
proteins was further investigated by using cross-linking
agents. Following incubation of the two proteins with
N-ethyl-3-(3-dimethylaminopropyl)carbodiimide, FprA was
fully converted to protein adducts of about 66 kDa as
determined by SDS/PAGE. This is the expected value
for a 1 : 1 cross-linked complex between the flavopro-
tein and Fd I [12]. The cross-linked species acquired the
capacity to reduce directly cytochrome c as judged by
measuring the cytochrome c reductase activity in the
absence of added Fd I. The same type of experiments were
repeated replacing the spinach protein with the 7Fe
ferredoxin. A cross-linked protein of about 66 kDa was
also obtained, although at a lower rate of formation with
respect to the plant ferredoxin.
Anaerobic reduction of FprA with NAD(P)H
Bovine AdR shows peculiar behavior when anaerobically
reduced by NADPH [21]. We therefore tried to verify
whether FprA presented the same reduction pattern when
treated with physiological reductants. Identification of
reduced intermediates could help in elucidating the mech-
anism of action of FprA. The titration of FprA with the less
efficient substrate NADH practically superimposed to that

+
of the fully
Fig. 4. NADPH reduction of FprA. The titration was performed in
10 m
M
Tris/HCl, pH 7.4 under anaerobiosis. 47 l
M
FprA was titrated
with NADPH. The spectra recorded at 0, 0.3, 0.45, 0.6, 0.7, 0.9, 1 (A)
and at 1.3, 1.6, 1.9, 3, 6 (B) NADPH/FAD molar ratios are reported.
The inset shows the plot of the absorbance at 625 nm due to SQ,
obtained by subtracting the contribution of charge-transfer species as
detailed in Materials and methods, as a function of NADPH/FAD
molar ratio.
Table 3. Kinetic parameters for the 2Fe and 7Fe ferredoxin reductase reactions of FprA.
Electron acceptor
k
cat
(e
–
Æs
)1
)
K
NADPH
m
(l
M
)
k

M. smegmatis
ferredoxin
3.4 ± 0.27 3.5 ± 0.72 0.97 ± 0.21 0.03 ± 0.004 110 ± 17
3010 F. Fischer et al. (Eur. J. Biochem. 269) Ó FEBS 2002
reduced enzyme obtained by photoreduction was performed
(Fig. 5A). The spectra resemble those already observed
during the early steps of NADPH titration of oxidized
enzyme (Fig. 4A). It can be noted that both the absorbance
at 450 and 340 nm of the solution increased at each addition
of NADP
+
up to 1 NADP
+
per FAD (see inset) and no SQ
was formed. Thus, the spectrum of the CT formed in this
experiment, which is superimposable to that formed in the
titration of the oxidized enzyme with a molar amount of
NADPH, is mostly due to FAD-NADPH charge transfer,
as can be judged from the high absorbance at 340 and
450 nm, and low absorbance at 750 nm. The SQ amount
present during NADPH titration could then be calculated
by subtracting from the spectra the contribution of the CT
species as obtained from the experiment shown in Fig. 5A.
In the inset of Fig. 4B, the absorption changes due to SQ
accumulation are plotted against the NADPH/flavin molar
ratio. It can be observed that the SQ built up only after one
NADPH/flavin was added, reached its maximum after
addition of slightly more than two NADPH/flavin, and
then remained at this level notwithstanding the high amount
of NADPH added. Indeed, full reduction of the bound

red
-NADPH $ 2E
sq
-NADPH
where CT indicate an equilibrium mixture of the two
charge-transfer species FAD-NADPH and FADH
2
-
NADP
+
.
DISCUSSION
The functional annotation of proteins identified in genome
sequencing projects is based on protein sequence similarities
to homologs in the databases. However, due to the
possibility of divergent evolution, homologous enzymes
may not catalyze the same reaction. Thus, a biochemical
characterization of the gene product is required to establish
the protein’s real function in that organism. This was
particularly necessary in the case of the fprA gene product of
M. tuberculosis, because of the absence in the bacterial
genome of genes coding for [2Fe)2S] ferredoxins, the
expected protein substrate for an adrenodoxin reductase-
like enzyme. To our knowledge, this is the first adrenodoxin
reductase-like protein from a bacterium to be characterized.
The recombinant enzyme was shown to be a flavoprotein
containing noncovalently bound FAD, whose fluorescence
was nearly fully quenched. This is a remarkable difference
from the mammalian enzyme, the flavin of which is
fluorescent [22,23]. The fprA gene product did not show

addition after or before FprA photoreduction.
Photoreduction of FprA was performed in 10 m
M
Hepes-KOH,
pH 7.0, in the presence of 15 m
M
EDTA and 1.8 l
M
5-deazaribofla-
vin. (A) NADP
+
titration of 26.5 l
M
photoreduced FprA. The spectra
recorded at 0, 0.1, 0.3, 0.4, 0.5, 0.7, 0.8, 1 NADP
+
/FAD molar ratios
are reported. The inset shows the absorbance changes at 452 (s), 550
(d)and750nm(h) as a function of NADP
+
/FAD molar ratio. The
absorbance change at 750 nm has been multiplied by four for clarity.
(B) photoreduction of 20 l
M
FprA in the presence of 30 l
M
NADP
+
.
The spectra recorded before and after 1.5, 2.5, 3.5 min irradiation

because a homolog is present in M. leprae, whose genome is
greatly downsized and degraded [25]. On the basis of the
high similarity of FprA with mammalian AdR (Table 1), its
enzymatic function may be inferred. In mitochondria, AdR,
with a [2Fe)2S] ferredoxin, is part of an electron chain
which delivers electrons from NADPH to cytochrome P450
enzymes, mainly involved in hydroxylation reactions
[9,19,20]. The M. tuberculosis genome is rich in genes
encoding P450 cytochromes (22 genes, see [1]), whereas it
lacks genes coding for Adx-type ferredoxins and it contains
only genes encoding 7Fe and 3Fe ferredoxins [1]. In
bacteria, different systems for P450 cytochrome reduction
are employed. Well known is the system comprising
putidaredoxin reductase, a NADH-dependent flavoprotein,
and putidaredoxin (2Fe ferredoxin), which transfers elec-
trons to P450
cam
[26]. This system is similar to the
mammalian AdR-Adx. A microsomal-type P450 reductase
instead is present in Bacillus megaterium [27]. Apparently,
purification of the reductase from other bacteria was
unsuccessful due to protein instability and low expression
level. A microbial cytochrome P450 reduction system was
purified from Streptomyces griseus grown in a soybean
flour-enriched medium [28]. The ferredoxin reductase was
a NADH-dependent flavoprotein of 60 kDa with a
N-terminal sequence comprising a FAD binding consensus
sequence (GXGXXG), which is typical of the glutathione
reductase large family [29], to which AdR also belongs.
They showed that this enzyme can couple electron transfer

the reduction of P450 enzymes as is the case of the other
bacerial reductases cited above. Recently, the P450
14a-demethylase of M. tuberculosis has been characterized
and suggested to be involved in the cholesterol biosynthetic
pathway [39]. Cholesterol has been shown to be essential to
M. tuberculosis infection [40]. Furthermore, some of the
cytochrome P450 enzymes could be involved in the synthesis
of the complex cell wall components. Thus, if FprA
provides electrons to several pathways through the inter-
action with several ferredoxins, it represents a potential
target for antimycobacterial drugs. Crystals of FprA have
been obtained and the three-dimensional structure is being
currently determined.
ACKNOWLEDGEMENTS
This work was carried out with funds from the Ministero dell’Univer-
sita
`
e della Ricerca Scientifica e Tecnologica (Prin 1999) and European
Union (EU Cluster QLK2-2000–01761). We thank Dr G. Riccardi
(University of Genova), Dr R. Cantoni and Dr M. Branzoni
(University of Pavia) for help in cloning and DNA sequencing,
Dr A. Negri and Dr G. Tedeschi for protein microsequencing, and
Dr M. A. Vanoni and Dr B. Curti for helpful discussions.
REFERENCES
1. Cole, S.T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C.,
Harris, D., Gordon, S.V., Eiglmeier, K., Gas, S., Barry III, C.E.
et al. (1998) Deciphering the biology of Mycobacterium tuber-
culosis from the complete genome sequence. Nature 393, 537–544.
2. McKinney, J.D. (2000) In vivo veritas: the search for TB drug
targets goes live. Nat. Med. 6, 1330–1333.

characterization of a ferredoxin from Mycobacterium smegmatis
Takeo. Biochim. Biophys. Acta 743, 91–97.
12. Zanetti, G., Curti, B. & Aliverti, A. (1984) A cross-linked complex
between ferredoxin and ferredoxin-NADP
+
reductase. J. Biol.
Chem. 259, 6153–6157.
13. Aliverti, A., Curti, B. & Vanoni, M.A. (1999) Identifying and
quantitating FAD and FMN in simple and iron-sulfur-containing
flavoproteins. In Methods in Molecular Biology, Vol. 131 Flavo-
protein Protocols (Chapman, S.K. & Reid, G.A., eds), pp. 9–23.
Humana Press Inc., Totowa, NJ.
14. Coves, J. & Fontecave, M. (1993) Reduction and mobilization of
iron by a NAD(P)H: flavin oxidoreductase from Escherichia coli.
Eur.J.Biochem.211, 635–641.
15. Massey, V. & Hemmerich, P. (1977) A photochemical procedure
for reduction of oxidation-reduction proteins employing deazari-
boflavin as catalyst. J. Biol. Chem. 252, 5612–5614.
16. Massey, V. & Palmer, G. (1962) Charge-transfer complexes
of lipoyl dehydrogenase and free flavins. J. Biol. Chem. 237,
2347–2358.
17. Sakuray, T. & Hosoya, H. (1966) Charge-transfer complexes of
nicotinamide-adenine dinucleotide analogues and flavin mono-
nucleotide. Biochim. Biophys. Acta 112, 459–468.
18. Massey, V. & Hemmerich, P. (1980) Active-site probes of flavo-
proteins. Biochem. Soc. Trans. 8, 246–257.
19. Hanukoglu, I. (1992) Steroidogenic enzymes: structure, function,
androleinregulationofsteroidhormonebiosynthesis.J. Steroid
Biochem. Mol. Biol. 43, 779–804.
20. Bernhardt, R. (1996) Cytochrome P450: structure, function, and

28. Ramachandra, M., Seetharam, R., Emptage, M.H. & Sariaslani,
F.S. (1991) Purification and characterization of a soybean flour-
inducible ferredoxin reductase of Streptomyces griseus. J. Bacteriol.
173, 7106–7112.
29. Dym, O. & Eisenberg, D. (2001) Sequence-structure analysis of
FAD-containing proteins. Protein Sci. 10, 1712–1728.
30. O’Keefe, D.P., Gibson, K.J., Emptage, M.H., Lenstra, R.,
Romesser,J.A.,Litle,P.J.&Omer,C.A.(1991)Ferredoxinsfrom
two sulfonylurea herbicide monooxygenase systems in Strepto-
myces griseolus. Biochemistry 30, 447–455.
31. Sielaff,B.,Andreesen,J.R.&Schra
¨
der, T. (2001) A cytochrome
P450 and a ferredoxin isolated from Mycobacterium sp. strain
HE5 after growth on morpholine. Appl. Microbiol. Biotechnol. 56,
458–564.
32. Manzella, L., Barros, M.H. & Nobrega, F.G. (1998) ARH1 of
Saccharomyces cerevisiae: a new essential gene that codes for a
protein homologous to the human adrenodoxin reductase. Yeast
14, 839–846.
33. Barros, M.H. & Nobrega, F.G. (1999) YAH1 of Saccharomyces
cerevisiae: a new essential gene that codes for a protein homo-
logous to human adrenodoxin. Gene 233, 197–203.
34. Lacour, T., Achstetter, T. & Dumas, B. (1998) Characterization of
recombinant adrenodoxin reductase homologue (Arh1p) from
yeast. Implications in in vitro cytochrome P45011b monooxyge-
nase system. J. Biol. Chem. 273, 23984–23992.
35. Lange, H., Kaut, A., Kispal, G. & and Lill, R. (2000) A
mitochondrial ferredoxin is essential for biogenesis of cellular iron-
sulfur proteins. Proc. Natl Acad. Sci. USA 97, 1050–1055.