Eur. J. Biochem.

220, 439-446 (1994)

0 FEBS 1994

A purified ferredoxin from Giardia duodenalis

Steven M. TOWNSON', Graeme R. HANSON', Jacqueline A. UPCROFP and Peter UPCROFT'
' Queensland Institute of Medical Research, The Bancroft Center, Brisbane, Australia
* Centre for Magnetic Resonance, The University of Queensland, Brisbane, Australia
(Received September 13November 30, 1993) - EJB 93 1401/3

A ferredoxin has been purified to homogeneity from the ancient protozoan parasite Giardia
duodenatis. As far as we know, this is the first electron transport protein to be characterised from
the organism. The ferredoxin exhibits absorption maxima at 296 and 406 nm with molar absorption
coefficients of E~~~ = 16650?240 M-' cm-' and E~~~ = 131002 370 M-' cm-' respectively. The
A406/A296ratio ranged over 0.78-0.82. The molecular mass of the apoprotein calculated by mass
spectrometry was 5730 -+ 100Da and the minimum molecular mass by amino acid analysis was
5926Da. There were four cysteine residues/molecule protein but no methionine, arginine, histidine
or tyrosine. The absence of these latter residues is consistent with the amino acid content of most
ferredoxins. The N-terminal amino acid sequence exhibited greatest similarity to Desulfovibrio gigas
ferredoxin I1 and indicated the potential to coordinate an iron-sulfur cluster. There were
3.21 ? 0.41 mol sulfide and 2.65 ? 0.06 mol iron/mol protein. Electron paramagnetic resonance
studies of this protein have indicated the presence of an iron-sulfur centre consistent with those
of known ferredoxins. Ferredoxin serves as a biological electron acceptor from giardial pyruvate
dehydrogenase with metronidazole as a terminal electron acceptor. Such a pathway may serve as a
possible mechanism for the reductive activation of metronidazole in this parasite. A second ferre-
doxin has been purified to homogeneity, but at this stage there is insufficient material to fully
characterise this protein. No other low-molecular-mass electron transport proteins have been iden-
tified in Giardia under the growth conditions described.

Giardia duodenalis is an intestinal tract parasite of mam-
mals and one of a few anaerobic protozoa described which
lack morphologically identifiable mitochondria. Some of the
amitochondrial microaerophilic eukaryotes which have been
examined, Tritrichomonasfoetus and Trichomonas vaginalis,
have evolved a specialised organelle, the hydrogenosome,
which contains a simple electron transport chain similar to
components of the electron transport chains in anaerobic bac-
terial systems (Gorrell et al., 1984; Marczak et al., 1983). G.
duodenalis, which is regarded as an evolutionarily ancient
eukaryote (Sogin et al., 1989; Upcroft et al., 1991), has not
evolved such an organelle, suggesting that electron transport
in this organism may be primitive and unique among the
eukaryotes.
It has been proposed that G. duodenalis has a respiratory
chain consisting of iron-sulfur proteins and flavins and is
deficient in heme-containing proteins such as cytochromes
(Lindmark, 1980; Weinbach et al., 1980), although very little
is known about the individual components of the electron
transport chain. Ellis et al. (1993) have recently reported an
EPR study on a soluble acidic and on crude membrane frac-
tions of Giardia lamblia (strain H-1-P) and have identified a
Correspondence fo P. Upcroft, Queensland Institute of Medical
Research, The Bancroft Center, 300 Herston Rd.; Brisbane, Queens-
land, Australia 4029
Fax: +61 7 362 0105.
Abbreviations. HRLC, high-resolution liquid chromatography;
ICP-MS, inductively coupled plasma mass spectrometry; LSE, least-
squares-error parameter.
Enzyme. Pyruvate:ferredoxin oxidoreductase (EC 1.2.7.1).
number of signals arising from iron-sulfur clusters. There is
some evidence to suggest that G. duodenalis contains a par-
ticulate respiratory chain different from that in mitochondria
and not associated with oxidative phosphorylation, ATP
production being dependent on substrate level phosphoryla-
tion (Weinbach et al., 1980). More recent evidence has sug-
gested that this organism relies on pyrophosphate as an en-
ergy source which is consistent with G. duodenalis being
regarded as a primitive organism (Mertens, 1993).
Iron-sulfur proteins have been classified into two classes,
namely simple and complex proteins (Nomenclature Com-
mittee IUB, 1991), where the latter class of proteins contain
additional cofactors, such as molybdenum in the molybdben-
um-iron cofactor associated with nitrogenase (Kim and Rees,
1992). Rubredoxins and ferredoxins, classified as simple
iron-sulfur proteins, are usually involved in electron transfer
reactions (for recent reviews, see Beinert, 1990; Cammack,
1992; Bruschi and Guerlesquin, 1988; Howard and Rees,
1991). Whde rubredoxins contain one or more F' , - ~ C~S]('-.~-)
(S = 5/2, 2) centers, other types of iron-sulfur clusters
have been identified in ferredoxins: [2Fe-2S](2'."), (S = 0,
112); [3Fe-4S]('+.0), (S = 1/2, 2) and [4Fe-"](3'~2'.2'.''',
(S = 1/2, 0, 1/2). More recently, a putative [6Fe-
6S](h+.5 + .5+,4+ .4+,3+)cluster has been suggested for a protein
purified from Desulfovibrio vulguris strain Hildenborough
(Pierik et al., 1992) and Desulfovibrio desulfuricans (ATCC
27774) (Moura et al., 1992). In clusters with more than a
single iron atom, there are two or more p-sulfido bridges
with cysteine completing the tetrahedral coordination geome-
try about the high-spin iron ions. In the Rieske protein, two
440
histidines are coordinated to one iron and two cysteines are
coordinated to the other iron atom in the [2Fe-2S] cluster
(Gurbiel et al., 1989).
Ferredoxin- or flavodoxin-like low-potential redox pro-
teins were proposed to play a metabolic role in amitochon-
drial anaerobic protozoa similar to that in anaerobic prokary-
otes (Yoch and Carithers, 1979). Subsequently, ferredoxins
were purified from the anaerobic protozoa, 7: foetus (Marc-
zak et al., 1983), 7: vaginalis, (Gorrell et al., 1984) and Enta-
moeba histolytica (Reeves et al., 1980) and were shown to
be involved in electron transport in these organisms. G. duo-
denalis contains iron-sulfur centers (Weinbach et al., 1980)
and a pyruvate :ferredoxin oxidoreductase activity (or similar
enzyme) capable of reducing ferredoxin (Lindmark, 1980).
The proposed selectivity of nitroimidazole and nitrofuran
drugs against anaerobic organisms such as G. duodenalis is
based on the low redox potentials of components of their
electron transport chains (Muller, 1986). It is thought that
metronidazole, a 5-nitroimidazole, is reduced to its cytotoxic
nitroso radical by the chemical oxidation of ferredoxin or
flavodoxin in anaerobic organisms. Ferredoxin or flavodoxin
can be reduced by a number of enzymes including hydrogen-
ase I during normal metabolism in C. pasteurianum (Church
et al., 1990). It was therefore considered likely that a ferre-
doxin- or flavodoxin-like protein exists in G. duodenalis and
the study of such a protein is important in the context of
electron transport and drug activation.
In this report, we describe the purification and properties
of a ferredoxin from G. duodenalis which serves as a biologi-
cal electron acceptor from pyruvate dehydrogenase. The
study of this protein may help to elucidate the evolutionary
patterns of electron transport that have developed in anaero-
bic eukaryotes to compensate for the absence of an identifi-
able respiratory organelle.
MATERIALS AND METHODS
Materials
All chemicals were of analytical grade or better and were
used without further purification. Water was deionised with
a MilliQ water treatment system. Ammonium sulfate was
Schwartz Mann ultra-pure grade from Beckman. Methyl vio-
logen, coenzyme A, pyruvate, 2-mercaptoethanol and metro-
nidazole were from Sigma. Dithionite was from Fluka.
Organism and culture
G. duodenalis strain BRIS/HEPU/83/106 was cultured in
TYI-S-33 medium (Capon et al., 1989) using Belco dual-
sided roller bottles or Nunc cell factories, to late-log phase.
Trophozoites were harvested, washed twice with degassed
NaClP, (145 mM NaCl, 10 mM Na,HPO,, 3 mM KH,PO,,
pH 7.2) and stored in liquid nitrogen anaerobically as a pellet
until use.
Preparation of trophozoite extract
Trophozoites (approximately 2.5 g wet mass pellet) were
freeze-thawed four times and resuspended in anaerobic
buffer (SO mM Tris/HCl pH 7.5, 2 mM 2-mercaptoethanol)
to a final volume of 12 ml. Buffers were made anaerobic by
prereduction for 72 h in Oxoid anaerobic jars and stored
under nitrogen. Trophozoites were sonicated on ice for two
15-s bursts with a Branson sonifier at 20% output power
(level 2) and examined microscopically for disruption. The
crude cell extract was centrifuged under a nitrogen atmos-
phere at 40000Xg for 30 min at 4°C in a Beckman 52-21/
Me centrifuge. The pellet was discarded.
Ferredoxin purification
Ferredoxin was purified by a modification of the pro-
cedure of Schonheit et al. (1978) for the rapid purification of
ferredoxin from C. pasteurianum. All purification steps were
carried out at 0-4°C with anaerobic buffers except for the
HRLC step which was carried out at room temperature with
nitrogen-sparged buffers. The centrifuged cell extract was ad-
justed to 22.5 ml with buffer (50 mM Tris/HCl pH 7.5, 2 mM
2-mercaptoethanol) and 27.5 ml acetone (at -20°C) was
added and rapidly mixed on ice. Precipitated proteins were
centrifuged at 40000Xg for 30min at 4°C under nitrogen.
The pellet was discarded and acetone was removed from the
supernatant by rotary evaporation in a Sorvall Speedivac for
40 min. The supernatant was then adjusted to 60% saturated
ammonium sulfate with saturated ammonium sulfate and
centrifuged as described above. Non-precipitated proteins
were loaded onto a column (1 .OX 1.5 cm) of DEAE-cellulose
(DE 52, Whatman) column previously equilibrated with 60%
saturated ammonium sulfate. The ferredoxin bound as a
brown band at the top of the column. The column was
washed with 20ml 60% saturated ammonium sulfate and
ferredoxin was eluted with 10% saturated ammonium sulfate.
Ammonium sulfate solutions were buffered with 1 M Tris to
pH 7.5. Elution with 20 mM Tris/HCl pH 7.5 did not remove
all ferredoxin from the column. The eluate was desalted
through a Sephadex G25 medium column (Pharmacia LKB
Biotechnology Inc.) equilibrated with 20 mM Tris/HCl
pH 7.5. The ferredoxin was then loaded onto a Bio-Rad high-
resolution liquid chromatography (HRLC) system with a
Pharmacia MonoP 5/20 anion-exchange column equilibrated
with 20 mM Tris/HCl pH 7.5 and eluted with a linear NaCl
gradient. Ferredoxin constituted the only major protein peak.
Fractions containing ferredoxin were collected and desalted
with a Sephadex G25 medium column equilibrated with
50 mM Hepes, pH 7.5. Pure ferredoxin was stored anaerobi-
cally at 77 K until used. When required, ferredoxin was con-
centrated using Centricon 3 tubes (Amicon) under anaerobic
conditions. Approximately 300 pg pure ferredoxin was ob-
tained from 2.5 g (wet mass) starting material.
Analytical procedures
Electronic absorption spectroscopy was carried out at
20°C with a Cary 4 E spectrophotometer in 50-p1 cuvettes
using version 3 of the Cary spectral processing software
package. N-terminal amino acids were sequenced using an
Applied Biosystems 470A gas-phase sequencer and the phe-
nylthiohydantoin derivatives identified with an on-line Ap-
plied Biosystems liquid chromatograph. Molecular mass was
determined bv matrix-assisted laser desomtion ionisation
mass spectrometry (Karas and Hillenkamp, 1588) with a Fin-
nigan Mat Lasermat mass spectrometer using a 3,5-dime-
thoxy-4-hydroxy-cinnamicacid matrix (70% acetonitrile,
30% HPLC-grade water, 0.1% trifluoroacetic acid, 10 mg/ml
3,5-dimethoxy-4-hydroxy-cinnamic acid). Horse heart cyto-
chrome c (type Ill, Sigma) was used as an internal calibrant.
The molecular mass of Clostridium pasteurianum apofer-
redoxin was calculated to be within 20Da (0.35%) of that
determined from the gene sequence minus methionine

(Graves et al., 1985). Mass spectrometry was also used to
determine the purity of protein preparations. Purity was also
determined by SDSE'AGE (Laemmli, 1970) and gels were
silver stained by the method of Blum et al. (1987). Inorganic
sulfide was determined by the micro method of Beinert
(1983) and iron was determined by inductively coupled
plasma mass spectrometry (ICP-MS) with a Perkin Elmer
Sciex Eland 5000 mass spectrometer. Protein concentration
was determined either by the enhanced bicinchoninic acid
method (Pierce Chemical Co.) or by a modification of the
procedure of Lowry et al. (1951) where reaction with reagent
A was carried out at 50°C for 30min (Aono et al., 1989).
This gave a value for the protein concentration which was
within 10% of that using the bicinchoninic method. Other
protein concentrations were determined by the method of
Bradford (1976) using the Bio-Rad dye reagent kit.
Determination of biological activity
The ability of giardial ferredoxin to accept electrons from
pyruvate dehydrogenase was investigated using metronida-
zole as an artificial electron acceptor (Chen and Blanchard,
1979). Pyruvate dehydrogenase was partially purified ana-
erobically at 4°C as follows. A centrifuged cell extract was
produced as described above for the purification of ferre-
doxin from approximately 0.8 g (wet mass) trophozoite pellet
in buffer (50 mM Tris/HCl, pH 7.5, 20mM 2-mercapto-
ethanol, 0.1 % Triton X-100, 2 mM phenylmethylsulphonyl
fluoride). The extract was loaded onto a DE52 column and
the effluent and wash collected which contained the pyruvate
dehydrogenase activity. Ferredoxin remained bound to the
column. The purification was followed by a standard assay
for the enzyme (Lindmark and Muller, 1973) using E~~~ =
13600 M-' cm-' for methyl viologen (Mayhew, 1978). The
assays contained ferredoxin-depleted extract, 10 mM sodium
pyruvate, 200 pM CoASH, 100 pM metronidazole and giar-
dial ferredoxin in 50 mM Tris/HCl, pH 7.5, 20 mM 2-mer-
captoethanol in a final volume of 200 p1 and were set up in
an anaerobic glove box. Pyruvate was added through a stop-
per to start the reaction. C. pasteurianurn ferredoxin was used
as a control. The absorption coefficient for metronidazole
was taken as E~~~ = 9310 M-' cm-' (Wilson et al., 1974).
Assays were recorded with a Cary 4 E spectrophotometer
with a kinetics program in the spectral processing software.
EPR spectroscopy
X-band EPR spectra were recorded on a Bruker ESP300E
EPR spectrometer equipped with an Oxford Instruments
ESR910 liquid helium flow-through cryostat. Version 3.01 of
esp300e software was employed for the measurement of the
spectra. Calibration of the microwave frequency and the
magnetic field were performed with an EIP 548B microwave
frequency counter and a Bruker 035 M gauss meter, respec-
tively.
Computer simulation of EPR spectra
Computer simulation of the orthorhombic S = 1/2 EPR
spectrum from the [3Fe-4SI1+center was Ijerformed in fre-
quency space and included the effects of g-strain (Pilbrow,
1984, 1990). A nonlinear least-squares simulation program
epr50fit.f running on a SUN SPARCstation 10130 worksta-
tion was used to refine the fit. The quality of the final simu-
lated spectrum can be estimated from the least-squares-error
441
1.o 0.6
0.5
0.75 -
- 0.1
B
Time (min)
Fig. 1. MonoP HRLC chromatograph of G. duodenalis ferre-
doxin. Protein was measured from its absorbance at 280 nm. Ferre-
doxin was eluted with two sucessive linear NaCl gradients and con-
stituted the only major protein peak. A, ferredoxin; B second minor
ferredoxin. Arrows indicate injections and washes.
parameter (LSE) (Martinelli et al., 1989). Spectral subtrac-
tions, comparisons and the determination of LSE were car-
ried out with the EPR software package EPR-PLOT running
on the University of Queensland's Prentice Computer Cen-
ter's VAX 8550 computer.
RESULTS
Purification of ferredoxin
A protein characteristic of anaerobic bacterial ferredoxins
was purified to homogeneity from the aerotolerant anaerobe
G. duodenalis. A second protein has been purified and iden-
tified as a second ferredoxin by electronic spectroscopy of
the oxidized and dithionite reduced protein. The characteriza-
tion of this ferredoxin is in progress.
A number of modifications to the original protocol for
the purification of ferredoxin from C. pasteurianum
(Schonheit et al., 1978) were required to purify ferredoxin
from G. duodenalis. The use of a vacuum desiccator to re-
move the acetone was found more practical than a poly
(ethyleneimine) precipitation for small quantities of protein.
The addition of the mono P ion-exchange column step
(Fig. 1) removed final contaminating proteins and enabled
the separation of the second ferredoxin which ammonium
sulfate precipitation failed to achieve. Aerobic purification
of ferredoxin resulted in a decreased absorbance at 406 nm
consistent with the loss of the iron-sulfur center. Ferredoxin
can be purified from as little as 0.3 g (wet mass) starting
material (approximately 6 X 10' parasites) with this tech-
nique.
The internal membrane structures and the function of
vesicles in G. duodenalis are not defined. However, we have
found that the ferredoxin is not tightly associated with any
identifiable cytoplasmic organelle or membrane structure as
even careful disruption of the cell, leaving vesicles and or-
ganelles intact, released the protein into the soluble cell frac-
tion.
No contaminating proteins were found in purified prepa-
rations of ferredoxin by mass spectrometry or when samples
where tested by silver-stained SDSPAGE gels. The molecu-
lar mass of the apoprotein, approximately 7000Da calculated
by SDSPAGE, differs from the 5730 -C 100Da calculated by
442
1 Table 1. Amino acid composition of G. duodenalis ferredoxin.
The amino acid content of giardial ferredoxin and (in parentheses)
proposed number of residues; n.d. = not determined.
0.75
Amino acid Amount in protein
0 residueslmolecule
5 0.5
;
n
Alanine
Arginine
11.8 (12)
0.0 (0)
a Aspartic acid 3.8 (4)
0.25 Cysteine 3.5 (4)
Glutamic acid 6.1 (6)
Glycine 4.8 (5)
Histidine 0.2 (0)
0
Isoleucine 2.7 (3)
220 320 420 520
Leucine 5.2 ( 5 )
Wavelength (nm) Lysine 2.1 (2)
Methionine 0.2 (0)
Fig. 2. Electronic absorption spectra of ferredoxin from G. duo- Phenylalanine 1.0 (1)
denalis. A, spectrum of air-oxidized ferredoxin (0.268 mglml in Proline 4.1 (4)
50 rnM Hepes pH 7.5); B, ferredoxin reduced anaerobically with a Serine 4.4 (4)
slight e x e s of sodium dithionite; C, spectrum as in A after Threonine 3.8 (4)
acidification to pH 3 with HC1. Tryptophan n.d.
Tyrosine 0.2 (0)
Valine 7.1 (7)
mass spectromery and 5926Da determined by amino acid Total 61
analysis. This is consistent with previous findings ; ferredox- Minimum M , 5926
ins are thought not to bind an appropriate amount of sodium
dodecyl sulfate during gel electrophoresis (Lode et al., 1976;
Schlatter et al., 1985; Yang et al., 1977; Yoch, 1973). The
protein was not stable when exposed to air for extended peri-
ods of time, and was completely denatured when stored aero-
bically for 48 h at 4°C. However, G. duodenalis ferredoxin 1.
was stable for several months when stored anaerobically in
2.
liquid nitrogen.

Spectral properties
The electronic absorption spectra of the oxidized and re-
duced ferredoxin (Fig. 2) were consistent with the spectra of
proteins containing an iron-sulfur cluster. The protein exhib-
its absorption peaks at 296nm and 406nm with molar ab-
*
sorption coefficients of 166502 240 and 13100 370 M-'
cm-', respectively. These molar absorption coefficients are
similar to the values reported for the single 3Fe cluster ferre-
doxin from Methanosarcina thermophila (Terlesky and Ferry,
1988b). The A,,/A,,, ratio ranged over 0.78-0.82 which is
typical of ferredoxins and supports our conclusion that the
protein is of high purity. The G. duodenalis ferredoxin was
readily reduced anaerobically with sodium dithionite. The re-
duction resulted in a 45% decrease in absorbance at 406 nm.
After acidification to pH 3 with HC1, the absorbance de-
creased in the 406-nm and 296-nm regions indicating that a
significant proportion of the iron-sulfur clusters were re-
moved from the protein. Concomitant with this decrease in
absorbance was the release of hydrogen sulfide. The absor-
bance decrease after acidification is similar to the decrease
observed at 406 nrn when the protein is denatured at 80°C
(data not shown). No other chromophores were detected.
Amino acid analysis and N-terminal sequence
Table 1 shows the amino acid composition of ferredoxin
isolated from G. duodenalis. Methionine, arginine, histidine
and tyrosine residues were absent which is consistent with
the amino acid profile of ferredoxins. In addition there is
a smaller number of acidic residues (16.4%) present when
4.
5.
6.
Fig.3. Alignment of G. duodenalis N-terminal amino acid se-
quence with several other ferredoxins. Numbering is according
to G. duodenalis ferredoxin except for Desutfovibrio desuljiiricans
Norway ferredoxin 11. 1, G. duodenalis; 2, D. gigus ferredoxin 11
(Bruschi, 1979); 3, D. desulfuricans ferredoxin I1 (Guerlesquin et
al., 1984); 4, Methanosarcina thermophila (Terlesky and Ferry,
1988b); 5, Methanosarcina barkeri (DSM 800) (Hausinger et al.,
1982); 6, M . barkeri (DSM 804) (Hatchikian et al., 1982).
compared with ferredoxins from other sources such as Meth-
anosarcina barkeri (31.7%) (Hatchikian et al., 1982). There
were only 4 mol cysteine/mol protein. The calculated mini-
mum molecular mass of the apoprotein from the amino acid
analysis was 5926Da which compared well with the mass
calculated for the apoferredoxin by mass spectrometry. The
N-terminal amino acid sequence is presented in Fig. 3. The
arrangement of cysteines in the N-terminus indicates the po-
tential to coordinate one iron-sulfur cluster. The iron-sulfur
binding motif (-C,-X-X-C;-X-X-C, ......C,-P-) found in [4Fe-
4S]('+,'+)ferredoxins (Cammack, 1992) is modified in G. du-
odenalis. There is a non-conservative substitution of alanine
for the second cysteine (C;) in the N-terminal motif. The
iron-sulfur binding motif in the ferredoxin from Giardia is
then modified to (-C,-X-X-A-X-X-C, ......C,-P-). Only one
polypeptide chain was found by N-terminal amino acid se-
quencing of purified ferredoxin. The protein databases were

Table 2. EPR parameters for [3Fe-4SI1+clusters. crRI(i = x,y,z) are residual linewidths due to dipolar broadening and/or unresolved metal
and ligand hyperfine interactions, while Ag,/g, represent the half width of the strain-induced distribution of g values.
kHZ
Giardia
ferredoxin 1.9500 1.9799 2.0210 14.98
E2, lactyl-CoA 1.982 1.995 2.109 0.0
dehydratase
searched with the N-terminal amino acid sequence of G. duo-
denalis ferredoxin using the Pearson and Lipman programme
FASTA (Pearson and Lipman, 1988) in the GCG suite of
programmes (Devereux et al., 1984). A number of methano-
gen and Desulfovibrio ferredoxins were shown to have se-
quence similarity with the N-terminus of the ferredoxin from
G. duodenalis. There was a 44% similarity with D. gigas
ferredoxin I1 and Desulfovibrio desulfuricans ferredoxin I1
(Fig. 3).
Iron-sulfur determination
There were 3.21 20.41 (n = 5) mol sulfide and
2.65 -C 0.06 (n = 2) mol irordmol ferredoxin. Although there
are a number of inherent problems associated with the deter-
mination of s*-(Beinert, 1983), the data are consistent with
the presence of a [3Fe-4S]('+.O)cluster for this ferredoxin.
Biological activity
Metronidazole is an artificial electron acceptor that is
non-enzymically reduced by ferredoxin and was used to as-
sess electron transfer between giardial pyruvate dehydroge-
nase and ferredoxin. The ferredoxin from G. duodenalis read-
ily accepts electrons from giardial pyruvate dehydrogenase
when metronidazole is the final electron acceptor. Metronida-
zole reduction did not occur in the absence of ferredoxin. As
little as 90 pmol ferredoxin was required to show electron
transfer to metronidazoie. Further purification of pyruvate
dehydrogenase has been achieved, but enzymic activity
decreases due to the lability of the enzyme. We considered
that electron transfer reactions were best carried out with a
highly active enzyme preparation. The limited amount of
protein material meant assays were carried out at non-saturat-
ing ferredoxin levels. C. pasteurianum ferredoxin was an
equivalent electron acceptor in the same assay system. The
pyruvate dehydrogenase activity was assayed in a ferredoxin-
depleted extract (the cell extract eluant from the DE52 col-
umn). At a specific pyruvate dehydrogenase activity of
2058 nmol methyl viologen reduced . min-' . mg protein-',
metronidazole was reduced at the rate of 68 and 141 nmol
. min-' . mg protein-' with 200 and 400 pmol, respectively,
of added G. duodenalis ferredoxin.
EPR spectroscopy
The X-band EPR spectrum of oxidized Giardia ferre-
doxin at 4 K (Fig. 4a) reveals resonances around g = 2 typi-
cal of species containing a single unpaired electron. No other
443
MHz
7.310 0.1259 2.97 6.66 24.10 this work
0.0 0.0 52.7 39.1 17.0 Kuchta
et al., 1986
3.20 3.35 3.50 3.65 3.80
Magnetic field (mT)
Fig. 4. EPR spectra of oxidized G. duodenalis ferredoxin. (a) Ex-
perimental spectrum, in 50mM Hepes pH = 7.5, 4.2 K, u =
9.4447GHz; (b) computer simulation (LSE = 2.237X10-').
signals were observed over wider scan widths, 1.5 T. Com-
puter simulation (see Materials and Methods) of the g = 2
resonances with an orthorhombic S = 1/2 electron Zeeman
spin Hamiltonian and the anisotropic g and linewidth param-
eters listed in Table 2 yields the spectrum shown in Fig. 4b.
The least-squares-error parameter (LSE = 2.237X indi-
cates a satisfactory interpretation of the data. Anaerobic re-
duction of the protein with sodium dithionite results in the
loss of the EPR spectrum. The anisotropic g values (Table 2)
and the redox behaviour of the iron-sulfur center are consis-
tent with a cubane-like [3Fe-4SI1+cluster.
DISCUSSION
The purification and characterisation of a ferredoxin from
G. duodenalis represents an important starting point for the
study of electron transport pathways in this parasite. Ferre-
doxins from other anaerobic amitochondrial protozoa have
been isolated and characterized but are very different from
Giardia ferredoxin in size, amino acid sequence and iron-
sulfur clusters. Ferredoxin from E. histolytica contains two
[4Fe-4S](2+,'+ ) clusters and is closely related to clostridial
ferredoxins (Huber et al., 1988) while T vaginalis (Gorrell
et al., 1984) and T foetus (Marczak et al., 1983) ferredoxins
have a single [2Fe-2S](2+.'+)center. The ferredoxin from T
vaginalis shows some relatedness to ferredoxins from mixed
oxidase systems (Johnson et al., 1990) and serves as the bio-
logical electron acceptor for pyruvate :ferredoxin oxidoreduc-
444

tase. This enzyme is found in anaerobic eubacteria and ar- an ancestral peptide of 26 amino acids, the gene for which
chaebacteria and is considered to be the ancient anaerobic underwent various duplications and deletions producing a di-
equivalent of the pyruvate dehydrogenase multienzyme com- verse range of functionally related proteins including the typ-
plex of respiratory eubacteria and mitochondria (Kerscher ical bacterial 2[4Fe-4S](2+,'+)ferredoxins (Otaka and Ooi,
and Oesterhelt, 1982). 1987, 1989). Further evolution of 2[4Fe-4S](2+~"'ferredox-
Enzymic activity from Giardia trophozoite homogenates ins led to loss of one of the clusters in a number of organisms
has been previously referred to as pyruvate synthase, pyru- (Fukuyama et al., 1988). In genera such as Methanosarcina,
vate :ferredoxin oxidoreductase and pyruvate :acceptor oxi- Clostridium and Desulfovibrio, there are species containing
doreductase in the absence of a purified and characterized either dual or monocluster ferredoxins. It is interesting that,
enzyme (Lindmark, 1980; Lindmark and Miller, 1988; Ellis in monocluster ferredoxins, it is the N-terminal cluster that
et al., 1992). Many enzyme activities have been partially is apparently always maintained; the evolutionary reasons for
characterized in cell homogenates that have been called pyru- this are unclear at present. The effect on electron transfer
vate :ferredoxin oxidoreductase, e.g. Bacteroides fragilis efficiency, metabolic pathways and the evolutionary conse-
(Narikawa et al., 1991). In Halobacterium halobium quences of loss of the second cluster are yet to be defined.
(Kerscher and Oesterhelt, 1981) and C. pasteurianum (Sauer The purification of a ferredoxin from Giardia has al-
et al., 1976) the enzymic activity attributable to the oxidative lowed us to examine the possible role of this protein in the
decarboxylation of pyruvate or the reverse reaction was activation of metronidazole in the parasite since redox path-
found to be associated with more than one enzyme. On this ways have been implicated in activation of nitroheterocyclic
basis, we have called the enzyme from G. duodenalis pyru- drugs. By coupling electron transport with pyruvate dehydro-
vate dehydrogenase until the enzyme(s) is further charac- genase and using metronidazole as an artificial electron ac-
terized. ceptor, we have shown that such a mechanism of metronida-
The arrangement and number of cysteines in a ferredoxin zole reduction can occur in vitro as proposed, although the
molecule can be used to predict the type of cluster accommo- overall biological relevance of this observation needs to be
dated by a protein (Beinert, 1990; Cammack, 1992). The placed in its correct perspective. It is not known whether the
amino acid analysis of G. duodenalis indicates that there are reduction seen in vitro occurs in vivo since metronidazole can
four cysteine residues in the protein molecule and the N- be used as an artificial electron acceptor in other ferredoxin
terminal sequence reveals that cysteine (C;) is replaced by systems, e.g. the carbon monoxide dehydrogenase system
alanine in the iron-binding motif for a typical [4Fe-4S](z+.'+) from Methanosarcina thermophila (Terlesky and Ferry,
cluster. This indicates that only a 3Fe cluster may be accom- 1988a). Furthermore the pyruvate dehydrogenase ferredoxin
modated by this protein. This distribution of cysteine resi- pathway may not be the sole system for metronidazole acti-
dues for the incorporation of an iron-sulfur cluster into a vation, although at present there is no evidence for the exis-
ferredoxin is similar to that found in the ferredoxins from tence of other ferredoxin-reducing pathways or enzymes such
Streptomyces griseolus where the second cysteine in the iron- as xanthine oxidase which are capable of reducing
sulfur cluster motif is also replaced by alanine (O'Keefe et nitroheterocyclic drugs (Tatsumi et al., 1976).
al., 1991). These ferredoxins are biologically active just as
we have shown that the ferredoxin from G. duodenalis is We would like to thank Mackay Edmundson, Chris Cook and
capable of electron transfer from giardial pyruvate dehydro- Paul Hieschler for help with amino acid sequencing, amino acid
genase to an artificial electron acceptor. analysis and ICP-MS, respectively. S. M. T. is a recipient of an
The EPR spectra of G. duodenalis ferredoxin exhibit res- Australian Postgraduate Research Award. This research was funded
onances around g = 2 (Table 2) which are inconsistent with by the National Health and Medical Research Council of Australia.
a rubredoxin center ([Fe-Cys,]'+) as the ferric ion has an S =
5/2 ground state producing resonances around g = =9, ~ 4 . 3
and ~ 0 . 6 Although
. the [2Fe-2SI1+, [3Fe-4SI1+ and [4Fe-
4S]'+,'+ clusters have S = 1/2 ground states, the [2Fe- REFERENCES
2S]'+,'+ and the [4Fe-4SI2+.' clusters exhibit different redox Aono, S., Bryant, F. 0. & Adams, M. W. W. (1989) A novel and
+

behaviour (i.e. oxidised proteins are EPR silent and upon remarkably thermostable ferredoxin from the hypothermophilic
reduction EPR signals are produced). Of the two remaining archaebacterium Pyrococcus furiosus, J. Bacteriol. 171, 3433 -
clusters ([3Fe-4SI1+and [4Fe-4SI3+)the EPR spectrum of G. 3439.
duodenalis ferredoxin can be assigned to an [3Fe-4SI1+clus- Beinert, H. (1983) Semi-micro methods for the analysis of labile
ter on the basis of the observed g anisotropy and spectral sulfide and of labile sulfide plus sulfane sulfur in unusually sta-
lineshape (Kuchta et al., 1986, Kent et al., 1980). The origin ble iron-sulfur proteins, Anal. Biochem. 131, 373-378.
Beinert, H. (1990) Recent developments in the field of iron-sulfur
of the S = 1/2 ground state has been described in terms of proteins, FASEB J. 4 , 2483 -2491.
isotropic antiferromagnetic Heisenberg exchange between Blum, H., Beier, H. & Gross, H. J. (1987) Improved silver staining
three high-spin iron(II1) ions (S = 5/2) (Kent et al. 1980). of plant proteins, RNA and DNA in polyacrylamide gels, Elec-
The iron and sulfide content (2.65 -t 0.06 mol iron and trophoresis 8, 93-99.
3.21 -t 0.41 mol sul€ide/mol protein) indicates the presence Bradford, M. M. (1976) A rapid and sensitive method for the quanti-
of a [3Fe-4S]('+.O)cluster in the ferredoxin from G. duode- tation of microgram quantities of protein utilizing the principle
nalis as purified. Recent EPR evidence (Ellis et al., 1993) of protein-dye binding, Anal. Biochem. 72, 248-254.
has suggested that the putative ferredoxin in G. lamblia may Bruschi, M. (1979) Amino acid sequence of Desulfovibrio gigas
contain one or two [4Fe-4S]('+.'+)clisters. ferredoxin: revisions, Biochem. Biophys. Res. Commun. 91,
623-628.
Ferredoxin from G. duodenalis exhibited greatest amino Bruschi, M. & Guerlesquin, F. (1988) Structure, function and evolu-
acid sequence similarity to methanogen and Desulfovibrio tion of bacterial ferredoxins, FEMS Microbiol. Rev. 54, 155-
ferredoxins of which the ferredoxins from D. gigas and 176.
Methanosarcina thermophila contain a single iron-sulfur cen- Cammack, R. (1992) Iron-sulfur clusters in enzymes: themes and
ter. All ferredoxins are thought to have been derived from variations, Adv. Inorg. Chem. 38, 281 - 322.

Capon, A. G., Upcroft, J. A., Boreham, P. F. L., Cottis, L. E. &
Bundesen, P. G. (1989) Similarities of Giardia antigens derived
from human and animal sources, Int. J. Parasitol. 19, 91-98.
Chen, J.-S. & Blanchard, D. K. (1979) A simple hydrogenase-linked
assay for ferredoxin and flavodoxin, Anal. Biochem. 93, 216-
222.
Church, D. L., Rabin, H. R. & Laishley, E. J. (1990) Reduction of
2-, 4- and 5-nitroimidazole drugs by hydrogenase 1 in Clostrid-
ium pasteurianum, J. Antimicrob. Chemother: 25, 15-23.
Devereux, J., Haeberli, P. & Smithies, 0. (1984) A comprehensive
set of sequence analysis programs for the VAX, Nucleic Acids
Res. 12, 387-395.
Ellis, J. E., Williams, R., Cole, D., Cammack, R. & Lloyd, D. (1993)
Electron transport components of the parasitic protozoan Giardia
lamblia, FEBS Lett. 325, 196-200.
Ellis, J. E., Wingfield, J. E., Cole, D., Boreham, P. F. L. & Lloyd,
D. (1992) Oxygen affinities of metronidazole-resistant and -sen-
sitive stocks of Giardia intestinalis, Int. J. Parasitol. 23, 35-39.
Fukuyama, K., Nagahara, Y., Tsukihara, T., Katsube, Y., Hase, T. &
Matsubara, H. (1988) Tertiary structure of Bacillus thermopro-
teolyticus [4Fe-4S] ferredoxin : evolutionary implications for
bacterial ferredoxins, J. Mol. Biol. 199, 183- 193.
Gorrell, T. E., Yarlett, N. & Miiller, M. (1984) Isolation and charac-
terisation of Trichomonas vaginalis ferredoxin, Carlsberg Res.
Commun. 49, 259-268.
Graves, M. C., Mullenbach, G. T. & Rabinowitz, J. C. (1985) Clon-
ing and nucleotide sequence determination of the Clostridium
pasteurianurn ferredoxin gene, Proc. Natl Acad. Sci. USA 82,
1653- 1657.
Guerlesquin, F., Bruschi, M. & Bovier-Lapierre, G. (1984) Electron
transfer mechanism and interaction studies between cytochrome
C, and ferredoxin, Biochimie 66, 93-99.
Gurbiel, R. J., Batie, C. J., Sivaraja, M., True, A. J., Fee, J. A.,
Hoffman B. M. & Ballou, D. P. (1989) Electron-nuclear double
resonance spectroscopy of IsN-enriched phthalate dioxygenase
from Pseudomonas cepacia proves that the two histidines are
coordinated to the [2Fe-2S] Rieske-type clusters, Biochemistry
28, 4861 -4871.
Hatchikian, E. C., Bruschi, M., Forget, N. & Scandellari, M. (1982)
Electron transport components from methanogenic bacteria: the
ferredoxin from Methanosarcina barkeri (strain Fusaro), Bio-
chem. Biophys. Res. Commun. 109, 1316-1323.
Hausinger, R. P., Moura, I., Moura, J. J. G., Xavier, A. V., Santos,
M. H., LeGall, J. & Howard, J. B. (1982) Amino acid sequence
of a 3Fe:3S ferredoxin from the “archaebacterium” Methanosar-
cina barkeri (DSM 800), J. Biol. Chem. 257, 14192-14197.
Howard, J. B. & Rees, D. C. (1991) Perspectives on non-heme iron
protein chemistry, Adv. Protein Chem. 42, 199-280.
Huber, M., Garfinkel, L., Gitler, C., Mirelman, D., Revel, M. &
Rozenblatt, S. (1988) Nucleotide sequence analysis of an Enta-
moeba histolytica ferredoxin gene, Mol. Biochem. Parasitol. 31,
27 - 34.
Johnson, P. J., D’Oliveira, C. E., Gorrell, T. E. & Miiller, M. (1990)
Molecular analysis of the hydrogenosomal ferredoxin of the an-
aerobic protist Trichomonas vaginalis, Proc. Natl Acad. Sci. USA
87, 6097-6101.
Karas, M. & Hillenkamp, F. (1988) Laser desorption ionization of
proteins with molecular masses exceeding 10000 daltons, Anal.
Chem. 60, 2299-2301.
Kent, T. A,, Huynh, B. H. & Munck, E. (1980) Iron-sulfur proteins:
spin-coupling model for three-iron clusters, Proc. Natl Acad. Sci.
USA 77, 6574-6576.
Kerscher, L. & Oesterhelt, D. (1981) Purification and properties of
two 2-oxoacid :ferredoxin oxidoreductases from Halobacterium
halobium, Eur. J. Biochem. 116, 587-594. . from Desulfovibrio vulgaris (Hildenborough). Characterization
Kerscher, L. & Oesterhelt, D. (1 982) Pyruvate:ferredoxin oxidore-
ductase - new findings on an ancient enzyme, Trends Biochem.
Sci. 7 , 371-374.
Kim, J. & Rees, D. C. (1992) Structural models for the metal centers
in the nitrogenase molybdenum-iron protein, Science 257,
1677-1682.
445
Kuchta, R. D., Hanson, G. R., Holmquist, B. & Abeles, R. H. (1986)
Fe-S centers in lactyl-CoA dehydratase, Biochemistry 2.5, 7301 -
7307.
Laemmli, U. K. (1970) Cleavage of structural proteins during assem-
bly of the head of bacteriophage T4, Nature 227, 680-685.
Lindmark, D. G. (1980) Energy metabolism of the anaerobic proto-
zoon Giardia lamblia, Mol. Biochem. Parasitol. I , 1- 12.
Lindmark, D. G. & Miller, J. J. (1988) in Advances in Giardia re-
search (Wallis, P. M. & Hammond, B. R., eds) pp. 187-189,
University of Calgary Press, Calgary.
Lindmark, D. G. & Miiller, M. (1973) Hydrogenosome, a cyto-
plasmic organelle of the anaerobic flagellate Tritrichomonas foe-
tus, and its role in pyruvate metabolism, J. Biol. Chem. 248,
7724-7728.
Lode, E. T., Murray, C. L. & Rabinowitz, J. C. (1976) Apparent
oxidation-reduction potential of Clostridium acidi-urici ferre-
doxin. Effect of pH, ionic strength, and amino acid replacements,
J. Biol. Chem. 251, 1683-1687.
Lowry, 0. H., Rosenbrough, N. J., Farr, A. L. & Randall, R. J.
(1951) Protein measurement with the folin phenol reagent, J.
Biol. Chem. 193, 265-275.
Marczak, R., Gorrell, T. E. & Miiller, M. (1983) Hydrogenosomal
ferredoxin of the anaerobic protozoan, Tritrichomonas foetus, J.
Biol. Chem. 258, 12427-12433.
Martinelli, R. A., Hanson, G. R., Thompson, J. S., Holmquist, B.,
Pilbrow, J. R., Auld, D. S. & Vallee, B. L. (1989) Characteriza-
tion of the inhibitor complexes of cobalt carboxypeptidase A by
electron paramagnetic resonance spectroscopy, Biochemistry 28,
2251 -2258.
Mayhew, S. G. (1978) The redox potential of dithionite and SO;
from equilibrium reactions with flavodoxins, methyl viologen
and hydrogen plus hydrogenase, Eur: J. Biochem. 85, 535-547.
Mertens, E. (1993) ATP versus pyrophosphate: glycolysis revisited
in parasitic protists, Parasitol. Today 9, 122- 126.
Moura, I., Tavares, P., Moura, J. J. G., Ravi, N., Huynh, B. H., Liu
M.-Y., & LeGall, J. (1992) Direct spectroscopic evidence for the
presence of a 6Fe cluster in an iron-sulfur protein isolated from
Desulfovibrio desulfuricans (ATCC 27774), J. Biol. Chem. 267,
4489-4496.
Miiller, M. (1986) Reductive activation of nitroimidazoles in anaero-
bic microorganisms, Biochem. Pharmacol. 35, 37-41.
Narikawa, S., Suzuki, T., Yamamoto, M. & Nakamura, M. (1991)
Lactate dehydrogenase activity as a cause of metronidazole resis-
tance in Bacteroides fragilis NTCT 11295, J. Antimicrob. Chem-
other: 28, 47-53.
Nomenclature Committee of the International Union of Biochemis-
try (1991) Nomenclature of electron-transfer proteins, Eur: J.
Biochem. 200, 599-611.
O’Keefe, D. P., Gibson, K. J., Emptage, M. H., Lenstra, R., Romes-
ser, J. A., Litle, P. J. & Omer, C. A. (1991) Ferredoxins from two
sulfonylurea herbicide monooxygenase systems in Streptomyces
griseolus, Biochemistry 30, 447-455.
Otaka, E. & Ooi, T. (1987) Examination of protein sequence ho-
molgies : IV. Twenty-seven bacterial ferredoxins, J. Mol. Evol.
26, 257-267.
Otaka, E. & Ooi, T. (1989) Examination of protein sequence homol-
ogies: V. New perspectives on evolution between bacterial and
chloroplast-type ferredoxins inferred from sequence evidence, J.
Mol. Evol. 29, 246-254.
Pearson, W. R. & Lipman, D. J. (1 988) Improved tools for biological
sequence comparison, Proc. Natl Acad. Sci. USA 85, 2444-
2448.
Pierik, A. J., Hagen, W. R., Dunham, W. R. & Sands, R. H. (1992)
Multi-frequency EPR and high-resolution Mossbauer spectros-
copy in a putative [6Fe-6S] prismane-cluster-containing protein
of a supercluster and superspin model protein, Eur: J. Biochem.
206, 705-719.
Pilbrow, J. R. (1 984) Lineshapes in frequency-swept and field-swept
EPR for spin 112, J. Magn. Res. 58, 186-203.
Pilbrow, J. R. (1990) Transition ion electron paramagnetic reso-
nance, Clarendon Press, Oxford.
446
Reeves, R. E., Guthrie, J. D. & Lobelle-Rich, P. (1980) Entamoeba
histolytica: isolation of ferredoxin, Exp. Parasitol. 49, 83- 88.
Sauer, F. D., Bush, R. S. & Stevenson, I. L. (1976) The separation
of pyruvate- ferredoxin oxidoreductase from Clostridium paste-
urianum into two enzymes catalyzing different reactions, Bio-
chim. Biophys. Acta. 445, 518-520.
Schlatter, D., Waldvogel, S., Ziilli, F., Suter, F., Portmann, W. &
Zuber, H. (1985) Purification, amino-acid sequence and some
properties of the ferredoxin isolated from Bacillus acidocal-
darius, Biol. Chem. Hoppe-Seyler 366, 223 -231.
Schonheit, P., Wascher, C. & Thauer, R. K. (1978) A rapid pro-
cedure for the purification of ferredoxin from clostridia using
polyethyleneimine, FEBS Lett. 89, 219-222.
Sogin, M. L., Gunderson, J. H., Elwood, H. J., Alonso, R. A. &
Peatie, D. A. (1989) Phylogenetic meaning of the kingdom con-
cept: an unusual ribosomal RNA from Giardia lamblia, Science
243, 75-77.
Tatsumi, K., Kitamura, S. & Yoshimura, H. (1976) Reduction of
nitrofuran derivatives by xanthine oxidase and microsomes. Iso-
lation and identification of reduction products, Arch. Biophys.
Biochem. 17.5, 131-137.
Terlesky, K. C. & Ferry, J. G. (1988a) Ferredoxin requirement for
electron transport from the carbon monoxide dehydrogenase
complex to a membrane-bound hydrogenase in acetate-grown
Methanosarcina themophila, J. Biol. Chem. 263, 4075 -4079.
Terlesky, K. C. & Ferry, J. G. (1988b) Purification and characteriza-
tion of a ferredoxin from acetate-grown Methanosarcina themo-
phila, J. Biol. Chem. 263, 4080-4082.
Upcroft, J. A., Healey, A., Mitchell, R., Boreham, P. F. L. & Upcroft,
P. (1991) Antigen expression from the ribosomal DNA repeat
unit of Giardia intestinalis, Nucleic Acids Res. 18, 7077-7081.
Weinbach, E. C., Elwood-Claggett, C., Keister, D. B., Diamond, L.
S. & Kon, H. (1980) Respiratory metabolism of Giardia lamblia,
J. Parasitol. 66, 347-350.
Wilson, R. L., Cramp, W. A. & Ings, R. M. J. (1974) Metronidazole
('Flagyl') : mechanisms of radiosensitization, Int. J. Radiat. Biol.
26, 557-569.
Yang, S. S., Ljungdahl, L. G. & LeGall, J. (1977) A four-iron, four-
sulfide ferredoxin with high thermostability from Clostridium
thennoaceticum, J. Bacteriol. 130, 1084- 1090.
Yoch, D. C. (1973) Purification and properties of two ferredoxins
from the nitrogen-fixing bacterium Bacillus polyrnyxa, Arch.
Biochem. Biophys. 1.58, 633-640.
Yoch, D. C. & Carithers, R. P. (1979) Bacterial iron-sulfur proteins,
Microbiol. Rev. 43, 384-421.