Content uploaded by Illarion Turko
Author content
All content in this area was uploaded by Illarion Turko on Jan 03, 2016
Content may be subject to copyright.
Available via license: CC BY 4.0
Content may be subject to copyright.
Structural Basis for Inactivation of Giardia lamblia Carbamate
Kinase by Disulfiram
*
Received for publication, January 27, 2014, and in revised form, February 12, 2014 Published, JBC Papers in Press, February 20, 2014, DOI 10.1074/jbc.M114.553123
Andrey Galkin
‡1
, Liudmila Kulakova
‡1
, Kap Lim
‡1
, Catherine Z. Chen
§
, Wei Zheng
§
, Illarion V. Turko
‡¶
,
and Osnat Herzberg
‡储2
From the
‡
Institute for Bioscience and Biotechnology Research, University of Maryland, Rockville, Maryland 20850,
§
Therapeutics
for Rare and Neglected Diseases, National Center for Advancing Translational Sciences, National Institutes of Health, Bethesda,
Maryland 20892, the
¶
National Institute of Standards and Technology, Gaithersburg, Maryland 20899, and the
储
Department of
Chemistry and Biochemistry, University of Maryland, College Park, Maryland 20742
Background: Carbamate kinase is an essential Giardia lamblia enzyme, and the anti-alcoholism drug disulfiram kills the
trophozoites and inhibits the enzyme.
Results: Disulfiram acts by modifying Cys-242 adjacent to the active site and cures giardiasis in mice.
Conclusion: G. lamblia CK is a good drug target and disulfiram may be repurposed as antigiardiasis drug.
Significance: We need new antigiardiasis drugs because current treatments fail frequently.
Carbamate kinase from Giardia lamblia is an essential
enzyme for the survival of the organism. The enzyme catalyzes
the final step in the arginine dihydrolase pathway converting
ADP and carbamoyl phosphate to ATP and carbamate. We pre-
viously reported that disulfiram, a drug used to treat chronic
alcoholism, inhibits G. lamblia CK and kills G. lamblia tropho-
zoites in vitro at submicromolar IC
50
values. Here, we examine
the structural basis for G. lamblia CK inhibition of disulfiram
and its analog, thiram, their activities against both metronida-
zole-susceptible and metronidazole-resistant G. lamblia iso-
lates, and their efficacy in a mouse model of giardiasis. The crys-
tal structure of G. lamblia CK soaked with disulfiram revealed
that the compound thiocarbamoylated Cys-242, a residue
located at the edge of the active site. The modified Cys-242 pre-
vents a conformational transition of a loop adjacent to the ADP/
ATP binding site, which is required for the stacking of Tyr-245
side chain against the adenine moiety, an interaction seen in
the structure of G. lamblia CK in complex with AMP-PNP.
Mass spectrometry coupled with trypsin digestion confirmed
the selective covalent thiocarbamoylation of Cys-242 in solu-
tion. The Giardia viability studies in the metronidazole-resis-
tant strain and the G. lamblia CK irreversible inactivation mech-
anism show that the thiuram compounds can circumvent the
resistance mechanism that renders metronidazole ineffectiveness
in drug resistance cases of giardiasis. Together, the studies suggest
that G. lamblia CK is an attractive drug target for development of
novel antigiardial therapies and that disulfiram, an FDA-approved
drug, is a promising candidate for drug repurposing.
The enteric protozoan, Giardia lamblia, causes the human
intestinal disease giardiasis, a severe diarrheal disease com-
monly acquired from contaminated freshwater and public
water supplies and by a direct fecal-oral route. Giardiasis is
highly prevalent in the developing world with the latest esti-
mates of 280 million infected people worldwide (1). The infec-
tive Giardia cysts are not destroyed by chemical treatment of
public water sources; thus, the disease is difficult to control in
poor countries lacking adequate water management. Reinfec-
tion reaches as high as 90% in regions where infection is highly
endemic and where environmental contamination is high.
Moreover, treatment failures with standard care drugs such as
metronidazole, tinidazole, and albendazole occur at ⬃20% rate
(2–6). Giardiasis has negative economical impact in underde-
veloped and developing countries. In particular, chronically
infected children suffer from malnutrition, growth retardation,
poor cognitive function, and death. The spread of strains resis-
tant to currently available drugs is a growing concern, and the
unpleasant side effects of these drugs lead to non-compliance.
Finally, prevention of infections through vaccines has proven a
challenge because G. lamblia evade the host immune system by
displaying variant-specific surface proteins. Clearly, there is a
need for new alternative anti-giardia drugs that are not subjects
to current resistance mechanisms.
G. lamblia utilizes the arginine dihydrolase pathway to pro-
duce ATP from ADP and L-arginine (7), a pathway that is absent
in high eukaryotes, including humans. The arginine dihydro-
lase pathway employs three enzymes, arginine deiminase, orni-
thine transcarbamoylase, and carbamate kinase (CK;
3
EC
2.7.2.2). CK catalyzes the last step of the pathway, converting
carbamoyl phosphate and ADP into carbamate and ATP. We
have shown that the enzyme from G. lamblia (G. lamblia CK) is
essential for the survival of the trophozoites, and determined
three crystal structures of the enzyme, one in complex with the
non-hydrolysable ATP analog, AMP-PNP, the second with a
*This work was supported by the National Institutes of Health Grant
R56AI059733 (to O. H.) and Science Applications International Corpora-
tion/NCI, National Institutes of Health contract 11XS049 (to O. H.), and the
Intramural Research Programs of the Therapeutics for Rare and Neglected
Diseases, National Center for Advancing Translational Sciences, National
Institutes of Health (to C. Z. C. and W. Z.).
The atomic coordinates and structure factors (code 4OLC) have been deposited in
the Protein Data Bank (http://wwpdb.org/).
1
These authors contributed equally to this work.
2
To whom correspondence should be addressed: Institute for Bioscience and
Biotechnology, University of Maryland, 9600 Gudelsky Dr., Rockville, MD
20850. Tel.: 240-314-6245; Fax: 240-314-6255; E-mail: osnat@umd.edu.
3
The abbreviations used are: CK, carbamate kinase; MLC, minimum lethal
concentration; AMP-PNP,
␥
-imino-ATP.
THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 289, NO. 15, pp. 10502–10509, April 11, 2014
Published in the U.S.A.
10502 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 289•NUMBER 15• APRIL 11, 2014
by guest on September 18, 2015http://www.jbc.org/Downloaded from
carbamate phosphate analog, citric acid, and the third in the
unbound state (8, 9). These structures revealed several modes
of enzyme conformational flexibility.
The essentiality of G. lamblia CK, the absence of the enzyme
in the human genome, and the ample precedence of “drugga-
blility” of kinases, have led us to target G. lamblia CK for drug
development. To enable high throughput compound library
screening, we have developed two bioluminescence-based as-
says that monitor ATP production. One assay monitors the cellu-
lar ATP content, which correlates with the viability of G. lam-
blia trophozoites, and the second assay measures the ATP pro-
duced by the G. lamblia CK reaction. We used these assays to
identify compounds that both kill the organism and inhibit
G. lamblia CK. We screened the LOPAC
1280
library of pharma-
ceutical active compounds and the NIH Chemical Genomics
Center Pharmaceutical Collection library of approved drugs
(10, 11). Disulfiram (tetraethylthiuram disulfide) was found the
most potent compound that inhibited G. lamblia CK (IC
50
⫽
0.58
M) and killed the Giardia trophozoites (IC
50
, 0.9
M),
while exhibiting no toxicity in HepG2 mammalian cells at the
highest employed concentration (40
M).
Disulfiram, also known as antabuse, is a commonly used drug
for long term treatment of chronic alcoholism. The drug inhib-
its acetaldehyde dehydrogenase by specifically modifying one of
the active site cysteine residues of the enzyme. The cysteine
modification is followed by elimination and formation of a
disulfide bond between two active site cysteine residues (12,
13). Enzyme inactivation leads to an aversive reaction to alcohol
consumption (severe hangover-like symptoms), which is
avoided by abstaining from alcohol. Although disulfiram con-
tains a thiocarbamate group that can potentially interacts with
multiple targets, studies have concluded that the drug has an
acceptable side effect profile for long term treatments of alco-
holism at the daily doses of 250–500 mg (14, 15). Because dis-
ulfiram has been identified as the first submicromolar inhibitor
of G. lamblia CK as well as a compound that kills Giardia tro-
phozoites, we have undertaken structural and mass spectrom-
etry studies to characterize the enzyme/inhibitor complex.
A previous potential antigiardiasis drug search using com-
pounds known to bind to zinc finger proteins discovered that
disulfiram was effective against Giardiasis in adult mice but did
not identify the molecular target (16). Through the above two
high throughput screening assays, we have independently iden-
tified disulfiram as a Giardia-cidal compound that acts through
the essential G. lamblia CK enzyme (10, 11). We also validated
the result in vivo using an improved adult mouse model that
enables quantitative determination of the reduced trophozoite
load by coupling the drug treatment to in vitro proliferation
assay of axenic G. lamblia GS cultures. Here, we report the
results of the structural and in vivo studies.
EXPERIMENTAL PROCEDURES
Protein Preparation, Crystallization, and Structure Determi-
nation—Pure G. lamblia CK was produced as described previ-
ously (8). The purified protein was concentrated to 30 mg/ml in
solution containing 50 mMTris-HCl, pH 8.0, 0.1 MNaCl, 5 mM
MgCl
2
,and1mMDTT (dithiothreitol) and stored in aliquots at
⫺80 °C.
Crystals of G. lamblia CK were grown at room temperature
by the vapor diffusion methods in hanging drops. The reservoir
solutions contained 0.4 Mammonium citrate dibasic, pH 5.0,
and 21% PEG 3350. The hanging drops consisted of 1:1 protein
and reservoir solutions. This condition yielded the structure of
G. lamblia CK with bound citric acid (9). 100 mMdisulfiram
(Sigma-Aldrich) dissolved in dimethyl sulfoxide was diluted in
mother liquor to a final concentration of 2 mM. Crystals were
soaked in this drug solution for 16 h and then transferred to
mother liquor containing 20% glycerol and flash-cooled in liq-
uid nitrogen for x-ray diffraction data acquisition.
X-ray diffraction data were collected at the GM/CA-CAT
synchrotron beamline 23ID at the Advanced Photon Source in
Argonne National Laboratory (Argonne, IL). The beamline was
equipped with the MARmosaic 300 CCD detector (Marre-
search GmbH) controlled with the JBluIce user interface
program.
Diffraction data were integrated with the XDS program (17)
and scaled with the AIMLESS program, the successor to
SCALA (18), as implemented in CCP4 (19). The initial struc-
ture was determined by Furrier synthesis using the coordinates
and calculated phases of the previously determined protein
structure (9). Rigid body minimizations and refinements were
carried out with the Phenix program (20). A fragment of the
disulfiram crystal structure coordinates (21) was added to a
modified cysteine residue seen in the electron density map. The
Coot graphics program (22) was used for model building and
visual inspection of the structures. Structure figures were gen-
erated with Raster3D linked to Molscript (23, 24) and PyMOL
(DeLano Scientific).
Mass Spectrometry—Mass spectrometry (MS) analysis cou-
pled with trypsin digestion was performed to identify residues
modified by disulfiram in solution. The G. lamblia CK was
mixed with disulfiram at 1:150 molar ratio in 50 mMNH
4
HCO
3
buffer (pH 7.7). The sample was treated with trypsin (Sigma-
Aldrich) at 1:25 (w/w) enzyme/substrate ratio for 14 h at 37 °C.
Following proteolysis, the protein sample was treated with 50
mMiodoacetamide for 60 min at room temperature. Aliquots of
the sample were then mixed with equal volumes of 10 mg/ml
␣
-cyano-4-hydroxycinnamic acid dissolved in 50% acetonitrile
and 0.1% trifluoroacetic acid. The MS analysis was performed
with an AB4700 Proteomics Analyzer (Applied Biosystems,
Framingham, MA).
The MS mode acquisitions consisted of 1,000 laser shots
averaged over 20 sample positions. For MS/MS mode acquisi-
tions, 3,000 laser shots were averaged over 30 sample positions
for post source decay fragments. Combined acquisition of MS
and MS/MS data were automatically controlled with the 4000
Series Explorer software (version 3.0). Data analysis was per-
formed with the GPS Explorer software utilizing Mascot
(MatrixScience, version 2.0, London, UK) as the search engine.
During the search, the mass tolerance was 0.08 Da for the pre-
cursor ions and 0.2 Da for the fragment ions.
G. lamblia Cultures—Trophozoites of G. lamblia Assem-
blage A isolate WB, Assemblage B isolate GS/H7 (25), and
Assemblage A metronidazole-resistant isolate 713M3 (26) were
grown anaerobically in borosilicate glass screw-cap culture
tubes (Thermo Fisher) at pH 7.0 in modified TYI-S-33 medium
G. lamblia Carbamate Kinase Inactivation by Disulfiram
APRIL 11, 2014•VOLUME 289• NUMBER 15 JOURNAL OF BIOLOGICAL CHEMISTRY 10503
by guest on September 18, 2015http://www.jbc.org/Downloaded from
(11, 27). The medium was supplemented with 10% heat-inacti-
vated bovine serum (Sigma-Aldrich) and 0.05% bovine bile
(Sigma-Aldrich). To attain low-oxygen-tension conditions, the
tubes were filled to 85 to 90% of their total volume capacity and
incubated without shaking at 37 °C. Subcultures (2 ⫻10
5
tro-
phozoites per tube) were made three times a week. For enumer-
ation and for mouse infection, the trophozoites were detached
from the wall of the tube by chilling the cultures on ice for
20 min.
Minimum Lethal Concentration (MLC) Determination—
MLC values were determined by incubating G. lamblia tropho-
zoites with the drugs in 96-well culture plates (Corning, Inc.),
followed by transferring the trophozoites into 8-ml tubes
(Fisher Scientific) for proliferation. Assays were performed in
duplicates. Dry compounds were dissolved in dimethyl sulfox-
ide (Sigma) at stock concentration of 10 mMand then diluted
1:100 fold in growth medium to a final compound concentra-
tion of 100
M. 100-
l aliquots of compound solutions were
prepared by 2-fold serial dilutions (12 concentrations) in
medium. This was followed by the addition of 10
lofG. lam-
blia culture containing 10,000 organisms into medium with the
compound to be tested. All growth medium used contained
reduced cysteine concentration (2.9 mMcysteine) because the
usual concentration (11.6 mM) blunts the activities of thiuram
compounds (16). Metronidazole mixed with the same medium
served as positive control, and the medium with dimethyl sulf-
oxide alone served as negative control. Plates were incubated
under anaerobic condition in sealed bags (Becton Dickinson
and Company) at 37 °C for 3 days and surveyed visually under
the microscope to check trophozoite survival, mobility, and
attachment. The plates were chilled on ice for 30 min, and the
entire contents of each of 4 wells in the growth/death transition
were transferred into the 8-ml tubes containing growth
medium and no drug. The tubes were incubated under anaer-
obic condition for 3 days at 37 °C and checked under micro-
scope. The MLC value was attributed to the lowest concentra-
tion without any live organisms.
ATP Content Assay—The ATP content assay was described
in detail previously (11). The assays were performed in dupli-
cates. Tubes containing G. lamblia trophozoites were placed
on ice for 30 min. 100-
l aliquots of each tube were transferred
into 96-well black clear-bottom assay plate followed by the
addition of 70
l/well of the ATPLite reagent (PerkinElmer Life
Science) to initiate a one-step cell lysis and detection of the ATP
level. The luminescence signals were measured on an EnSpire
2300 plate reader (PerkinElmer Life Science).
Giardiasis Animal Model and Drug Treatment—Fifteen
adult (4-week-old) C57BL/6J female mice (The Jackson Labo-
ratory, Bar Harbor, ME) were infected with G. lamblia GS/H7
trophozoites, the only human Giardia isolate known to infect
adult mice (28). The trophozoites (500,000 suspended in 200
l
of TYI-S-33 medium) were administered by oral gavage. On
days 3 through 6 after infection, five mice were treated once
daily with 5 mg of disulfiram (or dithiodimethylthiuram),
whereas the remaining untreated five mice served as control. A
comparison experiment with metronidazole was carried out
with 15 mice, treating 10 animals and leaving five animals
untreated for control. The drugs were suspended in 200
lof
corn oil and were administered by oral gavage. All mice were
euthanized on day 7. Two inches of the upper small intestine
were dissected and washed with 2 ml of medium supplemented
with antibiotics (piperacillin 1 mg/ml, moxalactam 1 mg/ml).
The harvested small intestines were opened longitudinally and
minced in a Petri dish containing 10 ml of ice-chilled medium.
Plates were placed on ice for 30 min to allow the trophozoites to
detach from the intestine and were surveyed under microscope
for estimation of trophozoites population by the plate survey
method described previously (16). The trophozoites were enu-
merated in several random fields at all depths not obscured by
intestines at a magnification of 20⫻with an Axiovert 40 C
microscope (Zeiss). The University of Maryland College Park
The Institutional Animal Care and Use Committee approved
the animal studies.
Mouse Trophozoite Load Quantification by Proliferation of
Axenic Cultures—The entire contents of each Petri dish was
transferred into a 15-ml glass tube, the volume was adjusted to
14 ml by adding medium containing 1 mg/ml piperacillin and 1
mg/ml moxalactam, and the tube was vigorously vortexed to
separate the trophozoites from the intestine debris. For the fol-
lowing 2 h, tubes were kept at 37 °C to allow trophozoite attach-
ment to the wall of the glass tube, after which the medium
containing the intestine debris was decanted and replaced by
fresh medium. Tubes were kept at 37 °C for another 15 min
followed by a second medium replacement, but this time the
medium was supplemented with 1 mg/ml piperacillin, 1 mg/ml
moxalactam, 5
g/ml amoxicillin/clavulanic acid, and 10
g/ml
nalidixic acid. The tubes were kept at 37 °C for 6 days for pro-
liferation. Trophozoites growth was determined once daily by
direct cell counting using a Beckman Coulter Z1 counter (Beck-
man Coulter, Inc.) and by monitoring the ATP content using
the ATPLite reagent.
The proliferation measurements were used to calculate
growth curves and trophozoite load of drug-treated mice com-
pared with untreated mice. Trophozoite load was quantified by
calculating the initial population from the growth curves, fit-
ting the luminescence or the trophozoite counting data to the
Malthusian model of exponential growth N
t
⫽N
0
e
rt
, where N
t
is the signal of the population at time t,ris the growth rate, and
N
0
is the value we seek to determine, the signal of the initial
population at the end of treatment. The % load was determined
relative to N
0
of untreated mouse samples. The two methods of
measurement yielded consistent results.
RESULTS AND DISCUSSION
Disulfiram Impairs the Viability of Metronidazole-suscepti-
ble and Metronidazole-resistant G. lamblia Isolates—The bio-
luminescence assay that measures the cellular ATP content is a
cell viability assay amenable to high throughput screening of
compound libraries. However, this assay does not discriminate
between cell killing and metabolically inactive trophozoites. In
contrast, MLC assays, first incubating the trophozoites with the
drug and then transferring the culture to drug-free medium for
proliferation, determine the drug concentration when no sur-
viving organisms remain. MLC values were determined for dis-
ulfiram and thiram using three G. lamblia isolates that infect
human: the metronidazole-susceptible WB and GS/H7, and the
G. lamblia Carbamate Kinase Inactivation by Disulfiram
10504 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 289•NUMBER 15• APRIL 11, 2014
by guest on September 18, 2015http://www.jbc.org/Downloaded from
metronidazole-resistant 713M3. The data are summarized in
Table 1 together with the G. lamblia CK IC
50
values, confirm-
ing the potency of disulfiram and thiram against both metron-
idazole-susceptible and metronidazole-resistant isolates. Hence,
these compounds and possibly future G. lamblia CK inhibitors
would not be subject to the metronidazole resistance mecha-
nism, supporting the hypothesis that G. lamblia CK is a novel
target for the development of new antigiardiasis drugs.
Disulfiram Interferes with Nucleotide Binding by Covalently
Modifying Cys-242—Recently, we have reported the crystal
structures of G. lamblia CK in complexes with the non-hydro-
lysable ATP analog AMP-PNP and with citric acid, which mim-
ics carbamoyl phosphate binding (8, 9). These structures reveal
how the phosphoryl group transfer occurs and provide the
structural basis for understanding how disulfiram inhibits en-
zyme activity. Briefly, elongated active site architecture enables
the accommodation of the two substrates, ADP and carbamoyl
phosphate, in line for direct transfer of the phosphoryl group
to produce ATP and carbamate. Superposition of the two
enzyme-ligand structures shows that the binding site of the
AMP-PNP
␥
-phosphoryl group overlaps with a carboxylate
group of the citrate, the surrogate of the carbamoyl phosphate
phosphoryl group. An auxiliary domain encompassing amino
acid residues 134–164 adopts a closed conformation when the
citric acid binds (Fig. 1A) but adopts an open or disordered
conformation when AMP-PNP binds and displaces the citric
acid (Fig. 1B). Another region of enzyme flexibility, located at
the nucleotide binding site, comprises a loop carrying Tyr-245
(amino acid residues 244–250). The loop adjusts upon nucleo-
tide binding such that the aromatic ring of the tyrosine stacks
above the adenine group (Fig. 1B). In contrast, in the absence of
nucleotide the loop either lacks well defined conformation, or it
adopts an open conformation that places Tyr-245 more
remotely from the adenine site (Fig. 1A).
The crystal structure of G. lamblia CK soaked with disul-
firam was determined to the resolution limit of 2.6 Å. Data
collection and refinement statistics are summarized in Table 2.
FIGURE 1. Difference Fourier electron density maps associated with ligands bound in the active site of G. lamblia CK: the coefficients F
o
ⴚF
c
and
calculated phases omitting the ligands from the calculation are used. A, citric acid bound in the carbamoyl phosphate binding site. B, AMP-PNP bound in
the ADP/ATP binding site. C, the disulfiram thiocarbanoylation product modifying Cys-242 adjacent to the ADP/ATP binding site and the citric acid in the
carbamoyl phosphate binding site. Atomic colors are as follows: gray, carbon; red, oxygen; blue, nitrogen; and green, phosphor. Active site regions that undergo
conformational transitions are highlighted in yellow, and their amino acid residue ranges are labeled. The 244 –250 loop is disordered in C.Aand Bare adopted
from Ref. 9.
TABLE 1
Inhibition of G. lamblia CK activity and growth of G. lamblia isolates by
drugs
Compound CK IC
50
a
G. lamblia MLC
b
WB GS/H7 713M3
M
M
Disulfiram 0.64 3.1 1.5 0.75
Thiram 0.15 0.75 0.8 0.38
Metronidazole Not inhibited 3.1 3.1 50
a
Values were reported previously (10).
b
Values for all three compounds were determined with medium containing 2.9
mMcysteine rather than the usual 11.6 mMincluded in optimal laboratory me-
dium to prevent masking of the activity of thiuram compounds.
TABLE 2
X-ray data collection and structure refinement statistics
r.m.s.d., root mean square deviation.
Data collection
Space group P2
1
Cell dimension a⫽70.3, b⫽98.2, and c⫽102.7 Å,

⫽107.6°
Wavelength (Å) 1.0332
Resolution (Å) 2.6
No. of observed reflections 137,419
Completeness (%)
a
99.1 (99.7)
No. of unique reflections 40,662
R
mergeb
0.082 (0.314)
具I/
(I)典10.8 (3.8)
Redundancy 3.4 (3.4)
Refinement
No. of reflections used 40,453
No. of protein atoms 9,168
No. of ligand atoms 76
No. of water atoms 265
R
crystc
0.199 (0.232)
R
freed
0.265 (0.320)
r.m.s.d. from ideal geometry
Bond length (Å) 0.009
Bond angle 1.2°
Average B factor (Å
2
)
Protein 43
Ligand 65
Water 36
Ramachandran plot (%)
e
86.9, 13.1, 0.0, 0.0
a
The values in parentheses are for the highest resolution shell, 2.71–2.60 Å.
b
R
merge
⫽⌺
hkl
[(⌺
j
兩I
j
⫺具I典兩)/⌺
j
兩I
j
兩].
c
R
cryst
⫽⌺
hkl
储F
o
兩⫺兩F
c
储/⌺
hkl
兩F
o
兩, where F
o
and F
c
are the observed and calculated
structure factors, respectively.
d
R
free
is computed with 2,013 randomly selected reflections omitted from the
refinement.
e
Ramachandran plot categories are most favored, allowed, generously allowed,
and disallowed (32, 33).
G. lamblia Carbamate Kinase Inactivation by Disulfiram
APRIL 11, 2014•VOLUME 289• NUMBER 15 JOURNAL OF BIOLOGICAL CHEMISTRY 10505
by guest on September 18, 2015http://www.jbc.org/Downloaded from
The disulfiram reacted with a single G. lamblia CK cysteine
residue, Cys-242, to form a covalent product seen in one of the
four subunits in the crystal asymmetric unit. The electron den-
sity map is consistent with a thiodiethylcarbamoyl adduct,
which suggests the dithiocarbamoylation reaction depicted in
Fig. 2, except that there is no electron density to account for the
sulfur of thioketone (Fig. 1C). The degradation of the dithiodi-
ethylcarbamoyl adduct into thiodiethylcarbamoyl adduct may
be attributed to radiation damage during x-ray data collection.
X-ray radiation, in particular high-energy synchrotron radia-
tion, has long been known to cause disulfide bond breakage and
decarboxylation of acidic side chains in protein crystals, even at
cryogenic temperature (29, 30). A modified Cys-242 contains a
disulfide bond and a thioketone, both prone to radiation dam-
age. Although missing in the electron density map, the position
of the thioketone sulfur is defined by the stereochemistry, as
depicted in the model shown in Fig. 3. Note that the electron
density map shows thiocarbamoylation of only one of four of
the G. lamblia CK Cys-242 in the crystal asymmetric unit. The
data cannot distinguish between cleavage of the thiocarbamate
group by the x-ray radiation and an unmodified thiol group.
Interestingly, in addition to the covalently linked thiocarbam-
oyl adduct, the electron density map contained two extensive
peaks ⬃11 Å away from two free Cys-242 residues, which were
modeled as free dithiodiethylcarbamate molecules that may
account for degradation of other modified cysteine residues. As
described below, the mass spectrometry studies confirmed that
disulfiram selectively modifies the majority of the Cys-242 thiol
groups.
Modification of Cys-242 interferes with the conformational
transition that accompanies nucleotide binding, whereas the
citric acid binding remains unaltered and the auxiliary domain
is in the closed conformation. The affected region is shown in
Fig. 3, which depicts the superposition of the 244–250 loop
environment in the absence and presence of nucleotide (gray
and magenta, respectively), and in the presence of the thiocar-
bamoylated Cys-242 (green). In the absence of AMP-PNP, the
Cys-242 thiol group interacts with Lys-252 amino group (3.5
Å), which in turn, forms a salt bridge with Glu-250. Upon nucle-
otide binding the loop undergoes adjustments that bring Cys-
242 closer to Glu-250 (3.0 Å). Thus, the Cys-242-Glu-250-Lys-
252 triad plays crucial role in defining the loop conformational
adjustments and the correlated 3.8 Å shift of Tyr-245 (Fig. 3).
Dithiocarbamoylation of Cys-242 leads to steric clashes with
both Glu-250 and Lys-252 side chains. To avoid these clashes,
Lys-252 backbone flips concomitantly with disordering of the
244–250 loop. Moreover, the position of the modified Cys-242
overlaps with Tyr-245 side chain at the ATP-PNP bound state.
Thus, the crystal structure is consistent with inhibition mech-
anism due to irreversible modification of Cys-242 that prevents
nucleotide binding.
The mass spectrometry analysis confirmed that disulfiram
modified the Cys-242 thiol, selectively. The enzyme contains
eight cysteine residues, three of which are exposed to solvent
and only Cys-242 is located in the vicinity of the active site.
MALDI-TOF analysis showed that mixing G. lamblia CK with
150-fold molar excess of disulfiram yielded a single molecular
peak shifted by 151 mass units relative to the untreated protein
Et
N
Et
S
S
Et
N
Et
S
S
CK-Cys242-SH +Et
N
Et
S
SH
Et
N
Et
S
S+
CK-Cys242-S
Mcalc = 33,936 Mcalc = 34,083
FIGURE 2. The proposed thiocarbamoylation reactions leading to the modification of Cys-242.
FIGURE 3. Stereoscopic representation of the environment of Cys-242 superposed in three conformational states. The AMP-PNP bound structure
(magenta), the citric acid bound structure (gray), and the thiocarbamoylated structure, which exhibits disordered 244 –250 loop (green). The 3.8 Å shift
of the Tyr-245 side chain in response to AMP-PNP binding is highlighted by the magenta dashed line. Steric clashes contacts between the modified
Cys-242 and other side chains are indicated by black dashed lines. The transparent protein surface corresponds to the citric acid bound structure.
G. lamblia Carbamate Kinase Inactivation by Disulfiram
10506 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 289•NUMBER 15• APRIL 11, 2014
by guest on September 18, 2015http://www.jbc.org/Downloaded from
peak (Fig. 4). This mass difference is consistent with a single
cysteine modification by a dithiodiethylcarbamate group that
should increase the protein mass by 148 units. Trypsin diges-
tion followed by MS/MS mass spectrometry analysis identified
the Cys-242-containing peptide as the sole dithiocarbamoy-
lated peptide. Thus, both the crystal structure and mass spec-
trometry in solution suggest that Cys-242 thiol group is the
most reactive of all G. lamblia CK thiol groups. The electro-
static microenvironment of the Cys-242, including the Glu-
250-Lys-252 pair, may enhance the reactivity of this thiol
group.
Adult Mouse Model and Culture Axenization—The human
G. lamblia GS/H7 is the only known isolate that also infects
adult mice. Other human isolates infect neonatal mice, which
are more delicate and prone to accidental injuries during oral
gavage. We therefore elected to evaluate the efficacy of antigiar-
diasis compounds in adult mice infected with G. lamblia
GS/H7 trophozoites. Of the sources of C57BL/6J mice tested,
only mice from The Jackson Laboratory were consistently
infected, and after a week (corresponding to the entire period of
the in vivo infection and drug treatment experiments), the tro-
phozoite population could be enumerated using a plate survey
method (16). However, following drug treatment the remaining
trophozoites were too few for reliable estimation. To quantify
determine parasite survival, the in vivo studies were followed by
in vitro proliferation assays.
Because the gut flora is populated by anaerobic bacterial spe-
cies that thrive on the medium required for in vitro optimal
anaerobic trophozoite growth, obtaining axenic G. lamblia cul-
tures reproducibly is crucial for the proliferation experiments.
Recently, axenization of Giardia cultures harvested from the
intestines of suckling mice has been reported (31). The pub-
lished method consists of harvesting the intestine of the treated
mice and transferring the luminal washes to the peritoneal cav-
ity of adult mice for 24 h and then euthanizing these mice to
obtain the intraperitoneal content used in the in vitro prolifer-
ation assay. We have developed a simpler axenization method
that avoids the doubling of sacrificed mice by identifying an
antibiotic mixture that does not interfere with G. lamblia
growth. This mixture includes the standard antibiotics used to
avoid bacterial contamination in in vitro studies (piperacillin
and moxalactam) as well as the combination drug amoxicillin/
clavulanic acid and the quinolone nalidixic acid. We have used
this mouse model and axenization protocol extensively for two
years and rarely encountered bacterial contamination. The
requirement for clavulanic acid, a class A

-lactamase inhibitor,
suggests that the commensal bacteria in the gut of The Jackson
Laboratory C57BL/6J mice have acquired this enzyme, not a
FIGURE 4. MALDI-TOF mass spectroscopy of G. lamblia CK before (left panel) and after (right panel) addition of disulfiram. The small discrepancies
between the calculated molecular masses (see Fig. 2) and those measured by mass spectrometry are well within the experimental error. The mass
difference of 151 units is consistent with a single cysteine modified by a dithiodiethylcarbamate group (calculated mass difference of 148 units).
TABLE 3
In vivo mice studies of trophozoite load following disulfiram, thiram, and metronidazole treatments (once daily dose of 5 mg/day for 4 days)
Drug
Trophozoite load, pvalue
a
Visual plate survey
Proliferation assay
b
Trophozoite count ATP content
%
Disulfiram 1.2 (0.004) 0.45 ⫾1.01 (0.0032) 0.55 ⫾0.97 (0.00037)
Thiram 0.4 (0.004) 0.⫾0. (0.0032) 0.04 ⫾0.08 (0.00038)
Metronidazole 1.9 (0.003) 0.01 ⫾0.02 (0.00023) 0.002 ⫾0.0052 (0.000006)
a
Trophozoite load is calculated relative to the untreated mice. The pvalue is calculated based on a t-test relative to the trophozoite load of the control untreated mice.
b
The trophozoite load calculated from the growth curves at time 0, i.e. the end of animal treatment, either by cell counting or ATP luminescence (see “Experimental Proce-
dures” for detail).
G. lamblia Carbamate Kinase Inactivation by Disulfiram
APRIL 11, 2014•VOLUME 289• NUMBER 15 JOURNAL OF BIOLOGICAL CHEMISTRY 10507
by guest on September 18, 2015http://www.jbc.org/Downloaded from
surprising finding considering the wide use of

-lactam antibi-
otics. In case of emergence of new antibiotic resistance in com-
mensal bacteria, alternative inhibitors may be identified that
enable axenization of the in vitro G. lamblia cultures.
Disulfiram and Thiram Cure Giardiasis in Adult Mice—Pre-
viously, Nash and Rice (16) reported cure or reduced Giardia
trophozoite load in adult mice treated daily with 25-mg disul-
firam doses for 4 days. We reduced the doses to 5 mg/day and
observed comparable drug efficacies of disulfiram, thiram, and
metronidazole, in contrast to the untreated mice that remained
infected (Table 3). For all three drugs, trophozoites could not be
detected in most mice, and only a few trophozoites could be
detected in 20–40% of the treated mice. Moreover, both pro-
liferation quantification methods (direct trophozoite count and
ATP content determination) yielded similar results. Thus,
these experiments demonstrate that disulfiram, an FDA-ap-
proved drug, is a potential antigiardiasis therapeutic agent.
(Thiram is not an approved drug.) Further dose and schedule
studies in animals followed by human clinical studies will be
necessary to validate the effectiveness of disulfiram treatment.
Future Prospects—The arginine dihydrolase pathway en-
zyme, CK, is essential for G. lamblia trophozoite survival, and
the enzyme is irreversibly inactivated by disulfiram and its ana-
log, thiram. These compounds kill metronidazole-susceptible
and metronidazole-resistant G. lamblia trophozoites in vitro
and exhibit efficacy in vivo in mouse model. CK has not been
exploited as antigiardial drug target; thus, future inhibitors will
not be subject to drug resistance mechanisms against current
standard care drugs. Unlike treatment of chronic alcoholism,
which requires long term use of the drug, antigiardiasis therapy
is expected to span only a short period, reducing the risk of side
effects. Nevertheless, compound modification that increases
selectivity toward G. lamblia CK would avoid undesirable side
effects due to off target inhibition of the human acetaldehyde
dehydrogenase. Improved selectivity and potency may be
achieved if one of the ethyl substituents is replaced by a
larger group that occupies more of the adenine binding site.
Moreover, G. lamblia trophozoites attach to the wall of the
intestine but do not invade the cells. Therefore, antigiardia-
sis drugs need not cross the intestinal epithelial burier.
Chemical modifications of disulfiram that minimize absorp-
tion through the gut but still enable transport into the tro-
phozoites will further reduce the risks of off target inhibition
and side effects.
Because ultimately resistance develops against any drug, the
use of combination therapies reduces the rate of emerging
resistance. This strategy may be employed in combating the
spread of drug resistance in G. lamblia. Combination therapy
using one of the current standard care drugs with disulfiram
would also facilitate dose reduction and therefore the undesir-
able side effects of drugs such as metronidazole.
Finally, the adult mouse model offers convenience and
reduces accidental injuries during oral gavage. The caveat is
that no G. lamblia strain other than GS has been reported to
infect adult mice in the laboratory whereas so far, no G. lamblia
GS isolate exhibiting metronidazole resistance has been
reported. Nonetheless, the new G. lamblia axenization method
may be applied also to the neonatal mouse model, simplifying
the currently used procedure and reducing the number of scar-
ified animals (31) while enabling in vivo drug trials against met-
ronidazole-resistant G. lamblia isolates.
Acknowledgements—We thank the staff at GM/CA-CAT of the
Advanced Photon Source and Dr. Chen Chen for assistance in the
x-ray data acquisition. We also thank Dr. Lars Eckmann and Dr.
Yukiko Miyamoto for providing the metronidazole-resistant G. lam-
blia isolate (716 M) and Dr. Theodore Nash for valuable advice about
the mouse model.
REFERENCES
1. Ankarklev, J., Jerlström-Hultqvist, J., Ringqvist, E., Troell, K., and Svärd,
S. G. (2010) Behind the smile: cell biology and disease mechanisms of
Giardia species. Nat. Rev. Microbiol. 8, 413–422
2. Cañete, R., Escobedo, A. A., González, M. E., Almirall, P., and Cantelar, N.
(2006) A randomized, controlled, open-label trial of a single day of
mebendazole versus a single dose of tinidazole in the treatment of giardia-
sis in children. Curr. Med. Res. Opin. 22, 2131–2136
3. Escobedo, A. A., Alvarez, G., González, M. E., Almirall, P., Cañete, R.,
Cimerman, S., Ruiz, A., and Pérez, R. (2008) The treatment of giardiasis in
children: single-dose tinidazole compared with 3 days of nitazoxanide.
Ann. Trop. Med. Parasitol. 102, 199–207
4. Lemée, V., Zaharia, I., Nevez, G., Rabodonirina, M., Brasseur, P., Ballet,
J. J., and Favennec, L. (2000) Metronidazole and albendazole susceptibility
of 11 clinical isolates of Giardia duodenalis from France. J. Antimicrob.
Chemother. 46, 819–821
5. Upcroft, P., and Upcroft, J. A. (2001) Drug targets and mechanisms of
resistance in the anaerobic protozoa. Clin. Microbiol. Rev. 14, 150–164
6. Wensaas, K. A., Langeland, N., and Rortveit, G. (2009) Prevalence of re-
curring symptoms after infection with Giardia lamblia in a non-endemic
area. Scand. J. Prim. Health Care 27, 12–17
7. Schofield, P. J., Costello, M., Edwards, M. R., and O’Sullivan, W. J. (1990)
The arginine dihydrolase pathway is present in Giardia intestinalis.Int. J.
Parasitol. 20, 697–699
8. Galkin, A., Kulakova, L., Wu, R., Nash, T. E., Dunaway-Mariano, D., and
Herzberg, O. (2010) X-ray structure and characterization of carbamate
kinase from human parasite Giardia lamblia.Acta Crystallogr. Sect. F
Struct. Biol. Cryst. Commun. 66, 386–390
9. Lim, K., Kulakova, L., Galkin, A., and Herzberg, O. (2013) Crystal struc-
tures of carbamate kinase from Giardia lamblia bound with citric acid and
AMP-PNP. PLoS One 8, e64004
10. Chen, C. Z., Southall, N., Galkin, A., Lim, K., Marugan, J. J., Kulakova, L.,
Shinn, P., van Leer, D., Zheng, W., and Herzberg, O. (2012) A homogenous
luminescence assay reveals novel inhibitors for Giardia lamblia carba-
mate kinase. Curr. Chem. Genomics 6, 93–102
11. Chen, C. Z., Kulakova, L., Southall, N., Marugan, J. J., Galkin, A., Austin,
C. P., Herzberg, O., and Zheng, W. (2011) High throughput Giardia lam-
blia viability assay using bioluminescent ATP content measurements. An-
timicrob. Agents Chemother. 55, 667–675
12. Kitson, T. M. (1983) Mechanism of inactivation of sheep liver cytoplasmic
aldehyde dehydrogenase by disulfiram. Biochem. J. 213, 551–554
13. Lipsky, J. J., Shen, M. L., and Naylor, S. (2001) In vivo inhibition of aldehyde
dehydrogenase by disulfiram. Chem. Biol. Interact. 130, 93–102
14. Fuller, R. K., and Gordis, E. (2004) Does disulfiram have a role in alcohol-
ism treatment today? Addiction 99, 21–24
15. Malcolm, R., Olive, M. F., and Lechner, W. (2008) The safety of disulfiram
for the treatment of alcohol and cocaine dependence in randomized clin-
ical trials: guidance for clinical practice. Expert Opin. Drug Saf. 7,
459–472
16. Nash, T., and Rice, W. G. (1998) Efficacies of zinc-finger-active
drugs against Giardia lamblia.Antimicrob. Agents Chemother. 42,
1488–1492
17. Kabsch, W. (2010) XDS. Acta Crystallogr. D Biol. Crystallogr. 66, 125–132
18. Evans, P. R. (2011) An introduction to data reduction: space-group deter-
G. lamblia Carbamate Kinase Inactivation by Disulfiram
10508 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 289•NUMBER 15• APRIL 11, 2014
by guest on September 18, 2015http://www.jbc.org/Downloaded from
mination, scaling and intensity statistics. Acta Crystallogr. D Biol. Crystal-
logr. 67, 282–292
19. Winn, M. D., Ballard, C. C., Cowtan, K. D., Dodson, E. J., Emsley, P.,
Evans, P. R., Keegan, R. M., Krissinel, E. B., Leslie, A. G., McCoy, A.,
McNicholas, S. J., Murshudov, G. N., Pannu, N. S., Potterton, E. A.,
Powell, H. R., Read, R. J., Vagin, A., and Wilson, K. S. (2011) Overview
of the CCP4 suite and current developments. Acta Crystallogr. D Biol.
Crystallogr. 67, 235–242
20. Adams, P. D., Grosse-Kunstleve, R. W., Hung, L. W., Ioerger, T. R.,
McCoy, A. J., Moriarty, N. W., Read, R. J., Sacchettini, J. C., Sauter, N. K.,
and Terwilliger, T. C. (2002) PHENIX: building new software for auto-
mated crystallographic structure determination. Acta Crystallogr. D Biol.
Crystallogr. 58, 1948–1954
21. Karle, I. L., Estlin, J. A., and Britts, K. (1967) The crystal and molecular
structure of tetraethylthiuram disulfide, C
10
N
2
H
20
S
4
.Acta Crystallogr. 22,
273–280
22. Emsley, P., Lohkamp, B., Scott, W. G., and Cowtan, K. (2010) Features
and Development of Coot. Acta Crystallogr. D. Biol. Crystallogr. 66,
486–501
23. Bacon, D. J., and Anderson, W. F. (1988) A fast algorithm for rendering
space-filling molecule pictures. J. Mol. Graph. 6, 219–220
24. Kraullis, P. J. (1991) A program to produce both detailed and schematic
plots of protein structures. J. Appl. Crystallogr. 24, 946–950
25. Nash, T. E., McCutchan, T., Keister, D., Dame, J. B., Conrad, J. D., and
Gillin, F. D. (1985) Restriction-endonuclease analysis of DNA from 15
Giardia isolates obtained from humans and animals. J. Infect. Dis. 152,
64–73
26. Townson, S. M., Laqua, H., Upcroft, P., Boreham, P. F., and Upcroft, J. A.
(1992) Induction of metronidazole and furazolidone resistance in Giardia.
Trans. R Soc. Trop. Med. Hyg. 86, 521–522
27. Keister, D. B. (1983) Axenic culture of Giardia lamblia in TYI-S-33
medium supplemented with bile. Trans. R Soc. Trop. Med. Hyg. 77,
487–488
28. Byrd, L. G., Conrad, J. T., and Nash, T. E. (1994) Giardia lamblia infec-
tions in adult mice. Infect Immun. 62, 3583–3585
29. Garman, E. F. (2010) Radiation damage in macromolecular crystallogra-
phy: what is it and why should we care? Acta Crystallogr. D Biol. Crystal-
logr. 66, 339–351
30. Weik, M., Ravelli, R. B., Kryger, G., McSweeney, S., Raves, M. L., Harel, M.,
Gros, P., Silman, I., Kroon, J., and Sussman, J. L. (2000) Specific chemical
and structural damage to proteins produced by synchrotron radiation.
Proc. Natl. Acad. Sci. U.S.A. 97, 623–628
31. Tejman-Yarden, N., Millman, M., Lauwaet, T., Davids, B. J., Gillin, F. D.,
Dunn, L., Upcroft, J. A., Miyamoto, Y., and Eckmann, L. (2011) Impaired
parasite attachment as fitness cost of metronidazole resistance in Giardia
lamblia.Antimicrob. Agents Chemother. 55, 4643–4651
32. Ramachandran, G. N., Ramakrishnan, C., and Sasisekharan, V. (1963)
Stereochemistry of polypeptide chain configurations. J. Mol. Biol. 7,
95–99
33. Laskowski, R. A., MacArthur, M. W., Moss, D. S., and Thornton, J. (1993)
PROCHECK: a program to check the stereochemical quality of protein
structures. J. Appl. Crystallogr. 26, 283–291
G. lamblia Carbamate Kinase Inactivation by Disulfiram
APRIL 11, 2014•VOLUME 289• NUMBER 15 JOURNAL OF BIOLOGICAL CHEMISTRY 10509
by guest on September 18, 2015http://www.jbc.org/Downloaded from
Turko and Osnat Herzberg
Catherine Z. Chen, Wei Zheng, Illarion V.
Andrey Galkin, Liudmila Kulakova, Kap Lim,
Carbamate Kinase by Disulfiramlamblia GiardiaStructural Basis for Inactivation of
Protein Structure and Folding:
doi: 10.1074/jbc.M114.553123 originally published online February 20, 2014
2014, 289:10502-10509.J. Biol. Chem.
10.1074/jbc.M114.553123Access the most updated version of this article at doi:
.JBC Affinity SitesFind articles, minireviews, Reflections and Classics on similar topics on the
Alerts:
When a correction for this article is posted• When this article is cited•
to choose from all of JBC's e-mail alertsClick here
http://www.jbc.org/content/289/15/10502.full.html#ref-list-1
This article cites 33 references, 11 of which can be accessed free at
by guest on September 18, 2015http://www.jbc.org/Downloaded from
