Mannitol bis-phosphate based inhibitors of fructose 1,6-bisphosphate aldolases

Article abstract of DOI:10.1021/ml200129s

Several 5-O-alkyl- and 5-C-alkyl-mannitol bis-phosphates were synthesized and comparatively assayed as inhibitors of fructose bis-phosphate aldolases (Fbas) from rabbit muscle (taken as surrogate model of the human enzyme) and from Trypanosoma brucei. A l

Full text of DOI:10.1021/ml200129s

LETTER
pubs.acs.org/acsmedchemlett
Mannitol Bis-phosphate Based Inhibitors of Fructose
1,6-Bisphosphate Aldolases
Charles-Gabin Mabiala-Bassiloua,† Guillaume Arthus-Cartier,‡ Vꢀeronique Hannaert,§ Hꢀelꢁene Thꢀerisod,†
Jurgen Sygusch,*,‡ and Michel Thꢀerisod*,†
†ECBB, ICMMO (UMR 8182), LabEx LERMIT, Universitꢀe Paris-Sud, UMR 8182, F-91405 Orsay, France
‡Biochimie, Universitꢀe de Montrꢀeal, CP 6128, Stn Centre-Ville Montrꢀeal, PQ H3C 3J7 Canada
§Research Unit for Tropical Diseases, de Duve Institute, TROP 74.39, Avenue Hippocrate 74, B-1200 Brussels, Belgium
S Supporting Information
b
ABSTRACT: Several 5-O-alkyl- and 5-C-alkyl-mannitol bis-
phosphates were synthesized and comparatively assayed as
inhibitors of fructose bis-phosphate aldolases (Fbas) from
rabbit muscle (taken as surrogate model of the human enzyme)
and from Trypanosoma brucei. A limited selectivity was found in
several instances. Crystallographic studies confirm that the 5-O-
methyl derivative binds competitively with substrate and the
5-O-methyl moiety penetrating deeper into a shallow hydro-
phobic pocket at the active site. This observation can lead to the
preparation of selective competitive or irreversible inhibitors of the parasite Fba.
KEYWORDS: Selective inhibitors, microbial enzymes, fructose bis-phosphate aldolase, glycolysis, Trypanosoma brucei
arasitic diseases caused by Trypanosomatidae (Chagas’ dis-
ease, sleeping sickness, leishmaniasis, ...) are members of
One competitive inhibitor of FBA is hexitol bis-phosphate,
which was originally chemically prepared by reduction of fruc-
tose bis-phosphate (FBP)10 as a mixture of diastereoisomers,
glucitol bis-phosphate (GBP) and mannitol bis-phosphate
(MBP). We previously reported the separate synthesis and
biochemical evaluation of these two products.11 MBP is signifi-
cantly more active (14-fold) as inhibitor of FBA than the glucitol
counterpart. This is in agreement with the structural study of
rabbit-muscle FBA crystals soaked with hexitol bis-phosphate
solutions, where only the manno constituent of the mixture was
bound in the active site.12 MBP interacts strongly with several
residues of the mammalian enzyme: the P1-phosphate13 makes
strong hydrogen bonds with Ser 271, Arg 303, Gly 302, and Gly
272 while the P6- phosphate moiety makes hydrogen bonds with
Ser 35, Ser 38, and Lys 107 and with the peptide backbone at Glu
34 and Ser 35. The three hydroxyl groups at C-2, C-3, and C-4
also interact with active site residues but, interestingly, not the
hydroxyl at C-5, where there seems to be a large empty pocket.
The closest residue is situated 5.6 Å from the C-5 hydroxyl. A
similar situation is found in Trypanosoma brucei FBA, where the
putative C-5 binding locus would have a still larger pocket.14 The
absence of active site residues interacting with the C-5 hydroxyl
of MBP or in the natural substrate fructose bis-phosphate is
consistent with other compounds that are substrates of FBA yet
differ in substituents at the C-5 hydroxyl (Figure 1).15À18 On the
basis of this observation, we reasoned that by adding bulky alkyl
P
“neglected diseases” that include tuberculosis and malaria.1 The
two human infective subspecies of Trypanosoma brucei, T. b.
gambiense and T. b. rhodesiense, cause 66 000 deaths every year in
many sub-Saharan countries, and another 50 million people are
at risk of contracting the disease.2 The social impact of the disease
is worsened by massive infection of cattle in these countries by
another subspecies, T. b. brucei. The parasites reside in the
bloodstream of infected patients, and glycolysis represents the
only metabolic process through which their ATP is synthesized.3
This pathway is considered as a promising therapeutic target in
the struggle against T. brucei infections.4 Recent reports have
shown that only a partial depletion of some glycolytic enzymes
leading to the reduction of the glycolytic flux to approximately
50% of the wild-type levels is sufficient to kill the parasites in
culture.5,6 In bloodstream-form T. brucei, fructose bis-phosphate
aldolase (FBA) has been genetically validated as a drug target by
RNA interference. Humans have three aldolase isozymes loca-
lized in different tissues: type A in muscle and erythrocytes, type
B in the liver, and type C in the brain. Therefore, drugs that
target aldolase will need to be specific for the T. brucei aldolase
versus the human aldolases. The overall percentage sequence
identity between each of the human aldolases and T. brucei
aldolase is 47%.7 The crystal structure of T. brucei FBA has been
determined and shows important similarities with the human
enzymes, making the design and synthesis of selective inhibitors
challenging.8 Despite only subtle differences between the FBAs
from these two organisms, synthesis of a few selective inhibitors
has been reported.9
Received: May 26, 2011
Accepted: September 2, 2011
Published: September 03, 2011
r
2011 American Chemical Society
804
dx.doi.org/10.1021/ml200129s ACS Med. Chem. Lett. 2011, 2, 804–808
|

ACS Medicinal Chemistry Letters
LETTER
Scheme 2. Synthesis of 5-C-Alkyl Derivatives of Mannitol
Bis-phosphatea
Figure 1. Substrates of FBA with different substituents at C-5 (or a
missing hydroxyl) compared to FBP: L-SBP, L-sorbose bis-phosphate;
D-XP, D-xylulose bis-phosphate; D-SHBP, D-sedoheptulose bis-phosphate;
5-d-FBP, 5-deoxy-fructose bis-phosphate.15À18
a Reagents and conditions: (a) pyridinium chlorochromate, CH2Cl2, rt,
3 h; (b) R-MgBr, Et2O, rt; (c) BnBr/NaH, CH2Cl2, rt; (d) trifluor-
oacetic acid/n-butanol 1:1, CH2Cl2, rt; (e) iPr2NP(OBn)2/imidazole/
t
triazole, AcCN, rt, 12 h, then BuOOH/H2O; (f) H2/Pd-C, NEt3,
EtOH, then Dowex-50 (H+); (g) H2/Pd-C, EtOH.
Table 1. Enzymatic Inhibition Data Measured in the Pre-
sence of Substrate/Inhibitors21
Figure 2. Hexitol bis-phosphate derived inhibitors of FBA.
IC50 (KM), μM
Scheme 1. Synthesis of 5-O-Alkyl Derivatives of Mannitol
Bis-phosphatea
compd
T. brucei
rabbit muscle
FBP (substrate)
10 ( 0.6
9.4 ( 0.6
2.0 ( 0.12
51.0 ( 3
13 ( 0.8
8.0 ( 0.5
7.5 ( 0.5
64.0 ( 4
MBP
1
2
3
4
5
6
7
117.0 ( 7
31.0 ( 2
157.0 ( 9
21.0 ( 1.3
244.0 ( 15
17.0 ( 1
130.0 ( 8
15.0 ( 1
56.0 ( 3.5
137.0 ( 8
As was reported previously, the two-step deprotection to the
final product is made necessary by a possible migration of
phosphate groups if they were debenzylated after the
hydroxyls.11
Similarly, the strategy leading to C-5 derivatives starting from
the key intermediate of Scheme 1 is given in Scheme 2.19 The
alkylation step involves the reaction of an organometallic reagent
on a carbonyl function that could result in a mixture of two
diastereoisomers. On the basis of NMR spectra, a single diaster-
eoisomer was obtained in all cases and was assigned the D-manno
configuration, in accordance with the FelkinÀAhn rule.
These compounds were tested by standard methods for their
inhibitory properties on FBA from rabbit muscle (taken as
surrogate model of human FBA, with which it has 98% identity)
and on FBA from Trypanosoma brucei (His-tagged recombinant
enzyme expressed in E. coli).20 The results are reported in
Table 1.
a Reagents and conditions: (a) BnBr/NaH/Bu4NI cat., THF, rt; (b)
AcOH/H2O 7:3, 40 °C, 3 h; (c) TrCl/dimethylaminopyridine, DMF,
25 °C, 18 h; (d) BnBr/NaH 1 equiv/Bu4NI cat., THF, rt; (e) R-Br/
NaH, CH2Cl2; (f) trifluoroacetic acid/n-butanol 1:1, CH2Cl2, rt; (g)
t
iPr2NP(OBn)2/imidazole/triazole, AcCN, rt, 12 h, then BuOOH/
H2O; (h) H2/PdÀC, NEt3, EtOH, then Dowex-50 (H+); (i) H2/Pd-
C, EtOH.
groups on O-5 or C-5 of MBP it could be possible to prepare
selective inhibitors of parasitic FBA unable to enter the active
site of the human enzyme (Figure 2). Alternatively, selective
irreversible inhibitors bearing an electrophilic group could be
conceived.
The first step in this strategy is presented hereby. It consists of
validating that derivatives of MBP bearing an alkyl group at O-5
retain their inhibitory capacity.19 The synthetic strategy of this
class of compounds is summarized in Scheme 1.
Compounds 1À7 were shown by double-reciprocal plots to
act as competitive inhibitors on T. brucei FBA, consistent with
results previously obtained on the parent compound MBP.
Competitive inhibition for compound 1 is corroborated by the
805
dx.doi.org/10.1021/ml200129s |ACS Med. Chem. Lett. 2011, 2, 804–808

ACS Medicinal Chemistry Letters
LETTER
Figure 3. Electron density showing compound 1 trapped in the active
site of rabbit muscle aldolase. Compound 1 could be fitted into the
electron density using two orientations which are symmetrical with
respect to a 2-fold axis of rotation centered on the C3ÀC4 bond in
compound 1. The electron density encompassing compound 1 was
calculated from a simulated annealled Fo À Fc omit map and contoured
at 2σ to reveal the electron density of the 5-O-methyl, which is at half-
occupancy in both orientations. The electron density corresponding to
the 5-O-methyl group of compound 1 is explicitly labeled (Met) for the
two orientations. Phosphate oxyanions are identified as Pi. Hydrogen
bonds formed between compound 1 and active site residues are depicted
as green dotted lines. Water molecules are shown as red spheres. View is
looking into the β-barrel from the carboxyl side of the β-strands.
crystallographic studies. Compared to 1, inhibitors 2À7 show a
limited selectivity for Tb-FBA with respect to mammalian-FBA.
A surrogate strategy was undertaken to probe the structural
basis for the enhanced discrimination of compound 1 between
the two FBA enzymes, and it was the following: The structure of
rabbit muscle aldolase in complex with compound 1 was first
determined and then modeled by superimposition onto the
structure of native T. brucei aldolase (PDB entry code 1F2J) to
gain structural insight for the affinity differences in compound 1
binding to the two enzymes.
The crystal structure determination of the enzymatic complex
formed by soaking rabbit muscle aldolase crystals in the presence
of compound 1 is described in the Supporting Information and
summarized in Table S1. The crystal structure of the resultant
complex formed with compound 1 is shown in Figure 3.
The electron density corresponding to compound 1 binding is
continuous and was readily interpreted as compound 1 binding
in the active site in two symmetric and fully overlapping
orientations, each of equal occupancy that corresponded to full
active site occupancy based on B-factor comparison with neigh-
boring residues. Compound 1 bound in the active site of rabbit
muscle aldolase, and its binding mode was similar with that of
MBP, as previously described12 and shown in Figure 4a.
Compound 1 penetrates slightly deeper into the active site at
the 5-O-methyl position compared to the binding by MBP;
however, it makes no additional contact with active site residues.
Superposition of the rabbit muscle aldolase structure bound
with compound 1 onto the crystal structure of T. brucei aldolase
(rmsd 0.603 Å) revealed that for one binding mode, shown in
Figure 4b, the 5-O-methyl group of compound 1 would make an
additional hydrophobic (van der Waals) contact with the methyl
group of active site residue Ala-312.
Figure 4. Superimpositions of the crystal structure of rabbit muscle
aldolase and T. brucei aldolase showing the binding interactions of
compound 1. (a) Comparison of ligand binding in the active site of
rabbit muscle aldolase. Superposition of the rabbit muscle aldolase
structure bound with compound 1 (green) and that with bound MBP
(magenta) (PDB entry code 1ZAJ) yielded an rms value of 0.145 Å
consistent with identical subunit conformations. The root mean squared
value calculated for identical atoms in both ligands was 0.613 Å.
Hydrogen bonding patterns (dashes) made by compound 1 with active
site residues were conserved and similar when compared with those in
the MBP enzyme adduct. (b) Binding mode of compound 1 in the active
site of T. brucei aldolase (PDB entry code 1F2J) shown in green. Also
shown is subunit A of rabbit muscle aldolase (yellow) that was used to
superimpose rabbit muscle aldolase bound with compound 1 onto
subunit A of the T. brucei aldolase tetramer (rms value of 0.603 Å). The
hydrogen bonding patterns (green dashes) made by compound 1 in T.
brucei aldolase active site were conserved in the two aldolases. The only
significant difference was an additional hydrophobic interaction (3.0 Å)
made between Ala-312 in T. brucei aldolase and the 5-O-methyl (Met) in
compound 1 that is depicted by a dashed black line. Orientation is
similar to Figure 1.
In rabbit muscle aldolase, Gly-302 is the residue equivalent to
Ala-312 in T. brucei aldolase and does not make close contact
with the 5-O-methyl of compound 1. It represents the only
amino acid difference between the two active sites and is shown
in the alignment of amino acid sequences for rabbit muscle and T.
brucei aldolases (Table S2 of the Supporting Information). The
additional van der Waals interaction made between the 5-O-
methyl of compound 1 and the Ala-312 methyl group would thus
be responsible for the more potent inhibition in T. brucei aldolase
806
dx.doi.org/10.1021/ml200129s |ACS Med. Chem. Lett. 2011, 2, 804–808

ACS Medicinal Chemistry Letters
LETTER
by compound 1, shown in Table 1. The deeper penetration of the
aliphatic 5-O-methyl group in compound 1 into a region of the
active site lined predominantly by aliphatic carbons would not
be inconsistent with this interpretation. Furthermore, using the
same modeling procedure and superposing the liganded struc-
tures of rabbit muscle aldolase bound with FBP and MBP12
predicts identical interactions by FBP and MBP in T. brucei
aldolase, and neither ligand makes close contact with Ala-312
(data not shown).
the Natural Science and Engineering Research Council (Canada)
and Canadian Institutes for Health Research.
’ REFERENCES
(1) Feasey, N.; Wansbrough-Jones, M.; Mabey, D. C. W.; Solomon,
A. W. Neglected tropical diseases. Br. Med. Bull. 2010, 93, 179–200.
(2) Malvy, D.; Chappuis, F. Sleeping sickness. Clin. Microbiol. Infect.
2011, 17, 986–995.
The indistinguishable interactions by FBP and MBP in both
aldolases are consistent with their similar Km and IC50 values in
both aldolases, shown in Table 1, and corroborate the surrogate
modeling strategy. Modeling of compounds 2, 3, and 4 using
compound 1 as template would result in steric clashes with active
site residues in both aldolases, necessitating binding modes that
are different from those observed for FBP, MBP, and compound
1 and presumably of lower affinity, as shown in Table 1. Similarly,
modeling of the binding modes for compounds 5, 6, and 7 based
on that of compound 1 could not be accommodated due to steric
clashes in the active site of both aldolases and consistent with
active site interactions resulting in poorer affinity by these
compounds, as seen in Table 1.
(3) Michels, P. A. M.; Hannaert, V.; Bringaud, F. Metabolic aspects
of glycosomes in trypanosomatidae—new data and views. Parasitol.
Today 2000, 16, 482–489.
(4) Verlinde, C. L. M. J.; Hannaert, V.; Blonski, C.; Willson, M.;
Perie, J. J.; Fothergill-Gilmore, L. A.; Opperdoes, F. R.; Gelb, M. H.; Hol,
W. G. J.; Michels, P. A. M. Glycolysis as a target for the design of new
anti-trypanosome drugs. Drug Resist. Updates 2001, 4, 50–65.
(5) Albert, M.-A.; Haanstra, J. R.; Hannaert, V.; Van Roy, J.;
Opperdoes, F. R.; Bakker, B. M.; Michels, P. A. M. Experimental and
in Silico Analyses of Glycolytic Flux Control in Bloodstream Form
Trypanosoma brucei. J. Biol. Chem. 2005, 280, 28306–28315.
(6) Cꢀaceres, A. J.; Michels, P. A. M.; Hannaert, V. Genetic validation
of aldolase and glyceraldehyde-3-phosphate dehydrogenase as drug
targets in Trypanosoma brucei. Mol. Biochem. Parasitol. 2010,
169, 50–54.
The present selectivity, not exceeding a factor of 4, is obviously
too small for therapeutic utilization. One positive result, how-
ever, is that MBP derivatives bearing bulky groups at O-5 (C-5)
retain appreciable inhibitory properties. Modeling of the 5-O-
CH2F group in compound 1 suggests binding by such a
compound upon active site adjustment; however, the gain in
interaction energies rarely exceeds 1 kcal/mol.22 In this regard,
derivatives bearing short reactive functions at C-5 could be
introduced to act as irreversible selective inactivators. The
proposed studies would take advantage of docking calculations
of the compounds bound in the active site of the two enzymes
and followed up by structure determination of appropriate en-
zymeÀinhibitor complexes. Additional studies to be performed
would include the use of mutational analyses in concert with
structural and/or kinetic data to corroborate the role of active site
residues responsible for selectivity, as described previously.23
(7) Marchand, M.; Poliszczak, A.; Gibson, W. C.; Wierenga, R. K.;
Opperdoes, F. R.; Michels, P. A. Characterization of the genes for
fructose-bisphosphate aldolase in Trypanosoma brucei. Mol. Biochem.
Parasitol. 1988, 65–75.
(8) Chudzik, D. M.; Michels, P. A.; de Walque, S.; Hol, W. G.
Structures of type 2 peroxisomal targeting signals in two trypanosomatid
aldolases. J. Mol. Biol. 2000, 300, 697–707.
(9) Dax, C.; Duffieux, F.; Chabot, N.; Coincon, M.; Sygusch, J.;
Michels, P. A. M.; Blonski, C. Selective Irreversible Inhibition of
Fructose 1,6-Bisphosphate Aldolase from Trypanosoma brucei. J. Med.
Chem. 2006, 49, 1499–1502.
(10) Ginsburg, A.; Mehler, A. H. Specific Anion Binding to Fructose
Diphosphate Aldolase from Rabbit Muscle. Biochemistry 1966,
5, 2623–2634.
(11) Mabiala-Bassiloua, C.-G.; Zwolinska, M.; Therisod, H.; Sygusch,
J.; Therisod, M. Separate synthesis and evaluation of glucitol bis-
phosphate and mannitol bis-phosphate, as competitive inhibitors of
fructose bis-phosphate aldolases. Bioorg. Med. Chem. Lett. 2008, 18,
1735–1737.
(12) St-Jean, M.; Lafrance-Vanasse, J.; Liotard, B.; Sygusch, J. High
Resolution Reaction Intermediates of Rabbit Muscle Fructose-1,6-bi-
sphosphate Aldolase: Substrate Cleavage And Induced Fit. J. Biol. Chem.
2005, 280, 27262–27270.
(13) The two phosphate groups of MBP are indistinguishable
because of C2 symmetry in the molecule with respect to a symmetry
axis centred on the C3ÀC4 bond. Assignment of the P1- and P6-
phosphates in MBP was made on the basis of the distinguishable binding
loci of the P1- and P6-phosphate moieties of FBP in the active site
environment.10
(14) Chudzik, D. M.; Michels, P. A.; de Walque, S.; Hol, W. G. J.
Structures of type 2 peroxisomal targeting signals in two trypanosomatid
aldolases. J. Mol. Biol. 2000, 300, 697–707.
’ ASSOCIATED CONTENT
S
Supporting Information. Detailed syntheses and char-
b
acterizations of compounds 1À7, and crystallographic experi-
mental procedures. This material is available free of charge via the
Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: jurgen.sygusch@umontreal.ca or michel.therisod@
u-psud.fr.
Author Contributions
(15) Mehler, A. H.; Cusic, M. E., Jr. Aldolase Reaction with Sugar
Diphosphates. Science 1967, 155, 1101–1103.
C.-G.M-B.: Chemical syntheses, kinetic measurements; G.A-C.:
Crystallographic studies; V.H.: preparation of recombinant T.b.
FBA; H.T.: kinetic measurements, preparation of enzymes; J.S.:
Crystallographic studies, preparation of manuscript; M.T.: ki-
netic measurements, preparation of manuscript, and supervision
of C.-G.M.-B. Ph.D.
(16) Richards, O. C.; Rutter, W. J. Comparative Properties of Yeast
and Muscle Aldolase. J. Biol. Chem. 1961, 236, 3185–3192.
(17) Horecker, B. L.; Smyrniotis, P. Z.; Hiatt, H. H.; Marks, P. A.
Tetrose phosphate and the formation of sedoheptulose diphosphate. J.
Biol. Chem. 1955, 212, 827–836.
(18) Midelfort, C. F.; Gupta, R. K.; Rose, I. A. Fructose 1,6-bispho-
sphate: isomeric composition, kinetics, and substrate specificity for the
aldolases. Biochemistry 1976, 15, 2178–2185.
Funding Sources
This work was supported by a scholarship to C.-G.M.-B. from
Government of the Rꢀepublique du Congo. J.S. was supported by
(19) Analytical data available in Supporting Information.
807
dx.doi.org/10.1021/ml200129s |ACS Med. Chem. Lett. 2011, 2, 804–808

ACS Medicinal Chemistry Letters
LETTER
(20) Dax, C.; Duffieux, F.; Chabot, N.; Coinc-on, M.; Sygusch, J.;
Michels, P. A. M.; Blonski, C. Selective irreversible inhibition of fructose-
1,6-bisphosphate aldolase from Trypanosoma brucei. J. Med. Chem. 2006,
49, 1499–1502.
(21) Rabbit muscle aldolase was from Fluka. Enzymatic assays:
DHAP formed by cleavage of FBP by aldolase was estimated by
measuring spectrophotometrically (at 340 nm) the consumption of
NADH in a coupled system employing a 300-fold excess of glyceropho-
sphate dehydrogenase and triose phosphate isomerase in glycyl-glycine
buffer 0.1 M, pH 7.4. IC50 values were determined at a concentration in
substrate = KM
(22) Dimagno, S. G.; Sun, H. The strength of weak interactions:
aromatic fluorine in drug design. Curr. Top. Med. Chem. 2006,
6, 1473–82.
(23) Blonski, C.; de Moissac, D.; Periꢀe, J.; Sygusch, J. Inhibition of
rabbit muscle aldolase by phosphorylated aromatic compounds. Bio-
chem. J. 1997, 323, 71–77.
808
dx.doi.org/10.1021/ml200129s |ACS Med. Chem. Lett. 2011, 2, 804–808